Chemical Factors in Neural Growth, Degeneration and Repair

CHEMICAL FACTORS IN NEURAL GROWTH, DEGENERATION AND REPAIR r>^^

CHEMICAL FACTORS IN NEURAL GROWTH, DEGENERATION AND REPAIR r>^^

This Page Intentionally Left Blank

CHEMICAL FACTORS IN NEURAL GROWTH, DEGENERATION AND REPAIR r>^

C H R I S T O P H E R BELL Department of Physiology Faculty of Health Sciences Trinity College Dubliny Republic of Ireland

1996 ELSEVIER Amsterdam - Lausanne - New York - Oxford - Shannon - Tokyo

ELSEVIER SCIENCE B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam, The Netherlands

ISBN: 0 444 82529 0 © 1996 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science B.V., Copyright & Permissions Department, P.O. Box 521, 1000 AM Amsterdam, The Netherlands. Special regulations for readers in the U.S.A. - This publication has been registered with the Copyright Clearance Center Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the U.S.A. All other copyright questions, including photocopying outside of the U.S.A., should be referred to the copyright owner, Elsevier Science B.V., unless otherwise specified. No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. This book is printed on acid-free paper. Printed in The Netherlands.

List of Contributors S.E. Alves, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA F.J. Antonawich, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA E. Arenas, Laboratory of Molecular Neurobiology, Karolinska Institute, Stockholm S-17177, Sweden K.A. Bailey, Department of Genetics and Developmental Laboratory, Monash University, Clayton, Victoria 3168, AustraUa P.P. Bartlett, The Walter and Eliza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia S. Bieger, Department of Anatomy and Neurobiology, Faculty of Medicine, Sir Charles Tuooer Medical Building, HaUfax, Nova Scotia, Canada, B3H 4H7 G.W. Glazner, Department of Physiology and Department of Biochemistry and Molecular Biology, Colorado State University, Fort Colhns, CO 80523, USA M.E. Gotz, Clinical Neurochemistry, Department of Psychiatry, University of Wiirzburg, D-W97080, Wurzburg, Germany LA. Hendry, Neurobiology Research Group, Division of Neuroscience, The John Curtin School of Medical Research, The Australian National University, G.P.O. Box 334, Canberra, A.C.T. 2601, Australia L. lacovittti. Institute of Neuroscience, Hahnemann University, Broad and Vine Streets, Philadelphia, PA 19102, USA D.N. Ishii, Department of Physiology and Department of Biochemistry and Molecular Biology, Colorado State University, Fort ColHns, CO 80523, USA A. Jaworowski, Department of Medicine, Melbourne University, Royal Melbourne Hospital, Parkville, 3050, Australia U. Junghans, Molecular Neurobiology Laboratory, Department of Neurology, University of Dusseldorf, D-40225 Dusseldorf, Germany J. Kappler, Molecular Neurobiology Laboratory, Department of Neurology, University of Dusseldorf, D-40225 Dusseldorf, Germany T.J. Kilpatrick, The Walter and Eliza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia J. Kume, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA G. Kiinig, Clinical Neurochemistry, Department of Psychiatry, University of Wurzburg, D-W97080, Wurzburg, Germany T.S. Lee, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA S.J. Lee, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA D.J. Marsh, Department of Physiology and Department of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, CO 80523, USA R. Mayo, Department of Human Physiology and Centre for Neuroscience, GPO Box 2100, Adelaide 5001, South Australia, Australia

VI

A. Messina, Department of Medicine, Melbourne University, Royal Melbourne Hospital, Parkville, 3050, Australia H.W. Miiller, Molecular Neurobiology Laboratory, Department of Neurology, University of Dusseldorf, D-40225 Dusseldorf, Germany M. Murphy, The Walter and Eliza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia D.F. Newgreen, Embryology Laboratory, Murdoch Institute, University of Melbourne, Parkville, Victoria 3052, Australia V. Nurcombe, Department of Anatomy and Cell Biology, University of Melbourne, Parkville, Victoria 3052, Australia J.W. Olney, Department of Psychiatry, Washington University School of Medicine, St. Louis, MO 63110, USA S.F. Pu, Department of Physiology and Department of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, CO 80523, USA LJ. Richards, The Walter and EUza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia P.M. Richardson, Division of Neurosurgery, McGill University and Montreal General Hospital, 1650 Cedar Avenue, Montreal, Canada H3G 1A4 P. Riederer, Clinical Neurochemistry, Department of Psychiatry, University of Wurzburg, D-W97080, Wurzburg, Germany N. Rocamora, Department of Animal and Vegetal Cell Biology, Faculty of Biology, University of Barcelona, 08028-Barcelona, Spain R.A. Rush, Department of Human Physiology and Centre for Neuroscience, GPO Box 2100, Adelaide 5001, South Australia, Australia F.L. Strand, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA M.C. Subang, Division of Neurosurgery, McGill University and Montreal General Hospital, 1650 Cedar Avenue, Montreal, Canada H3G 1A4 P.S. Talman, The Walter and Eliza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia S.S. Tan, Embryology Laboratory, Department of Anatomy and Cell Biology, University of Melbourne, Parkville, Victoria 3052, Australia K. Unsicker, Institut fur Anatomic und Zellbiologie, Im Neuenheimer Feld 307, 69120 Heidelberg, Germany P.M. Whitington, University of New England, Department of Zoology, Armidale, NSW 2351, Australia K.A. WiUiams, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA T. Yamamori, National Institute for Basic Biology, 38 Nishigonaka, Myodaijicho, Okazaki, 444, Japan M.B.H. Youdim, Department of Pharmacology, Faculty of Medicine, Technion, Haifa, Israel C. Zettler, Department of Human Physiology and Centre for Neuroscience, GPO Box 2100, Adelaide 5001, South Australia, Australia H.-X. Zhuang, Department of Physiology and Department of Biochemistry and Molecular Biology, Colorado State University, Fort ColUns, CO 80523, USA C.F. Zorumski, Department of Psychiatry, Washington University School of Medicine, St. Louis, MO 63110, USA L.A. ZuccareUi, Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA

Preface Specific chemical factors participate in a multitude of processes that occur during normal development and maintenance of the nervous system, during the neurodegenerative changes associated with ageing, axonal injuries and various pathological processes and during the repair that may or may not follow nerve damage incurred by chemical, traumatic or infective insults. In some of these instances, our knowledge is in a relatively (to zero) advanced state; in others it is still largely suggestive or, at best, fragmentary. Over the last 10 years, particularly with the advent of gene cloning technologies, the rate of acquisition of new data has meant that the status and number of recognized chemical factors are continually changing and there is reason to anticipate that the list of endogenous molecules with apparently discrete functional roles will continue to increase by at least one per year, for some time to come. It would therefore be fruitless to presume that any survey of current knowledge will remain an exhaustive source for very long. In fact, under these circumstances, no matter how rapidly a reference work can be produced, everyone involved has an underlying suspicion that much of what is said will require some degree of qualification by the time it lands on a library shelf. Nevertheless, I believe there is strong justification at this time for a volume that spans the breadth of current thinking about how endogenous and exogenous chemical factors affect neuronal growth, integrity and repair. A regular stream of books is published in this area, but most of these are oriented towards specific types of chemical factor (e.g. Loughlin and Fallon, 1993; Ransohoff, 1996) or specific processes (e.g. Yurchenco et al., 1994; Schwartz and Osborne, 1995; Aschner, 1996) or specific neural pathways (e.g. Hendry and Hill, 1992). In these cases it is often difficult, particularly for the naive reader, to place the topics in question appropriately within the global context of neural function and dysfunction. In the current volume, with cross-referenced coverage spanning a wide range of themes, I hope that these transitions may be made more comfortably. A second characteristic of most books in this area is that each topic is usually treated purely from the viewpoint of a single group of authors. In the present volume, a conscious attempt has also been made to allow authors to summarize the relevance to their own themes of material dealt with primarily in other chapters. Thus, for example, a reader turning to the chapter on astroglial factors will find a summary of how various neurotrophins and adhesion molecules affect neural growth and repair, without having to consult each of the separate chapters dealing with these factors unless more detailed information is required. The book has been divided into four Sections, considering in turn factors involved with formation of axon pathways (Chapters 1-3), factors involved with developmental survival and specialization of particular neuron populations (Chapters 4-12), factors involved in general maintenance and repair of neurons (Chapters 13-16) and, finally, factors that are implicated as mediators of pathological neuronal damage (Chapters 17-18).

For vertebrate nervous systems, understanding the process of formation of axon pathways is hindered by the large numbers of neurons involved and the difficulties of dealing with embryonic tissues. These factors are far less of a problem with invertebrates, where the numbers of neurons are low and individual neurons can be reliably identified throughout development. For certain invertebrates such as C. elegans and Drosophila, analysis of the molecular basis of neural growth is additionally facilitated by the fact that the genomes are already well-characterized. In at least most respects, the events that take place in these species are likely to parallel those occurring in vertebrate systems and the specific chemical factors involved have, in some cases (for instance, the fascilin family), been shown to correspond to those functioning in vertebrates. In others, no vertebrate homologues are known but, even there, their structures may provide valuable insights into the vertebrate situation. Paul Whitington's survey of invertebrate axon pathway formation (Chapter 1) is therefore an appropriate way to introduce the theme of chemical factors in neural growth. Vertebrate nervous system development is characterized by the need not only for axonal guidance but also for post-neurulatory migration of many neuron populations from their sites of origin to more peripheral loci. This migratory behaviour depends on a range of protein factors that create or reduce adhesion between neurons and their substrates; the nature of these processes and the adhesion factors so far identified are discussed in Chapter 2 by Don Newgreen and Seong-Seng Tan. Some adhesion factors are also implicated in the process of axon guidance and laminin is one of these in which the effects are especially well-documented. In Chapter 3, Victor Nurcombe examines the roles of this protein in formation of a number of neural pathways. Development of the nervous system is characterized by two obligatory features in addition to directional migration of neurons and growth of axons. One of these is the death of a substantial proportion of the initially formed nerve cells by an active, genetically determined process. This so-called programmed cell death is thought to be an important factor in optimizing the intensity of synaptic inputs and removing aberrant synaptic connections; indeed, it has been suggested that the persistence of excess neuronal connections because of insufficient cell death during development may underlie some types of psychosis (Weinberger, 1987). Because of its seminal position in neural development and because similar cellular events appear to occur during neurodegenerative cell death, the molecular details of the death process and its genetic basis are of great practical importance. Aurora Messina and Anthony Jaworowski discuss these areas in Chapter 4. The other characteristic feature is that the initially formed neuroblasts are pluripotential, being able to act as precursors both for supporting cells and for a large range of mature neuron phenotypes. For vertebrate nervous systems, the survival of individual neurons during development and the commitment of individual cells to specific phenotypes both depend on the effects of a range of large peptide molecules. These peptides are collectively termed growth factors, although this is a rather misleading description of substances whose biological effects extend far beyond regulation of cell growth per se. In Chapter 5, Perry Bartlett and his colleagues provide an overview of the interactions between growth factors and developing neurons, presaging the more detailed coverage of individual factors in subsequent chapters. Many of these factors originate in the vicinity of the target cell, implying that, in order to influence the nuclear machin-

ery and regulate protein transcription, a signal must be conveyed retrogradely from axon terminal to cell body. Traditionally, this process has been envisaged as involving internalization and transport of the growth factor itself (see e.g. Thoenen et al., 1979). More recent experimental evidence, however, suggests that this concept is an over-simplification. In Chapter 6, Ian Hendry analyses the existing data and concludes that retrograde signalling is mediated indirectly by second messengers rather than by the growth factor molecules, with different factors utilizing distinct types of signalling machinery. By convention, the growth factors that protect neurons from programmed cell death during development are generically termed neurotrophins or neurotrophic factors. Four of these neurotrophins, nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3) and neurotrophin 4-5 (NT-4/5), represent products of a single gene family and exhibit a high degree of homology. Nevertheless, the individual family members manifest quite distinct biological activities. For example, within the neural crest-derived spinal sensory ganglia, protection from programmed cell death for nociceptive, mechanoreceptive and proprioceptive neuron populations has been shown to depend preferentially on NGF, BDNF and NT-3 respectively. Recent studies have demonstrated that there is an even more complex interaction between these factors than originally thought, with neuron populations changing their reliance from one factor to another as development progresses (see Davies, 1994). Because NGF has been studied for far longer than the other neurotrophins, its characteristics and actions have been reviewed repeatedly and yet another survey of that type would be redundant. By contrast, the regulation of NGF production, which is by definition essential to its biological activity, has been able to be studied only very recently, because of technical considerations. Robert Rush and his co-workers describe in Chapter 6 the provocative results of investigations in this area, which suggest that there is feedback regulation of NGF synthesis by neural inputs. The more recently discovered neurotrophins of the NGF family are reviewed in Chapters 8 and 9, where Karen Bailey discusses the properties of BDNF and Nati Rocamora and Ernest Arenas review those of NT-3 and NT-4/5. These authors also consider the existing evidence for deficiencies of these factors as causative influences in various neurodegenerative disorders. The final three Chapters in Section II deal with the phenomenon of neuronal phenotypic instruction. In Chapter 10, Lorraine lacovitti discusses the processes that underlie conversion of pluripotential neuroblasts to a catecholaminergic phenotype and provides evidence that this involves the synergistic interaction of traditional neurotrophins with other molecules. Expression of catecholaminergic enzymic pathways can be induced by this means even in neurons that normally are not catecholaminergic and can upregulate enzyme expression in adult as well as developing tissue. These findings may have parallels in other neurotransmitter systems and raise the prospect of novel approaches to specific neurotransmitter deficits such as occur in Parkinsonism and motor neuron disease. The two following Chapters deal with two closely related proteins, both of which promote dedication of neurons to a cholinergic phenotype, as well as exerting trophic survival effects on a variety of cell types. In Chapter 11, Tetsuo Yamamori surveys the state of knowledge about leukemia inhibitory factor (LIE) while Chapter 12, by Peter Richardson and M.C. Subang, deals with ciliary neurotrophic factor (CNTE).

A number of molecules are known to enhance neuronal viability and repair processes, without their being absolutely essential for neuron survival. These properties make the molecules obvious and attractive candidates for therapeutic use and the identities and properties of several such factors are examined in Section III, together with assessments of their potentials for therapeutic application. In some cases, these neurosupportive molecules are blood-borne; for instance, the 1-30 peptide sequence of adrenocorticotropin, which accelerates formation of somatic neuromuscular synapses as well as being able to protect other neuron phenotypes against neurotoxininduced damage. This topic is discussed by Fleur Strand and her colleagues in Chapter 13. Other factors are produced by neurons or their supporting cells. One group of such molecules that appears to exert particularly wide-ranging actions is the fibroblast growth factor (FGF) family which, as detailed by Sophie Bieger and Klaus Unsicker (Chapter 14), can modulate cell proliferation, survival and differentiation in both neuronal and glial lineages. Bioavailability issues may prevent effective administration of peptide growth factors for therapeutic purposes. An alternative strategy is to enhance tissue production of endogenous factor and, where neuronal repair is the aim, then the supporting cells are the obvious tissue elements to target. In Chapter 15, Hans Muller and associates review briefly the range of neurostimulant factors derived from astroglial cells, revisiting in part the material introduced in previous chapters but now with an overall emphasis on optimizing nerve repair. A final and distinct group of factors to be considered in relation to maintenance of neuronal function is the insulin-like growth factors (IGFs), which exhibit a mixture of properties of the different trophic molecules discussed earlier. The IGFs are derived from neurons and glia as well as from target cells and, unlike other factors from these sources, are present in detectable concentrations in the bloodstream as well as in tissue. They exert effects reminiscent of classical neurotrophins, binding to specific membrane receptors and enhancing neurite outgrowth and neuron survival and proliferation, but are far less selective in their actions than the NGF family. IGFs therefore may act as generalized support factors for neurons throughout the body, with the correlate that a decline in IGF levels may have generalized deleterious effects on neural function. These important issues are addressed by Douglas Ishii and his colleagues in Chapter 16. In order to optimize therapeutic approaches to neural damage, it is essential to understand the molecular mechanisms by which degenerative changes occur. Analysis of this area has been facilitated by the availability for animal use of a range of neurotoxins, including 6-hydroxydopamine and excitatory amino acids such as kainic acid, together with the fortuitous clinical experiment which resulted from MPTP ingestion (Langston et al., 1983). It now seems clear that many types of brain injury, including those due to hypoglycaemia and ischaemia, result from excessive activation of excitatory amino acid receptors, with subsequent deleterious surges in intracellular free calcium. Local production of free radical species, either inside or adjacent to the cell, is a second common mediator of death. This knowledge opens the possibility that substantial degrees of neuroprotection might be afforded by administration of agents that scavenge or prevent production of free radicals or which prevent intracellular calcium fluxes. Such approaches would have the potential advantage of bypassing the practical difficulties associated with administration of peptides. In Chapter 17, Mario Gotz and his colleagues describe the processes of free radical production and

provide a detailed survey of how these pathways may be implicated in various types of neural damage. Finally, in Chapter 18, Charles Zorumski and John Olney discuss the mechanisms of excitatory amino acid receptor-mediated neuronal damage and examine the evidence for involvement of this process in a variety of neuropsychiatric states. What of the next few years? It is certain that the field will become ever more complex and that many of the data discussed in the current volume will require reassessment. Probably, many of the hypotheses made by the authors will also require revision. Already, studies using administration of an NT-3 antibody to neonatal rats (Zhou and Rush, 1995) have challenged the accepted view that NGF is the only sympathetic survival factor over the period of programmed cell death. Progressive introduction and application of new technical approaches will undoubtedly improve the precision with which hypotheses can be tested. For example, the availability of fibroblasts transfected with specific neurotrophins now allows endogenous neurotrophin production to be selectively enhanced at localized sites in the intact animal (Arenas and Persson, 1994), while genetic engineering will continue to provide more precise information on the effects of selectively deleting single chemical factors from the body. As the multifarious neurally active growth factors become more and more closely studied, the breadth of their biological roles will also become apparent. The FGFs and LIF are already known to stimulate growth of various non-neuronal as well as neuronal cell types, but the true functional significance of these and other factors in vivo may be even more wide-ranging. In addition to its neurotrophic actions, NGF is now acknowledged to exert potent effects on the behaviour of peripheral nociceptive nerve terminals and appears to constitute an endogenous proinflammatory agent with the potential capacity to modulate the immune system (e.g. Amann et al., 1996; McMahon, 1996). It seems highly likely that gene deletion studies will soon demonstrate that other growth factors also have functions quite distinct from their accepted ones. Some final words of caution need to be given here in relation to cross-species comparisons in the experimental paradigms that are, of necessity, employed in studies of these areas. Virtually all studies of mammalian neural development utilize the rat, with the exception of gene deletion studies which, for technical reasons, are routinely performed in mice. Although most events involved in neural function are probably at least broadly similar between these species, the possibility that this is not so must temper extrapolation of observations from mouse to rat, until adequate supportive data are obtained. In this context, recent studies of a glial-derived factor (GDNF) should be mentioned. Developmentally, GDNF has been demonstrated to possess survival-enhancing neurotrophic activity on several populations of rat central nervous system neurons, both in vitro and in vivo. Mouse gene deletion studies, however, show no impairment of central neuron maturation, although dramatic deficits are seen in some peripheral neural populations (Moore et al., 1996; Sanchez etal., 1996). A more philosophical problem arises with the question of extrapolating from animal studies on neurodegeneration and repair to the clinical situation. Once again, there is no inherent reason to doubt that similar mechanistic events exist in rats (or mice) to those in humans. Nevertheless, given the fact that the respective lifespans

differ by an order of magnitude, progressive deteriorative or restorative changes may take place in human neurons that can never have time to occur in rodents. Conversely, the high metabolic rate of rodents may mean that manipulations of the neural environment in these species induce effects that are vastly different in amplitude to those seen in man. In this field of applied research perhaps more than any other, progress towards the goal of improved patient care will rely on the closest possible contact between preclinical and clinical scientists and on constant cross-correlation between the observations made in in vitro, intact animal and human situations. Christopher Bell July, 1996 References Amann, R., Schuligoi, R., Herzeg, G. and Donnerer, J. (1996) Intraplantar injection of nerve growth factor into the rat hind paw: local edema and effects on thermal nociceptive threshold. Pain 64: 323329. Arenas, E. and Persson, H. (1994) Neurotrophin-3 prevents the death of adult central noradrenergic neurons in vivo. Nature 376: 368-371. Aschner, M. (Ed.) (1996) The Role ofGlia in Neurotoxicity, CRC Press, London. Davies, A.M. (1994) Switching neurotrophin dependence. Curr. Biol. 4: 273-276. Hendry, I.A. and Hill, C.E. (Eds.) (1992) Development, Regeneration and Plasticity of the Autonomic Nervous System, Harwood, Berkshire. Langston, J.W., Ballard, P.A., Tetrud, J.W. and Irwin, I. (1983) Chronic parkinsonism in human due to a product of meperidine-analog synthesis. Science 225: 1480-1482. Loughlin, S.E. and Fallon, J.H. (Eds.) (1993) Neurotrophic Factors, Academic Press, New York. McMahon, S.B. (1996) NGF as a mediator of inflammatory pain. Philos. Trans. R. Soc. London Ser. B 351:431-440. Moore, M.W., Klein, R.D., Farinas, I., Sauer, H., Armanini, M., Phillips, H., Reichardt, L.F., Ryan, A.M., Carver-Moore, K. and Rosenthal, A. (1996) Renal and neuronal abnormalities in mice lacking GDNF. Nature 382: 76-79. Ransohoff, R.M. (Ed.) (1996) Cytokines and the CNS: Development, Defense and Disease, CRC Press, London. Sanchez, M.P., Silos-Santiago, I., Frisen, J., He, B., Lira, S.A. and Barbacid, M. (1996) Renal agenesis and the absence of enteric neurons in mice lacking GDNF. Nature 382: 70-73. Schwartz, L.M. and Osborne, B.A. (eds) (1995) Cell Death, Academic Press, London. Thoenen, H., Otten, U. and Schwab, M. (1979) Orthograde and retrograde signals for the regulation of neuronal gene expression: the peripheral sympathetic nervous system as a model. In F.O. Schmitt and E.G. Worden (Eds.), The Neurosciences: Fourth Study Program, MIT Press, Cambridge, MA. Weinberger, D.R. (1987) Implications of normal brain development for the pathogenesis of schizophrenia. Arch. Gen. Psychiat. 44: 660-669. Yurchenco, P.D., Birk, D.E. and Mecham, R.P. (eds) (1994) Extracellular Matrix Assembly and Structure, Academic Press, San Diego, CA. Zhou, X.F. and Rush, R.A. (1995) Sympathetic neurons in neonatal rats require endogenous neurotrophin-3 for survival. J. Neurosci. 15: 6521-6530.

Contents List of contributors Preface

v vii

Section I. Factors Implicated in Neural Pathway Formation 1. Axon guidance factors in invertebrate development P.M. Whitington (Armidale, Australia)

3

2. Adhesion molecules in neural crest development D.F. Newgreen and S.S. Tan (Parkville, Australia)

45

3. Laminin in neural development V. Nurcombe (Parkville, Australia)

67

Section II. Factors Implicated in Neuron Survival and Specialization 4. Mechanisms of developmental cell death A. Messina and A. Jaworowski (Parkville, Australia)

89

5. Regulation of the early development of the nervous system by growth factors P.F. Bartlett, T.J. Kilpatrick, L.J. Richards, P.S. Talman and M. Murphy (Parkville, Australia)

123

6. Retrograde factors in peripheral nerves LA. Hendry (Canberra, Australia)

149

7. The regulation of nerve growth factor synthesis and delivery to peripheral neurons R.A. Rush, R. Mayo and C. Zettler (Adelaide, Australia)

171

8. Brain-derived neurotrophic factor K.A. Bailey (Parkville, Australia)

203

9. Neurotrophin-3 and neurotrophin-4/5 N. Rocamora and E. Arenas (Barcelona, Spain and Stockholm Sweden)

219

10. Centrally-active differentiation factors in the nervous system L. lacovittti (Philadelphia, PA, USA)

251

XIV

11. Leukemia inhibitory factor and phenotypic specialization T. Yamamori (Okazaki, Japan)

265

12. Ciliary neurotrophic factor P.M. Richardson and M.C. Subang (Montreal, Canada)

293

Section III. Factors Implicated in Neuronal Support and Repair 13. Melanocortins as factors in somatic neuromuscular growth and regrowth F.L. Strand, K.A. Williams, S.E. Alves, F.J. Antonawich, T.S. Lee, S.J. Lee, J. Kume and L.A. Zuccarelli (New York, USA)

311

14. Functions of fibroblast growth factors (FGFs) in the nervous system S. Bieger and K. Unsicker (Heidelberg, Germany)

339

15. Astroglial neurotrophic and neurite-promoting factors H.W. Miiller, U. Junghans and J. Kappler (Dusseldorf, Germany)

377

16. Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival D.N. Ishii, S.F. Pu, G.W. Glazner, H.-X. Zhuang and D.J. Marsh (Fort Collins, CO, USA)

399

Section IV. Factors Implicated in Neuronal Damage 17. Oxidative stress: free radical production in neural degeneration M.E. Gotz, G. Kunig, P. Riederer and M.B.H. Youdim (Haifa, Israel)

425

18. Excitotoxic neuronal damage and neuropsychiatric disorders C.F. Zorumski and J.W. Olney (St. Louis, MO, USA)

511

Subject Index

531

Section I Factors Implicated in Neural Pathway Formation

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved. CHAPTER 1

Axon guidance factors in invertebrate development Paul M. Whitington University of New England, Department of Zoology, Armidale, NSW 2351, Australia

1. Introduction The question of how axons are guided to their synaptic targets presents us with one of the most challenging problems in developmental biology. This is a problem with a long evolutionary history, as even primitive metazoans, such as the nematodes, depend upon precise neural connectivity to generate functionally appropriate behaviour. We might therefore reasonably expect cellular and molecular mechanisms for axon guidance that evolved in one animal group to be re-deployed later, during the evolution of higher forms such as the vertebrates. This expectation has been largely borne out by recent discoveries. Cellular phenomena that were first documented in insects, such as the pioneering of axon pathways by specific neurons and selective fasciculation between later growing axons, have subsequently been reported in a number of vertebrate embryos (reviewed in Goodman and Shatz, 1993), while several of the molecules implicated in axon guidance in invertebrates have been found to have homologues performing apparently similar functions in vertebrates (reviewed in Hortsch and Goodman, 1991; Goodman, 1994). The purpose of this review is to present some of the insights concerning the nature and action of axon guidance factors that have been gleaned from recent studies in the invertebrates. Most of the research activity in this area has focussed on a small, but diverse group of species: the nematode Caenorhabditis elegans, the leech, and two insects, the grasshopper and the fruitfly Drosophila melanogaster. (Common names are used in this

review for those invertebrates where different workers have chosen to study distinct, but closely related species, such as the four grasshoppers, Schistocerca americana, Schistocerca nitens, Schistocerca gregaria and Locusta migratorid). These invertebrates share a feature, common to many of their relations, that greatly assists the analysis of the cellular and molecular basis for axon guidance: in each species, individual neurons can be identified and reliably recognised at different developmental stages. This affords a high degree of precision in descriptions of the cellular basis for axon guidance and considerably assists the interpretation of experiments in which these events are perturbed by surgical or genetic means. In addition, the pattern of expression of gene products can be characterised at the level of individual neurons, providing insights into the role of these molecules in axon guidance. Two species, C. elegans and Drosophila, possess another common feature of decisive value for investigations into the genetic/molecular basis for axon guidance: a well-characterized genome. This genetic tool can be applied in a variety of ways (see Marx, 1984; Rubin, 1988; Thomas and Crews, 1990). One of its most useful applications is to generate null mutants for specific genes. These mutants completely lack the product of that gene in their tissues and thus provide a definitive test of its role in axon guidance. This review draws together information about axon guidance at the cellular level in different invertebrates, most of which has come from studies of embryonic systems, to consider the following issues: What strategies do growing axons

Axon guidance factors in invertebrate development

use to navigate to their targets? Is axon growth highly directed from the outset or does it involve exploration of alternative pathways? What sorts of cellular and extracellular structures are used by axons as guidance cues? How specific are these cues for different types of neurons? Is there any redundancy in the cues used? How do guidance cues act to elicit directed axon growth? The value of such information in uncovering the identity of axon guidance molecules is discussed and information currently available about such molecules from invertebrate studies is reviewed. 2. Cellular mechanisms for axon guidance 2.1. Growth of invertebrate axons is often accurate Studies in a wide range of neural systems in invertebrate embryos have pointed to the generally stereotypic nature of the axon growth process. Axons often extend, from the outset, down pathways corresponding to their mature trajectories. Specific axon growth has been described for insect (specifically, grasshopper and fruitfly) motoneurons (Bastiani et al., 1984; Ball et al, 1985; du Lac et al., 1986; Anderson and Tucker, 1988; Johansen et al., 1989; Whitington, 1989; Myers et al., 1990; Halpem et al., 1991; Sink and Whitington, 1991a; Van Vactor et al., 1993), intemeurons (Raper et al. 1983a,b; Murray et al., 1984; Bastiani and Goodman, 1986; Jacobs and Goodman, 1989b; Klambt et al., 1991; Myers and Bastiani, 1993; Lin et al, 1994) and sensory neurons (Heathcote, 1981; Ho and Goodman, 1982; Bentley and Keshishian, 1982a; Caudy and Bentley, 1986c; Hartenstein, 1988); crustacean motoneurons and intemeurons (Whitington et al., 1993); leech motoneurons (Kuwada, 1984) and sensory neurons (Kuwada and Kramer, 1983; Johansen et al., 1992; Briggs et al., 1993; Jellies et al., 1994); and nematode intemeurons and motoneurons (Mclntire et al., 1992). Clearly this is a widespread strategy for navigation of axons to their targets; it is not restricted to particular classes of neurons or to particular species.

2.2. Directed axon growth involves active recognition of guidance cues Two main classes of mechanisms could underlie the directed axon growth seen during normal development. The first, so-called "passive" mechanisms, suggest that an axon follows a specific trajectory because this is the only one physically available at that time in development. Later appearing axons are assumed to follow different routes because the earlier growth substrates are either absent, inaccessible or already occupied. This type of mechanism predicts that dismptions to the relative timing of axon growth should lead to altered pathway choices. The second class of axon guidance mechanisms assumes that an axon actively "chooses" a particular pathway from amongst a number of equally accessible altematives. Whilst the availability of permissive growth substrates may indeed constrain the pathway options of an axon and contribute to the final specific projection, it is highly unlikely that passive mechanisms by themselves could explain the directed growth behaviour of invertebrate axons described above. This conclusion is based on three main lines of evidence: a. Axons often have access via the filopodia of their growth cones to pathways other than the one ultimately followed; b. In experimental situations, axons may grow down some of those alternative pathways; c. The relative timing of axon growth during normal development is not absolutely precise and experimental alteration of the timing of outgrowth does not necessarily result in errors in growth trajectories. Studies in the grasshopper embryo have provided some of the most conclusive evidence of this type. Bastiani and Goodman (1986) showed that the axon of the pCC neuron in the central nervous system (CNS) of this insect embryo grows along a specific axon bundle or fascicle which is pioneered by the axons of the dMP2 and MPl neurons (Fig. 1). In so doing, it ignores the immediately adjacent vMP2 fascicle, which is within reach of the filopodia of its growth cone. The pCC axon ceases its

P.M. Whitington 28

29

30

31

& Q^.

rig. 1. Specific axon fasciculation in the grasshopper embryo. (A) The pattern of axonal outgrowth from the identified neurons dMP2, MPl, U, aCC and pCC in the grasshopper embryo from 28 to 37% of development. The anteriorly directed pCC axon fasciculates with the MPl/dMP2 axons. The aCC axon follows the U axons into the intersegmental nerve. Scale bar = SOjum. (B) The effect of ablation of MPl/dMP2 neurons in the same and next most anterior segment on axon growth from the pCC neuron. The p e c axon on the control side (CON) has advanced along the MPl/dMP2 fascicle, whereas the pCC axon on the operated side (EXP) has failed to grow. The debris of the ablated MPl/dMP2 neurons is indicated with an asterisk. Neurons were injected intracellularly with Lucifer yellow (LY). Scale bar = 20yum (reproduced with permission from Bastiani et al., 1986).

extension when it arrives at the anterior limit of the dMP2/MPl axons, despite the fact that the vMP2 fascicle has extended anterioriy beyond this limit. It only continues advancing anterioriy when the posterioriy growing dMP2/MPl axons from the next most anterior segment arrive in its vicinity. Ablation of the dMP2 and MPl neurons causes the p e c axon to behave abnormally; axon formation is aborted or the axon grows in aberrant directions. The grasshopper motoneuron aCC shows a similar behaviour. Its axon specifically follows the axons of the U motoneurons to the edge of the CNS (Fig. 1), even though other axons such as vMP2 and MPl are within filopodial reach (du Lac et al., 1986). If the U neurons are ablated, the aCC axon fails to advance posteriorly along its normal trajectory. If the neuroblast that generates the aCC neuron is ablated, the aCC neuron is bom later in development after this neuroblast is replaced. Despite the subsequent delay in the outgrowth of the aCC

axon, the trajectory followed by this axon is normal (Doe et al., 1986). However, if the delay in axon outgrowth is increased, by ablating the neuroblast at later stages of development, the aCC axon does make growth errors. This result shows that the timing of axon outgrowth is not absolutely irrelevant to the outcome of the growth process: the environment encountered by the axon may change during development such that guidance cues available to it earlier may no longer be present or accessible. 2.3. Initial axon growth in invertebrates is not always error-free While the navigation of axons in invertebrates is often accurate, it is not invariably error-free. In many situations, axon branches have been found on occasion to extend down pathways in which they are not found in the mature animal. Such er-

Axon guidance factors in invertebrate

rors are corrected at later developmental stages by the retraction of the inappropriate branches. (Note that since the axonal anatomy of the mature neuron is usually not absolutely stereotypic, the term "inappropriate" means that such a branch does not correspond with what is generally observed in the adult, and therefore refers to a statistical rather than an absolute concept.) Early studies in grasshopper (e.g. Raper et al., 1983b,c; Bastiani et al., 1986) and leech (e.g. Kuwada, 1984) embryos emphasised the accuracy of axon growth, although when the data in these reports is re-examined, cases of inappropriate projections can be seen. For example, the MPl neuron in the grasshopper embryo forms an anteriorly directed branch in addition to its appropriate posterior branch (Bastiani et al., 1986). The G neuron

development

sends lamellipodia and short branches down several other routes as it is turning posteriorly into its correct path, the longitudinal connective (see Fig. 4 in Raper et al., 1983b). Heathcote (1981) reported that the stretch receptor axon in the grasshopper embryo sometimes forms additional branches in the periphery and his figures also show that projection errors occur, albeit less frequently, in the CNS. Indications that the formation of inappropriate projections might be a more widespread phenomenon came from later studies in leech and grasshopper embryos. Leech motoneurons (Baptista and Macagno, 1988) and sensory neurons (Gao and Macagno, 1987a,b) were found to branch extensively along inappropriate peripheral and central pathways (Fig. 2). Myers et al. (1990) found that

9

[9

B

Fig. 2. Inappropriate axonal branching in the leech embryo. (A) Rostral penile evertor (RPE) motoneurons in a 50-day juvenile leech, showing the morphology typical of the mature adult. The pair of RPE motoneurons has been stained in this preparation. Each neuron sends an axonal branch out the contralateral, anterior nerve root in the same segment and another out the homologous nerve in the next most anterior segment, via the longitudinal connective. Both branches terminate on the target, the male genitalia. (B) The morphology of the RPE motoneuron in a 15-day-old embryo. The neuron possesses many extra-ganglionic branches which are later withdrawn. (C) The morphology of the pair of RPE motoneurons in a 50-day juvenile leech whose male genitalia were removed at 15 days of development. The neurons retain many axonal branches in aberrant pathways and their branching is more extensive than at any time during normal development. Camera lucida drawings of HRP intracellular injections. Scale bar = 200 ^m (reproduced with permission from Baptista and Macagno, 1988 © Cell Press).

P.M. Whitington

Fig. 3. The axon morphology of the SETi motoneuron in the grasshopper embryo at 48% of embyrogenesis. The branches indicated by arrows are inappropriate branches, never observed in animals older than 58% of embryogenesis. Camera lucida drawing of a LY intracellular fill. Scale bar =lOOjum (reproduced with permission from Myers et al., 1990).

the extensor tibiae motoneurons of the grasshopper embryo send inappropriate axonal branches down a variety of nerve pathways in the limb bud (Fig. 3). Halpem et al. (1991) and Sink and Whitington (1991a) have observed that the RP motor axons in Drosophila extend over a number of inappropriate muscles in addition to their appropriate targets and retract these branches later in embryogenesis. 2.4. Under what circumstances does inappropriate branching occur? The reasons why inappropriate branching occurs in some situations and not in others are unclear. The phenomenon is not restricted to a particular group of animals or to a particular class of neurons. It has been reported in a wide variety of

invertebrate species, in all classes of neurons (motor, sensory and intemeurons), in both the CNS and in the periphery, and for both pioneering neurons (e.g. Kuwada and Kramer, 1983) and neurons that grow out later (e.g. Myers et al., 1990). A given neuron may show inappropriate branching at some points along its axon trajectory and not at others; for example, the axons of the RP neurons in the Drosophila embryo consistently turn in an exclusively posterior direction when entering the longitudinal connective, yet branch over a variety of muscles in the periphery (Sink and Whitington, 1991a). Even at sites where a neuron consistently forms inappropriate branches, the distribution of these branches may vary between individual embryos; for instance, RP motor axon branching over muscles in Drosophila (Sink and Whitington, 1991a). A common error made by axons when they fasciculate with a pre-existing axon bundle is to send a branch in the incorrect direction along that fascicle. Such branches are often of substantial length and tend to persist longer than other types of branching errors. The high frequency of such errors might stem from an inherent ambiguity in polarity cues along axons and/or a fundamental difficulty faced by growth cones in interpreting such cues. Amongst the insects, there appears to be a trend for more accurate initial axon navigation within the CNS: inappropriate branching is more often seen in the periphery (Myers et al., 1990; Sink and Whitington, 1991a). However, this does not hold for the leech. Widespread inappropriate axon branching within the CNS is a feature of axon development for the RPE motoneurons (Baptista and Macagno, 1988). A possible explanation for this variable behaviour is that all growth cones have the potential to form inappropriate branches in any given situation. Whether a filopodium contacting a particular substrate develops into an axonal branch might depend upon the relative affinity between molecules on the filopodium and that substrate. If the affinity is weak, the filopodium might only develop into a thin branch or be quickly retracted; if it is strong, the filopodium might give rise to a substantial

Axon guidance factors in invertebrate development

axon (see Section 4.2 for a discussion on the relationship between filopodial and axonal development). If affinity of the growth cone for the appropriate substrate is only slightly stronger than for alternative routes, inappropriate axonal branches might be formed along those alternative paths. Whether accurate axon growth or inappropriate branching is seen could also depend upon how closely the axon growth process is monitored. Inappropriate branches may be invariably formed but withdrawn quickly at a specific location, in which case a short time interval between observations will be required to detect inappropriate branches. Time-lapse observations of the dynamics of axon growth in the CNS of the Drosophila embryo (Murray and Whitington, unpublished) show that filopodia up to 10/^m in length can be extended and retracted over a time interval of 5 min. Many such events would be missed using the conventional method of following axon growth by fixing embryos at various developmental stages. Whatever the reasons for the initial formation of inappropriate axonal branches, the neuron must have a mechanism for their removal. Whether the environmental cues responsible for this pruning process are the same as those that initially guide the growth cone in a particular direction remains to be determined, as does the operation of the subcellular machinery involved in the response of the neuron to those cues. This matter is discussed in more depth in Section 4.2. 3. The nature of axon guidance factors 3.1. Are axon guidance factors substrate-bound or diffusible? An issue that has received much attention in the vertebrate literature is whether axons are attracted by a gradient of diffusible chemo-attractant released from the target or whether they are guided by substrate-bound molecules in the immediate vicinity of the growth cone. This is clearly an important question because it dictates where we look for guidance cues - on cells along the pathway or in cells far removed. One might also expect that the cellular machinery for reading and responding

to diffusible guidance cues would be different to that used for substrate-bound molecules. The most definitive evidence for a role for chemotropism in axon guidance in vertebrates has come from in vitro systems in which neural explants or individual neurons have been shown to send neurites in the direction of the source of a diffusible chemical. For example, the growth cones of chick dorsal-root ganglion neurons orient towards increasing concentrations of nerve growth factor (Letoumeau, 1978; Gunderson and Barrett, 1979). Commissural axons from the embryonic rat spinal cord extend preferentially toward floorplate explants in a manner suggestive of a chemotropic mechanism (Tessier-Lavigne et al., 1988; Placzek et al., 1990). Recently, two closely-related proteins isolated from embryonic chick brains, netrin-1 and netrin-2, have been shown to possess chemotropic activity in this spinal cord assay system (Serafini et al., 1994; Kennedy et al., 1994). A variety of observations of axon growth both during normal development and in response to surgical transplantation of neuronal targets (reviewed in Goodman and Shatz, 1993) also provide indirect support for chemotropic guidance cues in vertebrates. Little direct evidence for chemotropism is available for invertebrates, due largely to the fact that tissue/organ culture systems have not been developed and exploited to the same extent as in vertebrates. Hay don et al. (1984) have shown that neurons of the snail Helisoma respond to exogenous application in vitro of the neurotransmitter serotonin by retraction of their growth cones. However, the precise role of this substance in vivo has yet to be determined (see Section 4.1). In vivo observations provide little evidence for chemotropism in invertebrates: on the contrary, most of the behaviour of growth cones in these embryos is suggestive of local, substrate-bound guidance. Reorientation of growth cones takes place when filopodia are in contact with a welldefined substrate feature and turning behaviour of axons is generally associated with a change in the substrate. In some situations, invertebrate axons show predictable turning behaviour in the absence of obvious, local cellular cues. For example, the

P.M. Whitington

growth cone of the Ql neuron in the grasshopper embryo turns towards the midline of the CNS without contacting any apparent neuronal or glial cells (Myers and Bastiani, 1993). However, the Ql growth cone may be responding to local cues on the overlying basal lamina, rather than to a diffusible attractant. It has been shown in several insect systems that axons can grow right into the target area in the absence of the target tissue, excluding a role for the target as a source of a chemotropic factor for long-distance axon guidance. For example, Berlot and Goodman (1984) showed that sensory axons in the grasshopper antenna follow a normal pathway to the base of the antenna, even when it is detached from the rest of the embryo. Similar results were obtained in the grasshopper limb bud for navigation of the afferent Trl axons (Lefcort and Bentley, 1987) and for navigation of efferent motor axons of the Drosophila embryo in the absence of their target muscles (Sink and Whitington, 1991b; Cash et al., 1992). Such findings do not exclude a role for short-range chemotropic guidance of axons once they have been guided into the vicinity of the target tissue by other means. Nor do they rule out the release of chemotropic factors from sites along the trajectory of the axon. However, they do speak strongly against long-range chemotropic guidance by target-derived cues. If a chemotropic effect is in operation in invertebrates, it may be revealed by the orienting response of axons in vivo following transplantation of their targets to an ectopic site. Rostral penile evertor motor axons in the leech embryo can successfully innervate their target tissue after it is transplanted to ectopic segments (Baptista and Macagno, 1988). However, they appear to do so by expanding their growth in several directions simultaneously, both towards and away from the target, rather than by directed turning towards the ectopic target from a distance. Lin and Goodman (1994) have found that motor axons in the Drosophila embryo which have been misrouted from their normal growth pathways by inducing ectopic or increased expression of the protein Fasciclin II (see Section 5.1.1), are often able to correct their pathfinding errors and take

circuitous routes to find their target muscles. However, it is conceivable that these motomeurons locate their muscles by a process of widespread axon branching, followed by retraction of branches on inappropriate muscles, rather than by following a target-muscle derived chemotropic factor. In summary, the possibility of chemotropic guidance of axons in invertebrates is intriguing but little direct evidence exists to support it. However, it would be foolish at this stage to regard chemotropism as an exclusively vertebrate phenomenon, particularly as the netrins, strong candidates for a chemotropic factor in the vertebrates (Serafini et al., 1994; Kennedy et al., 1994), show a high degree of sequence similarity to the nematode UNC6 protein (see Section 5.2.1). 3.2. Cellular identity of substrate-bound axon guidance factors In contrast to the paucity of evidence for diffusible, chemotropic factors in invertebrates, a large number of observations support a role for substrate-associated factors in axon guidance. Such factors appear to be located on a variety of cellular and extracellular structures. Precise identification of these structures is indispensable for characterising the molecular nature of the axon guidance factors and elucidating their function. The various classes of cellular structures implicated in axon guidance in invertebrates are considered in turn, below. 3.3. Guidance cues on pre-existing axons 3.3.1. Selective fasciculation and the ''labelled pathways" hypothesis Studies in many invertebrate systems have pointed to the importance of pre-existing axons in guiding axon growth. The fasciculation of axons growing in the CNS of the grasshopper embryo (Raper et al., 1983a-c; Bastiani et al., 1984; Bastiani et al., 1986; Doe et al., 1986) has already been discussed, and a similar behaviour is displayed by sensory and motor axons navigating in the periphery in this insect (Bate, 1976; Keshishian and Bentley, 1983a; Whitington, 1989; Meier and

10

Reichert, 1991), in the cricket embryo (Edwards and Chen, 1979), in the Drosophila embryo (Goodman et al., 1984; Anderson and Tucker, 1988; Hartenstein, 1988; Jacobs and Goodman, 1989b; Johansen et al., 1989; Halpem et al., 1991; Sink and Whitington, 1991a) and pupa (Murray et al., 1984; Jan et al., 1985; Lienhard and Stocker, 1991), in the leech CNS and periphery (Johansen et al., 1992; Briggs et al., 1993; Jellies et al., 1994), in the dorsal and ventral nerve cords of the nematode C. elegans (Mclntire et al., 1992) and in the central nervous system of two crustaceans, the woodlouse Porcellio scaber and the crayfish Cherax destructor (Whitington et al., 1993). In some of these situations, the choice of axon pathway appears to be specific, as a particular preexisting fascicle is chosen over other nearby fascicles that are within filopodial reach. This applies especially to axon growth within the CNS (Bastiani et al., 1984; Bastiani, et al, 1986). Such behaviour indicates that the pre-existing axonal path is more than a permissive substrate, chosen merely because of its physical proximity. Rather, it suggests that the axon can discriminate between the chosen pathway and other neighbouring permissible routes. Goodman and co-workers developed the "labelled pathways" hypothesis to explain such specific growth cone behaviour (Raper et al., 1983a,b). This hypothesis proposes that "axonal pathways are differentially labelled on their cell surfaces" and that "later growth cones are differentially determined in their ability to make specific choices of which labelled pathways to follow" (Raper etal., 1983c). In other situations, the axonal pathway followed by a growth cone appears to be the only one within filopodial reach of that growth cone at that developmental stage. For example, the behaviour of early differentiating peripheral neurons in the limb bud of the grasshopper embryo is consistent with the growth cones of these neurons following the first axon fascicle they happen to contact (Keshishian and Bentley, 1983a). A similar explanation may underlie the pathway choices made by peripheral sensory neurons in the body wall of the Drosophila embryo (Hartenstein, 1988). In such cases, there is no need to assume that different ax-

Axon guidance factors in invertebrate development

onal pathways are differentially labelled. Rather, the growth cone may simply recognise a general molecular characteristic of the axonal pathway shared by all, or a large subset, of neurons of that species. 5.5.2. Pre-existing axons can function as necessary guidance cues While growth cones can utilise pre-existing axon pathways as a growth substrate, this does not necessarily indicate that such pathways are essential for directed axon growth. Experimental evidence that fasciculation with pre-existing axons can play a necessary role in axon guidance has been obtained for several neurons in the CNS of the grasshopper embryo, du Lac et al. (1986) reported that the aCC axon, which normally follows the U axons into the intersegmental nerve, fails to extend out this, or any other, pathway if the U neurons are ablated. Bastiani et al. (1986) ablated the MPl and dMP2 neurons and found that the pCC axon, which normally fasciculates with these axons, failed to extend along the connective (Fig. 1). Raper et al. (1984) ablated the A and P axons and showed that the axon of the G neuron, which normally fasciculates with the A/P fascicle in the longitudinal connective, only extended a short distance anteriorly along the connective. In a different insect, the cricket Acheta domesticus, laser ablation of pioneer neurons located in the distal part of the cercus, an abdominal appendage, resulted in defects in axon growth from sensory neurons that developed subsequent to the ablation (Edwards et al., 1981). In other cases, axon fasciculation, while a feature of normal development, does not appear to be necessary for directed axon growth. Ablation of the Til cells in the grasshopper limb bud deprives the more proximally located Fl and F2 neurons of the opportunity of fasciculating with the Til axons, as they normally do (Keshishian and Bentley, 1983b). Nonetheless, the Fl and F2 axons follow normal trajectories in these experimental embryos (Fig. 4). This finding suggests that there is some redundancy of axon guidance cues in this system. The possible identity of the alternative guidance factors is discussed in Section 3.5.

P.M. Whitington

11

choice of particular axons as growth substrates appears to be absolute since, in the absence of these substrates, axon extension is halted. The G neuron, on the other hand, often continues to extend an axon in a variety of aberrant directions when its normal growth substrate, the A/P axon fascicle, is ablated (Raper et al., 1984). These aberrant branches have a limited length, usually less than SOjum, The choice of central axonal pathway followed appears to be random: there is no clear preference for a particular alternative. A similar behaviour is shown by the axon of the grasshopper motoneuron FETi (Fast Extensor Tibiae), which consistently extends down aberrant pathways in the body wall when its normal exit route from the CNS, nerve 5, is missing (Whitington and Seifert, 1984). However, the growth of the FETi axon in the periphery in this

Fig. 4. The effect of ablation of pioneer Til neurons on axonal growth from peripheral neurons in the grasshopper limb bud. (A) Operated limb bud in which the Til neurons had been ablated prior to axonogenesis (30% stage) and the embryo allowed to develop in culture medium to the 40/45% stage. The pattern of axon growth from neurons developing proximal to the ablated Til neurons is normal. (B) The pattern of axon growth in the corresponding control, contralateral limb bud to (A). The neuron pair labelled CTl is referred to as Cxi in the text. Camera lucida drawings of embryos stained using anti-HRP inmiunohistochemistry. Scale bar= lOOyum (reproduced with permission from Keshishian and Bentley, 1983b).

More recently, Lin et al. (1994) have found that pCC, MPl, dMP2 and vMP2 axons in the CNS of fasciclin II mutant Drosophila embryos fail to show the normal pattern of fasciculation seen in wild-type embryos. Nonetheless, these axons all grow in the correct direction along the nerve cord, showing that fasciculation with pre-existing axons is not necessary for accurate axon navigation by this set of neurons. 3.3.3. Does the selection of a specific axon pathway involve an absolute or relative choice? In the case of the aCC and pCC neurons, the

Fig. 5. The axon morphology of the motoneuron FETi in the grasshopper embryo following ablation of its target limb bud prior to axonogenesis. These two examples illustrate the variability observed in the pattern of axon branching from this motoneuron. Camera lucida drawings of intracellular LY fills. Scale bar = 50/im, 75% embryo (reproduced with permission from Whitington and Seifert, 1984).

Axon guidance factors in invertebrate development

12

experimental situation is much more extensive than that displayed by the G axon growing within the CNS. The difference correlates with the degree of inappropriate branching shown by axons growing in the CNS versus the periphery during normal development in this insect (see Section 2.3). The pattern of FETi axon growth in different experimental embryos is highly variable, both in terms of the degree of branching of the FETi axon and the particular routes followed (Fig. 5). As for the G neuron, the FETi axon appeared to follow virtually any other axon within reach. Similar conclusions were reached in a study of the growth of motoneurons in the nervous system of the fleshfly Sarcophaga bullata during metamorphosis (Nassel et al., 1986). The leg disc was removed from prepupae, thereby depriving the leg motoneurons of their target muscles. Operated flies examined after metamorphosis were missing the

leg nerve and the axons of leg motoneurons left the CNS via novel routes. The choice of alternative exit route was variable, with no indication of a strict hierarchy. The clear conclusion from these studies is that the selection of a particular axonal pathway involves a relative, rather than an absolute choice but that there is no strict hierarchy of preferences amongst the alternative axonal pathways. 3.4. "Guidepost" cells While selective fasciculation can play an important role in the growth of later-appearing axons along specific pathways, it cannot be responsible for the directed growth of the first axons, the pioneers, which navigate through an axonless environment. What cues direct these pioneer axons, and can the same cues also be used by later axons?

^HCV dKCvTSM(^);E-T

T?M; E-T

L3-V ; '. ; Acv

^^^ L3-2 L3-3

Fig. 6. The role of peripheral neuron somata in sensory axon guidance in the wing disc of Drosophila. (A-C) The sequence of axon growth from sensory neurons in the wing disc. (A) 6 h after pupariation (AP). (B)12 h AP. (C) 30 h AP. Axons from distal neurons contact the somata of more proximal neurons (e.g. L3-2 axon contacts L3-v soma) en route to the base of the wing disc. For example, arrow in (B) indicates where the axon of L3-2 contacts the L3-v soma. Camera lucida drawings of embryos stained using antiHRP inununohistochemistry. Scale bars: (A) 85 ^m; (B) 50yum; (C) 100/^m (reproduced with permission from Murray et al., 1984).

P.M. Whitington

One possibility is that axons navigate along a discontinuous path of "guidepost" cells. A variety of different cell types has been suggested to serve this function. 3.4.1. Neuronal somata Descriptions of neurogenesis and axonogenesis in the limb bud of the grasshopper embryo have revealed a stereotypic sequence of development of peripheral neuron somata and axons. It has been suggested that these somata function as a series of axon guidance cues (Bate, 1976; Keshishian, 1980; Bentley and Keshishian, 1982a,b; Ho and Goodman, 1982; Keshishian and Bentley, 1983a). For example, the growth cones of the Til neurons associate closely with the somata of the Trl neurons, located at the coxal/trochanteral border, then with the Cxi neurons, located just proximal to this border (Fig. 4, 20A). In a similar fashion, sensory afferent axons in the wing disc of Drosophila contact the somata of more proximally located sensory neurons during their navigation to the base of the appendage (Murray et al., 1984) (Fig. 6). Efferent motor axons in the grasshopper embryo are apparently guided into the limb bud by contact with the peripheral sensory Cxi neuron somata

Fig. 7. RP axons in the Drosophila embryo are guided into the connective by contact with their contralateral, homologous soma. During normal development, the axons of the RPl (A) and RP3 (B) motoneurons wrap around the soma of their contralateral homologue after crossing the midline and prior to entering the longitudinal connective, ac, anterior commissure; pc, posterior commissure. Scale bar = 5 ^m. Camera lucida drawings of LY intracellular fills. (A reproduced with permission from Sink and Whitington, 1991a, © Company of Biologists, Ltd.).

13

(Whitington, 1989). Neuronal somata may also function as guidepost cells for efferent axons in the leech. Braun and Stent (1989a) have shown that motor axons contact peripheral neuron somata during outgrowth and that the rate of growth of the motor axons transiently decreases when they contact these somata. Within the CNS, Sink and Whitington (1991a) have suggested that axons of the RPl and RP3 motoneurons in the Drosophila embryo are guided into the longitudinal connective by the somata of their contralateral homologues, with which they associate closely (Fig. 7) (see also Jacobs and Goodman, 1989b). A similar role in central guidance of the posteriorly directed MPl/dMP2 axons has been advanced for the somata of the aCC/pCC neurons in Drosophila and the grasshopper (Bastiani et al., 1986; Jacobs and Goodman, 1989b). The axons that pioneer the posterior transverse commissure in the Drosophila embryo have been reported to associate closely with the most anterior of the midline VUM neuronal somata (Klambt et al., 1991), and it has been suggested that contact with this soma may be important in guiding these axons across the midline (Fig. 8). Pioneering axons growing within the embryonic CNS of two crustaceans, the woodlouse and the crayfish, also associate closely with other neuron somata, although in a somewhat different fashion (Whitington et al., 1993). Central somata are closely packed in these embryos and the pioneering axons grow around the region of contact between cells, rather than over their dorsal surfaces (Fig. 9). A re-examination of figures of axon growth in the embryonic grasshopper CNS reveals a similar phenomenon in some cases (e.g. growth of the pCC axon around the soma of the aCC neuron, see Figs. 4 and 7 in Bastiani et al. (1986)). Early differentiating central neurons in the Drosophila and silverfish embryos also show the same behaviour (M. Murray, K.-L. Harris and P. Whitington, unpublished observations). Since, at early stages, soma arrangement is relatively constant in these species, this tendency of axons to grow between somata may contribute to their final stereotypic morphology. This would represent another

14

Axon guidance factors in invertebrate

pioneering of posterior commissure

pioneering of anterior commissure

migration of glia during commissure formation

development

separation of commissures

MGM MGA MP1

PC

VUMs

stage 12/5

stage 12/3

stage 12/0

stage 13

B MGM MGM MGA

VUM

MGA VUM

MGM

MGM MGA

otd

MGA

Fig. 8. The role of midline cells in the formation of commissures in the embryo of Drosophila. (A) Diagrammatic representation of the relationship between the midline cells in the Drosophila embryo and axons pioneering the anterior and posterior commissure at various stages of development. Dorsal view. The axons pioneering the posterior commissure (PC) grow towards the midline and contact the anteriormost of the VUM neuron somata. The axons then grow around the anterior side of the VUM cells and cross the midline. At the same time, the axons pioneering the anterior commissure (AC) extend toward and contact the posterior side of the MGA glia. Later in development, the AC and PC separate as the MGM glia migrate posteriorly. (B) Schematic representation of the formation of the AC and PC in wild-type embryos as viewed laterally, illustrating the processes described above. (C) In homozygous orthodenticle (otd) mutant embryos, the VUM neurons fail to differentiate and no posterior commissure forms. Many of the axons which would normally have extended in the PC now cross in the AP, which is therefore thicker than normal, (reproduced with permission from Klambt et al., 1991, © Cell Press).

15

P.M. Whitington

if

Fig. 9. The morphology of an identified interneuron in the embryo of the crustacean Porcellio scaber (woodlouse). The axon of this neuron passes between the regions of contact of neuron somata in its path, rather than growing over their dorsal or ventral surfaces. Photomicrograph of a LY intracellular fill. Scale bar = 15 ^m.

example of a specific axon morphology arising from a relatively non-specific pathway choice.

3,4,2. Ablation of guidepost neurons has varied effects on growing axons Experimental evidence that contact with neuronal somata plays an important role in axon guidance comes from a number of sources. Bentley and Caudy (1983) found that ablation of the Cxi neuron in the grasshopper limb bud causes aberrations in the growth of the Til axons: the latter form multiple axonal branches or project straight across this border instead of turning posteriorly (Fig. 10). The axons of the distally located sensory neurons of the subgenual organ normally contact the Til cell bodies immediately after they traverse the tibial/femoral segment border. If differentiation of the Til somata is prevented by heat-shocking the embryo, the subgenual axons fail to cross the tibial/femoral border (Klose and Bentley, 1989). These experiments show that filopodial contact with neuronal somata on the proximal side of segment boundaries in the locust limb bud is necessary if that border is to be traversed by the growth cone, a topic discussed further in Section 3.5. Sink and Whitington (1991c) found that abla-

Fig. 10. Effect of ablation of the Cxi neurons on axon growth from the Til neurons in the grasshopper embryo. (A) Operated leg in which the Cxi neurons had been killed at the onset of Til neuron axonogenesis and the embryo allowed to develop further in culture medium. Asterisk indicates debris of Cxi cells. The Til axons possess multiple abnormal branches (arrowheads). (B) Contralateral control limb for (A) The Til axons take a normal pathway to the CNS. (C) Operated limb bud. The Til axons project straight ahead (arrowhead) towards efferent axons from the CNS, rather than turning posteriorly towards the Cxi cell site. (D) Contralateral control limb bud for (C), showing a normal Til trajectory. Camera lucida drawings from anti-HRP immunohistochemistry preparations. Scale bar = 100/^m (reproduced with permission from Nature 304, © (1983) Macmillan Magazines Limited).

Axon guidance factors in invertebrate development

16

tion of the RPl or RP3 neurons in the Drosophila embryo caused defects in the growth of the axons of their contralateral homologues: the latter consistently failed to grow down the longitudinal connective, the path normally followed after contact with their contralateral somata. Some of the RP

nz3

Fig. 11. Effect of ablation of putative guidepost cells on outgrowth of axons from the CNS in the leech embryo. (A) Control hemisegment in an operated embryo. Axons of the UP nerve contact the nz3 and nzl,2 sensory neuron cell bodies when growing into the periphery. (B) Operated hemisegment in the same embryo as (A), in which the stem cells which generate the nz3 and nzl,2 neurons had been ablated earlier in development (arrowhead). The UP nerve is missing in this hemisegment. (C) Operated hemisegment in another embryo in which the stem cell for the nz3 and nzl,2 neurons had been ablated (asterisks) with the consequent absence of the UP nerve. (D) Control hemisegment in the same embryo as (C). The UP nerve is present, as are the nz3 and nzl,2 neurons. Scale bar = 20^m. Tracings from preparations in which axons and peripheral neurons were stained by injection of fluorescent dyes into early embryonic stem cells (reproduced with permission from Braun and Stent, 1989b).

neurons failed to form a lamellipodium, while others formed a short axon which grew down a variety of aberrant pathways. In a minority of cases, the RP axon followed its normal route into the anterior commissure, but in a more ventral position than normal. In another study of the Drosophila embryo, Klambt et al. (1991) found that genetic deletion of midline cells results in the failure of commissural formation, supporting the claim that these cells are crucial in guiding the pioneering commissural axons across the midline (Fig. 8). Braun and Stent (1989b) found that ablation of peripheral neurons in the leech embryo has a variety of effects on axon outgrowth from central neurons. Ablation of individual peripheral neurons that form part of a chain of closely spaced peripheral neurons with which efferent axons associate, has no effect on the formation of those efferent axons. However, ablation of peripheral neurons nz 1, 2, 3 and pz8, which lie along nerve paths where putative landmark cells are spaced widely apart, results in the failure of nerve path formation (Fig. 11). The conclusion from this study was that a variety of peripheral neurons normally provide guidance information for outgrowing nerves, but that this role is only a necessary one in situations where alternative cues are out of filopodial reach. The implication is that there is no intrinsic difference between peripheral neurons that can be experimentally shown to be necessary for axon guidance and those that are not. In contrast, other studies report that the presence of putative axon guidepost cells is not indispensable for axon growth along a normal trajectory. Using a combination of surgical and genetic ablation, Palka and co-workers (Schubiger and Palka, 1985; Blair and Palka, 1985) found that in Drosophila the absence of sensory neurons from the wing has no effect on the ability of several, more distal sensory neurons to navigate along a normal trajectory to the wing base. One sensory neuron, L3-v, does show an abnormal axon polarity in the absence of its normal neural neighbours and may therefore be dependent upon guidepost cells for reliable axon navigation. Similarly, the axons of the grasshopper Til neurons continue to show a normal proximal orientation in limb buds

P.M. Whitington

17

Fig. 12. The effect of ablation of the identified glial cell, SBC, on axon outgrowth from the motoneuron aCC in the grasshopper embryo. (A) Photomicrograph of the SBC cell in a control 45% embryo injected with LY to show its position with respect to the intersegmental nerve (ISN). Scale bar = 20 ^m. (B) A segment from an operated embryo in which the SBC cell on the right hand side had been ablated prior to the formation of the intersegmental nerve and the embryo allowed to develop further in culture. The axon of the aCC neuron on the operated side (EXP), injected here with LY, fails to turn into the intersegmental nerve and instead, grows further posteriorly down the connective. The aCC axon on the contralateral, control side (CON) extends into the periphery along its normal course, the intersegmental nerve (IS). Scale bar = 10 /^m (reproduced with permission from Bastiani and Goodman, 1986).

in which the more proximal guidepost neuron somata Fel and Trl differentiate later than normal, after the Til axons have passed their location (Caudy and Bentley, 1986c). Therefore, it appears that while guidepost neurons may be used as cues to guide the growth of later appearing axons, such neurons are not always necessary. Other axon guidance cues must be available in these situations. 3.4.3. Non-neuronal guidepost cells: primitive glial cells The strongest evidence that undifferentiated or "primitive" glial cells may be important in guiding axons comes from the grasshopper embryo. Bastiani and Goodman (1986) found that, as the axons of the motoneurons Ul, U2 and aCC, the pioneers of the intersegmental nerve, turn away from the longitudinal connective, they associate closely with the ventral surface of two large, flattened primitive glial cells, SBC and SBC2. When the SBC was mechanically ablated before arrival of

the U growth cones, the aCC axon failed to turn laterally to found the intersegmental nerve. Instead, it either stopped growing or continued to grow some distance posteriorly along the connective (Fig. 12). In the Drosophila embryo, a similar close relationship was reported between the growth cones of the RPl and RP3 motoneurons and the ventral surfaces of longitudinal glial cells and between the SBC, which lie dorsal to the developing longitudinal connectives, and the intersegmental nerve (Jacobs and Goodman, 1989a,b). The axons of the early differentiating Drosophila neurons, pCC, MPl, dMP2 and vMP2, contact the ventral surface of one of these longitudinal glial cells, LGX, as they begin to extend up and down the nerve cord (Lin etal., 1994). Initial growth of the axons that pioneer the longitudinal connectives is almost error-free in the Drosophila mutant glial cells missing (gem) which lacks longitudinal glial cells (Hosoya et al., 1995; Jones et al., 1995), although defects are seen in late

18

Axon guidance factors in invertebrate

embryos. On the other hand, Hidalgo et al. (1995) report that ricin induced ablation of the longitudinal glial cells leads to a failure of formation of longitudinal connectives in 71% of embryos. Menne and Klambt (1994) used a temperaturesensitive allele of the neurogenic gene Notch to delete midline glial cells in Drosophila and found a correlation between absence of these cells and MP133a

lOOum Fig, 13. The effect of ablation of the muscle pioneer cell MP133a on axon outgrowth from the Df motoneuron in the grasshopper embryo. (A) During normal development, filopodia of the Df growth cone enwrap the MP133a cell as the axon of this neuron diverges from the main nerve 5 pathway within the coxa. Scale bar = 20yum. (B,C) Operated embryos in which the MP 133a cell had been ablated prior to axon outgrowth from the Df neuron. The Df axon fails to diverge from the nerve 5 pathway in the coxa, and instead continues growing distally into the limb bud. The axon of the sibling of Df, Dfsib, also continues along the main nerve 5 pathway. Star marks the site of MP 133a ablation. The dotted line shows the path of nerve 5B1, which the axons of Df and DfSib do not follow in experimental embryos. Scale bar = 100/4m. Camera lucida drawings from Df neurons intracellularly injected with HRP (reproduced with permission from Ball et al., 1985).

development

failure of commissural formation. Midline neurons were also displaced to a more dorsal position in these mutants. While this result is suggestive of an axon guidance role for the midline cells, it remains to be shown that the effect is due solely to the defects in midline cells, rather than to some other effect of the Notch mutation. It is also unclear whether the effect is due to the absence of the midline glia or to the abnormal positioning of the midline neurons. 3.4.4, Non-neuronal guidepost cells: muscle cells Another class of non-neuronal cells, muscle cells or their precursors, has also been implicated in growth cone guidance in invertebrates. In the leech embryo, Kuwada and Goodman (1985) and Jellies and Kristan (1988) have reported the presence of large, flattened mesodermal cells in the vicinity of the CNS with which the axons of outgrowing motoneurons become closely associated. These cells have a number of characteristics that indicate that they may be precursors of muscles. Ablation of one of the cells, the axon runway cell, results in the absence of the nerve normally associated with that cell (Jellies and Kristan, 1988). In the grasshopper embryo. Ball et al. (1985) have found that, when the motoneurons Df and Ds turn away from nerve 5 towards their target muscle 133a, their growth cones become closely associated with an identified mesodermal cell, the 133a muscle pioneer, which appears to erect a scaffold for the formation of muscle 133a: a mass of mesodermal cells later fuses with this cell. Ablation of the 133a muscle pioneer results in failure of the Df motor axon to leave the nerve 5 pathway at its normal site; instead, this axon advances further into the embryonic limb bud (Fig. 13). Halpem et al. (1991) and Sink and Whitington (1991a) have also reported a close association between the axons of embryonic motoneurons and muscles in the Drosophila embryo. As the RP motor axons advance to their target muscles, they traverse three sheets of muscle fibres. Each motor axon branches extensively over these muscles, including non-target muscles. It thus appears as if the muscles in this sheet contain a guidance cue or cues, which can be recognised by all of the RP

P.M. Whitington

growth cones. A class of undifferentiated embryonic mesodermal cells in the Drosophila embryo, the PT cells, which are precursors of the adult muscles (Bate et al., 1991), has been implicated in motor axon guidance. The growth cone of the pioneering motomeuron aCC contacts three of these PT cells in the dorsal muscle field en route to its target muscle (Van Vactor et al., 1993). There is evidence from this same system that the target muscle plays a more specific role in signalling the motoneuron to withdraw axonal branches to other, inappropriate muscles (see Section 2.3). Ablation of the target muscles for the RP3 motoneuron causes that neuron to retain inappropriate branches well after the stage when these should have been withdrawn (Sink and Whitington, 1991b). Similarly, the motomeurons that innervate muscle 5 in the Drosophila embryo form functional synapses on adjacent muscle fibres following genetic and surgical ablation of their normal targets (Cash et al., 1992). Baptista and Macagno (1988) have found a similar phenomenon in the leech embryo. Elimination of supernumerary axonal branches of the RPE neuron depends upon contact with its target, the male genitalia. These workers also suggested that the target provides a signal to stop further growth of axonal branches, since in the absence of the target, the branching of the RPE neuron is more extensive than that seen at any time during normal development. The grasshopper motoneuron FETi and the Drosophila motoneuron RP3 often show a similar exaggerated branching behaviour when they are deprived of their target muscles (Whitington and Seifert, 1984; Sink and Whitington, 1991b) (Fig. 5). In the numb mutant, which lacks most of the ventral muscles, the Drosophila motoneuron RPl grows past its normal ventral termination site into the dorsal region of the body wall (Chiba et al., 1993). Thus the provision of a signal to halt further advance of the axon may be a general feature of targets. 5.5. Guidance cues on epithelial cells and/or associated basal laminae In a number of situations, axons are required to

19

navigate over epithelial cell sheets in the absence of other obvious cellular guidance cues, such as axons or neuronal guidepost cells. The default explanation is that the guidance cues are located on the epithelium. A pair of sensory neurons in the antenna of the grasshopper embryo, the DP cells, follows an axial trajectory towards the base of this appendage (Berlot and Goodman, 1984). There are no apparent guidepost neurons along this path. The BP cell, a neuron soma, which lies at the base of the antenna, and which might conceivably act as a guidepost neuron, is not within filopodial grasp of the DP growth cones. The DP axons remain in close contact with the epidermal cell surfaces during this axial growth and the guidance cues presumably reside within this epithelium. A set of motoneurons in the nematode C. elegans have dorsally projecting axons which grow along the epithelium of the body wall with its associated basal lamina (Hedgecock et al., 1990). Mutants have been isolated in which this growth is disrupted and analysis of these mutants provides evidence for an axon guidance factor in the basal lamina (see Section 5.2.1). The initial stretch of the trajectory pioneered by the Til neurons in the grasshopper limb bud is along an epithelial sheet overlain with a basal lamina (Keshishian and Bentley, 1983a) (Fig. 4). Based upon the increasing complexity of growth cone structure in more proximal parts of the limb bud region, Caudy and Bentley (1986b) proposed that there is a proximal-distal gradient of affinity of the substrate for sensory axons and that this causes the axons to grow in an axial manner. They suggested involvement of a gradient of adhesiveness for growth cones, but did not rule out other possibilities. The existence of a proximally increasing gradient of affinity of epithelia for sensory axons was indicated in a study in the moth pupal wing (Nardi, 1983). Following transplantation of patches of epithelium from distal to proximal regions of the wing, axons encountering the grafts failed to cross them. Grafts from proximal to distal regions were, however, readily crossed (Fig. 14). There appears to be a discontinuity in axon

20

Axon guidance factors in invertebrate

development

boundary, the growth cones of the sensory neurons must establish filopodial contact with the Cxi guidepost neuron somata just proximal to the boundary (Bentley and Caudy, 1983). The axon of the grasshopper SETi motoneuron follows a similar circumferential trajectory when encountering the base of the coxa, showing that motoneurons respond to this feature of the epithelium in a similar way to sensory neurons (Whitington, 1989).

Fig. 14. Effect of exchange of epithelial substrata on axon growth from sensory neurons in the wing of the moth Manduca sexta. (A) A piece of epithelium in the distal region of the pupal wing, corresponding to the dashed rectangle, was transferred to a more proximal region of the wing. Scale bar = 5 mm. (B) The trajectories of sensory nerves in the operated wing shown in (A) were examined at the adult stage and are indicated by dotted lines. The solid lines represent wing veins. Axons in the lp-3 and lp-2 nerve pathways which enter the graft fail to cross it, while lp-1 axons diverge posteriorly around the graft. (C) In this case, a piece of wing from the proximal part of the pupal wing was transferred to the distal wing region indicated. (D) Sensory nerves which encounter the proximal graft shown in C (IVa-6, IVa-5 and IVa4) cross it successfully. Drawings from methylene blue stained preparations (reproduced with permission from Nardi, 1983).

guidance cues on the epithelium at limb segment boundaries. When growth cones of sensory neurons in the grasshopper embryo encounter the coxa-trochanter border, they re-orient and grow circumferentially around this boundary (Caudy and Bentley, 1987). Cells on either side of the coxatrochanter boundary differ in a number of features to their neighbours, including morphology, presence of anti-horseradish peroxidase (HRP) immunoreactivity (Caudy and Bentley, 1986a) and expression of a number of proteins including annulin (Singer et al., 1992; Bastiani et al., 1992), fasciclin IV (Kolodkin et al., 1992) and glycosylphosphatidylinositol-anchored alkaline phosphatase (Chang et al., 1993). In order to cross this

3.5.7. Do guidance cues reside on the epithelial cells or on their basal laminae? There is some debate as to whether the factors directing axon growth over epithelia reside on the epithelial cells themselves or on the basal lamina associated with them. An electron-microscopic study showed that growth cones and filopodia of the Til neurons in the embryonic grasshopper limb bud contact the basal lamina, rather than the ectoderm of the limb bud (Anderson and Tucker, 1988). These authors concluded that guidance cues for these axons reside in the basal lamina, rather than the ectoderm. They also found that there is a proximal-distal gradient in basal lamina structure, with thinner basal lamina at the tip and thicker at the base (Anderson and Tucker, 1989). This correlates with the proximal-distal orientation of the pioneering Trl axons and it was suggested that this correlation may be a causative one for guidance of the Trl axons. Condic and Bentley (1989) have tested whether factors for guidance of the Til axons are located in the extracellular matrix (ECM) of the limb bud by removing the ECM with the enzymes, elastase or ficin. The majority of Til axons followed a normal trajectory in the absence of the basal lamina. In those cases where the Til axon morphology was abnormal, the Til soma had been displaced to an abnormal position. These findings show that the basal lamina is not essential for axon guidance in this system. However, they do not exclude an axon guidance role for the ECM during normal development. There may be guidance cues in both the ECM and the epithelium, which are redundant in respect of each other.

P.M. Whiting ton

Nardi and Vernon (1990) provided evidence that, in the moth, the epithelium of the wing disc rather than the basal lamina is involved in guiding axons, since the latter appears only after pioneering sensory axons have begun to grow towards the base of the wing bud. Topographical features of this substrate may contribute to the proximally directed growth of these pioneer sensory axons: the substrate area encountered by axons increases as they grow proximally, due to an increased number of basal processes on the epithelial cells in more proximal regions. On the other hand, the proximally directed growth of axons along the epithelium of the wing disc of Drosophila, which has also been shown to occur reliably in the absence of guidepost cells or other axons (Blair et al., 1985), does not depend upon simple physical factors such as free space. Separating the dorsal and ventral epithelia of the wing disc, which prevents the formation of veins and other physical channels, does not affect the growth of sensory axons. These authors concluded that a specific region of the wing epithelium must be chemically marked in some way that makes it a preferred substrate for axon growth. This affinity seems to be general for all sensory neurons, as sensory axons which are transplanted to the wing

21

disc from ectopic sites still follow this pathway preferentially (Blair et al., 1987). The basal lamina is also associated with axon growth in the CNS. Growth cones navigating in the CNS of the grasshopper embryo often contact the basal lamina which overlies the dorsal side of the CNS. When initial contact is made with the basal lamina, filopodia are extended in a planar array, forming a flattened sheet just under the basal lamina (du Lac et al., 1986; Whitington, 1989) (Fig. 15). Motor axons that exit via the segmental nerve in the grasshopper embryo then project laterally in a parallel bundle, just beneath the basal lamina (Whitington, 1989). Since the behaviour of all motor axons is the same when growing along the basal lamina, with growth in a consistent medial to lateral direction, their growth cones may be responding to a common cue distributed in a medial-to-lateral gradient on this structure. There is no need to invoke independent and specific guidance factors for each of the motor axons. In summary, there is strong evidence from observations of axon behaviour that factors located on epithelia and/or basal laminae can guide axons. Experimental studies have shown that the role of the basal lamina may be redundant when it is associated with an epithelium. 3.6. Guidance cues on trachea

Fig. 15. The growth cone of the motoneuron SETi in the grasshopper embryo extends a planar array of filopodia when encountering the basal lamina. Virtually all of the filopodia of this neuron, which is drawn from the dorsal perspective, lie in a single plane, just beneath the basal lamina. Camera lucida drawing from a neuron intracellularly injected with LY. Scale bar = 25 jum (reproduced with permission from Whitington, 1989, © Company of Biologists, Ltd.).

There is an emerging body of evidence that axon guidance cues are resident on the cells of the tracheal system in insects. In the Drosophila embryo, the growth cones of both motor and sensory axons of the intersegmental nerve adhere closely to the apical surface of the tracheal epithelial cells in the lateral region of the body wall (Hartenstein, 1988; Giniger et al., 1993; Van Vactor et al., 1993; Younossi-Hartenstein and Hartenstein, 1993). In the mutant wrong way (Kolodziej et al., 1995), axons of the lateral chordotonal sensory neurons make incorrect pathway choices at precisely those points along their growth trajectory where they would normally associate with trachea (K.-L. Harris and P. Whitington, unpublished observations). A variety of defective sensory axon morphologies have been reported in the mutants trachealess and

22

breathless, in which tracheal morphogenesis is disrupted (Giniger et al., 1993; Younossi-Hartenstein and Hartenstein, 1993), although it would appear from these studies that the trachea is not absolutely necessary for accurate sensory axon pathfinding. Nevertheless, the finding that the axons of the intersegmental nerve invariably follow residual tracheal material in the breathless mutant, however irregular that tracheal remnant might be (Giniger et al., 1993), suggests that the trachea is a preferred substrate for sensory axons. 3.7. Summary: what do cellular studies say about the molecular nature of axon guidance factors? When comparing cellular mechanisms for axon growth across a diverse selection of invertebrate phyla, from nematodes to annelids to arthropods, a remarkably consistent view emerges. Navigation of axons to their synaptic targets is well directed, but it is not invariably accurate. Branches are often sent down inappropriate pathways, to be retracted later in development to yield the final stereotypic projection pattern. No general explanation is available for why inappropriate branching occurs in one situation and not in another. However, it is possible that all neurons have the potential to form inappropriate branches and whether a branch is sent down a particular path depends upon the relative affinity of the growth cone for that path in comparison to the correct route. Such a model could explain why inappropriate branches vary in length, from lamellipodia to long axon collaterals. It might also suggest that different growth pathways are recognized by growth cones because they possess different combinations of a set of molecular labels, rather than each being labelled with a unique molecule. Many observations show that invertebrate axons are guided, in large part, by factors associated with the substrate in the immediate vicinity of the growth cone. In some situations, such as within the CNS of insect embryos, the axon appears to make highly specific choices of a growth substrate. In other situations, where the options available to a growth cone are limited by the physical availability of alternative routes (such as in the periphery of

Axon guidance factors in invertebrate development

insect embryos), relatively non-specific choices (e.g. neuron versus epithelia) could give rise to specific axon morphologies. It would certainly appear unwarranted to assume that each individual axon pathway bears a specific molecular label which distinguishes it from all other pathways of the same cell type. When axons fail to follow their normal pathways after ablation of guidance cues, they show a variety of responses. Axons that apparently navigate accurately during normal development, as for example those growing within the CNS of the grasshopper embryo, do not grow down alternative pathways when their normal guidance cues are ablated. Those that normally show inappropriate branching also explore aberrant routes in the absence of their normal guidance cues. Again, this behaviour supports a model in which choice of pathway is made on a relative, rather than an absolute, basis. A wide variety of cellular and extracellular structures has been implicated in axon guidance in invertebrates. Several of these may be used by a given axon at different points along its trajectory; for instance other axons, neuronal somata and affinity gradients on epithelia have all been implicated in guidance of sensory axons in the grasshopper limb bud. This means that when seeking to understand how neurons attain their final axon morphologies, it is necessary to dissect the overall growth process into its component parts. In particular, attention should be focussed on those points where the axon changes direction or substrate. The cellular and molecular mechanisms at work at one part of the axon trajectory may be entirely different to those at another part. No single cellular axon guidance cue appears to be unique to a particular invertebrate group or class of neuron and hence one cannot a priori exclude any specific cellular structure as the site of axon guidance molecules. Therefore, the cellular identity of guidance factors needs to be established anew every time one begins to work with a new system. Growth cones can be considered to be "opportunistic" in their choice of guidance cues: any structure in the environment of the growth

23

P.M. Whitington

cone can potentially influence its growth if it happens to bear molecules that can interact with receptors on the growth cone. Thus, some of the factors that guide a particular axon may have other developmental functions and may even be expressed in organs other than the nervous system. It would therefore be inadvisable to use expression outside of the nervous system, or expression outside of the region of the CNS where the molecules are thought to act, as a basis for eliminating candidates in a screen for axon recognition molecules. Indeed, almost all of the putative axon guidance molecules that have been characterized to date are also expressed in a variety of different tissues unrelated to the nervous system (see Section 5.1). A large number of transplantation studies in insects and leeches support the view that factors that guide a particular axon can be found in regions outside of the normal territories explored by that axon, where they presumably play other roles. These studies show that sensory axons entering foreign regions of the CNS, after transplantation of their somata to ectopic body positions, form central arborizations which are similar in morphology to those seen in their native locations (Passani et al. (1991) in the leech embryo; Sivasubramanian and Nassel (1985, 1989), Schmid et al. (1986) in fly during metamorphosis; Anderson and Bacon (1979) in adult grasshopper; Murphey et al. (1985) in larval crickets). In a number of cases, putative axon guidance factors have been shown to be essential for accurate axon navigation. In others, ablation of cells closely associated with the growth cone has little, if any, effect on axon growth, showing that such cells are redundant to other guidance cues. The task of characterizing axon guidance molecules is likely to be considerably more difficult in situations where redundancy of such factors exists at a cellular level. 4. How do guidance factors mediate directed axon growth? Cellular studies of axon growth in invertebrates have shown that axons acquire their stereotypic, mature morphologies by both accurate growth

cone turning and by more expansive initial branching followed by retraction of inappropriate branches. How do axon guidance factors elicit these cellular responses? An answer to this question will require an understanding of how the various organelles and molecular components of the growth cone interact to produce axon advancement and retraction. Our current understanding of this problem is limited and has come largely from studies of neurite growth in vitro in vertebrates. However, invertebrate systems have made some notable contributions. 4.7. A role for repulsion in axon pathfinding? A variety of studies in the vertebrates have provided a persuasive case for the involvement of repulsion in growth cone guidance (reviewed in Goodman and Shatz, 1993). Two main lines of evidence support such a mechanism: growth cones in tissue culture collapse after exposure to specific cell types (e.g. Kapfhammer and Raper, 1987; Cox et al., 1990; Schwab and Caroni, 1988), and surgical or genetic ablation of a structure from a region normally avoided by a specific set of axons leads to the growth of the axons across that region (e.g. Hatta et al., 1991; Bernhardt et al., 1992). If, indeed, axon repulsion is an important mechanism for growth cone guidance in vertebrates, it is highly likely that it also plays a role in invertebrates. To date, however, invertebrate systems have provided little direct evidence for axon repulsion in vivo. Kater and co-workers, working with identified neurons of the snail Helisoma in vitro, have shown that environmental factors can act to inhibit neurite extension rather than to stimulate it. The neurotransmitter serotonin, when added to the culture medium surrounding these neurons, causes an immediate cessation of neurite elongation (Haydon et al., 1984). This effect is associated with the direct inhibition of growth cone motile activities and the retraction of filopodia and lamellipodia, as shown by focal applications of serotonin to the growth cone with a micropipette. It has since been shown that the action of serotonin may be more complex, as it can stimulate neurite outgrowth if applied to

24

neurons which possess stable, non-elongating neurites (Goldberg et al., 1991). Reduction of endogenous serotonin levels by treatment of embryos with 5,7-DHT is associated with the development of aberrant neurite morphologies, specifically in neurons that are known to respond to serotonin in vitro (Goldberg and Kater, 1989). This result, taken together with the finding that serotonin shows a widespread distribution in the embryonic nervous system of Helisoma (Goldberg and Kater, 1989), supports the view that this chemical could play a role in the shaping of axon morphologies in vivo. However, whether the serotonin is released from the target cell for the neurons upon which it acts or from other cells along the pathway to the target, and whether it acts directly to determine the final morphology of these neurons, remain to be established. In a study of chemically induced mutations in Drosophila, Seeger et al. (1993) identified a mutant roundabout (robo) in which axons that remain on one side of the CNS in wild-type embryos, cross the midline. One interpretation of this phenotype is that it results from a deletion of a molecule which is normally repellent to certain axons that grow near the midline. Alternatively, the phenotype could result from the absence of cues that attract the axon in the correct direction. Deciding between these alternatives will require cloning of the robo gene and a detailed study of the localization and mode of action of its protein product. Another indication that repulsion may play a role in axon guidance in insects comes from the finding that H-Sema III, a human protein closely related to the family of insect semaphorins (see Section 5.2.1) is the likely homologue of the chick protein collapsin (Kolodkin et al., 1993). The latter protein has been shown to cause the collapse of growth cones of chick sensory neurons in vitro (Luo et al., 1993). While one insect semaphorin, G-Sema I (previously known as fasciclin IV) has been shown to be involved in the guidance of pioneer sensory axons in the grasshopper limb bud (Kolodkin et al., 1992), there is as yet no evidence that this function involves growth cone repulsion. The strongest evidence to date for axon repul-

Axon guidance factors in invertebrate development

sion in insects comes from a recent study (Nose et al., 1994) in which Drosophila embryos were transformed with a construct containing the entire open reading frame of the gene connectin (see Section 5.2.3) and the enhancer for the gene Toll, which is normally expressed in a subset of ventral longitudinal muscles. The resulting transformants show ectopic expression of connectin protein on those ventral longitudinal muscles. Motor axons that would normally innervate those muscles either bypass them, or make a detour and stall in the ventral region. Assuming that the response of the motor axons to connectin in an ectopic location faithfully mirrors the function of that molecule in its normal position (see Section 5.2.3), this result strongly supports a role for repulsion of motor axon growth cones in normal development. 4,2. Is differential filopodial adhesion the basis for growth cone turning? Observations of the behaviour of growth cones of vertebrate neurons in vitro in the early 1980s led to the hypothesis that growth cone turning is due to the selective adhesion of filopodia to different substrates in the path of the growth cone, with a subsequent mechanical pulling by the filopodia of the growth cone in the direction of high adhesion (e.g. Bray and Chapman, 1985). This model was widely accepted and heavily influenced interpretation of the pioneering observations of axon growth in insect embryos. Recent observations in a variety of animals, but notably in several invertebrates, have begun to cast doubt on this model as a sufficient explanation for growth cone turning. The mollusc Aplysia has been a favoured organism for in vitro studies of growth cone structure and function because several identified neurons of this species, such as the buccal ganglion neurons Bl and B2, form large, expanded growth cones in culture. Burmeister and Goldberg (1988) found that turning of growth cones at a border between preferred and non-preferred substrates was attributable to the selective regression of filopodia and lamellipodia on the non-preferred substrate, rather than to the selective attachment or advance of these structures on the preferred substrate. Such

25

P.M, Whitington

observations do not fit well with a model based exclusively on selective filopodial adhesion. Rather, they suggest that filopodial contact with favourable substrates might generate a signal leading to the selective survival and transformation of those filopodia into an neuritic branch, and the regression of filopodia on unfavourable substrates. One mechanism by which this could be accomplished is the differential partitioning of axonally transported cytoplasmic materials between branches. This possibility is supported by studies of the dynamics of advancement of Til sensory neuron growth cones in the grasshopper limb bud (O'Connor et al., 1990). It was observed that one mechanism for growth cone steering is the selective dilation of a single filopodium in contact with a guidepost cell. This single filopodium could thereby reorient an entire growth cone, despite the fact that many other filopodia remain in contact with non-selected substrates during turning of the growth cone. The filopodium thus seems to act as an "antenna" rather than a "tow-rope"; interactions between molecules on its surface and the substrate lead to the production of signals which ultimately cause the axon to grow in one direction, rather than another. The concept of selective filopodial adhesion as a model for growth cone turning has the added attraction that it provides a common mechanism for both directed growth cone turning and for expansive axon branching followed by retraction (see Section 2.3). In the former case, the appropriate substrate would be markedly dissimilar to any structure in the vicinity of the growth cone. Filopodial contact with anything other than the appropriate substrate would therefore lead to minimal flow of cytoplasmic materials into those regions of the growth cone and axon branch formation would be restricted to the appropriate route. In the latter case, alternative paths would be sufficiently similar in character to the correct route that substantial flow of materials and advancement of the axon in these directions takes place. However, such flow would not be as strong as that directed to branches running in the appropriate direction and these would therefore be starved of materials and eventually retract.

This model predicts that the formation or retraction of an axonal branch by a neuron on a particular substrate does not result primarily from competition with other neurons occupying that substrate. Rather, neurons behave autonomously, with competition taking place between different branches of an individual neuron. Indeed, targetdeprived motoneurons in the Drosophila embryo form and maintain connections with adjacent nontarget muscles, even though such muscles continue to receive innervation from their appropriate motoneurons (Cash et al., 1992). On the other hand, the model does not exclude the possibility that an independent mechanism, involving competition between the terminals of different neurons for occupation of synaptic space on target cells, might operate. An indication that competition between motor axon terminals might operate in the Drosophila embryo comes from experiments involving genetic or laser ablation of the RP3 motomeuron (Keshishian et al., 1993). Ablating the soma of this neuron, before its axon has arborized over its target muscle 6/7, results in the expansion of motor axon branches over muscle 6/7 from the foreign transverse nerve or from motor endings on adjacent ventral longitudinal muscles. 4.3. A role for receptor mediated signal transduction in growth cone navigation? Another important implication of this model is that the facilitation of mechanical adhesion of filopodia or lamellipodia to the substrate is not the only mechanism by which axon guidance factors could exert their effects. Rather, binding of such substrate-associated molecules to receptors in the plasma membrane of the growth cone could elicit a variety of intracellular changes, perhaps mediated by second messenger systems, which lead to axon advancement (see Doherty and Walsh, 1994 for a review of evidence in support of this concept from vertebrate in vitro systems). Such changes could include the extension of microtubules (Sabry et al., 1991) and the flow of vesicles from central to peripheral regions of the growth cone (Burmeister et al., 1988) as well as the breakdown of actin filament networks in the distal part of the growth cone

26

(Forscher and Smith, 1988). In this view, any molecule on the membrane of the growth cone that takes part in a receptor-ligand interaction leading to intracellular changes of the type referred to above could be considered to be an axon guidance molecule - a definition that encompasses a far wider variety of molecules than those traditionally considered to be cell adhesion molecules. Molecular components of signalling pathways are present in neurons in the Drosophila embryo at the time of axon outgrowth. Wolfgang et al. (1991) have shown that the a-subunit of the G protein is expressed at high levels in the neuropil of the Drosophila embryo at this time, and Tian et al. (1991) and Yang et al. (1991) report that a number of receptor-linked protein-tyrosine phosphatases are selectively expressed on CNS axons in the embryo. Such proteins may interact with receptors for axon guidance molecules on the nerve cell membrane and stimulate intracellular second messenger systems responsible for axon growth. Pulido et al. (1992) have identified a receptor tyrosine kinase, closely related to the mammalian trk neurotrophin receptor, which is expressed strongly in central and peripheral axons from early stages of axon growth. This molecule, gpl60^''^^ has structural homology with cell adhesion molecules of the immunoglobulin superfamily and promotes cell adhesion in a homophilic, Ca^^ independent manner. gpl60^^'^^, or a related molecule, could combine the dual functions of being a receptor for axon guidance molecules on surrounding cells and the first element in an intracellular signalling cascade. While the identity of "cell adhesion" receptor molecules remains unclear, there is considerable evidence that the intracellular signal transduction pathways which link reception of an extracellular signal with changes in growth cone behaviour involve changes in the level of intracellular Ca^"*" ([Ca^+]i). Again, invertebrate systems have made a significant contribution in this area. Identified neurons from the buccal ganglia of the snail Helisoma were placed into tissue culture and exposed to various treatments which resulted in a focal (Davenport and Kater, 1992) or global change in [Ca^+Jj (Rehder and Kater, 1992). Tran-

Axon guidance factors in invertebrate development

sient changes in [Ca^+li, monitored with the indicator fura-2, correlated closely with changes in filopodial morphology, involving an initial filopodial elongation, followed by retraction. A role for calcium in regulating axon growth is further suggested by the in vivo finding that pioneer neurons in the grasshopper limb bud which have begun axonogenesis show elevated [Ca^+]i' compared to those that have not (Bentley et al., 1991). Recently, VanBerkum and Goodman (1995) have performed an elegant series of experiments in which the function of the major intracellular receptor for calcium, calmodulin, was blocked in the growth cones of specific pioneering neurons in the CNS of the Drosophila embryo. This was accomplished by creating transformant flies in which the neurogenic control element of the fushi tarazu gene was used to drive expression of either a functionless form of calmodulin or a calmodulin antagonist. Expression of the protein was restricted to the growth cones by fusing the gene construct with the motor domain of the kinesin gene. In both types of transformants, the neurons expressing the fusion proteins showed stalls in axon extension and errors in axon guidance, fasciculating with inappropriate axons. Arguably, these results provide the most definitive evidence yet that calcium plays a central role in axon guidance in vivo. 5. The molecular basis for axon growth in the invertebrates Using the solid base of information about the cellular events underlying directed axon growth in invertebrates, a number of research groups have made concerted efforts to identify molecules involved in this process. Three main strategies have been used: generation of monoclonal antibodies using nervous systems as an immunogen and screening for antibodies that bind to restricted subsets of axons or nerve cells (e.g. Bastiani et al., 1987), using transposons to insert the lacZ gene randomly into the genome and screening for lines showing expression of the reporter gene in restricted subsets of axons or nerve cells (e.g. Nose et al., 1992) and undertaking a genetic screen for

P.M. Whitington

mutants showing aberrant axon morphologies (e.g. Seeger et al., 1993; Van Vactor et al., 1993; Salzberg et al., 1994; Kolodziej et al., 1996). Each approach has its own limitations but each has been successful in uncovering candidate axon guidance molecules. Recent reviews have described in detail the molecular features of these factors in insects (Goodman et al., 1991; Hortsch and Goodman, 1991; Harrelson, 1992), so this aspect is dealt with only briefly here. The likely biological roles of such molecules are discussed. 5.1. Molecules implicated in axon fasciculation The "labelled pathways" hypothesis, developed by Goodman and co-workers to explain the specific choices of axonal pathways made by growth cones in insect embryos (see Section 3.3.1), predicts the existence of a set of axon guidance molecules with a highly specific and restricted distribution on central and peripheral axonal pathways.

Fig. 16. The distribution of the putative axon guidance proteins, fasciclin I and fasciclin II in the grasshopper embryo. (A) Fasciclin I is expressed on subsets of axons in the anterior and posterior commissures (large arrowhead), as well as a few axons in the longitudinal connective (small arrow) and intersegmental nerve (small arrowhead). (B) Fasciclin II is expressed on large numbers of axons in the longitudinal connectives (small arrow) and the intersegmental nerve (small arrowhead, out of focus), but on few axons in the commissures (large arrowhead). Scale bar = 50yum. Photomicrographs of preparations stained with monoclonal antibodies directed against the respective fasciclin proteins (reproduced with permission from Bastiani et al., 1987, © Cell Press).

27

Several workers have searched for molecules that meet these criteria and these searches have resulted in identification of a number of candidate axon guidance molecules, the best characterized being the fasciclin family. 5.1.1. The fasciclins Using a monoclonal antibody screen, Bastiani et al. (1987) identified two glycoproteins which are expressed on subsets of axons in the CNS of the grasshopper embryo. Fasciclin I appears on the surface of a subset of axons in the anterior and posterior commissures, as well as on a group of neuron cell bodies and their axons which grow out the peripheral nerves (Fig. 16). The homologous gene and its product in Drosophila was subsequently identified and shown to have a similar pattern of expression to that in the grasshopper (Zinn et al., 1988). The deduced amino acid sequence of the gene product predicts that it is an extrinsic membrane protein, possibly linked to the membrane by a phosphotidylinositol (Pl)-lipid anchor. Transfection of non-adhesive S2 cells in tissue culture with fasciclin I cDNA causes these cells to aggregate with other cells expressing the same gene, suggesting that the protein can mediate homophilic cell adhesion. These features of the protein are consistent with a role in axon guidance although, as argued in Section 4.3 above, the ability to mediate cell adhesion may not be a necessary property of an axon guidance molecule. Fasciclin II is expressed on a subset of longitudinal axons in the connective and the intersegmental nerve (Harrelson and Goodman, 1988 (Fig. 16). Some axons appear to express fasciclin I within the commissures and fasciclin II when they turn into the longitudinal connective (Grenningloh et al., 1990). The appearance of fasciclin II expression on individual axons correlates with the switch in growth along non-neuronal substrates (glial cells) to axons. Again, the Drosophila homologue to fasciclin II has been identified and shown to be expressed on a homologous set of neurons, although with differences in the relative timing of expression which parallel differences between these species in the timing of axon outgrowth (Grenningloh et al., 1991; Lin et al., 1994). Both

28

the grasshopper and Drosophila gene products show extensive sequence similarity to vertebrate neural adhesion molecules in the immunoglobulin superfamily, being most closely related to N-CAM (Harrelson and Goodman, 1988; Grenningloh et al., 1991). Two forms of fasciclin II have been described: one has a transmembrane domain, while the other appears to be linked to the cell membrane by a phosphodylinositol anchor (Fig. 17). These two products are derived from the same gene by alternative splicing. The pattern of expression and molecular analysis of the fasciclin I and fasciclin II proteins suggest that their most likely role is to enable the recognition of a particular fascicle (or fascicles) within the anterior commissure (fasciclin I) or the longitudinal connectives (fasciclin II). This suggestion is supported by the result of antibody blocking experiments and mutant analysis. When antibodies against fasciclin II are applied to the grasshopper embryo, and the embryo allowed to develop further in a tissue culture medium, the initial growth of the MPl axon along non-neuronal substrates is not affected. However, when the antibodies are applied just before, or after the MPl axon fasciculates with the MPl/dMP2 fascicle, the posterior growth of the axon is stalled and the axon often explores other axon fascicles, such as the transverse commissures (Harrelson and Goodman, 1988). It was originally reported in null mutants for the fasciclin II gene in Drosophila that the vMP2 axon begins to grow anteriorly, but then stalls before it reaches the next segment (Grenningloh et al., 1991). The MPl/dMP2 axons fail to grow posteriorly along the longitudinal pathway; instead they extend a short distance laterally. It has subsequently been reported that this is an artefactual phenotype, resulting from the storage of the mutant embryos at low temperatures prior to fixation (Lin et al., 1994). A revised analysis of iht fasciclin II loss-of-function phenotype reveals that the timing and polarity of initial axon extension is normal, but that there are defects in axon fasciculation (Lin et al., 1994). Specifically, the vMP2, dMP2, MPl and pCC axons, which fasciculate

Axon guidance factors in invertebrate

development

I neuroglian

fasciclin ii

immunoglobulin domain

fibronectin - ^ type III domain

Fig. 17. The structure of the neuroglian and fasciclin II and III proteins, as deduced from their DNA sequences. Each possesses several immunoglobulin domains, a protein structure found in a variety of cell adhesion and immunoglobulin molecules in vertebrates. The neuroglian and fasciclin II proteins also possess a number of fibronectin type III domains. All three proteins exist in a transmembrane form with a cytoplasmic domain, while fasciclin II also appears to exist in a phosphotidylinositol-linked form (reproduced, with permission, from the Annual Review of Cell Biology, Vol. 7, © 1991 by Annual Reviews Inc.).

closely with each other in wild-type embryos, remain defasciculated in the mutant (Fig. 18). The more lateral FN3 bundle of axons, which also normally expresses Fasciclin II, is similarly affected. Furthermore, while the vMP2, dMP2, MPl and pCC axons extend along the space beneath the LGX glial cell, as in wild-type embryos, they fail to adhere to its ventral surface. An analysis of gain-of-function mutations, generated by using the fushi-tarazu neurogenic control element to drive enhanced expression of Fasciclin II protein in central neurons, shows that the FN3

P.M. Whitington

29

Fig. 18. Axon growth from neurons pioneering longitudinal axon pathways in the CNS of wild-type and fasciclin II null mutant Drosophila embryos. (A) Schematic diagram of a dorsal view of pioneering neurons in a wild-type embryo at embryonic stage 13. The posteriorly projecting MPl and dMP2 axons fasciculate with the anteriorly projecting axons from pCC and vMP2 in regions 1 and 2. All of these axons adhere to the ventral surface of the longitudinal glial cell LGX. (B) Fasciclin II null mutant at the same stage as (A). MPl and dMP2, pCC and vMP2 neurons initiate and send axons in the normal direction but these fail to fasciculate properly. (C) Cross-sectional views of the CNS in wild type embryos, showing the close fasciculation between the dMP2, pCC and MPl axons in region 1, between these axons and the LGX cell in region 2 and between the pCC and vMP2 axons in region 4. (D) Cross-sectional views of afasciculin II mutant, showing the failure of fasciculation between these axons (reproduced with permission from Lin et al., 1994, © Cell Press).

fascicle often fuses with the MPl fascicle. In addition, the defasciculation of the MPl/dMP2 axons from the pCC/vMP2 axons which is seen in wildtype embryos within the segment, often fails to occur in the gain-of-function mutant (Lin et al., 1994). These results show thai fasciclin II is indeed involved in axon fasciculation but is not responsible for the direction of axon outgrowth or the polarity of axon extension along the nerve cord. Furthermore, fasciclin II is clearly not the sole determinant of axon fasciculation for all neurons, as other central neurons which express Fasciclin II protein strongly in wild-type embryos, such as the motomeurons aCC and the three U neurons, do not fasciculate with the MPl fascicle.

Further clues as to the function of the Fasciclin II protein have come from a study of the effects of ectopic and increased expression of the protein on motor growth cone guidance in the body wall of the Drosophila embryo (Lin and Goodman, 1994). A GAL4 effector strain was used to drive expression of Fasciclin II on all embryonic neural tissues, including peripheral motor and sensory axons and sensory cell bodies. Motor axons, all of which express Fasciclin II in wild-type embryos, show a wide variety of aberrant growth responses in these transformed embryos including; failure to defasciculate from the nerve bundle at normal positions; delayed de-fasciculation, such that axons enter their target muscles from a more distal trajectory than normal; and stalling of growth.

30

These results are consistent with fasciclin II mediating axon-axon fasciculation in the PNS of wild-type embryos, in the same way as it appears to act in the CNS (assuming that the response of axons to increased levels of Fasciclin II is simply an exaggerated form of their response to wild-type concentrations of this protein). A number of observations in Lin and Goodman's (1994) study suggest that, as in the CNS, fasciclin II does not act alone in this regard. These include the facts that in many segments axon trajectories are apparently normal, that the motor axon growth errors observed are corrected at later stages of embryonic development and that the motor axons do not fasciculate with sensory axons or cell bodies in the transformed embryos, even though they express high levels of Fasciclin II. Taken together with the observation that fasciclin II null mutant embryos do not display obvious axon growth defects, these results suggest that fasciclin II normally acts in concert with other, as yet unidentified genes to mediate axon fasciculation in the periphery and that the actions of these gene products are redundant with respect to each other. Null mutants of the fasciclin I gene have been generated, but show no obvious abnormalities in axon growth in the CNS during embryogenesis. Whitlock (1993) has reported a slight increase in overall branching of Drosophila wing sensory neurons within the adult CNS of fasciclin I null mutants, although the central branching of embryonic sensory neurons is indistinguishable from wildtype animals (D.J. Merritt and P. Whitington, unpublished observations). On the other hand, double null mutants for/a5'ciclin I and the abl gene, which encodes a tyrosine kinase, shows gross defects in axon growth for the identified motoneuron, RPl (Elkins et al., 1990b). In 83% of double mutant embryos, this neuron extends an axon down aberrant routes. Since abl null mutants alone show no obvious neural defects, it appears that these two gene products effect axon guidance by parallel pathways, which are functionally redundant. If correct, this interpretation implies that the action of the fasciclin I protein may be more than simply cell adhesion, because the abl gene product, to which the fasciclin I pro-

Axon guidance factors in invertebrate development

tein is apparently redundant, forms part of a signal transduction pathway and is not a cell adhesion molecule. Keshishian and co-workers have used thermal denaturation of fasciclin I protein by laser irradiation of embryos treated with anti-fasciclin I antibodies, to determine the role of this protein in the grasshopper limb bud (Jay and Keshishian, 1990). Fasciclin I protein is normally expressed by the pair of Til peripheral neurons in the limb. When the fasciclin I protein is denatured by laser irradiation, the sibling Til axons often grow separate from one another, rather than tightly fasciculated. However, they still navigate correctly in a proximal direction, recognise Cxi cells and cross segment boundaries. Therefore, Til axon fasciculation is dependent upon fasciclin I expression, but the protein is apparently not necessary for directed axon growth. It is unclear why axons in the CNS of the Drosophila embryo do not show a similar dependence on fasciclin I activity for normal fasciculation. This may be a consequence of the different techniques used to inactivate the molecule in the two studies, or may reflect a different role of the protein in the CNS versus the periphery. A subsequent monoclonal screen in the Drosophila embryo uncovered another glycoprotein, Fasciclin III, whose pattern of expression is also suggestive of a role in axon fasciculation (Patel et al., 1987). Fasciclin III is expressed on a small subset of neurons, including the growth cones, axons and cell bodies of the RP neurons, and some axons in the anterior and posterior commissures. The Fasciclin III protein is composed of three immunoglobulin domains which are much more divergent than those found in fasciclin II (Fig. 17). As mth fasciclin /, transfection of non-adhesive S2 cells in tissue culture with the fasciclin III gene causes these cells to adhere to other cells expressing the gene, suggesting that the protein can mediate homophilic cell adhesion (Elkins et al., 1990a). However a null mutant for the fasciclin III gene shows no obvious defects in the organization of the embryonic CNS (reported in Elkins et al., 1990b) and the central projections of embryonic sensory neurons in this mutant are indistinguish-

P.M. Whitington

able from wild-type embryos (D.J. Merritt and P. Whitington, unpublished observations). Whitlock (1993) has reported a dramatic alteration in the pattern of wing axons of fasciclin III null mutants at the adult stage. Sensory axons that normally run in a medial tract of the CNS apparently deviate to a more lateral tract, along a pathway that is present in normal animals. Monoclonal antibody screens on grasshopper embryos uncovered a protein which, like Fasciclin I, II and III, is expressed on subsets of axons in the embryonic CNS. It was originally christened Fasciclin IV, but has more recently been renamed Semaphorin, as it has become clear that it differs in a number of fundamental respects from the Fasciclins (Kolodkin et al., 1992). The semaphorins are discussed in Section 5.2.1, 5.1.2. Is the carbohydrate component of glycoproteins involved in axon guidance? Many of the proteins which have been implicated in axon guidance in both vertebrates and invertebrates (including all of the fasciclin proteins discussed above) possess carbohydrate moieties. What is the evidence that this carbohydrate component is important in axon guidance? One line of evidence has come from the leech embryo. Sensory axons growing into the CNS of this animal arrive as a tightly fasciculated bundle and subsequently defasciculate into the synaptic neuropile. One of these afferents expresses a 130 kDa surface glycoprotein. Exposing embryos to Fab fragments of antibodies against this protein leads to failure of the sensory afferents to defasciculate (Zipser et al., 1989). A similar effect is produced by cleaving the asparagine-linked carbohydrate moieties from surface proteins with the glycosidase N-glycanase, or by competing for a putative mannose-binding protein with the neoglycoprotein mannose-bovine serum albumin. These experiments indicate the presence of a mannosebinding protein in the neuropile region explored by the sensory afferents. The identity and cellular localization of this protein remain to be established. A number of proteins expressed on CNS and PNS axons in insects, including Fas I and Fas II (Snow et al., 1987), neurotactin and neuroglian

31

(see Section 5.3) and two receptor-linked tyrosine phosphatases, DPTPIOD and DPTP99A (Desai et al., 1994) possess a carbohydrate moiety which is recognized by anti-HRP antibodies. The tissue distribution of the proteins bearing this carbohydrate group might suggest that it plays some function in axon guidance. The mutation nac, which eliminates expression of this epitope in imaginal tissues (Katz et al., 1988), displays severe misroutings in the projection of wing sensory neurons into the CNS: a medial axon tract is missing (Whitlock, 1993). On the other hand, another mutation which removes the carbohydrate epitope in embryos has no obvious effect on the structure of the CNS (cited in Desai et al., 1994). 5.1.3. Genes involved in axon fasciculation in C. elegans A study of the phenotype of mutants of the nematode worm C. elegans with uncoordinated motor function {unc mutants) has uncovered three genes involved in the guidance of axons along the longitudinal nerve cords (Mclntire et al., 1992). The neurons which were examined, the HSN motoneurons, normally extend anteriorly along the ventral nerve cord, following other axons. In the unc-34, unc-7I, and unc-76 mutants, the HSN axons terminate anterior growth at variable positions within the ventral nerve cord (Fig. 19). A number of other anteriorly projecting axons in the ventral cord are similarly affected, as are the longitudinally projecting processes of the DD and VD motoneurons in the dorsal nerve cord. These defects in longitudinal elongation are associated with abnormalities in axon fasciculation: single axon bundles in the dorsal and ventral nerve cords of wild-type animals are split into separate bundles in the mutants. The primary defect in these mutants may be a failure to fasciculate with other axons and this may result secondarily in cessation of longitudinal extension, as occurs for longitudinally projecting axons in the CNS of insect embryos (see Section 3.3.2). The relevant gene products have not yet been characterized. However, it is significant that the axons continue to extend for some distance before they cease growth. This may suggest that

32

guidance cues on other axonal pathways bear some similarity to the product normally provided by the mutant gene. Furthermore, it is apparent that these genes encode relatively non-specific axon guidance factors, as the growth of several axons is similarly affected in each mutant. Mclntire et al. (1992) speculate that genes responsible for guiding axonal growth of a single neuron type do not exist. All of the 11 mutants which show defects in DD and VD outgrowth also show defects in the outgrowth of the HSN and other neuron types. Conversely, none of the 35 identified genes which affect the HSN neurons specifically disrupts axonogenesis of this neuron type (Desai et al., 1988). If this hypothesis is correct, specificity of pathway choice by axons is likely to depend on the action of unique combinations of broadly expressed molecules. An opposing argument is that genes directing growth of single neuron types do exist, but that these are not revealed by mutant analysis because they are functionally redundant. 5.2. Molecules responsible for axon guidance on non-neuronal substrates As discussed in Sections 3.4 and 3.5, axons can grow along a variety of substrates other than preexisting axons. What are the molecules that guide growth cones along non-axonal substrates? Are these factors the same as those which guide axons along axonal substrates? Mclntire et al. (1992) have provided genetic evidence that at least some of the genes involved in axon growth along epithelial substrates in C. elegans (see Sections 3.5 and 5.2.1) are different to those involved in fasciculation with other axons, and work on the identification of non-axonal guidance molecules is proceeding in a number of different systems. 5.2.1. Epithelia or extracellular matrix The cellular studies outlined in Section 3.5 point to a role for factors in the extracellular matrix, and particularly the basal lamina, in axon guidance. A number of studies in invertebrate species have begun to characterize molecules that may subserve axon guidance on such substrates.

Axon guidance factors in invertebrate

development

wild type

longitudinal-elongation defective

B Fig. 19. The morphology of the HSN motoneurons in wildtype and unc-76(e911) mutant adult Caenorhabditis elegans worms. The axons of the HSN motoneurons terminate prematurely in the ventral nerve cord in the mutant (B), before reaching the nerve ring. Schematic diagrams from immunofluorescence photomicrographs of worms stained with antiserotonin antibodies, (reproduced with permission from Mclntire et al., 1992, © Cell Press).

Wang and Denburg (1992) found that treatment of cockroach embryos with exogenous heparin or heparan sulphate, or with the glycosaminoglycandegradative enzymes heparinase II and heparitinase, causes defects in the trajectories taken by the Til axons in the limb buds. Binding sites for heparin and heparan sulphate are found on the basal lamina and the epidermis of the limb bud. The same molecular species are implicated in axon recognition of guidepost cells, since enzyme treatment causes the axon of the central neuron MP4, which pioneers the medial axon tract, to grow in aberrant directions. A large midline cell at the point of bifurcation of the MP4 axons possesses heparan sulphate- and heparin-binding sites. One factor involved in guidance of pioneering axons along an epithelium in C. elegans has been characterized at the molecular level. In unc-5 and unc-6 mutants, motor axons which normally grow dorsally between the lateral body wall hypodermis and its associated basal lamina grow in aberrant oblique or longitudinal directions, and fail to reach the dorsal nerve cord (Hedgecock et al., 1990; Mclntire et al., 1992). In unc-6 and unc-40 mutants the HSN axons, which normally grow in a ventral

P.M. Whiting ton

direction along the epithelium to the ventral nerve cord, project in an anterior direction along the lateral epidermal edge of the animal. Analysis of the phenotypes of double mutants reveals that all known unc-d functions require unc5 gene function, and conversely, all unc-S functions require unc-d gene function, suggesting that these gene products act in a common path or as a single complex in dorsal axon guidance. The unc-5 gene may encode a ligand on the epithelial surface or basal lamina while the unc-6 gene may encode a receptor for that ligand on the motor axons, or vice versa. The unc-6 gene has now been shown to encode a novel secreted protein related to the B2 sub-unit of laminin, a major component of basal laminae (Ishii et al., 1992) while unc-5 encodes a novel cell adhesion receptor of the immunoglobulin superfamily (Leung-Hagesteijn et al., 1992). A simple model for axon guidance in this system is that UNC-5 protein is a receptor on the growth cones of neurons, and that an interaction between this receptor and a dorso-ventral gradient of UNC-6 ligand guides the growth cones dorsally. Analysis of genetic mosaics shows that the wild-type unc-5 gene must be expressed in the neurons for normal dorsally-directed axon growth to occur (LeungHagesteijn et al., 1992), a result which is consistent with this model. Furthermore, ectopic expression of UNC-5 receptor in touch receptor neurons causes them to extend axons dorsally, rather than following the normal longitudinal or ventral trajectories, and these abnormal phenotypes depend upon unc-6 expression (Hamelin et al., 1993). The Fasciclin IV protein, which was originally isolated from a monoclonal screen in grasshopper embryos (Kolodkin et al., 1992), also appears to be involved in the guidance of axons over an epithelial substrate. Grasshopper embryos cultured in the presence of anti-Fasciclin IV protein exhibit aberrant axon morphologies of the pioneer sensory neuron Til. This axon normally turns ventrally at the trochanter-coxa boundary, along a stripe of epithelial cells that express Fasciclin IV (Kolodkin et al., 1992). Studies of the dynamics of growth cone advance in this region suggest that the growth cone is guided by cues associated with the epithe-

33

lial substrate (O'Connor et al., 1990). In the presence of blocking antibody, the Til axon develops multiple, ventrally directed axon branches distal to the trochanter-coxa boundary (Kolodkin et al., 1992). This suggests that Fasciclin IV normally functions to constrain growth to the trochantercoxa boundary but that other, as yet unidentified cues, are responsible for the tendency to grow in a ventral direction. Cloning of the grasshopper fasciclin IV gene reveals that it encodes a novel, integral membrane glycoprotein which appears to mediate cell adhesion in vitro in a heterophilic, rather than a homophilic, fashion (Kolodkin et al., 1992). A family of proteins related to Fasciclin IV has been cloned in a range of insect and vertebrate species. These proteins, renamed the semaphorins (Kolodkin et al., 1993), share a conserved extracellular domain of approx. 500 amino acids and a signal sequence, suggesting they are either transmembrane or secreted proteins. Some members of the family possess a transmembrane domain with a modestly conserved cytoplasmic domain, while others possess a single Ig domain instead of this transmembrane domain. As noted in Section 4.1 above, the human semaphorin gene shows strong sequence similarity with the chick collapsin protein. Mutations of the semaphorin genes in Drosophila provide few definitive clues as to their function in this animal, beyond the fact that they are apparently required for certain aspects of visuallyguided behaviour and survival at the adult stage (Kolodkin et al., 1993). As the semaphorin proteins are expressed widely within the CNS of the Drosophila embryo, they may be involved in aspects of axon pathfinding and synaptogenesis unrelated to their demonstrated role in the grasshopper of navigation along epithelial substrates. The slit protein of Drosophila is another extracellular factor that may be involved in axon guidance. This protein is expressed in a subset of glial cells in the midline of the embryonic CNS (Rothberg et al., 1990). Slit protein is secreted by Schneider S2 cells in tissue culture and it has been suggested that it is secreted by the midline glial cells and binds to proteins in the ECM. The sequence of flanking regions to four leucine-rich

34

repeats present in the slit protein, as well as the presence of two epidermal growth factor-like repeats, are supportive of this idea. Null mutants of the slit gene show defects in the formation of transverse axonal pathways that cross the midline. Such defects might suggest a role for extracellular slit protein in guidance of axons across the midline of the CNS in wild-type embryos. Alternatively, the defects could result from the absence or abnormal positioning of midline glial cells which is also a feature of the slit null mutant phenotype (see also Section 3.4.3). The problems implicit in establishing a link between gene activity and axonal pattern is illustrated by the case of tht faint little ball (fib) mutation. Null mutants of this gene, which encodes the Drosophila homolog of the epidermal growth factor receptor protein (DER), show defects in the morphology of axon pathways in the CNS. The DER protein is expressed on midline glial cells and it was originally suggested that it may be a receptor involved in the interaction of pioneer axons crossing the midline with midline glial cells (Zak et al., 1990). Indeed, Rothberg et al. (1990) suggested that the DER protein may be the receptor for the slit molecule. More recently, it has been shown that the fib gene functions at several different times during development, including the phase of neural determination (Raz and Shilo, 1992). The abnormal axonal phenotype seen in the fib null mutant may be due, at least in part, to a failure of these determinative cell interactions and the subsequent absence of certain neuron types, rather than to a direct effect upon the glial cells which form the substrate for commissural axon growth. 5.2.2. Glia The cellular studies discussed in Section 3.4.3 provide evidence for a role for glia in guiding axons in insect embryos. To date, however, there are no well-documented examples of molecules that may subserve this function. The case of the DER protein in the Drosophila embryo has been discussed above. Some proteins, such as fasciclin in (see Section 5.1.1) and neuroglian (Section 5.3), are expressed on glia as well as on neurons, but their glial function has not been determined. Oth-

Axon guidance factors in invertebrate development

ers appear to be specific to glia. The REGA-1 antigen is expressed on a subset of glial cells in the CNS of the grasshopper embryo (Carpenter and Bastiani, 1991). The protein, which has a molecular weight of 60 kDa, is expressed from the earliest stages of axonogenesis on lamellar processes which line the edges of certain axonal pathways in the CNS. Its spatio-temporal pattern of expression indicates that it may serve to delineate boundaries of axon growth. The connectin gene (see Section 5.2.3) has recently been found to be expressed strongly in longitudinal glial cells in the CNS of the Drosophila embryo, at the time when pioneering central axons are exploring their surfaces (Meadows et al., 1994). However, embryos cultured in the presence of an anti-connectin antibody do not show any obvious defects in axon patterning in the CNS. 5.2.3. Muscles Cellular studies indicate that the peripheral muscle targets for motoneurons play a variety of roles in shaping the axon morphology of those neurons: causing motor axons to diverge away from a fascicle, providing a substrate for axon advancement, and signalling motoneurons to withdraw inappropriate branches. Studies of the molecular basis for these phenomena are in their infancy. Nose et al. (1992) have identified a protein, connectin, which is expressed on a subset of muscles and the motoneurons that innervate them, in the Drosophila embryo. The connectin gene has been cloned and found to encode a signal sequence, ten leucine-rich repeats, and a putative phosphodylinositol membrane linkage. The protein can mediate homophilic cell adhesion in S2 tissue culture cells. Expression of connectin correlates temporally with the formation of connections between these motoneurons and their muscles. A loss-of-function mutation in the connectin gene does not result in any apparent defects in neuromuscular development (Nose et al., 1994). However, ectopic expression of connectin on ventral longitudinal muscles of the embryo, driven by a Toll enhancer element, produces clear defects in axon growth by SNb motomeurons over these

P.M. Whitington

muscles, their normal synaptic targets. The identified motomeurons RP3 and RPl often stall at the point where they normally leave the intersegmental nerve to branch over the most ventral muscles in this group. Alternatively, they may grow past this branch point, following the ISN, then either branch into the ventral muscle group from a more distal position or continue advancing along the ISN. Interestingly, in those cases where RP3 manages to find its target muscle 6/7, it ultimately forms normal synaptic connections, despite the fact that the muscle continues to express high levels of connectin. As discussed in Section 4.1 above, these findings are consistent with a repulsive role for connectin in preventing SNb motor axons from branching over muscles that express it. However the relevance of this effect to normal developmental events is unclear, since many of the connectin expressing muscles are in regions of the body wall not normally explored by the growth cones of the SNb motomeurons (Sink and Whitington, 1991a). In addition, the fact that the loss-of-function connectin mutations do not result in any overt axon growth defects shows that connectin expression is not the sole determinant of normal motor axon growth. As yet unidentified axon-attractive factors, working in competition with connectin-mediated repulsion, may play a vital role. The growth cones of the Drosophila motoneurons RPl, RP3 and RP4 and their muscle targets express a variety of membrane proteins, including Toll, fasciclin I and fasciclin III (Halpem et al., 1991; Keshishian et al., 1993). However, null mutants or extreme hypomorphs for these genes fail to show errors in axon growth or neuromuscular connectivity either in single or double mutant combinations. 5.3. Putative axon guidance proteins of widespread distribution A number of proteins have been identified in insect embryos which have structural features suggestive of a role in axon guidance but which, unlike the fasciclins, show a widespread distribution within the nervous system. It should be noted that

35

a highly restricted pattern of expression may not be a necessary criterion for an axon guidance molecule. As discussed in Section 3.7, where the options available to a growth cone are limited by the physical or temporal availability of alternative routes, a relatively widely distributed guidance factor could be responsible for specific axon morphologies. Furthermore, two molecules, each distributed widely but overlapping in a restricted fashion with each other could, by acting in combination, specify a restricted growth trajectory. Piovant and Lena (1988), de la Escalera et al. (1990) and Hortsch et al. (1990) have identified a protein, called neurotactin, which is apparently expressed on all neurons in the CNS of the Drosophila embryo. In the peripheral nervous system, the neurotactin protein is expressed strongly on a subset of sensory neurons, particularly those with multiple dendrites. Neurotactin is an integral membrane protein with an extracellular domain which has homology to serine esterases, although the serine residue characteristic of the active site of known esterases is missing. The role of this esterase activity in axon guidance, if indeed it is possessed by the molecule, is unclear. The extracellular domain of the protein also has three copies of the leucine-arginine-glutamate motif which forms part of the adhesive site for s-laminin and may thus interact with the extracellular matrix. Neurotactin has been shown to mediate adhesion of S2 cells in vitro although, unlike the fasciclin proteins, this adhesion seems to proceed in a heterophilic manner (Barthalay et al., 1990). The intracellular domain has several putative phosphorylation sites, which could potentially interact with cytoskeletal components. Another membrane glycoprotein, called neuroglian, is widely expressed at high levels on the surfaces of glia, neuronal cell bodies and axons in the CNS of the Drosophila embryo (Bieber et al., 1989). The gene encoding this protein was cloned and found to have homology with the immunoglobulin superfamily. The extracellular portion of the protein consists of six immunoglobulin domains, followed by five fibronectin type III domains (Fig. 17). The deduced amino acid sequence shows high homology with the mouse cell adhe-

36

sion molecule LI. However, null mutants for the neuroglian gene give no clues as to a possible role of neuroglian in axon guidance: such mutants show an apparently normal pattern of axonal tracts in the CNS. In the peripheral nervous system, the null mutant shows a disruption in the normally regular, parallel arrangement of five neuron cell bodies in the pentascolopodial chordotonal organ (cited in Hortsch and Goodman, 1991). Whether axon growth from these neurons is also affected has not been determined. Seeger et al. (1988) have characterized a gene called amalgam which is widely expressed in the CNS of the Drosophila embryo. The amalgam protein shows amino acid sequence similarity with vertebrate cell adhesion molecules and other molecules of the immunoglobulin superfamily. Using monoclonal antibodies, Seaver et al. (1991) identified an antigen which is expressed on the growth cones, filopodia and axons of virtually all neurons in the grasshopper CNS. While the antigen is widely distributed, its expression is Umited to the period of active axon growth. Experimental evidence for a role in axon growth has come from antibody blocking experiments. When cultured with antibodies against this protein, embryos display a range of defects in axon growth: growth cones are either blocked in growth or they extend along aberrant pathways. The effect is Umited to growth cones that use other axons as a

Fig. 20. The effect of treatment with phosphatidylinositolspecific phospholipase C (PI-PLC) on axon growth from the Til pioneer neurons in the grasshopper limb bud. (A) A 3 1 32% control embryo, cultured without phospholipase C for 24 h, showing the normal Til trajectory. The growth cones (open arrowhead) have contacted guidepost cells Fel (F) and Trl (T) before turning ventrally along the Tr-Cx segment boundary (large arrowheads) to contact the Cxi cells (small arrowhead). (B) Abnormal Til growth in the presence of PI-PLC. The axon has turned distally in the mid-femur and grown to the limb tip (arrowhead). The Til, Trl (T) and Fel (F) cell bodies have also been displaced. (C) In this PI-PLC treated embryo, the Til axons failed to turn ventrally at the Tr-Cx border and instead grew straight ahead into the coxa. Open arrowhead, Cxi cells. Photomicrographs of neurons stained with the anti-HRP antibody. Scale bar = 50yum (reproduced with permission from Chang et al., 1992, © Company of Biologists, Ltd.).

Axon guidance factors in invertebrate

development

growth substrate: axons of pioneer neurons, such as MPl, which grow on non-neuronal substrates, are not affected. This protein may be an example of a relatively non-specific marker for axonal vs. non-axonal substrata. The apparent lack of specificity in its action does not reduce its significance: it merely indicates that it must function in conjunction with other molecules to define specific growth pathways.

37

P.M. Whiting ton

Giniger et al. (1993) have presented evidence that the transmembrane proteins Notch and Delta, which are widely expressed in tissues of the Drosophila embryo, and which play diverse roles in cell determination, may be involved in the guidance of sensory axons along the trachea. Temperature-sensitive alleles were used to eliminate functional Notch and Delta protein expression during the specific developmental period when sensory axons are growing along the trachea: in treated embryos, dorsal sensory neurons often failed to extend axons or extended them in aberrant directions. 5.4. Role ofphosphotidylinositol-linked proteins Chang et al. (1992) have recently shown that treatment of grasshopper limb buds with enzymes that cleave the phosphodylinositol molecule causes abnormalities in axon growth of the pioneer sensory neuron Til. These defects include failure to fasciculate with its sibling axon, distal rather than proximal growth in the first part of its trajectory, when the axon is believed to be navigating along the epithelium and/or basal lamina, growth across rather circumferentially around segmental boundaries within the limb bud and failure to bind to "guidepost" neuron cell bodies within the limb (Fig. 20). This would suggest that phosphotidylinositol-linked membrane molecules are responsible for a diverse array of axon guidance phenomena, involving a number of different substrates. A likely candidate for one of these molecules is fasciclin I, which apparently mediates fasciculation of the Til axons and which also appears to be a phosphotidylinositol-linked protein (see Section 5.1.1). 6. Conclusions Work on identifying the molecules responsible for axon guidance in invertebrate nervous systems has been underway for only a few years, yet already a number of candidate recognition molecules have been characterized. We can confidently expect this list to grow rapidly as an arsenal of powerful genetic and molecular techniques is brought to bear on the problem.

Once a putative axon guidance factor has been identified, a major challenge remains in determining its specific role in the developing nervous system. This will require going back to the embryo and re-examining the cellular events underlying specific axon growth, preferably at the level of individually identified cells. In particular, the following sorts of questions will have to be addressed: What is the pattern of expression of this molecule on growing axons and their substrates? How does this pattern of expression change as axons advance over new substrates? Does expression of the molecule correlate with accurate axon turning, widespread axon branching, withdrawal of inappropriate branches or cessation of axon growth? What effect does absence of expression of this molecule in a null mutant or in the presence of specific antagonists have on the afore-mentioned cellular processes? What is the effect on these processes of ectopic expression of the molecule: is expression of the molecule per se sufficient to guide the axon? Is there redundancy in the action of the molecule and if so, what other factors must be deleted to produce an observably abnormal phenotype? How specific is the action of the factor: does absence of expression affect the growth of single axons or groups of axons? Is there cooperativity in the action of guidance factors: does ablating two putative axon guidance factors together have a greater effect than ablating either alone? Having clarified the cellular role of the factor, the next step will be to elucidate how it carries out this role. Does the molecule act to simply mediate cell adhesion? Is it involved in a cellular signalling pathway and if so, at what level: as a ligand, a receptor, or further down the signalling chain? Assaying the effects of specific pharmacological agents on the dynamic behaviour of identified growth cones in vivo will provide one avenue of approach to this difficult problem. Acknowledgements I am grateful to David Merritt for his insightful criticisms on various drafts of this manuscript. Helen Sink contributed substantially to the devel-

38

opment of many of the ideas contained therein. I also thank Sandy Hamdorf and Becky Francis for help in compiling references and Zoltan Enoch for photographic assistance. References Anderson, H. and Bacon, J. (1979) Developmental determination of neuronal projection patterns from wind-sensitive hairs in the locust, Schistocerca gregaria. Dev. Biol. 72: 78-87. Anderson, H. and Tucker, R.P. (1988) Pioneer neurones use basal lamina as a substratum for outgrowth in the embryonic grasshopper limb. Development 104: 601-608. Anderson, H. and Tucker, R.P. (1989) Spatial and temporal variation in the structure of the basal lamina in embryonic grasshopper limbs during pioneer neurone outgrowth. Development 106: 185-194. Ball, E.E., Ho, R.K. and Goodman, C.S. (1985) Development of neuromuscular specificity in the grasshopper embryo: guidance of motoneuron growth cones by muscle pioneers./. Neurosci. 5: 1808-1819. Baptista, C.A. and Macagno, E.R. (1988) Modulation of the pattern of axonal projections of a leech motor neuron by ablation or transplantation of its target. Neuron 1: 949962. Barthalay, Y., Hipeaujacquotte, R., de la Escalera, S., Jimenez, F. and Piovant, M. (1990) Drosophila neurotactin mediates heterophilic cell adhesion. EMBO J. 9: 3603-3609. Bastiani, M.J. and Goodman, C.S. (1986) Guidance of neuronal growth cones in the grasshopper embryo. III. Recognition of specific glial pathways. J. Neurosci. 6: 35423551. Bastiani, M.J., Raper, J.A. and Goodman, C.S. (1984) Pathfinding by neuronal growth cones in grasshopper embryos III. Selective affinity of the G growth cone for the P cells within the A/P fascicle. J. Neurosci. 4: 2311-2328. Bastiani, M.J., du Lac, S. and Goodman, C.S. (1986) Guidance of neuronal growth cones in the grasshopper embryo. I. Recognition of a specific axonal pathway by the pCC neuron. J. Neurosci. 6: 3518-3531. Bastiani, M.J., Harrelson, A.L., Snow, P.M. and Goodman, C.S. (1987) Expression of fasciclin I and II glycoproteins on subsets of axon pathways during neuronal development in the grasshopper. Cell 48: 745-755. Bastiani, M.J., Decouet, H.G., Quinn, J.M.A., Karlstrom, R.O., Kotrla, K., Goodman, C.S. and Ball, E.E. (1992) Position-specific expression of the annulin protein during grasshopper embryogenesis. Dev. Biol. 154: 129-142. Bate, CM. (1976) Pioneer neurones in an insect embryo. Nature 260: 54-56. Bate, M., Rushton, E. and Currie, D.A. (1991) Cells with persistent twist expression are the embryonic precursors of adult muscles in Drosophila. Development 113: 79-89. Bentley, D. and Caudy, M. (1983) Pioneer axons lose directed

Axon guidance factors in invertebrate development growth after selective killing of guidepost cells. Nature 304: 62-65. Bentley, D. and Keshishian, H. (1982a) Pathfinding by peripheral pioneer neurons in grasshoppers. Science (NY.) 218: 1082-1088. Bentley, D. and Keshishian, H. (1982b) Pioneer neurons and pathways in insect appendages. Trends Neurosci. 5: 364367. Bentley, D., Guthrie, P.B. and Kater, S.B. (1991) Calcium ion distribution in nascent pioneer axons and coupled preaxonogenesis neurons in situ. J. Neurosci. 11: 1300-1308. Berlot, J. and Goodman, C.S. (1984) Guidance of peripheral pioneer neurons in the grasshopper: adhesive hierarchy of epithelial and neuronal surfaces. Science (NY.) 223: 493496. Bernhardt, R.R., Nguyen, N. and Kuwada, J.Y. (1992) Growth cone guidance by floor plate cells in the spinal cord of zebrafish embryos. Neuron 8: 869-882. Bieber, A.J., Goodman, C.S., Traquina, Z.R., Snow, P.M., Jacobs, J.R., Hortsch, M., Patel, N.H. and Schilling, J. (1989) Drosophila neuroglian - a member of the immunoglobulin superfamily with extensive homology to the vertebrate neural adhesion molecule LI. Cell 59: 447-460. Blair, S.S. and Palka, J. (1985) Axon guidance in cultured wing discs and disc fragments of Drosophila. Dev. Biol. 108:411^19. Blair, S.S., Murray, M.A. and Palka, J. (1985) Axon guidance in cultured epithelial fragments of Drosophila wing. Nature 315:406-409. Blair, S.S., Murray, M.A. and Palka, J. (1987) The guidance of axons from transplanted neurons through aneural Drosophila wings. J. Neurosci. 7: 4165-4175. Braun, J. and Stent, G.S. (1989a) Axon outgrowth along segmental nerves in the leech. 1. Identification of candidate guidance cells. Dev. Biol. 132: 471-485. Braun, J.B. and Stent, G.S. (1989b) Axon outgrowth along segmental nerves in the leech II. Identification of actual guidance cells. Dev. Biol. 132: 486-501. Bray, D. and Chapman, K. (1985) Analysis of microspike movements on the neuronal growth cone. /. Neurosci. 5: 3204-3213. Briggs, K.K., Johansen, K.M. and Johansen, J. (1993) Selective pathway choice of a single central axonal fascicle by a subset of peripheral neurons during leech development. Dev. Biol. 158: 380-389. Burmeister, D. and Goldberg, D. (1988) Micropruning: the mechanism of turning of Aplysia growth cones at substrate borders in vitro. J. Neurosci. 8: 3151-3159. Burmeister, D.W., Chen, M., Bailey, C.H. and Goldberg, D.J. (1988) The distribution and movement of organelles in maturing growth cones: correlated video-enhanced and electron microscopic studies. J. Neurocytol. 17: 783-795. Carpenter, E.M. and Bastiani, M.J. (1991) Developmental expression of REGA-1, a regionally expressed glial antigen in the central nervous system of grasshopper embryos. J. Neurosci. 11:277-286.

P.M. Whiting ton Cash, S., Chiba, A. and Keshishian, H. (1992) Alternate neuromuscular target selection following the loss of single muscle fibers in Drosophila. J. Neurosci. 12: 2051-2064. Caudy, M. and Bentley, D. (1986a) Epithelial cell specialization at a limb segment boundary in the grasshopper embryo. Dev.Biol. 118:399^02. Caudy, M. and Bentley, D. (1986b) Pioneer growth cone morphologies reveal proximal increases in substrate affinity within leg segments of grasshopper embryos. J. Neurosci. 6: 364-379. Caudy, M. and Bentley, D. (1986c) Pioneer growth cone steering along a series of neuronal and non-neuronal cues of different affinities. /. Neurosci. 6: 1781-1795. Caudy, M. and Bentley, D. (1987) Pioneer growth cone behavior at a differentiating limb segment boundary in the grasshopper embryo. Dev. Biol. 119: 454-465. Chang, W.S., Serikawa, K., Allen, K. and Bentley, D. (1992) Disruption of pioneer growth cone guidance in vivo by removal of glycosyl-phosphatidylinositol-anchored cell surface proteins. Development 114: 507-519. Chang, W.S., Zachow, K.R. and Bentley, D. (1993) Expression of epithelial alkaline phosphatase in segmentally iterated bands during grasshopper limb morphogenesis. Development 118: 651-663. Chiba, A., Hing, H., Cash, S. and Keshishian, H. (1993) Growth cone choices of Drosophila motoneurons in response to muscle fiber mismatch. /. Neurosci. 13: 714732. Condic, M.L. and Bentley, D. (1989) Removal of the basal lamina in vivo reveals growth cone-basal lamina adhesive interactions and axonal tension in grasshopper embryos. J. Neurosci. 9: 2678-2686. Cox, E.C., Muller, B. and Bonhoeffer, F. (1990) Axonal guidance in the chick visual system: posterior tectal membranes induce collapse of growth cones from the temporal retina. Neuron 4: 31-31. Davenport, R.W. and Kater, S.B. (1992) Local increases in intracellular calcium elicit local rilopodial responses in Helisoma neuronal growth cones. Neuron 9: 405-416. de la Escalera, S., Bockamp, E.O., Moya, P., Piovant, M. and Jimenez, F. (1990) Characterization and gene cloning of neurotactin, a Drosophila transmembrane protein related to cholinesterases. EMBOJ. 9: 3593-3601. Desai, C, Carriga, G., Mclntire, S.L. and Horvitz, H.R. (1988) A genetic pathway for the development of the C. elegans HSN motor neurons. Nature 336: 638-646. Desai, C.J., Popova, E. and Zinn, K. (1994) A Drosophila receptor tyrosine phosphatase expressed in the embryonic CNS and larval optic lobes is a member of the set of proteins bearing the "HRP" carbohydrate epitope. J. Neurosci. 14: 7272-7283. Doe, C.Q., Bastiani, M.J. and Goodman, C.S. (1986) Guidance of neuronal growth cones in the grasshopper embryo. IV. Temporal delay experiments. J. Neurosci. 6: 35523563. Doherty, P. and Walsh, F.S. (1994) Signal transduction events

39 underlying neurite outgrowth stimulated by cell adhesion molecules. Curr. Opin. Neurobiol. 4: 49-55. du Lac, S., Bastiani, M.J. and Goodman, C.S. (1986) Guidance of neuronal growth cones in the grasshopper embryo. II. Recognition of a specific axonal pathway by the aCC Neuron. J. Neurosci. 6: 3532-3541. Edwards, J.S. and Chen, S. (1979) Embryonic development of an insect sensory system, the abdominal cerci of Acheta domesticus. Roux's Arch. Dev. Biol. 186: 151-178. Edwards, J.S., Chen, S.W. and Berns, M.W. (1981) Cereal sensory development following laser microlesions of embryonic apical cells in Acheta domesticus. J. Neurosci. 1: 250-258. Elkins, T., Hortsch, M., Bieber, A.J., Snow, P.M. and Goodman, C.S. (1990a) Drosophila fasciclin I is a novel homophilic adhesion molecule that along with fasciclin III can mediate cell sorting. 7. Cell Biol. 110: 1825-1832. Elkins, T., Zinn, K., McAllister, L., Hoffmann, F.M. and Goodman, C.S. (1990b) Genetic analysis of a Drosophila neural cell adhesion molecule: interaction of fasciclin I and abelson tyrosine kinase mutations. Cell 60: 565-575. Forscher, P. and Smith, S.J. (1988) Actions of cytochalasins on the organization of actin filaments and microtubules in a neuronal growth cone. /. Cell Biol. 107: 1505-1516. Gao, W.-Q. and Macagno, E.R. (1987a) Extension and retraction of axonal projections by some developing neurons in the leech depends upon the existence of neighbouring homologues. I. The HA cells. J. Neurobiol. 18: 43-59. Gao, W.-Q. and Macagno, E.R. (1987b) Extension and retraction of axonal projections by some developing neurons in the leech depends upon the existence of neighbouring homologues. II The AP and AE neurons. J. Neurobiol. 18: 295-313. Giniger, E., Jan, L.Y. and Jan, Y.N. (1993) Specifying the path of the intersegmental nerve of the Drosophila embryo - a role for Delta and Notch. Development 117: 431440. Goldberg, J.I. and Kater, S.B. (1989) Expression and function of the neurotransmitter serotonin during development of the Helisoma nervous system. Dev. Biol. 131: 483-495. Goldberg, J.I., Mills, L.R. and Kater, S.B. (1991) Novel effects of serotonin on neurite outgrowth in neurons cultured from embryos oi Helisoma trivolvis. J. Neurobiol. ll: 182194. Goodman, C.S. (1994). The likeness of being: Phylogenetically conserved molecular mechanisms of growth cone guidance. Cell 78: 353-356. Goodman, C.S. and Shatz, C.J. (1993) Developmental mechanisms that generate precise patterns of neuronal connectivity. Cell 72: 77-98. Goodman, C.S., Bastiani, M.J., Doe, C.Q., du Lac, S., Helfand, S.L., Kuwada, J.Y. and Thomas, J.B. (1984) Cell recognition during neuronal development. Science (N.Y.) 225: 1271-1279. Goodman, C.S., Grenningloh, G. and Bieber, A.J. (1991) Molecular genetics of neural cell adhesion molecules in

40 Drosophila. In: P.C. Letourneau (Ed.), The Nerve Growth Cone, Raven Press, New York, pp. 283-303. Grenningloh, G., Bieber, A.J., Rehm, E.J., Snow, P.M., Traquina, Z.R., Hortsch, M., Patel, N.H. and Goodman, C.S. (1990) Molecular genetics of neuronal recognition in Drosophila - evolution and function of immunoglobulin superfamily cell adhesion molecules. Cold Spring Harbor Symp. Quant. Biol. 55: 327-340. Grenningloh, G., Rehm, E.J. and Goodman, C.S. (1991) Genetic analysis of growth cone guidance in Drosophila: Fasciclin II functions as a neuronal recognition molecule. Cell 67: 45-57. Gunderson, R.W. and Barrett, J.N. (1979) Neuronal chemotaxis: chick dorsal-root axons turn toward high concentrations of nerve growth factor. Science (N.Y.) 206: 10791080. Halpern, M.E., Chiba, A., Johansen, J. and Keshishian, H. (1991) Growth cone behavior underlying the development of stereotypic synaptic connections in Drosophila embryos. J. Neurosci. 11: 3227-3238. Hamelin, M., Zhou, Y.W., Su, M.W., Scott, I.M. and Culotti, J.G. (1993) Expression of the UNC-5 guidance receptor in the touch neurons of C elegans steers their axons dorsally. Nature 364: 327-330. Harrelson, A.L. (1992) Molecular mechanisms of axon guidance in the developing insect nervous system. J. Exp. Zool. 261:310-321. Harrelson, A.L. and Goodman, C.S. (1988) Growth cone guidance in insects: Fasciclin II is a member of the immunoglobulin superfamily. Science (N.Y.) 242: 700-707. Hartenstein, V. (1988) Development of Drosophila larval sensory organs: spatiotemporal pattern of sensory neurones, peripheral axonal pathways and sensilla differentiation. Development 102: 869-886. Hatta, K., Kimmel, C.B. and Walker, C. (1991) The cyclops mutation blocks specification of the floor plate of the zebrafish central nervous system. Nature 350: 339341. Haydon, P.G., McCobb, D.P. and Kater, S.B. (1984) Serotonin selectively inhibits growth cone motility and synaptogenesis of specific identified neurons. Science (N.Y.) 226: 561-564. Heathcote, R.D. (1981) Differentiation of an identified sensory neuron (SR) and associated structures (CTO) in grasshopper embryos. /. Comp. Neurol. 202: 1-18. Hedgecock, E.M., Culotti, J.G. and Hall, D.H. (1990) The unc-5y unc-6 and unc-40 genes guide circumferential migrations of the pioneer axons and mesodermal cells on the epidermis in C. elegans. Neuron 4: 61-85. Hildago, A., Urban, J. and Brand, A.H. (1995) Target ablation of glia disrupts axon tract formation in the Drosophila CNS. Development 121: 3703-3712. Ho, R.K. and Goodman, C.S. (1982) Peripheral pathways are pioneered by an array of central and peripheral neurones in grasshopper embryos. Nature 297: 404-406. Hortsch, M. and Goodman, C.S. (1991) Cell and substrate

Axon guidance factors in invertebrate development adhesion molecules in Drosophila. Annu. Rev. Cell Biol. 7: 505-557. Hortsch, M., Patel, N.H., Bieber, A.J., Traquina, Z.R. and Goodman, C.S. (1990) Drosophila neurotactin, a surface glycoprotein with homology to serine esterases, is dynamically expressed during embryogenesis. Development 110: 1327-1340. Hosoya, T., Takizawa, K., Nitta, K. and Hotta, Y. (1995) glial cells missing: a binary switch between neuronal and glial determination in Drosophila. Cell 82: 1025-1036. Ishii, N., Wadsworth, W.G., Stern, B.D., Culotti, J.G. and Hedgecock, E.M. (1992) UNC-6, a laminin-related protein, guides cell and pioneer axon migrations in C. elegans. Neuron 9: S73-SSI. Jacobs, J.R. and Goodman, C.S. (1989a) Embryonic development of axon pathways in the Drosophila CNS. 1. A glial scaffold appears before the first growth cones. J. Neurosci. 9:2402-2411. Jacobs, J.R. and Goodman, C.S. (1989b) Embryonic development of axon pathways in the Drosophila CNS. 2. Behavior of pioneer growth cones. J. Neurosci. 9: 24122422. Jan, Y.N., Ghysen, A., Christoph, I., Barbel, S. and Jan, L.Y. (1985) Formation of neuronal pathways in the imaginal discs of Drosophila melanogaster. J. Neurosci. 5: 24532464. Jay, D.G. and Keshishian, H. (1990) Laser inactivation of fasciclin I disrupts axon adhesion of grasshopper pioneer neurons. Nature 348: 548-550. Jellies, J. and Kristan, W.B. (1988) An identified cell is required for the formation of a major nerve during embryogenesis in the leech. /. Neurobiol. 19: 153-165. Jellies, J., Johansen, K. and Johansen, J. (1994). Specific pathway selection by the early projections of individual peripheral sensory neurons in the embryonic medicinal leech. J. Neurobiol. 25: 1187-1199. Johansen, J., Halpern, M.E. and Keshishian, H. (1989) Axonal guidance and the development of muscle-fiber specific innervation in Drosophila embryos. J. Neurosci. 9: 43184332. Johansen, K.M., Kopp, D.M., Jellies, J. and Johansen, J. (1992) Tract formation and axon fasciculation of molecularly distinct peripheral neuron subpopulations during leech embryogenesis. Neuron 8: 559-572. Johansen, K.M., Kopp, D.M., Jellies, J. and Johansen, J. (1992) Tract formation and axon fasciculation of molecularly distinct peripheral neuron subpopulations during leech embryogenesis. Neuron 8: 559-572. Jones, B.W., Fetter, R.D., Tear, G. and Goodman, C.S. (1995) glial cells missing: a genetic switch that controls glial versus neuronal fate. Cell 82: 1013-1023. Kapfhammer, J.P. and Raper, J.A. (1987) Collapse of growth cone structure on contact with specific neurites in culture. J. Neurosci. 7: 201-212. Katz, F., Moats, W. and Jan, Y.N. (1988) A carbohydrate epitope expressed uniquely on the cell surface of Droso-

P.M. Whitington phila neurons is altered in the mutant nac (neurally altered carbohydrate). EMBO J. 11: 3471-3477. Kennedy, T.E., Serafini, T., Torre, J.R. and Tessier-Lavigne, M. (1994) Netrins are diffusible chemotropic factors for commissural axons in the embryonic spinal cord. Cell 78: 425-435. Keshishian, H. (1980) The origin and morphogenesis of pioneer neurons in the grasshopper metathoracic leg. Dev. Biol 80: 388-397. Keshishian, H. and Bentley, D. (1983a) Embryogenesis of peripheral nerve pathways in grasshopper legs I. The initial nerve pathway to the CNS. Dev. Biol 96: 89-102. Keshishian, H. and Bentley, D. (1983b) Embryogenesis of peripheral nerve pathways in grasshopper legs III. Development without pioneer neurons. Dev. Biol 96: 116-124. Keshishian, H., Chiba, A., Chang, T.N., Halfon, M.S., Harkins, E.W., Jarecki, J., Wang, L.S., Anderson, M., Cash, S., Halpern, M.E. and Johansen, J. (1993) Cellular mechanisms governing synaptic development in Drosophila melanogasterJ. Neurobiol 24: 757-787. Klambt, C, Jacobs, J.R. and Goodman, C.S. (1991) The midline of the Drosophila central nervous system - A model for the genetic analysis of cell fate, cell migration, and growth cone guidance. Cell 64: 801-815. Klose, M. and Bentley, D. (1989) Transient pioneer neurons are essential for formation of an embryonic peripheral nerve. Science (N. Y.) 245: 982-984. Kolodkin, A.L., Matthes, D.J., O'Connor, T.P., Patel, N.H., Admon, A., Bentley, D. and Goodman, C.S. (1992) Fasciclin IV - sequence, expression, and function during growth cone guidance in the grasshopper embryo. Neuron 9: 831845. Kolodkin, A.L., Matthes, D.J. and Goodman, C.S. (1993) The semaphorin genes encode a family of transmembrane and secreted growth cone guidance molecules. Cell 75: 13891399. Kolodziej, P.A., Jan, L.Y. and Jan, Y.N. (1995) Mutations that affect the length, fasciculation, and ventral orientation of specific sensory axons in the Drosophila embryo. Neuron 15: 273-286. Kuwada, J.W. (1984) Normal and abnormal development of an identified leech motor neuron. J. Emhryol Exp. Morphol 79: 125-137. Kuwada, J.Y. and Goodman, C.S. (1985) Neuronal determination during embryonic development of the grasshopper nervous system. Dev. Biol 110: 114-126. Kuwada, J.Y. and Kramer, A.P. (1983) Embryonic development of the leech nervous system: primary axon outgrowth of identified neurons. J. NeuroscL 10: 2098-2111. Lefcort, F. and Bentley, D. (1987) Pathfinding by pioneer neurons in isolated, opened and mesoderm-free limb buds of embryonic grasshoppers. Dev. Biol 119: 466-480. Letourneau, P.C. (1978) Chemotactic response of nerve fiber elongation to nerve growth factor. Dev. Biol 66: 183-196. Leung-Hagesteijn, C, Spence, A.M., Stern, B.D., Zhou, Y., Su, M-W., Hedgecock, E.M. and Culotti, J.G. (1992) UNC-

41 5, a transmembrane protein with immunoglobulin and thrombospondin type 1 domains, guides cell and pioneer axon migrations in C. elegans. Cell 71: 289-299. Lienhard, M.C. and Stocker, R.F. (1991) The development of the sensory neuron pattern in the antennal disc of wild-type and mutant {lz3, ssa) Drosophila melanogaster. Development Wl: 1063-1075. Lin, D.M. and Goodman, C.S. (1994) Ectopic and increased expression of fasciclin II alters motoneuron growth cone guidance. Neuron 13: 507-523. Lin, D.M., Fetter, R.D., Kopczynski, C, Grenningloh, G. and Goodman, C.S. (1994) Genetic analysis of fasciclin II in Drosophila: defasciculation, refasciculation and altered fasciculation. Neuron 13: 1055-1069. Luo, Y.L., Raible, D. and Raper, J.A. (1993) Collapsin - a protein in brain that induces the collapse and paralysis of neuronal growth cones. Cell 75: 217-227. Marx, J.L. (1984) Caenorhabditis elegans: getting to know you. Science (NY.) 225: 40-41. Mclntire, S.L., Garriga, G., White, J., Jacobson, D. and Horvitz, H.R. (1992) Genes necessary for directed axonal elongation or fasciculation in C. elegans. Neuron 8: 307-322. Meadows, L.A., Gell, D., Broadie, K., Gould, A.P. and White, R.A.H. (1994) The cell adhesion molecule, connectin, and the development of the Drosophila neuromuscular system. J.CellScl 107:321-328. Meier, T. and Reichert, H. (1991) Serially homologous development of the peripheral nervous system in the mouthparts of the grasshopper. J. Comp. Neurol 305: 201-214. Menne, T.V. and Klambt, C. (1994) The formation of commissures in the Drosophila CNS depends on the midline cells and on the Notch gene. Development 120: 123-133. Murphey, R.K., Bacon, J.P. and Johnson, S.E. (1985) Ectopic neurons and the organization of insect sensory systems. J. Comp. Physiol A -Sens. Neural Behav. Physiol 156: 381389. Murray, M.A., Schubiger, M. and Palka, J. (1984) Neuron differentiation and axon growth in the developing wing of Drosophila melanogaster. Dev. Biol 104: 259-273. Myers, P.Z. and Bastiani, M.J. (1993) Cell-cell interactions during the migration of an identified commissural growth cone in the embryonic grasshopper. J. NeuroscL 13: 115126. Myers, CM., Whitington, P.M. and Ball, E.E. (1990) Embryonic development of the innervation of the locust extensor tibiae muscle by identified neurons: formation and elimination of inappropriate axon branches. Dev. Biol 137: 194206. Nardi, J.B. (1983) Neuronal pathfinding in developing wings of the moth Manduca sexta. Dev. Biol 95: 163-174. Nardi, J.B. and Vernon, R.A. (1990) Topographical features of the substratum for growth of pioneering neurons in the Manduca wing disc. J. Neurobiol 21: 1189-1201. Nassel, D., Helgee, A. and Sivasubramanian, P. (1986) Development of axon paths of motoneurons after removal of target muscles in a holometabolous insect. Dev. Brain Res.

42

26:211-219. Nose, A., Mahajan, V.B. and Goodman, C.S. (1992) Connectin - a homophilic cell adhesion molecule expressed on a subset of muscles and the motoneurons that innervate them in Drosophila. Cell 70: 553-567. Nose, A., Takeichi, M. and Goodman, C.S. (1994) Ectopic expression of connectin reveals a repulsive function during growth cone guidance and synapse formation. Neuron 13: 525-539. O'Connor, T.P., Duerr, J.S. and Bentley, D. (1990) Pioneer growth cone steering decisions mediated by single filopodial contacts in situ. J. Neurosci. 10: 3935-3946. Passani, M.B., Peinado, A., Engelman, H., Baptista, C.A. and Macagno, E.R. (1991) Normally unused positional cues guide ectopic afferents in the leech CNS. J. Neurosci. 11: 3868-3876. Patel, N., Snow, P. and Goodman, C. (1987) Characterization and cloning of fasciclin 111: a glycoprotein expressed on a subset of neurons and axon pathways in Drosophila. Cell 48: 975-988. Piovant, M. and Lena, P. (1988) Membrane glycoproteins immunologically related to the human insulin receptor are associated with presumptive neuronal territories and developing neurones in Drosophila melanogaster. Development 103: 145-156. Placzek, M., Tessier-Lavigne, M., Jessell, T. and Dodd, J. (1990) Orientation of commissural axons in vitro to a floor plate-derived chemoattractant. Development 110: 19-30. Pulido, D., Campuzano, S., Koda, T., Modolell, J. and Barbacid, M. (1992) Dtrk, a Drosophila gene related to the trk family of neurotrophin receptors, encodes a novel class of neural cell adhesion molecule. EMBO 7. 11: 391-404. Raper, J.A., Bastiani, M.J. and Goodman, C.S. (1983a) Guidance of neuronal growth cones: selective fasciculation in the grasshopper embryo. Cold Spring Harbor Symp. Quant. Biol. 48: 587-598. Raper, J.A., Bastiani, M.J. and Goodman, C.S. (1983b) Pathfinding by neuronal growth cones in grasshopper embryos: I Divergent choices made by the growth cones of sibling neurons. J. Neurosci. 3: 20-30. Raper, J.A., Bastiani, M.J. and Goodman, C.S. (1983c) Pathfinding by neuronal growth cones in grasshopper embryos: II Selective fasciculation onto specific axonal pathways. J. Neurosci. 3: 31-41. Raper, J.A., Bastiani, M.J. and Goodman, C.S. (1984) Pathfinding by neuronal growth cones in grasshopper embryos IV. The effects of ablating the A and P axons upon the behavior of the G growth cone. J. Neurosci. 4: 2329-2345. Raz, E. and Shilo, B.Z. (1992) Dissection of the faint little ball (fib) phenotype - determination of the development of the Drosophila central nervous system by early interactions in the ectoderm. Development 114: 113-123. Rehder, V. and Kater, S.B. (1992) Regulation of neuronal growth cone filopodia by intracellular calcium. J. Neurosci. 12: 3175-3186. Rothberg, J.M., Jacobs, J.R., Goodman, C.S. and Artavanis-

Axon guidance factors in invertebrate development Tsakonas, S. (1990) Slit - an extracellular protein necessary for development of midline glia and commissural axon pathways contains both EGF and LRR domains. Genes Dev. 4: 2169-2187. Rubin, G.M. (1988) Drosophila melanogaster as an experimental organism. Science (New York) 240: 1453-1459. Sabry, J.H., O'Connor, T.P., Evans, L., Toroian-Raymond, A., Kirshner, M. and Bentley, D. (1991) Microtubule behavior during guidance of pioneer neuron growth cones in situ. J. Cell Biol. 115:381-395. Salzberg, A., D'Evelyn, D., Schulze, K.L., Lee, J.K., Strumpf, D., Tsai, L. and Bellen, H.J. (1994) Mutations affecting the pattern of the PNS in Drosophila reveal novel aspects of neuronal development. Neuron 13: 269-287. Schmid, H., Gendre, N. and Stocker, R.F. (1986) Surgical generation of supernumerary appendages for studying neuronal specificity in Drosophila melanogaster. Dev. Biol. 113:160-173 Schubiger, M. and Palka, J. (1985) Genetic suppression of putative guidepost cells: effect on establishment of nerve pathways in Drosophila wings. Dev. Biol. 108: 399-410. Schwab, M.E. and Caroni, P. (1988) Oligodendrocytes and CNS myelin are nonpermissive substrates for neurite growth and fibroblast spreading in vitro. J. Neurosci. 8: 2381-2393. Seaver, E.C., Karlstrom, R.O. and Bastiani, M.J. (1991) The restricted spatial and temporal expression of a nervoussystem-specific antigen involved in axon outgrowth during development of the grasshopper. Development 111: 881893. Seeger, M.A., Haffley, L. and Kaufman, T.C. (1988) Characterization of amalgam: a member of the immunoglobulin superfamily from Drosophila. Cell 55: 589-600. Seeger, M., Tear, G., Ferres-Marco, D. and Goodman, C.S. (1993) Mutations affecting growth cone guidance in Drosophila - genes necessary for guidance toward or away from the midline. Neuron 10: 409-426. Serafini, T., Kennedy, T.E., Galko, M.J., Mirzayan, C, Jessell, T.M. and Tessier-Lavigne, M. (1994) The netrins define a family of axon outgrowth-promoting proteins homologous to C. elegans UNC-6. Cell 78: 409-424. Singer, M.A., Hortsch, M., Goodman, C.S. and Bentley, D. (1992) Annulin, a protein expressed at limb segment boundaries in the grasshopper embryo, is homologous to protein cross-linking transglutaminases. Dev. Biol. 154: 143-159. Sink, H. and Whitington, P.M. (1991a) Pathfinding in the central nervous system and periphery by identified embryonic Drosophila motor axons. Development 112: 307-316. Sink, H. and Whitington, P.M. (1991b) Early ablation of target muscles modulates the arborisation pattern of an identified embryonic Drosophila motor axon. Development 113: 701-707. Sink, H. and Whitington, P.M. (1991c) Pathfinding by identified motoraxons in the embryo of Drosophila melanogaster. Proc. Austr. Neurosci. Soc. 2,46.

P.M. Whitington Sivasubramanian, P. and Nassel, D.R. (1989) Sensory projections from ectopic appendages in an insect - inherent specificity and influence of location. Dev. Growth Diff. 31: 341-349. Sivasubramanian, P. and Nassel, D.R. (1985) Axonal projections from transplanted ectopic legs in an insect. J. Comp. Neurol. 239: 247-253. Snow, P.M., Patel, N.H., Harrelson, A.L. and Goodman, C.S. (1987) Neural-specific carbohydrate moiety shared by many surface glycoproteins in Drosophila and grasshopper embryos. J. Neurosci. 1: 4137-4144. Tessier-Lavigne, M., Placzek, M., Lumsden, A.G.S., Dodd, J. and Jessell, T.M. (1988) Chemotropic guidance of developing axons in the mammalian central nervous system. Nature 336:115-11%. Thomas, J.B. and Crews, S.T. (1990) Molecular genetics of neuronal development in the Drosophila embryo. Fed. Am. Soc. Exp. Biol. 4: 2476-2482. Tian, S.S., Tsoulfas, P. and Zinn, K. (1991) Three receptorlinked protein tyrosine phosphatases are selectively expressed on central nervous system axons in the Drosophila embryo. Cell 67: 675-685. VanBerkum, M.F.A. and Goodman, C.S. (1995) Targeted disruption of Ca^'''-calmodulin signaling in Drosophila growth cones leads to stalls in axon extension and errors in axon guidance. Neuron 14: 43-56. Van Vactor, D., Sink, H., Fambrough, D., Tsoo, R. and Goodman, C.S. (1993) Genes that control neuromuscular specificity in Drosophila. Cell 73: 1137-1153. Wang, L.S. and Denburg, J.L. (1992) A role for proteoglycans in the guidance of a subset of pioneer axons in cultured embryos of the cockroach. Neuron 8: 701-714. Whitington, P.M. (1989) The early development of motor axon pathways in the locust embryo - the establishment of

43

the segmental nerves in the thoracic ganglia. Development 105: 715-721. Whitington, P.M. and Seifert, E. (1984) Axon growth from limb motoneurons in the locust embryo: the effect of target limb removal on the pattern of axon branching in the periphery. Dev. Biol. 106: 438-449. Whitington, P.M., Leach, D. and Sandeman, R. (1993). Evolutionary change in neural development within the arthropods - axonogenesis in the embryos of two crustaceans. Development I IS: 449-461. Whitlock, K. (1993) Development of Drosophila wing sensory neurons in mutants with missing or modified cell surface molecules. Development 117: 1251-1260. Wolfgang, W.J., Quan, F., Thambi, N. and Forte, M. (1991) Restricted spatial and temporal expression of G-protein alpha-subunits during Drosophila embryogenesis. Development 113: 521 ~53S. Yang, X.H., Seow, K.T., Bahri, S.M., Don, S.H. and Chia, W. (1991) Two Drosophila receptor-like tyrosine phosphatase genes are expressed in a subset of developing axons and pioneer neurons in the embryonic CNS. Cell 67: 661-673. Younossi-Hartenstein, A. and Hartenstein, V. (1993) The role of the tracheae and musculature during pathfinding of Drosophila embryonic sensory axons. Dev. Biol. 158: 430-447. Zak, N.B., Wides, R.J., Schejter, E.D., Raz, E. and Shilo, B. (1990) Localization of the DER/flb protein in embryos: implications on the faint little ball lethal phenotype. Development 109: 865-874. Zinn, K., McAlhster, L. and Goodman, C.S. (1988) Sequence analysis and neuronal expression of fasciclin I in grasshopper and Drosophila. Cell 53: 577-587. Zipser, B., Morell, R. and Bajt, M.L. (1989) Defasciculation as a neuronal pathfinding strategy: involvement of a specific glycoprotein. Neuron 3: 621-630.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 2

Adhesion molecules in neural crest development D.F. Newgreeni and S.S. Tan^ ^Embryology Laboratory, Murdoch Institute and ^Embryology Laboratory, Department of Anatomy and Cell Biology, University of Melbourne, Parkville, Victoria 3052, Australia

1. Introduction The morphogenesis of the neural crest is marked by a developmentally regulated epithelio-mesenchymal transformation, cell dispersal along stereotyped routes and final localization to produce the ganglia of the peripheral nervous system, as well as other derivatives. These have been described and their developmental mechanics investigated using a combination of "classic" embryological techniques combined with cell biology and biochemistry. The execution of this behavior involves a changing relationship between the crest cells and their environment, primarily involving molecules affecting and modifying cell-cell and cell-matrix adhesion and cell motility. These molecular mechanisms are not unique to the neural crest, but the particular pattern of its morphogenesis depends on the degree of expression and especially the spatiotemporal expression, of these in particular combinations, probably in a redundant fashion. The genetic controls specifying the expression patterns of these molecular effectors are only now being sought, using molecular biological approaches. 2. Peripheral ganglia are established from distant precursors The peripheral nervous system, including the autonomic nervous system, is derived mainly from the neural crest. This structure arises during early embryogenesis in the primitive ectoderm along the

border between the central neural plate (which produces the neural tube, the forerunner of the central nervous system) and the more peripheral epidermal ectoderm (which produces the epidermis of the skin). Following neurulation, neural crest cells are initially found along the entire longitudinal neural axis, typically on the dorsal surface of the forming neural tube. Mapping studies using radioisotopes, fluorescent dyes and chick-quail chimeras have established that these epithelial crest cells convert to mesenchyme with a defined timetable, then translocate along defined, often complex routes and re-assemble in characteristic sites. The derivatives of the crest are not restricted to neural cells, such as neurons and support cells, but also include a range of endocrine cells, such as adrenal chromaffin and ultimobranchial body cells, non-retinal pigment cells and, especially in the head, neck and great vessels of the heart, a variety of cells forming dense and loose connective tissues (Horstadius, 1950; Weston, 1970; Le Douarin, 1982). Each spatial locus of the neural crest (e.g. a single segmental level) normally gives rise to a variety of derivatives, such as the dorsal root and sympathetic ganglia (Le Douarin, 1982). In addition, different loci can give rise to different derivatives, like sympathetic versus parasympathetic ganglia, while sharing other common derivatives, such as dorsal root ganglia. The position and distance of each structure from the contributing neural crest is variable; this, therefore, focusses attention on how distance, direction and timing of cell migration is achieved.

46

3. Role of adhesion during initiation of migration The onset of neural crest cell migration involves two basically different events, which are simultaneous in birds and mammals, but which are sequential in urodele amphibians. The events are (a) egress of crest cells from the neural epithelium and (b) dispersal of these cells into adjacent tissues. The events governing these can be analyzed from the standpoint of what could inhibit them and how and in what sequence these inhibitions are lifted. The control of the timetable of the initiation of migration would then reside in the last inhibitory control. Four general categories of restraints can be recognized: (i) physical barriers to migration in the environment of the crest cells; (ii) compositional qualities unfavorable to migration in the environment; (iii) locomotory incompetence in the crest cells; (iv) inability of the crest cells to detach from the neural tube. 3.1. Physical barriers Physical barriers postulated to limit crest cell emigration are basal laminae and lack of extracellular space (see Newgreen, 1989). In vitro and in vivo tests with avian crest-derived cells showed that basal laminae can, indeed, prevent cell egress from an epithelium (Erickson, 1985). At trunk levels of avian, mammalian and amphibian embryos, however, a basal lamina never actually forms over the premigratory crest, due to the mode of creation of new basal surface in the neural epithelium (Martins-Green, 1988). Thus, the absence of a basal lamina is a precondition that is met well before the event occurs; it is, therefore, not a decisive controlling factor in the timing of the egress of crest cells at these axial levels in these species. An inhibitory role for basal laminae could still be preserved in a different form by limiting dispersal, because the continuity (originally and wrongly termed "fusion" by Newgreen and Gibbins (1982) and Lofberg et al. (1985)) of the basal laminae of the trunk neural tube and epidermal ectoderm (Martins-Green, 1988) lateral to the crest itself could confine crest cells to a local, sealed dorsal

Adhesion molecules in neural crest development

compartment. However, this putative compartment is reportedly opened several hours prior to emigration in avian embryos (Newgreen, 1989). Therefore, this also seems to be a precondition met before emigration, rather than a strict control point. Moreover, this basal lamina continuity may not occur at all trunk levels. In cranial levels of rodents, in contrast, the disintegration of a basal lamina correlates precisely with the egress of crest cells (Nichols, 1986, 1987; Tan and Morriss-Kay, 1985, 1986) and the deformation of cells as they protrude through small gaps in the lamina strongly suggests that the lamina is a physical barrier. Removal of this barrier may be, in some cases, by cellular protrusive force, but the ultrastructural appearance (Nichols, 1987) is more consistent with enzymatic attack. Mouse crest cells produce plasminogen activator (PA), probably chiefly urokinase-type both in vitro and in vivo (Menoud et al., 1989a,b), which can attack basal lamina components. Metalloproteinases (Matrisian, 1990) could also be suspected, given the collagen type IV-content of basal laminae. The timed activation of such enzymes (Pittman, 1990; Matrisian, 1990) would be required to account for the sudden loss of basal laminae. Alternatively, or in addition, the basal lamina could be disrupted by expansionary extracellular matrix (ECM) pressure and the usual candidate for this role, hyaluronan (HA), has been reported by Poelmann et al. (1990) using a labelled HA-ligand although not by Morris-Wiman and Brinkley (1990) using histochemical methods. This difference may be due to difficulties in preserving HA in embryonic tissues. Lack of gross space between tissues and cells could physically prevent crest cells from dispersing. In vitro modelling has revealed that cylindrical pores, less than l/^m in diameter, permitted the passage of avian crest cells (Newgreen, 1989), whereas most inter-tissue spaces in vivo exceeded this. In addition, although increases in extracellular spaces are correlated with initiation of migration, collapse of these spaces by digestion of HA at cranial (Anderson and Meier, 1982) and trunk levels (Schoenwolf and Fisher, 1983) failed to prevent crest migration. Thus, it seems unrealistic that gross spatial insufficiency prevents premature

D.F. Newgreen and S.S. Tan

migration in the crest system; it may, however, reduce the flux of cells. 3.2. Unfavorable environments One way in which crest cells could disperse is by locomotion, which would be dependent on the appearance of structural and adhesive environmental features. The only candidate for these is the ECM, which contains a variety of interstitial and basal lamina-related adhesion molecules, such as fibronectin (FN), laminin (LN) and collagens (see Section 6). That such ECM is necessary is indicated in avian cranial tissues by experimental interference with cell-ECM adhesion in vivo (Boucaut et al., 1984; Bronner-Fraser, 1985, 1986b; Poole and Thiery, 1986) or by explanting avian trunk neural anlagen (neural tube plus neural crest) into ECM gels of low adhesivity (Bilozur and Hay, 1989; see Section 6). In both cases, dispersal of crest cells from the neural anlage was curtailed, yet presumptive crest cells still lost their position in the epithelium, either bulging internally into the neurocoele, or collecting just outside the neural tube. In all species examined, ECM molecules, such as FN, LN and collagen types I and IV antedate the start of migration (Newgreen and Thiery, 1980; Duband and Thiery, 1982, 1987; Sternberg and Kimber, 1986; Krotoski et al, 1986; Duband et al, 1986; Martins-Green and Erickson, 1987; MartinsGreen, 1988; Epperlein et al., 1988; Krotoski and Bronner-Fraser, 1990). Therefore, in order that changes in local ECM be considered as controlling the crest migration timetable, additional appropriately staged qualitative or quantitative ECM changes are required. Evidence that such changes can occur is provided by emplacement of filters in older axolotl embryos, which coats the filters with ECM. Transplantation of those filters to contact the premigratory crest string in younger axolotl embryos instigated precocious migration, which was restricted to within the borders of the filter (Lofberg et al., 1985). The cellular responses indicate an adhesive agent, but the nature of the active ECM components are, as yet, unknown. Epperlein et al.

47

(1988) report that the ECM component cytotactin (CT) appears here around the time of emigration, but in vitro assays have not clarified the role of this molecule (Tan et al, 1987; Mackie et al, 1988; Halfteretal., 1989). These experiments clearly point to changes in local microenvironment, presumably ECM, as triggering crest cell dispersal in the axolotl, even though the molecules responsible are not known. Moreover, avian migratory stage crest cells do not migrate into early somitic tissue that is heterochronically transplanted, in contrast to their response to older somites (Bronner-Fraser and Stem, 1991), indicating likewise some microenvironmental change from non-permissive to permissive. Nevertheless, in avian embryos, premigratory neural anlagen do not produce prematurely migrating cells when explanted onto ECM substrates known to favor cell locomotion, even though the crest cells adhere to the ECM (Newgreen and Gibbins, 1982). Thus, at trunk levels in birds, changes additional to any that might occur in the microenvironment must be made before emigration proceeds and these changes must be within the neural anlage itself. 3.3. Locomotory incompetence Crest cell migration might be delayed until the cells mature to a degree where they can actively perform locomotion. The most likely maturational changes would be in the development of cell surface adhesion molecules for ECM and in the extensile and contractile machinery of the cytoskeleton (see Newgreen, 1990). The filter transplant experiments of Lofberg et al. (1985) indicate that these qualities are adequate for locomotion in premigratory axolotl trunk crest cells. Immunocytochemical methods indicate that at least some integrin-type cell-ECM adhesion molecules are present on avian premigratory crest cells (Duband et al., 1986; Krotoski et al, 1986) and TEM reveals cell extension processes indistinguishable from those found later on migratory cells (Tosney, 1978, 1982; Newgreen and Gibbins, 1982). Moreover, premigratory avian trunk crest cells can adhere to favorable ECM substrates in vitro, even

48

though they cannot leave the neural anlage (Newgreen and Gibbins, 1982). These results do not indicate that locomotory competence changes but, recently, more sophisticated analyses of avian neural crest cells in culture have shown that the ability of these cells to spread on ECM substrates is greater in migratory-phase cells compared with premigratory-phase cells and that the phase change of this property could be prematurely advanced by TGF-)8 (Delannet and Duband, 1992). Coordinated and rapid increases in locomotory ability also result from protein kinase C inhibition in culture (Newgreen and Minichiello, 1995), suggesting that the responses to the growth factor may involve this pathway. Cell spreading is a complex cytoskeletal response that requires adhesion as a preliminary and is itself seen as being a preliminary to cell locomotion. These experiments, therefore, suggest that properties of locomotory competence, in fact, do alter at the time of onset of crest cell migration. It would be tempting, therefore, to predict that the onset of migration would be inhibited by suppressing cell locomotion. However, avian crest cells can still leave the parent epithelium even when locomotion is suppressed by actin microfilament depolymerization in vivo (Schoenwolf et al., 1988). These results indicate that innate locomotory competence does increase before crest migration begins in trunk levels of avian embryos, but such locomotory ability does not necessarily ensure that crest cells leave the neural epithelium. 5.4. Inability to de-adhere Well before migration begins, presumptive crest cells are integrated into the ectodermal epithelium via prominent junctional cell adhesion complexes. Since these are lacking at the migratory phase, their loss could control the timing of emigration. This has focussed attention on cell adhesion molecules (CAMs). Many cell-cell adhesion molecules are now known in developing tissues (reviewed by Hynes and Lander, 1992) and those that have been investigated with respect to neural crest cell morphogenesis are adherens CAM (A-CAM), liver CAM (L-CAM) and neural CAM (N-CAM). ACAM (also known as N-cadherin) and L-CAM

Adhesion molecules in neural crest development

(also known as E-cadherin) are both transmembrane glycoproteins of about 120-135 kDa. They mediate strong homophilic adhesion, which is absolutely Ca^+-dependent and despite close sequence similarity, binding between A-CAM and L-CAM is very limited. These molecules form the adhesive basis of and are enriched at the adherens junctions seen with the electron microscope, but they are not exclusive to this junctional complex. The cytoplasmic domain of these molecules interacts with actin cytoskeleton via several specific proteins termed catenins, which integrate cytoskeletal events with cell-cell adhesion. N-CAM is a homophilic adhesion glycoprotein of the immunoglobulin superfamily and, in contrast to A- and L-CAM, is Ca^+-independent. Several different splicing forms of the N-CAM polypeptide exist, ranging from about 180 to 120 kDa, all with adhesive function, the cytoplasmic portion being progressively shortened. The smallest variant entirely lacks a transmembrane domain and is attached to the cell surface via a phosphotidyl-inositol linkage. It is suggested that this version is important in rapidly changing situations, due to its high lateral mobility in the cell membrane. N-CAM is heavily and variably glycosylated, especially with sialic acid and this is developmentally regulated. High levels of sialic acid substitution correlate with lower binding affinity and tend to occur on early embryonic N-CAM, in contrast to N-CAM in adult tissues. At cranial and trunk levels of avian embryos, the separation of the neural anlage from the epidermal ectoderm appears to involve the consolidation of L-CAM to the epidermal primordium and A-CAM plus N-CAM to the neural anlage (Thiery et al., 1982; Duband et al., 1988b). The later dissociation of the crest cells from the neural epithelium is closely related to the loss of both adherens junctions (Tosney, 1978, 1982; Newgreen and Gibbins, 1982) and cell surface A-CAM (Duband et al., 1988b; Akitaya and Bronner-Fraser, 1992). A reduction of N-CAM (Thiery et al, 1982; Duband et al, 1985) appears to occur also, but slightly after the onset of migration (Akitaya and Bronner-Fraser, 1992). This is consistent with migration being cued by a generalized lowering of

49

D.F. Newgreen and S.S. Tan

cell-cell adhesion, firstly between crest cells and epidermal cells and then, immediately prior to emigration, between individual crest cells and their crest and neural tube neighbors. The critical event, from these correlational studies, would presumably involve the down-regulation of A-CAM. In the cranial levels of rodents, loss of adherens junctional components (Innes, 1985; Nichols, 1987) likewise accompanies, or immediately precedes, de-epithelialization. The loss of cell-cell adhesion among these cells also correlates with the increase in intercellular space (possibly mediated by HA-induced hydration) (Poelmann et al, 1990, cf. Morris-Wiman and Brinkley, 1990), but the local appearance of an adhesion-masking molecule, chondroitin sulfate-proteoglycan (CS-PG) (Morriss-Kay and Tuckett, 1989) may be an important promoter of epithelial disintegration at the crest region. In the axolotl, the morphology of crest-string formation indicates that cell adhesion is lost between crest and epidermis and then between crest and neural tube, but that adhesions between the crest cells themselves still function. However, this homotypic adhesion is of a residual character; presentation of older ECM immediately results in dissolution of the crest string (Lofberg et al., 1985). When explanted in vitro on a mechanically unstable, migration-permissive ECM substrate, avian premigratory trunk level neural anlagen showed a distinct lag in the onset of crest cell migration, in contrast to similar explants from migratory stages. Physical detachment of the neural anlagen resulted in the migratory levels invariably leaving a footprint of crest cells adherent to the ECM, while crest cells never detached from premigratory anlagen. Instead, in many cases, the entire explant detached not only with the entire complement of crest cells, but also with large fragments of ECM. Since the physical resilience of the ECM was likely to be constant, this indicates that a functional decrease of cell-cell adhesion does occur in the crest cell population at exactly the time of normal initiation of migration (Newgreen and Gibbins, 1982). Furthermore, artificial weakening of cell-cell adhesion by proteolytic attack of Ca^"^dependent adhesion molecules (a class including

the A-CAM/adherens junctions) caused premature crest cell migration in vitro (Newgreen and Gooday, 1985). Interestingly, although the coherence of the entire neural anlage was reduced by this, only the crest cells responded by rapid cell spreading and migration. It would seem from these in vitro functional studies that latent migratory ability of avian trunk crest cells is released by coordinate loss of CAMs, especially of the A-CAM type. As yet, however, experiments to functionally test the role of adhesions in vivo have not been reported. In summary, the same four general properties outlined above seem, in all vertebrates, to change from non-permissive to permissive, during the passage of neural crest cells from the premigratory to the migratory stage. In particular, the importance of the loss of cell-cell adhesions is suggested by several different lines of evidence. Nonetheless, the timing of change of these four properties can vary between species and even between different axial levels in the same species. In addition, the dominant molecular mechanisms mediating functionally equivalent changes in each of these four characters may vary between species. Thus, a major morphogenetic event, the onset of crest cell migration, is highly conserved throughout the vertebrates, but the timing, nature and importance of the many contributory mechanisms may not be highly conserved. Moreover, there is also marked heterochrony between the onset of crest migration and the morphogenetic programs carried out by neighboring tissues. This implies that the, as yet, obscure genetic regulation of the quality, quantity and timing of these events involves some degree of parallel independent processing, rather than a rigid cascade of events. 4. Intrinsic versus extrinsic influences on cell migration 4.7. Intrinsic The release of neural crest cells from their parent epithelium, of itself, does not necessarily mean that the cells of this population will undergo significant net displacement, yet these cells translo-

50

cate directionally along stereotyped pathways. Heart, somite and limb-bud mesenchyme cells, for example, placed on avian crest cell migration routes (Fisher and Solursh, 1979b; Noden, 1978; Erickson et al., 1980), remain at the site of insertion, whereas crest-derived cells soon distribute themselves along these pathways (Le Douarin et al., 1978; Bronner and Cohen, 1979; Erickson et al., 1980; Ayer-Le Lievre and Le Douarin, 1982). Moreover, crest cells grafted to abnormal sites also show dispersive abilities superior to other mesenchyme cells (Fisher and Solursh, 1979b; Erickson et al., 1980). These results indicate that crest cells possess attributes facilitating cell displacement and that these are less well developed in many other cells. However, such properties are not entirely specific for crest cells; mouse sarcoma S 180 cells can also show crest-like distribution patterns, when implanted into chick embryos (Erickson et al., 1980). Many of the cellular qualities thought to promote migration in crest cells center on cell adhesion. Neural crest cell behavior has been studied in detail in tissue culture, often in comparison with that of other cells (Newgreen et al., 1979; Fisher and Solursh, 1979a; Rovasio et al, 1983). Although able to adhere to many ECM molecules (see Section 6), crest cells in vitro are less flattened, are more easily sheared from culture substrates and show briefer cell-cell and cellsubstrate contacts, in comparison with somite cells (Newgreen et al., 1979). They also show very little development of focal contacts to the substrate; instead, they chiefly maintain the more generalized close contacts (Duband et al, 1986). This indicates that the adhesions of crest cells are weak or unstable and at least in tissue culture, the FN adhesion receptor is highly mobile in crest cells, correlating with the less organized state of their cytoskeleton. This contrasts with the low mobility of the same receptor in less migratory cells, which have focal contacts and highly ordered microfilament cytoskeletal arrays (Duband et al, 1988a). Adhesive instability may be related, in part, to proteolytic processing of cell-substrate contacts (Pollanen et al., 1987). Proteases are important in many developmental events and especially in cell

Adhesion molecules in neural crest development

migration (Matrisian, 1990; Pittman, 1990). The protease PA is strongly associated with crest cells (Valinsky et al., 1990; Menoud et al., 1989a,b) and inhibition of urokinase-PA activity using antibodies decreases crest cell migration on FN substrates in vitro (Valinsky et al., 1990). Likewise, a range of chemical protease inhibitors inhibits crest outgrowth in collagen gel cultures (Erickson and Isseroff, 1989). This suggests that an optimal degree of protease activity (which itself reflects a balance between protease and inhibitors) is poised between too great an activity, which would compromise the ability to form adhesions and too little activity, which would prevent adhesions being relinquished. Adhesive modulation could also be achieved by masking of adhesive molecules (see Newgreen, 1990) and crest cells produce CS-PG (Pintar, 1978), which has this function in cell culture assays (Newgreen, 1982). Cell traction is exerted on adhesions by the cytoskeleton. Fibroblasts in vitro have actin bundles terminating at focal contacts, which apply force sufficient to deform the surroundings (Harris et al., 1981), but this also impairs locomotion on deformable substrates. In contrast, cells that are known to translocate in vivo develop much weaker tractions, associated with a more diffuse, less oriented actin microfilament network in the cell cortex (Tucker et al., 1985; Duband et al., 1986). This, as well as allowing drastic changes in cell shape (Newgreen, 1989), permits crest cells to move on substrates of very low rigidity, such as low concentration collagen gels and silicone microsheets (Tucker et al., 1985). In addition, crest cells show particularly high levels of vinculin and talin, molecules mediating the attachment of the cytoskeleton to integral membrane adhesion receptors (Duband and Thiery, 1990). One of these, the FN receptor, is highly mobile in crest cells, correlating with the less organized state of their cytoskeleton. This contrasts with the low mobility of the same receptor in less motile cells, which have focal contacts and highly ordered microfilament cytoskeletal arrays (Duband et al., 1988a). The extension of projections is the least well understood facet of cell locomotion (Newgreen,

D.F. Newgreen and S.S. Tan

1990), but migrating crest cells in vitro extend processes much more rapidly than do, for example, somite cells. Extensions of crest cells occur on a narrow front, in contrast to those of somite cells (Newgreen et al., 1979; Tucker et al, 1985) and similar narrow projections are seen on crest cells in in vivo (Newgreen et al., 1990). Individual crest cell projections in vitro tend to be short-lived (correlating with adhesive instability) and co-exist with other projections at various stages of extension and retraction. In summary, crest cells are intrinsically fitted for a locomotory mode by having a broad spectrum substrate-adhesive capacity. This alone, of course, is insufficient to achieve mobility, but is coupled to a cytoskeleton permitting a highly deformable cell shape (including the ability to extend cell processes) which gives flexibility of force application at the expense of absolute contractile strength. In addition, crest cells' adhesions are relatively unstable and cellular de-adhesion may be actively regulated (reviewed by Newgreen, 1992). At least in vitro, even this does not ensure consistent cell displacement because, in the absence of external guidance, this locomotory capacity becomes dissipated both by simultaneous extensions in divergent directions and by rapid sequential and random directional alterations (Newgreen et al., 1979). Therefore, conversion of this capacity for active locomotion into real displacement would require almost constant directional instruction, derived from extrinsic sources (as outlined in the next section). 4.2. Extrinsic The influence of the environment on crest morphogenesis has been tested by heterotopic grafting, whereby a spatially defined region of the premigratory neural crest is removed from an embryo and replaced by a different region of neural crest, which is known to exhibit, at later stages, a different pattern of migration (Le Douarin and Teillet, 1974; Noden, 1975, 1978, 1983). Likewise, the effect on the development of the neural crest of microsurgical additions or deletions of neighboring tissues have been observed (Detwiler, 1937;

51

Keynes and Stem, 1984; Noden, 1986; Stem and Keynes, 1987; Tosney, 1987; Kalchiem and Teillet, 1989). Results of these experiments imply that the tissues surrounding the neural crest, in part, dictate their migration routes and impose a morphogenetic pattern on the neural derivatives. In addition, heterochronic cell labelling and grafting indicated that the developmental age of surrounding tissues also influences crest cell migration (Weston and Butler, 1966; Serbedzija et al., 1990; Bronner-Fraser and Stem, 1991). Taken together, these observations and experiments indicate that the local microenvironment provides cues to promote and control cell migration and that neural crest cells are specialized to follow these cues. In other words, the nature and degree of interactions between the cell and its microenvironment determine, in a large part, whether cell migration will occur, what routes will be followed and at what speed. 5. Cell adhesion is a major control mechanism of migration Cell locomotion in tissue culture clearly involves a rolling cycle of cell process extension, adhesion, cytoplasmic contraction (or flow) in the direction of adhesion, and de-adhesion (see Newgreen, 1990). Thus, the interactions between cultured cells and their micro-environment that control the process of cell movement have, as their hub, an adhesive component. Manipulation of cell adhesion in tissue culture assays leads to predictable changes in cell movements. Could the complex process of neural crest cell migration in vivo also have an important adhesive basis? First, in favorable in vivo situations where the dynamics of cell movement can be observed, the process of locomotion of crest-derived cells resembles that seen in the simplified cell culture models (see Newgreen, 1990). Secondly, ultrastructural observations of migrating crest cells in the embryo reveal contacts between the surface of migrating cells and microenvironmental features, such as ECM fibrils and the surfaces of other cells, which are identical to the adhesive contacts observed in vitro (Bancroft and Bellairs, 1976; Newgreen et al., 1982).

52

Thirdly, adhesion receptors are found on neural crest cells and their ligands are present along the cells' migration pathways (see Section 6). Fourthly, neural crest cells respond appropriately to these molecules in in vitro assays. Fifthly, interference with some adhesive interactions perturbs neural crest cell migration in vivo. In total, the general observations outlined above strongly suggest that cell adhesion is one of the major control mechanisms for neural crest cell migration, with the general model being that the ability to move per se is maximal at some intermediate level of adhesivity, because too great a level would impede the de-adhesion phase of the locomotory cycle, while too little would impede the adhesion phase. The course of migration, in this view, would be defined by the layout of adhesive substrates in the microenvironment. Much recent work has concentrated on putting molecular flesh on the bones of this notion of adhesive control. 6. Cell adhesion is mediated by specific molecules Studies on pathway choice will be outlined below under the headings of each adhesion-related molecule that might be involved. 6.1. Fibronectins Fibronectins (FN) are glycoprotein multimers (monomer molecular mass around 230 kDa) existing in several forms produced by differential splicing of a single gene (Komblihtt et al., 1985; Ffrench-Constant, 1988). FN is very widely distributed in practically all areas rapidly populated by crest cells and tissue culture experiments indicate that cells of every tissue surrounding the crest migration pathways can produce it (Newgreen and Thiery, 1980). The avian crest cell population, however, shows a patterned variation in its ability to produce FN. The cells migrating earliest from cranial and sacral levels of the crest produce it, while the later migrating cells, as well as almost all crest cells at cervical to lumbar levels, fail to produce it (Newgreen and Thiery, 1980).

Adhesion molecules in neural crest development

In vitro assays show that FN is a very favorable substrate for adhesion, spreading and locomotion of crest cells of all vertebrates tested (Greenberg et al., 1981; Newgreen, 1982; Newgreen et al., 1982; Erickson and Turley, 1983; Rovasio et al, 1983; Ferris and Johansson, 1987; Menoud et al., 1989b; Smith-Thomas and Fawcett, 1989). This ability resides mainly in the major cell-binding domain of the molecule, which is centered on the arginineglycine-aspartic acid-serine (RGDS) motif. However, limited responses are also associated with other domains, in particular the alternatively sphced connecting sequence-1 near the disulfide bond at the carboxy-terminal (Dufour et al., 1988) and also with the heparin-binding region near the amino-terminal (Ferris et al., 1989). Lower degrees of migration were also elicited by the collagenbinding domain with or without the adjacent heparin fibrin-binding region near the amino-terminal (Ferris et al., 1989). These results, therefore, account for almost all the recognized domains of FN and it is clear that the response of crest cells to FN in vitro is complex. The precise relevance for crest cells of such diverse adhesive modalities is unknown. Crest cells can accurately follow tracks of FN flanked by low-adhesion tracks (Newgreen et al., 1982; Rovasio et al., 1983), but stray from FN tracks onto adjacent LN tracks (Newgreen, 1984). Crest cells, however, do prefer tracks of fibronectin to adjacent tracks of collagen type I. FNrich regions in vivo could, therefore, provide adhesively favored routes. Crest cell adhesion and spreading increases with the substrate density of FN up to a plateau level (Ferris et al., 1989), suggesting that it could also provide haptotactic guidance whereby cells move directionally up a concentration gradient (see Newgreen, 1990 for review). The adhesion of avian crest cells to the RGDSsite is due to an integrin receptor (see Akiyama et al., 1990) of the ^psubfamily, which can be demonstrated on crest cells in situ (Duband et al., 1986; Krotoski et al., 1986). Similar immunoreactivity is seen on amphibian crest cells (Krotoski and Bronner-Fraser, 1990). It is likely that the re-

D.F. New green and S.S. Tan

ceptor for the connecting sequence-1 site is also a ySj-integrin (Dufour et al., 1988; Humphries et al, 1988). The functional importance of such integrins has been demonstrated by perturbation using antibodies (Bronner-Fraser, 1985) and anti-sense oligonucleotides (Lallier and Bronner-Fraser, 1992) in tissue culture. The responses to the heparinbinding site is presumably mediated by heparinsulfate (HS)-PG at the cell surface. The presence of FN in the embryonic ECM, along with the presence of relevant receptors on crest cells, suggests that FN distribution influences the migration pathways of crest cells, but its breadth of distribution is far wider than the actual pathways. Splicing out (Komblihtt et al., 1985) or masking in vivo (Brauer and Markwald, 1988) may mean that the real adhesive choices on FN in vivo are more limited than the distribution of the whole molecule suggests. The effect of FN on crest cell behavior in vitro is also consistent with it playing a role in vivo. In addition, the inability of many crest cells to produce FN has been interpreted as a means of ensuring greater accuracy in the following of FN pathways (Newgreen and Thiery, 1980). Direct attempts at proving this relationship in vivo have involved the use of agents known to perturb cell-FN interactions. Introduction of antibodies that block the adhesive sites of FN (Poole and Thiery, 1986), or its major receptor (Bronner-Fraser, 1985, 1986b) and competitive inhibition with soluble RGDS-containing peptides (Boucaut et al., 1984) all produced a profound localized diminution of crest outgrowth when microinjected into the cranial crest pathways of the chick embryo. The receptor-blocking antibodies and the RGDS competition approaches cannot be considered specific for interactions with FN. Nevertheless, combined with the anti-FN results, these indicate that FN is a necessary component in avian cranial crest migration. When the same reagents were applied at trunk levels, in contrast, inhibition of migration was, at most, slight, although each agent was effective on trunk crest cells in culture. If FN is involved in defining crest migration pathways at trunk levels, it may be one of several redundant adhesion components. The placement of FN as one of a spectrum of migratory cues is also

53

suggested by the observations that in situ it is found before and after, as well as during, crest cell migration, that it also occurs at high levels in ECM not occupied by crest cells (Newgreen and Thiery, 1980) and that crest cells can move from areas rich in FN to FN-poor regions (Rickmann et al., 1985). 6.2. Laminin Laminin (LN), another adhesive glycoprotein, is usually isolated from the mouse EHS-sarcoma, where it consists of an A-chain (about 400 kDa) and Bl and B2 chains (each about 200 kDa) linked to form a cruciform molecule. Another molecule, entactin, is non-covalently linked to LN. This tumor LN configuration is not the only one possible and LN in different tissues and stages of development may show deletions of chains or other alterations (Cooper and MacQueen, 1983). Immunolocalization reveals LN as being concentrated in basal laminae around embryonic epithelial organs, but immunoreactivity can also be found interstitially in the mesenchymal compartments (Rogers et al., 1986; Simon-Assman et al., 1988) around the time and position of crest cell migration. Because of its placement, it had been supposed that epithelia were responsible for LN synthesis, but in the gut, at least, the source of epithelial LN is the mesenchyme cells (SimonAssman etal., 1988, 1990). In vitro assays showed that LN is a favorable substrate for crest cell migration (Newgreen, 1984; Erickson, 1985; cf. Rovasio et al., 1983) and avian crest cells readily cross a border between FN and LN in short-term migration assays (Newgreen, 1984). In longer term dispersion assays, however, the spreading of the crest population is less on LN than on FN for avian (Ferris et al., 1989) and, more markedly, for mouse crest cells (Boisseau and Simonneau, 1989). Unlike crest cell responses to FN, which reach a plateau with increasing substrate density, the response of amphibian and avian crest cells to LN shows an optimum, past which displacement is reduced (Ferris and Johansson, 1987; Ferris et al, 1989). Dispersal of crest cells on LN is also increased when it was presented in association with entactin, or when it is co-absorbed

54

with a HS-PG or collagen type IV, molecules normally distributed with LN in the basal lamina. LN has many sites potentially available for interaction with cells (see Akiyama et al., 1990). Several proteolytic cleavage domains have been tested to identify the regions favored by crest cells, using dispersion assays along with peptide and LN-ligand inhibition tests (Bilozur and Hay, 1989; Penis et al., 1989). The major responses to adsorbed LN substrates were directed to the E8 fragment, but the terminal globular heparinbinding subfragment of E8, termed E3, did not mediate this. The El fragment, which possesses the adhesive peptide sequence YIGSR, was less effective and peptide inhibition suggested that the effect of El depended on an RGDS-like sequence and not the YIGSR sequence (Perris et al., 1989). In contrast, crest cell migration in a LN-rich 3dimensional basement membrane gel was inhibited by YIGSR (Bilozur and Hay, 1989), suggesting that this domain is recognized in this spatial context. Cell surface receptors to LN also show considerable complexity (Lallier et al., 1992). Antibodies to the avian integrin jSj chain were highly effective in inhibiting crest cell responses to LN (BronnerFraser, 1985). Integrin-type LN receptors are known in other systems (see Akiyama et al., 1990). Responses to LN in vitro, under some conditions, were also inhibited by the monoclonal antibody HNK-1, although responses to FN were unaffected, probably due to an HNK-1 site on an a-integrin chain (Lallier and Bronner-Fraser, 1991). These results point to at least two ^ j integrin receptors that bind to LN with different affinities and under different conditions. Modification or competitive inhibition of )3-l,4-galactosyl transferase molecules of 67 and 77 kDa (Hathaway and Shur, 1992) reduced migration but not initial adhesion of crest cells on LN substrates, but were ineffective on FN. In addition, the magnitude of the migratory response to LN was increased by addition of an exogenous galactosyl transferase catalytic substrate (Runyan et al., 1986). These results in vitro point to the involvement of a cellsurface galactosyl transferase in the interaction of crest cells with A^-acetylglucosamine groups on

Adhesion molecules in neural crest development

LN, most likely on the E8 fragment (Begovac et al., 1991). Interestingly, the inhibitory effects were chiefly felt by the cranial crest cell population rather than the trunk-level cells and of the cranial crest cells, those that migrated first were mainly affected. This corresponds to the FN-producing crest cells (Newgreen and Thiery, 1980), which may be the mesenchyme-producing crest subpopulation, as opposed to the neurogenic subpopulation. The action of the YIGSR-recognizing LN receptor (typically a molecule of 67-70 kDa) was not detected on adsorbed LN, but may operate in the basement membrane gels (Bilozur and Hay, 1989; cf. Perris et al., 1989). This inconsistency in results between adsorbed and gelled LN requires clarification, but the simplest explanation is that the two modes of presentation of LN conformationally expose different domains. Which is closer to reality is unknown. The role of LN in vivo in avian embryos has been probed using microinjection of a monoclonal antibody termed INO (inhibitor of neurite outgrowth), which recognizes a LN-HS-PG complex (Bronner-Fraser and Lallier, 1988). Unlike antibodies against fii integrins, which block both FN and LN receptors and which restricted overall crest outgrowth at cranial levels (see above) (BronnerFraser, 1985, 1986b), INO permitted most crest cells to move distally, but retarded a subgroup that aggregated ectopically next to the neural tube, or within the lumen of the neural tube. The window of sensitivity to INO was brief and injections after the 9-somite stage were ineffective: this corresponds to the end of the period of emigration from the neural tube. To confuse interpretation, antibodies to LN alone, which are effective in vitro, had no effect in in vivo applications. Injection of HNK-1 monoclonal antibody, which recognizes crest cell surface molecules (including LN receptors) rather than LN, gave identical results to INO (Bronner-Frasers, 1987). Interference with cell surface )3-1,4 galactosyltransferase by microinjection of antibodies into the avian neural crest pathways also interfered with neural crest cell migration, although again only when administered before the 10-somite stage (Hathaway and Shur, 1992). It should be noted that embryos were more

D.F, Newgreen and S.S. Tan

likely to show reduction of crest cell emigration on injection of whole antibody compared with Fab (67% abnormal crest emigration versus 17%, respectively). This was not due to a generalized ineffectiveness of the particular Fab, since unlike the whole antibody, it was highly effective in producing neural tube dysmorphogenesis. These differing results raise the possibility that the divalent antibodies were hampering migration by artifactually promoting cell-cell aggregation via cross-linking of molecules on adjacent cells rather than by interfering specifically with crest cell interactions with LN. Similar approaches using INO and HNK-1 at trunk levels did not alter crest cell distribution but this could have been due to leakage or to insufficient antibody, since, for technical reasons, injections were smaller than at cranial levels. In addition, the injections were made into the 10th last somite, when the first crest cells have already reached the dorsal aorta (Newgreen et al., 1990); this may have been too late to affect crest cell distribution. In the succeeding chapter, Nurcombe discusses the role of laminin in more detail. 6.3. Vitronectin Vitronectin is an adhesive glycoprotein found in early avian embryos (Delannet et al., 1994). It is distributed quite differently to FN, occurring on or at the surface of most cells, rather than in the ECM; in addition, its expression is reduced after neural crest cells have completed their migration. The responses of neural crest cells to VN are directed against the RGD sequence, but appear from tissue culture studies to be complex. The aV integrin appears to combine with ^ 1 , 3 and 5 integrins to give three active VN receptors, and it has been suggested that the different combinations differentially affect cell adhesion and cell migration (Delannet et al., 1994). The role of VN, however, is still obscure and no in vivo testing has been performed. 6.4. Collagens Collagens are major components of embryonic

55

as well as adult ECM. Collagen type IV in avian embryos is found with immuno-electron microscopy in typical basal laminae, but it also occurs in interstitial bodies and associated with interstitial fibrils (Martins-Green and Erickson, 1987; Martins-Green, 1988). It is widely distributed in avian and mouse embryos (Sternberg and Kimber, 1986; Duband and Thierye, 1987) and extensively overlaps the distribution of FN and, more exactly, LN. Interstitial collagen types I and III can be found at stages of crest cell migration, overlapping FN in distribution. In avian embryos, collagen type I is the predominant species (Duband and Thiery, 1987) but, in mice, collagen type III is probably the major early collagen (Leivo et al., 1980). Collagen type IV substrates in vitro permit dispersion of crest cells to a similar extent as does LN (Ferris et al., 1989), but this result could be influenced by the neural anlage conditioning the substrate. A variety of collagen type IV-binding cell surface molecules are known, including ySj integrins (Akiyama et al., 1990) and these are present on crest cells (Lallier et al., 1992). Collagen type I substrates in vitro support rapid attachment of avian crest cells, independently of FN (Newgreen, 1982). Despite their adhesion, migration of crest cells is delayed and they are less flattened on fibrillar collagen (Newgreen et al., 1982). Mouse crest cells fail entirely to migrate on substrates of fibrillar collagen type I (Boisseau and Simonneau, 1989). In competition experiments, avian crest cells clearly preferred tracks of FN in comparison with tracks of collagen type I (Newgreen, 1992). The collagens bind to other ECM molecules (such as FN); therefore, it is, at present, unclear whether in vitro migration of avian crest cells on collagen type I (as opposed to adhesion) is caused by the collagen itself or by molecules adsorbed to the collagen. The poor locomotive responses of crest cells to collagen in comparison with FN have been interpreted as indicating that collagen type I is a structural matrix component whose function (if any) in the control of migration is indirect, by acting as a support for biologically active molecules and. as a regulator of extracellular space. Indeed, suppression of collagen type I in early mouse embryos by silencing the

56

gene does not obviously affect organogenesis of crest or other systems (Schnieke et al., 1983). Many collagen receptors have been characterized, including integrins (Ogle et al., 1989; Akiyama et al., 1990) and antibody perturbation experiments indicate that neural crest cells possess an a^jSi, integrin type receptor to collagen I (Lallier et al., 1992). 6.5. Cytotactin Cytotactin (CT), also known as tenascin, myotendinous antigen, hexabrachion, GMEM or Jl200/220 (see Erickson and Bourdon, 1989), is a large glycoprotein consisting of six similar arms (molecular mass each arm, approx. 200 kDa) joined at a central globular domain. The appearance of CT in embryonic tissues is much more tightly regulated than the above-mentioned ECM molecules. It is chiefly restricted to axial tissues, with a spatiotemporal pattern more closely related to crest cell distribution at trunk and cranial levels in avian embryos (Crossin et al., 1986; Tan et al., 1987; Bronner-Fraser, 1988; Mackie et al., 1988), mammals (Mackie et al., 1988) and amphibians (Epperlein et al., 1988). For this reason, intense interest has focussed on this molecule as a positive regulator of crest cell migration. More recent reevaluations of the precise relationship of CT to crest cell migration in avian embryos (Stem et al., 1989; Newgreen et al., 1990) has suggested a diametrically different interpretation. For example, for early avian somites, CT was maximally expressed along the caudal border and in the caudal half, regions avoided by crest cells (Stem et al., 1989; Newgreen et al., 1990). Later, this distribution is reversed CT becoming restricted largely to the rostral half of the sclerotome (Tan et al., 1987; Mackie et al., 1988), where crest cells are also found (Rickmann et al., 1985; Bronner-Fraser, 1986a; Loring and Erickson, 1987). However, since crest cells had already entered the rostral half of the somite prior to the appearance of CT, the expression of this molecule subsequently in the rostral somite is functionally more related to the cessation, or inhibition, of migration (Stem et al., 1989; Newgreen et al., 1990). This view is easier

Adhesion molecules in neural crest development

to correlate with the results of in vitro crest cell migration assays in which CT itself was a poor substrate for locomotion (Tan et al., 1987; Mackie et al., 1988) and could reduce migration on substrates of FN (Tan et al., 1987; cf. Halfter et al., 1989). Ablation of the dorsal neural tube and crest by Stem et al. (1989) produced somites that not only lacked crest cells, but also failed to develop the restriction of CT distribution to the rostral half of the somites. Using CT antibodies for immunoaffinity chromatography followed by polyacrylamide gel electrophoresis, young somites, prior to crest cell entry, yielded only molecules of molecular mass far below that of authentic monomeric CT, but more appropriately sized molecules (200 kDa) appeared in older somites. More importantly, older somites from dorsal neural tubeablated embryos also provided only the low molecular mass range of molecules. This led Stem et al. (1989) to propose that both the appearance of high molecular mass CT and its rostral hemisegmental localization in the somites required an interaction between somite cells and crest cells; given the effect of CT on crest cell locomotion, this acted to limit further crest cell input. This hypothesis neatly explained the result of Weston and Butler (1966), in which crest cell migration into older somites was curtailed. As a mechanism. Stem et al. (1989) tentatively suggested that somites and crest cells could each supply precursors to build up full-sized CT. These results were re-examined by Tan et al. (1991), who found that CT sometimes failed to become restricted to the rostral somite-half even when neural crest cells were present (e.g. in shamoperated embryos) and sometimes became normally restricted to the rostral somite half even when neural crest cells were absent after ablation. Moreover, immunoblotting showed that CT was always of large size (approximately 200 kDa) in young somites, in older sham-operated somites and in older somites from ablated embryos: in short, no small size CT-immunotypes were seen under any conditions. It is, therefore, likely that the spatial abnormalities of CT expression observed by Stem et al. (1989) were due to a non-specific effect after

D.F. New green and S.S. Tan

embryonic surgery, rather than to a specific response to the loss of crest cells. The biochemical abnormalities could result from technical difficulties in handling small samples on large absorptive surfaces, compounded by the specifically immunoabsorbed CT being itself a ligand for other molecules. This could result in a collection of heterologous molecules, with the specifically sought molecule (i.e. CT) at undetectable levels (Tan et al., 1991). The function of CT in vivo has been probed, somewhat inconclusively, by injection of antibodies (Bronner-Fraser, 1988) in avian embryo, but experiments testing the effect of the antibodies on crest cells in vitro have not been reported. The results in vivo were very similar to those induced by injection of INO and HNK-1 monoclonal antibodies. At cranial levels, proximal ectopic groups of crest cells (in this case in or on the epidermal ectoderm) occurred with injections prior to the 9somite stage, but no serious retardation of distal crest outgrowth was noted. Similar injections at trunk levels produced no observable effects on crest cell outgrowth. Another strategy produced knockout mice with null mutations of the CT gene (Saga et al., 1992). Since the mice appeared normal, it would appear that CT does not have an indispensable role in development. 6.6. Proteoglycans Proteoglycans (PG) consist of glycosaminoglycan (GAG) side-chains linked to a core protein. These molecules gain diversity from various core proteins, different types of GAGs (e.g. CS, dermatan sulfate, keratan sulfate, HS) of different chain lengths, chain numbers and degrees of sulfation. As yet, little information has been gained to tie these sub-groups to crest cell migration, but the major embryonic sulfated molecule, CS-PG, has been linked spatiotemporally to crest cell migration (Newgreen and Thiery, 1980). In avian, mammalian and amphibian embryos, PG molecules, chiefly CS-PG, have been localized with positively charged dyes, such as alcian blue (Pintar, 1978; Derby, 1978; Erickson and Weston, 1983; Tucker, 1986) and ruthenium red (Hay,

57

1978; Lofberg et al., 1980; Newgreen et al., 1982), immunocytochemically (Morriss-Kay and Tuckett, 1989; Poelmann et al., 1990; Perris et al., 1991) and using metabolic sulfate labelling (Pratt et al, 1975; Manasek, 1975; Brauer and Markwald, 1988). These studies show, in avian and amphibian embryos at least, that although CS-PG is widely distributed, there is a strong correlation between local maxima and areas of ECM not occupied by crest cells. Tests of the function of CS-PG have been made in tissue culture, using molecules derived not from the appropriate embryos, but from more accessible sources, bovine nasal cartilage (Newgreen, 1982; Perris and Johansson, 1987, 1990) and avian brain (Tan et al., 1987). The bovine nasal cartilage CSPG is a large molecule (molecular mass about 2.5 X 10^) with fully sulfated CS and keratan sulfate chains, capable of binding to collagen and aggregating specifically with HA. The avian brain CS-PG is also large (molecular mass approximately 1 X 10^) and is a specific ligand for CT (Hoffman and Edelman, 1987). The properties of the CS-PG in young embryonic tissues is not known, but the ultrastructure suggests that it binds HA (Hay, 1978; Newgreen et al., 1982; Tennyson etal., 1990). In vitro tests on plane substrates containing various ECM molecules showed that crest cell adhesion and migration are substantially reduced by CS-PG in the culture medium. In the case of collagen type I, this is mediated by CS-PG attaching to it. On ECM molecules, which bind little or no CSPG, it interferes with migration by linking to cell surface HA (Perris and Johansson, 1990). This broad inhibitory spectrum suggests that the effects of CS-PG are non-specific. Experiments using artificial GAG-protein molecules (Yamagata et al., 1989) show that the inhibition is mediated by the cluster of large, negatively charged GAGs sterically blockading interactions between adhesion molecules in the ECM and their cell surface receptors. The binding properties of the CS-PG molecule as a whole probably serve to localize it either at the ECM and/or the cell surface. The report that CS itself interferes with migration (Newgreen et al., 1982) may have depended on the complex

58

ECM substrate used in these assays being able to bind CS, even when not in the PG form. In that example, a further complication was noted in that moderately undersulfated CS was more inhibitory than either fully sulfated or fully unsulfated CS. By contrast, crest cells cultured in threedimensional collagen gels showed increased speed of movement when CS-PG was included (Tucker and Erickson, 1984). It was suggested that, in this case, CS-PG facilitated movement by increasing the cell-collagen de-adhesion cycle. In an attempt to study more closely the role of embryonic PG, avian crest cell migration assays have been performed in the presence of the notochord, a producer of various ECM molecules including CS-PG (Newgreen et al., 1986). Crest cells in these assays avoided the region close to the notochord, as they do in vivo. This avoidance was abolished by chondroitin lyases, but was unaffected by specific hyaluronidase digestion, suggesting that the CS moiety of notochordal PGs was involved and that binding to cell-surface HA was not necessary under conditions in which binding to the substrate was likely. Inhibition of sclerotomal mesenchyme cell movement near the notochord was not evident, indicating an unexplained cell type specificity. Experiments using chondroitin lyase have been carried out in vivo on amphibian (Tucker, 1986) and rat embryos (Morris-Kay and Tuckett, 1989). Treatment of Xenopus embryos resulted in the appearance of crest-derived chromatophores under the epidermis, using a normally forbidden CS-PGrich pathway. This indicates that CS-PG prevented migration in this region, either by restricting space (in opposition to HA) and/or by inhibiting cell adhesion (Tucker, 1986). In the experiments with rat embryos, reduction of CS caused several abnormalities, one of which was a partial inhibition of crest cell emigration from the cranial neuroepithelium. Given that CS-PG normally becomes very widely distributed, but appears first in the neuroepithelium at the presumptive crest-region, the interpretation of these results was that CS-PG facilitates crest migration by aiding cell-cell deadhesion in the initial separation of crest cells (see Nichols, 1987).

Adhesion molecules in neural crest development

The patterns of CS-PG distribution, its effect on crest cells in vitro and the responses to CS-PG reduction in vivo have prompted apparently contradictory views on its role, although all arguments hinge on its reducing cell adhesion. This conflict can be resolved by examining the context of proposed CS-PG action. Since cell movement declines above and below an optimal adhesive level, then an overall decrease in adhesion, via CS-PG, from a sub-optimally high level could favor cell movement. This may occur in the collagen gel cultures of avian crest cells (Tucker and Erickson, 1984) and in cranial levels of rat embryos (MorrisKay and Tuckett, 1989). On the other hand, when encountering landscape of higher and lower adhesivity (mediated by lower and higher CS-PG levels, respectively), cells will not enter or remain in the lower adhesivity zone, even if it is closer to the optimum adhesive value for cell movement. This may be the case when crest cells encounter CS-PG maxima perichordally (Newgreen et al., 1986) and subepidermally (Tucker, 1986). Thus, CS-PG can both facilitate cell movement in general and inhibit cell movement into discrete regions. 6.7. Hyaluronan Hyaluronan (HA) was the first ECM component to be related to crest cell migration (Pratt et al, 1975; Derby, 1978; Pintar, 1978). HA is a nonsulfated GAG of large but variable size, which, unlike the PGs, is not necessarily covalently bound to protein. Its retention in aldehyde-fixed tissues is capricious (Singley and Solurshs, 1980; Tuckett and Morriss-Kay, 1988). but, with adequate fixations it can be stained at light and electron microscope-level using cationic dyes as for PGs. Fortunately, its contribution can be separated from that of PG by the use of the highly specific Streptomyces hyaluronidase, or by staining at critical electrolyte concentrations (see Derby, 1978; Pintar, 1978). More recently, the use of specific ligands for HA have also enabled its localization in tissues (Poelmann et al., 1990). At the ultrastructural level, HA appears as a network of fibrils about 3 nm in diameter (Hay, 1978; Newgreen et al., 1982; Tennyson et

59

D.F. Newgreen and S.S. Tan

al, 1990). An increase in staining is frequently noted in regions of crest cell migration, and, in vitro, crest cells can synthesize HA (Pintar, 1978). However, the precise distribution and timing of HA in vivo suggests that most is produced by the cells adjacent to crest migration routes (Pratt et al., 1975). Estimates based on densitometry place the maximum HA concentration in mouse crest cell pathways as up to 20 mg/cm^ (Derby, 1978). The swelling of gross intercellular space prior to or concomitant with avian crest cell migration is dependent on HA, since enzyme degradation causes spatial reduction (Anderson and Meier, 1982; Schoenwolf and Fisher, 1983). Despite this, crest cell migration, at least in its early phases, is not prevented at either cranial or trunk levels by this treatment, indicating that even the residual space is sufficient for cell migration (see also Newgreen, 1989). In vitro assays show that HA is not a substrate for crest cell migration (Newgreen et al., 1982) but, over a certain concentration range, it aids migration in collagen gels (Tucker and Erickson, 1984). The most likely explanation for this is that it expands the interstices of the collagen fibril network, effectively increasing the microspaces that facilitate the insinuation of cells (see Newgreen, 1989). Cell surface HA may also modulate cell movement by acting as an anchor for CS-PG, which directly interferes with cell adhesion (Perris and Johansson, 1990). More recently, a series of studies have implicated HA directly in the promotion of active cell movement, via interaction with specific cell surface receptors (Lacy and Underbill, 1987; Turley, 1989). It may well be that this is an additional function of HA, as a signal for crest cell migration. Addition of HA also altered the number of crest cells leaving neural anlagen in culture (LuckenbillEdds and Carrington, 1988). Since low concentrations reduced the numbers and high concentrations increased them, it was proposed that cell-cell crosslinking via HA at low concentrations impeded outgrowth, but swamping of HA-binding sites at high concentrations favored cell detachment from the neural explant.

6,8, Cell adhesion molecules Cell adhesion molecules (CAMs; see Section 3.4) could also be used to permit crest cells to move on the surface of other cells; use of the cell surface as a migratory substrate occurs in the development of the CNS (e.g. Lindner et al., 1983; Rakic, 1985). Such a mechanism has received less interest for crest cells, where interest has focussed chiefly on cell-ECM interactions. Indeed, most work on the involvement of CAMs in crest morphogenesis has emphasized the loss of such molecules at the onset of migration (see Section 3.4) and their reappearance at the end of migration (see Section 7; Thiery et al., 1982; Duband et al., 1988b; Lallier and Bronner-Fraser, 1988). Nevertheless, electron microscopy reveals that migrating crest cells show 20-nm membrane apposition, not only with other crest cells (Bancroft and Bellairs, 1976; Tosnev. 1978, 1982, 1988) but also with somite cells (Newgreen et al., 1990). This appearance in migratory cells is consistent with the engagement of cell surface molecules between the contacting cells (see Springer, 1990). Crest cells entering the rostral half of the somites might develop this pattern by avoidance of putative adhesion-inhibitory molecules (e.g. CSPG; CT) in the caudal half-somite (Stem et al., 1986, 1989; Tan et al., 1987). However, the observation that the crest cells entering the rostral somite leave the perisomitic ECM, which is rich in adhesive ECM molecules (Newgreen and Thiery, 1980; Rickmann et al., 1985), suggests that the preference for the rostral somite-half is also based on a positive affinity rather than just the avoidance of the caudal half. As yet, the nature of the interaction between crest cells and the rostral halfsomite is unknown, but heterotypic cell-cell contacts appear to be involved (Newgreen et al., 1990). In quail embryos, the HNK-1 epitope appears on the surface of somite cells shortly before infiltration by crest cells and this disappears as the crest cells form homotypic assemblies within the somitic realm (Newgreen et al., 1990). This exact spatiotemporal congruence of the HNK-1 epitope on somite cells with the phase of cell contact of crest and somite cells, along with the association

60

of this epitope with many cell adhesion molecules (Kruse et al., 1984; Keilhauer et al., 1985; Riopeile et al., 1986; Cole and Schachner, 1987; Kumemund et al., 1988) suggests that HNK-1 may mark a molecule on the somites that mediates their interaction with crest cells. The sensitivity of this epitope to lipophilic solvents (Newgreen et al., 1990) suggests that its parent molecule(s) might be a glycolipid (see Roberts and Ginsburg, 1988) or a glycoprotein not firmly integrated with submembranous proteins (such as those with phosphotidylinositol linkage; see Geiger, 1989). The putative complementary molecule on the surface of crest cells is unknown. The possibility exists, therefore, that the preference of crest cells for the rostral half-somite is ensured by positive affinity to this region based on cell-cell adhesions, as well as by an inhibitory effect in the caudal half-somite due, in part, to adhesion inhibitory molecules, such as CT and CS-PG. In summary, the basic concept of the course of neural crest cell migration being controlled by adhesive interactions is now well established. However, the molecular events are likely to be complicated, featuring a balance between adhesive molecules (e.g. FN) and anti-adhesive molecules (e.g. CS-PG). Numerous other molecules in the microenvironment, which are recognized by numerous receptors on the cell surface, have not had precise functions assigned to them, but they could play a part in the fine control of migration in in vivo situations. 7. Adhesions in the cessation of migration The cessation of neural crest cell migration has been the least well-documented phase of this morphogenetic process, with few in vitro or in vivo functional studies. Hypothetically, Weston (1970) has suggested that this phase arises from an increase in the ratio of adhesion between crest cells versus adhesion between crest cells and their microenvironment. Correlative studies using antibodies to A-CAM and N-CAM (Duband et al, 1988b; Lallier and Bronner-Fraser, 1988; Akitaya and Bronner-Fraser, 1992) suggest that these molecules (which are expressed prior to migration;

Adhesion molecules in neural crest development

Section 3.4) are re-expressed by crest cells during re-aggregation to form peripheral ganglia, consistent with an elevation of cell-cell adhesion. NCAM definitely re-appears too late to be important in initiation of ganglionation and, although this is not certain, A-CAM may also occur slightly after the initial accumulation of cells in the ganglionic position. This raises the possibility of the existence of further CAMs that initiate the passage of migratory neural crest cells to sedentary ganglion cells, this process being amplified in succession by increases in A-CAM and N-CAM. At the same time, adhesive molecules, such as FN, in the microenvironment show decreased immunoreactivity (Newgreen and Thiery, 1980), while anti-adhesive molecules, like CS-PG, are increasingly expressed (Perris et al., 1991). In addition, ECM-receptor molecules are down-regulated in aggregated neural crest cells (Duband et al., 1986; Krotoski et al, 1986). All these observations are broadly consistent with a reduction of the adhesion between neural crest cells and their microenvironment playing a part in the cessation of migration. Although the groundwork for the role of adhesion in the cessation of neural crest cell migration has been laid, further, more detailed correlational observations are needed, along with the construction of functional in vitro and in vivo tests. References Akitaya, T. and Bronner-Fraser, M. (1992) Expression of cell adhesion molecules during initiation and cessation of neural crest cell migration. Dev. Dyn. 194: 20. Akiyama, S.K., Nagata, K. and Yamada, K.M. (1990) Cell surface receptors for extracellular matrix components. Biochim. Biophys. Acta 1031: 91-110. Anderson, C.B. and Mieier, S. (1982) Effect of hyaluronidase treatment on the distribution of cranial neural crest cells in the chick embryo. /. Exp, Zool 211: 329-335. Ayer-le Lievre, C.S. and Le Douarin, N.M. (1982) The early development of cranial sensory ganglia and the potentialities of their component cells studied in quail-chick chimeras. Dev. 5/o/. 94: 291-310. Bancroft, M. and Bellairs, R. (1976) The neural crest cells of the trunk region of the chick embryo studied by SEM and TEM. Zoon 4: 73-85. Begovac, P.C, Hall, D.E. and Shur, B.D. (1991) Laminin fragment E8 mediates PC 12 cell neurite outgrowth by

D.F. New green and S.S. Tan binding to cell surface p\A galactosyltransferase. J. Cell Biol. 113:637-644. Bilozur, M.E. and Hay, E.D. (1989) Cell migration into neural tube lumen provides evidence for the "fixed cortex" theory of cell motility. Cell Motil. Cytoskeleton 14: 469^84. Boisseau, S. and Simonneau, M. (1989) Mammalian neuronal differentiation: early expression of a neuronal phenotype from mouse neural crest cells in a chemically defined culture medium. Development 106: 665-674. Boucaut, J.C, Darribere, T., Poole, T.J., Aoyama, H., Yamada, D.M. and Thiery, J.P. (1984) Biologically active peptides as probes of embryonic development: a competitive peptide inhibitor of fibronectin function inhibits gastrulation in amphibian embryos and neural crest cell migration in avian embryos. /. Cell Biol. 99: 1822-1830. Brauer, P.R. and Markwald, R.R. (1988) Specific configurations of fibronectin-containing particles correlate with pathways taken by neural crest cells at two axial levels. Anat. Rec. 111'. 69-82. Bronner, M.E. and Cohen, A.M. (1979) Migratory patterns of cloned neural crest melanocytes injected into host chicken embryos. Proc. Natl. Acad. Sci. USA 76: 1843-1848. Bronner-Fraser, M. (1985) Alterations in neural crest migration by a monoclonal antibody that affects cell adhesion. J. Cell Biol. 101:610-617. Bronner-Fraser, M. (1986a) Analysis of the early stages of trunk neural crest migration in avian embryos using monoclonal antibody HNK-I. Dev. Biol. 115: 44-55. Bronner-Fraser, M. (1986b) An antibody to a receptor for fibronectin and laminin perturbs cranial neural crest development in vivo. Dev. Biol. 117: 528-536. Bronner-Fraser, M. (1987) Perturbation of cranial neural crest migration by the HNK-1 antibody. Dev. Biol. 123: 321331. Bronner-Fraser, M. (1988) Distribution and function of tenascin during cranial neural crest development in the chick. /. Neurosci. Res. 21: 135-147. Bronner-Fraser, M. and Laller, T. (1988) A monoclonal antibody against a laminin-heparan sulfate proteoglycan complex perturbs cranial neural crest migration in vivo. J. Cell Biol. 106: 1321-1329. Bronner-Fraser, M. and Stern, C. (1991) Effects of mesodermal tissues on avian neural crest cell migration. Dev. Biol. 143: 213-217. Cole, G.J. and Schachner, M. (1987) Localization of the L2 monoclonal antibody binding site on chicken neural cell adhesion molecule (NCAM) and evidence for its role in NCAM-mediated cell adhesion. Neurosci. Lett. 78: 227232. Cooper, A.R. and MacQueen, H.A. (1983) Subunits of laminin are differentially synthesized in mouse eggs and embryos. Dev. Biol 96: 461^11. Crossin, K.L., Hoffman, S., Grumet, M., Thiery, J.P. and Edelman, G.M. (1986) Site-restricted expression of cytotactin during development of the chicken embryo. /. Cell Biol. 102: 1917-1930.

61 Delannet, M. and Duband, J.-L. (1992) Transforming growth factor-^ control of cell-substratum adhesion during avian neural crest cell emigration in vitro. Development 116: 275-287. Delannet, M., Martin, F., Bossy, B., Cheresh, D.A., Reichardt, L.F. and Duband, J.-L. (1994) Specific roles of the aV/31, aVj53 and aVp5 integrins in avian neural crest adhesion and migration on vitronectin. Development 120: 26872702. Derby, M.A. (1978) Analysis of glycosaminoglycans within the extracellular environments encountered by migrating neural crest cells. Dev. Biol. 66: 321-336. Detwiler, S.R. (1937) Observations upon the migration of neural crest cells and upon the development of the spinal ganglia and vertebral arches in Ambystoma. Am. J. Anat. 61: 63-94. Duband, J.L. and Thiery, LP. (1982) Appearance and distribution of fibronectin during chick embryo gastrulation and neurulation. Dev. Biol. 94: 337-350. Duband, J.L. and Thiery, J.P. (1987) Distribution o/laminin and collagens during avian neural crest development. Development 101: 461-478. Duband, J.L. and Thiery, J.P. (1990) Spatial and temporal distribution of vinculin and talin in migrating avian neural crest cells and their derivatives. Development 108: 421434. Duband, J.L., Tucker, G.C., Poole, T.J., Vincent, M., Aoyama, H. and Thiery, J.P. (1985) How do the migratory and adhesive properties of the neural Crest govern ganglion formation in the avian peripheral nervous system? J. Cell. Biochem. 11'. 189-203. Duband, J.L., Rocher, S., Chen, W.T., Yamada, K.M. and Thiery, J.P. (1986) Cell adhesion and migration in the early vertebrate embryo: location and possible role of the putative fibronectin receptor complex. J. Cell Biol 102: 160178. Duband, J.-L., Nuckolls, G.H., Ishihara, A., Hasegawa, T., Yamada, K.M., Thiery, J.-P. and Jacobson, K. (1988a) Fibronectin receptor exhibits high lateral mobility in embryonic locomoting cells but is immobile in focal contacts and fibrillar streaks in stationary cells. J. Cell Biol 107: 13851396. Duband, J.-L., Volberg. T., Sabanay, I., Thiery, J.P. and Geiger, B. (1988b) Spatial and temporal distribution of the adherens junction-associated adhesion molecule A-CAM during avian embryogenesis. Development 103: 325-344. Dufour, S., Duband, J.L., Humphries, M.J., Obara, M., Yamada, K.M. and Thiery, J.P. (1988) Attachment, spreading and locomotion of avian neural crest cells are mediated by multiple adhesion sites on fibronectin molecules. EMBO J. 7:2661-2671. Epperlein, H.H.E., Halfter, W. and Tucker, R.P. (1988) The distribution of fibronectin and tenascin along migratory pathways of the neural crest in the trunk of amphibian embryos. Development 103: 743-756. Erickson, C.A. (1985) Control of neural crest cell dispersion

62 in the trunk of the avian embryo. Dev. Biol. HI: 138157. Erickson, C.A. and Isseroff, R.R. (1989) Plasminogen activator is associated with neural crest cell motility in tissue culture. J. Exp. Zool. 251: 123-133. Erickson, C.A. and Turley, E.A. (1983) Substrata formed by combinations of extracellular matrix components alter neural crest cell motility in vitro. J. Cell. Sci. 61: 299-323. Erickson, C.A. and Weston, J.A. (1983) An SEM analysis of neural crest cell migration in the mouse. J. Embryol. Exp. Morphol. 14: 91-nS. Erickson, C.A., Tosney, K.W. and Weston, J.A. (1980) Analysis of migratory behavior of neural crest and fibroblastic cells in embryonic tissues. Dev. Biol. 77: 142-156. Erickson, H.P. and Bourdon, M.A. (1989) Tenascin: an extracellular matrix protein prominent in specialized embryonic tissues and tumors. A. Rev. Cell. Biol. 5: 71-92. Ffrench-Constant, C. (1988) Patterns of fibronectin gene expression and splicing during cell migration in chicken embryos. Development 104: 369-382. Fisher, M. and Solursh, M. (1979a) The influence of the substratum on mesenchyme spreading in vitro. Exp. Cell Res. 123: 1-14. Fisher, M. and Solursh, M. (1979b) The influence of local environments on the organization of mesenchyme cells. J. Embryol. Exp. Morphol. 49: 295-306. Geiger, B. (1989) Cytoskeleton-associated cell contacts. Curr. Opin. Cell Biol. \: 103-109. Greenberg, J.H., Seppa, S., Seppa, H. and Hewitt, T. (1981) Role of collagen and fibronectin in neural crest cell adhesion and migration. Dev. Biol. 87: 259-266. Halfter, W., Chiquet-Ehrismann, R. and Tucker, R.P. (1989) The effect of tenascin and embryonic basal lamina on the behavior and morphology of neural crest cells in vitro. Dev. Biol. 132:14-25. Harris, A.K., Stopaic, D. and Wild, P. (1981) Fibroblast traction as a mechanism for collagen morphogenesis. Nature 290:249-251. Hathaway, H.J. and Shur, B.D. (1992) Cell surface P 1,4galactosyltransferase functions during neural crest cell migration and neurulation in vivo. J. Cell Biol. 117: 369-382. Hay, E.D. (1978) Fine structure of embryonic matrices and their relation to the cell surface in ruthenium-red fixed tissue. Growth 42: 399-423. Hoffman, S. and Edelman, G.M. (1987) A proteoglycan with HNK-I antigenic determinants is a neuron-associated ligand for cytotactin. Proc. Natl. Acad. Sci. USA 84: 2523-2521. Horstadius, S. (1950) The Neural Crest: Its Properties and Derivatives in the Light of Experimental Research. Oxford University Press, London. Humphries, M.J., Akiyama, S.K., Komoriya, A., Olden, K. and Yamada, K.M. (1988) Neurite extension of chicken peripheral nervous system neurons on fibronectin: relative importance of specific adhesion sites in the central cellbinding domain and the alternatively spliced type 111 connecting segment. /. Cell Biol. 106: 1989-1297.

Adhesion molecules in neural crest development Hynes, R.O. and Lander, A.D. (1992) Contact and adhesive specificities in the associations, migrations and targeting of cells and axons. Cell 68: 303-322. Innes, P.B. (1985) The ultrastructure of early cephalic neural crest cell migration in the mouse. Anat. Embryol. 172: 3338. Kalcheim, C. and Teillet, M.A. (1989) Consequences of somite manipulation on the pattern of dorsal root ganglion development. Development 106: 85-93. Keilhauer, R., Faissner, A. and Schachner, M. (1985) Differential inhibition of neuron-neuron, neuron-astrocyte and astrocyte-astrocyte adhesion by LI, L2 and N-CAM antibodies. Nature 316: 728-730. Keynes, R.J. and Stern, C. (1984) Segmentation in the vertebrate nervous system. Nature 310: 786-789. Kornblihtt, A.R., Umezawa, K., Vibe-Pedesen, K. and Baralle, R.E. (1985) Primary structure of human fibronectin: differential splicing may generate at least ten polypeptides from a single gene. EMBO J. 4: 1755-1759. Krotoski, D. and Bronner-Fraser, M. (1990) Distribution of integrins and their ligands in the trunk of Xenopus laevis during neural crest cell migration. J. Exp. Zool. 253: 193150. Krotoski, D.M., Domingo, C. and Bronner-Fraser M. (1986) Distribution of a putative cell surface receptor for fibronectin and laminin in the avian embryo. /. Cell Biol. 103: 1061-1071. Kruse, J., Mailhammer. R., Wernecke, H., Faissner, A., Sommer, I., Goridis, C. and Schachner. M. (1984) Neural cell adhesion molecules and myelin-associated glycoprotein share a common carbohydrate moiety recognized by monoclonal antibodies L2 and HNK-1. Nature 311: 153155. Kiinemund, V., Jungalwala, F.B., Fischer, G., Chou, D.K.H., Keilhauer, G. and Schachner, M. (1988) The L2/HNK-1 carbohydrate of neural cell adhesion molecules is involved in cell interactions. J. Cell Biol. 106: 213-203. Lacy, B.E. and Underbill, C.B. (1987) The hyaluronan receptor is associated with actin filaments. J. Cell Biol. 105: 1395-1400. Lallier, T. and Bronner-Fraser, M. (1988) A spatial and temporal analysis of dorsal root and sympathetic ganglion formation in the avian embryo. Dev. Biol. 127: 99-112. Lallier, T. and Bronner-Fraser, M. (1991) Avian neural crest cell attachment to laminin: involvement of divalent cation dependent and independent integrins. Development 113: 1069-1084. Lallier, T and Bronner-Fraser, M. (1992) Inhibition of neural crest cell attachment by integrin antisense oligonucleotides. Science 259: 692-695. Lallier, T., Leblanc, G., Artinger, K.B. and Bronner-Fraser, M. (1992) Cranial and trunk neural crest cells use different mechanisms for attachment to extracellular matrices. Development 116:531-541. Le Douarin, N.M. (1982) The Neural Crest. Cambridge University Press, Cambridge, UK.

D.F. New green and S.S. Tan Le Douarin, N.M. and Teillet, M.A. (1974) Experimental analysis of the migration and differentiation of neuroblasts of the autonomic nervous system and of neurectodermal mesenchymal derivatives, using a biological cell marking technique. Dev. Biol. 41: 169-184. Le Douarin, N.M., Teillet, M.A., Ziller, C. and Smith, J. (1978) Adrenergic differentiation of cells of the cholinergic ciliary and Remak ganglia in avian embryos after in vivo transplantation. Proc. Natl. Acad. Sci. USA 75: 2030-2034. Leivo, I., Vaheri, A., Timpl, R. and Wartiovaara, J. (1980) Appearance and distribution of collagens and laminin in the early mouse embryo. Dev. Biol. 76: 100-114. Lindner, J., Rathjen, F.J. and Schachner, M. (1983) LI mono and polyclonal antibodies modify cell migration in early postnatal mouse cerebellum. Nature 305: 427-430. Lofberg, J., Ahlfors, K. and Fallstrom, C. (1980) Neural crest cell migration in relation to extracellular matrix organization in the embryonic axolotl trunk. Dev. Biol. 75: 148-167. Lofberg, J., Nynas-McCoy, A., Olsson, C , Jonsson, L. and Perris, R. (1985) Stimulation of initial neural crest cell migration in the axolotl embryo by tissue grafts and extracellular matrix transplanted on microcarriers. Dev. Biol. 107: 442-459. Loring, J.F. and Erickson, C.A. (1987) Neural crest cell migratory pathways in the trunk of the chick embryo. Dev. Biol. 121:220-236. Luckenbill-Edds, L. and Carrington, J.L. (1988) Effect of hyaluronic acid on the emergence of neural crest cells from the neural tube of the quail, Coturnix coturnix japonica. Cell Tissue Res. 252: 573-579. Mackie, E.J., Tucker, R.P., Halfter, W., Chiquet-Ehrisman, R. and Epperlein, H.H. (1988) The distribution of tenascin coincides with pathways of neural crest cell migration. Development 102: 237-250. Manasek, F.J. (1975) The extracellular matrix: a dynamic component of the developing embryo. Curr. Top. Dev. Biol. 10: 35-102. Martins-Green, M. (1988) Origin of the dorsal surface of the neural tube by progressive delamination of the epidermis ectoderm and neuroepithelium: implications for neurulation and neural tube defects. Dev. Biol. 82: 267-286. Martins-Green, M. and Erickson, C.A. (1987) Basal lamina is not a barrier to neural crest cell emigration: documentation by TEM and by immunofluorescent and immunogold labelling. Development 101: 517-533. Matrisian, L.M. (1990) Metalloproteinases and their inhibitors in matrix remodeling. Trends Genet. 6: 121-125. Menoud, P.A., Debrot, S. and Schowing, J. (1989a) Localization of urokinase-type and tissue-type plasminogen activator mRNA during organogenesis in the mouse. Roux's Arch. Dev. Biol. 198: 219-226. Menoud, P.A., Debrot, S. and Schowing, J. (1989b) Mouse neural crest cells secrete both urokinase-type and tissuetype plasminogen activators in vitro. Development 106: 685-690. Morriss-Kay, G. and Tuckett, F. (1989) Immunohistochemical

63 localization of chondroitin sulphate proteoglycans and the effects of chondroitinase ABC in 9- to 11-day rat embryos. Development 106: 787-798. Morris-Wiman, J. and Brinkley, L.L. (1990) Changes in mesenchymal cell and hyaluronate distribution correlate with in vivo elevation of the mouse mesencephalic neural folds. Anat. Rec. 226: 383-395. Newgreen, D.F. (1982) Adhesion to extracellular materials by neural crest cells at the stage of initial migration. Cell Tissue Res. 227: 291-311. Newgreen, D. (1984) Spreading of explants of embryonic chick mesenchymes and epithelia on fibronectin and laminin. Cell Tissue Res. 236: 265-277. Newgreen, D.F. (1989) Physical influences on neural crest cell migration in avian embryos: contact guidance and spatial restriction. Dev. Biol. 131: 136-148. Newgreen, D.F. (1990) Control of the directional migration of mesenchyme cells and neurites. Semin. Dev. Biol. 1: 3 0 1 311. Newgreen, D.F. (1992) Establishment of the form of the peripheral nervous system. In: LA. Hendry and C.E. Hill (Eds.), Development, Regeneration and Plasticity of the Autonomic Nervous system. Harwood Academic, Switzerland, pp. 1-94. Newgreen, D.F. and Thiery, J.P. (1980) Fibronectin in early avian embryos: synthesis and distribution along the migration pathways of neural crest cells. Cell Tissue Res. 211: 269-291. Newgreen, D.F. and Gibbins, I.L. (1982) Factors controlling the time of onset of the migration of neural crest cells in the fowl embryo. Cell Tissue Res. 224: 145-160. Newgreen, D.F. and Gooday, D. (1985) Control of the onset of migration of neural crest cells in avian embryos. Role of Ca^'*"-dependent cell adhesions. Cell Tissue Res. 239: 3 2 9 336. Newgreen, D.F. and Minichiello, J. (1995) Control of epitheliomesenchymal transformation. I. Events in the onset of neural crest migration are separable and inducible by protein kinase inhibitors. Dev. Biol. 170: 91-101. Newgreen, D.F., Ritterman, M. and Peters, E.A. (1979) Morphology and behaviour of neural crest cells of chick embryos in vitro. Cell Tissue Res. 203: 115-140. Newgreen, D.F., Gibbins, I.L., Sauter, J., Wallenfels, B. and Wiitz, R. (1982) Ultrastructural and tissue culture studies on the role of fibronectin, collagen and glycosaminoglycans in the migration of neural crest cells in the fowl embryo. Cell Tissue Res. 221:521-549. Newgreen, D.F., Scheel, M. and Kastner, V. (1986) Morphogenesis of sclerotome and neural crest in avian embryos. In vivo and in vitro studies on the role of notochordal extracellular material. Cell Tissue Res. 224: 299-313. Newgreen, D.F., Powell, M.E. and Moser, B (1990) Spatiotemporal changes in HNK-I/L2 glycoconjugates on avian embryo somite and neural crest cells. Dev. Biol. 139: 100120. Nichols, D.H. (1986) Formation and distribution of neural

64 crest mesenchyme to the first pharyngeal arch region of the mouse embryo. Am. J. Anat. 176: 221-231. Nichols, D.H. (1987) Ultrastructure of neural crest formation in the midbrain/rostral hindbrain and preotic hindbrain regions of the mouse embryo. Am. J. Anat. 179: 143-154. Noden, D.M. (1975) An analysis of the migratory behavior of avian cephalic neural crest cells. Dev. Biol. 42: 106-130. Noden, D.M. (1978) The control of avian cephalic neural crest cytodifferentiation. II. Neural tissues. Dev. Biol. 67: 313329. Noden, D.M. ( 1983) The role of the neural crest in patterning of avian cranial skeletal, connective and muscle tissues. Dev. Biol. 96: 257-276. Noden, D.M. (1986) Patterning of avian craniofacial muscles. Dev. Biol 116:347-356. Ogle, R.C., Potts, A.J., Yacoe, M. and Little, D.D. (1989) Embryonic chicken fibroblast collagen binding proteins: distribution, role in substratum adhesion and relationship to integrins. J. Cell Sci. 94: 361-369. Perris, R. and Johansson, S. (1987) Amphibian neural crest cell migration on purified extracellular matrix components: a chondroitin sulfate proteoglycan inhibits locomotion on fibronectin substrates. 7. Cell Biol. 105: 2511-2521. Perris, R.. and Johansson, S. (1990) Inhibition of neural crest cell migration by aggregating chondroitin sulfate proteoglycans is mediated by their hyaluronan-binding domains. Dev. Biol. 137: 1-12. Perris, R., Pausson, M. and Bronner-Fraser, M. (1989) Molecular mechanisms of avian neural crest cell migration on fibronectin and laminin. Dev. Biol. 136: 222-239. Perris, R., Krotoski, D., Lallier, T., Domingo, C, Sorrell, J.M. and Bronner-Fraser, M. (1991) Spatial and temporal changes in the distribution of proteoglycans during avian neural crest development. Development 111: 583-599. Pintar, J.E. ( 1978) Distribution and synthesis of glycosaminoglycans during quail neural crest morphogenesis. Dev. Biol. 67: 444-^64. Pittman, R.N. (1990) Developmental roles of proteases and inhibitors. Semin. Dev. Biol. 1: 65-74. Poelmann, R.E., Gittenberger-de Groot, A. C, Mentinck, M.M.T., Delpech, B., Girard, N. and Christ, B. (1990) The extracellular matrix during neural crest formation and migration in rat embryos. Anat. Embryol. 182: 29-39. Pollanen, J., Saicsela, O., Salonen, E.M., Andreasen, P., Nielsen, L., Dano, K. and Vaheri, A. (1987) Distinct localization of urokinase-type plasminogen activator and its type I inhibitor under cultured human fibroblasts and sarcoma cells. 7. Cell Biol. 104: 1085-1096. Poole, T.J. and Thiery, J.P. (1986) Antibodies and synthetic peptides that block cell-fibronectin adhesion arrest neural crest migration in vivo. In: H. Slavkin (Ed.), Progress in Developmental Biology, Alan R. Liss, New York, pp. 235238. Pratt, R.M., Larsen, M.A. and Johnston, M.C. (1975) Migration of cranial neural crest cells in a cell-free hyaluronaterich matrix. Dev. Biol. 44: 298-305.

Adhesion molecules in neural crest development Rakic, P. (1985) Contact regulation of neuronal migration. In: G.M. Edelman and J.P. Thiery (Eds.), Cell in Contact: Adhesions and Junctions as Morphogenetic Determinants, Neurosciences Research Foundation, Wiley, New York, pp. 67-91. Rickmann, M., Fawcett, J. and Keynes, R.J. (1985) The migration of neural crest cells and the growth of motor axons through the rostral half of the chick somite. J. Embryol. Exp. Morphol. 90: 437-455. Riopelle, R.J., McGarry, R.C. and Roder, J.C. (1986) Adhesion properties of a neuronal epitope recognised by the monoclonal antibody HNK-1. Brain Res. 367: 20-25. Roberts, D.D. and Ginsburg, V. (1988) Sulphated glycolipids and cell adhesion. Arch. Biochem. Biophys. 267: 405-415. Rogers, S.L., Edson, K.J., Letourneau, P.C. and McLoon, S.C. (1986) Distribution of laminin in the developing peripheral nervous system of the chick. Dev. Biol. 113: 429-435. Rovasio, R.A., Delouvee, A., Yamada, K.M., Timpl, R. and Thiery, J.P. (1983) Neural crest cell migration: requirements for exogenous fibronectin and high density. J Cell Biol. 96: 462^73. Runyan, R.B., Maxwell, G.D. and Shur, B.D. (1986) Evidence for a novel enzymatic mechanism of neural crest cell migration on extracellular glycoconjugate matrices. J. Cell Biol. 102:432-441. Saga, Y., Yagi, T., Ikawa, Y., Sakaicura, T. and Aizawa, S. (1992) Mice develop normally without tenascin. Genes Dev. 6: 1821-1831. Schnieke, A., Harbers, R. and Jaenisch, R. (1983) Embryonic lethal mutation in mice induced by retrovirus insertion into the alpha 1 (I) collagen gene. Nature 304: 315-320. Schoenwolf, G.C. and Fisher, M. (1983) Analysis of the effects of Streptomyces hyaluronidase on formation of the neural tube. J. Embryol. Exp. Morphol. 73: 1-15. Schoenwolf, G.C, Folsom, D. and Moe, A. (1988) A reexamination of the role of microfilaments in neurulation in the chick embryo. Anat. Rec. 220: 87-102. Serbedzija, G.N., Eraser, S.E. and Bronner-Fraser, M. (1990) Pathways of trunk neural crest cell migration in the mouse embryo as revealed by vital dye labelling. Development 108: 605-612. Simon-Assmann, R., Bouziges, F., Arnold, C, Haffen, K. and Kedinger, M. (1988) Epithelial-mesenchymal interactions in the production of basement membrane components in the gut. Development 102: 339-347. Simon-Assmann, P., Bouziges, F., Freund, J.N., PerrinSchmitt, F. and Kedinger, M. (1990) Type TV collagen mRNA accumulates in the mesenchymal compartment at early stages of murine developing intestine. J. Cell Biol. 110:849-857. Singley, C.T. and Solursh, M. (1980) The use of tannic acid for ultrastructural visualization of hyaluronic acid. Histochemistry 65: 93-102. Smith-Thomas, L.C. and Fawcett, J.W. (1989) Expression of Schwann cell markers by mammalian neural crest cells in vitro. Development 105: 251-262.

D.F. Newgreen and S.S. Tan Springer, T.A. (1990) Adhesion receptors of the immune system. Nature 346: 425-434. Stern, C. and Keynes, R.J. (1987) Interactions between somite cells-the formation and maintenance of segment boundaries in the chick embryo. Development 99: 261-272. Stern, C, Sisodiya, S.M. and Keynes, R.J. (1986) Interactions between neurites and somite cells: inhibition and stimulation of nerve growth in the chick embryo. /. Embryol. Exp. Morphol. 92: 209-226. Stern, CD., Norris, W.E., Bronner-Fraser, M., Carlson, G.J., Faissner, A., Keynes, R.J. and Schachner, M. (1989) Jl/tenascin-related molecules are not responsible for the segmented pattern of neural crest cells or motor axons in the chick embryo. Development 107: 309-319. Sternberg, J. and Kimber, S.I. (1986) Distribution of fibronectin. laminin and entactin in the environment of migrating neural crest cells in early mouse embryos. J. Embryol. exp. Morphol 91: 267-282. Tan, S.S. and Morris-Kay, G.M. (1985) The development and distribution of the cranial neural crest in the rat embryo. Cell Tissue Res. 240: 403-416. Tan, S.S. and Morris-Kay, G.M. (1986) Analysis of cranial neural crest cell migration and early fates in postimplantation rat chimeras. J. Embryol. exp. Morphol. 98: 21-58. Tan, S.S., Crossin, K.L., Hoffman, S. and Edelman, G.M. (1987) Asymmetric expression in somites of cytotactin and its proteoglycan ligand is correlated with neural crest cell distribution. Proc. Natl. Acad. Set. USA 84: 7977-7981. Tan, S.S., Prieto, A.L., Newgreen, D.F., Crossin, K.L. and Edelman, G.M. (1991) Cytotactin expression in somites after dorsal neural tube and neural crest ablation in chick embryos. Proc. Natl. Acad Set. USA 88: 6398-6402. Tennyson, V.M., Payette, R.F., Rothman, T.P. and Gershon, M.D. (1990) Distribution of hyaluronic acid and chondroitin sulfate proteoglycan in the presumptive aganglionic terminal bowel of Is/ls fetal mice: an ultrastructural analysis. /. Comp. Neurol. 291: 345-362. Thiery, J.P., Duband, J.L., Rutishauser, U. and Edelman, G. (1982) Cell adhesion molecules in early chick embryogenesis. Proc. Natl. Acad. Set. USA 79: 6737-6741. Tosney, K. (1978) The early migration of neural crest cells in the trunk region of the avian embryo: an electron microscopic study. Dev. Biol. 62: 317-333.

65 Tosney, K.W. (1982) The segregation and early migration of cranial neural crest cells in the avian embryo. Dev. Biol. 89: 13-24. Tosney, K.W. (1987) Proximal tissues and patterned neurite outgrowth at the lumbosacral level of the chick embryo: deletion of the dermamyotome. Dev. Biol. 122: 540-558. Tosney, K.W. (1988) Proximal tissues and patterned neurite outgrowth at the lumbosacral level of the chick embryo: partial and complete deletion of the somite. Dev. Biol. 127: 266-286. Tucker, R.P. (1986) The role of glycosaminoglycans in anuran pigment cell migration. J. Embryol. Exp. Morphol. 92: 145-164. Tucker, R.P. and Erickson, C.A. (1984) Morphology and behavior of quail neural crest cells in artificial threedimensional extracellular matrices. Dev. Biol. 104: 390405. Tucker, R.P., Edwards, B.F. and Erickson, C.A. (1985) Tension in the culture dish: microfilament organization and migratory behavior of quail neural crest cells. Cell Motil. 5: 225-237. Tucker, F. and Morris-Kay, G.M. (1988) Alcian blue staining of glycosaminoglycans in embryonic material: effect of different fixatives. Histochem. J. 20: 174-182. Turley, E.A. (1989) The role of a cell-associated hyaluronanbinding protein in fibroblast behaviour. In: The Biology of Hyaluronan, Ciba Foundation Symposium 143, Wiley, Chichester, pp. 121-137. Valinsky, J.E., Grossmann, G., Wong-Schneider, S. and Quigley, J.P. ( 1990) Involvement of plasminogen activator in the migration of avian neural crest cells. In: B.W. Festoff (Ed.), NATO Series: Regulation of Extravascular Fibrinolysis in Nervous System, Development and Disease, Plenum, New York. Weston, J.A. (1970) The migration and differentiation of neural crest cells. Adv. Morphogr. 8: 41-114. Weston, J.A. and Butler, S.L. (1966) Temporal factors affecting localization of neural crest cells in the chicken embryo. Dev. Biol. 14: 246-266. Yamagata, M., Susuki, S., Akiyama, S.K., Yamada, K.M. and Kimata, K. (1989) Regulation of cell-substrate adhesion by proteoglycans immobilized on extracellular substrates. J. Biol. Chem. 264: 8012-8018.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 3

Laminin in neural development Victor Nurcombe Department of Anatomy and Cell Biology, University of Melbourne, Parkville, Victoria 3052, Australia

1. Introduction The survival and development of neurons are influenced not only by soluble molecules such as neurotransmitters and trophic factors, but also by cell adhesion molecules anchored either on cell membranes or in the extracellular matrix. Laminin has received a great deal of attention in recent years by virtue of its powerful effects on the course of nerve fibre outgrowth. As such, it has been the subject of a number of fine reviews in preceding years (see Timpl et al., 1979; Martin and Timpl, 1987; Sanes, 1989; Mercurio, 1990; Reichardt and Tomaselli, 1991; Hynes and Lander, 1992; Yurchenco, 1994; Yurchenco and O'Rear, 1994). The aim of the current overview is to focus on some of the work that has occurred in just the last few years, particularly as it pertains to neuronal development. After the first discovery of laminin's stimulatory effects on neuronal cells (Baron-Van Evercooren et al., 1982), interest quickly focussed on the mechanisms by which this large, extracellular matrix (ECM) molecule elicits its effects. The first experiments trying to dissect these effects indicated that laminin had important synergistic influences on the way that neuronal cells were able to see crucial soluble factors in their micro-environment (Edgar and Thoenen, 1982), key results that are only now just starting to be followed up with any vigour. When mouse embryonal carcinoma cells are grown in culture their death can be prevented by the addition of fibroblast growth factors (Schubert and Kimura, 1991); however, they do not divide. If

placed on substrata of laminin, these growth factors become potent mitogens (see also Drago et al., 1991). Laminin also potentiates the differentiation of embryonal carcinoma into neurons, strongly indicating that specific interactions between an early precursor cell population and extracellular laminin are required during early neural differentiation (Sweeney et al., 1990). The data demonstrate that the substrata to which cells are exposed can regulate their mitogenic response to growth factors. Feedback loops are also known to exist between neurotrophic factors and neural cells. For example, ECM-nerve growth factor interaction upregulates only those receptors required for cells to bind to laminin, resulting in neurite outgrowth (Rossini et al., 1990). Other types of growth factor-induced changes in laminin receptor activity occur in the absence of changes in receptor number, and rely rather on receptor activation (Schubert, 1992). Due to the synergistic nature of growth factor-ECM interactions it is likely that each tissue has devised a unique set of rules for use during development. The great upsurge in interest in laminin stems largely from the observation that it is almost unique among defined ECM molecules for the rapidity with which it is able to stimulate the growth and regeneration of neuronal processes: neurite outgrowth can be initiated within minutes of exposure of neurons to tissue culture substrata containing laminin. Furthermore an analogous role in vivo seems to be likely; as discussed later, laminin immunoreactivity appears only transiently in the developing optic nerve, coinciding both with axonal growth in the optic tract and with the period in

68

which retinal ganglion cells are most sensitive to laminin in vitro (Cohen et al., 1986, 1987). Perhaps even more strikingly, laminin can act synergistically with appropriate target-derived neurotrophic factors to promote neuronal survival during early development both in vivo and in vitro (Edgar, 1989). Subsequently, it appears that laminin may play important roles in maintaining the transmitter phenotype of neurons: both a rapid activation and an enhanced expression of the enzyme tyrosine hydroxylase have been demonstrated in chromaffin cells after culture on laminin substrata (Acheson et al., 1986). The intracellular mechanisms responsible for these various activities of laminin remain incompletely understood, although the potency and specificity of its effects on neurons point to the existence of a specific transduction mechanism on neural cell surfaces. 1.1. The basal lamina The ECM is a complex and dynamic meshwork that is assembled outside the cell from the specialized glycoproteins and proteoglycans secreted by them. It influences many cell biological processes including cell adhesion, migration, tissue morphogenesis, cell proliferation and differentiation. Basal laminae (BLs) are specialized regions of ECM that separate cells, primarily but not exclusively of ectodermal origin from underlying connective tissue. The molecular components of these thin sheets interact non-covalently to form a dense network, thereby forming a barrier to cells within the nervous system and so compartmentalizing it. BLs throughout the vertebrate body contain a number of common components - notably collagen type IV, laminin, nidogen/entactin, SPARC/BM40 and a heparan sulphate proteoglycan (HSPG; Martin and Timpl, 1987; Yurchenco and O'Rear, 1994). These molecules are important for maintaining the structural integrity of BLs, and for mediating their attachment to other components of the ECM and to cell membranes. Each of these molecules are members of larger supergene families which have variations in their structures and tissue distributions. While initial molecular analyses of BL quite naturally focussed on their major, com-

Laminin in neural development

mon constituents, recent studies have increasingly highlighted less abundant, tissue-specific components. 1.2. Laminin in the basement membrane Basement membranes are formed from the glycoprotein and proteoglycan protomers mentioned above, which interact with each other to produce defined supramolecular assemblies (Yurchenco and Schittny, 1990). The assembly of basement membrane from its components appears, to a large extent, to be one of mass action-driven "self assembly". These interactions consist of protomers binding to themselves to produce homologous oligomers and polymers, and binding to each other to form heterologous complexes. The process is complex, involving reversible interactions as well as covalent stabilizations and cross-linking. Both weak and strong cross-links can be identified between laminin and collagen IV, and there may be coupling of such interactions such that strong ones drive protomers to high local concentration, allowing weaker interactions to operate and contribute to final structure: this would allow the reshuffling of bonds affecting structure to meet different functional needs (Siebold et al., 1988). Much of the work on basement membrane structure has derived from interactions studied in vitro using components isolated from the Engelbroth-Holm-Swarm (EHS) tumour: in some cases, it has been possible to verify such structural information about extracellular molecules through direct visualization of the molecular architecture in situ (Yurchenco and Schittny, 1990). It has now become clear that different components of a tissue-specific BL may derive from very different cell types abutting it; for example, BL at epithelial-mesenchymal interfaces are assembled from components variably derived from both cell types (Yurchenco, 1994). Laminin itself will aggregate in vitro into large polymers in a temperature-, time- and concentration-dependent manner. Aggregation exhibits both concentration and thermal reversibility and there is a critical concentration for polymerization of about 60 nM, reflecting cooperative nucleation-propagation type assembly (Yurchenco et al, 1985). Diva-

69

V. Nurcombe

lent cation, calcium in particular, is required for polymerization, and self-assembly can be separated into an initial temperature-dependent oligomer-forming step, followed by a calciumdependent polymer-forming step (Paulsson, 1988; Yurchenco and Cheng, 1993). 2. The laminin molecule 2.1. Nomenclature It has now been well established that laminin belongs to a supergene family of proteins with genetically distinct subunit chains that associate into many, generally trimeric, isoforms. Nine subunit chains have been described thus far, which are known to associate into 6 known hetertrimeric isoforms (Tryggvason, 1993). Because it is almost certain that more forms of laminin will be found, a new and more systematic nomenclature has arisen in which the previously designated A, Bl and B2 chains have been renamed a, ^ and x- The isoforms of each are now numbered chronologically in order of discovery, as are the complete laminins (Burgeson et al, 1994). 2.2. EHS laminin structure Prototypical, EHS tumour laminin is a large, flexible, three-chained glycoprotein (M, 900 kDa) consisting of 3 short arms (37 nm) and one long arm (77 nm) (Fig. 1). Each laminin molecule is made up of a a-^-x heterotrimer joined in a coiledcoil domain. Each short arm possesses a pair of globular domains and the long arm a larger globule at its end (Martin and Timpl, 1987). Tumour laminin then is composed of 3 distinct peptide chains, designated a l (440 kDa), ^l (225 kDa) and %1 (205 kDa); the cDNAs for these chains have been completely sequenced (Sasaki and Yamada, 1987; Sasaki et al., 1987, 1988) and the predicted domain structure is in good agreement with the electron microscopic rotary shadowing and physical data (Yurchenco and Schittny, 1990). The carboxyl-terminal moieties of all 3 chains join to form the long arm through triple-coiled alphahelices, and the chains are held together by disul-

fide bridges at the vertex and near the long arm globule. Ionic interactions determine the specificity of chain assembly into the coiled-coil structure; correct assembly is absolutely necessary to create the conformation required for cell recognition and heparan sulfate binding (Sung et al., 1993). Laminin has a fairly high carbohydrate content (12-15%), with nine forms of N-linked oligosaccharides, mostly of the complex variety (Fujiwara et al, 1988). The function of this carbohydrate is not well understood although suppression of glycosylation, while decreasing secretion, has not been found to adversely affect disulfide bonding between subunits or to alter heparin-binding (Howe, 1984). Two distinct regions of laminin bind cells with high affinity and are proposed as being important for cell regulation: short arm reA or M Chain

Bl or 5 Chain

B2 Chain

Fig. 1. Model of the prototypical EHS laminin-1 isoform with the location of active sites. Large proteolytic fragments (PI, E8) lie within the boxes and locations of bioactive peptides are indicated by arrows. The PI region is thought to bind the a\pi integrin during cell adhesion and the E8 region is thought to contain binding sites for both the a3>pi and «6^1 integrins during adhesion and neurite outgrowth.

70

gions near the intersection of the cross (fragment PI) and a site near the end of the long arm (fragment E8). 2.3. Laminin isoforms Several isoforms of laminin subunits have been described, after having been identified using monoclonal antibodies and subsequently shown by sequence homology to be related to the prototypical EHS tumour laminin (laminin-1). One isoform, called merosin (laminin-2), differs from EHS laminin by the replacement of the a\ (A) chain with a distinctly different, but related a2 (M) chain (Engvall et al., 1990; Vuolteenaho et al., 1994). The other isoform, called S-laminin (laminin-3), has a related chain in place of the ^31 (Bl) chain in laminin-1 (Sanes et al., 1990). Through these discoveries a consensual laminin family has emerged. It consists of trimers containing an a 1-3 chain, a ^1-3 chain, and a%l-2 chain. Unlike fibronectin, the different chains arise from different genes and not by alternative splicing. The diversity of laminins promises to be biologically significant, as BL on different cells in different regions express unique laminins. The laminin surrounding epithelial derivatives, laminin-5 (kalinin/nicein) consists of a3^3x2, whereas merosin (laminin-2) found associated with muscle consists of a2)81%l. The neuromuscular junction is further specialized in that although the a l and %1 chains are present in laminin-3 (s-laminin), the)32 chain replaces the^Sl (Sanes et al., 1990). Interestingly, although isolated recombinant laminin-3 is adhesive for motor neurons, it inhibits neurite outgrowth on EHS laminin. Four-armed laminins of strikingly similar a and P chain structure have been found in basement membranes of many species, including the insects. On the other hand, it is becoming increasingly apparent that there are variant forms of laminin within a given species, that are expressed at different stages of development and in different tissue locations. Schwannoma cell laminin, for example, possesses only three arms in a Y shape and lacks an identifiable a chain (Edgar et al., 1988); however, the presence of a large globular domain at the

Laminin in neural development

end of the long arm argues for the presence of a third truncated chain with a similar carboxylterminal region. At the present time the binding functions of these variant forms of laminin are not well understood, although they are likely to provide differential cell signalling or alter molecular architecture. Engvall and her colleagues originally purified laminin-2 from human placenta both in intact form and as pepsin fragments (Ehrig et al., 1990). They demonstrated that the a2-chain molecule also bound heparin, and that fragments from the long arm also promoted neurite outgrowth from neuronal cells, just as laminin does. Cells with various integrin-like receptors for laminin attached equally well to laminin-2 and laminin-1, suggesting that several of the known laminin-binding receptors also bind to laminin-2. Antibodies to the fi-l subunit of the integrin receptor family (see below) inhibited the neurite outgrowth stimulated by laminin-2, confirming the involvement of integrinmediated interaction of cells with both laminin-1 and -2. Schwannoma cells, previously shown to secrete a laminin-like neurite-promoting factor, were shown to synthesize laminin-2 in vivo and in vitro (Engvall et al., 1992); these results show that laminin-2, which is the most abundant basement membrane protein known in the laminin family, has properties very like laminin-1 despite differences in the structure of the heavy chain. Furthermore, laminin-2 may be identical to or a component of the neurite-promoting factors variously reported from heart, muscle, and Schwann cells. Northern analysis of fetal tissue revealed significant expression of a2 chains in heart, meninges, choroid plexus and pancreas; in contrast, a\ expression was most prominent in kidney, cerebellum, olfactory bulb and the neuroretina (Vuoteenaho et al., 1994). Reduction of laminin-2 levels has been noted in forms of congenital muscular dystrophy (Hayashi et al., 1993). This suggests that a2 chains have a role in maintaining skeletal muscle tissue. 3. Laminin binding by neural cells The importance of laminin in development is

V. Nurcombe

highlighted by the fact it is the first detectable component of the ECM, and maintains its pivotal role during the processes of subsequent cell proliferation, differentiation and tissue organisation. The identification of true, physiological laminin receptors is of obvious interest as they provide the mechanism by which the laminins exert their influence; however, the search for definitive receptors has been problematical. Several putative laminin receptors have been suggested over the past few years as mediating some or all of the various developmental processes. The study of any binding event that involves the central mechanism of adhesion is that it is notoriously difficult, either in vivo or especially in vitro, owing both to the real and constant danger of non-specific interaction as well as the possibility of receptor redundancy. The other compounding problem is the multi-domain structure of the laminin molecule itself; several of the distinct functional domains on the molecule may indeed have their own distinct surface receptors; indeed, each of the different laminin chains may have individual receptors (Nurcombe et al, 1989). It is also a distinct possibility that the authentic neuronal cell surface receptor binds to a functional site on the laminin molecule that is formed by the folding of some or all three chains (Deutzmann et al., 1990). It also appears that laminin has a number of "cryptic" sites that are only activated from the native trimeric structure after protease degradation (Aumailley et al., 1987). It is still not entirely clear which of the cell surface receptors identified so far actually represent mechanisms of import for the in vivo nervous system. Given that the structural domains of laminin that interact with cellular receptors have polypeptide sequences apparently not shared by other ECM molecules, then it is reasonable to expect that the cellular receptors for laminin might also be unique. Nevertheless, the results of antibody inhibition experiments may be taken together as an indication that at least one site located on the long arm of laminin interacts with integrins or integrin-associated molecules on the membranes of neurons and other cells (Mercurio, 1990).

71

3.1, Identification of putative neural cell binding sites Early studies to analyse the molecular basis of laminin-neuron interactions used antibodies, directed against either laminin or components of the neuronal membrane, in attempts to block laminin-induced neurite outgrowth. It was thus determined that the site on the laminin molecule responsible for its interactions with neural cells is likely to be at or near the end of the long arm in the E8 fragment, since domain-specific inhibitory antibodies all recognize epitopes located in this general area (Edgar et al., 1984; Sung et al., 1993). This result has been confirmed with domainspecific activation of the neuronal migration of olfactory neuroepithelial cells (Calof et al., 1994). However, such blocking antibodies might not necessarily recognize the neurite-promoting site of laminin per se, but might rely on the presence of epitopes either spatially or functionally associated with the site, with blockade occurring indirectly either by steric hindrance or allosteric inhibition. It is clear the site is RGD-independent. Indeed, this is the picture that is increasingly emerging from the study of laminin variants, many of which can still stimulate neurite outgrowth even though, due to differing subunit composition, they lack the epitopes recognised by blocking antibodies. Similar experiments with a variety of non-neuronal tissues, however, have confirmed that many cell types also interact with the same or closely associated sites on the long arm of laminin. It follows that, if both neural and non-neural cells interact with the same sites on laminin, then any receptor will not necessarily be found only on neural cells. Not unexpectedly, given its multiple activities, four distinct cell-binding sites have been identified within the native conformation of EHS laminin, with putative synthetic peptides derived from these sequences also showing biological activity. It should be noted that experiments with short synthetic peptides that attempt to model adhesive molecules have been strongly questioned (Yamada, 1991). The "F9 peptide", which corresponds to an amino acid sequence in the globular

72

domain within the cross arms of the^Sl chain, exhibits both heparin- and cell-binding characteristics (Sephel et al., 1989). An adjacent domain with homology to epidermal growth factor has a rod-like structure with many disulfide bonds and contains the amino acid sequence YIGSR, which, when present in synthetic peptides, apparently shows cell- and receptor-binding activity. Cyclic forms of this peptide have been reported to show higher activity than linear peptides, reflecting the many disulfide bonds that occur in this domain of the native molecule (Graf et al., 1987). Further evidence is provided by type-1 astrocytes, whose migration towards laminin is apparently inhibited by YIGSR (Armstrong et al., 1990). However, even cyclized YIGSR peptides are much less potent than whole laminin on a molar basis. It is therefore possible that two adjacent binding sites may act cooperatively in the native molecule (Goodman et al., 1991). A third putative cell attachment site within the cross arms of laminin is located in the short arm of the a\ chain and contains an RGD fibronectin-like binding sequence (Sasaki et al., 1988). Preliminary studies have shown that synthetic peptides to amino acid sequences derived from this region are active in cell attachment (Goodman et al., 1991). Further work appears to demonstrate that at least two putative receptors mediate attachment to the short arms of laminin. One, related to the al^Sl integrin member of the integrin superfamily, recognises RGDS-independent sites in PI fragments, and one is an RGD-dependent molecule recognising sites in PI and is not a^Sl integrin (Goodman et al., 1991). It should be stressed, however, that cell attachment may depend strongly on the conformation of a multichain structure (Deutzmann et al., 1990). Single chains in randomly coiled conformation have been found to be inactive in vitro, suggesting that synthetic peptides modelled on sequences in this region cannot exhibit comparable activities. By the use of well-defined fragments comprising the cell binding site at the end of the long arm of laminin, it has been shown that the presence of all constituent chains in their native conformations is required for activity (Deutzmann

Laminin in neural development

et al., 1990). This work demonstrates the limitations of assigning cellular functions to short, linear peptide sequences which lack correct conformational structure (see Engel, 1991 for review). It is also important to note that no conserved RGD sequences have been identified in laminin chain isoforms. An IKVAV sequence in the laminin a\ chain has also been advanced as a cell-binding and neurite-promoting site (Tashiro et al., 1989). The corresponding sequence in laminin-2 is IKVSV, which may be sufficiently conserved to be an active site. The similar activities of laminin-1 and -2 also suggested that the cell-binding sites may be located in one of the common subunits, the ^ chains, rather than in the different subunits. Since the x2-containing and ySl-containing laminin-2 preparations are equally active in vitro, it might be hypothesized that the active sites are in the% chain. A neurite-promoting site in the C-terminal end of the;^ chain has been proposed (Liesi et al., 1989), but much further work is needed to confirm the activity of this site. The laminin-3 homologue concentrated in synaptic clefts of the neuromuscular junction (Hunter et al., 1989) has been reported to contain a crucial tripeptide, LRE that mediates binding to neuronal cells (Hunter et al., 1992). Inhibition studies with a series of 20 tripeptide LRE analogues demonstrated that cells exhibit a high degree of selectivity for LRE, and suggested that ligand-binding requires a combination of electrostatic and hydrophobic interactions; the adhesion did not appear to require the presence of calcium. The authors postulated that LRE constitutes a motoneuron-selective adhesion site accessible in native basal laminae, which can also act to inhibit neurite outgrowth, thereby stabilizing the synaptic contact. Wujek and colleagues (1990) have provided evidence that cortical astrocytes secrete lamininlike molecules into the ECM, but that they only seem to synthesize the %1 chain of laminin; they suggest that the % chain may suffice to stimulate neurite outgrowth. A more likely explanation is that these cells secrete other chain homologues of laminin not probed for in this study.

V. Nurcombe

3.2. Ligand-binding studies Many of the ambiguities inherent in indirect attempts to characterize laminin-cell interactions by antibody inhibition have been circumvented by assaying the ability of laminin to interact directly with its responsive cells. Dissection of the laminin molecule by proteolytic fragmentation has confirmed that a site near the end of the long arm mimics the effects of the whole molecule on neural cells (Edgar et al., 1984). The use of such fragments has also demonstrated that the short arms of laminin contain at least two other cell binding sites, but that these short arm sites are latent within the native molecule (Aumailley et al, 1987; Nurcombe et al, 1989; Goodman et al, 1991). That is, their activities can only be demonstrated after release from intact laminin by proteolysis. However, changes in cell behaviour (such as neurite outgrowth) cannot be expected to simply reflect the initiating molecular interactions between laminin and its cellular receptor; changes in cell behaviour are the end points of complex, incompletely understood biological processes, any one of which may be modified by events unrelated to the receptor-laminin interaction (interaction with trophic factors, for example). Maximally, some 10^ high-affinity binding sites specific for laminin (K^ = 10~^ M) have been demonstrated on the membranes of responsive neuronal cells by ligand-binding assays (Aumailley et al, 1987; Nurcombe et al, 1989); binding of the whole molecule can be competitively blocked by long arm, but not short arm fragments. Although ligand-binding studies in themselves do not prove that the high-affinity binding sites for laminin molecules in solution are the actual physiological receptors mediating responses to the ECM, the presence of binding sites correlates extremely well with the ability of cells to attach and spread on laminin substrata (Aumailley et al, 1987; Goodman et al, 1989). In particular, it has been shown that the number of high-affinity laminin receptors on embryonic retinal ganglion cells decreases during maturation, this loss being reflected by the decline in laminin-stimulated neurite outgrowth from these neurons (Cohen et al, 1986, 1989).

73

This sort of information has been useful in estimating how many receptors can be expected from a given cellular source, their relative affinity for laminin, and details of the specificity to be expected in their molecular interactions with laminin. In addition to the difficulties arising from the use of inhibitory antibodies to identify putative receptor molecules, the demonstration that an isolated protein can bind laminin obviously does not guarantee that the protein acts as a physiologically relevant laminin receptor. One of the major problems with the affinity chromatography techniques used to analyse integrin receptors has been the lack of precise knowledge of the peptide sequences of laminin involved in the binding of the native molecule to its neuronal receptors, knowledge that was instrumental in identifying the integrin receptors for fibronectin (Hynes, 1987). Indeed, such a strategy may well turn out to be impossible because evidence has been advanced (Deutzmann et al, 1990) that the binding site for neurons within the long arm E8 fragment of laminin is not formed by a simple stretch of amino acid sequence, as in fibronectin, but requires the native conformation provided by the maintenance of secondary and tertiary structure in its 3-chain rod and globular domains. It has, therefore, not been possible to convincingly and specifically elute laminin affinity columns with short peptides. Consequently, the only methods available for elution of putative neuronal laminin receptors have relied upon changes in ionic conditions (Douville et al, 1988; Ignatius and Reichardt, 1988), conditions where specificity may be at a premium. The binding of integrins to ECM molecules is known to be dependent on divalent cations; however, it remains to be shown that the cation requirements of neurite outgrowth are due to a cation-sensitive interaction of laminin with the integrins present in the membranes of these neuronal cells. The use of laminin fragments containing the neurite-promoting site, possibly together with synthetic oligopeptides derived from the sequences involved, may improve the effectiveness of affinity chromatography by both reducing the chances of non-specific interactions and increasing the specificity of elution. However, as stressed by Edgar

74

(1989), it will only be certain that a receptor has been identified when it can be shown that a cell transfected with the cloned receptor for the candidate results in the expression of functional receptors and in the appearance of cellular responses to laminin. 3.3. The integrin receptor family The integrin receptor superfamily is to date the best characterized of the ECM receptors, and several excellent reviews have recently appeared detailing their structure and function (Hynes, 1987; Buck and Albelda, 1990; Mosher, 1991; Ruoslahti, 1991; Haas and Plow, 1994). The name of this receptor family reflects the integral membrane nature of the transmembrane receptor and its ability to link macromolecular assemblies in the extracellular environment with actin filament arrays inside the cell. The functional receptor is a noncovalently linked heterodimer of two subunits, a and )3, both of which have multiple types, so that a particular receptor for a particular ECM protein can be generated by combining a particular a chain with a particular p chain. The integrins are divided into family groups, based upon the ^ chain component. Some subunits are subject to alternative splicing, some integrins interact with more than one adhesive ligand and some integrins interact with different adhesive ligands when expressed by different cell types (Mosher, 1991). Therefore, a cell's complement of integrins constitutes an extremely flexible and sensitive system for detecting extracellular molecules in the immediate environment of the cell. What is less well understood is how the integrins influence changes in neuronal (and particularly growth cone) movement, shape, growth and eventual synaptic differentiation. Using antibodies directed against cell membrane components, it has been shown that anti-integrin antibodies block neurite outgrowth on laminin (Bozyczko and Horwitz, 1986; Cohen et al., 1987). Integrin complexes are involved in many cell-cell and cell-matrix interactions, as a consequence of which the anti-integrin antibodies reported to block neurite outgrowth on laminin are not highly specific, in the sense that they also block the at-

Laminin in neural development

tachment and spreading of many cell types on a variety of ECM molecules. The reason for this is that these antibodies recognize epitopes present on the)3-l subunit common to many integrins. The last few years have seen decisive progress towards an understanding of how the integrins bind their ligands. Recent NMR studies have determined the conformation of the tripeptide ROD sequence (Johnson et al., 1993), a consensus binding sequence for many integrin ligands, and shown that its secondary loop structure controls its biological activity. It is to be expected that such considerations will govern the binding to other ECM proteins. Current thought holds that when several integrins bind to their target amino acids, they cluster. The cytoplasmic tails of either subunit may be phosphorylated, causing changes in the activities and locations of a common set of receptor kinases (Kornberg et al., 1991). Ligated, clustered integrins can become tethered by cytoskeletal structures into focal adhesions, thus becoming foci for the tension-generating structures within neural cells. The largest integrin family is composed of the P'\ family. Within this family there are seven different a chains (a 1-6, a-v) that appear to combine with the P-\ chain. Four of these )8-l combinations are able to mediate binding to laminin in vitro; a-1, 2, 3, 6 with ^8-1 (Mosher, 1991). Many integrins, including the laminin-binding integrins a-1)8-1, a-iP'l and a-ZpA, may not require highly conserved binding sites on the laminin molecule, as evidenced by their binding to two or more of such unrelated proteins as laminin, collagen and fibronectin (Ruoslahti, 1991; Haas and Plow, 1994). Interestingly, none of the integrins binds laminin in an RGD-dependent manner although, within the PI fragment (Fig. 1), there is a cryptic cell adhesion site that can apparently be inhibited by the fibronectin-binding tripeptide RGD (Goodman et al., 1991). It has not been reliably determined with what affinities the various integrin receptors bind to laminin, making informed guesses as to which of the various ligand-receptor interactions are the most physiologically relevant very difficult.

75

V. Nurcombe

The)3-1 integrins were directly implicated in the binding of neural cells to laminin when such interactions were found to be strongly inhibited by anti^-1 antibodies (Bozyczko and Horwitz, 1986; Cohen et al., 1987; Tomaselli et al., 1987). Using specific ( chain antibodies, it has been possible to further map the binding of a-1)3-1 heterodimer receptor binding on neural cells to laminin fragment El (Hall et al., 1990), and both a-G^-l and a-3a-l to E8 (Hall et al., 1990; Sonnenberg et al., 1990). The integrin a-1^8-1 heterodimer can function either as a collagen receptor or as a laminin receptor in the PC 12 neural cell line (Tomaselli et al, 1988; Ignatius et al., 1990). However, the use of function-blocking antibodies has indicated that, in primary sympathetic neurons in vitro, this integrin works as a collagen IV receptor by recognising the NCI domain of collagen type IV (Lein et al., 1991). It is thus unclear whether this receptor can also function as a laminin receptor in these primary neurons, as opposed to cell lines. Although the a-6)8-4 and a-G^A integrins share a common alpha subunit chain, the recognition site on laminin for the )8-4 integrin has not been mapped; it appears that cells that express it do not bind to the E8 fragment (Sonnenberg et al., 1990), suggesting that its means of interaction is quite different. The specificity of the a-l^-X integrin seems to be determined by the cell type that expresses the molecule. On platelets, this receptor appears to mediate binding to collagen whereas PC 12 cells, like endothelial cells and melanoma cells, appear to use it to bind to laminin (Tomaselli et al., 1987). The reason for such receptor behaviour is unknown, but it does suggest that there is celldependent regulation of binding, which might be an important determinant for the adhesion properties of cells; different cells within a subpopulation may express different receptors which transduce fundamentally different signals. The a-3)3-l integrin also appears to bind fragments of laminin that have neurite-promoting activity in vitro (Gehlsen et al, 1989; Tomaselli et al., 1990). Most intriguingly, this receptor binding is inhibited by an antibody capable of blocking the neurite outgrowth-activity of laminin, an antibody

known to bind near the junction of the long arm and its terminal domain (Horwitz, 1991). The a-6)8-l receptor has been perhaps the most convincingly implicated in transducing the laminin signal for developing neuronal cells (de Curtis et al., 1991; see also Aumailley et al., 1990); it has also been shown to be the major laminin receptor of mouse embryonic stem cells and chick ciliary ganglion neurons (Cooper et al., 1991; Weaver et al., 1995). Further evidence (Deutzman et al, 1990; Skubitz et al., 1991) seems to demonstrate that, although cell adhesion and spreading of neurons can be blocked by antibodies to the a-6 integrin subunit, neurite outgrowth is unaffected, indicating distinct receptors for these two activities. The active sites on the integrin dimers are still being determined. The so-called "I" domains, sequences of approximately 190 amino acids found in a chains, have been implicated in both the ligand- and cationic-binding functions of these receptors (Michishita et al., 1993; Bilsland et al., 1994). The ^-1 subfamily also appears to bind multiple ligands. A region lying between residues 207 and 218 has been identified as crucial for the binding of both activating and inhibiting antibodies (Takada and Puzon, 1993; Shih et al, 1993). It lies between two putative ligand-binding regions. It has not yet been determined whether it works independently, or acts as a point where the a chain generates its contribution to ligand specificity. The structural stability, ligand specificity and affinity of integrins are also known to be dependent on divalent cations, especially calcium and manganese (Masumoto and Hemler, 1993). 3.4. Disintegrins A very interesting new class of proteins, the disintegrins, have been isolated as short ligands for the integrins (Weskamp and Blobel, 1994). The transmembrane forms of the family are characterised by metalloproteinase, disintegrin, cysteinerich and EOF domains. These molecules are able to couple and thus inactivate normal integrins. Current ideas posit that these molecules are able to modulate cell-cell and cell-ECM interactions, both at the cell membrane and into the BL. Their ex-

76

Laminin in neural development

pression and role during neural development has yet to be elucidated. 3.5. The high-affinity 67 kDa laminin-binding protein The first putative laminin cell surface receptors were purified from membrane preparations of various tumour cells by laminin-affinity chromatography. A protein of 67-70 kDa was identified (Rao et al., 1983), which had the astonishing property of retaining a high binding affinity for laminin after its isolation {K^ approximately 2 nM), even in the presence of detergent. The protein has been subsequently identified in a range of tissues, including neuronal cells (Douville et al., 1988; Kleinman et al, 1991). The true nature of this molecule has been hard to identify unequivocally, however. Recent experiments with antibodies to the "68 kDa laminin receptor" have shown that it may be used as a positional marker for dorsal embryonic retina, in a variety of vertebrates (Rabacchi et al., 1990). Similar work by McCaffery et al. (1990) has shown that the dorso-ventral asymmetry of this protein has 4 peculiarities: in immunoblots the molecular mass of the receptor is not 68 kDa, but 43 kDa; the molecular mass of the protein deduced from cDNA is only 33 kDa (see also Auth and Brawerman, 1992); the antibodies stain a cytoplasmic antigen, not a cell surface protein and, despite the pronounced dorso-ventral difference seen after immunostaining, the 43 kDa protein identified by immunoblotting appears evenly distributed throughout the retina. The authors concluded from sequence homology that the molecule may constitute a translation-initiation factor that reflects asymmetries in some aspect of protein translation. Attempts to further characterize this receptor have been both controversial and difficult, although it is possible that it shares identity with a molecule identified as a joint elastin/laminin receptor (Martin and Timpl, 1987). 3.6. Other binding proteins Another series of putative neuronal laminin re-

ceptors is represented by the cell surface glycosyltransferases (Begovac and Shur, 1990; Thomas et al., 1990; Shur, 1993), molecules which are functionally similar to lectins. Some of these enzymes he on the cell surface and recognize specific carbohydrate side-chains on proteins and will, in the presence of the appropriate nucleotide-sugar, catalyze the addition of an additional monosaccharide. In the absence of such a complex, the enzyme remains bound to its substrate, acting much like a lectin. The cell-surface glycosyltransferases that have been identified in the nervous system include ^1,4-galactosyltransferase (GalTase) and A^-acetyl-galactosaminyltransferase. The former enzyme binds to terminal A^-acetylglucosamine residues, and may participate in the responses of neurons and neural crest cells to laminin by binding to its oligosaccharide side chains (Begovac and Shur, 1990). Cells can make two forms of GalTase: the longer form has a 13 amino acid extension on its cytoplasmic tail and can thus function as a cell adhesion molecule by binding to the oligosaccharide chains on laminin that terminate in A^-acetylglucosamine. The amino terminal extension may allow the transmembrane GalTase to communicate with the actin cytoskeleton (Eckstein and Shur, 1992). Studies with a neuronal cell line suggest that cell-surface GalTase interacts with the E8 domain of laminin, the region recognized by several integrins, and is involved in laminin's ability to stimulate the initiation of neurites (Begovac et al., 1991). Interestingly, this enzyme is specifically localized to neuromuscular junctions in vivo (Scott et al, 1990). Thomas et al. (1990) have confirmed the thrust of these results with evidence that suggests that differential expression of GalTase at the growth cone might contribute to axonal guidance through glycoconjugate-rich environments. These data, together with the observation that the GalTase effects on neurite outgrowth are independent of and subsequent to initial cell attachment (Riopelle and Dow, 1991), indicate the necessity for further studies to dissect the process of neuronal laminin adhesion and establish the part played by each of the different laminin receptor processes involved. Recent work has also shown biological roles for the carbohydrate moieties of

11

V. Nurcombe

laminin (Dean et al., 1990; Chandrasekaran et al., 1991). These workers demonstrated that carbohydrate-binding lectins could block neurite outgrowth, but not cell spreading, on laminin substrates. As well, unglycosylated laminin produced from tunicamycin-treated cultures of an embryonal carcinoma cell line, although conformationally indistinguishable from glycosylated laminin, was unable to stimulate neurite outgrowth from PC 12 cells. They concluded that, after cells have bound to laminin, carbohydrate residues must be available to enable the cells to spread and extend neuritic processes. Other transmembrane receptor systems for laminin receiving recent prominence are the cell surface proteoglycans (Dow et al., 1991; Hynes and Lander, 1992). There is evidence that in fibronectin the integrin recognition site only permits complete cell spreading, including the formation of stress fibres and focal contacts, in conjunction with a heparin-binding site (Ruoslahti, 1988). Dow et al. (1991) have demonstrated that HSPGs can play a subsidiary role in promoting neurite growth on laminin, presumably by bridging between lamininbinding domains and the ECM. This idea concords well with the way the laminin molecule is thought to configure itself within the BL; the short arms of the molecule binding to collagen type IV and leaving the long arm free to contact the cell surface (Schittny et al., 1988). It is also interesting to note that an HSPG core molecule has been cloned that has extensive similarity to the a\ chain of laminin1 (Noonan et al., 1991). How might the integrins transduce extracellular signals? It has been found that the growth of neurites from ciliary ganglion neurons on laminin is enhanced by 12-0-tetradecanoylphorbol-13acetate and inhibited by protein kinase C inhibitors. This suggests that the neuronal response to laminin is mediated by activation of protein kinase C (Bixby, 1989). It is known that the organisation of the cytoskeleton in the developing neuron plays a fundamental role in the formation of neurites. ECM receptors may play an important role in the activation of the intracellular signals involved, because integrins colocalize with several cytoskeletal components at the tip of filopodia extend-

ing from the growth cone (Neugebauer et al., 1988; Letoumeau and Shattuck, 1989). Other putative laminin-binding proteins have been described from a number of sources (for example, Smalheiser and Schwartz, 1987). Kleinman et al. (1991) used a synthetic peptide derived from the a chain of laminin to elute a llOkDa protein from brain tissue; however, the resulting molecule had sequence similarity to nucleolin. This putative receptor appears to be expressed in postnatal brain, particularly the hippocampus (Luckinbilledds et al., 1995). Albini et al. (1992) have described the isolation of a similar 100 kDa molecule from Y-79 retinoblastoma cells that is eluted from laminin affinity columns by 20 mM EDTA; this binding protein has only intermediate affinity for laminin, however, but may act to influence gene expression independent of attachment. Laminin has been reported to bind to Ng-CAM on the basis of antibody inhibition studies, an interaction important for the formation of neuron-glia adhesion (Grumet etal., 1993). 4. Laminin cell biology 4.7. Laminin and neural precursor proliferation At early stages of neural cell proliferation, neuroepithelial cells are particularly sensitive to the fibroblast growth factors (FGFs) (Drago et al., 1991, 1992). One effect of FGF on cultured neuroepithelial cells is to enhance the amount of laminin expressed at the protein level of these cells; this upregulation correlates with a significant upregulation of steady-state levels of laminin Bl and B2 chain expression at the mRNA level. When the precursor population is split into precursor neuronal and glial subpopulations on the basis of differential expression of major histocompatibility class-1 antigens, only the glial precursor fraction is found to be capable of laminin synthesis. It is thus possible that a major action of FGF is to increase the synthesis and release of extracellular molecules from neural cells which act in a paracrine manner to further stimulate differentiation. This result is supported by the findings of Reh and Radke (1988), who found that continuous contact of the

78

neural progenitors in the retina with underlying ECM was necessary to maintain normal rates of precursor proliferation. 4.2. Laminin and neural cell migration Cell migration is a particularly striking feature of nervous system morphogenesis, and migration of neuronal precursors and immature neurons is responsible for establishing structural features such as the laminae of the cerebral cortex, and the ganglia that are a feature of the peripheral nervous system. Much of what is known about the molecules that mediate the migration of neural cells has come from studies of the neural crest (Hynes and Lander, 1992), the cell population that gives rise to most of the peripheral nervous system. Some recent studies looking at the relationship of laminin to neural crest migration will be discussed below. Experiments performed in vivo and in vitro over many years indicate that the ECM through which neural precursor cells move constitutes a rich source of molecular cues that stimulate and guide migration. Calof and Lander (1991) have recently shown that neuronal cells of the embryonic olfactory epithelium (OE), which undergo extensive migration within the central nervous system during normal development, are highly migratory in culture as well. This migration is strongly dependent on substratum-bound ECM molecules, being specifically stimulated and guided by both laminin-1 and -2 in preference to fibronectin. Interestingly, quantitative assays of adhesion of OE neuronal cells to substrata treated with different ECM molecules demonstrated no correlation, either positive or negative, between the migratory preferences of cells and the strength of cell-substratum adhesion (see also Lemmon et al., 1992). Indeed, measurements of cell adhesion to substrata containing combinations of ECM proteins revealed that laminin-1 and -2 are antiadhesive for OE neuronal cells, causing these cells to adhere poorly to substrata that would be otherwise be strongly adhesive. These effects, which resemble the effects of other anti-adhesive molecules such as tenascin and thrombospondin (de Curtis, 1991), are not due to interactions between

Laminin in neural development

the laminin and the other ECM molecules but, rather, to an effect of laminin on cells which altered the way the cells subsequently adhered. Consistent with this view, laminin-1 was found by Calof and Lander (1991) to interfere strongly with the formation of focal contacts by OE neuronal cells. Treatment of laminin with antibodies against its short arms (domains El or PI) changed the migration behaviour of the OE neurons from a preference for the E8 domain via a6p\ receptors to another p\ receptor type (Calof et al., 1994). Laminin also seems to be major adhesion system used by granule neurons in their migration within the developing cerebellum (Fischman and Hatten, 1993). 4.3. Laminin at the neuromuscular junction Approximately 0.1% of the BL that ensheathes muscle fibres lines the synaptic cleft separating nerve from muscle at the neuromuscular junction; the remaining 99.9% is extrasynaptic and separates muscle fibre membrane from endomysial connective tissue. When motor axons regenerate to denervated muscles after nerve injury, they preferentially reinnervate the original synaptic sites and subsequently differentiate into nerve terminals (Sanes, 1989). On BL sheaths from which muscle fibres have been eliminated, axons also prefefentially form contacts and differentiate at synaptic sites, implicating ECM components in the remarkable specificity of reinnervation. Similar experiments, in which myotubes were allowed to regenerate within BL sheaths in the absence of nerves, demonstrated that synaptic sites on BL sheaths can also direct postsynaptic differentiation (reviewed in Sanes, 1989). These experiments hinted at the existence of synapse-specific components in muscle fibre BL that axons and myotubes recognize and to which they respond. Immunohistochemical methods have demonstrated molecules which may serve these functional specializations. Whereas antibodies to several components of BL (collagen type IV, laminin, nidogen, HSPG) stain both synaptic and extrasynaptic BL (Eldridge et al., 1986), other antibodies stain synaptic BL selectively, and still others stain

V. Nurcombe

extracellular BL more intensely than synaptic BL (Sanes and Chui, 1983). Molecular cloning has recently revealed that one of these synaptic antigens, subsequently dubbed "S-laminin" (laminin-3), is a homologue of the^l chain of laminin-1 (Hunter et al., 1989). It is thus an intriguing possibility that different tissues may use a variety of laminin-like molecules with different bioactivities to regulate axonal behaviour in complex ways. These findings subsequently inspired histological studies which sought to examine the distribution of laminin-3, laminin-2 and the EHS laminin subunits (i.e. a-^-x chains) in both muscle and non-muscle BLs (Sanes et al., 1990; Green et al., 1992) and have led to revision of the view that all BLs contain an identical set of major structural components. Strong evidence for the important role of laminin-3 in synapse formation was provided by gene knockout experiments (Noakes et al., 1995). The selective ablation of the laminin %2 gene by targeted mutation greatly disrupted synapse formation at the neuromuscular junction. Clearly, such detailed molecular study is necessary if knowledge of developmental processes is to be exploited for therapeutic use in instances, say, requiring nerve regeneration. After nerve injury, axons regenerate through denervated nerve trunks and into denervated muscles, where they form new neuromuscular junctions. The axons frequently grow along BLs and evidence obtained both in vivo and in vitro indicates that laminin is an important promoter of their extension (Sanes, 1989; Shewan et al, 1995). Particularly noteworthy is the lack of effect of denervation on the abundance of the laminin a chain; this subunit is thought to bear a neurite outgrowthpromoting site (Tashiro et al., 1989) but is not detectable in endoneurial BL along which axons preferentially regenerate (Sanes, 1989). Such immunohistochemical studies have helped to extend our views about the heterogeneity of BLs in general and the molecular architecture of synaptic BL in particular. Although BLs have been subjected to biochemical study for many years, their analysis was long complicated by the extreme insolubility of adult BLs in aqueous solutions and by the extensive covalent cross-linking of their components.

79

5. Laminin in pathway formation It has become abundantly clear that axonal outgrowth in vitro can be directed by the spatial patterning of ECM components and their presentation to growth cone filopodia (Clark et al., 1993; Gomez and Letoumeau, 1994). These experiments have led inevitably to the hypothesis that analogous laminin-rich pathways may be operating in vivo to direct axons to their targets. 5.1. The retinotectal pathway Major insights as to some of the roles of laminin and how it may be acting have been obtained by examining a particularly well-studied pathway, the one leading axons from the retina of the vertebrate eye to the optic tectum. Retinal ganglionic axons exit the retina through the optic nerve head and then grow in contact with the laminin-containing BL (Cohen et al., 1987; Drazba and Lemmon, 1990), which may play a role in promoting, if not directing their growth. It is interesting to note that the ganglion cells themselves are synthesizers of laminin Bl chain (Sarthy and Fu, 1990). Laminin is unlikely to provide the information to guide these axons to either the optic nerve head or further, however, since it is known to be incapable of conferring directionality on axon growth in vitro, even in relatively steep concentration gradients (McKenna and Raper, 1988). After axons reach the optic nerve head, they must traverse an environment consisting of neuroepithelial cells and early glia. Laminin is expressed in abundance on glial endfeet just inside the glial limitans of the ECM (Cohen et al., 1987). The glial limitans is found at the interface between the meninges and the central nervous system and is formed wherever astrocytes and meningeal cells come into contact (Abnet et al., 1991). In vitro studies of retinal ganglion cells taken from chick embryos at these early stages reveal that laminin binding appears to be through an integrin (Cohen et al., 1987; de Curtis et al., 1991). The presence of laminin all along the optic pathway is particularly interesting because its appearance is transient, corresponding precisely to the time when retinal axons are trav-

80

ersing this route, and because laminin appears to be absent from most other brain regions. At the end of the optic tract lies the target region for the ganglion cells, the optic tectum. To enter the tectum, the axons must leave the laminin-rich pathways and arborize within regions which appear largely devoid of laminin expression. The ability of retinal axons to abandon a laminin-containing environment may result from a developmental change they undergo. In vitro, retinal ganglion neurons taken from stages at which axons have not yet reached their target in the tectum can be triggered by laminin to extend neurites; they are able to bind exogenous laminin with a large number of receptors (Cohen et al., 1989). In contrast, ganglion cells isolated from stages after the tectum has been reached appear unresponsive to laminin in vitro. This may be in part due to down-regulation by retinal ganglion cells of the integrin subunit a6 which, in combination withal, appears to form the major retinal laminin receptor (de Curtis et al., 1991). In the other retinal neurons, expression of a6 remains high even though these cells have also lost their ability to interact with laminin, suggesting that the activation state of the receptor in these cells is regulated by post-translational mechanisms. This hypothesis is corroborated by the finding that, when retinal neurons which have lost responsiveness to laminin are incubated with a monoclonal antibody against the p\ integrin subunit, they regain the ability to bind to laminin in culture, indicating that integrin function at the surface of the neuron is modulated (Neugebauer and Reichardt, 1991). The mechanisms that induce the transcriptional regulation of the a6 laminin receptor during the development of retinal ganglion cells in culture are still obscure. The maintenance of neuronal adhesion to laminin in culture after ablation of the tectum suggests that a targetmediated mechanism, possibly the availability of essential neurotrophic factors, may be involved (Cohen et al, 1989). Several growth factors are known which support neuronal survival and neurite extension; the effects of nerve growth factor in the regulation of ECM receptor expression has been reported in PC 12 cells, where it induces a specific increase in the expression of the integrin

Laminin in neural development

a 1^1, the purported laminin/collagen-binding receptor for this neuronal cell line (Rossini et al., 1990). 5.2. Laminin and the neural crest It has become quite clear that cell-matrix interactions mediated by integrins play a major role during neural crest migration (see Hynes and Lander, 1992, for review). Neural crest cells arise from the dorsal part of the neural tube and, at stages after they lose expression of cell-cell adhesion factors such as N-CAM and N-cadherin, they migrate into the spaces around the neural tube. These spaces are full of ECM, rich in both laminin and fibronectin. Cultured neural crest cells adhere and migrate on these ECM proteins, and can be blocked by their respective antibodies, or by antibodies to integrins (Bronner-Fraser, 1985, 1986). Antibodies directed against a laminin-proteoglycan complex can also block cranial neural crest migration in vivo (Lallier et al., 1990). This cannot explain directed migration, however, as laminin is also expressed in adjacent territories in the ECM where neural crest cells do not migrate. Schwann cell migration can be blocked with antibodies specific for laminin-2 (Anton et al., 1994). This leads to the idea that molecules such as laminin might always be found expressed in conjunction with molecules that are anti-adhesive (ChiquetEhrisman, 1991), and that it co-joint expression of at least two signals is needed to define a migratory or axonal pathway. Thus the latest picture to emerge is that cells that disperse may lose cell-cell adhesion receptors and then acquire cell-matrix receptors, including integrins, which then mediate cell migration in response to ECM cues. Reassociation of precursor ganglionic cell populations may then be mediated by re-expression of cell-cell adhesion receptors. This sequence of events requires temporal and spatial regulation of the expression and/or function of adhesion receptors during successive phases. Neural crest derivatives such as Schwann cells appear to use lamininintegrin associations for later stages of their differentiation (Obremski and Bunge, 1995). Using specific anti-j81 integrin antibodies Femandezvalle

V. Nurcombe

and colleagues (1994) showed that an integrin binds laminin to Schwann cell surfaces and transduces signals critical for the initiation of Schwann cell differentiation into myelinating cells. 5.3. Laminin in other pathways A series of interesting experiments was recently carried out by Yip and Yip (1992), who examined the possible role of laminin in axon guidance and outgrowth in vivo. They did this by examining the expression of laminin and its relationship to the outgrowth of sensory, motor and sympathetic axons in the chick embryo, and by evaluating the changes in the pattern of sympathetic preganglionic projections subsequent to injections of exogenous laminin, anti-laminin antibodies, or other laminin function-blocking antibodies. Microinjection of these blockers or enhancers into the pathways of axons were found, surprisingly, to have no effect on the spatio-temporal pattern of axonal projection; the results suggested that laminin may not be absolutely crucial to the initial outgrowth of peripheral axons, but may be more important in maintaining the structure of the peripheral nerve. In this context it is interesting to note that mutation of the gene for UNC-6, a laminin homologue found in C. elegans, severely disrupts axonal outgrowth in dorsal and ventral directions along the body wall (Ishii et al., 1992). Other experiments have demonstrated that the relative adhesive capacities of different substrates in vitro are poor predictors of either axon growth rate or the degree of fasciculation (McKenna and Raper, 1988; Lemmon et al, 1992; Rivas et al., 1992). This suggests that adhesion molecules such as laminin may serve primarily as permissive substrates, defining axonal pathways but not providing information about which path to take at a choice point, or about which direction to go along the path. These results are buttressed by the results of Buettner and Pittman (1991), who demonstrated that neurite outgrowth from superior cervical ganglionic neurons in culture is relatively insensitive to changes in laminin concentration, showing only a two-fold increase in neuritic branching for a 100-fold increase in laminin con-

81

centration. All of these results support the idea that laminin forms a permissive rather than an instructive substrate. Before laminin expression becomes localized to the BL in adult brain, it is found throughout the developing brain in several different forms (Zhou, 1990). Small and large punctate deposits are found, which disappear when the brain matures. Two other forms, sheath laminin and somal laminin, reduce in intensity but persist throughout adult life. Each has a unique spatial and temporal distribution. Small puncta appear first in development and disappear after fibre pathways have formed; sheath laminin is associated with the microvasculature, the ependyma, the choroid plexus and the surface of the brain. Somal laminin, the last appearing, is found associated mainly with the cell body of neurons and is correlated with the appearance of compacting brain nuclei. Ultrastructural localizations of laminin have also been carried out at the electron microscope level during development and regeneration of the mouse sciatic nerve (Kuecherer-Ehret et al., 1990), in order to ascertain the distribution of laminin during elaboration of the peripheral nervous system. Laminin distribution is not restricted to the BL secreted by the Schwann cells, but is also found in direct contact with developing axons, as well as on the surface of the Schwann cells; such a distribution reinforces the idea that laminin is involved in the outgrowth process in vivo. 6. Conclusions This short and selective review indicates that research into laminin structure and function has become increasingly sophisticated. The immediate future promises discovery of an increasing family of laminin isoforms, allowing the characterization of their biological activities, a more definitive map of the active domains and the peptides which subtend them, the working out of anti-laminin modulating influences, and characterisation of the receptors interacting with these domains in vivo to mediate cellular responses. Such study should clarify the contexts by which laminin exerts its powerful influence on neuronal development.

82

Acknowledgements The author would like to acknowledge his former colleagues David Edgar and Rupert Timpl for their support during his initiation into the laminin field. He is supported by the National Health and Medical Research Council of Australia, the Australian Research Council, the Multiple Sclerosis Society of Australia and the A.L.S.-Motor Neurone Disease Society of Australia. References Abnet, K., Fawcett, J.W. and Dunnett, S.B. (1991) Interactions between meningeal cells and astrocytes in vivo and in vitro. Dev. Brain Res. 59: 187-196. Acheson, A.L., Edgar, D., Timpl, R. and Thoenen, H. (1986) Laminin increases levels and activity of tyrosine hydroxylase in calf adrenal medullary cells. /. Cell Biol. 102: 151159. Albini, A., Noonan, D.M., Melchiori, A., Fassina, G.F., Percario, M., Gentleman, S., Toffenetti, J. and Chader, G.J. (1992) Laminin-induced retinoblastoma cell differentiation: possible involvement of a 100 kDa cell-surface lamininbinding protein. Proc. Natl. Acad. Sci. USA 89: 22572261. Anton, E.S., Sandrock, A.W. and Matthew, W.D. (1994) Merosin promotes neurite growth and Schwann cell migration in vitro and nerve regeneration in vivo - evidence using an antibody to merosin, ARM-1. Dev. Biol. 164: 133-146. Armstrong, R.C., Harvath, L. and Dubois-Dalcq, M.E. (1990) Type 1 astrocytes and oligodendrocyte-type 2 astrocyte glial progenitors migrate toward distinct molecules. J. Neurosci. Res. 27: 400-407. Aumailley, M., Nurcombe, V., Edgar, D., Paulsson, M. and Timpl, R. (1987) The cellular interactions of laminin fragments./. Biol. Chem. 262: 11532-11538. Aumailley, M., Timpl, R. and Sonnenberg, A. (1990) Antibody to integrin alpha-6 subunit specifically inhibits cell binding to laminin fragment 8. Exp. Cell Res. 188: 55-60. Auth, D. and Brawerman, G. (1992) A 33 kDa polypeptide with homology to the laminin receptor: component of translation machinery. Proc. Natl. Acad. Sci. USA 89: 4368-4372. Baron-Van Evercooren, A., Kleinman, H.K., Ohno, S., Marangos. P., Schwartz, J.P. and Dubois-Dalcq, M.E. (1982) Nerve growth factor, laminin and fibronectin promote neurite outgrowth in human fetal sensory ganglia cultures. J. Neurosci. Res. 8: 179-194. Begovac, P.C and Shur, B.D. (1990) Cell surface galactosyltransferase mediates the initiation of neurite outgrowth from PC12 cells on laminin. J. Cell Biol. 110: 461^70. Begovac, P.C, Hall, D.E. and Shur, B.D. (1991) Laminin

Laminin in neural development fragment E8 mediates PC 12 cell neurite outgrowth by binding to cell surface beta 1,4 galactosyltransferase. J. Cell Biol. 113:637-644. Bixby, J.L. (1989) Protein kinase C is involved in laminin stimulation of neurite outgrowth. Neuron 3: 287-297. Bozyczko, D. and Horwitz, A. (1986) The participation of a putative cell surface receptor for laminin and fibronectin in peripheral neurite extension. J. Neurosci. 6: 12411251. Bronner-Fraser, M. (1985) Alterations in neural crest migration by a monoclonal antibody that affects cell adhesion. J. Cell Biol. 101:610-617. Bronner-Fraser, M. (1986) An antibody to a receptor for fibronectin and laminin perturbs cranial neural crest migration in vivo. Dev. Biol. 117: 528-536. Buck, C.A. and Albelda, S.M. (1990) Integrins and other cell adhesive molecules. FASEB J. 4: 2868-2880. Buettner, H.M. and Pittman, R.N. (1991) Quantitative effects of laminin concentration on neurite outgrowth in vitro. Dev. Biol. 145: 266-276. Burgeson, R.E., Chiquet, M., Deutzmann, R., Ekblom, P., Engel, J., Kleinman, H., Martin, G.R., Meneguzzi, G., Paulsson, M., Sanes, J. and Timpl, R. (1994) A new nomenclature for laminins. Matrix Biol. 14: 209-211 Calof, A.L. and Lander, A.D. (1991) Relationship between neuronal migration and cell-substratum adhesion: laminin and merosin promote olfactory neuronal migration but are anti-adhesive. 7. Cell Biol. 115: 779-794. Calof, A.L., Campanero, M.R., O'Rear, J.J., Yurchenco, P.D. and Lander, A.D. (1994) Domain-specific activation of neuronal migration and neurite outgrowth-promoting activities of laminin. A^ewrow 13: 117-130. Chandrasekaran, S., Dean, J.W., Giniger, M.S. and Tanzer, M.L. (1991) Laminin carbohydrates are implicated in cell signalling. 7. Cell Biol. 46: 115-124. Chiquet-Ehrismann, R. (1991) Anti-adhesive molecules of the extracellular matrix. Curr. Opin. Cell Biol. 3: 800-804. Clark, P., Britland, S. and Connelly, P. (1993) Growth cone guidance and neuron morphology on micropatterned laminin surfaces. J. Cell Sci. 105: 203-212. Cohen, J., Burne, J.F., Winter, J. and Bartlett, P.F. (1986) Retinal ganglion cells lose response to laminin with maturation. Nature 322: 465-467. Cohen, J., Burne, J.F., McKinley, C. and Winter, J. (1987) The role of laminin and the laminin/fibronectin receptor complex in the outgrowth of retinal ganglion cell axons. Dev. Biol. 122:407-418. Cohen, J., Nurcombe, V., Jeffrey, P. and Edgar, D. (1989) Developmental loss of functional laminin receptors on retinal ganglion cells is regulated by their target tissue, the optic tectum. Development 107: 381-387. Cooper, H.M., Tamura, R.N. and Quaranta, V. (1991) The major laminin receptor of mouse embryonic stem cells is a novel isoform of the alpha-6 beta-1 integrin. J. Cell Biol. 115:843-850. Dean, J.W., Chandrasekaran, S. and Tanzer, M.L. (1990) A

V. Nurcombe biological role of the carbohydrate moieties of laminin. /. Biol. Chem. 265: 12553-12562. de Curtis, I. (1991) Neuronal interactions with the extracellular matrix. Curr. Opin. Cell Biol. 3: 824-831. de Curtis, I., Quaranta, V., Tamura, R.N. and Reichardt, L.F. (1991) Laminin receptors in the retina: sequence analysis of the chick integrin alpha-6 subunit. J. Cell Biol. 113: 405416. Deutzmann, R., Aumailley, M., Wiedemann, H., Pysny, W., Timpl, R. and Edgar, D. (1990) Cell adhesion, spreading and neurite stimulation by laminin fragment E8 depends on maintenance of secondary and tertiary structure in its rod and globular domain. Eur. J. Biochem. 191: 513-522. Douville, P.J., Harvey, W.J. and Carbonetto, S. (1988) Isolation and partial characterization of high affinity laminin receptors in neural cells. J. Biol. Chem. 263: 14964-14969. Dow, K.E., Riopelle, R.J. and Kisilevsky, R. (1991) Domains of neuronal heparan sulfate proteoglycans involved in neurite growth on laminin. Cell Tissue Res. 265: 345351. Drago, J., Nurcombe, V. and Bartlett, P.P. (1991) Laminin through its long arm E8 fragment promotes the proliferation and differentiation of murine neuroepithelial cells in vitro. Exp. Cell Res. 192: 256-265. Drago, J., Nurcombe, V., Pearse, M.J., Murphy, M. and Bartlett, P.P. (1992) Basic fibroblast growth factor upregulates steady-state levels of laminin Bl and B2 chain mRNA in cultured neuroepithelial cells. Exp. Cell Res. 196: 246254. Drazba, J. and Lemmon, V. (1990) The role of cell adhesion molecules in neurite outgrowth on Mueller cells. Dev. Biol. 138: 82-93. Eckstein, D.J. and Shur, B.D. (1992) Cell surface )3-l,4galactosyltransferase is associated with the detergentinsoluble cytoskeleton on migrating mesenchymal cells. Exp. Cell Res. 201: S3-90. Edgar, D. (1989) Neuronal laminin receptors. Trends Neurosci. 12:248-251. Edgar, D. and Thoenen, H. (1982) Modulation of NGFinduced survival of chick sympathetic neurons. Dev. Brain Res. 5: 89-92 Edgar, D., Timpl, R. and Thoenen, H. (1984) The heparinbinding domain of laminin is responsible for its effects on neurite outgrowth and neuronal survival. EMBO J. 3: 14631468. Edgar, D., Timpl, R. and Thoenen, H. (1988) Structural requirements for the stimulation of neurite outgrowth by two variants of laminin and their inhibition by antibodies. J. Cell Biol. 106: 1299-1306. Ehrig, K., Leivo, I., Argraves, W.S., Ruoslahti, E. and Engvall, E. (1990) Merosin, a tissue-specific basement membrane protein, is a laminin-like protein. Proc. Natl. Acad. Sci. USA 87: 3264-3268. Eldridge, C.F., Sanes, J.R., Chui, A.Y., Bunge, R.P. and Cornbrooks, C.J. (1986) Basal lamina-associated heparan sulfate proteoglycan in the rat peripheral nervous system:

83 characterization and localization using monoclonal antibodies. J. Neurocytol. 15: 37-51. Engel, J. (1991) Common structural motifs in proteins of the extracellular matrix. Curr. Opin. Cell Biol. 3: 779-785. Engvall, E., Earwicker, D., Haaparant, T., Ruoslahti, E. and Sanes, J.R. (1990) Distribution and isolation of four laminin variants: tissue restricted distribution of heterotrimers assembled from five different subunits. Cell Reg. 1: 731-740. Engvall, E., Earwicker, D., Day, A., Muir, D., Manthorpe, M. and Paulsson, M. (1992) Merosin promotes cell attachment and neurite outgrowth and is a component of the neuritepromoting factor of RN22 schwannoma cells. Exp. Cell Res. 198: 115-123. Fernandezvalle, C, Gwynn, L., Wood, P.M., Carbonetto, S. and Bunge, M.B. (1994) Anti-beta-1 integrin antibody inhibits Schwann cell myelination. J. Neurobiol. 25: 12071226. Fischman, R.B. and Hatten, M.E. (1993) Multiple receptor systems promote CNS neural migration. J. Neurosci. 13: 3485-3495. Fujiwara, S., Shinkai, H., Deutzmann, R., Paulsson, M. and Timpl, R. (1988) Structure and distribution of N-linked oligosaccharide chains on various domains of mouse tumour laminin. Biochem. J. 252: 453^61. Gehlsen, K.R., Dickerson, K., Argraves, W.S., Engvall, E. and Ruoslahti, E. (1989) Subunit structure of a laminin-binding integrin and localization of its binding site on laminin. J. Biol. Chem. 264: 19034-19038. Gomez, T.M. and Letourneau, P.C. (1994) Filopodia initiate choices made by sensory neuron growth cones at laminin fibronectin borders in vitro. J. Neurosci. 14: 5959-5972. Goodman, S., Deutzmann, R. and Nurcombe, V. (1989) Locomotory competence and laminin-specific cell surface binding sites are lost during myoblast differentiation. Development 106: 795-802. Goodman, S.L., Aumailley, M. and von der Mark, H. (1991) Multiple cell surface receptors for the short arms of laminin: integrin and RGD-dependent proteins mediate cell attachment only to domains III in murine tumor laminin. J. Cell Biol. 113:931-941. Graf, J., Iwamoto, Y., Sasaki, M., Martin, G.R., Kleinman, H.K., Robey, F.A. and Yamada, Y. (1987) Identification of an amino acid sequence in laminin mediating cell attachment chemotaxis and receptor binding. Cell 48: 989-996. Green, T.L., Hunter, D.D., Chan, W., Merlie, J.P. and Sanes, J.R. (1992) Synthesis and assembly of the synaptic cleft protein S-laminin by cultured cells. J. Biol. Chem. 267: 2014-2022. Grumet, M., Friedlander, D.R. and Edelman, G.M. (1993) Evidence for the binding of NG-CAM to laminin. Cell Adhesion Commun. 1: 177-190. Haas, T.A. and Plow, E.F. (1994) Integrin-ligand interactions: a year in review. Curr. Opin. Cell Biol. 6: 656-662. Hall, D.E., Reichardt, L.F., Crowley, E., Holley, B., Moezzi, H., Sonnenberg, A. and Damsky, C.H. (1990) The alpha-1 beta-1 and alpha-6 beta-1 integrin heterodimers mediate

84 cell attachment to distinct sites on laminin. J. Cell Biol 110:2175-2184. Hayashi, Y.K., Engvall, E., Arikawa-Hirasawa, E., Goto, K., Koga, R., Nonaka, I., Sugita, H. and Arakata, K. (1993) Abnormal localisation of laminin subunits in muscular dystrophies. J. Neurol ScL 119: 53-64. Horwitz, A.F. (1991) More than just scaffolding....Cwrr. Biol 1:6-7. Howe, C.C. (1984) Functional role of laminin carbohydrate. Mol Cell Biol 4:1-1. Hunter, D.D., Shah, V., Merlie, J.P. and Sanes, J.R. (1989) A laminin-like adhesive protein concentrated in the synaptic cleft of the neuromuscular junction. Nature 338: 229-234. Hunter, D.D., Cashman, N., Morris-Valero, R., Bulock, J.W., Adams, S.P. and Sanes, J.R. (1992) An LRE (leucinearginine-glutamate)-dependent mechanism for adhesion of neurons to S-laminin. J. Neuroscl 11: 3960-3971. Hynes, R.O. (1987) Integrins: a family of cell surface receptors. Ce//48: 549-554. Hynes, R.O. and Lander, A.D. (1992) Contact and adhesive specificities in the associations, migrations, and targeting of cells and axons. Cell 68: 303-322. Ignatius, M.J. and Reichardt, L.F. (1988) Identification of a neuronal laminin receptor: an M^ 200K/120K integrin heterodimer that binds laminin in a divalent cation-dependent manner. Neuron 1: 713-725 Ignatius, M.J., Large, T.H., Houde, M., Tawil, J.W., Barton, A., Esch, F., Carbonetto, S. and Reichardt, L.F. (1990) Molecular cloning of the rat integrin alpha-1 subunit: a receptor for laminin and collagen. J. Cell Biol 111: 709720. Ishii, N., Wadsworth, W.G., Stern, B.D., Culotti, J.G. and Hedgecock, E.M. (1992) UNC-6, a laminin-related protein, guides cell and pioneer axon migrations in C. elegans. Neuron 9: S13-SS\. Johnson, W.C, Pagano, T.G., Basson, C.T., Madri, J.A., Gooley, P. and Armitage, J.M. (1993) Biologically active Arg-Gly-Asp oligopeptides assume a type II /5-turn in solution. Biochemistry 32: 268-273. Kleinman, H.K., Weeks, B.S., Cannon, F.B., Sweeney, T.M., Sephel, G.C., Clement, B., Zain, M., Olson, M.O., Jucker, M. and Burrous, B.A. (1991) Identification of a llOkDa nonintegrin cell surface laminin-binding protein which recognizes an A chain neurite-promoting peptide. Arch. Biochem. Biophys. 290: 320-325. Kornberg, L.L., Earp, S.H., Turner, C.E., Prockop, C. and Juliano, R.L. (1991) Signal transduction by integrins: increased protein tyrosine phosphorylation caused by clustering of beta-1 integrins. Proc. Natl Acad. ScL USA 88: 8392-8396. Kuecherer-Ehret, A., Graeber, M.B., Edgar, D., Thoenen, H. and Kreutzberg, G.W. (1990) Immunoelectron microscopic localization of laminin in normal and regenerating mouse sciatic nerve. J. Neurocytol 19: 101-109. Lallier, T., Artinger, M., Matthew, W. and Bronner-Fraser, M. (1990) Distribution and biochemical characterization of the

Laminin in neural development INO antigen during chick neural crest cell migration. Neuroscl Res. Suppl. 13: S126-140. Lein, P.J., Higgins, D., Turner, D.C., Rier, L.A. and Terranova, V.P. (1991) The NCI domain of type IV collagen promotes axonal growth in sympathetic neurons through interaction with the alpha-1 beta-1 integrin. J. Cell Biol 113:417-428. Lemmon, V., Burden, S.M., Payne, H.R., Elmslie, G.J. and Hlavin, H.L. (1992) Neurite growth on different substrates: permissive versus instructive influences and the role of adhesive strength. /. Neuroscl 12: 818-826. Letourneau, P.C. and Shattuck, T.A. (1989) Distribution and possible interactions of actin-associated proteins and celladhesion molecules of nerve growth cones. Development 105:505-519. Liesi, P., Narvanen, A., Soos, J., Sariola, H. and Snounou, G. (1989) Identification of a neurite outgrowth-promoting domain of laminin using synthetic peptides. FEBS Lett. 244: 141-148. Luckinbilledds, L., Kaiser, C.A., Rodgers, T.R. and Powell, D.D. (1995) Localisation of the llOkDa receptor for laminin in brains of embryonic and postnatal mice. Cell Tissue Res. 219: 311-311. Martin, G.R. and Timpl, R. (1987) Laminin and other basement membrane components. Annu Rev. Cell Biol 3: 5785. Masumoto, A. and Hemler, M.E. (1993) Mutation of putative divalent cation sites in the a-4 subunit of the integrin VLA4: distinct effects on adhesion to CSl/fibronectin, V-CAMl and invasin. J. Cell Biol 123: 245-253. McCaffery, P., Neve, R.L. and Drager, U.C. (1990) A dorsoventral asymmetry in the embryonic retina defined by protein conformation. Proc. Natl Acad. Scl USA 87: 85708574. McKenna, M.P. and Raper, J.A. (1988) Growth cone behaviour on gradients of substratum-bound laminin. Dev. Biol 130: 232-236. Mercurio, A.M. (1990) Laminin: multiple forms, multiple receptors. Curr. Opin. Cell Biol 2: 845-849. Michishita, M., Videm, V. and Arnauout, M.A. (1993) A novel divalent integrin cation-binding site in the A domain of the p2 integrin CR3 (CDllc/CD18) is essential for ligand binding. Cell 72: 857-867. Mosher, D.F. (1991) How do integrins integrate? Curr. Biol. 1:394-396. Neugebauer, K.M. and Reichardt, L.F. (1991) Cell-surface regulation of beta-1 integrin activity on developing retinal neurons. Nature 350: 68-71. Neugebauer, K.M., Tomaselli, K.J., Lillien, J. and Reichardt, L.F. (1988) N-cadherin, N-CAM and integrins promote retinal neurite outgrowth on astrocytes in vitro. J. Cell Biol 107:1177-1187. Noakes, P.G., Gautam, M., Mudd, J., Sanes, J.R. and Merlie, J.P. (1995) Aberrant differentiation of neuromuscular junctions in mice lacking s-laminin (laminin P-2) Nature 374: 258-262.

V. Nurcombe Noonan, D.M., Fulle, A., Valente, P., Cai, S., Horigan, E., Sasaki, M., Yamada, Y. and Hassell, J.R. (1991) The complete sequence of perlecan, a basement membrane heparan sulfate proteoglycan, reveals extensive similarity with laminin A chain, low density lipoprotein receptor, and the neural cell adhesion molecule. J. Biol. Chem. 266: 2293922947. Nurcombe, V., Aumailley, M., Timpl, R. and Edgar, D. (1989) The high affinity binding of laminin to cells. Eur. J. Biochem. 180: 9-14. Obremski, V.J. and Bunge, M.B. (1995) Addition of purified basal lamina molecules enables Schwann cell ensheathment of sympathetic neurites in culture. Dev. Biol. 168: 124-137. Paulsson, M. (1988) The role of Ca^"^ binding in the selfaggregation of laminin-nidogen complexes. J. Biol. Chem. 263: 5424-5430. Rabacchi, S.A., Neve, R.L. and Drager, U.C. (1990) A positional marker for the dorsal embryonic retina is homologous to the high-affinity laminin receptor. Development 109: 521-531. Rao, N.C., Barsky, S.H., Terranova, V.P. and Liotta, L.A. (1983) Isolation of a tumor cell laminin receptor. Biochem. Biophys. Res. Commun. Ill: 804-808. Reh, T.A. and Radke, K. (1988) A role for the extracellular matrix in retinal neurogenesis in vitro. Dev. Biol. 129: 283293. Reichardt, L.F. and Tomaselli, K.J. (1991) Extracellular matrix molecules and their receptors: functions in neural development. AAZ^IM. Rev. Neurosci. 14: 531-570. Riopelle, R.J. and Dow, K.E. (1991) Neurite formation on laminin: effects of a galactosyltransferase on primary sensory neurons. Brain Res. 541: 265-272. Rivas, R.J., Burmeister, D.W. and Goldberg, D.J. (1992) Rapid effects of laminin on the growth cone. Neuron 8: 107-115. Rossini, P., Gavazzi, I., Timpl, R., Aumailley, M., Abbadini, M., Giancotti, F., Silengo, L., Marchisio, P.C. and Tarone, G. (1990) Nerve growth factor induces increased expression of a laminin-binding integrin in rat pheochromocytoma PC12 cells. Exp. Cell Res. 189: 100-108. Ruoslahti, E. (1988) Structure and biology of proteoglycans. Annu. Rev. Cell Biol. 4: 229-255. Ruoslahti, E. (1991) Integrins. J. Clin. Invest. 87: 1-5. Sanes, J.R. (1989) Extracellular matrix molecules that influence neural development. Annu. Rev. Neurosci. 12: 491516. Sanes, J.R. and Chui, A.Y. (1983) The basal lamina of the neuromuscular junction. Cold Spring Harbour Symp. Quant. Biol. 48: 667-678. Sanes, J.R., Engvall, E., Butkowski, R. and Hunter, D.D. (1990) Molecular heterogeneity of basal laminae: isoforms of laminin and collagen IV at the neuromuscular junction and elsewhere. 7. Cell Biol. I l l : 1685-1699. Sarthy, P.V. and Fu, M. (1990) Localization of laminin Bl mRNA in retinal ganglion cells by in situ hybridization. J. Cell Biol. 110:2099-2108.

85 Sasaki, M. and Yamada, Y. (1987) The laminin B2 chain has a multidomain structure homologous to the Bl chain. J. Biol. Chem. 262: 17111-17117. Sasaki, M., Kato, S., Kohno, K., Martin, G.R. and Yamada, Y. (1987) Sequence of the cDNA encoding the laminin Bl chain reveals a multidomain protein containing cysteinerich repeats. Proc. Natl. Acad. Sci. USA 84: 935-939. Sasaki, M., Kleinman, H.K., Huber, H., Deutzman, R. and Yamada, Y. (1988) Laminin, a multidomain protein: the A chain has a unique globular domain and homology with the basement membrane proteoglycan and the laminin B chain. J. Biol. Chem. 263: 16536-16544. Schittny, J.C, Timpl, R. and Engel, J. (1988) High resolution immunoelectron microscopic localization of functional domains of laminin, nidogen, and heparan sulfate proteoglycan in epithelial basement membrane of mouse cornea reveals different topological orientations. J. Cell Biol. 107: 1599-1610. Schubert, D. (1992) Collaborative interactions between growth factors and the extracellular matrix. Trends Cell Biol. 2: 63-66. Schubert, D. and Kimura, H. (1991) Substratum-growth factor collaborations are required for the mitogenic activities of activin and FGF on embryonal carcinoma cells. J. Cell Biol. 114:841-846. Scott, L.J.C., Balsamo, J., Sanes, J.R. and Lillien, J. (1990) Synaptic localization and neural regulation of an A^acetylgalactosaminyltransferase in skeletal muscle. Neuroscience 10: 346-350. Sephel, G.C., Tashiro, K., Sasaki, M., Kandel, S., Yamada, Y. and Kleinman, H.K. (1989) A laminin-pepsin fragment with cell attachment and neurite outgrowth activity at distinct sites. Dev. Biol. 135: 172-181. Shewan, D., Berry, M. and Cohen, J. (1995) Extensive regeneration in vitro by early embryonic neurons on inmiature and adult CNS tissue. J. Neurosci. 15: 2057-2062. Shih, D.T., Edelman, J.M., Horowitz, A.F., Grunwald, G.B. and Buck, C.A. (1993) Structure/function analysis of the integrin p-l by epitope mapping. J. Cell Biol. 122: 13611371. Shur, B.D. (1993) Glycosyltransferases as cell adhesion molecules. Curr. Opin. Cell Biol. 5: 854-863. Siebold, B., Deutzmann, R. and Kuehn, K. (1988) The arrangement of intra- and intermolecular disulfide bonds in the carboxyterminal, noncollagenous aggregation and cross-linking domain of basement membrane type IV collagen. Eur. J. Biochem. 176: 617-624. Skubitz, A.P., Letourneau, P.C, Wayner, E. and Furcht, L.T. (1991) Synthetic peptides from the carboxy-terminal globular domain of the A chain of laminin: their ability to promote cell adhesion and neurite outgrowth, and interact with heparin and the)3-l integrin subunit. J. Cell Biol. 115: 1137-1148. Smalheiser, N.R. and Schwartz, N.B. (1987) Cranin: a laminin-binding protein of cell membranes. Proc. Natl. Acad. Sci. USA 84: 6457-6461.

86 Sonnenberg, A., Linders, C.J.T., Modderman, P.W., Damsky, C.H., Aumailley, M. and Timpl, R. (1990) Integrin recognition of different cell binding fragments of laminin (PI, E3, E8) and evidence that alpha-6 beta-1 but not alpha-6 beta-4 functions as a major receptor for fragment E8. /. Cell Biol 110:2167-2174. Sung, U., O'Rear, J.J. and Yurchenco, P.D. (1993) Cell and heparin-binding sites in the distal long arm of laminin: identification of active and cryptic sites with recombinant and hybrid glycoprotein. 7. Cell Biol 123: 1255-1268. Sweeney, T.M., Ogle, R.C. and Little, CD. (1990) Laminin potentiates differentiation of PCC4uva embryonal carcinoma into neurons. J. Cell ScL 97: 23-31. Takada, Y. and Puzon, W. (1993) Identification of a regulatory region of integrin y3-l subunit using activating and inhibiting antibodies. 7. Biol Chem. 268: 17597-17601. Tashiro, K., Sephel, G.C., Weeks, B., Sasaki, M., Martin, G.R., Kleinman, H.K. and Yamada, Y. (1989) A synthetic peptide containing the IKVAV sequence from the A chain of laminin mediates cell attachment, migration, and neurite outgrowth. 7. Biol Chem. 264: 16174-16182. Thomas, W.A., Schaefer, A.W. and Treadway, R.M. (1990) Galactosyltransferase-dependence of neurite outgrowth on substratum-bound laminin. Development 110: 1101-1114. Timpl, R., Rohde, A., Gehron-Robey, P., Rennard, I.T., Foidart, J.M. and Martin, G.R. (1979) Laminin, a glycoprotein from basement membrane. 7. Biol Chem. 254: 9933-9937. Tomaselli, K.J., Damsky, C.H. and Reichardt, L.F. (1987) Interactions of a neuronal cell line (PC 12) with laminin, collagen IV, and fibronectin: identification of integrinrelated glycoproteins involved in attachment and process outgrowth. 7 Cell Biol 105: 2347-2358. Tomaselli, K.J., Damsky, C.H. and Reichardt, L.F. (1988) Purification and characterization of mammalian integrins expressed by a rat neuronal cell line (PC 12): evidence that they function as alpha/beta heterodimeric receptors for laminin and type IV collagen. 7 Cell Biol 107: 12411252. Tomaselli, K.J., Hall, D.E., Rier, L.A., Gehlsen, K.R., Turner, D.C., Carbonetto, S. and Reichardt, L.F. (1990) A neuronal cell line (PC 12) expresses two beta 1-class integrins-alpha1 beta-1 and alpha-3 beta-1 that recognize different neurite outgrowth-promoting domains in laminin. Neuron 5: 651662.

Laminin in neural development Tryggvason, K. (1993) The laminin family. Curr. Biol 5: 877-882. Vuolteenaho, R., Nissenen, M., Saino, K., Byers, M., Eddy, R., Hirvonen, H., Shows, T.B., Sariola, T.B., Engvall, E. and Tryggvason, K. (1994) Human laminin M chain (merosin): complete primary structure, chromosomal alignment and expression of the M and A chains in human fetal tissue. 7 Cell Biol 124: 381-394. Weaver, CD., Yoshida, C.K., De Curtis, I. and Reichardt, L.F. (1995) Expression and in vitro function of beta(l) integrin laminin receptors in the developing avian ciliary ganglion. 7 NeuroscL 15: 5275-5285. Weskamp, G. and Blobel, CP. (1994) A family of cellular proteins related to snake venom disintegrins. Proc. Natl Acad. Scl USA 91: 2748-2751. Wujek, J.R., Haleem-Smith, H., Yamada, Y., Lipsky, R., Lan, Y.T. and Freese, E. (1990) Evidence that the B2 chain of laminin is responsible for the neurite outgrowth-promoting activity of astrocyte extracellular matrix. Dev. Brain Res. 55: 237-247. Yamada, K.M. (1991) Adhesive recognition sequences. 7. Biol Chem. 266: 12809-12812. Yip, J.W. and Yip, Y.P.L. (1992) Laminin - developmental expression and role in axonal outgrowth in the peripheral nervous system of the chick. Dev. Brain Res. 68: 23-33. Yurchenco, P.D., Tsilibary, E.C, Charonis, A.S. and Furthmayr, H. (1985) Laminin polymerization in vitro: evidence for a two-step assembly with domain specificity. 7. Biol Chem. 260: 7636-7644. Yurchenco, P.D. and Schittny, J.C (1990) Molecular architecture of basement membranes. FASEB J. 4: 1577-1590. Yurchenco, P.D. and Cheng, Y.S. (1993) Self assembly and calcium binding sites in laminin: a 3 arm interaction model. 7 Biol Chem. 268: 17286-17299. Yurchenco, P.D. (1994) Assembly of laminin and type IV collagen into basement membrane networks. In: P.D. Yurchenco, D.E. Birk and R.P. Mecham (Eds.), Extracellular Matrix Assembly and Structure. Academic Press, San Diego, CA, pp. 351-388. Yurchenco, P.D. and O'Rear, J.J. (1994) Basal lamina assembly. Curr. Opin. Cell Biol 6: 674-681. Zhou, F.C (1990) Four patterns of laminin-immunoreactive structure in developing rat brain. Dev. Brain Res. 55: 191201.

Section II

Factors Implicated in Neuron Survival and Specialization

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 4

Mechanisms of developmental cell death A. Messina and A. Jaworowski Department of Medicine, Melbourne University, Royal Melbourne Hospital, Parkville, 3050, Australia

1. Introduction 1.1, Initial descriptions of developmental cell death It has been obvious to developmental biologists for some decades that cell death is a prominent part of development (Glucksmann, 1951) and that cells are ^capable of effecting their own demise' and disposing of their own remains, or are able to be earmarked for death and slain by extrinsic agents (Saunders, 1966). However, it has been revealed only relatively recently that the mode by which this cell death and disposal is effected differs vastly from necrosis. In 1965, Kerr described two morphologically distinct types of cell death that occurred in rat liver, following ischaemia. In the first type the morphology was typical of necrosis; cells died in patches within hours and this was associated with inflammation and intracellular rupture of lysosomes. The second type of cell death presented a different scenario; it occurred at a later stage in only scattered individual hepatocytes and was not associated with inflammation or rupture of cell lysosomes. In this second type of cell death, the dying cells were converted into small round masses which were phagocytosed by other cells within 24 h. Importantly, these dying cells were sometimes detected in normal liver, as well as following ischaemia (Kerr, 1965). Kerr (1971) studied the development of these small, round masses, a process which he referred to as shrinkage necrosis, by electron microscopy; The earliest recognisable stage in their formation involves condensation of cytoplasm of parenchymal

cells and separation of the plasma membranes from those of adjacent cells. There is sometimes extensive villous transformation of the cell surface. The nuclei show either aggregation of their chromatin beneath the nuclear envelope or fragmentation to produce dense masses of granules that are only occasionally surrounded by membranes'. Degradation was seen to occur only after ingestion by parenchymal cells and histiocytes (Kerr, 1971). By the end of 1972, this type of cell death had been identified morphologically in a surprisingly large number of physiological and pathological conditions, and Kerr and colleagues coined the term apoptosis to describe the phenomenon (Kerr et al., 1972). These authors stressed its importance as a general mechanism of controlled cell deletion, complementary to mitosis in the regulation of cell populations, and that it could be triggered by physiological stimuli, noxious agents, or occur spontaneously. They also suggested that it was an active, inherently programmed phenomenon and speculated that it might involve stimulation of mRNA and protein synthesis. The programmed or developmental cell death described many decades earlier by developmental biologists showed all the morphological hallmarks of apoptosis (Kerr et al., 1972). 1.2. Apoptosis is defined by characteristic morphological changes Accurately applied, apoptosis is therefore a descriptive term which refers to a widely occurring phenomenon of cell death, characterized by a distinct set of morphological features. These are:

90

(a) condensation of chromatin into uniformally dense, circumscribed masses with or without margination towards the nuclear envelope. (b) detatchment of cells from their neighbours (where applicable). (c) condensation of the cytoplasm. (d) convolution of nuclear and cellular outlines. (e) fragmentation of the nucleus. (f) blebbing of the plasma membrane leading to budding off of the cellular contents as membrane-bound vesicles. These morphological changes can be completed in several minutes for example in the case of apoptosis induced in vitro by T-cells in their target cells. Although it is not known how quickly the process occurs in vivo, apoptotic cells with characteristic condensed chromatin may persist for a few hours, before being phagocytosed (Russell et al., 1972; Sanderson, 1976; Matter, 1979; Bursch et al., 1990) but in other circumstances they are phagocytosed at the time of, or even before chromatin condensation (Savill et al., 1993; Lagasse and Weissman, 1994; Matsubara et al., 1994). 1.3. Apoptosis occurs in widely differing physiological circumstances

Mechanisms of developmental cell death

(b) Apoptosis occurs as a protective or defensive mechanism by which cells that are aberrant, injured, or perceived as being potentially harmful are eliminated. This may occur as a spontaneous event triggered by inherent mechanisms following mutation or viral infection of a cell, or else it may be mediated by the immune system, whereby cytotoxic T lymphocytes actively induce apoptosis in target cells. (c) Apoptosis can be triggered experimentally in cells by non-physiological stimuli. Examples of this include experimentally-induced DNA damage or derangements of metabolism following irradiation or treatment with cytotoxic drugs. Apoptosis is also triggered by inappropriate oncogene expression, receptor activation and/or changes in signal transduction pathways. Thus, cell death is accompanied by similar morphological changes in a wide variety of circumstances. However, as we shall see below, they do not involve the same mechanisms. In this context, one should be careful not to use the functional terms programmed or developmental cell death, to denote other types of cell death of which the end result is apoptosis but which are clearly not developmental. 1.4. Are there common mechanisms of apoptosis?

The circumstances of cell death in which the final outcome is apoptosis can be loosely, but not exclusively, separated into 3 groups. (a) Apoptosis occurs during embryogenesis, development and ageing as a normal physiological process that results in deletion of apparently normal cells. During embryogenesis, apoptosis occurs in cells as a pre-programmed fate determined by cell lineage and is an essential part of the orderly development of the embryo. This phenomenon is called programmed cell death. During development, apoptosis mediates refinement of the nervous system, as cells compete for limited amounts of survival-inducing trophic factors, as well as maturation of the immune system, where deletion of potentially autoreactive cells occurs via a process of negative selection. During ageing, apoptosis maintains homeostasis by removal of cells that are redundant or senescent.

In all of the above instances in which cell death occurs, the cell is triggered initially by an external or internal signal or else by some derangement of intracellular homeostasis. The message is then processed and either ignored or executed. If executed, a number of effector pathways are activated which lead to cytoplasmic condensation, chromatin condensation and fragmentation and packaging of cellular contents, followed by signals which facilitate recognition of the apoptotic cell and phagocytosis. It is clear that the effector pathways leading to apoptosis can differ between cells and even within cells when triggered by different stimuli. What is not clear at present is whether these eventually converge into one common pathway. The fact that the morphological features of apoptosis are similar in a wide variety of cases suggests that the down-

A. Messina and A. Jaworowski

Stream mechanisms are conserved but, even here, there are differences depending on the system studied and, to date, no one feature of apoptosis has been found to be indispensable. 2. Models of apoptosis Many models are currently being used to study apoptosis. These include the study of growth factor-deprived cell cultures, cells treated with radiation or cytoxic agents, virus-infected cells and many models in vivo and in vitro of programmed or developmental cell death. In this review, we will concentrate on models of developmental cell death. These include developmental cell death in the nematode, C. elegans, primary cultures of neurotrophic-dependent neurons, the pheochromocytoma cell line, PC 12, and primary cultures of immature thymocytes. Studies involving these models of cell death have defined molecules and pathways that may be important in a wide variety of types of apoptosis. 2.1. C. elegans C. elegans has proven to be an important model for programmed cell death and an invaluable tool in identifying genes involved in this process. C. elegans is a microscopic worm consisting of exactly 1090 cells. The origin and fate of each cell has been tracked to reveal that, during the worm's development, 131 cells undergo apoptosis (Sulston and Horvitz, 1977; Sulston et al., 1983). Horvitz and colleagues have studied mutations that influence apoptosis in C-elegans and identified a number of 'cell death' (ced) genes. Fourteen such genes have been identified, however only two, ced-3 and ced'4, are essential for cell death, (Ellis and Horvitz, 1986; Yuan and Horvitz, 1990; Ellis et al., 1991b) in that recessive mutations of these genes prevent apoptosis in most of the 131 cells that normally die. A third gene, ced'9 protects C. elegans cells from apoptosis by antagonising ced-3 and ced-4. It is only necessary for survival when ced-3 and ced-4 are functional (Hengartner etal., 1992).

91

These genes have been cloned and sequenced. The cloning and sequencing of Ced-4 did not reveal an homology with any known mammalian protein (Yuan and Horvitz, 1992) however the predicted amino acid sequence of Ced-9 was shown to have 23% homology with Bcl-2 (Hengartner et al, 1992) and that of Ced-3 was shown to have 28% homolgy with Interleukin l^S-converting enzyme (ICE) (Yuan et al, 1993). The roles of these proteins in apoptosis will be discussed below. 2.2. Neuronal models Neurons in both the developing and adult peripheral nervous systems require neurotrophic support for growth and survival (Oppenheim, 1991). In many cases this is mediated by a family of neurotrophins which includes nerve growth factor (NGF), brain derived neurotrophic factor (BDNF), ciUary neurotrophic factor (CNTF), neurotrophin-3 (NT-3) and neurotrophin-4/5 (NT-4/5). During development, a large proportion of sympathetic, motor and sensory neurons die via an apoptotic mechanism and this is thought to be mainly due to a failure of these neurons to obtain sufficient amounts of targetderived neurotrophic factors. NGF exposure increases expression of its cognate receptors in responsive neurons so that they become dependent on NGF for survival (Miller et al, 1991). Withdrawal of NGF from primary cultures of these neurons forms the basis of models of neuronal developmental cell death in vitro. NGF withdrawal was first shown to induce apoptosis in cultures of sympathetic neurons and this was prevented by inhibitors of protein and RNA synthesis such as cycloheximide and actinomycin D respectively (Martin et al., 1988). Protein and RNA synthesis inhibitors were subsequentially shown to rescue CNTF-, BDNFand NGF-dependent neurons from apoptosis in vitro (Scott and Davies, 1990) and to reduce naturally occurring neuronal cell death in chick somatic motor and sensory neurons (Oppenheim et al., 1990). The pheochromocytoma cell line PC 12 is widely used to study the mechanisms of neuronal cell death following NGF withdrawal. PC 12 cells have the functional receptors for NGF, p75 (p75ngfr)

92

and pi40^^ (TrkA) and these cells respond to NGF by forming *neurite' outgrowths, reminiscent of those found in neurons in vivo, and by differentiating into cells with a sympathetic neuronal phenotype (Huff et al., 1981). Furthermore, when deprived of NGF these neuronal PC 12 cells undergo apoptosis (Greene, 1978; Rukenstein et al., 1991) with similar kinetics to sympathetic neurons (Mesner et al., 1992). Like the process induced in primary cell culture models, apoptosis in neuronal PC 12 cells is prevented by inhibitors of protein and mRNA synthesis (Mesner et al., 1992). It should be noted at this point that naive PC 12 cells also undergo apoptosis when cultured in serumfree medium. These cells can be rescued by NGF but, unlike neuronally differentiated PC 12 cells, they can also be rescued by serum (Greene, 1978). Since naive PC 12 cells can be rescued from apoptosis by NGF they too have been used as a model for developmental cell death. However, these cells behave differently from their post-mitotic neuronal counterparts in that they can be transformed and are capable of mitosis (Greene, 1978) and in that they are not rescued from NGF-deprivation induced apoptosis by protein and RNA synthesis inhibitors (Rukenstein et al., 1991). 2.3. The immune system Perhaps the most widely studied processes of cell death are those associated with T cells. Apoptosis occurs during development to remove potentially autoreactive thymocytes in the thymus and also occurs later, following T cell activation, in both the T cell itself and in the target cell. During development, cortical thymocytes undergo a selection process before being released into the periphery. Early in this process they express a pre-T cell receptor which, when activated, induces expression of the co-receptors CD4 and CD8. These cells complete a limited number of divisions and then express the mature ap T-cell receptor (TCR). The TCR consists of two components, an invariant CD3 (signal transducing) component and a variable Tj (antigen binding) component. At this point the thymocytes are often designated CD4+CD8+. These cells are programmed to die within 3 ^ days

Mechanisms of developmental cell death

unless they are able to interact in a specific manner with the MHC antigens in the thymus. Thymocytes whose TCR fails to interact with MHC, die. Thymocytes that react with 'self antigens complexed with MHC are eliminated by an apoptotic mechanism in a process termed negative selection. Thymocytes that do not express autoreactive Ti antigens differentiate into mature CD4+/CD8" or CD4~ /CD8"^ T-cells. (see Kisielow and von Boehmer, 1995 for a review). These mechanisms of cell death are poorly understood but probably include (1) an antigenspecific mechanism which involves triggering of the TCR, CD4 and CD8, as well as MHC/antigen presenting cells and (2) non-antigen-specific mechanisms which involve the membrane receptor Fas (see Section 3.2) (Kisielow et al., 1988a,b; Sha et al, 1988; Smith et al., 1989; Swat et al., 1991; Debatin, 1994; Debatin et al., 1994; McConkey et al., 1994; Kisielow and von Boehmer, 1995). Activation of the TCR and Fas on thymocytes forms the basis of models of developmental cell death. Mature T cells may also undergo apoptosis if they interact with 'self antigens in the periphery. In mature T cells, activation of the TCR induces proliferation via an autocrine mechanism by stimulating the synthesis of both IL-2 and IL-2 receptor (Meuer et al., 1984). On withdrawal of IL-2 these cell undergo apoptosis by a mechanism that requires synthesis of both protein and RNA (Duke and Cohen, 1986). However, if IL-2 or other growth factors are available, these cells proliferate for 3-4 days, after which time they undergo apoptosis, a phenomenon referred to as activation-induced cell death (AICD). Protein and RNA synthesis is required at the time of activation for AICD to proceed (Shi et al, 1989; Ucker et al, 1989; Green and Scott, 1994). Although not strictly developmental cell death, AICD has been used to study the involvement of molecules such as Fas (see Section 3.2) and Myc (see Section 5.2) in the T cell. CD8+ cytotoxic T-cells (CTL) bind to virally infected, tumour or allogenic cells and induce apoptosis in these target cells via a mechanism that is not prevented by inhibitors of protein and RNA synthesis (Duke et al., 1983; Ostergaard and Clark,

A. Messina and A. Jaworowski

1989). Two mechanisms have been described by which apoptosis is induced in the target cell, which can be used independently by different effector cells or sequentially by the same cell (Berke, 1994). One mechanism involves expression of Fas ligand (see Section 3.2) which induces apoptosis in Fas-expressing targets, while the other is a membranolytic mechanism in which apoptosis is initiated by a number of enzyme-bearing granules that are transferred into the target cell via perforin holes (Berke, 1994, Kagi et al., 1994). 2,4. The effect of metabolic inhibitors on apoptosis in the above models In the different models described above, apoptosis is triggered and mediated by different mechanisms. In some models, protein and RNA synthesis inhibitors rescue cells from death (Duke and Cohen 1986; Martin et al, 1988; Mesner et al, 1992), and demonstrations of this in early studies were useful in promoting the concept that apoptosis is an active process. However, in many later studies using other systems, these same inhibitors either had no effect or even stimulated cell death (for example see Shi et al., 1992; Vaux et al., 1992). These observations can be explained as follows: (1) In cells that either do not constitutively express the machinery for apoptosis or rapidly turn over the components of this machinery, metabolic inhibitors lead to depletion of apoptosisinducing proteins and thereby offer protection (McConkey et al., 1990; Gaido and Cidlowski, 1991). (2) In cells that constitutively express the apoptotic machinery (now thought to be the majority of cells) but suppress its activity via mechanisms that require continued protein synthesis, metabolic inhibitors enhance apoptosis. The situation, however, is not so clear cut as the above descriptions imply, since multiple mechanisms appear to operate in some cells, depending on the stimulus and the state of the cell (Wyllie et al., 1984; Rukenstein et al., 1991). A similar situation exists for the mechanisms that suppress or protect from apoptosis (Edwards et al., 1991; Ni et al., 1994; Rukenstein et al., 1991). In recent years it has become obvious that a number of initiation

93

pathways for apoptosis exist. These may feed into a common final pathway or else can be activated at different points, which may bypass the requirement for de novo protein synthesis (Vaux and Wisemman, 1993; Lindenboim et al., 1995) In the preceding sections, we have described the major models used to study developmental cell death. We will now discuss the cell death pathways implicated by these models, which lead to apoptosis. For convenience, apoptosis may be subdivided into the initial signalling events, the apoptotic signal transduction pathways and the terminal events involved in the observed phenotypic changes. 3. Initial events In the case of programmed cell death that occurs during embryogenesis, there is evidence that the mechanism of cell death is inherent and independent of interactions with other cells. Recent evidence for this assertion comes mainly from studies in C. elegans (see Ellis et al., 1991b for a review). In other cases of developmental cell death, however, such as that which occurs in neuronal and Tcell development, the initial signal to undergo apoptosis is provided via cell surface receptors in response to external signals. These cell surface receptors may have either a positive or a negative influence on apoptosis. 3.L Receptors for growth and survival inhibit apoptosis It has been proposed that cells have a tendency to undergo apoptosis unless prevented from doing so by survival factors, cell-cell contact or autocrine signals (Raff, 1992; Raff et al., 1993). Cells differentiate or proliferate in response to growth factors, which may also act as survival factors by inhibiting apoptosis. It is becoming clear that different signal transduction pathways may be activated via different regions of a given growth factor receptor to promote proliferation or differentiation on the one hand, or to inhibit apoptosis and promote survival on the other. The actions of NGF on neuronal cells are mediated by the high-affinity TrkA and

94

low-affinity p75"^^^ receptors (Meakin and Shooter, 1992). NGF prevents apoptosis when bound to either receptor, however the two receptors differ in that expression of p75"^^^ in immortalised neural cells actively induces cell death (Rabizadeh et al., 1993b; Rabizadeh and Bredesen, 1994). There is recent evidence that the ability of TrkA to prevent apoptosis involves activation of phosphatidylinositol 3-OH kinase (PI-3 kinase) and is not dependent on Ras- or MAP-kinase activation whereas its ability to induce differentiation is Ras-dependent (Yao and Cooper, 1995). Similarly, the C-terminal domain of the y3 subunit of the IL3/GM CSF receptor is required for growth factor-mediated cell survival (Sato et al, 1993; Kinoshita et al., 1995) but is not involved in mediating short-term proliferation (i.e. progression from Gl to S). Undoubtedly, many more such studies will appear in the near future to identify regions of growth factor receptors involved in triggering signal transduction pathways that promote cell survival by suppressing apoptosis. 5.2. Fas receptor Cell surface receptors may also induce apoptosis when activated following ligand binding. The two best-characterised receptors which do this are the receptor, Fas, and the tumour necrosis factor (TNF) receptor. Of these. Fas is involved in developmental cell death and will be considered in detail here (for a recent review on Fas-mediated apoptosis see Nagata and Golstein, 1995). Monoclonal antibodies raised against a human cell surface marker known variously as Apo-1 or Fas were shown to kill cell lines carrying this marker (Trauth et al, 1989; Yonehara et al., 1989). The antigen was later shown to be a 45 kDa type 1 membrane protein (Itoh et al., 1991; Oehm et al., 1992) consisting of a 319 amino acid polypeptide chain which belongs to the TNF/NGF receptor family. It has an overall 28% similarity to the 55 kDa TNF-1 receptor, extending over an 80 residue sequence in the intracellular domain of the TNF-IR known as the intracellular *death domain' (Itoh and Nagata, 1993; Tartaglia et al., 1993). The FAS/APO-1 ligand is a glycosylated 40 kDa type

Mechanisms of developmental cell death

II transmembrane protein member of the TNF family (Suda et al, 1993) with an extracellular domain similar to those of TNF-a and TNF-y8. Fas mediates apoptosis when cross-linked with either anti-Fas antibody or Fas ligand and, as we shall see below, plays a role in apoptosis of immature thymocytes and peripheral T cells, CTL cytotoxicity and AICD (reviewed in Kisielow and von Boehmer, 1995; Nagata and Golstein, 1995). The role of Fas in T cell development has been addressed by the use of Ipr and gld strains of mice which have loss of function mutations in the genes coding for Fas and Fas ligand respectively. Fas is expressed in high levels on immature CD4+CD8+ thymocytes but does not appear to play an indispensible role in either positive or negative selection, since these processes are normal in the Ipr strain of mice (Sidman et al, 1992; Herron et al., 1993). These mice, however, show defective peripheral deletion of T cells and AICD (Russell and Wang 1993; Russell et al, 1993; Alderson et al, 1995). The mature T-lymphocyte expresses Fas and,when activated by antigen. Fas ligand (Trauth et al, 1989; Owen-Schaub et al., 1992; Suda et al., 1995; Tanaka et al., 1995; Vignaux et al., 1995). However, after antigen activation, Fas/FasL induced apoptosis, or AICD, occurs, in a cell autologous manner, only after 3-4 days (Owen-Schaub et al., 1992; Brunner et al., 1995; Dhein et al., 1995; Ju et al., 1995). It is not clear what prevents the immediate induction of apoptosis in this system although it has been proposed that a mechanism may exist to prevent Fas clustering (Boldin et al., 1995). Fas-mediated apoptosis also plays an important role in the mechanism of CTL killing of target cells which express Fas. In this process, cell-cell contact is required to allow the binding of Fas ligand expressed on the CTL to Fas on the surface of the target cell in order to trigger the cell death pathway in the latter (Rouvier et al., 1993; Kagi et al., 1994; Lowin et al., 1994). In cultured mouse hepatocytes, anti-Fas antibody induces apoptosis in cells only in the presence of inhibitors of protein and RNA synthesis or protein kinase C (see Section 4.2) suggesting the

95

A. Messina and A. Jaworowski

presence of proteins which supress the cell death signal of Fas (Ni et al., 1994). Although the mechanism by which Fas mediates cell death is not delineated, several things are known: (1) the pathway is independent of extracellular calcium (Rouvier et al., 1993; Ju et al., 1994) and reactive oxygen intermediates (Hug et al., 1994), (2) it occurs even in the absence of a nucleus (Schulze Osthoff et al., 1994), (3) it is only partially if at all blocked by overexpression of Bcl2 (Itoh et al., 1993; Chiu et al., 1995) but completely blocked by co-expression of Bcl-2 and Bag-1 (Takayama et al., 1995) and (4) it involves an ICE-like cysteine proteinase (Enarl et al., 1995; Los et al., 1995; Tewari and Dixit, 1995). 4. Signal transduction pathways involved in apoptosis Reasearch on the signal transduction pathways activated following the initial death signal is still in its infancy. Calcium, cAMP, PKA, PKC and, recently, PI-3 kinase and ceramide have all been implicated using various experimental models of apoptosis. These factors may induce various immediate early genes, which presumably activate genes involved in specific downstream apoptotic events. The picture emerging is one of a multitude of signalling events that can be activated in response to the initial stimulus, depending upon the cell type and the stimulus involved. 4.7. Calcium Ionized calcium plays a role in many early signalling events and was one of the early second messengers investigated with respect to apoptosis. Mammalian cells possess two intracellular pools of Ca^"^; a high-affinity, low-capacity endoplasmic reticulum (ER) store which can be discharged by thapsigargin and a low-affinity, high-capacity mitochondrial store which can be released by other agents. Increases in intracellular calcium concentration can be a result of redistribution of intracellular stores or influx from extracellular sources, although it should be noted that extracellular influx of Ca^"^ may also cause redistribution of intracellu-

lar stores (Trump and Berezesky, 1992). Both Ca^'^-dependent and Ca^"*"-independent mechanisms of apoptosis have been documented, the former case involving redistribution of both intracellular and extracellular Ca^"^ stores. In thymocytes, an increase in cytosolic free Ca^"*" preceeds apoptosis induced by calcium ionophores, glucocorticoids and the TCR (reviewed in McConkey and Orrenius, 1991) and is associated with endonuclease activation (see Section 9.1). Both the increase in [Ca^'^lj and endonuclease activation are prevented by protein and RNA synthesis inhibitors. The increase in [Ca^"*"]} however, does not appear to be sufficient to induce either apoptosis or endonuclease activity in this system, since phorbol esters, IL1 and bcl-2 prevent apoptosis and decrease endonuclease activity without affecting [Ca^"*"]i. In mature T-cells, TCR activation and calcium ionophores also increase [Ca^+]i but do not induce apoptosis (Wyllie et al., 1984; McConkey and Orrenius, 1991). In primary neuron cultures, apoptosis induced by neurotrophin withdrawal is associated with a decrease in [Ca^+Jj (Eichler et al., 1992, 1994). In addition, membrane depolarisation induced by KCl and thapsigargin treatment, which results in elevated [Ca^"^]i, protects neurons and neuronal PC 12 from apoptosis (Edwards et al., 1991; Rukenstein et al., 1991; Franklin et al., 1995; Lampe et al., 1995). In the case of Fasmediated apoptosis, evidence suggests that it is independent of Ca^^ (Rouvier et al., 1993). The above evidence suggests that changes in cytosolic free Ca^+ is not indispensible for apoptosis, but may play a role in activating Ca^"*"dependent enzymes associated with downstream events of apoptosis such as nuclear degradation and cytoplasmic packaging.(see Sections 9.1 and 9.4). 4.2. Protein kinases Cyclic AMP (cAMP) has been linked to apoptosis in a number of models but, as is the case with cell proliferation, it has opposite effects in different cell systems. Treatment of thymocytes with prostaglandin E2, and other agents that increase cAMP levels, has been shown to induce apoptosis

96

via activation of cAMP-dependent protein kinase A (McConkey and Orrenius 1991; Mastino et al., 1992). Conversely it has been reported that increases in cAMP levels induced by prostaglandin E2 protect CD4+CD8+ thymocyte-like cells from anti-Fas- and anti-TCR-mediated apoptosis (Goetzl et al., 1995) and AICD (Lee et al., 1993), and also that cAMP prevents apoptosis of neurons and delays apoptosis in neuronal PC 12 cells deprived of NGF (Rydel and Greene, 1988; Rukenstein et al., 1991). Agents such as phorbol esters or IL-1, that increase protein kinase C (PKC) activity either directly or via induction of its activator diacylglycerol (DAG), prevent apoptosis in thymocytes (McConkey and Orrenius, 1991; Batistatou and Green, 1993) but phorbol esters do not support long-term survival of PC 12 cells or rat sympathetic neurons (Rydel and Greene, 1988; Rukenstein et al., 1991): this may be attributable to the eventual down-regulation of PKC by prolonged phorbol ester treatment in these models (Batistatou and Green, 1993). PKC activation has been implicated in AICD and anti-Fas mediated apoptosis of hepatocytes (Jin et al., 1992; Ni et al., 1994). Furthermore, activation of PKC may have a role in ceramide-induced apoptosis (see next Section). 43. Ceramide Recently, evidence has accumulated to implicate ceramide (A^-acyl sphingosine) as a second messenger involved in apoptosis in a number of systems (Obeid et al., 1993; Jarvis et al., 1994). Interest in this compound derived initially from the finding that it might be a second messenger induced following stimulation of cells with TNFa. Subsequently it was shown that the cell-permeant analogue C2-ceramide, but not a closely related analogue dihydroceramide, induces apoptosis in the U937 monocytoid leukemic cell line (Obeid et al., 1993). Treatment of both haematopoietic and non-haematopoietic cells with exogenously added ceramide analogues or sphingomyelinase induces Gl arrest and apoptosis (Jarvis et al., 1994a,b; Jayadev et al., 1995; Ji et al., 1995) whereas phosphatidylcholine (the other product of sphingomyelin hydrolysis by sphingomyelinase) does not (Jarvis

Mechanisms of developmental cell death

et al., 1994). TNF-, IL-1- and Fas-induced apoptosis all share components of this signalling pathway (Cifone et al., 1994; Schutze et al., 1994). In these cases, a membrane-associated, Mg^"*"independent neutral sphingomyelinase is activated, which catalyses the hydrolysis of sphingomyelin to produce ceramide and phosphorylcholine. This receptor-activated sphingomyelinase is distinct from cytosolic and Mg^^-dependent enzymes which may also be present in the cell (Okazaki et al., 1994; Hannun and Obeid, 1995). The involvement of these other shingomyelinases in apoptosis is unclear; the acidic, Mg^+-dependent enzyme has been shown to be activated by, on the one hand, DAG (Schutze et al., 1994), which does not induce apoptosis in many cells, and, on the other hand, by Fas (Cifone et al., 1994). Several experimental findings indicate that there may be cross-talk between the sphingosine cycle and the classical diacylglycerol pathway. Ceramide-induced apoptosis was inhibited by exposure of cells to either phospholipase C or activators of PKC (Jarvis et al., 1994). In leukemic cells, apoptosis induced by serum deprivation was associated with an increased intracellular content of both ceramide and diacylglycerol and was inhibited by exogenously added diacylglycerol (Jayadev et al., 1995). Sphingosine, which is a metabolic breakdown product of ceramide, has also been shown to inhibit PKC activity and to induce apoptosis (Ohta etal., 1994). The downstream targets of ceramide that are involved in apoptosis remain to be identified. A few potential targets have, however, been described. In A431 cells, Davis and co-workers showed that sphingomyelin induced phosphorylation of the epidermal growth factor receptor at a specific site, Thr 669 (Davis et al., 1988; Faucher et al., 1988; Countaway et al., 1989). This led to the description of a membrane-bound, Ca^'^-independent protein kinase that is stimulated by ceramide (Mathias et al., 1991). Conversely, Hannun and co-workers have demonstrated that ceramide activates a member of the the heterotrimeric protein phosphatase 2A family in rat T9 glioma cells and have coined the term ceramide-activated protein phosphatase (CAPP) for this activity (Dobrowsky and Hannun,

A. Messina and A. Jaworowski

1992; Dobrowsky et al., 1993). Evidence for the involvement of CAPP in apoptosis comes from the fact that okadaic acid, a potent inhibitor of CAPP, also blocks both ceramide-induced c-myc downregulation (see Section 5.2) and apoptosis in these cells. One downstream effector for ceramide-induced apoptosis may be the retinoblastoma protein (Rb). Serum withdrawal induced cell cycle arrest and dephosphorylation of Rb in MOLT-4 cells associated with increased levels of endogenous ceramide, whereas sphingosine levels were unaffected. Addition of exogenous ceramide also induced dephosphorylation of Rb in these cells. Ceramide did not induce cell cycle arrest in cells that had defective or inactive Rb (Dbaibo et al., 1995). These data are consistent with the current models of Rb action, in which hypophosphorylation of Rb protein correlates with its ability to bind to transcription factors such as those of of the E2F family and to repress progression through the Gl phase of the cell cycle (Ewen, 1994). 5. Immediate early genes and other transcription factors involved in apoptosis Immediate early genes are defined as genes which are rapidly expressed following stimulation of cells by, for example, growth factors and whose expression is independent of protein synthesis. This implies that expression of these genes is controlled directly via the activation of pre-existing transcription factors. Many immediate early genes encode transcription factors which, in turn, stimulate the cell response by activating other genes. 5.7. Fos andJun Examples of immediate early genes include the genes encoding the transcription factors Fos and Jun. Unlike the transient induction seen following growth factor stimulation, a delayed but sustained increase in c-Fos and c-Jun has been associated with apoptosis. A sustained elevation of Fos was reported to occur in some foetal cells, including neurons, that were presumed to be destined to die, although this elevation was not directly shown to

97

be associated with cells which had an apoptotic morphology (Smeyne et al, 1993). Withdrawal of NGF in primary cultures of neurons induced c-jun and c-myb and, at a later time, c-fos mRNA (Estus et al, 1994). The expression of c-jun was sustained, whereas that of c-fos was transient, occurred around the time of chromatin condensation and was shown by in situ hybridisation to be restricted to neurons exhibiting apoptotic morphology. In this study, NGF deprivation and protein synthesis inhibition had an additive effect on c-jun expression, but blocked c-fos induction. Neutralizing antibodies to c-Jun, c-Fos ,FosB, Fra1 and Fra-2, but not to Jun B or Jun D, protected the neurons from apoptosis. In lymphocytes stimulated to undergo apoptosis by withdrawal of IL-2 or IL-6, c-fos and c-jun mRNAs were transiently elevated (Colotta et al, 1992). In this study, evidence was obtained that the induction of fos and/or jun is necessary for apoptosis to occur since co-injections of Fos and Jun mRNA antisensense oligonucleotides were shown to reduce apoptosis. In the IL-2-dependent CTLL T-cell-line, c-jun is expressed constitutively and the cells undergo apoptosis when deprived of IL-2. Cells can be protected by the addition of either IL-2 or phorbol ester, treatments which both induce c-fos expression and DNA binding activity to API sites in target promoters. The formation of a functional API transcription factor correlates with the repression of apoptosis induced by factor deprivation (Walker et al, 1993). The protection by phorbol ester and IL-2 against apoptosis induced by factor-deprivation also correlates with the ability of both treatments to induce PKC. In this system, the factor-deprived cells are also sensitive to glucocorticoid-induced apoptosis, but rescue from this form of cell death is obtained only by IL-2 treatment. This has been suggested to correlate with the ability of IL-2, but not phorbol ester, to induce tyrosine kinase activity in CTLL cells (Walker etal., 1993). These studies suggest a role for Fos and/or Jun in apoptosis in cell culture. Induction of c-fos and c-myc however, has been shown not to correlate with apoptosis in NGF-deprived embryonic rat sympathetic neurons in vitro (Martin et al., 1992).

98

with apoptosis in NGF-deprived embryonic rat sympathetic neurons in vitro (Martin et al., 1992). In addition, studies with c-fos null mice showed that in the majority of cells apoptosis proceeds normally (Johnson et al., 1992; Wang et al., 1992; Jain et al., 1994). C-jun knockout mice die at 11.5 days of gestation and therefore it is not known whether apoptosis proceed normally in the absence of c-Jun in vivo (Hilberg et al., 1993; Johnson et al., 1993). Using a technique which allows the study of the role of c-jun specifically in lymphocytes, mice generated by a RAG-2 deficient blastocyst complementation method (Chen et al., 1993) have thymuses which are small and have a drastically reduced number of thymocytes but have a normal T-cell subset composition and normal peripheral T-cell function (Chen et al., 1994). Recently, we have shown that c-Jun but not c-Fos protein is expressed in rat sympathetic neurons undergoing normal developmental cell death in vivo. C-Jun was expressed exclusively in apoptotic neurons that had proceeded to the stage of chromatin condensation (Messina, 1994). From the above discussion, it appears that c-Fos is not indispensible in developmental cell death although its transient induction correlates with apoptosis in a number of experimental systems. On the other hand, c-Jun is essential during development which may involve a role in developmental cell death. Although such a role has yet to be proven, sustained induction of c-Jun correlates with apoptosis in a number of experimental systems. Another early gene, nur77 (Winoto, 1994), is induced in apoptotic thymocytes and apoptotic T-cell hybridomas following TCR activation, but not in proliferating T-cells. Inhibition of nur77 induction by expression of dominant negative Nur77 or antisense oligonucleotides to nur77 prevented TCR, but not glucocorticoid-induced apoptosis, in these cells (Liu et al., 1994; Woronicz et al., 1994), suggesting that it may play a role in the former. Inhibitors of TCR-induced apoptosis and AICD such as cyclosporin A downregulate Nur77 DNA-binding activity (Shi et al., 1989; Yazdanbakhsh et al., 1995).

Mechanisms of developmental cell death

5.2. C-Myc C-Myc is a transcription factor which has been shown to be required for cell cycle progression but is also a potent activator of apoptosis (Bissonnette et al., 1992; Fanidi et al., 1992). In immortalised fibroblasts, up-regulation of c-myc by addition of serum growth factors is accompanied by cell proliferation, while down-regulation of c-myc by serum growth factor removal prevents division but leaves the cells viable. Conversely, if the cells are engineered to express c-Myc constitutively, they eventually undergo apoptosis (Evan et al., 1992). Thus, myc induction can lead either to proliferation or to apoptosis, depending on the circumstances. C-Myc binds to a specific DNA sequence and regulates gene transcription as a heterodimer with other transcription factors including Max (Amati et al., 1992; Littlewood et al., 1992). The Max-Max homodimer counteracts c-Myc function by competing for the same DNA-binding motifs in target promoters without the ability to activate transcription (Amati et al., 1992). Thus, Myc function may be controlled by the relative levels of Myc and Max within the cell. Interestingly, the Myc-Max heterodimer induces both proliferation and apoptosis in non-transformed fibroblasts (Amati et al., 1993), indicating that other signals must decide the eventual fate of the cell. It has not yet been established whether c-Myc is involved in developmental cell death in vivo. Circumstantial evidence for c-Myc involvement comes from the following observations: (1) c-myc mRNA is increased following activation of the TCR in immature thymocytes (Riegel et al, 1990), (2) c-myc antisense ohgonucleotides inhibit TCR-induced apoptosis in T cell hybridomas (Green et al., 1992; Shi et al., 1992) and (3) Myc-Max heterodimers are necessary for AICD in these cells (Bissonnette et al, 1994). However, in NGF deprivation-induced apoptosis of neurons, c-myc expression is not correlated with apoptosis (Martin et al., 1992). 6. BCL-2 (B cell lymphoma/leukaemia 2 gene) As mentioned in Section 2.1, bcl-2 is the mammalian homologue of the ced-9 anti-apoptotic gene of

A. Messina and A. Jaworowski

(Vaux et al., 1992). It was first demonstrated as a gene activated following chromosomal translocation in the majority of follicular non-Hodgkins lymphomas (Tsujimoto et a l , 1985). In 1988, Vaux and co-workers showed that stable transfer of bcl-2 into IL-3-dependent pre-B cells inhibited apoptosis and permitted prolonged survival, without proliferation (Vaux et al., 1988). This finding was confirmed in bcl-2 transgenic mice (Hockenbery et al., 1990). In the adult, bcl-2 expression is restricted to germinal centres in B cell follicles, surviving T cells in the thymic medulla, renewing stem cell-derived hematopoietic cells and long-lived postmitotic neurons (Hockenbery et al., 1991; Merry etal., 1994). In bcl-2 knockouts, excessive apoptosis occurs in the thymus and kidney but other tissues, including the nervous system, develop normally (Vies et al., 1993b). Bcl-2 protein has been locaUsed to the inner membrane of mitochondria, the nuclear envelope and the endoplasmic reticulum (Hockenberry et al., 1990; Jacobson et al., 1993; Krajewski et al., 1993) and this membrane localisation is presumed to be important for its function (Tanaka et al., 1993). However, a Bcl-2 construct in which the Cterminal 33 residues were deleted, and which is located predominantly in the soluble fraction of the cell, was still able to protect cells from cell death (Borner et al., 1994). 6.1. Bcl-2 involvement in models of developmental cell death Bcl-2 expression prevents or delays apoptosis under a variety of circumstances. These include (a) withdrawal of the neurotrophins NGF, BDNF, and NT-3 from primary neuron cultures and from PC12 cells (Garcia et al., 1992; Allsopp et al., 1993; Mah et al., 1993), (b) induction of apoptosis in thymocytes by glucocorticoids, irradiation, anti CD3-Ab (anti-TCR ab) and Ca^^ ionophores (Sentman et al., 1991; Siegal et al., 1992), (c) some models of AICD (Strasser et al., 1991; Siegal et al., 1992) and (d) withdrawal of growth factor from IL-3-and IL-4-or GM-CSF-dependent haematopoietic cell lines (Vaux et al., 1988; Nunez et al., 1990).

99

Bcl-2 does not rescue cells from apoptosis in all cases, however, and a number of Bcl-2-insensitive mechanisms of apoptosis have been documented. These include (a) some cases of CTL- and TNFmediated apoptosis, (b) withdrawal of IL-2 and IL6 from cultures of their respective dependent cell lines (Vaux et al., 1988, 1992; Nunez et al.,1990), (c) withdrawal of CNTF from ciliary neurons (Allsopp et al., 1993), (d) negative selection of thymocytes (Strasser et al., 1991; Sentman et al., 1991) and (e) Fas-mediated apoptosis in T cells (Itoh et al., 1993 ). In some of these cases, however, it is possible that some of the recently identified homologues of Bcl-2 may be involved (see Section 6.2). Cellular Bcl-2 levels have been shown to change during the development of T cells such that they decrease at a point in development when the cells are undergoing apoptosis. In thymocytes, Bcl-2 is present in many of the cells that are CD4" /CD8- but in only 5-10% of CD4VCD8-^ cells and this coincides with the stage when negative selection of the CD4VCD8+ cells takes place (Veis et al., 1993a; Andjelic et al., 1994). In addition, Bcl-2 is upregulated in all cells that survive negative selection, i.e. the mature CD8VCD4- and CD8/CD4+ cells (Gratiot Deans et al., 1993; Veis et al., 1993a). As discussed in Section 3.2, activationinduced cell death in peripheral T cells is mediated by Fas but occurs only after a delay of several days, during which time endogenous levels of Bcl2 decrease. Again, this correlation is consistent with an involvement of Bcl-2 in protecting T cells from apoptosis but, as Bcl-2 alone cannot prevent Fas-mediated apoptosis, other factors must also be implicated. A variety of evidence indicates that the protective action of Bcl-2 does not involve interaction with the effects of c-Myc. Thus, Bcl-2 expression blocks apoptosis in non-transformed fibroblasts without blocking their entry into the cell cycle (Bissonnette et al., 1992; Fanidi et al., 1992; Wagner et al., 1993) and myeloid cells transfected with bcl-2 down-regulate c-myc when deprived of trophic support. These cells undergo apoptosis directly from GQ when Bcl-2 levels decrease in a

100

directly from GQ when Bcl-2 levels decrease in a mechanism that is independent of c-Myc (Vaux andWiessman, 1993). 6.2. Bcl-2-related proteins involved in Bcl-2 action It is now known that Bcl-2 is one of a family of similar proteins which are characterised by the presence of 2 conserved domains termed the Bclhomology 1 (BHl) and 2 (BH2) domains respectively (reviewed in Williams and Smith, 1993). Boise and colleagues (1993) isolated the bcl-x gene that encodes two functionally different proteins via differential mRNA splicing. The long form, BC1-XL, shares 74% homology with Bcl-2 and forms heterodimers with Bcl-2 to promote cell survival. BC1-XL is found in long-lived cells such as neurons. The short form, Bcl-Xs, inhibits Bcl-2 function but does not bind to it and is found in short-lived cells such as those present in the immune system (Boise et al., 1993; Sato et al., 1994). Bcl-Xs blocks Bcl-2-mediated rescue of neurotrophin-deprived neurons but does not accelerate cell death (Martinou et al., 1995). Bax is another member of this family of proteins that forms both homodimers and heterodimers with Bcl-2. When overexpressed, Bax has been shown to accelerate cell death induced by growth factor deprivation in an IL-3-dependent cell line (Oltvai et al., 1993). However, overexpression of Bax is not sufficient to trigger apoptosis (Korsmeyer et al., 1993). Expression of Bax is more widespread than that of Bcl-2 and appears to be associated with cells that have a high rate of spontaneous or inducible apoptosis (Krajewski et al, 1994). It has been shown that Bax interacts with Bcl-2 via the BHl and BH2 domains. Evidence that the protection from apoptosis afforded by Bcl-2 is mediated via its heterodimers with Bax rather than Bcl-2 homodimers was obtained from experiments involving mutations of Bcl-2 that prevent the former, but not the latter (Yin et al., 1994). This has led to the model that the ratio of Bcl-2 to Bax in a cell may determine the sensitivity of that cell to apoptotic signals and thus determine whether these signals are ignored or executed

Mechanisms of developmental cell death

(Oltvai and Korsmeyer, 1994). For example, if Bcl-2 is in excess, this would lead to all of the Bax present in the cell being involved in heterodimer formation, and there would be sufficient Bcl-2 to form homodimers to promote cell survival. If, conversely, Bax is in excess, then there is sufficient to form Bax homodimers which mediate apoptosis (Oltvai and Korsmeyer, 1994). The involvement of Bcl-2-like proteins in developmental cell death has been extended to other model systems. In granulosa cells of ovarian follicles, apoptosis has been shown to be associated with an increase in bax mRNA and a decrease in bcl'Xi^ mRNA, but without change in bcl-l expression (Tilly et al., 1995), consistent with the presumed involvement of these proteins in the above-mentioned models. Other members of the Bcl-2 family have been described. Mcl-l and Al were isolated by differential screening of leukemic and normal cells. Mcl1 encodes a 37 kDa protein with 35% homology with Bcl-2 (Kozopas et al., 1993) and has been shown in Chinese hamster ovary cells to decrease apoptosis induced by enforced c-Myc overexpression (Reynolds et al., 1994). Al encodes a 20 kDa protein that has 40% homology with Bcl-2 (Lin et al., 1993) but as yet little is known about either of these two proteins. Bad is another Bcl-2 related protein that has recently been identified following yeast two-hybrid screening for proteins that interact with Bcl-2 (Yang et al., 1995). It binds to BclXL and, with weaker affinity, to Bcl-2, but not with Bax, Bcl-Xs, Mcl-l, Al or itself. Bad appears to promote apoptosis by binding to BC1-XL and displacing Bax (Yang et al., 1995). It reverses the repression of apoptosis by BC1-XL , but not Bcl-2, consistent with its higher affinity for the former protein. The most recently described Bcl-2 homolgue is Bak. Bak does not bind to Bcl-2, Bcl-x^, Bax or itself but does form heterodimers with BclXL- Bak tends to be present in tissues composed of long-lived cells, such as post-mitotic neurons, and there is some evidence that it may also be attached to membranes (Chittenden et al., 1995; Farrow et al., 1995; Kiefer et al., 1995). Overexpression of Bak in NGF-deprived sympathetic neurons accelerated apoptosis and blocked the protective actions

A. Messina and A. Jaworowski

of both ElB 19k (see Section 8) and Bcl-2 (Farrow etal., 1995). 6.3. Non-Bcl-2-related proteins capable of interacting with Bcl-2 In addition to interacting with Bcl-2 homologues, Bcl-2 interacts with structurally unrelated proteins. Bag-1 has recently been cloned following screening for proteins that bind to Bcl-2 (Takayama et al., 1995). When co-transfected with Bcl-2, it has been shown to protect Jurkat cells from apoptosis induced by a variety of stimuli, including stimulation with anti-Fas antibody, and apoptosis induced by CTL (Takayama et al., 1995). Bcl-2 binds R-ras (a monomeric G protein related to the H and Ki ras proto-oncogenes) and can be co-immunoprecipitated with this protein from extracts of various mammalian cells (FemandezSarabia and Bischoff, 1993; Wang et al., 1994a). In addition, Bcl-2 has been reported to be coimmunoprecipitated with Raf-l,a protein serine/threonine kinase activated by Ras and upstream of the MAP kinase pathway in many cell types (Wang et al., 1994a). These workers also reported that, when co-transfected into 32D.3 haematopoietic cells, Raf-1 synergises with Bcl-2 in protecting the cells from apoptosis induced by growth factor withdrawal. 6.4. Bcl-2 and reactive oxygen species There has recently been a large body of opinion that reactive oxygen species have a role to play in the mechanism of apoptosis. When neuron and lymphocyte cultures are exposed to reactive oxygen species, they undergo apoptosis (Hockenbery et al., 1993; Forrest et al., 1994; Whittemore et al., 1994) which can be prevented by treatment with antioxidants, exogenous catalase treatment and Bcl-2 (Forrest et al., 1994). A role for reactive oxygen species in apoptosis is further suggested by the finding that antisense oligonucleotides specific for superoxide dismutase induce apoptosis in neuronal PC 12 cells cultured either in the presence or the absence of NGF which can be prevented by

101

addition of anti-oxidants (Troy and Shelanski, 1994). Bcl-2 expression in neuronal cells, has been shown to prevent the rise in reactive oxygen species which occurs following the induction of apoptosis and has been suggested to protect cell membranes from oxidative injury by inhibiting the accumulation of lipid peroxides (Hockenbery et al., 1993; Kane et al., 1993). Given its intracellular location on the inner mitochondrial membrane, Bcl-2 is well placed to perform this role. Cu/Zndependent SOD delays apoptosis in trophic factordeprived neurons if injected into the cytoplasm or if overexpressed (Greenlund et al., 1995), further supporting a role for reactive oxygen species in neuronal apoptosis, and leading to the suggestion that reactive oxygen species act as signal transducers rather than as toxic agents. Recently, however, the claims that apoptosis involves the generation of reactive oxygen species and that Bcl-2 protects cells by inhibiting this generation have been disputed in a series of papers in which apoptosis was induced by hypoxia, i.e. under conditions where generation of reactive oxygen species is drastically reduced. In these experiments it was shown that Bcl-2 or BC1-XL prevent hypoxia-induced apoptosis in the absence of reactive oxygen species, lipid peroxidation or DNA damage (Jacobson and Raff 1995; Shimizu et al., 1995). Under these conditions apoptosis was not affected by either reactive oxygen species scavengers or inhibitors of reactive oxygen species scavengers. Furthermore, studies of apoptosis in a fibroblast cell line lacking intact mitochondia suggest that neither mitochondrial respiration nor oxidative phosphorylation are required for the induction of apoptosis or for Bcl-2-mediated protection (Jacobson et al., 1993). In this context, Bcl-2 overproduction does not affect intracellular ATP levels or O2 consumption in PCI2 cells (Mah et al., 1993). 6.5. BCL-2 in cell free systems A cell-free model of apoptosis has recently been developed (Newmeyer et al., 1994; see also Section 7.2) which offers prospects of dissecting the

102

biochemical pathways involved in nuclear condensation and fragmentation. Xenopus oocytes were obtained from eggs laid 14-28 days after treatment of frogs with pregnant mare serum gonadotrophin. When nuclei were incubated for 2-4 h with nuclear assembly extracts prepared from these oocytes, changes reminiscent of apoptosis were observed consistently (Newmeyer et al., 1994). The morphological changes in added nuclei were prevented by the addition of baculovirus-expressed Bcl-2 protein, but only when added during the during the latent phase prior to the observed morphological changes. Interestingly, the *apoptotic activity' of the nuclear assembly extracts had an absolute requirement for a subcellular fraction highly enriched in mitochondria. The system was used to test the effect of various treatments on apoptosis and it was found that the observed nuclear changes were inhibited by the addition of, amongst other things, inhibitors of calpain (a cysteine protease; see Section 7). 7. ICE: Interleukin ip converting enzyme As noted in Section 2.1, ced-3 is one of two genes identified in C. elegans essential for programmed cell death. Cloning and sequencing of ced-3 showed that the protein encoded by this gene is similar to a mammalian enzyme known as Interleukin 1^8 converting enzyme (ICE). ICE was initially identified as an enzyme responsible for cleaving the 31 kDa membrane-bound pro-interleukin-l)3 peptide into the 17 kDa biologically active cytokine involved in inflammatory responses. ICE is a cysteine protease that is formed by cleavage of an inactive 45 kDa pro-enzyme into two subunits of 20 and 10 kDa. The functional protein exists as a tetrameric cytoplasmic enzyme with the structure (p20)2:(pl0)2, in which both subunits contribute residues to the active site (Walker et al., 1994; Wilson et al., 1994). Although JL-ip mRNA is found predominantly in peripheral blood moncytes, ICE transcripts are detected in cells that do not make TL-l^ which has led to the supposition that other substrates for its protease activity exist (di Giovine and Duff, 1990; Dinarello, 1991; Cerreti et al., 1992) and that it there-

Mechanisms of developmental cell death

fore may have a role to play in other physiological processes. Recently, genes encoding several ICE-related proteases have been identified in mammalian tissues. These include the mouse gene Nedd-S (Kumar et al., 1992) and its human homologue, /c/i-7(Wang et al., 1994b) and a gene cloned from a human T cell line, CPP32 (Femandes-Alnemri et al., 1994). In addition, a *protein resembling ICE', prICE, has recently been identified in extracts prepared from chicken DU249 cells (Lazebnik et al., 1994). The structural relationships between the products of these genes and of ced-3 are discussed in a recent review (Kumar, 1995). All of the proteases for which the gene sequence is available share a region of homology, QACRG, which contains the catalytic cysteine residue (cys285 in the human ICE sequence). The genes for Nedd-2/Ich-l encode two protein products which are generated via alternative splicing of the mRNA. The products of the human gene have been called Ich-lL (the 51 kDa long form), and Ich-ls (the 39 kDa short form). Nedd-2 is expressed during embryonic development of mouse brain and kidney and is down-regulated in adults. The homology of ICE and ICE-related proteins to the protein product of ced-3 have led to studies of their potential involvement in apoptosis. Overexpression of Ced-3, ICE, Nedd-2, or Ich-1 were found to induce apoptosis in various cells such as Rat-1 cells and NGF-dependent neurons (Kumar et al., 1992, 1994; Miura et al., 1993; Gagliardini et al, 1994; Wang et al., 1994b), though not all cell types were equally sensitive (Kumar et al., 1994; Wang et al., 1994b). Apoptosis induced by ICE or Ced-3 could be prevented by Bcl-2 or the viral protein Crm A (see Sections 6 and 8) implying that they induce apoptosis via a common mechanism. Similar results have been obtained for Nedd-3 and Ich-1 (Kumar et al., 1994; Wang et al., 1994b). Furthermore, mutation of the catalytic cysteine residue completely abolished the ability of ICE to induce apoptosis, showing that aspartate-directed proteolysis is essential for this function (see Wilson etal, 1994). Although overexpression of ICE in various cell lines can induce apoptosis, the involvement of this

A. Messina and A, Jaworowski

protein in apoptosis in vivo is not clear. Initial studies using ICE knock-out mice indicates that there are no major defects in developmental cell death and that apoptosis occurs normally in both macrophages and thymocytes, suggesting that this protein does not have an essential and unique role in developmental cell death (Li et al., 1995). As expected from its role in catalysing DL-l maturation, ICE knockout mice have a defect in the production of mature IL-lj3 and resistence to endotoxic shock, as well as an unexpected defect in the production of IL-la (Li et al., 1995). The existence of both long and short forms of Ich-1, generated by alternative splicing, is reminiscent of the case of Bcl-x (see Section 6.2). Like Bcl-x, the long and the short form of Ich-1 appear to have opposing effects with respect to apoptosis. Overexpression of Ich-1 L induces apoptosis while overexpression of Ich-ls promotes survival in rat-1 fibroblasts cultured under conditions of serum deprivation (Wang et al., 1994b). The fact that Bcl-2 can protect cells from apoptosis induced by ICE-like proteins has led to a model in which there are 2 checkpoints in the apoptotic pathway, governed by the relative levels of both positive and negative signals (see Oltvai and Korsemeyer, 1994). The upstream checkpoint is governed by the relative influence of, for example, Bcl-2.Bax heterodimers versus Bax.Bax homodimers, while the second checkpoint is governed by the relative influence of signals from ICE-like proteases, for example, Ich-1 L homodimers versus Ich-1 L Ich-ls heterodimers. These checkpoints would regulate passage through a pathway which links the initial signals to undergo apoptosis (for example, those reviewed above in Sections 4 and 5) to the committed downstream events leading to the irreversible morphological changes characteristic of apoptosis (see Section 9). 7.1. An ICE-like activity involved in CTLmediated apoptosis Aspartate-directed proteases are a family of at least seven similar enzymes, including Granzyme B (also known as Fragmentin 2 or CPPl), which

103

have been shown to be present in granules in cytotoxic T cells and which are injected into target cells via the perforin holes involved in CTLmediated killing. These enzymes cleave after an aspartyl residue in the same context as that found in substrates for ICE. The optimum recognition sequence is YVAD (Shi et al., 1992; Thornberry et al., 1992). Granzyme B is not, however an ICE homologue, or even a cysteine protease, but rather a serine protease. Although the site specificity is similar to that for ICE, they apparently cannot cleave the similar sequences found in ICE itself, whose cleavage is necessary for ICE activation (Darmon et al., 1994). This indicates that the mechanism by which granzymes induce apoptosis is not via activation of ICE. Again, unlike the situation with ICE-mediated apoptosis, neither Bcl-2 nor viral genes offer protection from CTLmediated apoptosis (Vaux et al., 1992). 7.2. ICE in the cell-free system Lazebnik and co-workers developed a cell-free system derived from chicken hepatoma cell extracts, for studying nuclear changes in apoptosis (Lazebnik et al., 1993). Extracts displayed apoptotic activity only when derived from cells that were blocked in M phase by nocodazole after being initially arrested in S-phase by aphidicolin. In this system, a number of nuclear proteins were cleaved during apoptosis including an enzyme involved in DNA repair, poly-(ADP-ribose) polymerase (FAR?). FARF was shown to be cleaved at a sequence identical to the cleavage site of IL1^, and by an activity that was sensitive to inhibitors of ICE, implying that the protease is a member of the ICE family. The protease activity was named prICE (see Section 7) and was shown to be distinct from ICE itself, since exogenously added IL-1)8 was not cleaved by these extracts (Lazebnik etal, 1993). It has not been established whether prICE is the chicken homologue of ICE or another ICE family member. The potential role of FARF in apoptosis is intriguing. It specifically recognises, and binds to, single-stranded breaks in DNA via zinc fingerlike domains (Menissier de Murcia et al., 1989).

104

Binding to single-stranded, nicked DNA activates the enzyme to catalyze transfer of the ADP-ribosyl moiety from NAD to acceptor proteins, forming branched-chains of poly-(ADP-ribose). The acceptor proteins include PARP itself and proteins involved in chromatin architecture and DNA metabolism (for a review of PARP see de Murcia and Menissier de Murcia, 1994). The significance of poly-ADP-ribosylation in DNA repair is not yet fully understood; however, it is thought to result in a decreased affinity of chromatin proteins for DNA, allowing access of the damaged DNA strand to the repair machinery. PARP may therefore be an integral part of DNA repair. The connection between its role in DNA repair and its potential role in apoptosis is unclear. PARP synthesis is induced during cell death and both positive and negative roles for the enzyme in apoptosis have been put forward. PARP has been hypothesised to ADP ribosylate the endonuclease responsible for apoptotic DNA degradation, resulting in its activation (Jones et al., 1989) and, in Fas-stimulated U937 and B104 cells, inhibitors of PARP in combination with Zn^+ have been shown to inhibit apoptosis (Sumimoto et al., 1994). However, in other situations, PARPmediated endonuclease ribosylation may inactivate this enzyme and so depress apoptosis (Ferro et al., 1983). 8. Virally encoded proteins as inhibitors of apoptosis Since apoptosis has evolved in part as a mechanism for the elimination of virally-infected cells, viruses in turn have evolved mechanisms for counteracting apoptosis in their host cells. The mechanisms which viruses have evolved to block apoptosis have in turn provided researchers with valuable clues to apoptostic mechanisms in these cells. Adenoviruses produce two proteins which in hibit apoptosis, ElB 55 kDa protein and ElB 19kDa protein. The ElB 55 kDa protein inactivates p53, and blocks p53-mediated apoptosis in proliferating cells (Sarnow et al., 1982; Yew and Berk, 1992; Debbas and White, 1993). The ElB

Mechanisms of developmental cell death

19 kDa protein inhibits apoptosis induced by a wide range of stimuli including NGF deprivation, p53, TNF and Fas (Gooding et al., 1991; Rao et al., 1992; White et al., 1992 Debbas and White, 1993; Martinou et al., 1995; Sabbatini et al., 1995). Unlike the ElB proteins, expression of the adenovirus El A protein induces a round of proliferation that eventually results in apoptosis in a manner reminiscent of that following inappropriate expression oic-myc (Shenk and Flint, 1991; White et al., 1992). The mechanism by which ElA stimulates cell cycle progression is via the sequestering of Rb protein with the resultant activation of transcription factors of the E2F family (Whyte et al., 1988). Coupling of ElA-induced cell proliferation with the supression of cell death by ElB results in cell transformation (Debbas and White, 1993). The baculoviral protein p35 blocks apoptosis in insect cells and in neurons, but the site of action of this protein is not known (Clem et al., 1991; Rabizadeh et al., 1993a; Martinou et al., 1995). The cowpox viral gene crmA (cytokine response modifier gene A) encodes a serine protease inhibitor (SERPIN)-like protein that is a specific inhibitor of ICE (Ray et al., 1992; Komiyama et al., 1994). CrmA protein inhibits the activity of ICE and the production of mature IL-lyS, thereby suppressing immune responses to infection (Ray et al, 1992; Komiyama et al., 1994). Microinjection of CrmA protein into cells prevents cell death in a number of systems including Fas-induced apoptosis (Enarl et al., 1995; Los et al., 1995; Tewari and Dixit; 1995) and NGF-deprived sensory neurons (Gagliardini et al., 1994), implicating ICE or an ICE-like protease downstream of these processes. 9. Biochemical events involved in morphological changes in apoptosis 9.1. Endonuclease cleavage of DNA Since chromatin condensation is an early morphological event in apoptosis, early attention was focused on the biochemical changes that bring this about. Chromatin condensation is accompanied by fragmentation of the chromatin into an easily diag-

A. Messina and A. Jaworowski

nosable ladder of fragments, with a repeat unit of 180-200 bp, characteristic of internucleosomal fragmentation. Research has focussed on identifying endonucleases which are induced during cell death and has identified a Ca^VMg^'^-dependent endonuclease which may be involved in this process (Wyllie, 1980; Nikonova et al., 1982). In glucocorticoid- and calcium ionophore-induced apoptosis of thymocytes, cell death was associated with activation of a Ca^VMg^+ dependent Zn^"^ sensitive endonuclease that is constitutively expressed (Wyllie, 1980; Wyllie et al., 1981, 1984; Cohen and Duke 1984). Apoptosis in this model system is dependent upon protein synthesis and Ca^"^. The requirement for protein synthesis is not due to a requirement for the synthesis of the endonuclease, but the Ca^+ requirement is proposed to reflect that of the endonuclease. In mature T-cells, IL-2 deprivation also results in apoptosis, but the endonuclease could not be detected prior to IL-2 withdrawal (Duke and Cohen, 1986). Thus the endonuclease is apparently present constitutively in some cells but is inducible in other cells, such as those that are actively proliferating (Cohen and Duke, 1984). As usual, the situation is not so clear-cut, and the requirement for Ca^+ for the DNA cleavage observed in apoptosis is not universal. DNA fragmentation precedes apoptosis in PC 12 cell cultures and in primary cultures of neurons, but this does not require an elevation in [Ca^+Jj (Batistatou and Green, 1991, 1993; Edwards and Tolkovsky, 1994). Fragmentation is prevented by the nuclease inhibitor aurintricarboxylic acid and incubation with this inhibitor promotes survival of neurons following NGF withdrawal (Batistatou and Green, 1991). A few endonucleases that may mediate apoptosis have been identified. A Ca^VMg^'^-dependent 34kDa DNase 1-like endonuclease has been isolated from thymocytes (Peitsch et al., 1993) This enzyme is unusual in that it is initially complexed in the cell with actin and must be released in order to translocate into the nucleus. Other prospective candidates include DNase II, a 40 kDa Ca^-^/Mg^+independent endonuclease isolated from fibroblasts (Barry and Eastman, 1993) and NUC18, an

105

18 kDa endonuclease isolated from glucocorticoidtreated thymus (Gaido and Cidlowski, 1991). In summary, it is likely that the contribution of endonuclease cleavage of DNA to the morphological changes accompanying apoptosis, as well as the particular enzymes involved, vary from cell type to cell type. 9.2. DNA fragmentation is not absolutely required for apoptosis Despite the early popularity of using the extent of DNA fragmentation to track apoptosis, it is now clear that this event is not required for apoptosis to occur, nor does it do so in all cases of cell death. DNA fragmentation does not occur in some cases of CTL-mediated killing, and this appears to depend upon the target cell type (Duke et al., 1986; Sellins and Cohen, 1991; Ucker et al., 1992). Furthermore, Zn^"^ is able to block the nuclear changes associated with apoptosis in a number of systems, but the cytopolasmic changes proceed as normal (Giannnakis et al., 1991; Barbieri et al., 1992; Cohen et al., 1992). Enucleated cells show cytoplasmic changes characteristic of apoptosis on stimulation of Fas (Shulze Osthoff et al., 1994) and enucleated cytoplasts also undergo apoptosis in the absence of a nucleus and can be rescued by Bcl-2 and survival factors (Jacobson et al., 1994) indicating that nuclear signaling is not an absolute requirement for the late events of apoptosis to occur. Similar conclusions were reached by treating cells with inhibitors of endonucleases (Jacobson et al., 1994). 9.3. DNA condensation and cleavage into large fragments and involvement of topoisomerase II In some cells, such as thymocytes, cleavage of DNA into large fragments 300 or 50 kb is a critical step which occurs before or in the absence of typical DNA internucleosomal fragmentation (Brown et al, 1993; Oberhammer et al., 1993a,b). This event is believed to give rise to the characteristic chromatin condensation found universally in cell death (Oberhammer et al., 1993a,b). Although condensation of chromatin is often associated with

106

intemucleosomal fragmentation of the DNA, they are separate events (Sun et al., 1994) and not necessarily coupled. The size of the large fragments that appear prior to intemucleosomal degradation is consistent with cleavage of the radial-looped DNA domains by which DNA is assembled on the chromosome scaffold. The scaffold is known to contain high levels of topoisomerase II which binds to DNA with a periodicity of about 300 kb. Nuclear topoisomerases catalyze the addition or removal of supercoils in DNA, using reactions that involve cleavage and rejoining of DNA strands. Topoisomerase n has been suggested to be a candidate for the enzyme that cleaves DNA into large fragments during apoptosis. It has been pointed out that aurinotricarboxylic acid, which inhibits apoptosis in a number of systems, and which is assumed to act as an inhibitor of endonucleases, also inhibits topoisomerase II in vitro in low concentrations (Catchpoole et al., 1994). The involvement of topoisomerases in apoptosis has been suggested by many studies which show that inhibitors of topoisomerases induce apoptosis in proliferating cells. There have been a few studies showing that such compounds can also induce apoptosis in postmitotic neurons (Martin et al., 1990; Tomkins et al., 1994). Under the conditions used in these experiments, the DNA-bound topoisomerase-DNA intermediate is stabilised, which results in double strand breaks in the DNA. This, and other treatments that induce double strand breaks, promote apoptosis (Tomkins et al., 1994). 9.4. Transglutaminase Tissue transglutaminase is a Ca++-dependent enzyme that mediates cross-linking of proteins by catalysing the formation of inter-protein e (yglutamyl) lysyl isopeptide bonds to form SDSinsoluble bodies. The induction of transglutaminase activity has been found to occur during apoptosis in a wide variety of tissues such as in hepatocytes and also with glucocorticoid-induced apoptosis of thymocytes (Fesus et al, 1987; Piacentini et al, 1991). It has been suggested that the result of enhanced transglutaminase activity is

Mechanisms of developmental cell death

the formation of a network of cross-linked protein which may prevent excessive loss of cellular components as the integrity of the cell membrane is lost (Fesus, 1993). As in other features of apoptosis, induction of transglutaminase is not a universal feature in all cell types. 10. Phagocytosis One important difference that exists between models of apoptosis in vivo and in vitro is the final fate of the cell. Cells which undergo apoptosis in vitro lose their membrane integrity and then degenerate, whereas cells which undergo apoptosis in vivo are recognised and phagocytosed by their neighbours or by macrophages long before they lose membrane integrity. In the case of programmed cell death of the 131 cells in C elegans, it is known that most of the ced mutations, apart from ced 3,4 and 9, are in genes which code for proteins involved in recognition and phagocytosis of the dying cells by their neighbours (Hedgecock et al., 1983; Ellis et al., 1991a). It is now clear that cells undergoing cell death undergo specific membrane changes to facilitate recognition and phagocytosis. Several mechanisms have been described by which macrophages recognise cells undergoing cell death. Expression of 'eat me' molecules may involve loss of membrane phospholipid asymmetry which results in the exposure of phosphatidylserine, normally concentrated on the inner leaflet of the plasma membrane, on the outer surface of the apoptotic cell membrane. The exposed phosphatidylserine interacts with the CD35 receptor on macrophages (Savill et al., 1989a,b). A second mechanism, which has been implicated in the recognition of apoptotic rodent thymocytes by peritoneal macrophages in vitro, involves the exposure of carbohydrate moieties characteristic of immature glycoproteins on the target cell surface (Duvall et al., 1985). Recognition, via lectin-like molecules on the surface of the macrophage, can be blocked by addition of A^-acetyl glucosamine and similar sugars which are not normally found at the termini of mature glycoprotein sugar chains. Thirdly, bone marrow- or monocyte-derived macrophages recognise apoptotic lymphocytes or

A. Messina and A. Jaworowski

neutrophils via CD 36-mediated binding to the vitronectin (a cr-v-beta 3 integrin) receptor on the target cell or via thrombospondin receptors on the macrophage binding to as yet unidentified ligands (Savill et al., 1990, 1993; Fadok et al., 1992). Another macrophage surface antigen, defined by the monoclonal antibody 61D3, is involved in recognition of apoptotic cells via a pathway that is distinct from that of the vitronectin receptor (Flora and Gregory, 1994). It has also been reported that the asialoglycoprotein receptor aids in removal of apoptotic parenchymal cells in the liver (Dini et al., 1992). Monolayers of fibroblasts can also phagocytose apoptotic neutrophils by two distinct methods. One of these uses the vitronectin receptor, whereas the other uses mannose/fructosespecific lectins which are not involved in macrophage-mediated phagocytosis of apoptotic cells (Hall etal., 1994). Promotion of phagocytosis, independent of nuclear changes, may be as important a strategy for cell death in vivo as nuclear degradation. The importance of phagocytosis in vivo can be seen in the case of the thymus where, it has been estimated, up to one third of the thymocytes undergo cell death each day, but where histological examinations reveal very few 'apoptotic cells', although recently a number of cells with DNA strand breaks have been detected (Surh and Sprent, 1994). Clearly, in this case, the dying cell is rapidly phagocytosed following recognition by macrophages at an early stage of the death pathway. Even the induction of protective mechanisms blocking early events in apoptosis may not be sufficient to prevent recognition and engulfment by macrophages. In a recent study using transgenic mice in which the bcl-2 gene was expressed in neutrophils, although Bcl-2 rendered the neutrophils comparatively resistent to apoptosis in vitro, senescent cells were still recognised and eliminated by macrophages (Lagasse andWeissman, 1994). The packaging processes that accompany apoptosis, and which are currently used to identify this phenomenon, may therefore be a late event to either aid digestion of the dying cell after it has been phagocytosed or to prevent spillage of cellular contents in tissues where phagocytosis may not

107

be as rapid. The humble macrophage therefore plays a crucial role in apoptosis in vivo, recognising and engulfing the dying cell before it has progressed to the stage of releasing potentially immunoreactive intracellular contents. Interestingly, this has to occur without stimulating the normal proinflammatory responses of the macrophage. 11. The importance of cell-cell and cell-matrix contact in cell death A feature of developmental cell death in vivo, absent from many of the models used to study apoptosis in vitro, is that in vivo, cell death is accompanied by detachment of cells from their neighbours and from the extracellular matrix. The dependence of cells on specific anchorage in a tissue in order to survive is an important homeostatic mechanism controlling tissue development and has been suggested to be an important evolutionary development in metazoans to safeguard against cells growing in inappropriate places within the organism (Ruoslahti and Reed, 1994). The development of strategies to escape from detachment-induced cell death is essential for cellular transformation and tumour formation. The attachment of cells to each other, and to components of the extracellular matrix, is mediated by integrins, which suggests that signalling pathways downstream of integrins play an important role in apoptosis in vivo. Normal endothelial and epithelial cells undergo apoptosis when they are detached from substrate (Meredith et al., 1993; Frisch and Francis, 1994; Re et al., 1994), via a pathway that is blocked by expression of Bcl-2 (Frisch and Francis, 1994). That the prevention of apoptosis is specifically mediated by integrin signalling rather than attachment per se, was indicated by the finding that prevention of apoptosis and cell-spreading followed attachment of these cells to substrates coated with anti-integrin antibodies but not to surfaces coated with antibodies directed against non-integrin proteins (Meredith et al., 1993). It is probable that the prevention of apoptosis is due to the induction of cell spreading following binding of the integrins to the extracellular matrix. This suggests that focal adhesion kinase (or FAK) may be involved. In this

108

context, it is interesting that FAK has been shown to form a stable complex with PI-3 kinase following adhesion of fibroblasts via a pathway that is apparently different to stimulation of the same complex following PDGF stimulation (Chen and Guan, 1994a,b). These workers showed that activation of pathways downstream of either the integrins or of PDGF receptor caused phosphorylation of the regulatory subunit of PI-3 kinsase by FAK. We have seen that there is some evidence for the involvement of PI 3 kinase in prevention of cell death via the NGF receptor (see Section 3.1). The requirement for attachment to prevent cells undergoing apoptosis in vivo is a general one. Recently many model systems have been used to study this phenomenon. These studies usually examine apoptosis under conditions in which the target cells are either plated out onto various extracellular matrix components or onto supporting cell monolayers (Bendall et al., 1994; Koopman et al., 1994). In a recent study, basement membrane extracellular matrix (ECM), but not fibronectin or collagen, suppressed apoptosis in mammary epithelial cells in vivo and in vitro. Apoptosis was induced by antibodies to P-\ integrins or by overexpression of stromolysin-1 which degrades ECM. Loss of ECM correlated with ICE expression and addition of an inhibitor of ICE prevented apoptosis. Therefore ECM may regulate apoptosis through integrin-dependent regulation of ICE or an ICE homologue (Boudreau et al., 1995). 12. Is apoptosis linked to the cell cycle? 72.7. The cell cycle Cells that are not terminally differentiated and are in a state of quiescence either proliferate or undergo developmental cell death. Our current notions on cell proliferation regard this process as transition of the cell through a cell cycle in which a programmed sequence of gene expression is regulated by cyclin/cyclin dependent kinase (cdk) activity. Different cyclin/cdk complexes regulate passage through different phases of the cell cycle, the key events of which are DNA synthesis (S phase) and mitosis (M phase). The activity of vari-

Mechanisms of developmental cell death

ous cyclin/cdk complexes is controlled (1) at the level of their synthesis, (2) by phosphorylation events which activate and deactivate the cdk and (3) by the synthesis of inhibitor proteins such as pl^"'' and p2lwafi/cipi (mts-1) (for a recent review see Morgan, 1995). Checkpoints occur in the gap phases which occur before the S and M phases. In terms of making a commitment to enter the cell cycle and complete cell division, the cell has to pass the 'restriction point' of mid to late Gj which requires continued stimulation of the cell by growth factor, after which time progression through the cell cycle is growth factor-independent. As expected, passage through the cell cycle is tightly regulated, not only by the need for continued exposure to growth factor to reach the restriction point, but also at a mitotic checkpoint where mitosis is allowed to proceed only on successful replication of the genome. Recent work on yeast replication has also identified other checkpoint controls in which progression through the cell cycle is dependent on attaining a critical physical size. We shall see that failure to successfully pass through these checkpoints may result in apoptosis. 72.2. p53 12.2.1. p53 and DNA-damage-induced apoptosis p53, a tumour suppressor gene discovered in the late 1970s, is the most commonly-found mutated gene in human tumours (Levine et al., 1991). It encodes a transcription factor which activates genes responsible for cell cycle progression or the cell's response to DNA damage. Although not primarily a concern of this review, apoptosis induced by damage to DNA is directly linked to the cell cycle via p53 action. The activity of p53 is normally downregulated via the binding of MDM-2 to allow passage of cells through the cell cycle (Oliner, 1993; Oliner et al, 1993). Following DNA damage, p53 accumulates in the nucleus as a result of increased stability of the protein and a change in its subcellular location from the cytoplasm to the nucleus (Fritsche et al., 1993). The increase in nuclear p53 causes Gi arrest (Levine et al., 1991), which may be the result of

A. Messina and A. Jaworowski

increased expression of the cyclin-dependent kinase inhibitor, p21^af-i/cip-i (el Diery et al., 1993; Dulic et al., 1994 ). This has two consequences: (1) the cell is arrested in Gj, which allows DNA repair to occur prior to S phase and (2) if the DNA damage is too great, and repair cannot be completed, the cell undergoes apoptosis. Not all forms of DNA-damage-induced apoptosis are p53dependent, however. In studies using p53knockout mice, it was shown that cycling T lymphoma cells and activated T lymphocytes undergo apoptosis after irradiation and cytotoxic drug treatment, while lymphoblasts that express Bcl-2 growth arrest in Gj and G2 but resist apoptosis (Strasser et al, 1994). These studies suggest the existence of a p53-independent Gi checkpoint, which is invoked following DNA damage. 72.2.2. Cell cycle genes potentially regulated by p53 Many other genes have sequences in their promoters which are recognised by p53, for example c-myc (Moberg et al., 1992), Rb itself (Shiio et al., 1992), the multidrug-resistance gene MDRl (Chin et al., 1992), the proliferating cell nuclear antigen PCNA gene, encoding a component of the DNA replication machinery (Subler et al., 1992) the mdm-2 gene (Momand et al., 1992; Wu et al., 1993), and c-jun among others (Santhanam et al., 1991; Lechner etal, 1992; Agoff et al., 1993). The significance of these interactions for either p53induced cell cycle arrest or apoptosis is not yet known. The activity of p53 in controlling transcription may also be modulated by upstream signal transduction pathways. It has been shown, for example, that both casein kinase II and p34^^'^2 influence DNA replication by phosphorylating p53 (Herrmann et al., 1991; Milner et al., 1990; Sturzbecheretal, 1990). 12.2.3. Lack of involvement ofp53 in developmental cell death Studies using p53-knockout mice have helped to clarify the processes in which p53 is essential. p53-null animals show apparently normal development (Donehower et al., 1992). Many cells, including immature thymocytes, are resistant to ir-

109

radiation-induced and drug-induced DNA damage which normally causes apoptosis (Clarke et al., 1993; Lotem and Sachs, 1993; Lowe et al., 1993) and the p53-null animals develop thymic lymphomas (Donehower et al, 1992; Jacks et al., 1994). These results corroborate the findings that p53 is involved in mediating apoptosis in cells that sustain DNA damage. However, T cell development proceeds as normal and thymocytes remain sensitive to glucocorticoid- or TCR-induced apoptosis (Clarke et al., 1993; Lowe et al., 1993). Studies with viral proteins which inhibit p53 action (see Section 8) have shown that p53 does not appear to be involved in NGF withdrawal-induced apoptosis in primary cultures of rat sympathetic neurons (Martinou et al., 1995), supporting the observations quoted above that normal development occurs in the absence of p53. 12.2.4. p53 and modulation ofBcl-2/Bax checkpoint In some cells, it has been shown that the levels of Bcl-2 and Bax may be regulated by p53, thus linking these two important mediators of apoptosis. In Ml myeloid cells, activation of a temperature-sensitive p53 transgene, induced apoptosis and was associated with upregulation of bax expression and downregulation of bcl-2 expression (Selvakumaran et al., 1994). Ectopic overexpression of bcl'2 did not prevent p53-mediated apoptosis but did prevent TGF-)3-mediated apoptosis. P53 has recently been shown to transactivate reporter gene constructs containing elements of the bax promoter (Miyashita and Reed, 1995), raising the possibility that it is a direct transcriptional activator of the human bax gene. 12.3. Apoptosis as a consequence of inappropriate stimulation ofpassage through the cell cycle If a cell that is committed to division cannot successfully complete the cell cycle, a cell death pathway is activated. As discussed in Section 12.1, this may occur under circumstances in which the genome is damaged by external factors, such as chemical and irradiation-induced damage, or it

no may occur in vitro by enforced passage through the cell cycle. As we have seen above (see Section 5.2), fibroblasts that are engineered to express cmyc constitutively remain in cycle and may complete a round of cell division, but eventually undergo apoptosis (Evan et al., 1992). Given that apoptosis is induced by overexpression of a protooncogene that was originally assumed to drive the cell through the cell cycle, the question arises as to whether there is an intimate connection between these two processes. It has been suggested that apoptosis will occur in response to c-Myc if the cell fails to receive sufficient signals to traverse the cell cycle (Evan et al., 1992). In this model, apoptosis is a default pathway that is prevented in cycling cells by receipt of the full complement of proliferative signals. This is consistent with a role for apoptosis as a protective mechanism against transformation caused by inappropriate expression of one or a few proliferative signals. Evidence for the coupling of apoptosis to the cell cycle during developemental cell death is suggested by experiments which show that the cell undergoing cell death does so from a particular phase of the cell cycle. For example, apoptosis in thymocytes following T cell receptor activation, seems to require entry into S phase (Boehme and Lenardo, 1993), while apoptosis induced in target cells by CTL seem to require the target cell being in a quiescent, or GQ, phase (Nishioka and Welsh, 1994; Shi etal., 1994). 12.4. Apoptosis and mitosis Perhaps the most intriguing link between cell death and the cell cycle is the suggestion that the nuclear condensation observed during this process is in reality an aberrant or inappropriate form of mitosis (e.g. Meikrantz et al, 1994; Shi et al., 1994) in which the chromatin condensation and packaging normally activated during M phase is switched on in the absence of chromosomal division. Given the overt similarities between apoptotic chromatin condensation and normal mitotic events, this suggestion is appealing. For such a process to occur, the mitotic checkpoints would have to be by-passed. It must be remembered that

Mechanisms of developmental cell death

apoptosis can occur in the absence of nuclear events (see Section 9.2), but the nuclear changes that often accompany death may involve similar processes as those operating during mitosis. 13, Concluding remarks In multicellular organisms, the importance of the organism overrides that of an individual cell. This required the evolution of a process of cell death to both control the development of the organism and, as we now appreciate, to protect it from individual cells that are damaged or compromised in some way, such as by viral infection. We can regard both these cases as the development of an altruistic response by an individual cell to increase the survivability of the organism. As we have seen, studies of the nematode C. elegans confirm that aspects of the basic mechanism by which this cell death is accomplished, such as the involvement of aspartyl-directed proteases and Bcl-2-like molecules, arose early in metazoan evolution and have remained conserved throughout the evolution of more complex organisms. Although we have concentrated in the present review on developmental cell death, this process seems to share many morphological features with cell death induced by damage. It is probably too early to say whether the fundamental mechanisms of cell death which occur developmentally are the same as those which eliminate aberrent cells. It has been suggested, however, that the latter pathways were a later evolutionary refinement of the former (Vaux et al., 1994). A striking feature of cell death, in either the developmental or the protective sense, is that removal of the dying cell is accomplished by selective stimulation of the organism's immune system. There is currently a lack of information on how, for example, recognition by macrophages of a dying cell occurs without stimulating the inflammatory responses of the macrophages. Apoptosis can be conveniently divided into three phases: (1) the initial stimulus to undergo cell death,(2) the central signalling events and (3) the final morphological changes. In developmental cell death, the first phase may be intrinsic to the cell's genetic programming, or it may be induced

A. Messina and A. Jaworowski

by the environment, such as by lack of suitable survival factors. The signals which initiate the process are probably as varied as the cell types which undergo cell death, which has led to a multitude of signal transduction pathways being implicated as early events in apoptosis. In many cases, these early events converge into a single pathway involving the Bcl-2/bax and ICE checkpoints, at which point the decision to undergo apoptosis is carefully reviewed. As we have seen, this pathway does not operate in all cases and there remain others to be discovered, such as those involving the presumptive ced-4 homologue. In vitro, if the decision to undergo cell death is made at the apoptosis checkpoints, the cell undergoes a series of morphological changes whose outcome is to package intracellular contents in order to facilitate their disposal without leakage of potentially immunogenic products. The biochemical processes by which this occurs are poorly understood, but again may vary from cell to cell (e.g. the involvement of endonucleases). Whether packaging of the nuclear contents utilises the same machinery as is used in mitosis is an intriguing, but as yet unproven, concept. The development of an apoptotic morphology during cell death in vivo depends upon which tissues are considered. Apoptotic cells are easily discerned in the developing nervous system and, as Wyllie originally described in his seminal paper (1980), in the liver and other organs. However in the immune system, massive cell death occurs without discernible morphological changes. In both cases, the central process is one of phagocytosis of the dying cell or its packaged contents, either by macrophages or neighbouring cells. Presumably the rate of phagocytosis, determined perhaps by the accessibility to macrophages, determines how far the morphological changes proceed before the cell is eliminated. However, it is important to remember that it is a common convention to refer only to death which is accompanied by these morphological changes as 'apoptosis'. Thus, the anomaly may exist of instances of the same functional process being classified differently, dependent merely on the rate at which the events in question occur. An alternative description used in relation to the process of cell

111

death is 'programmed cell death', to imply that it occurs via a genetically programmed sequence of biochemical events and to distinguish it from accidental or traumatic death of a cell. However this term was originally coined by developmental biologists to describe the process of cell deletion which is a normal part of the developing multicellular organism. As a generality, we believe that the term apoptosis is to be preferred when describing cell death occuring via a specific pathway aimed at eliminating the cell without causing inflammation, irrespective of whether the dying cell aquires an 'apoptotic morphology'. References Agoff, S.N., Hou, J., Linzer, D.I. and Wu, B. (1993) Regulation of the human hsp70 promoter by p53. Science 259: 8487. Alderson, M.R., Tough, T.W., Davis Smith, T., Braddy, S., Falk, B., Schooley, K.A., Goodwin, R.G., Smith, C.A., Ramsdell, F. and Lynch, D.H. (1995) Fas ligand mediates activation-induced cell death in human T lymphocytes. J. Exp. Med. 181:71-77. Allsopp, I.E., Wyatt, S., Paterson, H.F. and Davies, A.M. (1993) The proto-oncogene bcl-2 can selectively rescue neurotrophic factor-dependent neurons from apoptosis. Cell 73: 295-307. Amati, B., Dalton, S., Brooks, M.W., Littlewood, T.D., Evan, G.I. and Land, H. (1992) Transcriptional activation by the human c-Myc oncoprotein in yeast requires interaction with Max. Nature 359: 423-426. Amati, B., Littlewood, T.D., Evan, G.I. and Land, H. (1993) The c-Myc protein induces cell cycle progression and apoptosis through dimerization with Max. EMBO J. 12: 5083-5087. Andjelic, S., Drappa, J., Lacy, E., Elkon, K.B. and Nikolic Zugic, J. (1994) The onset of Fas expression parallels the acquisition of CD8 and CD4 in fetal and adult alpha beta thymocytes. Int. Immunol. 6: 73-79. Barbieri, D., Troiano, L., Grassilli, E., Agnesini, C, Cristofalo, E.A., Monti, D., Capri, M., Cossarizza, A. and Franceschi, C. (1992) Inhibition of apoptosis by zinc: a reappraisal. Biochem. Biophys. Res. Commun, 187: 12561261. Barry, M.A. and Eastman, A. (1993) Identification of deoxyribonuclease II as an endonuclease involved in apoptosis. Arch. Biochem. Biophys. 300: 440-450. Batistatou, A. and Greene, L.A. (1991) Aurintricarboxylic acid rescues PC 12 cells and sympathetic neurons from cell death caused by nerve growth factor deprivation: correlation with suppression of endonuclease activity. /. Cell Biol. 115:461^71.

112 Batistatou, A. and Greene, L.A. (1993) Internucleosomal DNA cleavage and neuronal cell survival/death. J. Cell Biol 122: 523-532. Bendall, L.J., Daniel, A., Kortlepel, K. and Gottlieb, D.J. (1994) Bone marrow adherent layers inhibit apoptosis of acute myeloid leukemia cells. Exp. Hematol. 22: 12521260. Berke, G. (1994) The binding and lysis of target cells by cytotoxic lymphocytes: molecular and cellular aspects. Annu. Rev. Immunol. 12: 735-773. Bissonnette, R.P., Echeverri, F., Mahboubi, A. and Green, D.R. (1992) Apoptotic cell death induced by c-myc is inhibited by bcl-2. Nature 359: 552-554. Bissonnette, R.P., McGahon, A., Mahboubi, A. and Green, D.R. (1994) Functional Myc-Max heterodimer is required for activation-induced apoptosis in T cell hybridomas. /. Exp. Med. 180: 2413-2418. Boehme, S.A. and Lenardo, M.J. (1993) Ligand-induced apoptosis of mature T lymphocytes (propriocidal regulation) occurs at distinct stages of the cell cycle. Leukemia 1 (Suppl. 2): S45-S49. Boise, L.H., Gonzalez Garcia, M., Postema, C.E., Ding, L., Lindsten, T., Turka, L.A., Mao, X., Nunez, G. and Thompson, C.B. (1993) bcl-x, a bcl-2-related gene that functions as a dominant regulator of apoptotic cell death. Cell 74: 597-608. Boldin, M.P., Mett, I.L., Varfolomeev, E.E., Chumakov, I., Shemer Avni, Y., Camonis, J.H. and Wallach, D. (1995) Self-association of the "death domains" of the p55 tumor necrosis factor (TNF) receptor and Fas/APOl prompts signaling for TNF and Fas/APOl effects. J. Biol. Chem. 270: 387-391. Borner, C, Martinou, I., Mattmann, C, Irmler, M., Schaerer, E., Martinou, J.C. and Tschopp, J. (1994) The protein bcl-2 alpha does not require membrane attachment, but two conserved domains to suppress apoptosis. J. Cell Biol. 126: 1059-1068. Boudreau, N., Sympson, C.J., Werb, Z. and Bissell, M.J. (1995) Suppression of ICE and apoptosis in mammary epithelial cells by extracellular matrix. Science 267: 891-893. Brown, D.G., Sun, X.M. and Cohen, G.M. (1993) Dexamethasone-induced apoptosis involves cleavage of DNA to large fragments prior to internucleosomal fragmentation. J. Biol. Chem. 268: 3037-3039. Brunner, T., Mogil, R.J., LaFace, D., Yoo, N.J., Mahboubi, A., Echeverri, F., Martin, S.J., Force, W.R., Lynch, D.H., Ware, C.F. and Green, D.R. (1995) Cell-autonomous Fas (CD95)/Fas-ligand interaction mediates activation-induced apoptosis in T-cell hybridomas. Nature 373: 441-444. Bursch, W., Paffe, S., Putz, B., Barthel, G. and Schulte Hermann, R. (1990) Determination of the length of the histological stages of apoptosis in normal liver and in altered hepatic foci of rats. Carcinogenesis 11: 847-853. Catchpoole, D.R. and Stewart, B.W. (1994) Inhibition of topoisomerase II by aurintricarboxylic acid: implications for mechanisms of apoptosis. Anticancer Res. 14: 853-856.

Mechanisms of developmental cell death Cerretti, D.P., Kozlosky, C.J., Mosley, B., Nelson, N., Van Ness, K., Greenstreet, T.A., March, C.J., Kronheim, S.R., Druck, T., Cannizzaro, L.A., Huebner, K. and Black, R.A. (1992) Molecular cloning of the interleukin-1 beta converting enzyme. Science 256: 97-100. Chen, H.C. and Guan, J.L. (1994a) Stimulation of phosphatidylinositol 3-kinase association with focal adhesion kinase by platelet-derived growth factor. J. Biol. Chem. 49: 31229-31233. Chen, H.C. and Guan, J.L. (1994b) Association of focal adhesion kinase with its potential substrate phosphatidylinositol 3-kinase. Proc. Natl. Acad Sci. USA 91: 10148-10152. Chen, J., Lansford, R., Stewart, V., Young, F. and Alt, F.W. (1993) RAG-2-deficient blastocyst complementation: an assay of gene function in. Proc. Natl. Acad. Sci. USA 90: 4528-4532. Chen, Q., Kinch, M.S., Lin, T.H., Burridge, K. and Juliano, R.L. (1994) Integrin-mediated cell adhesion activates mitogen-activated protein kinases. J. Biol. Chem. 269: 2660226605. Chin, K.V., Ueda, K., Pastan, I. and Gottesman, M.M. (1992) Modulation of activity of the promoter of the human MDRl gene by Ras and p53. Science 255: 459-462. Chittenden, T., Harrington, E.A., O Conner, R., Flemington, C, Lutz, R.J., Evan, G.I. and Guild, B.C. (1995) Induction of apoptosis by the Bcl-2 homologue Bak. Nature 374: 733-736. Chiu, V.K., Walsh, CM., Liu, C.C, Reed, J.C. and Clark, W.R. (1995) Bcl-2 blocks degranulation but not fas-based cell-mediated cytotoxicity. J. Immunol. 154: 20232032. Cifone, M.G., De Maria, R., Roncaioli, P., Rippo, M.R., Azuma, M., Lanier, L.L., Santoni, A. and Testi, R. (1994) Apoptotic signaling through CD95 (Fas/Apo-1) activates an acidic sphingomyelinase. J. Exp. Med. 180: 1547-1552. Clarke, A.R., Purdie, C.A., Harrison, D.J., Morris, R.G., Bird, C.C, Hooper, M.L. and Wyllie, A.H. (1993) Thymocyte apoptosis induced by p53-dependent and independent pathways. Nature 362: 849-852. Clem, R.J., Fechheimer, M. and Miller, L.K. (1991) Prevention of apoptosis by a baculovirus gene during infection of insect cells. Science 254: 1388-1390. Cohen, G.M., Sun, X.M., Snowden, R.T., Dinsdale, D. and Skilleter, D.N. (1992) Key morphological features of apoptosis may occur in the absence of internucleosomal DNA fragmentation. Biochem. J. 286: 331-334. Cohen, J.J. and Duke, R.C (1984) Glucocorticoid activation of a calcium-dependent endonuclease in thymocyte nuclei leads to cell death. J. Immunol. 132: 38-42. Colotta, F., Polentarutti, N., Sironi, M. and Mantovani, A. (1992) Expression and involvement of c-fos and c-jun protooncogenes in programmed cell death induced by growth factor deprivation in lymphoid cell lines. J. Biol. Chem. 267: 18278-18283. Countaway, J.L., Northwood, I.C and Davis, R.J. (1989) Mechanism of phosphorylation of the epidermal growth

A. Messina and A. Jaworowski factor receptor at threonine 669. J. Biol. Chem. 264: 1082810835. Darmon, A.J., Ehrman, N., Caputo, A., Fujinaga, J. and Bleackley, R.C. (1994) The cytotoxic T cell proteinase granzyme B does not activate interleukin-1 beta-converting enzyme. J. Biol Chem. 269: 32043-32046. Davis, R.J., Girones, N. and Faucher, M. (1988) Two alternative mechanisms control the interconversion of functional states of the epidermal growth factor receptor. /. Biol. Chem. 263: 5373-5379. Dbaibo, G.S., Pushkareva, M.Y., Jayadev, S., Schwarz, J.K., Horowitz, J.M., Obeid, L.M. and Hannun, Y.A. (1995) Retinoblastoma gene product as a downstream target for a ceramide-dependent pathway of growth arrest. Proc. Natl. Acad. Sci. USA 92: 1347-1351. de Murcia, G. and Menissier de Murcia, J. (1994) Poly(ADPribose) polymerase: a molecular nick-sensor. Trends. Biochem. Sci. 19: 172-176. Debatin, K.M. (1994) APO-1 (CD95) and Bcl-2: determinants of cell death in the human thymus. Res. Immunol. 145: 146151. Debatin, K.M., Suss, D. and Krammer, P.H. (1994) Differential expression of APO-1 on human thymocytes: implications for.negative selection. Eur. J. Immunol. 24: 753758. Debbas, M. and White, E. (1993) Wild-type p53 mediates apoptosis by ElA, which is inhibited by ElB. Genes Dev. 7: 546-554. Dhein, J., Walczak, H., Baumler, C, Debatin, K.M. and Krammer, P.H. (1995) Autocrine T-cell suicide mediated by APO-l/(Fas/CD95). Nature 373: 438-441. di Giovine, F.S. and Duff, G.W. (1990) Interleukin 1: the first interleukin./mmMwo/. Today 11: 13-20. Dinarello, C.A. (1991) Inflammatory cytokines: interleukin-1 and tumour necrosis factor as effector molecules in autoimmune diseases. Curr. Opin. Immunol. 3: 941-948. Dini, L., Lentini, A., Mantile, G., Massimi, M. and Devirgiliis, L.C. (1992) Receptor-mediated endocytosis of galactose and mannose exposing ligands: an electron microscopic study on adult and neonatal cultured rat hepatocytes. Biol. CW/74: 217-224. Dobrowsky, R.T. and Hannun, Y.A. (1992) Ceramide stimulates a cytosolic protein phosphatase. J. Biol. Chem. 267: 5048-5051. Dobrowsky, R.T. and Hannun, Y.A. (1993) Ceramideactivated protein phosphatase: partial purification and relationship to protein phosphatase 2A. Adv. Lipid Res. 25: 9 1 104. Dobrowsky, R.T., Kamibayashi, C, Mumby, M.C. and Hannun, Y.A. (1993) Ceramide activates heterotrimeric protein phosphatase 2A. J. Biol. Chem. 268: 15523-15530. Donehower, L.A., Harvey, M., Slagle, B.L., McArthur, M.J., Montgomery, C.A.J., Butel, J.S. and Bradley, A. (1992) Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature 356: 215221.

113 Duke, R.C, Chervenak, R. and Cohen, J.J. (1983) Endogenous endonuclease induced DNA fragmentation: an early event in cell-mediated cytolysis. Proc. Natl. Acad. Sci. USA 80: 6361-6365. Duke, R.C. and Cohen, J.J. (1986) IL-2 addiction: withdrawal of growth factor activates a suicide program in dependent T cells. Lymphokine Res. 5: 289-299. Duke, R.C, Cohen, J.J. and Chervenak, R. (1986) Differences in target cell DNA fragmentation induced by mouse cytotoxic T lymphocytes and natural killer cells. J. Immunol. 137: 1442-1447. Dulic, v., Kaufmann, W.K., Wilson, S.J., Tlsty, T.D., Lees, E., Harper, J.W., EUedge, S.J. and Reed, S.I. (1994) p53dependent inhibition of cyclin-dependent kinase activities in human fibroblasts during radiation-induced Gl arrest. Ceine: 1013-1023. Duvall, E., Wyllie, A.H. and Morris, R.G. (1985) Macrophage recognition of cells undergoing programmed cell death (apoptosis). Immunology 56: 351-358. Edwards, S.N., Buckmaster, A.E. and Tolkovsky, A.M. (1991) The death programme in cultured sympathetic neurones can be suppressed at the posttranslational level by nerve growth factor, cyclic AMP and depolarization. J. Neurochem. 57: 2140-2143. Edwards, S.N. and Tolkovsky, A.M. (1994) Characterization of apoptosis in cultured rat sympathetic neurons after nerve growth factor withdrawal. /. Cell Biol. 124: 537546. Eichler, M.E., Dubinsky, J.M. and Rich, K.M. (1992) Relationship of intracellular calcium to dependence on nerve growth factor in dorsal root ganglion neurons in cell culture. /. Neurochem. 58: 263-269. Eichler, M.E., Dubinsky, J.M., Tong, J. and Rich, K.M. (1994) The ability of diphenylpiperazines to prevent neuronal death in dorsal root ganglion neurons in vitro after nerve growth factor deprivation and in vivo after axotomy. /. Neurochem. 62: 2148-2157. el Deiry, W.S., Tokino, T., Velculescu, V.E., Levy, D.B., Parsons, R., Trent, J.M., Lin, D., Mercer, W.E., Kinzler, K.W. and Vogelstein, B. (1993) WAFl, a potential mediator of p53 tumor suppression. Cell 75: 817-825. Ellis, H.M. and Horvitz, H.R. (1986) Genetic control of programmed cell death in the nematode C. elegans. Cell 44: 817-829. Ellis, R.E., Jacobson, D. and Horvitz, H.R. (1991a) Genes required for the engulfment of cell corpses during programmed cell death. Genetics 112: 591-603. Ellis, R.E., Yuan, J.Y. and Horvitz, H.R. (1991b) Mechanisms and functions of cell death. Annu. Rev. Cell Biol. 7: 663698. Enarl, M., Hug, H. and Nagata, S, (1995) Involvement of an ICE-like protease in Fas-mediated apoptosis. Nature 375: 78-80. Estus, S., Zaks, W.J., Freeman, R.S., Gruda, M., Bravo, R. and Johnson, E.M.J. (1994) Altered gene expression in neurons during programmed cell death: identification of c-jun as

114 necessary for neuronal apoptosis. J. Cell Biol 127: 17171727. Evan, G.I., Wyllie, A.H., Gilbert, C.S., Littlewood, T.D., Land, H., Brooks, M., Waters, CM., Penn, L.Z. and Hancock, D.C. (1992) Induction of apoptosis in fibroblasts by c-myc protein. Cell 69: 119-128. Ewen, M.E. (1994) The cell cycle and the retinoblastoma protein family. Cancer Metastasis Rev. 13: 45-66. Fadok, V.A., Savill, J.S., Haslett, C, Bratton, D.L., Doherty, D.E., Campbell, P.A. and Henson, P.M. (1992) Different populations of macrophages use either the vitronectin receptor or the phosphatidylserine receptor to recognize and remove apoptotic cells. J. Immunol. 149: 4029-4035. Fanidi, A., Harrington, E.A. and Evan, G.I. (1992) Cooperative interaction between c-myc and bcl-2 proto-oncogenes. Nature 359: 554-556. Farrow, S.N., Korsmeyer, S.J., Martinou, I., Raven, T., Pun, K.T., Grinham, C.J., Martinou, J.C. and Brown, R. (1995) Cloning of a Bcl-2 homologue by interaction with adenovirus ElB 19K. Nature 374: 731-733. Faucher, M., Girones, N., Hannun, Y.A., Bell, R.M. and Davis, R.J. (1988) Regulation of the epidermal growth factor receptor phosphorylation state by sphingosine in A431 human epidermoid carcinoma cells. /. Biol. Chem. 263: 5319-5327. Fernandes Alnemri, T., Litwack, G. and Alnemri, E.S. (1994) CPP32, a novel human apoptotic protein with homology to Caenorhahditis elegans cell death protein Ced-3 and mammalian interleukin-1 beta-converting enzyme. J. Biol. Chem. 269: 30761-30764. Fernandez Sarabia, M.J. and Bischoff, J.R. (1993) Bcl-2 associates with the ras-related protein R-ras p23. Nature 366: 274-275. Ferro, A.M., Higgins, N.P. and Olivera, B.M. (1983) Poly(ADP-ribosylation) of a DNA topoisomerase. J. Biol. Chem. 258: 6000-6003. Fesus, L. (1993) Biochemical events in naturally occurring forms of cell death. FEBS Lett. 328: 1-5. Fesus, L., Thomazy, V. and Falus, A. (1987) Induction and activation of tissue transglutaminase during programmed cell death. FEBS Lett. 224: 104-108. Rora, P.K. and Gregory, CD. (1994) Recognition of apoptotic cells by human macrophages: inhibition by a monocyte/macrophage-specific monoclonal antibody. Eur. J, Immunol. 24: 2625-2632. Forrest, V.J., Kang, Y.H., McClain, D.E., Robinson, D.H. and Ramakrishnan, N. (1994) Oxidative stress-induced apoptosis prevented by Trolox. Free Rad. Biol. Med. 16: 675-684. Franklin, J.L., Sanz Rodriguez, C, Juhasz, A., Deckwerth, T.L. and Johnson, E.M.J. (1995) Chronic depolarization prevents programmed death of sympathetic neurons in vitro but does not support growth: requirement for Ca2+ influx but not Trk activation. J. Neurosci. 15: 643-664. Frisch, S.M. and Francis, H. (1994) Disruption of epithelial cell-matrix interactions induces apoptosis. J. Cell Biol. 124: 619-626.

Mechanisms of developmental cell death Fritsche, M., Haessler, C and Brandner, G. (1993) Induction of nuclear accumulation of the tumor-suppressor protein p53 by DNA-damaging agents. Oncogene 8: 307-318. Gagliardini, V., Fernandez, P.A., Lee, R.K., Drexler, H.C, Rotello, R.J., Fishman, M.C and Yuan, J. (1994) Prevention of vertebrate neuronal death by the crmA gene. Science 263: 826-828. Gaido, M.L. and Cidlowski, J.A. (1991) Identification, purification and characterization of a calcium-dependent endonuclease (NUC18) from apoptotic rat thymocytes. NUC18 is not histone H2B. J. Biol. Chem. 266: 18580-18585. Garcia, I., Martinou, I., Tsujimoto, Y. and Martinou, J.C. (1992) Prevention of programmed cell death of sympathetic neurons by the bcl-2 proto-oncogene. Science 258: 302304. Giannakis, C, Forbes, I.J. and Zalewski, P.D. (1991) Ca2+/Mg(2+)-dependent nuclease: tissue distribution, relationship to inter-nucleosomal DNA fragmentation and inhibition by Zn2+. Biochem. Biophys. Res. Commun. 181: 915-920. Glucksmann, A. (1951) Cell, deaths in normal vertebrate ontogeny. Biol. Rev. 26: 59-86. Goetzl, E.J., An, S. and Zeng, L. (1995) Specific suppression by prostaglandin E2 of activation-induced apoptosis of human CD4+CD8+ T lymphoblasts. /. Immunol. 154: 10411047. Gooding, L.R., Aquino, L., Duerksen Hughes, P.J., Day, D., Horton, T.M., Yei, S.P. and Wold, W.S. (1991) The ElB 19,000-molecular-weight protein of group C adenoviruses prevents tumor necrosis factor cytolysis of human cells but not of mouse cells. J. Virol. 65: 3083-3094. Gratiot Deans, J., Ding, L., Turka, L.A. and Nunez, G. (1993) bcl-2 proto-oncogene expression during human T cell development. Evidence for biphasic regulation. J. Immunol. 151:83-91. Green, D.R. and Scott, D.W. (1994) Activation-induced apoptosis in lymphocytes. Curr. Opin. Immunol. 6: 476487. Green, D.R., Zheng, H. and Shi, Y. (1992) Antisense oligodeoxynucleotides as probes of T-lymphocyte gene function. Ann. N Y. Acad. Sci. 660: 193-203. Greene, L.A. (1978) Nerve growth factor prevents the death and stimulates the neuronal differentiation of clonal PC 12 pheochromocytoma cells in serum-free medium. J. Cell Biol. 78: 747-755. Greenlund, L.J., Deckwerth, T.L. and Johnson, E.M.Jr. (1995) Superoxide dismutase delays neuronal apoptosis: a role for reactive oxygen species in programmed neuronal death. iV^Mron 14: 303-315. Hall, S.E., Savill, J.S., Henson, P.M. and Haslett, C (1994) Apoptotic neutrophils are phagocytosed by fibroblasts with participation of the fibroblast vitronectin receptor and involvement of a mannose/fucose-specific lectin. J. Immunol. 153: 3218-3227. Hannun, Y.A. and Obeid, L.M. (1995) Ceramide: an intracellular signal for apoptosis. Trends Biochem. Sci. 20:13-11.

A. Messina and A. Jaworowski Hedgecock, E.M., Sulston, J.E. and Thomson, J.N. (1983) Mutations affecting programmed cell deaths in the nematode Caenorhabditis elegans. Science 220: 1277-1279. Hengartner, M.O., Ellis, R.E. and Horvitz, H.R. (1992) Caenorhabditis elegans gene ced-9 protects cells from programmed cell death. Nature 356: 494-499. Herrmann, C.P., Kraiss, S. and Montenarh, M. (1991) Association of casein kinase II with inmiunopurified p53. Oncogene 6: 877-884. Herron, L.R., Eisenberg, R.A., Kotzin, B.L., Roper, E., Kakkanaiah, V.N. and Cohen, P.L. (1993) Selection of the T cell receptor repertoire in Lpr mice. J. Immunol. 151: 34503459. Hilberg, R, Aguzzi, A., Howells, N. and Wagner, E.F. (1993) c-jun is essential for normal mouse development and hepatogenesis. Nature 365: 179-181. Hockenbery, D., Nunez, G., Milliman, C, Schreiber, R.D. and Korsmeyer, S.J. (1990) Bcl-2 is an inner mitochondrial membrane protein that blocks programmed cell death. Nature 348: 334-336. Hockenbery, D.M., Oltvai, Z.N., Yin, X.M., Milliman, C.L. and Korsmeyer, S.J. (1993) Bcl-2 functions in an antioxidant pathway to prevent apoptosis. Cell 75: 241-251. Hockenbery, D.M., Zutter, M., Hickey, W., Nahm, M. and Korsmeyer, S.J. (1991) BCL2 protein is topographically restricted in tissues characterized by apoptotic cell death. Proc. Natl. Acad. Sci. USA 88: 6961-6965. Huff, K., End, D. and Guroff, G. (1981) Nerve growth factorinduced alteration in the response of PC 12 pheochromocytoma cells to epidermal growth factor. J. Cell Biol. 88: 189198. Hug, H., Enari, M. and Nagata, S. (1994) No requirement of reactive oxygen intermediates in Fas-mediated apoptosis. FEES Lett. 351: 311-313. Itoh, N. and Nagata, S. (1993) A novel protein domain required for apoptosis. Mutational analysis of human Fas antigen./ Biol. Chem. 268: 10932-10937. Itoh, N., Tsujimoto, Y. and Nagata, S. (1993) Effect of bcl-2 on Fas antigen-mediated cell death. J. Immunol. 151: 621627. Itoh, N., Yonehara, S., Ishii, A., Yonehara, M., Mizushima, S., Sameshima, M., Hase, A., Seto, Y. and Nagata, S. (1991) The polypeptide encoded by the cDNA for human cell surface antigen Fas can mediate apoptosis. Cell 66: 233-243. Jacks, T., Remington, L., Williams, B.O., Schmitt, E.M., Halachmi, S., Bronson, R.T. and Weinberg, R.A. (1994) Tumor spectrum analysis in p53-mutant mice. Curr. Biol. 4: 1-7. Jackson, G.R., Apffel, L., Werrbach Perez, K. and Perez Polo, J.R. (1990) Role of nerve growth factor in oxidantantioxidant balance and neuronal injury. I. Stimulation of hydrogen peroxide resistance. J. Neurosci. Res. 25: 360368. Jacobson, M.D., Burne, J.F., King, M.P., Miyashita, T., Reed, J.C. and Raff, M.C. (1993) Bcl-2 blocks apoptosis in cells lacking mitochondrial DNA. Nature 361: 365-369.

115 Jacobson, M.D., Burne, J.F. and Raff, M.C. (1994) Programmed cell death and Bcl-2 protection in the absence of a nucleus. EM50 7. 13: 1899-1910. Jacobson, M.D. and Raff, M.C. (1995) Programmed cell death and Bcl-2 protection in very low oxygen. Nature 374: 814816. Jain, J., Nalefski, E.A., McCaffrey, P.G., Johnson, R.S., Spiegelman, B.M., Papaioannou, V. and Rao, A. (1994) Normal peripheral T-cell function in c-Fos-deficient mice. Mol. Cell Biol. 14: 1566-1574. Jarvis, W.D., Kolesnick, R.N., Fornari, F.A., Traylor, R.S., Gewirtz, D.A. and Grant, S. (1994) Induction of apoptotic DNA damage and cell death by activation of the sphingomyelin pathway. Proc. Natl. Acad. Sci. USA 91: 73-77. Jayadev, S., Liu, B., Bielawska, A.E., Lee, J.Y., Nazaire, F., Pushkareva, M.Y., Obeid, L.M. and Hannun, Y.A. (1995) Role for ceramide in cell cycle arrest. J. Biol. Chem. IIQ: 2047-2052. Ji, L., Zhang, G., Uematsu, S., Akahori, Y. and Hirabayashi, Y. (1995) Induction of apoptotic DNA fragmentation and cell death by natural ceramide. FEBS Lett. 358: 211-214. Jin, L.W., Inaba, K. and Saitoh, T. (1992) The involvement of protein kinase C in activation-induced cell death in T-cell hybridoma. Cell Immunol. 144: 217-227. Johnson, R.S., Spiegelman, B.M. and Papaioannou, V. (1992) Pleiotropic effects of a null mutation in the c-fos protooncogene. Cell 71: 577-586. Johnson, R.S., van Lingen, B., Papaioannou, V.E. and Spiegelman, B.M. (1993) A null mutation at the c-jun locus causes embryonic lethality and retarded cell growth in culture. Genes Dev. 7: 1309-1317. Jones, D.P., McConkey, D.J., Nicotera, P. and Orrenius, S. (1989) Calcium-activated DNA fragmentation in rat liver nuclei. J. Biol. Chem. 264: 6398-6403. Ju, S.T., Cui, H., Panka, D.J., Ettinger, R. and Marshak Rothstein, A. (1994) Participation of target Fas protein in apoptosis pathway induced by CD4+. Proc. Natl. Acad. Sci. USA 91:4185-4189. Ju, S.T., Panka, D.J., Cui, H., Ettinger, R., el Khatib, M., Sherr, D.H., Stanger, B.Z. and Marshak Rothstein, A. (1995) Fas(CD95)/FasL interactions required for programmed cell death after T-cell activation. Nature 373: 444-^48. Kagi, D., Vignaux, F., Ledermann, B., Burki, K., Depraetere, V. and Nagata, S. (1994) Fas and perforin pathways as major mechanisms of T cell-mediated. Science 265: 528530. Kane, D.J., Sarafian, T.A., Anton, R., Hahn, H., Gralla, E.B., Valentine, J.S., Ord, T. and Bredesen, D.E. (1993) Bcl-2 inhibition of neural death: decreased generation of reactive oxygen species. Science 262: 1274-1277. Kerr, J.F. (1965) A histochemical study of hypertrophy and ischaemic injury of rat liver with special reference to changes in lysosomes. J. Pathol. Bacteriol. 90: 419-435. Kerr, J.F. (1971) Shrinkage necrosis: a distinct mode of cellular death./. Pathol. 105: 13-20.

116 Kerr, J.F., Wyllie, A.H. and Currie, A.R. (1972) Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br. J. Cancer 26: 239-257. Kiefer, M.C., Brauer, M.J., Powers, V.C, Wu, J.J., Umansky, S.R., Tomei, D. and Ban, P.J. (1995) Modulation of apoptosis by the widely distributed Bcl-2 homologue Bak. Nature 374: 736-738. Kinoshita, T., Yokota, T., Arai, K. and Miyajima, A. (1995) Suppression of apoptotic death in hematopoietic cells by signalling through the IL-3/GM-CSF receptors. EMBO J. 14: 266-275. Kisielow, P., Bluthmann, H., Staerz, U.D., Steinmetz, M. and von Boehmer, H. (1988) Tolerance in T-cell-receptor transgenic mice involves deletion of nonmature. Nature 333: 742-746. Kisielow, P., Teh, H.S., Bluthmann, H. and von Boehmer, H. (1988) Positive selection of antigen-specific T cells in thymus by restricting MHC. Nature 335: 730-733. Kisielow, P. and von Boehmer, H. (1995) Development and selection of T-cells: facts and puzzles. Adv. Immunol. 58: 87-209. Komiyama, T., Ray, C.A., Pickup, D.J., Howard, A.D., Thornberry, N.A., Peterson, E.P. and Salvesen, G. (1994) Inhibition of interleukin-1 beta converting enzyme by the cowpox virus serpin CrmA. An example of cross-class inhibition./ Biol. Chem. 269: 19331-19337. Koopman, G., Keehnen, R.M., Lindhout, E., Newman, W., Shimizu, Y., van Seventer, G.A., de Groot, C. and Pals, S.T. (1994) Adhesion through the LFA-1 (GDIla/CD 18)ICAM-1 (CD54) and the VLA-4 (CD49d)-VCAM-l (CD 106) pathways prevents apoptosis of germinal center B cells. J. Immunol. 152: 3760-3767. Korsmeyer, S.J., Shutter, J.R., Veis, D.J., Merry, D.E. and Oltvai, Z.N. (1993) Bcl-2/Bax: a rheostat that regulates an anti-oxidant pathway and cell death. Semin. Cancer Biol. 4: 327-332. Kozopas, K.M., Yang, T., Buchan, H.L., Zhou, P. and Craig, R.W. (1993) MCLl, a gene expressed in programmed myeloid cell differentiation, has sequence similarity to BCL2. Proc. Natl. Acad. Sci. USA 90: 3516-3520. Krajewski, S., Krajewska, M., Shabaik, A., Miyashita, T., Wang, H.G. and Reed, J.C. (1994) Immunohistochemical determination of in vivo distribution of Bax, a dominant inhibitor of Bcl-2. Am. J. Pathol. 145: 1323-1336. Krajewski, S., Tanaka, S., Takayama, S., Schibler, M.J., Fenton, W. and Reed, J.C. (1993) Investigation of the subcellular distribution of the bcl-2 oncoprotein: residence in the nuclear envelope, endoplasmic reticulum and outer mitochondrial membranes. Cancer Res. 53: 47014714. Kumar, S. (1995) ICE-like proteases in apoptosis. Trends Biochem. Sci. 20: 198-202. Kumar, S., Kinoshita, M., Noda, M., Copeland, N.G. and Jenkins, N.A. (1994) Induction of apoptosis by the mouse Nedd2 gene, which encodes a protein similar to the product of the Caenorhabditis elegans cell death gene ced-3 and the

Mechanisms of developmental cell death mammalian IL-1 beta-converting enzyme. Genes Dev. 8: 1613-1626. Kumar, S., Tomooka, Y. and Noda, M. (1992) Identification of a set of genes with developmentally down-regulated expression in the mouse brain. Biochem. Biophys. Res. Commun. 185:1155-1161. Lagasse, E. and Weissman, I.L. (1994) bcl-2 inhibits apoptosis of neutrophils but not their engulfment by macrophages. J. Exp. Med. 179: 1047-1052. Lampe, P.A., Cornbrooks, E.B., Juhasz, A., Johnson, Jr., E.M. and Franklin, J.L. (1995) Suppression of programmed neuronal death by a thapsigargin-induced Ca^"*" influx. J. Neurohiol. 26: 205-212. Lazebnik, Y.A., Cole, S., Cooke, C.A., Nelson, W.G. and Earnshaw, W.C. (1993) Nuclear events of apoptosis in vitro in cell-free mitotic extracts: a model system for analysis of the active phase of apoptosis. J. Cell Biol. 123: 7-22. Lazebnik, Y.A., Kaufmann, S.H., Desnoyers, S., Poirier, G.G. and Earnshaw, W.C. (1994) Cleavage of poly(ADP-ribose) polymerase by a proteinase with properties like ICE. Nature 371:346-347. Lechner, M.S., Mack, D.H., Finicle, A.B., Crook, T., Vousden, K.H. and Laimins, L.A. (1992) Human papillomavirus E6 proteins bind p53 in vivo and abrogate p53-mediated repression of transcription. EMBO J. 11: 3045-3052. Lee, M.R., Liou, M.L., Yang, Y.F. and Lai, M.Z. (1993) cAMP analogs prevent activation-induced apoptosis of T cell hybridomas. J. Immunol. 151: 5208-5217. Levine, A.J., Momand, J. and Finlay, C.A. (1991) The p53 tumour suppressor gene. Nature 351: 453^56. Li, P., Allen, H., Banerjee, S., Franklin, S., Herzog, L., Johnston, C, McDowell, J., Paskind, M., Rodman, L., Salfeld, J., Towne, E., Tracey, D., Wardwell, S., Wei, F.-Y., Wong, W., Kamen, R. and Seshadri, T. (1995) Mice deficient in IL-1 beta-converting enzyme are defective in production of mature IL-1 beta and resistant to endotoxic shock. Cell 80: 401^11. Lin, E.Y., Orlofsky, A., Berger, M.S. and Prystowsky, M.B. (1993) Characterization of Al, a novel hemopoieticspecific early-response gene with sequence similarity to bcl-2. J. Immunol. 151: 1979-1988. Lindenboim, L., Haviv, R. and Stein, R. (1995) Inhibition of drug-induced apoptosis by survival factors in PC 12 cells. /. Neurochem. 64: 1054-1063. Littlewood, T.D., Amati, B., Land, H. and Evan, G.I. (1992) Max and c-Myc/Max DNA binding activities in cell extracts. Oncogene 7: 1783-1792. Liu, Z.G., Smith, S.W., McLaughlin, K.A., Schwartz, L.M. and Osborne, B.A. (1994) Apoptotic signals delivered through the T-cell receptor of a T-cell hybrid require the immediate-early gene nur77. Nature 367: 281-284. Los, M., Van der Craen, M., Penning, L.C., Schenk, H., Westendorp, M., Banuerle, P.A., Droge, W., Krammer, P.H., Rers, W. and Schulze Osthoff, K. (1995) Requirement of a ICE/CED-3 protease for Fas/Apo-1 mediated apoptosis. Namre 375: 81-83.

A. Messina and A. Jaworowski Lotem, J. and Sachs, L. (1993) Hematopoietic cells from mice deficient in wild-type p53 are more resistant to induction of apoptosis by some agents. Blood 82: 1092-1096. Lowe, S.W., Schmitt, E.M., Smith, S.W., Osborne, B.A. and Jacks, T. (1993) p53 is required for radiation-induced apoptosis in mouse thymocytes. Nature 362: 847-849. Lowin, B., Hahne, M., Mattmann, C. and Tschopp, J. (1994) Cytolytic T-cell cytotoxicity is mediated through perforin and Fas lytic. Nature 370: 650-652. Mah, S.P., Zhong, L.T., Liu, Y., Roghani, A., Edwards, R.H. and Bredesen, D.E. (1993) The protooncogene bcl-2 inhibits apoptosis in PC12 cells. J. Neurochem. 60: 1183-1186. Martin, D.P., Ito, A., Horigome, K., Lampe, P.A. and Johnson, E.M.J. (1992) Biochemical characterization of programmed cell death in NGF-deprived sympathetic neurons. J. Neurobiol. 23: 1205-1220. Martin, D.P., Schmidt, R.E., DiStefano, P.S., Lowry, O.H., Carter, J.G. and Johnson, E.M.J. (1988) Inhibitors of protein synthesis and RNA synthesis prevent neuronal death caused by nerve growth factor deprivation. J. Cell Biol. 106: 829-844. Martin, D., Wallace, T. and Johnson, Jr., E.M. (1990) Cytosine arabinoside kills post-mitotic neurons in a fashion resembling trophic factor deprivation: evidence that a deoxycytidine-dependent process may be required for nerve growth factor signal transduction. /. Neurosci. 10: 184-193. Martinou, I., Fernandez, P.A., Missotten, M., White, E., Allet, B., Sadoul, R. and Martinou, J.C. (1995) Viral proteins E1B19K and p35 protect sympathetic neurons from cell death induced by NGF deprivation. J. Cell Biol. 128: 201208. Mastino, A., Piacentini, M., Grelli, S., Favalli, C, Autuori, F., Tentori, L., Oliverio, S. and Garaci, E. (1992) Induction of apoptosis in thymocytes by prostaglandin E2 in vivo. Dev. Immunol. 2: 263-271. Mathias, S., Dressier, K.A. and Kolesnich, R.N. (1991) Characterization of a ceramide-activated protein kinase: stimulation by tumour necrosis factor alpha. Proc. Natl. Acad. Sci. USA 88: 10009-10013. Matsubara, K., Kubota, M., Adachi, S., Kuwakado, K., Hirota, H. and Wakazono, Y. (1994) Induction of apoptosis in childhood acute leukemia by chemotherapeutic. Eur. J. Haematol. 52: 47-52. Matter, A. (1979) Microcinematographic and electron microscopic analysis of target cell lysis induced by cytotoxic T lymphocytes. Immunology 36: 179-190. McConkey, D.J., Fosdick, L., D'Adamio, L., Jondal, M. and Orrenius, S. (1994) Co-receptor (CD4/CD8) engagement enhances CD3-induced apoptosis in thymocytes. Implications for negative selection. J. Immunol. 153: 2436-2443. McConkey, D.J., Hartzell, P. and Orrenius, S. (1990) Rapid turnover of endogenous endonuclease activity in thymocytes: effects of inhibitors of macromolecular synthesis. Arch. Biochem. Biophys. 278: 284-287. McConkey, D.J. and Orrenius, S. (1991) Cell,ular signalling in thymocyte apoptosis. In: L.D. Tomei and F.O. Cope

111 (Eds.), Apoptosis: The Molecular Basis of Cell Death, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 227-246. Meakin, S.O. and Shooter, E.M. (1992) The nerve growth factor family of receptors. Trends. Neurosci. 15: 323-331. Meikrantz, W., Gisselbrecht, S., Tam, S.W. and Schlegel, R. (1994) Activation of cyclin A-dependent protein kinases during apoptosis. Proc. Natl. Acad. Sci. USA 91: 37543758. Menissier de Murcia, J., Molinete, M., Gradwohl, G., Simonin, F. and de Murcia, G. (1989) Zinc-binding domain of poly (ADP-ribose) polymerase participates in the regulation of single strand breaks on DNA. J. Mol. Biol. 210: 229233. Meredith, J.E.J., Fazeli, B. and Schwartz, M.A. (1993) The extracellular matrix as a cell survival factor. Mol. Biol. Cell 4:953-961. Merry, D.E., Veis, D.J., Hickey, W.F. and Korsmeyer, S.J. (1994) bcl-2 protein expression is widespread in the developing nervous system and retained in the adult PNS. Development 120:301-311. Mesner, P.W., Winters, T.R. and Green, S.H. (1992) Nerve growth factor withdrawal-induced cell death in neuronal PC 12 cells resembles that in sympathetic neurons. J. Cell Biol. 119: 1669-1680. Messina, A. (1994) Expression of C-Jun protein during developmental cell death in sympathetic neurons. Proc. Aust. Physiol. Pharmacol. Soc, 25: 109P. Meuer, S.C, Hussey, R.E., Cantrell, D.A., Hodgdon, J.C, Schlossman, S.F., Smith, K.A. and Reinherz, E.L. (1984) Triggering of the T3-Ti antigen-receptor complex results in clonal T-cell proliferation through an interleukin 2dependent autocrine pathway. Proc. Natl. Acad. Sci. USA 81: 1509-1513. Miller, D.K., Calaycay, J.R., Chapman, K.T., Howard, A.D., Kostura, M.J., Molineaux, S.M. and Thornberry, N.A. (1994) The IL-IB converting enzyme as a therpeutic target. Ann. N. Y. Acad Sci. 744: 133-148. Miller, F.D., Mathew, T.C. and Toma, J.G. (1991) Regulation of nerve growth factor receptor gene expression by nerve growth factor in the developing peripheral nervous system. J. Cell Biol. 112:303-312. Milner, J., Cook, A. and Mason, J. (1990) p53 is associated with p34cdc2 in transformed cells. EMBO J. 9: 28852889. Miura, M., Zhu, H., Rotello, R., Hartwieg, E.A. and Yuan, J. (1993) Induction of apoptosis in fibroblasts by IL-1 betaconverting enzyme, a mammalian homolog of the C. elegans cell death gene ced-. Cell 75: 653-660. Miyashita, T. and Reed, J.C. (1995) Tumor suppressor p53 is a direct transcriptional activator of the human bax gene. Cell 80: 293-299. Moberg, K.H., Tyndall, W.A. and Hall, D.J. (1992) Wild-type murine p53 represses transcription from the murine c-myc promoter in a human glial cell line. J. Cell Biochem., 49: 208-215.

118 Momand, J., Zambetti, G.P., Olson, D.C., George, D. and Levine, A.J. (1992) The mdm-2 oncogene product forms a complex with the p53 protein and inhibits p53-mediated transactivation. Cell 69: 1237-1245. Morgan, D.O. (1995) Principles of CDK regulation. Nature 374: 131-134. Nagata, S. and Golstein, P. (1995) The Fas death factor. Science 267: 1449-1456. Newmeyer, D.D., Farschon, D.M. and Reed, J.C. (1994) Cell,free apoptosis in Xenopus egg extracts: inhibition by Bcl-2 and requirement for an organelle fraction enriched in mitochondria. Cell 79: 353-364. Ni, R., Tomita, Y., Matsuda, K., Ichihara, A., Ishimura, K., Ogasawara, J. and Nagata, S. (1994) Fas-mediated apoptosis in primary cultured mouse hepatocytes. Exp. Cell Res. 215: 332-337. Nikonova, L.V., Nelipovich, P.A. and Umansky, S.R. (1982) The involvement of nuclear nucleases in rat thymocyte DNA degradation after gamma-irradiation. Biochim. Biophys. Acta 699\2U-2%9. Nishioka, W.K. and Welsh, R.M. (1994) Susceptibility to cytotoxic T lymphocyte-induced apoptosis is a function of the proliferative status of the target. J. Exp. Med. 179: 769774. Nunez, G., London, L., Hockenbery, D., Alexander, M., McKearn, J.P. and Korsmeyer, S.J. (1990) Deregulated Bcl2 gene expression selectively prolongs survival of growth factor-deprived hemopoietic cell lines. J. Immunol. 144: 3602-3610. Obeid, L.M., Linardic, CM., Karolak, L.A. and Hannun, Y.A. (1993) Programmed cell death induced by ceramide. Science 259: 1769-1771. Oberhammer, F., Fritsch, G., Schmied, M., Pavelka, M., Printz, D., Purchio, T., Lassmann, H. and Schulte Hermann, R. (1993a) Condensation of the chromatin at the membrane of an apoptotic nucleus is not associated with activation of an endonuclease. J. Cell Sci. 104: 317-326. Oberhammer, F., Wilson, J.W., Dive, C, Morris, I.D., Hickman, J.A., Wakeling, A.E., Walker, P.R. and Sikorska, M. (1993b) Apoptotic death in epithelial cells: cleavage of DNA to 300 and/or 50 kb fragments prior to or in the absence of internucleosomal fragmentation. EMBO J. 12: 3679-3684. Oehm, A., Behrmann, I., Falk, W., Pawlita, M., Maier, G., Klas, C, Li Weber, M., Richards, S., Dhein, J., Trauth, B.C., Postingl, H. and Krammer, P.H. (1992) Purification and molecular cloning of the APO-1 cell surface antigen, a member of the tumor necrosis factor/nerve growth factor receptor superfamily. Sequence identity with the Fas antigen. J. Biol. Chem. 267: 10709-10715. Ohta, H., Yatomi, Y., Sweeney, E.A., Hakomori, S. and Igarashi, Y. (1994) A possible role of sphingosine in induction of apoptosis by tumor necrosis factor-alpha in human neutrophils. FEBSLett. 355: 267-270. Okazaki, T., Bielawska, A., Domae, N., Bell, R.M. and Hannun, Y.A. (1994) Characteristics and partial purification of

Mechanisms of developmental cell death a novel cytosolic, magnesium-independent, neutral sphingomyelinase activated in the early signal transduction of 1 alpha,25-dihydroxyvitamin D3-induced HL-60 cell differentiation. J. Biol. Chem. 269: 4070-4077. Oliner, J.D. (1993) Discerning the function of p53 by examining its molecular interactions. BioEssays 15: 703-707. Oliner, J.D., Pietenpol, J.A., Thiagalingam, S., Gyuris, J., Kinzler, K.W. and Vogelstein, B. (1993) Oncoprotein MDM2 conceals the activation domain of tumour suppressor p53. Nature 362: 857-860. Oltvai, Z.N. and Korsmeyer, S.J. (1994) Checkpoints of dueling dimers foil death wishes. Cell 79: 189-192. Oltvai, Z.N., Milliman, C.L. and Korsmeyer, S.J. (1993) Bcl-2 heterodimerizes in vivo with a conserved homolog, Bax, that accelerates programmed cell death. Cell 74: 609-619. Oppenheim, R.W. (1991) Cell, death during development of the nervous system. Annu. Rev. Neurosci. 14: 453-501. Oppenheim, R.W., Prevette, D., Tytell, M. and Homma, S. (1990) Naturally occurring and induced neuronal death in the chick embryo in vivo requires protein and RNA synthesis: evidence for the role of cell death genes. Dev. Biol. 138:104-113. Ostergaard, H.L. and Clark, W.R. (1989) Evidence for multiple pathways used by cytolytic T-lymphocytes. J. Immunol. 143: 2120-2123. Owen Schaub, L.B., Yonehara, S., Crump, W.L. and Grimm, E.A. (1992) DNA fragmentation and cell death is selectively triggered in activated human lymphocytes by Fas antigen engagement. Cell Immunol. 140: 197-205. Peitsch, M.C., Polzar, B., Stephan, H., Crompton, T., MacDonald, H.R., Mannherz, H.G. and Tschopp, J. (1993) Characterization of the endogenous deoxyribonuclease involved in nuclear DNA degradation during apoptosis (programmed cell death). EMBO J. 12: 371-377. Piacentini, M., Autuori, F., Dini, L., Farraee, M.G., Ghibelli, L., Piredda, L. and Fesus, L. (1991) 'Tissue' transglutaminase is specifically expressed in neonatal rat liver cells undergoing apoptosis upon epidermal growth factorstimulation. Cell Tissue Res. 263: 227-235. Rabizadeh, S. and Bredesen, D.E. (1994) Is p75N^^ involved in developmental neural cell death. Dev. Neurosci. 16: 207211. Rabizadeh, S., LaCount, D.J., Friesen, P.D. and Bredesen, D.E. (1993a Expression of the baculovirus p35 gene inhibits mammalian neural cell death. J. Neurochem. 61: 23182321. Rabizadeh, S., Oh, J., Zhong, L.T., Yang, J., Bitler, CM., Butcher, L.L. and Bredesen, D.E. (1993b) Induction of apoptosis by the low-affinity NGF receptor. Science 261: 345-348. Raff, M.C (1992) Social controls on cell survival and cell death. Nature 356: 397-400. Raff, M.C, Barres, B.A., Burne, J.F., Coles, H.S., Ishizaki, Y. and Jacobson, M.D. (1993) Programmed cell death and the control of cell survival: lessons from the nervous system. Science 262: 695-700.

A. Messina and A. Jaworowski Rao, L., Debbas, M., Sabbatini, P., Hockenbery, D., Korsmeyer, S. and White, E. (1992) The adenovirus ElA proteins induce apoptosis, which is inhibited by the ElB 19-kDa and Bcl-2 proteins. Proc. Natl. Acad. Sci. USA 89: 7742-7746. Ray, C.A., Black, R.A., Kronheim, S.R., Greenstreet, T.A., Sleath, P.R., Salvesen, G.S. and Pickup, D.J. (1992) Viral inhibition of inflammation: cowpox virus encodes an inhibitor of the interleukin-1 beta converting enzyme. Cell 69: 597-604. Re, F., Zanetti, A., Sironi, M., Polentarutti, N., Lanfrancone, L., Dejana, E. and Colotta, F. (1994) Inhibition of anchorage-dependent cell spreading triggers apoptosis in cultured human endothelial cells. J. Cell Biol. 127: 537-546. Reynolds, J.E., Yang, T., Qian, L., Jenkinson, J.D., Zhou, P., Eastman, A. and Craig, R.W. (1994) Mcl-l, a member of the Bcl-2 family, delays apoptosis induced by c-Myc overexpression in Chinese hamster ovary cells. Cancer Res. 54: 6348-6352. Riegel, J.S., Richie, E.R. and Allison, J.P. (1990). Nuclear events after activation of CD4^CD8'*" thymocytes. J. Immunol. 144:3611-3618. Rouvier, E., Luciani, M.F. and Golstein, P. (1993) Fas involvement in Ca(2+)-independent T cell-mediated cytotoxicity. J. Exp. Med. \11\ 195-200. Rukenstein, A., Rydel, R.E. and Greene, L.A. (1991) Multiple agents rescue PC 12 cells from serum-free cell death by translation- and transcription-independent mechanisms. J. Neurosci. 11:2552-2563. Russell, J.H., Rush, B., Weaver, C. and Wang, R. (1993) Mature T cells of autoimmune Ipr/lpr mice have a defect in antigen-stimulated suicide. Proc. Natl. Acad. Sci. USA 90: 4409-4413. Russell, J.H. and Wang, R. (1993) Autoimmune gld mutation uncouples suicide and cytokine/proliferation pathways in activated, mature T cells. Eur. J. Immunol. 23: 2379-2382. Russell, S.W., Rosenau, W. and Lee, J.C. (1972) Cytolysis induced by human lymphotoxin. Am. J. Pathol. 69: 103118. Rydel, R.E. and Greene, L.A. (1988) cAMP analogs promote survival and neurite outgrowth in cultures of rat sympathetic and sensory neurons independently of nerve growth factor. Proc. Natl. Acad. Sci. USA 85: 1257-1261. Sabbatini, P., Chiou, S.K., Rao, L. and White, E. (1995) Modulation of p53-mediated transcriptional repression and apoptosis by the adenovirus ElB 19K protein. Mol. Cell Biol. 15: 1060-1070. Sanderson, C.J. (1976) The mechanism of T cell mediated cytotoxicity. II. Morphological studies of cell death by time-lapse microcinematography. Proc. R. Soc. London B. Biol. Sci. 192: 241-255. Santhanam, U., Ray, A. and Sehgal, P.B. (1991) Repression of the interleukin 6 gene promoter by p53 and the retinoblastoma susceptibility gene product. Proc. Natl. Acad. Sci. USA 88: 7605-7609. Sarnow, P., Ho, Y.S., Williams, J. and Levine, A.J. (1982)

119 Adenovirus Elb-58 kd tumor antigen and SV40 large tumor antigen are physically associated with the same 54 kd cellular protein in transformed cells. Cell 28: 387-394. Sato, N., Sakamaki, K., Terada, N., Arai, K. and Miyajima, A. (1993) Signal transduction by the high-affinity GM-CSF receptor: two distinct cytoplasmic regions of the common beta subunit responsible for different signaling. EMBO J. 12:4181-4189. Sato, T., Hanada, M., Bodrug, S., Irie, S., Iwama, N., Boise, L.H., Thompson, C.B., Golemis, E., Fong, L., Wang, H.G. et al. (1994) Interactions among members of the Bcl-2 protein family analyzed with a yeast two-hybrid system. Proc. Natl. Acad. Sci. USA 91: 9238-9242. Saunders, J.W.J. (1966) Death in embryonic systems. Science 154: 604-612. Savill, J., Dransfield, I., Hogg, N. and Haslett, C. (1990) Vitronectin receptor-mediated phagocytosis of cells undergoing apoptosis. Nature 343: 170-173. Savill, J., Fadok, V., Henson, P. and Haslett, C. (1993) Phagocyte recognition of cells undergoing apoptosis. Immunol. Today 14: 131-136. Savill, J.S., Henson, P.M. and Haslett, C. (1989) Phagocytosis of aged human neutrophils by macrophages is mediated by a novel 'charge-sensitive' recognition mechanism. J. Clin. Invest. 84: 1518-1527. Savill, J.S., Wyllie, A.H., Henson, J.E., Walport, M.J., Henson, P.M. and Haslett, C. (1989) Macrophage phagocytosis of aging neutrophils in inflammation. Programmed cell death in the neutrophil leads to its recognition by macrophages. J. Clin. Invest. 83: 865-875. Schulze Osthoff, K., Walczak, H., Droge, W. and Krammer, P.H. (1994) Cell, nucleus and DNA fragmentation are not required for apoptosis. 7. Cell Biol. Ill: 15-20. Schutze, S., Machleidt, T. and Kronke, M. (1994) The role of diacylglycerol and ceramide in tumor necrosis factor and interleukin-1 signal transduction. J. Leukoc. Biol. 56: 533541. Scott, S.A. and Davies, A.M. (1990) Inhibition of protein synthesis prevents cell death in sensory and parasympathetic neurons deprived of neurotrophic factor in vitro. J. Neurobiol. 21:630-638. Sellins, K.S. and Cohen, J.J. (1991) Cytotoxic T lymphocytes induce different types of DNA damage in target cells of different origins. J. Immunol. 147: 795-803. Selvakumaran, M., Lin, H.K., Miyashita, T., Wang, H.G., Krajewski, S., Reed, J.C, Hoffman, B. and Liebermann, D. (1994) Immediate early up-regulation of bax expression by p53 but not TGF beta 1: a paradigm for distinct apoptotic pathways. Oncogene 9: 1791-1798. Sentman, C.L., Shutter, J.R., Hockenbery, D., Kanagawa, O. and Korsmeyer, S.J. (1991) bcl-2 inhibits multiple forms of apoptosis but not negative selection in thymocytes. Cell 67: 879-888. Sha, W.C, Nelson, C.A., Newberry, R.D., Kranz, D.M., Russell, J.H. and Loh, D.Y. (1988) Positive and negative selection of an antigen receptor on T cells in. Nature 336: 73-76.

120 Shenk, T. and Flint, J. (1991) Transcriptional and transforming activities of the adenovirus El A proteins. Adv. Cancer Res. 57: 47-85. Shi, L., Kraut, R.P., Aebersold, R. and Greenberg, A.H. (1992) A natural killer cell granule protein that induces DNA fragmentation and apoptosis. J. Exp. Med. 175: 5 5 3 566. Shi, L., Nishioka, W.K., Th'ng, J., Bradbury, E.M., Litchfield, D.W. and Greenberg, A.H. (1994) Premature p34cdc2 activation required for apoptosis. Science 263: 1143-1145. Shi, Y., Glynn, J.M., Guilbert, L.J., Cotter, T.G., Bissonnette, R.P. and Green, D.R. (1992) Role for c-myc in activationinduced apoptotic cell death in T cell hybridomas. Science 257: 212-214. Shi, Y.F., Sahai, B.M. and Green, D.R. (1989) Cyclosporin A inhibits activation-induced cell death in T-cell hybridomas and thymocytes. Nature 339: 625-626. Shiio, Y., Yamamoto, T. and Yamaguchi, N. (1992) Negative regulation of Rb expression by the p53 gene product. Proc. Natl Acad. Sci. USA 89: 5206-5210. Shimizu, S., Eguchi, Y., Kosaka, H., Kamiike, W., Masuda, H. and Tsujimoto, Y. (1995) Prevention of hypoxia-induced cell death by Bcl-2 and BCI-XL- Nature 374: 811-813. Sidman, C.L., Marshall, H. and von Boehmer, H. (1992) Transgenic T cell receptor interactions in the lymphoproliferative and autoimmune syndromes of Ipr and gld mutant mice. Eur. J. Immunol. 22: 499-504. Siegel, R.M., Katsumata, M., Miyashita, T., Louie, D.C., Greene, M.I. and Reed, J.C. (1992) Inhibition of thymocyte apoptosis and negative antigenic selection in bcl-2 transgenic mice. Proc. Natl. Acad. Sci. USA 89: 7003-7007. Smeyne, R.J., Schilling, K., Oberdick, J., Robertson, L., Luk, D., Curran, T. and Morgan, J.I. (1993) A fos-lac Z transgenic mouse that can be used for neuroanatomic mapping. A^v.A^^Mro/. 59:285-291. Smith, C.A., Williams, G.T., Kingston, R., Jenkinson, E.J. and Owen, J.J. (1989) Antibodies to CD3/T-cell receptor complex induce death by apoptosis in immature T cells in thymic cultures. Nature 337: 181-184. Strasser, A., Harris, A.W. and Cory, S. (1991) bcl-2 transgene inhibits T cell death and perturbs thymic self-censorship. Cell 67: 889-899. Strasser, A., Harris, A.W., Jacks, T. and Cory, S. (1994) DNA damage can induce apoptosis in proliferating lymphoid cells via p53-independent mechanisms inhibitable by Bcl-2. Cell 79: 329-339. Sturzbecher, H.W., Maimets, T., Chumakov, P., Brain, R., Addison, C , Simanis, V., Rudge, K., Philp, R., Grimaldi, M., Court, W. and Jenkins, J.R. (1990) p53 interacts with p34cdc2 in mammalian cells: implications for cell cycle control and oncogenesis. Oncogene 5: 795-781. Subler, M.A., Martin, D.W. and Deb, S. (1992) Inhibition of viral and cellular promoters by human wild-type p53. 7. Virol. 66: 4757-4762. Suda, T., Okazaki, T., Naito, Y., Yokoto, T., Arai, N., Ozaki, S., Nakao, K. and Nagata, S. (1995) Expression of the Fas

Mechanisms of developmental cell death ligand in cells of T-cell lineage. J. Immunol. 154: 38063813. Suda, T., Takahashi, T., Golstein, P. and Nagata, S. (1993) Molecular cloning and expression of the Fas ligand, a novel member of the tumor necrosis factor family. Cell 75: 11691178. Sulston, J.E. and Horvitz, H.R. (1977) Post-embryonic cell lineages of the nematode, Caenorhabditis elegans. Dev. Biol. 56: 110-156. Sulston, J.E., Schierenberg, E., White, J.G. and Thomson, J.N. (1983) The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev. Biol. 100: 64-119. Sumimoto, S., Ishigami, T., Horiguchi, Y., Yonehara, S., Kanazashi, S., Heike, T., Katamura, K. and Mayumi, M. (1994) Anti-Fas antibody induces different types of cell death in the human histiocytic cell line, U937 and the human B cell line, B104: the role of single-strand DNA breaks and poly (ADP-ribosyl)ation in cell death. Cell Immunol. 153:184-193. Sun, D.Y., Jiang, S., Zheng, L.M., Ojcius, D.M. and Young, J.D. (1994) Separate metabolic pathways leading to DNA fragmentation and apoptotic chromatin condensation. J. Exp. Med. 179: 559-568. Surh, C D . and Sprent, J. (1994) T-cell apoptosis detected in situ during positive and negative selection in the thymus. Nature 372: 100-103. Swat, W., Ignatowicz, L., von Boehmer, H. and Kisielow, P. (1991) Clonal deletion of immature CD4+8+ thymocytes in suspension culture by extrathymic antigen-presenting cells. Nature 351: 150-153. Takayama, S., Sato, T., Krajewski, S., Kochel, K., Irie, S., Millan, J.A. and Reed, J.C. (1995) Cloning and functional analysis of BAG-1: a novel Bcl-2-binding protein with anticell death activity. Cell 80: 279-284. Tanaka, M., Suda, T., Takahashi, T. and Nagata, S. (1995) Expression of the functional soluble form of human Fas ligand in activated lymphocytes. EMBO J. 14: 1129-1135. Tanaka, S., Saito, K. and Reed, J.C. (1993) Structure-function analysis of the Bcl-2 oncoprotein. Addition of a heterologous transmembrane domain to portions of the Bcl-2 beta protein restores function as a regulator of cell survival. J. Biol. Chem. 268: 10920-10926. Tartaglia, L.A., Ayres, T.M., Wong, G.H. and Goeddel, D.V. (1993) A novel domain within the 55 kd TNF receptor signals cell death. Cell 74: 845-853. Tewari, M. and Dixit, V.M. (1995) Fas- and tumor necrosis factor-induced apoptosis is inhibited by the poxvirus crmA gene product. J. Biol. Chem. 270: 3255-3260. Thornberry, N.A., Bull, H.G., Calaycay, J.R., Chapman, K.T., Howard, A.D., Kostura, M.J., Miller, D.K., Mohneaux, S.M., Weidner, J.R., Aunins, J., Elliston, K.O., Ayala, J.M., Casano, F.J., Chin, J., Ding, G.J.-F., Egger, L.A., Gaffney, E.P., Limjuco, G., Palyha, O.C, Raju, S.M., Rolando, A.M., Salley, J.P., Yamin, T.-T., Lee, T.D., Shivley, J.E., MacCross, M., Mumford, R.A., Schmidt, J.A. and Tocci, M.J. (1992) A novel heterodimeric cysteine protease is required

A. Messina and A. Jaworowski for interleukin-1 beta processing in monocytes. Nature 356: 768-774. Tilly, J.L., Tilly, K.L, Kenton, M.L. and Johnson, A.L. (1995) Expression of members of the bcl-2 gene family in the immature rat ovary: equine chorionic gonadotropin-mediated inhibition of granulosa cell apoptosis is associated with decreased bax and constitutive bcl-2 and bcl-xlong messenger ribonucleic acid levels. Endocrinology 136: 232241. Tomkins, C.E., Edwards, S.N. and Tolkovsky, A.M. (1994) Apoptosis is induced in post-mitotic rat sympathetic neurons by arabinosides and topoisomerase II inhibitors in the presence of NGF. 7. Cell Sci. 107: 1499-1507. Trauth, B.C., Klas, C, Peters, A.M., Matzku, S., Moller, P., Falk, W., Debatin, K.M. and Kranmier, P.H. (1989) Monoclonal antibody-mediated tumor regression by induction of apoptosis. Science 245: 301-305. Troy, CM. and Shelanski, M.L. (1994) Down-regulation of copper/zinc superoxide dismutase causes apoptotic death in PC12 neuronal cells. Proc. Nad. Acad. Sci. USA 91: 63846387. Trump, B.F. and Berezesky, I.K. (1992) The role of cytosolic Ca2+ in cell injury, necrosis and apoptosis. Curr. Opin. Cell Biol. 4: 227-232. Tsujimoto, Y., Cossman, J., Jaffe, E. and Croce, CM. (1985) Involvement of the bcl-2 gene in human follicular lymphoma. Science 228: 1440-1443. Ucker, D.S., Ashwell, J.D. and Nickas, G. (1989) Activationdriven T cell death. I. Requirements for de novo transcription and translation and association with genome fragmentation. /. Immunol. 143: 3461-3469. Ucker, D.S., Obermiller, P.S., Eckhart, W., Apgar, J.R., Berger, N.A. and Meyers, J. (1992) Genome digestion is a dispensable consequence of physiological cell death mediated by cytotoxic T lymphocytes. Mol. Cell Biol. 12: 30603069. Vaux, D.L., Aguila, H.L. and Weissman, I.L. (1992) Bcl-2 prevents death of factor-deprived cells but fails to prevent apoptosis in targets of cell mediated killing. Int. Immunol. 4: 821-824. Vaux, D.L., Cory, S. and Adams, J.M. (1988) Bcl-2 gene promotes haemopoietic cell survival and cooperates with cmyc to immortalize pre-B cells. Nature 335: 440-442. Vaux, D.L. and Weissman, I.L. (1993) Neither macromolecular synthesis nor myc is required for cell death via the mechanism that can be controlled by Bcl-2. Mol. Cell Biol. 13: 7000-7005. Vaux, D.L., Weissman, I.L. and Kim, S.K. (1992) Prevention of programmed cell death in Caenorhabditis elegans by human bcl-2. Science 258: 1955-1957. Veis, D.J., Sentman, C.L., Bach, E.A. and Korsmeyer, S.J. (1993) Expression of the Bcl-2 protein in murine and human thymocytes and in peripheral T lymphocytes. J. Immunol. 151:2546-2554. Veis, D.J., Sorenson, CM., Shutter, J.R. and Korsmeyer, S.J. (1993) Bcl-2-deficient mice demonstrate fulminant lym-

121 phoid apoptosis, polycystic kidneys and hypopigmented hair. Ce//75: 229-240. Vignaux, P., Vivier, E., Malissen, B., Depraetere, V., Nagata, S. and Golstein, P. (1995) TCR/CD3 coupling to Fas-based cytotoxicity. J. Exp. Med. 181: 781-786. Wagner, A.J., Small, M.B. and Hay, N. (1993) Myc-mediated apoptosis is blocked by ectopic expression of Bcl-2. Mol. Cell Biol. 13:2432-2440. Walker, N.P., Talanian, R.V., Brady, K.D., Dang, L.C, Bump, N.J., Ferenz, C.R., Franklin, S., Ghayur, T., Hackett, M.C, Hanmiill, L.D., Herzog, L., Hugunin, M., Houy, W., Mankovich, J.A., McGuiness, L., Orlewicz, E., Paskind, M., Pratt, C.A., Reis, P., Summani, A., Terranova, M., Welch, J.P., Xiong, L., Moller, A., Tracey, D.E., Kamen, R. and Wong, W.W. (1994) Crystal structure of the cysteine protease interleukin-1 beta-converting enzyme: a (p20/pl0)2 homodimer. Cell 78: 343-352. Walker, P.R., Kwast Welfeld, J., Gourdeau, H., Leblanc, J., Neugebauer, W. and Sikorska, M. (1993) Relationship between apoptosis and the cell cycle in lymphocytes: roles of protein kinase C, tyrosine phosphorylation and API. Exp. Cell Res. 201: 142-151. Wang, H.G., Miyashita, T., Takayama, S., Sato, T., Torigoe, T., Krajewski, S., Tanaka, S., Hovey, L., Troppmair, J. and Rapp, U.R. (1994a) Apoptosis regulation by interaction of Bcl-2 protein and Raf-1 kinase. Oncogene 9: 27512756. Wang, L., Miura, M., Bergeron, L., Zhu, H. and Yuan, J. (1994b) Ich-1, an Ice/ced-3-related gene, encodes both positive and negative regulators of programmed cell death. Cell 78: 739-750. Wang, Z.Q., Ovitt, C, Grigoriadis, A.E., Mohle Steinlein, U., Ruther, U. and Wagner, E.F. (1992) Bone and haematopoietic defects in mice lacking c-fos. Nature 360: 741745. White, E., Sabbatini, P., Debbas, M., Wold, W.S., Kusher, D.I. and Gooding, L.R. (1992) The 19-kilodalton adenovirus ElB transforming protein inhibits programmed cell death and prevents cytolysis by tumor necrosis factor alpha. Mol. Cell Biol. 12: 2570-2580. Whittemore, E.R., Loo, D.T. and Cotman, CW. (1994) Exposure to hydrogen peroxide induces cell death via apoptosis in cultured rat cortical neurons. NeuroReport 5: 1485-1488. Whyte, P., Buchkovich, K.J., Horowitz, J.M., Friend, S.H., Raybuck, M., Weinberg, R.A. and Harlow, E. (1988) Association between an oncogene and an anti-oncogene: the adenovirus El A proteins bind to the retinoblastoma gene product. Nature 334: 124-129. Williams, G.T. and Smith, CA. (1993) Molecular regulation of apoptosis: genetic controls on cell death. Cell 74: 777779. Wilson, K.P., Black, J.A., Thomson, J.A., Kim, E.E., Griffith, J.P., Navia, M.A., Murcko, M.A., Chambers, S.P., Aldape, R.A., Raybuck, S.A. and Livingstone, D.J. (1994) Structure and mechanism of interleukin-1 beta converting enzyme. Nature 370: 270-275.

122 Winoto, A. (1994) Molecular characterization of the Nur77 orphan steroid receptor in apoptosis. InU Arch, Allergy Immunol. 105: 344-346. Woronicz, J.D., Calnan, B., Ngo, V. and Winoto, A. (1994) Requirement for the orphan steroid receptor Nur77 in apoptosis of T-cell hybridomas. Nature 367: 277-281. Wu, X., Bayle, J.H., Olson, D. and Levine, A.J. (1993) The p53-mdm-2 autoregulatory feedback loop. Genes Dev. 7: 1126-1132. Wyllie, A.H. (1980) Glucocorticoid-induced thymocyte apoptosis is associated with endogenous endonuclease activation. Nature 284: 555-556. Wyllie, A.H., Beattie, G.J. and Hargreaves, A.D. (1981) Chromatin changes in apoptosis. Histochem. J. 13: 681692. Wyllie, A.H., Morris, R.G., Smith, A.L. and Dunlop, D. (1984) Chromatin cleavage in apoptosis: association with condensed chromatin morphology and dependence on macromolecular synthesis. J. Pathol. 142: 67-77. Yang, E., Zha, J., Jockel, J., Boise, L.H., Thompson, C.B. and Korsmeyer, S.J. (1995) Bad, a heterodimeric partner for Bcl-XL and Bcl-2, displaces Bax and promotes cell death. C^//80: 285-291. Yao, R. and Cooper, G.M. (1995) Requirement for phosphayidylinositol-3 kinase in the prevention of apotosis by nerve growth factor. Science 267: 2003-2006.

Mechanisms of developmental cell death Yazdanbakhsh, K., Choi, J.W., Li, Y., Lau, L.F. and Choi, Y. (1995) Cyclosporin A blocks apoptosis by inhibiting the DNA binding activity of the transcription factor Nur77. Proc. Natl. Acad. ScL USA 92: 437-441. Yew, P.R. and Berk, A.J. (1992) Inhibition of p53 transactivation required for transformation by adenovirus ElB 55KDa protein. Nature 357: 82-85. Yin, X.M., Oltval, Z.N. and Korsmeyer, S.J. (1994) BHl and BH2 domains of Bcl-2 are required for inhibition of apoptosis and heterodimerization with Bax. Nature 369: 321323. Yonehara, S., Ishii, A. and Yonehara, M. (1989) A cell-killing monoclonal antibody (anti-Fas) to a cell surface antigen codownregulated with the receptor of tumor necrosis factor. J. Exp. Med. 169: 1747-1756. Yuan, J. and Horvitz, H.R. (1992) The Caenorhabditis elegans cell death gene ced-4 encodes a novel protein and is expressed during the period of extensive programmed cell death. Development 116: 309-320. Yuan, J., Shaham, S., Ledoux, S., Ellis, H.M. and Horvitz, H.R. (1993) The C. elegans cell death gene ced-3 encodes a protein similar to mammalian interleukin-1 beta-converting enzyme. Cell 75: 641-652. Yuan, J.Y. and Horvitz, H.R. (1990) The Caenorhabditis elegans genes ced-3 and ced-4 act cell autonomously to cause programmed cell death. Dev. Biol. 138: 33-41.

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 5

Regulation of the early development of the nervous system by growth factors Perry F. Bartlett, Trevor J. Kilpatrick, Linda J. Richards, Paul S. Talman and Mark Murphy The Walter and Eliza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Parkville, Victoria 3050, Australia

1. Introduction

2. Early central nervous system development

Underlying early mammalian neural development are complex interactions which include intercellular signalling and differential gene expression. These influences are exerted upon a precursor cell population which can become committed to either the neuronal or glial lineages in the central nervous system (CNS), or to neurons, gha, melanocytes and numerous other mesenchymal cells in the peripheral nervous system (PNS). The lineage potential of individual precursor cells, however, remains the subject of intensive study, as does the relative influences which intrinsic commitments and epigenetic factors, such as growth factors, might play in both lineage determination and in instructing regionally specific development. Furthermore, whereas conventional wisdom suggests that it is the depletion of precursor cells which accounts for the cessation of neurogenesis in late development, recent studies indicate that neuronal precursor cells may persist in late embryogenesis and even within the adult mammalian brain. This implies that it is a change in the factor regulation of precursors, rather than a depletion of precursor cells per se, which is responsible for the failure to generate neurones in the adult brain. This review explores these issues, with particular emphasis on the role growth factors play in converting an apparently homogeneous population of precursor cells into the complex cellular and structurally diverse tissues of the adult mammalian nervous system.

In the developing embryo, the primordial neural tissue is first recognisable as a pseudostratified, columnar layer of precursor cells known as the neuroepithelium, which forms the neural tube and generates the entire nervous system (Theiler, 1972). Around the time of neural-tube closure, the dorsal segment of the tube gives rise to a migrating, cellular population, the neural crest, which generates most of the PNS and a variety of other cell types, including mesenchymal tissue, endocrine and melanocyte derivatives. The neural tube also generates the CNS which is comprised of two major lineages, neurons and glia, the latter being further subdivided into two principal cellular types, namely astrocytes and oligodendrocytes. The initial phase of neural development is characterized by cellular proliferation (Cowan, 1979). As cell numbers increase, the epithelium thickens becoming separated into a ventricular zone and a less well defined and transient outer region, the subventricular zone, which becomes populated by committed precursor cells. In mammalian development, lineage commitment of CNS neural precursor cells is first apparent at around midgestation (Bailey, 1987). For instance, in the rat, the first terminally differentiated (i.e. post-mitotic) neurons develop at about embryonic day 11 (Ell) (Nornes and Das, 1974). This, however, does not necessarily imply that uncommitted precursor cells are thereafter depleted from the neural anlage nor that committed precursor cells have necessarily

124

Regulation of the early development of the nervous system by growth factors

lost the ability to proliferate. In fact, in the rat cerebral cortex, neurogenesis, as determined by the proliferation and subsequent differentiation of committed, neuronal precursors, is well documented to continue until around E20 (Angevine and Sidman, 1961; Frederiksen and McKay, 1988). On the other hand, gliogenesis, as determined by the generation of glial fibrillary acidic protein (GFAP)-positive astrocytes, has been reported to commence at El6 in the rat (Abney et al., 1981) and continues well into postnatal development (Miller et al., 1985; Frederiksen and McKay, 1988). However, at least in some species, there may not be a temporal dispersion between the phases of neurogenesis and gliogenesis. For instance, in the monkey a specialized subset of glial cells, known as radial glia, are generated at the time of neurogenesis (Schmechel and Rakic, 1979) and it has been proposed that mature astrocytes may be derived from these cells (Schmechel and Rakic, 1979; Culican et al., 1990). Alternatively, astrocytes may be derived from morphologically undifferentiated precursor cells which nevertheless, are committed to the glial lineage from the earliest phase of neural development. 2.7. The concept of multipotentiality In order to study the interplay between intrinsic and environmental (growth factor) regulation of precursor cells within the developing nervous sytem, it is important to establish whether individual neuroepithelial cells retain the capacity to differentiate into both neurons and glia; in other words, to what extent do regulatory processes acting after neural induction exert an instructive effect upon neural cell lineage? There is evidence to suggest that the morphologically-homogeneous population of neuroepithelial cells is heterogeneous. This implies that, amongst the undifferentiated precursors, there are cells which are already restricted in their lineage potential. The concept of a heterogeneous population of precursor cells dates back to the studies of His (1889), in which two morphologically-distinct cell types were identified; round cells, or neuroblasts, thought to be neuronal precursors and co-

lumnar cells, or spongioblasts, hypothesized to be glial progenitors. However, it was subsequently shown that these morphologies reflected different phases of the mitotic cycle, rather than specifying lineage restriction (Schaper, 1897). Levitt et al. (1981) provided more convincing evidence of preexisting cellular determination by showing that, at midgestation in the foetal monkey, a subset of the morphologically-undifferentiated ventricular cells expressed GFAP, suggesting that subpopulations of these precursors already possessed discrete commitment potentials. The operational use of GFAP as a lineage marker was probably justified, as proliferating precursor populations which invariably give rise solely to neurons (e.g. cerebellar granule cells) never expressed detectable levels of GFAP. Current opinion favours the view that some neuroepithelial cells retain a multipotential capacity, with the ability to differentiate into both neurons and glia. Evidence for this has accumulated from immunohistochemical staining (De Vitry et al, 1980), retroviral infection (Bartlett et al., 1988; Frederiksen et al., 1988) and the single cell cloning (Temple, 1989; Kilpatrick and Bartlett, 1993a) of neural precursor cells. One of the earliest studies to claim multipotentiality (De Vitry et al., 1980), demonstrated that murine hypothalamic precursor cells expressed both glial and neuronally restricted antigens but this conclusion was dependent upon the absolute specificity of these lineage markers. More recently, fluorescent tracers have been injected into precursor cells in the developing frog retina and avian neural crest and provide strong evidence to support the presence of multipotential precursors in these tissues (Bronner-Fraser and Eraser, 1988; Holt et al, 1988; Wetts and Eraser, 1988). However, the technique is limited to shortterm studies and can only be used in vivo in readily accessible structures. Frederiksen et al. (1988) immortalised cerebral precursor cells with a temperature-sensitive mutant of the proto-oncogene SV40 large T antigen and investigated the lineage potential of the resultant immortalised cell lines. These cells retained the morphological characteristics of precursor cells at the permissive temperature but, at elevated tem-

P,F. Bartlett et al.

peratures, both GFAP-positive cells and neuronlike cells were generated. The precursor cell lines have also provided evidence to support a role for epigenetic factors in the regulation of lineage commitment. Using v-myc immortalized cells, Frederiksen et al. (1988) found that cells of both neuronal and glial-like phenotypes could be induced by culture with dibutyryl cAMP and retinoic acid. In addition, Bartlett et al. (1988), were able to induce the commitment of c-myc immortalized precursor cells, by adding fibroblast growth factor (FGF). The phenotypic potential of retrovirally-infected neural precursor cell lines has more recently been explored by transplantation experiments (Renfranz et al., 1991; Snyder et al, 1992). These studies confirm that these cells have a multipotential capacity and that they, or their differentiated progeny, may have the ability to migrate to specific zones of the postnatal rodent brain, including the dentate gyrus of the hippocampus and the internal granular layer of the cerebellum. However, alteration of the genome induced by retroviral infection could alter the developmental potential of the immortalised cells and thus it is unclear if these results reflect the lineage potential of primary cells. Replication-incompetent retroviruses containing the Escherichia coli LacZ gene, the ^galactosidase (BAG) retrovirus, detectable by the chromogenic substrate X-gal, have also been used to assess lineage, both in vivo and in vitro (Sanes et al., 1986; Price et al., 1987; Luskin et al., 1988; Price and Thurlow, 1988; Williams et al., 1991). When this technique has been applied to the developing chick spinal cord (Leber et al., 1990), tectum (Gray et al., 1988; Galileo et al., 1990) and forebrain (Gray et al., 1990) and to rat retinal development (Turner and Cepko, 1987), the presence of multipotential precursor cells has been established. However, when applied to the study of precursor cells in the mammalian cerebral cortex, the retroviral-labelling technique has produced disparate results. Problems with the technique include the variable expression of ^-galactosidase by infected cells and, until recently, the full migration potential of clonally related cells may not have been appreciated (Walsh and Cepko, 1992), mak-

125

ing interpretation of the reported data difficult. Luskin et al. (1988) reported that the vast majority of clones derived by the in vivo infection of E l 2 14 murine cortical cells with the BAG retrovirus were homogeneous with respect to cell type; most (81%) of the derived clones were of neuronal morphology, although a smaller percentage contained phenotypic oligodendrocytes (13%) and astrocytes (4%). Price and Thurlow (1988), studied cell lineage in the El6 rat cerebral cortex and reported similar findings. It was of note that, in both studies, glial clones were less frequent but contained more cells, suggesting that neurons and glia might accumulate from precursor cells at either different rates or by different mechanisms (Luskin et al., 1988; Price and Thurlow, 1988). Alternatively, it may be that glial precursors incorporate the BAG retrovirus at a lower efficiency than neuronal precursors. Williams et al. (1991) used retroviral infection to study the in vitro clonal development of cells derived from embryonic rat cerebral cortex. This study found that most clones derived from cells isolated from the cerebral cortex at El6 were once again restricted to a single lineage; after 78 days in vitro, 39% contained oligodendrocytes, 18% contained neurons and 3% contained GFAPpositive astrocytes. In addition, 36% of the clones were comprised of undifferentiated cells and 5% contained both neurons and oligodendrocytes but no astrocytes. Although, collectively, these results suggest significant lineage restriction within the developing rodent cerebral cortex, both the studies of Luskin and Williams identified cells of undifferentiated phenotype in some clones; consequently, the full lineage potential of the precursor cells may not have been apparent, possibly because the appropriate growth conditions for the differentiation of astrocytes were not provided. Clonal boundaries and, thus, lineage now have been more rigorously analyzed in this system, by infecting the developing brain with a library of genetically distinct viruses and amplifying single viral genomes using the polymerase chain reaction (Walsh and Cepko, 1992). This study has suggested that at least 5% of clones labeled from El5 rat neocortex contained both neurons and astrocytes and, thus, that they were derived from mul-

126

Regulation of the early development of the nervous system by growth factors

tipotential precursors. However, this result may still fall short of being definitive, given that its interpretation is dependent upon the statistical chance of the same retroviral tag infecting more than one precursor within the cortex (Guthrie, 1992). Clonal analysis of Xenopus and rat retinal cellular development also suggests that a single neuroepithelial cell can generate multiple cell types (Turner and Cepko, 1987; Holt et al., 1988; Turner et al., 1990). These results, together with the observation that clonally-related cells migrating radially within the retina encounter different microenvironments suggest that epigenetic factors may influence lineage (Holt et al., 1988). The molecular nature of this postulated epigenetic influence has not, however, been investigated. An alternative explanation for these results is that a cellautonomous programme may regulate cell type and that epigenetic factors influence migration and cytodifferentiation, rather than instructing commitment to a particular cell lineage (Raff, 1989; Watanabe and Raff, 1990). Temple (1989), in an in vitro study of El3.514.5 rat septal precursor cells, found that there was heterogeneity within the isolated population. By clonal analysis, it was shown that some cells exhibited multipotentiality, whilst others were restricted to a single lineage, a finding interpreted as reflecting some intrinsic commitment. Kilpatrick and Bartlett (1993) undertook clonal analysis of ElO murine neuroepithelial cells, and confirmed that there was heterogeneity within the population, in that two predominant and morphologically distinct types of clone were identified. The first type (type A) consisted of large, amorphous cells (37% of clones) and the second (type B) contained cuboidal, epithelial-like cells (54% of clones). In many of the type B clones, very large numbers (> 5x 10"^) of precursor cells were produced, suggesting that their differentiation was not due to a 'biological clock' that counts cell divisions. Although the vast majority of the clones contained undifferentiated cells, 24% of type B clones contained a small number of neurons (< 1% of the component cells). It was also found that 59% of the clones which contained neurons also contained

GFAP-positive astrocytes, indicating that many of the type B clones were derived from bipotential precursors. At a later stage of development, the lineage potential of glial progenitors has been extensively studied and the results suggest that one precursor can generate progeny of more than one glial cell type (Raff et al, 1983a; Temple and Raff, 1985). These studies have also suggested that there are two major types of astrocyte (Raff et al., 1983b). Type 1 astrocytes are thought to be important in the maintenance of the blood-brain barrier (Janzer and Raff, 1987) and in the production of neurotrophic factors (Raff, 1989), whilst type 2 astrocytes may be important in the maintenance of the nodes of Ranvier (ffrench-Constant and Raff, 1986a). In vitro lineage studies of cells isolated from the postnatal rat optic nerve, and more recently from the adult brain (ffrench-Constant and Raff, 1986b; Wolswijk and Noble, 1989), have suggested that these two cell types may be derived from different precursors. In particular, it has been thought that type 2 astrocytes share a common precursor, the oligodendrocyte-type-2 astrocyte precursor, with oligodendrocytes (Raff et al., 1983a). However, it remains uncertain if these in vitro observations are pertinent to in vivo development, in particular because it has been difficult to definitively identify type 2 astrocytes in vivo (Lillien and Raff, 1990). Further, the expression of GFAP by these cells in vitro is sometimes a transient event and, thus, is not by itself indicative of lineage commitment (Hughes et al., 1988). The recent observations of Williams et al. (1991), analyzing clones labeled with the lacZ reporter gene, have failed to identify astrocytes and oligodendrocytes in a single clone. However, these studies did reveal a previously unidentified clonal association, suggesting a common neuronal and oligodendrocyte precursor. To resolve the apparent discrepancies between the observed restricted nature of clones in vivo and the identification of proliferating precursors with multipotential capacity in vitro, it is necessary to postulate that cells with a multipotential capacity become progressively dormant in vivo, as development proceeds. This would explain why few

P.F.

Bartlettetal

clones with bipotential capacity are detected by labelling E12-E16 brain with retroviral markers (Luskin et al., 1988; Price and Thurlow, 1988) given that, with this technique, only actively dividing cells are detected. It is also possible that multipotential cells might be present but remain dormant both at later developmental stages and within the adult brain (see Sections 2.3 and 2.4 for further discussion). This would imply that proliferating progenitor cells and/or their differentiated progeny inhibit the mitotic division of the multipotential precursor, as development proceeds. 2.2. The role of growth factors in the instruction of the developmental fate of neural precursor cells Epigenetic factors may act upon neuroepithelial cells either to potentiate their survival, to induce their proliferation, or to facilitate their differentiation into mature cell phenotypes. Indeed, it has been found that insulin-like growth factor 1 is a necessary epigenetic requirement for neuroepithelial cell survival (Drago et al., 1991). Growth factors which regulate the in vitro proliferation of neuroepithelial cells have also been identified (Gensburger et al., 1987; Cattaneo and McKay, 1990; Murphy et al., 1990; Anchan et al, 1991) and, in particular, FGF has been shown to induce the proliferation of neuroepithelial cells isolated from ElO murine telencephalon and mesencephalon (Murphy et al., 1990). In addition, Cattaneo and McKay (1990) have reported that neuronal precursor cells isolated from E13.5-14.5 rat striatum not only respond to FGF but that proliferation was further potentiated by nerve growth factor (NGF), which correlates with the finding that embryonic striatal cells express the NGF receptor (Gage et al., 1989). Recent reports by Anchan et al. (1991) and Reynolds et al. (1992) also suggest that epidermal growth factor (EGF) provides a proliferative stimulus for CNS neuroepithelial cells isolated from El7 rat retina and E14 murine striatum, respectively. In dissociated cell cultures, at high cell density, FGF also induces the differentiation of neuroepithelial cells into neurons and astrocytes (Murphy et al., 1990). However, this latter response may have

ni resulted from the secondary production of other factors within the cultures. Clonal studies provide a more precise means of analyzing these potential influences. Using this analytical method, Kilpatrick and Bartlett (1993) found that the primary response of FGF was to induce the proliferation of the neuroepithelial population, given that the vast majority of cells in the derived clones were of the undifferentiated phenotype. Further, in these studies, the removal of the proliferative signal was inadequate by itself to induce differentiation, a process which required quite discrete and separate epigenetic signals. These findings differed from those of Cattaneo and McKay (1990) who found that in the rat, the withdrawal of FGF and NGF was sufficient to invoke the differentiation of colonies of precursor cells into neurons. The differences in the findings of the two studies may reflect either ontogenic or phylogenetic differences. Alternatively, in the culture system of Cattaneo and McKay (1990), accessory cells could have produced secondary differentiative factors, whose effects were inhibited by FGF and/or NGF. Indeed, it was apparent in the studies of Kilpatrick and Bartlett (1993) that, when the neuroepithelial precursor cells were exposed to soluble activities produced by an N-m>'cimmortalized, astrocyte-precursor cell line (Kilpatrick et al, 1993), significant numbers of differentiated neurons were generated. This finding is consistent with the observation that astroglial cells can inhibit precursor proliferation, further confirming that glial cells play a role in the terminal differentiation of CNS progenitor cells (Gao et al, 1991), although in the latter study the effect was dependent upon cell surface membranes. The differentiation of precursor cells into neurons does, however, also appear to require the removal of the proliferative signal, as the effectiveness of the astrocyte-derived factor was diminished by the presence of basic FGF (bFGF) (Kilpatrick and Bartlett, 1993). It is of interest that a recent report by Nurcombe et al. (1993) suggests that bFGF and acidic FGF (aFGF) may have different roles in early neural development. The expression patterns of these molecules show that bFGF is expressed at the time of precursor proliferation (ElO) but that aFGF is first expressed at

128

Regulation of the early development of the nervous system by growth factors

Ell, corresponding to the time of neurogenesis. However, it remains to be determined if aFGF is specifically involved in neuronal commitment and/or differentiation. Preliminary results, using clonal analysis, suggest that aFGF is not a factor which invokes neuronal differentiation in this system (Kilpatrick, 1993; unpublished observations). The nature of the regulation which invokes the differentiation of neuroepithelial cells into astrocytes remains largely unknown. Cell-cell interactions may be important in this process, as the generation of GFAP-positive astrocytes amongst clonal populations derived from multipotential precursors occurs predominantly in regions comprised of multiple layers of cells (Kilpatrick and Bartlett, 1993). In optic nerve cultures, the fate of committed glial progenitor cells has been shown to be dependent on epigenetic factors, in that their proliferation is stimulated by either plateletderived growth factor (Noble et al., 1988; Raff et al., 1988; Richardson et al, 1988), FGF (McKinnon et al., 1990) or a combination of both platelet derived growth factor and FGF (Bogler et al., 1990), and withdrawal of these factors induces oligodendrocyte differentiation. Alternatively, the cells can be stimulated to differentiate into type 2 astrocytes by culture with ciliary neurotrophic factor (CNTF) (Hughes et al., 1988) and the extracellular matrix (Lillien et al., 1990), in the presence of serum (Raff et al., 1983a). However, in the assay system established by Murphy et al. (1990), CNTF has no identifiable effect upon ElO murine neuroepithelial cells. Cell-signalling molecules may not only influence cell lineage but they could also instruct regionally specific development within the mammalian CNS, either by influencing cell fate or by potentiating the survival and/or proliferation of regionally-committed progenitors (McMahon et al., 1992). Circumstantial evidence links retinoic acid with the regulation of Hox genes but, as emphasized by McGinnis and Krumlauf (1992), the effects of retinoic acid are pleiotropic, making it difficult to conlcude that its influence upon Hox gene expression is necessarily direct. In contrast, there is good evidence to suggest that the Wnt protooncogenes (Nusse and Varmus, 1992) could repre-

sent a family of genes encoding signalling molecules which directly influence regional development: firstly, Wnt genes have highly restricted patterns of expression within the developing brain (Wilkinson et al., 1987) and secondly, mice homozygous for null alleles of Wnt-1 (int-1) exhibit loss of the midbrain and cerebellum (McMahon and Bradley, 1990). Of further interest is the finding that these mutant mice also exhibit perturbed transcriptional factor expression within the hindbrain, in particular of the homeobox-containing gene. En, In fact, the first abnormality identifiable in these mice is of loss of anterior expression of En within the midbrain, suggesting that although Wnt-l is not implicated in the activation of En, it may be necessary for the maintenance of En expression (McMahon et al., 1992). This interaction may be of great significance as, in Drosophila, the orthologue of En, engrailed, is implicated in the specification of posterior segment identity in response to the polarized expression of the orthologue of Wnt-1, wingless (Martinez Arias et al., 1988; McMahon et al, 1992). This potential signalling cascade serves to demonstrate how growth factors may interact with homeobox genes to influence pattern formation within the developing embryo. These findings do not, however, preclude the possibility that a cell-intrinsic developmental programme is important in guiding either regionally specific pattern formation or cell lineage. In terms of regional development, this influence could be mediated by the intrinsic control of Hox gene expression. This might be effected by either autoregulatory circuits (McGinnis and Krumlauf, 1992) or by the expression of transcription factors which directly enhance Hox gene expression, as exemplified by the regulation of HoxB2 (2.8) by the zinc finger gene KroxlO during hindbrain segmentation (Sham et al., 1993). At the single cell level, subsets of committed but morphologically homogeneous and undifferentiated progenitor cells could be produced by the asymmetrical division of multipotential precursors. If this were so, epigenetic factors may provide the necessary requirements to invoke differentiation but may not, by themselves, instruct cell type. The best evidence

P.F. Bartlett et al.

for such a combinational influence is derived from studies of the oligodendrocyte-type-2 astrocyte progenitor (Raff, 1989) and of rat retinal neuroepithelial cell development (Watanabe and Raff, 1990), in which it has been demonstrated that cell-cell interactions and a cell-autonomous developmental programme combine to instruct both the timing and the nature of precursor cell differentiation. 2.3. Age dependency: an ontogenic hierarchy of stem cells? It has been found that ElO murine neuroepithelial cells proliferate in response to bFGF but not to EGF (Murphy et al., 1990; Kilpatrick and Bartlett, 1993). In contrast, Reynolds et al. (1992) identified EGF and transforming growth factor-a (TGFa) as mitogens for multipotential precursors isolated from E14 murine striatum and Anchan et al. (1991) identified EGF as a mitogen for El7 rat retinal neuroepithelial cells. These disparate results suggest that the capacity of a precursor cell to respond to a given growth factor is dependent on developmental stage. Of relevance to this issue are the findings of Lillien and Cepko (1992) which demonstrated that rat retinal neuroepithelial cells change during development in terms of their responsiveness to mitogenic signals: in particular, progenitor cells from younger retinas were more responsive to FGF, whereas those isolated from older retinas were more responsive to TGFa. This issue has now been further addressed by cloning El7 murine cerebral cells (Kilpatrick and Bartlett, unpublished observations). The results indicate that it is possible to isolate proliferating precursor cells from this population with either bFGF or EGF, although the cloning efficiency is superior with bFGF. However, whereas El7 precursor cells gave rise to astrocytic progeny with either bFGF or EGF, it was only possible to identify proliferating multipotential precursors with bFGF, suggesting that the effects of EGF upon precursor cells are confined to committed glial progenitors. This contrasts with the findings of Reynolds et al. (1992) who have suggested that EGF and TGFa induce the proliferation of CNS multipotential progeni-

129

tors. However, these latter results were obtained from bulk cultures which contained 2500 cells/ cm^ and, when single-cell analysis was undertaken, neither the generation of GFAP- nor neurofilament-positive cells was reported. This observation suggests that the differentiative effects of EGF could be indirect and that neuronal differentiation is reliant upon the production of secondary factors by either committed or post-mitotic cells in the bulk cultures. Furthermore, it is unclear as to whether the proliferating cells in the clonal cultures had multipotential capacity, even though they stained positively for nestin, a putative precursor-cell marker (Frederiksen and McKay, 1988). Indeed, it remains possible that these cells were committed but GFAP-negative glial progenitors, an interpretation which would be consistent with our findings (Kilpatrick and Bartlett, 1995). We have also found that, unlike ElO precursors. El7 multipotential precursors cultured with bFGF exhibit the intrinsic capacity to generate large numbers of neurofilament positive cells, even in the continued presence of the proliferative signal. Although the majority of these latter cells remained morphologically undifferentiated, this finding indicates that multipotential precursors have an agedependent variability in their capacity to exhibit commitment to the neuronal lineage, in response to set epigenetic conditions. Furthermore, the prohferative potential of cells within clones derived from multipotential precursors isolated at El7 is much lower than from those isolated at ElO (Kilpatrick and Bartlett, unpublished observations). Lillien and Cepko (1992) also found that the proliferative potential of progenitors isolated from the retina tended to reduce with age. Thus, there may be an ontogenetic hierarchy of multipotential-precursor cells within the CNS, analogous to the well defined hierarchy of precursors within the haematopoietic system (Metcalf and Moore, 1971). 2.4. The identification of multipotential precursors in the adult mammalian brain In the mouse, the majority of neurons are formed prenatally, although there are well identi-

130

Regulation of the early development of the nervous system by growth factors

fied exceptions, including granule neurons in the hippocampus (Altman, 1963; Schlessinger et al., 1975; Altman and Bayer, 1990) and cerebellum (Altman, 1972), and the olfactory neuroepithelium (Graziadei and Graziadei, 1979). It is already well established that cells within the subventricular zones of the brains of adult fish, amphibians and some birds retain the capacity for proliferation (Alvarez-Buylla et al., 1988). In canaries, these proliferating cells have been shown to maintain the capacity to differentiate into neurons and to populate song-bird nuclei, in response to hormonal stimuli (Nordeen and Nordeen, 1989). Further, it has been shown that proliferating cells can be identified in the subventricular zone of adult mammals, as assessed by tritiated thymidine incorporation studies (Morshead and van der Kooy, 1992). When these cells were marked by LacZ infection, it was shown that the resultant clones were virtually never more than two cells in size, whereas some 33% of the subventricular cellular population was shown to be undergoing cell division. This finding suggests that the proliferating cells were dividing asymmetrically, with one progeny maintaining an undifferentiated phenotype, thus contributing to the self-renewal of the precursor, whereas the other daughter cell invariably died. If this were so, it could be that the proliferating cells are precursors which, in the absence of appropriate epigenetic conditions, lack the ability to differentiate into mature neural cell phenotypes. The elucidation of the epigenetic conditions which induce the neuronal differentiation of neuronal precursors in embryogenesis has provided the means to address this issue. Two recent reports suggest that neuronal precursor cells exist in the adult mammalian brain (Reynolds and Weiss, 1992; Richards et al., 1992). Reynolds and Weiss (1992) isolated cells from the striatum of the adult mouse brain and induced the in vitro proliferation of precursors with EGF in serum-free conditions. These cells were also demonstrated to have the capacity to differentiate into neurons and glia. Richards et al. (1992) reported that neuronal induction from murine cerebral precursors was optimal under in vitro conditions in which the cells were initially stimulated with

bPGF and then with medium conditioned by an Nmyc immortalized astrocyte-precursor cell line (Kilpatrick et al., 1993). In this study, neuronal induction was not potentiated by initial culture with EGF. The reason for the disparate results of the studies with regard to the efficacy of EGF remains unclear. It may be that under the culture conditions employed by Reynolds and Weiss (1992) there was production of secondary factors, including FGF, within the cultures (see Section 2.2). Alternatively, it is possible that the differences relate to a disparity in the expression of the EGF receptor by different subpopulations of precursor cells. One major challenge is to determine if the in vivo differentiation of these precursor cells can be achieved. In order to investigate this, it will be necessary to determine whether the precursors survive within the brain in a dormant, nonproliferative state or whether they represent an analogous population to the proliferating subependymal cells identified by Morshead and van der Kooy (1992). 2.5. Potential clinical applications The results of adult neural cell culture, together with the results of clonal analysis of embryonic neuroepithelial cells, suggest that the differentiation of precursor cells is invoked by specific epigenetic conditions. These observations raise the possibility of utilizing precursor cell populations, present in situ, to replace degenerating neurons by the in vivo administration of differentiative factors. The ability to generate large clonal populations from multipotential precursors in vitro suggests that neuronal repopulation could also be achieved by precursor cell transplantation. Renfranz et al. (1991) and Snyder et al. (1992), have recently transplanted retrovirally-immortalized multipotential cells into postnatal rat brain and have provided convincing evidence of in vivo differentiation of these cells into neuronal and glial elements. The elucidation of epigenetic conditions which are permissive for the proliferation of multipotential precursors has provided the methodology to repeat the above transplantation experiments with pri-

P.F, Bartlett et al.

mary cells. Such experiments would have obvious advantages over those using cell lines. Firstly, lineage could be studied using cells which have not been affected by genetic manipulation and, secondly, if precursor cell transplantation is to be applied to human degenerative disease, primary cells will be the preferred source. Transplantation experiments could also be performed using primary cells after they have been exposed to either specific growth factors or to antisense oligonucleotides designed to inhibit the effects of specific growth factors or their receptors. Such experiments would also provide further insight into how exogenous and endogenous mechanisms interact to regulate the differentiation of neural cells. Preliminary results indicate that transplanted neuroepithelial cells can successfully engraft into the postnatal mammalian brain (Kilpatrick et al., 1994). These experiments also suggest that cells injected into the hippocampus selectively migrate to regions of continuing neurogenesis in the host brain. In particular, some of the transplanted cells align within the granular layer of the dentate gyrus, where they elaborate processes which extend toward the CA2 region of the pyramidal layer, indicating that the donor cells can adopt similar morphologies to those of host granule neurons. An additional challenge will be to determine the lineage potential of the injected cells in other regions of the brain. Furthermore, if this technique is ever to be applied to human degenerative disease, it also will be necessary to establish that cells can engraft and differentiate within the brain of adult hosts, especially animals with cerebral lesions. 3. Early development of the peripheral nervous system The neural crest is a transient structure that arises from the dorso-lateral aspect of the closing neural tube. The cells migrate along several discrete tissue pathways and give rise to the majority of cells of the PNS, both neuronal and glial, melanocytes and adrenal medullary cells. In the cephalic region, crest cells give rise to additional cell types, in-

131

cluding facial mesenchyme derivatives and branchial mesenchymal structures. In addition to forming structures it also influences the development of structures such as the thymus. Thus, studying the regulation of neural crest development has not only profound relevance to neural development but also to the development of the whole animal. As a consequence, many of the major questions raised in crest development are identical to those concerning broader aspects of developmental biology. 5.7. Fate map of the neural crest One early problem in the study of neural crest ontogeny was tracing the neural crest cells as they migrated through the embryo. Early studies involving extirpation of the neural crest in a variety of experimental animals led to the identification of some of the neural crest derivatives (see LeDouarin, 1982). Later, Weston and Johnston pioneered the technique of tagging the neural crest cells, in this case with pHJthymidine, and following their fate through the embryo (Weston, 1963; Weston, 1986). Perhaps the most definitive studies of the fate of the neural crest came from the use of the chick-quail marker system (reviewed in Noden, 1978; Le Douarin, 1982, 1986; Le Douarin and Smith, 1988). In this system, chimeras were made by replacing a particular region of an embryo of one species with the same region from the other species. These chimeras remain viable at least until birth, and the cells of the donor can be identified on the basis of structural differences in the interphase nuclei between the two species. Le Douarin and colleagues transplanted fragments of either the entire neural primordium (i.e. fragments of the neural tube with associated neural crest) or, at the cephalic level, the neural folds alone containing the neural crest cells. This approach, has led over a period of ten years or more to the construction of a fate map of the neural crest. This map showed that there are discrete regions of the crest which give rise to particular ganglia and other neural crest-derived structures. The fate map demonstrates that the direction of most neural crest cell migration is lateral to the neural

132

Regulation of the early development of the nervous system by growth factors

tube and thus that the resultant neural crest derivatives correspond to their original position along a rostro-caudal axis. For example, the adrenal medullary cells originate from the spinal neural crest between the level of somites 18-24, the spinal neural crest caudal to somite 5 gives rise to the ganglia of the sympathetic chains; the ciliary ganglion is derived from the mesencephalic neural crest and the mesectodermal derivatives are derived from the rostral regions of the neural crest and are mainly located in the head and neck. 3.2. Commitment versus multipotentiality The fate map described above refers to the normal developmental potential of the neural crest, as the grafting experiments described were both isotopic and isochronic. However, in other experiments neural primordia or neural folds were transplanted to different regions (heterotopic) of the host embryo to determine their full developmental potential (see Le Douarin, 1982). The results showed that, in general, the location of the grafted cells in the chimeric embryo and not their origin determined their developmental fate. For example, vagal crest cells (which normally contribute to parasympathetic innervation of the gut) grafted to the level of somites 18-24 differentiated into adrenergic cells in the sympathetic ganglia and adrenal medulla (the normal derivatives of the crest of this region). In the reverse experiment, where the presumptive adrenomedullary neural crest cells were transplanted into the vagal region, enteric ganglia containing both cholinergic and peptidergic neurons were formed. A range of such experiments established that in most cases it is the embryonic environment of the neural crest cells that determines their differentiated phenotype and that the crest cells are multipotential. It is important to note, however, that this type of analysis reflects the differentiation potential of a population of cells and can be explained by either the selection of different populations of partially committed cells, or the multipotentiality of individual crest cells. There are some significant exceptions to the perceived multipotentiality of neural crest cells. It

is only the cephalic regions of the neural crest that can give rise to ectomesenchymal derivatives such as bone, smooth muscle, adipose tissue, meninges and endothelial cells, which are exclusively located in the head and upper body. Further, if chick midbrain neural crest, which normally migrates to the first (mandibular) arch, is grafted to the second (hyoid) arch, normal migration into the arch occurs, but first arch structures are formed (Noden, 1983, 1986). These results show that not only are there regional variations in the potential of the neural crest, but also that, to a limited extent, some neural crest cells may already be committed to a particular fate. In addition to these obvious differences there are some subtle differences in the capacity of different regions of the crest to replace other normal crest regions. For example, replacement of the mesencephalic neural crest with trunk neural crest results in the development of an abnormal trigeminal ganglion (Noden, 1978); and the potential for adrenergic differentiation and melanocyte formation is greater in the trunk neural crest than in the cephalic crest (Newgreen et al., 1980). Also, when cephalic crest is transplanted to the trunk, the crest cells migrate into the dorsal mesentery and colonize the gut, which does not normally happen (Le Douarin and Teillett, 1973). It is these exceptions which suggest that there are regional differences in the composition of the crest along its rostral-caudal axis. The cephalic regions of the crest appear to have the potential to give rise to all neural crest derivatives, whereas the trunk crest is restricted to PNS and melanocytic derivatives. This restriction may apply only after a particular developmental stage, as Lumsden has found that trunk neural crest cells from the mouse can participate in tooth formation when combined with mandibular epithelium (Lumsden, 1987, 1988), but only if the cells are taken from very early neural crest (6-12 somite stage). Thus, the restriction process presumably occurs after this time. The rostral-caudal gradient and segmentation pattern observed in vertebrate neural development may be primarily the result of mesenchymal influence, although the inductive agents are as yet unknown (see Dodd, 1992 for review).

P.F. Bartie tt et al.

3.3. Neural crest cells display multipotential and committed characteristics when grown in vitro The above transplantation experiments cannot give a clear picture of the differentiation potential of individual neural crest cells within a regional population. To do this, clonal analysis, either in vitro or in vivo , is required. In vitro, progeny of single neural crest cells (clones) can be influenced or manipulated relatively simply by adjusting the components of the medium. In this way, the differentiation potential of neural crest cells may be examined. An important bonus of this approach is that it also provides an assay for putative factors which may influence the development of the crest derivatives. A number of workers have developed these clonal cultures, and their results suggest that there are both committed and multipotential cells within the neural crest. Sieber-Blum and Cohen (1980) first used clonal analysis to study quail neural crest cells and found a proportion of clones which contained both catecholaminergic (neuronal lineage) and pigmented cells. More recent studies (Sieber-Blum 1989; Ito and Sieber-Blum, 1991) have revealed three classes of clones: clones exclusively of the melanogenic lineage, clones that were unpigmented, and clones containing both pigmented and unpigmented cells (mixed). The unpigmented and mixed clones all contained both catecholaminergic and sensory neurons. Thus, in this system, there is evidence for tripotent cells, cells restricted to two lineages, and fully committed cells. In the latter study (Ito and Sieber-Blum, 1991), a clonal analysis of the cardiac neural crest, pluripotential (mesenchymal, neuronal and melanocytic), bipotential (mesenchymal, neuronal) and fully restricted clones were found. Studies from the laboratory of Le Douarin found evidence for a similarly heterogeneous range of clones (Baroffio et al., 1988; Dupin et al., 1990; Baroffio et al., 1991). In these studies, besides the fully restricted clones, multipotent clones comprising neurons, pigmented cells and nonneuronal cells were found, as well as partially more restricted clones which contained Schwann cells, satellite cells and neurons, but not pigmented

133

cells. This pattern may indicate that neural and melanocytic cell precursors segregate early in the differentiation process. In one study (Baroffio et al., 1991), some multipotential clones were found to contain either the full array of neural crest derivatives, including mesenchymal elements, or were restricted to either neural and melanocytic, or neural and mesenchymal. These studies thus support the idea that within any crest population there are both multipotential and committed cells at the migratory stage. The observation of considerable heterogeneity in the clones is not necessarily an indication that there is innate heterogeneity in the neural crest cell's repertoire. It may be that all neural crest cells are initially multipotent but, at the time the cells are isolated, they have reached different stages of differentiation. If this is the case, then the actual lineage pathways or commitment steps may be inferred from the segregation pattern observed in the clones. Within the multipotent clones, there are some which are either neural/melanocytic or neural/mesenchymal, which suggests that this may be the first restrictive choice the neural crest cells make. Neurons, melanocytes and mesenchymal cells segregate from each other in the more restricted clones. Glial cells also segregate from the other cell types, but a significant number of clones show co-segregation of neurons and glial cells, indicating that there is a common glial/neuronal precursor which retains its bipotentiality late into the differentiation process. These findings, especially with the many intermediate, or partially committed clones, strongly supports the concept of sequential differentiation from a multipotential cell. However, the concept of identical pluripotential stem cells in all regions of the neural crest may be an oversimplification, given the differing potential found between cephalic and trunk crest to give rise to mesectodermal derivatives. A complication in these studies is that the clones are normally analysed after a number of weeks, when there can be thousands of cells in each clone. Under these conditions the microenvironment of each clone might itself vary; there might be endogenous production of different growth factors which could influence cell pheno-

134

Regulation of the early development of the nervous system by growth factors

type. One labor-intensive approach to this problem would be to subclone daughter cells as soon as they arise and to examine the resultant clonal phenotype. This would also provide a more detailed account of the sequence of lineage restrictions occuring during crest development. The other clonal approach used to determine the degree of multipotentiality or commitment of the neural crest cells has been undertaken in vivo. In these studies, single neural crest cells have been microinjected with a flourescent dye prior to migration from the neural tube (Bronner-Fraser and Fraser, 1988, 1989; Fraser and Bronner-Fraser, 1991). After 2 days, the clonal progeny of the cells were sometimes found to be distributed in many regions to which neural crest cells normally migrate. Although the phenotype of these cells could not be definitively ascribed, it was found, on the basis of morphology and antibody binding, that individual clones contained sensory neurons, presumptive melanoblasts, satellite cells in dorsal root ganglia, adrenomedullary cells and neural tube cells. Thus, these findings support the idea that there are multipotential neural crest cells in vivo. These studies also purported to show that some cells in the neural tube can give rise to both neural crest cells and neural tube cells destined to become mature CNS cells (Bronner-Fraser and Fraser, 1988, 1989). Whilst these studies indicate the diversity of cell products there has been concern as to whether they represent the progeny of a single cell. Confining the injection to just one cell appears to still present considerable technical difficulties to other workers. However, results obtained by following the progeny of neural crest cells infected with laC'Z containing retrovirus in vivo in the dorsal root ganglia (Frank and Sanes, 1991) do tend to support the multipotential concept. However, here again, problems with infecting a small cohort of dividing cells, rather than a single cell, makes this interpretation somewhat equivocal. 3.4. Mammalian neural crest cells lines display a variety of differentiated characteristics In mammals, far less is known about cell lineage and commitment of the neural crest than in

avian species. One approach to studying the mammalian neural crest is to make immortalized cell lines. If clonal cell lines can be obtained which represent neural crest cells or their derivatives, they may be useful in inferring cell lineage relationships in an analogous way to that described for clonal analysis of avian neural crest cells. Previous work from our and other laboratories has shown that retrovirus-mediated proto-oncogene transduction of the neural precursor cells from mouse neuroepithelium results in the production of stable neuroepithelial and neural cell lines (Bartlett et al., 1988; Bernard etal., 1989). In similar manner, we immortalized mouse neural crest cultures using retroviruses bearing the cmyc or the N-myc proto-oncogenes (Murphy et al., 1991a). The different lines could be broadly classified into three subgroups. Group 1 contained flat adherent cells, which looked like primary neural crest cells. Group 2 contained flat cells at low density, a proportion of which at higher density and longer time in culture, tended to become stellate or dendritic. Group 3 cells grew initially as flat cells but after a relatively short time in culture, most of the cells elongated and put out processes. These cell lines were examined for the expression of lineage-specific or lineage-related antigenic markers and for the expression of neural specific mRNAs. We examined the expression of NGF and its receptor, which are expressed by cells in the PNS, myelin basic protein (MB?) and the proteolipid protein (PLP) of myelin which, in the PNS, are specific to Schwann cells. A neuron-specific gene SCG-10 (superior cervical ganglion, see Anderson and Axel, 1985) was also used in the analysis. Group 1 cell lines not only morphologically resembled migrating neural crest cells, but this group also was largely devoid of phenotypic markers, both antigenic and mRNA, expressed by mature neural cells. These observations are consistent with the idea that some of the migrating neural crest cells are not yet committed to a single developmental pathway and probably represent stem cells. These stem cells have presumably been arrested at this stage by the immortalization process.

P.F. Bartlettetal

Some of the cell lines also displayed a plastic or at least a bipotential nature, especially those in group 2. Particular cell lines expressed their bipotentiality in the expression of markers associated with two lineages. In one case a cell line, NC 14.9.1, appeared to be bipotential since in a cloned population these cells expressed neurofilament as well as MBP and PLP, showing that it had characteristics of both neurons and Schwann cells. Likewise another cell line, NC14.4.9D, expressed both PLP mRNA and SCG-10 mRNA and all the cells express neurofilament. Similarly, multipotent neural cell lines have been isolated from newborn brain (Fredericksen et al., 1988; Ryder et al., 1990). These cell lines also share some other characteristics of our cell lines in that some of the antigenic markers examined were expressed on a small proportion of cells in particular cell lines. Cell lines from group 2 also had the properties of progenitor cells (Murphy et al., 1991a). For example NC 14.4.8 cells contained cells that differentiated, after 1-2 weeks in culture, into Schwann-like cells. Further, these older cultures expressed mRNA for MBP, PLP, NGF and NGF receptor. All these observations are consistent with this cell line comprising Schwann cell progenitors. Finally, one of the cell lines (in group 3) appears to represent differentiated neuronal cells. These cells (NC14.4.6E cells) have fine processes which contain neurofilament. In addition, these cells express mRNA for the neuronal protein SCG10, as well as for NGF. The multipotential nature of the neural crest cells which were originally infected with either cmyc or N-myc containing viruses was also demonstrated by the observation that cell lines which have the same myc integration pattern, and thus must have originated from the same cell, can have quite different phenotypes. It is possible that an immortalized multipotential cell divided a number of times before differentiation of the progeny cells into the different phenotypes took place. Thus, a single crest cell can give rise to an imature neural crest-like line, a Schwann cell progenitor and a bipotential cell line. Mammalian cell lines have also been derived from rat primary neural crest cultures (Lo et al..

135

1991). One cell line, NCM-1, was generated that displayed bipotential characteristics. NCM-1 has the characteristics of a glial progenitor and resembles Schwann cells in serum-free medium. In addition, some of these cells acquire sympathoadrenal characteristics in response to FGF and dexamethasone. Thus, this cell line contains cells with the potential to generate precursors in at least two neural crest sublineages. 3.5. Growth factor regulation of neural crest proliferation We have previously shown that FGF stimulates the proliferation of freshly-isolated neuroepithelial cells (Murphy et al, 1990). Given that the neural crest is initially contiguous with the neuroepithelium, FGF appears to be a good candidate for involvement in neural crest proliferation. We have recently shown that the great majority of NC cells which have migrated from the trunk regions of the neural tube are stimulated to divide by FGF (Murphy et al., 1994). Serum was also found to be required for NC proliferation; factors in serum which may be responsible include the insulin-like growth factors (IGFs), and in particular IGFl, which is required for FGF2 regulated proliferation of neuroepithelial cells (Drago et al., 1991a). However, other factors must also be required for NC division, since our serum free medium contains enough insulin to bind to the IGFl receptor and produce a biological signal. Our further studies demonstrated that a preponderance of sensory-like neurons could be generated from FGF treated cultures if they were subsequently treated with leukemia inhibitory factor (LIF) (Murphy et al., 1994; see also below Section 3.6.2.). This does not necessarily indicate that FGF was specifically stimulating the proliferation of a pool of sensory precursor cells, as it has been shown that the majority of sensory neurons are derived from multipotential stem cells (Sieber-Blum, 1989). Thus FGF may be acting to expand a multipotential stem cell pool. Previous results with subcloning of neuroepitheUal cells show that FGF can stimulate proliferation of multipotential cells within this population (Kilpatrick and Bartlett, 1993).

136

Regulation of the early development of the nervous system by growth factors

The effects of FGF are independent of the presence of other cell types in the cultures, which is in contrast to the effects of NTS, another factor which has been reported to stimulate the proliferation of NC cells. NTS weakly stimulates the proliferation of isolated NC cells, but in the presence of somite cells, more strongly stimulates proliferation in the NC (Kalcheim et al., 1992). Recent data indicate that NTS stimulates sympathetic neuroblast proliferation by promoting precursor survival (DiCiccoBloom et al., 199S) and thus it may be possible that a part of the effect seen in NC cultures is due to a similar mechanism. FGF2 immunoreactivity has been detected in the basement membrane around the dorsal neural tube (Kalcheim and Neufeld, 1990) as well as in association with a heparan sulphate proteoglycan (HSPG) within the neuroepithelium (Ford et al., 1994). This correlates with the expression of FGF2 mRNA within the neural tube from E9 (Drago et al., 1991a). In situ hybridisation analysis shows specific expression of FGF2 within the neural tube and developing dorsal root ganglia (DRG) at ElO (Murphy et al., 1994). Within the DRG, most of the cells are labelled with antisense probe and are probably synthesising FGF2. This corresponds to a time when the ganglia have just been formed from migrating neural crest cells and they are at a maximum stage of proliferation (Lawson et al., 1974), suggesting autocrine/paracrine regulation of proliferation within the DRG. Likewise, the cells in the neural tube which are positive for FGF2 mRNA are predominantly proliferating cells. Studies of the localisation of the FGF receptor also support a role for FGF in the early phases of sensory development. In situ hybridisation studies in the chicken show a significant level of FGFR mRNA in NC cells from 55 h to S.5 days, as well as in a subpopulation of sensory ganglion cells from ES.5 to E5 (Heuer et al., 1990), periods of active proliferation of NC cells and sensory neuroblasts (Carr and Simpson, 1978). In the rat, FGFR mRNA is present in the DRG from at least El2 (Wanaka et al., 1991), a period of maximal proliferation of sensory neuroblasts (Lawson et al., 1974).

3.6. Growth factor regulation of neural crest differentiation. 3.6.1. Sympathoadrenal lineage Developmentally, the best characterized cell lineage within the neural crest is probably the sympathoadrenal lineage. There are three cell types in this lineage, the sympathetic neuron, the adrenal chromaffin cell and a third cell of an intermediate phenotype, the so called small, intensely fluorescent cell (SIF cell) (see Anderson, 1989; Patterson, 1990). Although progenitors of this lineage have not been isolated from neural crest cultures, they have have been isolated from embryonic adrenal medulla as well as from both embryonic and neonatal sympathetic ganglia. These progenitors will differentiate into either chromaffin cells or sympathetic neurons depending on culture conditions (Doupe et al., 1985a,b; Anderson and Axel, 1986). FGF will initiate neuronal differentiation as well as a dependency of the cells on NGF for their survival. Glucocorticoids will stimulate the cells to differentiate into mature chromaffin cells. The evidence for the presence of FGF in the embryo around the neural tube has been presented above. The possibility that the developing sympathetic neuron precursors will find a supply of this factor at the site of ganglia is thus quite reasonable. In the adrenal medulla, on the other hand, when the precursors migrate into the adrenal gland they may be subject to a high concentration of steroids produced in the adrenal cortex. The role of NGF as a survival factor for sympathetic neurons has been demonstrated over the past forty years using numerous experimental systems (see Levi-Montalcini and Angeletti, 1968). It remains the only molecule to be unequivocally shown to be critical for neuron survival in vivo. The injection of anti-NGF antibodies into newborn mice results in the destruction of the sympathetic nervous system. Studies of the mechanism of action of NGF have resulted in it becoming the prototype of target-derived neurotrophic factors. In this model, the newly differentiated neurons sprout axons to their target fields, where there is a limited supply of a target-derived survival factor. It is

P,F. Bartlett et al.

postulated that only those neurons which have made the appropriate connections will obtain this factor and survive. Thus, this model provides a part of a mechanism for the control of the development of the nervous system into a threedimensional network. A number of other factors have been implicated in the development of the sympathoadrenal lineage and in particular the development of sympathetic neurons. IGF-1 stimulates proliferation in cultures of rat sympathetic ganglia (DiCicco-Bloom et al., 1990). Whether this is a direct effect of IGF-1 on the proliferation of the neuronal precursor cells, or whether the IGF-1 is acting principally as a survival agent and there are endogenous proliferative factors in these cultures, as we have observed in cultures of neuroepithelial cells (Drago et al., 1991), is unclear at present. Conversely, CNTF inhibits the proliferation of the neuroblasts and may provide a signal to initiate the differentiation of the cells (Emsberger et al., 1989). Other factors have been described which influence the transmitter phenotype of the sympathetic neurons. Most of the sympathetic neurons are adrenergic, except for those which innervate the sweat glands, which are cholinergic. One of the factors which may influence the switching of phenotype of these neurons to cholinergic has recently been purified and is equivalent to LIF (Yamamori et al., 1989). As discussed below, it is beginning to emerge that LIF has multiple activities within the nervous system as well as outside it. 3.6.2. Sensory lineage The processes which regulate the development of sensory neurons from their precursors in the embryonic neural crest have not been well characterized. We recently reported that LIF, a protein with multiple activities (Abe et al., 1986; Gearing et al., 1987; Williams et al., 1988; Baumann and Wong, 1989; Yamamori et al., 1989; see Section 3.5) stimulates the generation of neurons in cultures of NC cells (Murphy et al., 1991b). LIF belongs to a structurally related family of molecules (Bazan, 1991; Patterson and Nawa, 1993) including CNTF, oncostatin M (OSM) and probably another factor, growth promoting activity (Leung et

137

al., 1992), which share the same signalling mechanism (Gearing et al., 1992; Ip et al., 1992). The neurons generated in these cultures have the morphology of sensory neurons and contain neuropeptides, such as calcitonin gene-related peptide, found in mammalian sensory neurons. Consistent with these neurons being of the sensory lineage was the finding that they arose from non-dividing precursors, a characteristic previously observed for early arising sensory precursors in neural crest cultures (see Weston, 1991). In addition, LIF supported the generation of sensory neurons in cultures of cells obtained from embryonic DRG. The full differentiation of sensory neurons in these cultures is dependent on the presence of NGF (Murphy et al., 1993). Thus, the role of LIF early in the differentiation of sensory neurons appears to be primarily at the step of differentiation of neuronal precursor cell to newly differentiated neuron (Murphy et al., 1993). FGF also influences the differentiation of NC cells as prior treatment with FGF2 results in 50% of the cells differentiating into neurons when followed by treatment with LIF, a significantly higher proportion than that which differentiate in LIF alone (15%; Murphy et al., 1994). Brill et al. (1992) reported that in serum free conditions and in the presence of BHK21 cells, FGF2 causes up to a 4-fold increase in the number of neurons arising from NC cells. These findings suggest that FGF2 stimulates neuronal differentiation through interaction with this cell line or factors produced by this cell line. In our cultures, FGF2 may stimulate the production of other factors in the NC cultures which then prime some of the cells so that they differentiate under appropriate conditions. We have previously proposed such an FGF2 induced cascade to explain the effects of FGF2 on the differentiation of neuroepithelial cells, and have shown that FGF2 upregulates the synthesis of the neurite promoting molecule, laminin, in addition to stimulating prohferation (Drago et al., 1991b). At later stages in sensory development, LIF can act as a survival factor. In cultures of DRG isolated at postnatal day 2, a high proportion of neurons survived in the presence of LIF (Murphy et al., 1991b). Thus LIF may also be a neurotrophic

138

Regulation of the early development of the nervous system by growth factors

factor, like NGF. Binding studies on the DRG cultures from P2 mice showed that greater than 60% of the neurons bound significant amounts of [^2^I]LIF, which was completely inhibited by the addition of cold LIF (Hendry et al, 1992). Furthermore, there was negligible specific binding of [^^^I]LIF to non-neuronal cells in the culture. Thus, at this age, the only cells capable of responding to LIF in the DRG are the sensory neurons. One of the essential criteria to be fulfilled by a neurotrophic factor is that once the factor is taken up by the nerve terminals, it should be retrogradely transported back to the neural perikarion. The transport of the neurotrophic factor is the signal from the target tissue to the neuron that results in neuronal survival (Hendry et al., 1974). To test the possibility that LIF is retrogradely transported, mice were injected in the skin or muscle with ^'^^lLIF and, in those animals injected in the skin of the foot, there was a significant accumulation of radioactivity in the DRG (Hendry et al., 1992). The retrograde transport of LIF into the DRG was confirmed by autoradiographic examination of histological sections of ganglia from these animals, which revealed radioactive material only within the cell bodies of the sensory neurons. Thus LIF may have a dual role in the sensory nervous system, first as a differentiation stimulus for the sensory precursors and second as a neurotrophic factor for mature sensory neurons. Further supportive evidence that LIF has a role in sensory development in vivo comes from the finding of LIF mRNA in developing DRG from as early as El3 and possibly earlier (Murphy et al, 1993). In addition, LIF mRNA is present in the spinal cord region from El2, as well as at sites of peripheral sensory innervation. The best characterized factor shown to play a role in the development of sensory neurons is NGF (Levi-Montalcini and Angeletti, 1968; Thoenen and Barde, 1980; Levi-Montalcini, 1982). NGF most probably acts as a target-derived survival factor during the period of natural neuron death, as discussed in Section 3.6.1 for sympathetic neurons. The evidence for the time of action of NGF on sensory neurons comes from expression studies: mRNA for NGF is first observed in the target tis-

sue at the time of innervation of the newly formed neurons, which is concomitant with appearance of NGF receptors on nerve fibres as they innervate these target tissues (Bandtlow et al., 1987; Davies et al., 1987). In addition, the role of NGF in vivo has been established by Johnson and coworkers, who immunized female guinea pigs with NGF and showed that their offspring, which were exposed to NGF antibodies during the period of sensory development, lost up to 80% of their sensory neurons (see Johnson et al., 1986 for review). Other factors, such as the other neurotrophins, also act on developing sensory neurons. Brain derived neurotrophic factor (BDNF) in particular has been implicated in sensory neuron development and at similar stages to those described here (Kalcheim and Gandreau, 1988; Sieber-Blum, 1991). In recent studies from Davies' laboratory (Wright et al., 1993), it has been proposed that BDNF or neurotrophin-3 (NT-3) acts at the stage after neuronal differentiation, but before the neurons become dependent on NGF. BDNF or NT-3 accelerate the maturation of neurons before they become dependent on neurotrophic factors for survival, but the maturation process can still occur in the absence of these factors. If these findings are taken together, then a sequence of steps from neural crest precursor cell to mature sensory neuron can be proposed, each driven by a different factor. The first step, from precursor cell to immature neuron, requires LIF. The second step, from immature neuron to factordependent immature neuron, requires (or is stimulated by) BDNF or NT-3; the third step, the survival of the factor-dependent neuron during target innervation and further maturation, requires NGF. 3,6.3, Parasympathetic lineage While the identity of the factor(s) are as yet unknown, there is now evidence that a soluble factor can direct the differentiation of parasympathetic neurons from precursor cells in the neural crest. By the use of monoclonal antibodies to cell surface antigens, Barald and coworkers have identified a subpopulation of cephalic neural crest cells which are committed to a cholinergic neurogenic fate (Barald, 1988a,b). The monoclonal antibodies rec-

P.F. Bartlett et al.

ognize an antigen on the cell surface which is concerned with the high-affinity choline uptake. These antibodies label all the neurons in the chick and quail ciliary ganglion in vivo and in vitro. In addition, the antibodies label a subpopulation of earlymigrating cephalic neural crest cells. By the use of no-flow cytometry, Barald (1989) has isolated this subpopulation of cells from neural crest cultures and studied its behaviour under a variety of different culture conditions. The cells proliferate in the presence of 15% fetal bovine serum and high concentrations of chick embryo extract, but do not differentiate. However, in chick serum, elevated K+ or heart-, iris- or lungconditioned medium, the cells stopped proliferating and all of the cells became neuron-like within 10 days (Barald, 1989). These cells also stained positively for choline acetyl transferase (ChAT). These experiments were the first to demonstrate that the development of a presumably committed population of neural crest cells can be directly manipulated by culture conditions. The continued proliferation of the cells under one set of conditions indicates that the precursors can still divide, and the observation that they will all differentiate into ChAT-positive neuron-like cells suggests that they are indeed neuronal precursor cells. The conditions used to stimulate the differentiation of the cells are the same which promote the survival and/or cholinergic development of ciliary ganglion neurons. This reinforces the idea that the subpopulation of neural crest cells used in this study represents ciliary neuron precursors (Barald, 1989). 3.6.4. Melanocyte lineage The melanocyte lineage is apparently determined early in development in the mouse and whilst studies in chimeras suggest that thirty four primordial melanocytes are lined up in pairs longitudinally during neural crest formation (Mintz, 1967), this observation would appear to be due to the segregation of cohorts of like cells in metameric units along the spinal cord, as described in zebra fish, and not due to clonal expansion of a single primordial melanocyte. From related studies in the chick, the melanoblasts then undergo rapid

139

proliferation and migrate laterally to the skin (Rawles, 1944; Weston, 1963) where they differentiate into mature melanocytes. The processes which control the proliferation, migration and differentiation of these melanocyte precursors are not clearly understood. However, two classes of mouse mutants point to for the involvement of a growth factor-receptor interaction in this process. These are the White dominant-spotting (W) and Steel (SI) mice. Mice homozygous at either of these alleles are blacked-eyed white, anaemic and sterile; some of the mutations result in lethality (reviewed in Russel, 1979; Geissler et al., 1981). An analysis of the mutations in these mice has revealed a complementary molecular relationship between the two alleles. Firstly, it was found that W allele coded for a growth factor receptor-like tyrosine kinase and which was identical to the proto-oncogene c-kit (Chabot et al., 1988; Geissler et al., 1988). Subsequently, the ligand for c-kit was purified and cloned and was found to be encoded by the 5/ locus (Anderson et al., 1990; Copeland et al, 1990; Huang et al., 1990; Martin et al.,1990; Williams et al., 1990; Zsebo et al., 1990a,b). Thus, this SI factor and the c-kit receptor are strongly implicated in melanogenesis as well as germ-cell production and in haemopoiesis. Because of this range of involvements the SI factor has been variously called mast-cell growth factor, stem cell factor, and the c-kit ligand. In situ hybridisation studies have shown that the embryonic expression pattern of c-kit mRNA reflects the phenotype of WAV mutants: skin, gonads and haemopoietic tissues (Keshet et al., 1991). The tissue distribution of SLF mRNA shows a similar pattern by in situ hybridisation (Matsui et al., 1990; Keshet et al, 1991), in skin, gonads and haemopoietic tissues from around embryonic day 10 (El0) in the mouse, and in particular along the migratory routes for melanocyte precursors. However other tissues not apparently affected in WWor Sl/Sl mutants also express c-kit and SLF mRNA (Orr-Urtreger et al., 1990; Keshet, et al, 1991). In order to further define the action of SLF and c-kit, Nishikawa et al. used an anti-c-kit antibody, ACK-2, to perturb the development of melanocytes during development and during postnatal life

140

Regulation of the early development of the nervous system by growth factors

(Nishikawa et al., 1991). These studies showed a critical dependency of melanocyte development on C'kit at El3.5-14.5, and effects on coat colour were seen following intradermal injection of pregnant mice as early as El0.5. Even in the adult, intradermal injection of the anti-c-faY antibody when the hair cycle was activated resulted in unpigmented areas. In addition, studies in W^VW^^^ and Sl^/Sl"^ mice showed an almost complete loss of melanocyte precursors by by E13.5 and E12 respectively (Steel et al., 1992; Duttlinger et al., 1993). These studies show a dependency on SLF/c-kit interaction early in melanocyte development but the precise mechanism of action of SLF is not defined. Previous studies suggested a role in migration and/or proliferation (Huszar et al., 1991; Keshet et al, 1991), while our work (Murphy et al., 1992) and that of others (Steel et al, 1992; Morrison-Graham and Weston, 1993) implied a survival role for SLF. Our studies showed that SLF alone did not stimulate melanocyte production, but stimulated an increase in pigmented melanocyte numbers in the presence of 12-0-tetradecanoyl-phorbol-13-acetate (TPA) (Murphy et al., 1992). In combination, these data show a critical role for SLF in melanocyte precursor maintenance. We further studied the role of the SLF/c-kit signalling system in melanocyte development (Reid et al., 1995). We used the anti-c-/:iY antibody, ACK2, to identify c-kit^ cells in neural crest cultures and follow their development in vitro. These c-kit cells appear to give rise only to melanocytes and thus represent melanocyte progenitor cells. In addition, we studied the role of SLF in these cultures and found that it acts as both a survival factor and a proliferative factor for c-kit cells, but does not stimulate their differentiation (Reid et al., 1995). These findings provide a mechanism of regulation of melanocyte development, whereby C'kit is exclusively expressed by melanocyte progenitors within the neural crest precursor population, and subsequent survival and proliferation of these progenitors is regulated by SLF. Presumably, the activity of TPA in the stimulation of pigmentation is mimicking a function nor-

mally found postnatally in the skin at the time of melanocyte differentiation, which is. One possible hormone implicated in melanocyte differentiation is melanocyte-stimulating hormone (MSH) (Ito and Takeuchi, 1984). However, we have found no activity of MSH in the neural crest cultures either in the presence or absence of SI factor (Murphy et al., 1992). Another molecule which might be involved in the differentiation of melanocytes from their precursors is FGF, which enhances the development of pigment in DRG cultures and peripheral nerve (Stocker et al., 1991). In contrast, transforming growth factor-)8 1 inhibits the formation of melanocytes in these cultures and thus may act as a negative modulator in pigment development (Stocker et al, 1991). As stated earlier, it would be of interest to determine if this represents a separate lineage to that which migrates dorso-laterally. Recent results in our laboratory indicate that FGF overides the melanogenic capacity of TPA further suggesting separate identities (Murphy, Reid and Bartlett, unpublished observations). 4. Interaction between growth factors and transcriptional regulators The reductionist approach of attempting to identify discrete signals for differentiation is by necessity, an oversimplification of the regulation of the differentiation process. There is much evidence to suggest that neural differentiation is the result of a complex interplay between environmental signals and genetic predisposition. For example, it has been recently shown that the formation of rhombomeres in the developing hindbrain is not related to the clonal origin of such cells but reflects the position in which a cohort of cells, whose members may have originated at various locations, find themselves (Fraser et al., 1990). Boundaries between rhombomeres coincide with boundaries of expression of particular Hox genes. Whilst it may be that a certain number of the cells are precommitted to express these homeobox genes, it would appear that a process of recruitment is essential. Recruitment of this type requires some epigenetic signalling to take place between cells, and secreted

141

P.F. Bartlett et al.

growth factors such as those mentioned here are prime candidates for this role. The identity of the factors which might influence these processes have clearly not been identified. However there are some examples of growth factors which do influence the expression of particular homeobox genes. A case in vertebrate neurogenesis is the interaction of wnt-1 and engrailed (en). Wnt-1 has characteristics of a growth factor and is expressed at early times in the neural tube (Davis and Joyner, 1988). In mice containing deleted wnt-1 genes, major defects are observed in the midbrain and cerebellum and these defects have been partially correlated with a loss of expression of the en homeobox gene (McMahon et al., 1992), which is also implicated in cerebellum development (Joyner et al., 1991). Thus wnt-1 probably regulates the expression of en, as previously shown for its homologue in Drosophila (van derHeuvaletal., 1989). We propose that this two-way interaction between transcription factors, particularly homeobox genes, and growth factors may explain firstly, the process of lineage committment of nervous system precursor cells and secondly how morphogenesis of nervous system occurs. The demonstration of this awaits studies in which the influence of some of the various growth factors described above on individual neural crest cells can be followed during differentiation. Of course this model is not limited to growth factors, but could be applied to other cell-cell interactive molecules such as N-CAM and, indeed, recent findings (Jones et al., 1992) have shown that Hox gene products can either enhance or inhibit N-CAM production in vitro. Acknowledgements We wish to thank Kate Reid and Viki Likiardopoulos for much of the technical work described in work emanating from our laboratory, and the National Health and Medical Research Council of Australia, the Australian Medical Research and Development Corporation and the Cooperative Research Center for Cellular Growth Factors for providing financial support.

References Abe, E., Tanaka, H., Ishimi, Y., Miyaura, C , Hayashi, T., Nagasawa, H., Tomida, M., Yamaguchi, Y., Hozumi, M. and Suda, T. (1986) Differentiation-inducing factor purified from conditioned medium of mitogen-treated spleen cell cultures stimulates bone resorption. Proc. Natl. Acad. Sci. USA. 83: 5958-5962. Abney, E.R., Bartlett, P.F. and Raff, M.C. (1981). Astrocytes, ependymal cells and oligodendrocytes develop on schedule in dissociated cell cultures of embryonic rat brain. Dev. ^/o/. 83:301-310. Altman, J. (1963). Autoradiographic investigation of cell proliferation in the brains of rats and cats. Anat. Rec. 145: 573-591. Altman, J. (1972). Postnatal development of the cerebellar cortex in the rat. I. The external germinal layer and the transitional molecular layer. J. Comp. Neurol. 145: 353-398. Altman, J. and Bayer, S.A. (1990) Migration and distribution of two populations of hippocampal granule cell precursors during the perinatal and postnatal periods. /. Comp. Neurol. 301:365-381. Alvarez-Buylla, A., Theelen, M. and Nottebohm, F. (1988) Birth of projection neurons in the higher vocal center of the canary forebrain before, during, and after song learning. Proc. Natl. Acad. Sci. USA 85: 8722-8726. Anchan, R.M., Reh, T.A., Angello, J., Balliet, A. and Walker, M. (1991) EGF and TGF-a stimulate retinal neuroepithelial cell proliferation in vitro. Neuron 6: 923-936. Angevine, Jr., J.B. and Sidman, R.L. (1961) Autoradiographic study of cell migration during histogenesis of cerebral cortex in the mouse. Nature 192: 766-768. Anderson, D.J. (1989) The neural crest cell lineage problem: neuropoiesis? A^^Mron 3: 1-12. Anderson, D.J. and Axel, R. (1985) Molecular probes for the development and plasticity of neural crest derivatives. Cell 42, 649-662. Anderson, D.J. and Axel, R. (1986) A bipotential neuroendocrine precursor whose choice of cell fate is determined by NGF and glucocorticoids. Cell 47: 1079-1090. Anderson, D.M., Lyman, S.D., Baird, A., Wignall, J.M., Eisenman, J., Rauch, D., March, C.J., Boswell, H.S., Gimpel, S.D., Cosman, D. and Williams, D.E. (1990) Molecular cloning of mast cell growth factor, a hematopoietin that is active in both membrane bound and soluble forms. Cell 63: 235-243. Bailey, K.A. (1987) Identification of neural cell precursors in the developing mouse brain. PhD. Thesis. University of Melbourne, Australia. Bandtlow, C.E., Heumann, R., Schwabb, M.E. and Thoenen, H. (1987) Cellular localization of nerve growth factor synthesis by in situ hybridization. EMBO J. 4: 891-899. Barald, K.F. (1988a) Monoclonal antibodies made to chick mesencephalic neural crest cells and to ciliary ganglion neurons identify a common antigen on the neurons and a neural crest subpopulation. J. Neurosci. Res. 21: 107-118.

142

Regulation of the early development of the nervous system by growth factors

Barald, K.F. (1988b) Antigen recognized by monoclonal antibodies to mesencephalic neural crest and to ciliary ganglion neurons is involved in the affinity choline uptake mechanism in these cells. /. Neurosci. Res. 21: 119-134. Barald, K.F. (1989) Culture conditions affect the cholinergic development of an isolated subpopulation of chick mesencephalic neural crest cells. Dev. Biol. 135: 349-366. Baroffio, A., Dupin, E. and Le Douarin, N.M. (1988) Clone forming ability and differentiation potential of migratory neural crest cells. Proc. Natl. Acad. Sci. USA 85: 53255329. Baroffio, A., Dupin, E. and Le Douarin, N.M. (1991) Common precursors for neural and mesectodermal derivatives in the cephalic neural crest. Development 112: 301-305. Bartlett, P.P., Reid, H.H., Bailey, K.A. and Bernard, O. (1988) Immortalization of mouse neural precursor cells by the cmyc oncogene. Proc. Natl. Acad. Sci. USA 85: 3255-3259. Baumann, H. and Wong, G.G. (1989) Hepatocyte-stimulating factor III shares structural and functional identity with leukemia-inhibitory factor. 7./mmMno/. 143: 1163-1167. Bazan, J.F. (1991). Neuropoietic cytokines in the hematopoietic fold. Neuron 7: 197-208. Bernard, O., Reid, H.H. and Bartlett, P.F. (1989) The role of c-myc and N-myc proto-oncogenes in the immortalisation of neural precursors. J. Neurosci. Res. 24: 9-20. Bogler, O., Wren, D., Barnett, S.C, Land, H. and Noble, M. (1990). Cooperation between two growth factors promotes extended self-renewal and inhibits differentiation of oligodendrocyte-type-2 astrocyte (0-2A) progenitor cells. Proc. Natl. Acad. Sci. USA 87: 6368-6372. Brill, G., Vaisman, N., Neufeld, G. and Kalcheim, C. (1992) BHK-21-derived cell lines that produce basic fibroblast growth factor, but not parental BHK-21 cells, initiate neuronal differentiation of neural crest progenitors. Development 115: 1059-69. Bronner-Fraser, M. and Eraser, S.E. (1988) Cell lineage analysis reveals multipotency of some avian neural crest cells. Nature 335: 161-164. Bronner-Fraser, M. and Eraser, S.E. (1989) Developmental potential of avian trunk neural crest cells in situ. Neuron 3: 755-766. Carr, V.M. and Simpson, S.B. (1978) Proliferative and degenerative events in the early development of chick dorsal root ganglia. /. Comp. Neurol. 182: 727-740. Cattaneo, E. and McKay, R. (1990) Proliferation and differentiation of neuronal stem cells regulated by nerve growth factor. Nature 347: 762-765. Cepko, C. (1988) Retrovirus vectors and their applications in neurobiology. Neuron 1: 345-353. Cepko, C.L. (1989) Immortalisation of neural cell lines via retrovirus-mediated oncogene transduction. Annu. Rev. Neurosci. 12: 47-55. Chabot, B., Stephenson, D.A., Chapman, V.M., Besmer, P. and Bernstein, A. (1988) The proto-oncogene c-kit encoding a transmembrane tyrosine kinase receptor maps to the mouse W locus. Nature 335: 88-89.

Ciment, G., Glimelius, B., Nelson, D.M. and Weston, J.A. (1986) Reversal of a developmental restriction of neural crest-derived cells of avian embryos by a phorbol ester drug. Dev. Biol 118: 392-398. Copeland, N.G., Gilbert, D.J., Cho, B.C., Donovan, P.J., Jenkins, N.A., Cosman, D., Anderson, D., Lyman, S.D. and Williams, D.E. (1990) Mast cell growth factor maps near the steel locus on mouse chromosome 10 and is deleted in a number of steel alleles. Cell 63: 175-183. Cowan, W.M. (1979) The development of the brain. Sci. Am. 241:106-117. Culican, S.M., Baumrind, N.L., Yamamoto, M. and Pearlman, A.L. (1990) Cortical radial glia: identification in tissue culture and evidence for their transformation to astrocytes./. Neurosci. 10: 684-692. Davies, A.M., Bandtlow, C, Heumann, R., Korsching, S., Rohrer, H. and Thoenen, H. (1987) Timing and site of nerve growth factor in developing skin in relation to innervation and expression of the receptor. Nature 326: 353358. Davis, C.A. and Joyner, A.L. (1988) Expression patterns of the homeobox containing genes En-1 and EN-2 and the protooncogene int-1 diverge during mouse development. Genes Dev. 2: 1736-1744. De Vitry, F., Picart, R., Jacque, C, Legault, L., Dupouey, P. and Tixier-Vidal, A. (1980) Presumptive common precursor for neuronal and glial cell lineages in mouse hypothalamus. Proc. Natl. Acad. Sci. USA 11: 4165-4169. Dicicco-Bloom, E,. Townes-Anderson, E. and Black, LB. (1990) Neuroblast mitosis in dissociated culture: regulation and relationship to differentiation. J. Cell. Biol. 110: 20732086. DiCicco-Bloom, E., Friedman, W.J. and Black, LB. (1993) NT-3 stimulates sympathetic neuroblast proliferation by promoting precursor survival. Neuron 11: 1101-1111. Dodd, J. (1992) Mesodermal control of neural cell identity in vertebrates. Curr. Opin. Neurobiol. 2: 3-8 Doe, C.Q. and Scott, M.P. (1988) Segmentation and homeobox gene function in the developing nervous system of Drosophila. Trends Neurosci. 11: 101-106. Doupe, A.J., Patterson, P.H. and Landis, S.C. (1985a) Environmental influences in the development of neural crest derivatives: glucocorticoids, growth factors and chromaffin cell plasticity. J. Neurosci. 5: 2119-2142. Doupe, A.J., Patterson, P.H. and Landis, S.C. (1985b) Small intensely fluorescent (SIF) cells in culture: role of glucocorticoids and growth factors in their development and phenotypic interconversions with other neural crest derivatives. J. Neurosci. 5: 2143-2160. Drago. J., Murphy, M., Carroll, S.M., Harvey, R.P. and Bartlett, P.F. (1991a) Fibroblast growth factor-mediated proliferation of central nervous system precursors depends on endogenous procuction of insulin-like growth factor 1. Proc. Natl. Acad. Sci. USA 88: 2199-2203. Drago, J., Nurcombe, V., Pearse, M.J., Murphy, M. and Bartlett, P.F. (1991b) Basic fibroblast growth factor

P.F. Bartlett et al. upregulates steady-state levels of laminin Bl and B2 chain mRNA in cultured neuroepithelial cells. Exp. Cell Res. 196: 246-254. Dupin, E., Baroffio, A., Dulac, C, Cameron-Curry, P. and Le Douarin, N.M. (1990) Schwann-cell differentiation in clonal cultures of the neural crest as evidenced by the antischwann cell myelin protein monoclonal antibody. Proc. Natl. Acad. Sci USA 87: 1119-1123. Duttlinger, R., Manova, K., Chu, T.Y., Gyssler, C, Zelenetz, A.D., Bachvarova, R.F. and Besmer, P. (1993) W-sash affects positive and negative elements controlling c-kit expression: ectopic c-kit expression at sites of kit-ligand expression affects melanogenesis. Development 118: 705717. Eisenger, M. and Marko, O. (1982) Selective proliferation of normal human melanocytes in vitro in the presence of phorbol ester and cholera toxin. Proc. Natl. Acad. Sci. USA 79:2018-2022. Ekker, S.C, Young, K.E., Von Kessler, D.P. and Beachey, P.A. (1991) Optimal DNA sequence recognition by the ultrabithorax homeodomain of Drosophila. EMBO J. 10: 1179-1186. Ernsberger, U., Sendtner, M. and Rohrer, H. (1989) Proliferation and differentiation of embryonic chick sympathetic neurons: effects of ciliary neurotrophic factor. Neuron 2: 1275-1284. Ford, M.D., Bartlett, P.F. and Nurcombe, V. (1994) Colocalization of FGF-2 and a novel heparan sulphate proteoglycan in embryonic mouse brain. NeuroReport 5: 565568. Frank, E. and Sanes, J.R. (1991) Lineage of neurons and glia in chick dorsal root ganglia: analysis in vivo with a recombinant retrovirus. Development 111: 895-908. Eraser, S.E. and Bronner-Fraser, M. (1991) Migrating neural crest cells in the trunk of the avian embryo are multipotent. Development \ 12: 913-920. Frederiksen, K., Jat, P.S., Levy, D. and McKay, R. (1988) Immortalization of precursor cells from the mammalian CNS. Neuron 1:439-448. ffrench-Constant, C. and Raff, M.C. (1986a) The oligodendrocyte-type-2 astrocyte cell lineage is specified for myelination. Nature 323: 335-338. ffrench-Constant, C. and Raff, M.C. (1986b) Proliferating bipotential glial progenitor cells in adult rat optic nerve. Nature 319: 499-502. Eraser, S., Keynes, R. and Lumsden, A. (1990) Segmentation in the chick embryo hindbrain is defined by cell lineage restrictions. Nature 344: 431-435. Frederiksen, K. and McKay, R.D.G. (1988) Proliferation and differentiation of rat neuroepithelial precursor cells in vivo. J. Neurosci. S: 1144-1151. Frederiksen, K., Jat, P.S., Valtz, N., Levy, D. and Mckay, R. (1988) Immortalization of precursor Cells from the mammalian CNS. Neuron 1: 439^48. Gage, F.H., Batchelor, P., Chen, K.S., Chin, D., Higgins, G.A., Koh, S., Deputy, S., Rosenberg, M.B., Fischer, W.

143 and Bjorklund, A. (1989) NGF receptor reexpression and NGF-mediated cholinergic neuronal hypertrophy in the damaged adult neostriatum. Neuron 2: 1177-1184. Galileo, D.S., Gray, G.E., Owens, G.C., Majors, J. and Sanes, J.R. (1990) Neurons and glia arise from a common progenitor in chick optic tectum: demonstration with two retroviruses and cell type-specific antibodies. Proc. Natl. Acad. Sci. USA 87: 458-462. Gao, W.-Q., Heintz, N. and Hatten M.E. (1991) Cerebellar granule cell neurogenesis is regulated by cell-cell interactions in vitro. Neuron 6: 705-715. Gearing, D.P., Gough, N.M., King, J.A., Hilton, D.J., Nicola, N.A., Simpson, R.J., Nice, E.C., Kelso, A. and Metcalf, D. (1987) Molecular cloning and expression of cDNA encoding a murine myeloid leukaemia inhibitory factor (LIE). EMBO J. 6: 3995^002. Gearing, D.P., Comeau, M.R., Friend, D.J., Gimpel, S.D., Thut, C.J., McGourty, J., Brasher, K.K., King, J.A., Gillis, S., Mosley, B. et al. (1992) The IL-6 signal transducer, gpl30: an oncostatin M receptor and affinity converter for the LIF receptor. Science 255: 1434-1437. Gensburger, C, Labourdette, G. and Sensenbrenner, M. (1987). Brain basic fibroblast growth factor stimulates the proliferation of rat neuronal precursor cells in vitro. FEBS Lett. 217: 1-5. Geissler, E.N., McFarland, E.G. and Russell, E.S. (1981) Analysis of the pleiotropism at the dominant white-spotting (W) locus of the house mouse: a description of ten new W alleles. Genetics 97: 337-361. Geissler, E.N., Ryan, M.A. and Housman, D.E. (1988) The dominant-w/»'r^ spotting (W) locus of the mouse encodes the c-kit proto-oncogene. Cell 55: 185-192. Gray, G.E., Glover, J.C., Majors, J. and Sanes J.R. (1988) Radial arrangement of clonally related cells in the chicken optic tectum: lineage analysis with a recombinant retrovirus. Proc. Natl. Acad. Sci. USA 85: 7356-7360. Gray G.E., Leber S.M. and Sanes J.R. (1990) Migratory patterns of clonally derived cells in the developing central nervous system. Experientia 46: 929-940. Graziadei, G.A. and Graziadei, P.P. (1979) Neurogenesis and neuron regeneration in the olfactory system of mammals. II. Degeneration and reconstitution of the olfactory sensory neurons after axotomy. J. Neurocytol. 8: 197-213. Guthrie, S. (1992) Lineage in the cerebral cortex: when is a clone not a clone? Trends Neurosci. 15: 273-275. Halaban, R., Pomerantz, S.H., Marshall, S, Lambert, D.T. and Lerner, A.B. (1983) The regulation of tyrosinase in human melanocytes grown in culture. J. Cell. Biol. 97: 480-488. Hendry, I.A., Stockel, K., Thoenen, H. and Iversen, L.L. (1974) The retrograde axonal transport of nerve growth factor. Brain Res. 68: 103-121. Hendry, I.A., Murphy, M., Hilton, D.J., Nicola, N.A. and Bartlett, P.F. (1992) Binding and retrograde transport of leukemia inhibitory factor in the sensory nervous system. J. Neurosci. 12: 3427-3434 Heuer, J.G., von, B.C., Kinoshita, Y., Evers, P.C. and Both-

144

Regulation of the early development of the nervous system by growth factors

well, M. (1990) Alternating phases of FGF receptor and NGF receptor expression in the developing chicken nervous system. Neuron 5: 283-296. Huang, E., Nocka, K., Beier., D.R., Chu, T.-Y., Buck, J., Lahm, H.-W., Wellner, D., Leder, P. and Besmer, P. (1990) The haematopoietic growth factor KL is encoded by the SL locus and is the ligand of the c-kit recptor, the gene product of the \y locus. Cell 63: 225-233. His, W. (1889) Die neuroblasten und deren entstehung im embryonalen mark. Abh. Math. Phys. CI. Kgl. Sach. Ges. Wiss. 15: 313-372. Holt, C.E., Bertsch, T.W., Ellis, H.M. and Harris, W.A. (1988) Cellular determination in the Xenopus retina is independent of lineage and birth date. Neuron 1: 15-26. Hughes, S.M., Lillien, L.E., Raff, M.C., Rohrer, H. and Sendtner, M. (1988) Ciliary neurotrophic factor induces type-2 astrocyte differentiation in culture. Nature 335: 70-73. Huszar, D., Sharpe, A. and Jaenisch, R. (1991) Migration and proliferation of cultured neural crest cells in W mutant neural crest chimeras. Development 112: 131-41. Immergluck, K., Lawrence, P.A. and Bienz, L (1990) Induction across germ layers in Drosophila mediated by a genetic cascade. Cell 62: 261-268. Ip, N.Y., Nye, S.H., Boulton, T.G., Davis, S., Taga, T., Li, Y., Birren, S.J., Yasakawa, K., Kishimoto, T., Anderson, D. J., Stahl, N. and Yancopoulos, G.D. (1992) CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component gp130. Cell 69: 1121-1132. Ito, K. and Sieber-Blum, M. (1991) In vitro clonal analysis of quail cardiac neural crest development. Dev. Biol. 148: 95106. Ito, K. and Takeuchi, T. (1984) The differentiation of the neural crest cells of the mouse embryo. J. Embryol. Exp. Morph. 84: 49-62. Janzer, R.C. and Raff, M.C. (1987) Astrocytes induce bloodbrain barrier properties in endothelial cells. Nature 325: 253-257. Johnson, Jr., E.M., Rich, K.M. and Yip, H.K. (1986) The role of NGF in sensory neurons in vivo. Trends Neurosci. 9: 3337. Jones, F.S., Prediger, E.A., Bittner, D.A., De Robertis, E.M. and Edelman, G.M. (1992) Cell adhesion molecules as targets for Hox genes: neural cell adhesion molecule promoter activity is modulated by cotransfection with Hox-2.5 and 2.4. Proc. Natl. Acad. Sci. USA 89: 2091-2095. Joyner, A.L., Herrup, K., Aurbach, B.A., Davis, C.A. and Rossant, J. (1991) Subtle cerebellar phenotype in mice homozygous for a targeted deletion of the En-2 homeobox. Science 251: 1239-1243. Kalcheim, C. (1989) Basic fibroblast growth factor stimulates survival of non-neuronal cells developing from trunk neural crest. Dev. Biol. 134: 1-10. Kalcheim, C, Carmeli, C. and Rosenthal, A. (1992) Neurotrophin 3 is a mitogen for cultured neural crest cells. Proc. Natl. Acad. Sci. USA 89: 1661-1665.

Kalcheim, C. and Gandreau, M. (1988) Brain-derived neurotrophic factor stimulates survival and neuronal differentiation in cultured avian neural crest. Dev. Brain Res. 41: 79-86. Kalcheim, C. and Neufeld, G. (1990) Expression of basic fibroblast growth factor in the nervous system of developing embryos. Development 109: 203-215. Keshet, E., Lyman, S.D., Williams, D.E., Anderson, D.M., Jenkins, N.A., Copeland, N.G. and Parada, L.F. (1991) Embryonic RNA expression patterns of the c-kit receptor and its cognate ligand suggest multiple functional roles in mouse development. Embo J. 10: 2425-2435. Kessel, M. and Gruss, P. (1991) Homeotic transformations of murine vertebrae and concomitant alteration of Hox codes induced by retinoic acid. Cell 67: 89-104. Kilpatrick, T.J. and Bartlett, P.F. (1993) Cloning and growth of multipotential neural precursors: requirements for proliferation and differentiation. Neuron 10: 255-265. Kilpatrick, T.J. and Bartlett, P.F. (1995) Cloned multipotential precursors from the mouse cerebrum require FGF-2, whereas glial restricted precursors are stimulated with either FGF-2 or EGF. J. Neurosci. 15: 3653-3661. Kilpatrick, T.J., Talman, P.S. and Bartlett, P.F. (1993) The differentiation and survival of murine neurons in vitro is promoted by soluble factors produced by an astrocytic cell line.7. Neurosci. Res. 35: 147-161. Kilpatrick, T.J., Cheema, S.S., Koblar, S.A., Tan, S.S. and Bartlett, P.F. (1994) The engraftment of transplanted primary neuroepithelial cells within the postnatal mouse brain. Neurosci. Lett. 181: 129-133. Lawson, S.N., Caddy, K.W.T. and Biscoe, T.J. (1974) Development of rat dorsal root ganglion neurones. Studies of cell birthdays and changes in mean cell diameter. Cell Tissue Res. 153: 399^03. Leber, S.M., Breedlove, M. and Sanes, J.R. (1990) Lineage, arrangement and death of motoneurons in chick spinal cord. J. Neurosci. 10: 2451-2462. Le Douarin, N.M. (1982) The Neural Crest. Cambridge University Press, London. Le Douarin, N.M. (1986) Cell line segregation during peripheral nervous system ontogeny. Science 231: 1516-1522. Le Douarin, N.M. and Smith, J. (1988) Development of the peripheral nervous sytem from the neural crest. Annu. Rev. Cell. Biol. 4: 375^04. Le Douarin, N.M. and Tiellett (1973) Migration of neural crest cells to the wall of the digestive tract in avian embryo. J. Embryol. Exp. Morphol. 30: 31^8. Leung, D.W., Parent, A.S., Cachianes, G., Esch, F., Coulombe, J.N., Nikolics, K., Eckenstein, F.P. and Nishi, R. (1992) Cloning, expression during development, and evidence for release of a trophic factor for ciliary ganglion neurons. Neuron 8: 1045-1053. Levitt, P., Cooper, M.L. and Rakic, P. (1981) Coexistence of neuronal and glial precursor cells in the cerebral ventricular zone of the fetal monkey: an ultrastructural immunoperoxidase analysis. J. Neurosci. 1: 27-39.

P.F. Bartlett et al Levi-Montalcini, R. and Angeletti, P.U. (1968) The nerve growth factor. Physiol Rev, 48: 534-569. Levi-Montalcini, R. (1982) Developmental neurobiology and the natural history of nerve growth factor. Annu. Rev. Neurosci. 5: 341-362. Lillien, L.E. and Raff, M.C. (1990) Differentiation signals in the CNS: type-2 astrocyte development in vitro as a model system. Neuron 5: 111-119. Lillien, L.E., Sendtner, M. and Raff, M.C. (1990) Extracellular matrix-associated molecules collaborate with ciliary neurotrophic factor to induce type-2 astrocyte development. J. Cell Biol 111:635-644. Lillien, L. and Cepko, C. (1992) Control of proliferation in the retina: temporal changes in responsiveness to FGF and TGFa. Development 115: 253-266. Lo, L.-C, Birren, S.J. and Anderson, D.J. (1991) \-myc immortalization of early rat neural crest cells yields a clonal cell line which generates both glial and adrenergic progenitor cells. Dev. Biol 145: 139-153. Lumsden, A.G.S. (1987) The neural crest contribution to tooth development in the mammalian embryo. In: P.F.A. Maderson (Ed.), Developmental and Evolutionary Aspects of the Neural Crest, John Wiley, New York, pp. 261-300. Lumsden, A.G.S. (1988) Spatial organization of the epithelium and the role of neural crest in the initiation of the mammalian tooth germ. Development 103: 155-169. Luskin, M.B., Pearlman, A.L. and Sanes, J.R. (1988) Cell lineage in the cerebral cortex of the mouse studied in vivo and in vitro with a recombinant retrovirus. Neuron 1: 635647. Matsui, Y., Toksoz, D., Nishikawa, S., Nishikawa, S., Williams, D., Zsebo, K. and Hogan, B.L. (1991) Effect of steel factor and leukaemia inhibitory factor on murine primordial germ cells in culture. Nature 353: 750-752. McGinnis, W. and Krumlauf, R. (1992) Homeobox genes and axial patterning. Cell 68: 283-302. McKinnon, R.D., Matsui, T., Dubois-Dalcq, M. and Aaronson, S.A. (1990) FGF modulates the PDGF-driven pathway of oligodendrocyte development. Neuron 5: 603-614. McMahon, A.P. and Bradley, A. (1990) The Wnt-1 (int-1) proto-oncogene is required for development of a large region of the mouse brain. Cell 62: 1073-1085. McMahon, A.P., Joyner A.L., Bradley, A. and McMahon, J.A. (1992) The midbrain-hindbrain phenotype of Wnt-PAVnt1~ mice results from stepwise deletion of engrailedexpressing cells by 9.5 days post coitum. Cell 69: 581-595. Martinez Arias, A., Baker, N.E. and Ingham, P.W. (1988) Role of segment polarity genes in the definition and maintenance of cell states in the Drosophila embryo. Development 103: 157-170. Martin, F.H., Suggs, S.V., Langley, K.E., Lu, H.S., Ting, J., Okino, K.H., Morris, C.F., Mc Niece, I.K., Jacobsen, F.W., Mendiaz, E.A., Birkett, N.C., Smith, K.A., Johnson, M.J., Parker, V.P., Flores, J.C, Patel, A.C., Fisher, E.F., Eferjavec, J.H.O, Herrera, C.J., Wypych, J., Sachdev, R.K., Pope, J.A., Leslie, I., Wen, D., Lin, C.-H., Cupples, R.L.

145 and Zsebo, K.M. (1990) Primary structure and functional expression of rat and human stem cell factor DNA's. Cell 63:203-211. Matsui, Y., Zsebo, K.M. and Hogan, B.M. (1990) Embryonic expression of a haematopoietic growth factor encoded by the SI locus and the ligand for c-kit. Nature 347: 667-669. Metcalf, D. and Moore, M.A.S. (1971) Embryonic aspects of haemopoiesis. In: A. Neuberger and E.L. Tatum (Eds.), Haemopoietic Cells, North-Holland, Amsterdam, pp. 70271. Miller, R.H., David, S., Patel, R., Abney, E.R. and Raff, M.C. (1985) A quantitative immunohistochemical study of macroglial cell development in the rat optic nerve: in vivo evidence for two distinct astrocytic lineages. Dev. Biol. 111:35^1. Mintz, B. (1967) Gene control of mammalian pigmentary differentiation. 1. Clonal origin of melanocytes. Proc. Natl Acad. Scl USA 58: 344-351. Morrison-Graham, K. and Weston, J.A. (1993) Transient Steel factor dependence by neural crest-derived melanocyte precursors. Dev. Biol 159: 346-352. Morshead, CM. and van der Kooy, D. (1992) Postmitotic death is the fate of constitutively proliferating cells in the subependymal layer of the adult mouse brain. J. Neurosci. 12: 249-256. Mountcastle, V.B. (1957) Modality and topographic properties of single neurons of cat's sensory cortex. J. Neurophysiol 20: 408-434. Murphy, M., Drago, J. and Bartlett P.F. (1990) Fibroblast growth factor stimulates the proliferation and differentiation of neural precursor cells in vitro. J. Neurosci Res. 25: 463-475. Murphy, M., Bernard, O., Reid, K. and Bartlett, P.F. (1991a) Cell lines derived from mouse neural crest are representative of cells at various stages of differentiation. J. Neurobiol 11: 522-535. Murphy, M., Reid, K., Hilton, D.J. and Bartlett, P.F. (1991b) Generation of sensory neurons is stimulated by leukemia inhibitory factor. Proc. Natl Acad. Scl USA 88: 3498-3501. Murphy, M., Reid, K., Williams, D.E., Lyman, S.D. and Bartlett, P.F. (1992) Steel factor is required for maintenance, but not differentiation, of melanocyte precursors in the neural crest. Dev. Biol 153: 396-401. Murphy, M., Reid, K., Brown, M.A. and Bartlett, P.F. (1993) Involvement of leukemia inhibitory factor and nerve growth factor in the development of dorsal root ganglion neurons. Development ni: 1173-1182 Murphy, M., Reid, K., Ford, M., Furness, J. and Bartlett, P. F. (1994) Regulation of sensory neuron development from neural crest cells: FGF2 stimulates cell division division of precursors but LIF is required for full differentiation. Development 120: 3519-3528. Newgreen, D.F., Jahnke, I., Allan, l.J. and Gibbons, I.L. (1980) Differentiation of sympathetic and enteric neurons of the fowl embryo in grafts to the chorio-allantoic membrane. Cell Tissue Res. 208: 1-19.

146

Regulation of the early development of the nervous system by growth factors

Nishikawa, S., Kusakabe, M., Yoshinaga, K., Ogawa, M., Hayashi, S., Kunisada, T., Era, T., Sakakura, T. and Nishikawa, S. (1991) In utero manipulation of coat color formation by a monoclonal anti-c-kit antibody: two distinct waves of c-kit-dependency during melanocyte development. EMBOJ. 10:2111-2118. Noble, M., Murray, K., Stroobant, P., Waterfield, M.D. and Riddle, P. (1988) Platelet-derived growth factor promotes division and motility and inhibits premature differentiation of the oligodendrocyte/type-2 astrocyte progenitor cell. Nature 333: 560-562. Noden, D.M. (1978) The control of avian cephalic neural crest cytodifferentiation. Dev. Biol 67: 296-312. Noden, D.M. (1983) The role of the neural crest in patterning of avian cranial skeletal, connective, and muscle tissues. Dev Biol. 96: 144-165. Noden, D.M. (1986) Patterning of avian craniofacial muscles. Dev. Biol. 116:347-356. Nordeen, E.J. and Nordeen, K.W. (1989) Estrogen stimulates the incorporation of new neurons into avian song nuclei during adolescence. Dev. Brain Res. 49: 27-32. Nornes, H.O. and Das, G.D. (1974) Temporal pattern of neurogenesis in spinal cord of rat. I. An autoradiographic study-time and sites of origin and migration settling patterns of neuroblasts. Brain Res. 73: 121-138. Nurcombe, V., Ford, M., Wildshut, J.A. and Bartlett, P.P. (1993) Developmental regulation of neural response to FGF-1 and FGF-2 by heparan sulfate proteoglycan. Science 260: 103-106. Nusse, R. and Varmus, H.E. (1992) Wnt genes. Cell 69: 1073-1087. Orr-Urtreger, A., Avivi, A., Zimmer, Y., Givol, D., Yarden, Y. and Lonai, P. (1990) Developmental expression of c-kit, a proto-oncogene encoded by the W locus. Development 109:911-923. Panganiban, G., Reuter, R., Scott, M.P. and Hoffman, F.M. (1990) A Drosophila growth factor homolog, decapentaplegic, regulates homeotic gene expression within and across germ layers during midgut morphogenesis. Development UO: 1041-1050. Patterson, P.H. (1990) Control of cell fate in a vertebrate neurogenic lineage. Cell 62: 1035-1038. Patterson, P.H. and Nawa, H. (1993) Neuronal differentiation factors/cytokines and synaptic plasticity. Cell 72: 123-137. Price, J., Turner, D.L. and Cepko, C.L. (1987) Lineage analysis in the vertebrate nervous system by retrovirus-mediated gene transfer. Proc. Natl. Acad. Sci. USA 84: 156-160. Price, J. and Thurlow, L. (1988) Cell lineage in the rat cerebral cortex: a study using retro viral-mediated gene transfer. Development 104: 473^82. Raff, M.C. (1989) Glial cell diversification in the rat optic nerve. Science 243: 1450-1455. Raff, M.C, Miller, R.H. and Noble, M. (1983a) A glial progenitor cell that develops in vitro into an astrocyte or an oligodendrocyte depending on culture medium. Nature 303: 390-396.

Raff, M.C, Abney, E.R., Cohen, J., Lindsay, R. and Noble, M. (1983b) Two types of astrocytes in cultures of developing rat white matter: differences in morphology, surface gangliosides, and growth characteristics. /. Neurosci. 3: 1289-1300. Raff, M.C, Lillien, L.E., Richardson, W.D., Burne, J.F. and Noble, M.D. (1988) Platelet-derived growth factor from astrocytes drives the clock that times oligodendrocyte development in culture. Nature 333: 562-565. Raff, M.C. (1989) Glial cell diversification in the rat optic nerve. Science 243: 1450-1455. Rawles, M.E. (1944) The migration of melanoblasts after hatching into pigment-free skin grafts of the common fowl. Physiol. Zool. 17: 167-183, Reid, K., Nishikawa, S.-I., Bartlett, P.F. and Murphy, M. (1995) Steel factor directs melanocyte development in vitro through selective regulation of the number of c-kit^ progenitors. Dev. Biol 169: 568-579. Renfranz, P.J., Cunningham, M.G. and McKay, R.D. (1991) Region-specific differentiation of the hippocampal stem cell line HiB5 upon implantation into the developing mammalian brain. Cell 66: 713-729. Reuter, R., Panganiban, G.E.F., Hoffman, F.M. and Scott, M.P. (1990) Homeotic genes regulate the spatial expression of putative growth factors in the visceral mesoderm of Drosophila embryos. Development 110: 1031-1040. Reynolds, B.A., Tetzlaff, W. and Weiss, S. (1992) A multipotent EGF-responsive striatal embryonic progenitor cell produces neurons and astrocytes. J. Neurosci 12: 45654574. Reynolds, B.A. and Weiss, S. (1992) Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science 255: 1707-1710. Richards, L.J., Kilpatrick, T.J. and Bartlett, P.F. (1992) De novo generation of neuronal cells from the adult mouse brain. Proc. Natl Acad. Scl USA 89: 8591-8595. Richardson, W.D., Pringle, N., Mosley, M.J., Westermark, B. and Dubois-Dalcq, M. (1988) A role for platelet-derived growth factor in normal gliogenesis in the central nervous system. Ce//53: 309-319. Russel, E.S. (1979) Hereditary anemias of the mouse; a review for geneticists. Adv. Genet. 20: 357-459. Ryder, E.F., Snyder, E.Y. and Cepko, C.L. (1990) Establishment and characterization of multipotent neural cell lines using retrovirus vector-mediated oncogene transfer. /. NeuroMo/. 21:356-375. Sanes, J.R., Rubenstein, J.L.R. and Nicolas, J.-F. (1986) Use of a recombinant retrovirus to study post-implantation cell lineage in mouse embryos. EMBO J. 5: 3133-3142. Schaper, A. (1897) The earliest differentiation in the central nervous system of vertebrates. Science 5: 430-431. Schlessinger, A.R., Cowan, W.M. and Gottlieb, D.I. (1975) An autoradiographic tudy of the time of origin and the pattern of granule cell migration in the dentate gyrus of the rat. /. Comp. Neurol 159: 149-175. Schmechel, D.E. and Rakic, P. (1979) Arrested proliferation

P.F. Bartlett et al. of radial glial cells during midgestation in rhesus monkey. Nature 277: 303-305. Sham, M.-H., Vesque, C, Nonchev, S., Marshall, H., Frain, M., Gupta, R.-D., Whiting, J., Wilkinson, D., Charnay, P. and Krumlauf, R. (1993) The zinc finger gene KroxlO regulates HoxB2 (Hox2.8) during hindbrain development. Cell 12: 183-196. Sieber-Blum, M. (1989) Commitment of neural crest cells to the sensory lineage. Science 243: 1608-1610. Sieber-Blum, M. (1991) Role of the neurotrophic factors BDNF and NGF in the commitment of pluripotent neural crest cells. Neuron 6: 949-955. Sieber-Blum, M. and Cohen, A.M. (1980) Clonal analysis of quail neural crest cells: they are pluripotent and differentiate in vitro in the absence of noncrest cells. Dev. Biol. 80: 96-106. Simeone, A., Gulisano, M., Acampora, D., Stornaiuiolo, A., Rambaldi, M. and Boncinelli, E. (1992) Two vertebrate homeobox genes related to the Drosophila empty spiracles gene are expressed in the embryonic cerebral cortex. EMBO J. 11:2541-2550. Snyder E.Y., Deitcher D.L., Walsh C, Arnold-Aldea S., Hartwieg E.A., Cepko C.L. (1992) Multipotent neural cell lines can engraft and participate in development of mouse cerebellum. Ce//68: 33-51. Steel, K.P., Davidson, D.R. and Jackson, I.J. (1992) TRP2/DT, a new early melanoblast marker, shows that steel growth factor (c-kit ligand) is a survival factor. DevelopmentUS: 1111-1119. Stocker, K.M., Sherman, L., Rees, S. and Ciment, G. (1991) Basic FGF and TGF-beta 1 influence commitment to melanogenesis in neural crest-derived cells of avian embryos. Development 111: 635-645. Temple, S. (1989) Division and differentiation of isolated CNS blast cells in microculture. Nature 340: 471-473. Temple, S. and Raff, M.C. (1985) Differentiation of a bipotential glial progenitor cell in single cell microculture. A^«mre 313: 223-225. Theiler, K. (1972) The House Mouse, Springer-Verlag, Berlin. Thoenen, H. and Barde, Y.-A. (1980) Physiology of nerve growth factor. Physiol Rev. 60: 1284-1335. Turner, D.L. and Cepko, C.L. (1987) A common progenitor for neurons and glia persists in rat retina late in development. Nature 328: 131-136. Turner, D.L., Snyder, E.Y. and Cepko, C.L. (1990) Lineageindependent determination of cell type in the embryonic mouse retina. Neuron 4: 833-845. van der Heuval, M., Nusse, R., Johnson, P. and Lawrence P.A. (1989) Distribution of the wingless gene product in Drosophila embryos: a protein involved in cell-cell communication. Cell 59: 739-749. Walsh, C, and Cepko, C.L. (1992) Widespread dispersion of neuronal clones across functional regions of the cerebral coxiQX.Science 255: 434-440. Wanaka, A., Milbrandt, J. and Johnson, E. J. (1991) Expression of FGF receptor gene in rat development. Development 111:455^68.

147 Watanabe, T. and Raff, M.C. (1990) Rod photoreceptor development in vitro: intrinsic properties of proliferating neuroepithelial cells change as development proceeds in the rat XQXm2L.Neuron 4: 461-467. Weston, J.A. (1963) A radioautographic analysis of the migration and localization of trunk neural crest cells in the chick. Dev. Biol. 6: 279-310. Weston, J.A. (1986) Phenotypic diversification in neural crestderived cells: the time and stability of commitment during early development. Curr. Top. Dev. Biol. 20: 195-210. Weston, J.A. (1991) Sequential segregation of developmentally destricted intermediate cell populations in the neural crest lineage. Curr. Top. in Dev. Biol. 25: 133-153. Wetts, R. and Eraser, S.E. (1988) Multipotent precursors can give rise to all major cell types of the frog retina. Science 239: 1142-1145. Wilkinson, D.G., Bailes, J.A. and McMahon, A.P. (1987) Expression of the proto-oncogene int-\ is restricted to specific neural cells in the developing mouse embryo. Cell 50; 79-88. Williams, B.P., Read, J. and Price J. (1991) The generation of neurons and oligodendrocytes from a common precursor cell. Neuron 7: 685-693. Williams, D.E., Eisenman, J., Baird, A., Rauch, C, Van Ness, K.V., March, C.J., Park, L.S., Martin, U., Mochizuki, D.Y., Boswell, H.S., Burgess, G.S., Cosman, D. and Lyman, S.D. (1990) Identification of a ligand for the c-kit protooncogene. Cell 63: 167-174. Williams, R.L., Hilton, D.J., Pease, S., Willson, T.A., Stewart, C.L., Gearing, D.P., Wagner, E.F., Metcalf, D., Nicola, N.A. and Gough, N.M. (1988) Myeloid leukaemia inhibitory factor maintains the developmental potential of embryonic stem cells. Nature 336: 684-687. Wolswijk, G. and Noble, M. (1989) Identification of an adultspecific glial progenitor cell. Development 105: 387^00. Wright, E.M., Vogel, K.S. and Davies, A.M. (1993) Neurotophic factors promote the maturation of developing sensory neurons before they become dependent of these factors for survival. Neuron 9: 139-150. Yamamori, T., Fukada, K., Aebersold, R., Korsching, S., Fann, M.-J. and Patterson, P.H. (1989) The cholinergic neuronal differentiation factor from heart cells is identical to leukemia inhibitory factor. Science 241: 1412-1416. Zsebo, K.M., Williams, D.A., Geissler, E.N., Broudy, V.C., Martin, F.H., Atkins, H.L., Hsu, R.Y., Birkett, N.C., Okino, K.H., Murdock, D.C., Jacobsen, F.W., Langley, K.E., Smith, K.A., Takeishi, T., Cattanach, B.M., Galli, S.J. and Suggs, S.V. (1990a) Stem cell factor is encoded at the SI locus of the mouse and is the ligand for the c-kit tyrosine kinase receptor. Cell 63: 213-223. Zsebo, K.M., Wypych, J., McNiece, I.K., Lu, H.S., Smith., K.A., Karkare, S.B., Sachdev, R.K., Yuschenkoff, V.N., Birkett, N.C., Williams, L.R., Satyagal, V.B., Tung, W., Bosselman, R.A., Mendiaz, E.A. and Langley, K.E. (1990b) Identification, purification and biological characterization of haematopoietic stem cell factor from Buffalo rat livercondtioned medium. Cell 63: 195-201.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 6

Retrograde factors in peripheral nerves Ian A. Hendry Neurobiology Research Group, Division of Neuroscience, The John Curtin School of Medical Research, The Australian National University, G.P.O. Box 334, Canberra, A.C.T. 2601, Australia

1. Introduction The problem of understanding the mechanism by which a neuron achieves the correct innervation of its target tissue has remained unsolved by neurobiologists. During development a neuron must grow an axon from its cell body, choose the correct pathway and eventually synapse with its correct target. There have been many theories proposed to explain the extremely accurate wiring in the nervous system, and all involve some interaction between the target cell and its innervating neuron. The most extreme of these was the chemoaffinity hypothesis of Sperry (1951) that suggested a unique chemical coding between each innervating neuron and its target cell. It is unlikely, however, that there would be sufficient genetic information in a vertebrate to achieve such a coding within the central nervous system. It is more likely that there is a hierarchy of strategies used by the neuron to achieve these connections with different neuronal populations using various combinations of common themes. Guidance factors along the potential pathway and recognition factors in the target tissue will be critical for correct development Some of these elements may be lost in adults accounting for their impaired ability to regenerate. For example, pathway guidance for the developing axons may be along routes that are present only for short times during the formation of the embryo or consist of transiently expressed components of the extracellular matrix and cell surface adhesion molecules on the cells in the axon's path. One-on-one communication between the target cell and the potential innervating neuron com-

mences when the growth cone reaches the region of the target tissue and the first filopodial tip palpates the surface of the target cell. A complex exchange of information must then occur, ranging from affinities due to matching of cell surface adhesion molecules and thus selection of best matching pairs, through to two-way activation of receptors leading to alteration of nuclear expression of proteins. During regeneration, the growth of sprouts and elongation of the axon recapitulates the axon growth during embryonic development, although there are several significant differences between regeneration and development. The neuron can maintain an axon not connected to its target for a considerable time during regeneration and the guidance cues available in the embryo to the sprouting axon may not always still be present in the adult. Final maturity of the regenerating nerve and its presynaptic contacts are totally dependent on functional contact with the target tissue. Failure to innervate the target tissue results in persistence of chromatolysis and, eventually, death of the neuron (Hendry, 1975b; Matthews and Nelson, 1975; Purves, 1975). The main thrust of this review will be to examine in detail the potential ways in which a target cell can signal to the nucleus of its innervating neuron relevant information leading to the survival of the most appropriate cell to contact it and the resultant production of the exquisitely complex wiring of the adult vertebrate nervous system. 2. Developmental cell death No analysis of the formation of neuronal connec-

150

tions can be made without an understanding of developmental cell death. In most nerve centres there is a wave of cell death that occurs at the precise time that the target tissue is being innervated. At this stage a large proportion usually more than 50%, of the total neuronal population dies (Hamburger and Levi-Montalcini, 1949). In addition, if the target tissue is removed prior to innervation nearly all the neurons that are destined to supply it will die at the time when innervation would have taken place (Hamburger, 1934). An artificial increase in the size of the target field (Hamburger, 1939), or a reduction in the numbers of neurons innervating the same size of target, leads to an increase in the extent of neuronal survival (Pilar et al., 1980) showing that it is the ratio between the number of innervating neurons and the size of the target that is important for the final number of surviving neurons. Thus the target tissue must provide some message to the innervating neuron which is essential for its survival. The earliest experiments suggested that the number of surviving neurons innervating a target tissue was regulated by the size of that target tissue (Prestige, 1974). It was not known whether this was due to an increase in neuronal proliferation due to a mitogenic factor, or due to a decrease in cell death caused by a survival factor. While the final numbers of neurons in any nerve centre seem to depend on the size of the available target, regional variations can occur due to differential neuroblast proliferation, for example in the chick embryo spinal cord (Oppenheim et al., 1989). The earliest results leading to the discovery of nerve growth factor (NGF) suggested that this was a diffusible agent which caused an increase in the numbers of neurons in a nerve centre by an effect on proliferation (Levi-Montalcini, 1966). It was subsequently shown that NGF had its main action in causing neuronal survival rather than proliferation (Hendry, 1977b) and it was able to act not only via the general circulation but also by transport within the axons from the target tissue to the neuronal perikarya (Hendry, 1977a). It is the fact that this retrograde transport is confined within individual axons that provides the specificity of the communication between a target and its neuron.

Retrograde factors in peripheral

nerves

2.1. Mechanisms of survival and cell death Protein and RNA synthesis are required in the embryo for both naturally occurring and lesioninduced cell death (Oppenheim et al., 1990) and, in addition, the death of cultured neurons after the removal of NGF is prevented by inhibitors of RNA and protein synthesis (Martin et al., 1988). Thus neuronal cell death is an active process requiring biosynthetic events. It seems likely that neurotrophic factors have a dual function, both stimulating genes that promote survival and differentiation and suppressing genes that would kill the cell (Oppenheim et al., 1990). The finding that interferon retards cell death in cultures of sympathetic neurons after NGF withdrawal is intriguing and suggestive of a role for 2',5'-oligoadenosine synthetase in the process of neuronal rescue (Chang et al., 1990). As the product of this enzyme can activate an RNAse, interferon may interrupt the death program by causing the degradation of mRNA critical for this program. There are many factors involved in the regulation of cell death which may act synergistically and one attractive model for developmental control is the interaction between presynaptic connections and postsynaptic factors. For example, in the developing sympathetic nervous system both presynaptic and target tissue influences control the final numbers of neurons in the superior cervical ganglion (SCG) (Black et al., 1972; Hendry, 1973, 1975a). The number of embryonic chick spinal motoneurons that survive during the period of naturally occurring cell death is influenced by factors from both the target tissue (Hamburger, 1934, 1975; Oppenheim, 1981) and an intact descending afferent system (Okado and Oppenheim, 1984). Two distinct factors have been isolated from muscle and spinal cord which clearly promote motoneuron survival and there is synergy between these factors (Dohrmann et al., 1987). Naturally occurring cell death is enhanced in sympathetic and parasympathetic ganglia after blockade of ganglionic neurotransmission with pempidine (Hendry, 1973; Maderdrut et al., 1988). On the other hand blockade of activity of the target with curare leads to an increase in survival of mo-

LA. Hendry

151

toneurons (Oppenheim, 1981, 1984) and, conversely, increase in activity by electrical stimulation of the limb muscles in the chick embryo increases cell death (Oppenheim and Niifiez, 1982). Similarly, neuronal survival is enhanced by blockade of postsynaptic transmission. Some of these apparently activity-related phenomena may be due to changes in the target size, for example, sympathetic preganglionic neuron cell death is reduced by treatment with NGF (Oppenheim et al., 1982a) and hemicholinium (Oppenheim et al., 1982b). This may be due to the enlargement of the sympathetic ganglia and the supply of the preganglionic requirement for neurotrophic factors by the hypertrophied target. Taken together, these results suggest that increased activity in presynaptic neurons regulates cell survival enhancing the formation of correct connections. Increased activity, independent of neuronal firing, in the postsynaptic target can promote cell death as the target is already innervated by more appropriate neurons.

Nerve Cell Bodies Early axonal outgrowth

Critical Period

3. Mechanisms to achieve correct connections Developmental cell death occurs at the precise time of innervation and this has led most investigators to speculate that it must be involved in the selection of the correct neurons. This has led to the development of the neurotrophic theory of cell death (Hendry, 1976) which is outlined in Fig. 1. This theory suggests that the target releases limiting amounts of a neurotrophic factor such that the nerves innervating the tissues compete for the factor. Those that compete successfully will survive and those that do not make appropriate or sufficient contacts will die. In addition, neurons that do not make contact with their correct target, or contact an incorrect target providing an inappropriate neurotrophic factor, also die (Fig. 1). However, there are many lines of evidence that suggest the simple concept that neurons making the correct connections live while those that make incorrect connections die is unlikely to be the whole truth. There are several possibilities as to

Target Tissue

a

c^

NGF

<^

Correct connections leading to retrograde axonal transport of NGF Incorrect connection fail to transport NGF

Survival of neurones that make correct connections Post target innervation Neuronal death Fig. 1. Schematic diagram outlining the neurotrophic theory of cell death. Neurons extend axons in the presence of a low circulating level of neurotrophic factors. At a critical point in their development, usually corresponding to the period of target tissue innervation, the neuron becomes acutely dependent on the neurotrophic factor for its survival. At this stage those that make appropriate connections in the periphery obtain sufficient factor for their survival. Those not making correct connections do not obtain sufficient factor and die.

152

how a nerve makes the correct contact with its target and these are by no means mutually exclusive. 3. L Pathway to target tissue The mechanisms for guidance have been extensively reviewed (Tosney and Landmesser, 1985b; Hill and Vidovic, 1992) and will only be briefly touched on here. The main point is that for many neuronal populations there is a stereotype of patterned outgrowth, such that they all grow along the correct pathways to their target region with little or no deviations along the route (Mark, 1980; Tosney and Landmesser, 1985a). A large number of mechanisms for this has been proposed and all are likely to play a role (Landmesser, 1980; Purves, 1988). These include genetic preprogramming of fibre direction or gradients, topographical organisation, of connections, birthdates of neurons and the timing of axon outgrowth. Nerves in a nerve centre all grow to the same target as, for example, is seen for chick and amphibian motoneurons (Mark, 1980). There may be growth of axons along preformed anatomical pathways such as blood vessels. Axons may adhere to specific molecules along their route in the extracellular matrix or on surfaces of cells (such as guidepost cells in invertebrates). Gradients of trophic factors may be present in the pathways leading to chemotactic direction of growth, although stable gradients over long distances are unlikely and such gradients are more likely to be based on the matrix or on cell surfaces. In the chick motoneuron system, it has been shown that all the neurons in a motor nucleus destined to innervate a certain muscle, in fact, grow directly to that muscle and, even at the earliest stages of innervations do not appear to make any gross errors in the muscle they contact (Tosney and Landmesser, 1985a). The retino-tectal system has been extensively examined and it would appear that even at the earliest stages of the innervation of the tectum there are only minor mistakes in innervation (Marotte, 1990). Thus these neurons, at the grossly anatomical level, get to the correct target. It may be that there is a more subtle difference so that there is, within the muscle, a topography that requires to be established

Retrograde factors in peripheral nerves

and thus that it is not the innervation of the whole muscle but of individual fibres within it that is important. In the frog muscle such a change in the topographical innervation of the muscle has been shown (Noakes et al., 1983, 1988). This finding seems to be the exception rather than the rule, however, and the results on which the conclusions are based have been suggested to be ambiguous due to the potential electrical coupling of these immature muscles. 3.2. Target selection At the time of initial target tissue innervation the embryo is very small and the filopodia of the growth cones can contact much of the tissue in the target region. Thus it is possible that, over a few hundred microns, surface adhesion molecules could select the correct neuron-target match. Similarly over the same short distances, chemotactic agents could result in the growth cone being directed to the correct target. Target-derived neurotrophic agents may act in two ways: firstly, to stabilize appropriate synaptic contacts and, secondly, to promote the survival of appropriate neurons making correct connections. The above discussion of course leaves open the question as to which is the ^appropriate' neuron to make the 'correct' connection with a target, and it is a matter of semantics as to which neurons are correct. If there are embryonic guidance cues that result in all the neurons from a specific part of the spinal cord reaching a specific muscle, and we define the neurons within this part of the spinal cord as belonging to the motoneuron pool for that muscle, then there can only be 'correct' innervation. If. however, there are within that region of the spinal cord neurons that for various reasons are not suitable to innervate that muscle, then these 'correct' neurons will in fact be wrong. They may be anatomically correct, in that they come from the region of the spinal cord that provides the correct adult connections. However, not all of the embryonic neurons in this region of the cord may be functionally correct and. therefore, may be inappropriate, in spite of coming from the correct area. Furthermore, in order to be functionally correct, a

LA. Hendry

nerve must receive the appropriate presynaptic inputs as well as make contact with the appropriate target. The neurons from the correct part of the cord that make presynaptic connections appropriate for other muscles will need to be eliminated. Thus we must distinguish between anatomically correct and functionally correct, as some neurons that die have made anatomically correct connections (Prestige, 1976). The factors that led to cells from a specific area of the cord being guided to a specific muscle will relate to the guidance cues along the pathway and the response of the growth cones to it. On the other hand, the factors that control the presynaptic innervation of the cell in the cord will relate to the guidance cues for its innervating neuron, its own cell surface markers and neurotrophic signals it provides to its innervating neuron. It has been shown that spinal motoneurons in the chick embryo die following deafferentation and this can be reversed by neurotrophic agents (Yin et al., 1994). Cell death can therefore result from incorrect functional connections rather than anatomical ones under these circumstances. 4. Phenotypic determination by target tissue It is not known whether functional specificity is a property of the nerve fibre itself or is due to the end organ with which it makes contact. It is possible that the fibre and target are already specified, or that the target may confer specificity on its innervating neuron or, finally, that the neuron may guide formation of the target, as seen in the development of the taste buds. In the regenerating sympathetic nervous system the role of the presynaptic reinnervation with regard to function is less clear. Cross anastomosis of skin and muscle sympathetic neurons results in both appropriate and inappropriate changes in the reflexes shown by these neurons (Janig and Koltzenburg, 1991) and in this study it was concluded that the reflex organization of sympathetic neurons can change qualitatively following nerve lesion when sympathetic neurons regenerate and supply inappropriate target tissues. Thus there is a rearrangement of the central connections to generate these reflex changes. The sig-

153

nal for the central rearrangement must be derived from the target tissue to change either the phenotype of the postganglionic neuron or the recognition signals it generates so that it now accepts new and appropriate connections from the preganglionic neurons. Alternatively, the retrograde signal could be transmitted trans-synaptically through the ganglion and alter the phenotype of the preganglionic neuron to alter the connections it receives. Such a mechanism could be mediated by the transsynaptic transfer of a signal molecule derived from the target tissue. The only molecule to show such a trans-synaptic transfer has been tetanus toxin. An equivalent endogenous molecule has yet to be shown. The third mechanism for this alteration of connections could be by regulation of the neurotrophic factor produced by the ganglion neuron which could promote connection of the appropriate preganglionic neurons or result in a phenotypic change in these neurons so they regulate their connections. 4,1. Target-controlled regulation of central connections Sympathetic and sensory neurons from a given nerve centre may grow out more or less randomly and innervate targets of opportunity along their route (Langley, 1895). For most sympathetic innervation this appears to be the case, as there is no topographical representation in the sympathetic ganglion for the target, although the innervation of the mesenteric vessels and enteric neurons seem to form two populations with separate specificities for target innervation (Hill et al., 1983). It is possible that the target itself is able to modify the phenotype of its innervating neuron, in order for it to select the appropriate presynaptic innervation. Under these circumstances all neurons could be functionally appropriate. Neurons may, however, still be required to contact specific target types (for example, blood vessels or glandular tissues) in order to receive the appropriate neurotrophic factors for their survival, and this more subtle level of specificity may involve cell death. The regulation of the innervation of ganglionic neurons is clearly under neurotrophic control, as

154

can be seen by the abnormal innervation pattern resulting from NGF treatment during the perinatal period. NGF treatment results in the abnormal ingrowth of sensory neurons identified by their calcitonin gene-related peptide content. These abnormal sensory neurons apparently compete with the preganglionic fibres for their neurotrophic factor such that the preganglionic fibres fail to develop as seen by the failure of the preganglionic marker enzyme choline acetyltransferase (EC 2.3.1.6) (CAT) to develop normally (Hendry et al., 1992). Weiss (1942) proposed the hypothesis that sensory neurons become modulated during embryogenesis by the biochemical character of their endorgans and that this acquired characteristic directs the formation of appropriate central synaptic connections. There is a difference between the motor system and the sensory system. The evidence on the motor side favours a parallel differentiation of motoneuron pools in the spinal cord and muscles in the periphery. The correspondence between motoneurons and muscles is then made by prespecified motoneurons growing out to form connections with appropriate muscles (Mark, 1980). Sensory neurons do not have their peripheral connections controlled to such an extent, although their birthdates and segmental localizations do have an influence (Baker et al., 1978). Axolotls develop appropriate reflex wiping movements after cross anastomoses of sensorimotor nerves and reinnervation of the skin (Griersmith and Mark, 1982). Evidence suggests that the underlying central nervous system connectivity is not the result of autonomous differentiation in the dorsal root ganglia paralleling that of the skin, but is imposed by information that is not available to them until they receive morphogenetic signals from the skin (Griersmith and Mark, 1982). Thus, in the development of the normal reflex it would appear that there is an orthograde pattern of selection of connections such that the peripheral end-organ specifies the sensory neuron which, in turn, seeks out the motoneuron which has already sought out its correct muscle. The role of the target may become more important during regeneration. The effectiveness of

Retrograde factors in peripheral nerves

functional recovery may be influenced by the target tissue which, in some ways, may alter the innervating neuron to produce a more suitable input. For example. Leukemia inhibitory factor/cholinergic differentiating factor synthesis is increased after peripheral nerve injury in the adult (Banner and Patterson, 1994) and this can induce an injury response in the nerves but does not lead to a neurotransmitter switch (Rito et al., 1994). One case of regeneration best explained by this phenomenon is in the sympathetic nervous system where, after lesion of the postganglionic nerve trunk, there is a return of full function after reinnervation despite the finding that anatomically, many reinnervating neurons are incorrect in that they do not return to their original target (Hendry et al., 1986). To demonstrate this the retrogradely transported fluorescent dyes fast blue, was injected into the terminal regions of the neurons of the sympathetic ganglion. Fast Blue has the property of remaining within neuronal perikaryon for many months after its initial retrograde axonal transport from the peripheral tissue, even when the axon of the neuron is subsequently transected (Hendry et al., 1986). Thus, neurons can be labelled as to their initial target projection with Fast Blue and the specificity of subsequent regeneration can be followed using a second dye transported from the same or a different target. When such experiments were carried out on the regenerating SCG it became clear that a recovery of function can occur in a reinnervated target organ in the presence of a large component of fibres that were not originally to be found in that target. This can best be seen in an experiment where Fast Blue was injected into the anterior chamber of the eye and a week later the internal carotid nerve was cut. After regeneration of the sympathetic nervous system normal function returned to the iris and levator occuli. Another dye, Diamidino Yellow, was then injected into the submandibular gland to determine the projections of the regenerated axons. The surprising finding was that there was only a very small percentage of correct reinnervations in spite of a complete functional recovery (Hendry et al., 1986). This specificity of functional recovery must involve the presynaptic connections to the sympa-

I.A. Hendry

thetic neuron. It has been shown that at the segmental level there is correct presynaptic reinnervation (Langley, 1897; Nja and Purves, 1978). Axon branching allows a small number of neurons to have widespread peripheral effects (Langley, 1900). The repeated branching of sympathetic fibres at the periphery provides for axon reflexes which further widens the sympathetic influences at the periphery. Thus it is possible that few correct neurons may be responsible for the functional reinnervation appearing to be correct with the incorrect units being turned off. Alternatively, the incorrectly projecting neurons are capable of plasticity in their central connections, such that correct presynaptic connections are made. A further possibility is that the sympathetic system is much less, specific than, for example, the motor system and that most nerves mediate similar responses. This latter postulate is unlikely to be true for events such as the pupillary light reflex and can probably be discarded. 4.2. Mechanism ofphenotypic change The relationship between the innervating neuron and its target is not clear but the innervation of the sweat gland in the footpad may shed some light on this interaction. In vivo, the sweat gland is initially innervated by adrenergic neurons and then there is a phenotypic change, presumably under the influence of the target, when these cells become cholinergic (Landis and Keefe, 1983; Landis, 1988). Studies of postnatal rat sympathetic neurons developing in vitro have demonstrated that the environment can play a crucial role in the determination of the mature neurotransmitter phenotype. A series of detailed studies have shown that many non-neuronal cells produce a factor that results in a decrease in noradrenergic properties and the neurons acquire cholinergic properties (Patterson and Chun, 1974; 1977a,b; Kessler et al, 1984; Kessler. 1985a; Potter et al., 1986). In addition, environmental cues also influence the expression of neuropeptide levels in the sympathetic neurons, affecting substance P (Bohn et al., 1984; Kessler et al., 1984; Kessler, 1984b, 1985b; Adler and Black, 1985; Nawa and Sah, 1990), VIP, so-

155

matostatin (Nawa and Patterson, 1990) and enkephalin (Bohn et al., 1983; Henschen et al., 1986; Romagnano et al., 1989) levels. The target tissue itself has been implicated as a possible source of a putative environment factor (Schotzinger and Landis, 1988, 1990; Stevens and Landis, 1990). Thus there may be a rearrangement of the central connections to provide a new specificity related to the target tissue innervated by the neuron. A number of proteins influence neurotransmitter traits in sympathetic neurons (Fukada, 1985; Kessler et al., 1986; Saadat et al., 1989). Sympathetic cholinergic differentiating factor and leukemia inhibitory factor are the same 45 kDa glycoprotein molecule (Yamamori et al., 1989) which increases CAT activity and decreases tyrosine hydroxylase activity (Fukada, 1985). A 22 kDa protein isolated from rat sciatic nerve, ciliary neurotrophic factor (Manthorpe et al., 1986; Lin et al., 1989; Stockii et al., 1989) has a similar effect (Saadat et al., 1989). It appears that the two molecules act via the same receptor complex (Ip et al., 1992). Partially purified extracts of brain also induce cholinergic traits in sympathetic neurons (Kessler et al., 1986) and this is facilitated by plasma membranes and membrane-bound factor (Lee et al., 1990). Depolarization stimulates CAT in mouse spinal cord cultures (Ishida and Deguchi, 1983) but not in rat purified motoneurons (Martinou et al., 1989) suggesting a synergistic component from non-motoneurons. There is no obvious pattern of transmitter induction in sympathetic neurons but the timing of exposure to neurotrophic factors may alter the developmental profile of sympathetic neurons (Schwarting et al., 1990). Depolarization (Black et al., 1971; Black and Geen, 1973; Goodman et al., 1974; Walicke et al., 1977; Kessler et al., 1981, 1983; Kessler and Black, 1982; Kessler, 1986), cAMP analogues (Mackay and Iversen 1972; Goodman et al., 1974; Keen and Mclean, 1974), non-neuronal cells (Patterson and Chun, 1974; Kessler, 1984a, 1986; Lefebvre et al., 1991) all affect transmitter expression in different ways and result in diverse transmitter combinations, suggesting that co-localized transmitters are independently regulated. On the

156

Other hand, in vivo, the sympathetic ganglionic neurons do not appear to be affected by depolarization and remain adrenergic in spite of the section of preganglionic nerve trunks (Hill and Hendry, 1979). Substance P levels are increased in sympathetic ganglia by interleukin-I and stimulated splenocytes (Jonakait and Schotland, 1990). It is clear that depending on the interactions between factors, multiple agonists in different concentrations or acting sequentially in different orders can lead to the development of many different phenotypes. 5. Mechanisms to convey information from target to neuron There are a number of ways in which the target can convey information to its innervating neuron including provision of specific factors via the circulation, by retrograde axonal transport of neurotrophins (Lindsay et al., 1994), neurotrophin receptors or other messenger systems (Curtis and DiStefano, 1994), and by a variety of other modulatory effects on neuronal function. 5.1. Circulatory factors As mentioned earlier, the initial neuroembryological studies suggested that the size of a nerve centre was determined by the size of its target. When NGF was first described in tumour tissue it was clear that it acted systemically to affect not only cells innervating the tumour but also other sympathetic and sensory ganglia. Thus a target derived neurotrophic factor was acting via the circulation. Insulin-like growth factor-I (IGF-I) causes enhanced regeneration of the rat sciatic nerve after a freezing lesion (Kanje et al., 1989; Sjoberg and Kanje, 1989) There may be a general role for neurotrophic molecules to act via the circulation to increase the potential size of the neuronal pool (Fig. 2(1)) and both NGF and IGF-I have been suggested to increase mitogenic activity in neuroblasts (Rush et al., 1992). Such an action, however, does not explain the unilateral effects seen after limb removal or limb hyperplasia. IGF-I, which is important in the regulation of

Retrograde factors in peripheral

nerves

peripheral nerve regeneration (Linnemann and Bock, 1986) has been reported to be retrogradely transported in the sciatic nerve (Hansson et al., 1986, 1987) and it has been suggested that it may be responsible for the cell body reaction (Kanje et al., 1991). There has been failure, however, to detect high affinity uptake and retrograde axonal transport of IGFs by motoneuron processes in situ (Caroni and Grandes, 1990). There are likely to be many molecules that act directly on the cell bodies of developing neurons via the circulation. While many of these molecules could be produced in the target tissues, and may promote the survival of groups of neurons, it is unlikely that sufficient specificity could be achieved to enable accurate communication between the target tissue and the individual innervating neurons. 5.2. Retrograde transport of factor It is difficult to imagine an alternative mechanism for a specific target cell to convey information to a specific neuron cell body, other than by transport of a specific signal confined to the axon. In order for there to be a direct influence of the target only on the neuron innervating it, there must be a specific retrograde message carried via the axon from the target to the nucleus. The simplest mechanism for the target tissue to control the innervating neurons is for the target cell to make a neurotrophic molecule which is available only in limiting amounts. This retrogradely transported neurotrophin (retrophin) (Hendry and Hill, 1980) is then taken up by the nerve terminals or growth cones and retrogradely transported back to the neuronal perikarya, where it acts to promote the survival of the neuron that has transported it (Fig. 2(2)). The first neurotrophic molecule to be described to undergo retrograde axonal transport was NGF (Hendry et al., 1974). While the retrograde intra-axonal transport of NGF is now well established, the physiological significance of the specific retrograde axonal transport of NGF or, indeed, any other neurotrophic molecule is not resolved. The possible reasons for the transport of NGF include many of the mechanisms to convey

157

LA. Hendry

^-

Labile second messenger

IBR Receptor forB ©

Stable second messenger

|cR Receptor for C Extracellular IP

Binding to extracellular matrix

Matrix

Fig. 2. Diagrammatic representation of the ways in which the target can convey information to its innervating neuron. (1) Specific factor via circulation. (2) Transport of active neurotrophic factor. (3) Transport of active receptor. (4) Transport of second messenger. (5)Transport of factor-receptor complex. (6) Transport of permanently modified receptor. (7) Transport of second messenger. (8) Cessation of transport of death molecule. (9) Retrograde electrical activity. (10) Cessation of neurite outgrowth with build-up of molecules.

information from the terminal to the cell body that have already been outlined. 5.2.7. Transport of active neurotrophic factor Transport may be the means of getting NGF to the cell body where it or a breakdown product could exert its trophic effect. The site of action of NGF remains controversial with proponents for direct action on cytoskeletal elements and nuclear chromatin (Calissano and Cozzari, 1974; Andres et al., 1977; Bradshaw, 1983) or via second messenger mechanisms (Thoenen and Barde, 1980; Heumann et al., 1981). Some studies have suggested that free NGF reaches the cell cytoplasm and the nucleus where it interacts directly with the cytoskeletal elements leading to fibre outgrowth (Calissano and Cozzari, 1974; Nasi et al., 1982) or with nuclear binding sites (Andres et al., 1977; Bradshaw, 1983) leading to-regulation of expres-

sion of specific genetic information, such as neurotransmitter biosynthetic enzymes. Studies of the localization of retrogradely transported NGF have not resolved this conflict, as some studies with iodinated NGF have suggested that NGF may reach the nucleus, (Hendry et al., 1974; Iversen et al., 1975) but other studies, using either coupling products of NGF and horse radish peroxidase or autoradiographic localization of [P^^JNGF, did not provide any evidence that NGF reaches the cytoplasm or, subsequently, the nucleus (Schwab and Thoenen, 1977; Schwab, 1977). The problem is that no detection system can rule out the presence of a small proportion of NGF in the nucleus. Thus this type of localization experiment cannot demonstrate that there is not a small, undetectable amount of NGF that leaves the membrane-confined compartments into which it goes after its receptormediated uptake and that may be responsible for

158

the biological effects. In addition, NGF or its receptor could be degraded in the lysozomal compartment and a peptide could be released into the cytoplasm to act directly on the target elements. 5.2.2. Transport of active receptor Transport of NGF may be a means of getting the receptor to the perikaryon to act as the trophic message (Fig. 2(3)). The receptor for NGF is retrogradely transported by neurons both in the periphery and in the central nervous system, as shown by the intracellular localization of the antibody to the receptor (Johnson et al., 1987). NGF receptors are transiently expressed on motoneurons during development in newborn rats and NGF is retrogradely transported by these neurons (Yan et al., 1988). As these receptors are not capable of mediating any traditional neurotrophic effects on survival, this suggests that the retrograde axonal transport of NGF in neurons is not necessarily synonymous with the ability of NGF to exert trophic activity. 5.2.3. Transport of second messenger An appropriate second messenger, which is subsequently transported to the cell body, may be generated at the nerve terminal and the retrograde transport of NGF may merely reflect the presence of receptors capable of internalising NGF. It is unlikely that the neurotrophic factor itself is the message. It is more likely that the neurotrophic factor generates a second messenger, or a cascade of messengers, that is able to interact with the nucleus to result in neuronal survival. Any such second message that is to reach the cell body by its own retrograde transport would need to be very stable (Fig. 2(4)). 5.2.4. Transport of factor-receptor complex NGF binds to specific receptors, is internalized at nerve terminals and transported, together with its receptor, back to the cell body in coated vesicles which eventually fuse with lysosomes where the NGF is degraded (Schwab, 1977). It is unlikely that the internalized NGF itself is the active intracellular message, as intracellular NGF, where the cell surface receptors have been by-passed by di-

Retrograde factors in peripheral nerves

rect injection of NGF into the cytoplasm and thus able to directly reach the nuclear chromatin cannot mimic the NGF receptor-mediated response (Heumann et al., 1981, 1984; Rohrer et al., 1982). In addition, affinity purified antibodies to NGF injected directly into the cytoplasm of PCI2 cells do not affect the cellular response to NGF (Heumann et al., 1984). Fibre outgrowth also is not influenced by intracellular NGF (Heumann et al., 1981; Seeley et al., 1983). The blockade of the lysozomal degeneration of internalized NGF by the proteolytic enzyme inhibitor, leupeptin, influenced neither CAT induction nor fibre outgrowth, showing that it is also unlikely for degeneration products of NGF to be responsible for its biological activity. Thus it is clear that the mechanism of action of NGF is via a second messenger system, but it is not clear whether the various effects are via the same second messenger or a variety. The formation of a second messenger could be initiated by the interaction of NGF with its receptors either immediately after NGF binding at the cell membrane or following internalization and retrograde axonal transport of the receptor-NGF complex in vesicles. As the receptor-NGF complex in the transported vesicle is presumably continuously capable of generating a second messenger, it may be that the transport enables a short-lived or unstable second message to reach the cell body. It seems likely the rapid effects of NGF, such as chemotactic effects on growth cones (Gundersen and Barrett, 1980; Seeley and Greene, 1983) and initiation of neurite sprouting (Calissano and Cozzari, 1974; Nasi et al., 1982), could be mediated by locally acting second messengers activated by the interaction of NGF with its receptor at the surface and involving calcium ions and phosphorylation by receptor-associated tyrosine kinase (Kaplan et al., 1991). The longer term effects, involving the regulation of the synthesis of specific proteins in the cell body at the transcriptional or posttranscriptional level (Rohrer et al., 1978; Hefti et al., 1982), require the transport of the second messenger-generating system to the cell body or else the transport of a much more stable molecule. Peripheral administration of NGF to sympathetic nerve terminals leads to a specific increase in tyro-

LA. Hendry

sine hydroxylase in the cell bodies (Paravicini et al., 1975) and, subsequently, an increase in the size of the neurons that retrogradely transported the NGF (Hendry, 1977a). Therefore, the messenger responsible for the long-term effects must either be produced at the terminal and be stable enough to survive its own retrograde transport to the nucleus or be generated by the NGF-receptor complex as it is transported along the axon and into the cell body. Thus a much more labile second messenger could be responsible for the longer term actions of NGF (Crouch and Hendry, 1991). 5.2.5, Control of tissue retrophin levels If a neurotrophic factor is to have in vivo physiological relevance, there must be a mechanism for controlling its availability to neurons. This may be via competition between neurons for a limited amount of trophic factor released from the target which will be exacerbated if the neurons remove the factor. Retrograde transport may serve as a mechanism to remove NGF from the terminal region and therefore be a means to regulate the levels of NGF in the target. Neurotrophic factors for sympathetic and parasympathetic neurons are elevated in the rat ventricle after chemical denervation (Kanakis et al., 1985) and growth factors for ciliary neurons are elevated in skeletal muscle after denervation (Hill and Bennett, 1983). The effects of 6-hydroxydopamine and colchicine on levels of NGF in sympathetic ganglia and sympathetic target tissue strongly suggest that NGF levels in the target is controlled by retrograde axonal transport by the innervating neurons (Korsching andThoenen, 1985). The availability of trophic factors may be regulated by other mechanisms, such as binding to the substrate, as has been proposed for fibroblast growth factor (FGF) (Eccleston et al., 1985; Rogelj et al., 1989). If this is the case, then availability is regulated by the composition of the target tissue itself and not by the innervating neurons. The most plausible mechanism of action of NGF is that it is internalized together with its receptor into coated vesicles which are retrogradely transported to the cell body, where the NGFreceptor complex continues to generate the labile

159

second messenger required for its action (Fig. 2(5)). 5.3, No retrograde transport of factor In spite of the logic of the retrophin model, retrograde axonal transport has been described for only a very few putative neurotrophic molecules. While this mechanism is the simplest, it is not the only way to get a message from the periphery, and an in-depth analysis of some other mechanisms that may lead to the chromatolytic signal has been made previously (Cragg, 1970). If the neurotrophic molecule itself is not the second messenger, then there has to be translocation of some other message from the terminals in the target tissue to the nucleus. Before this can be considered some consideration must be given as to what is a neurotrophic factor. 5.3.1. Does a neurotrophic factor require retrograde transportation ? The NGF model has led to the suggestion that all neurotrophic factors must be retrogradely transported (Thoenen et al., 1985); a point of view that is much more rigid than the evidence warrants. If there are ways for a factor to signal from the target in the absence of the transport of the factor itself, then it should be still considered a neurotrophic factor. In order for a potential survival factor identified in culture to be seen as a viable neurotrophic factor it must act in vivo to promote neuronal survival. There are many neurotrophic factors that have been shown to have effects in vivo, but there is always the possibility that the effects of an agent are pharmacological in that it mimics the action of the endogenous factor, often at very high concentrations, rather than being the physiological factor itself. The physiological factor will need to be present in limiting amounts in order to achieve the desired control of survival. The classical studies on NGF have shown that this molecule can cause the survival of sensory neurons (Levi-Montalcini and Angeletti, 1968) and has provided a paradigm for the in vivo investigation of the other putative survival factors. NGF

160

also causes the survival of sympathetic neurons (Levi-Montalcini, 1965; Hendry and Campbell, 1976) and can rescue them after axotomy (Hendry, 1975a) and 6-hydroxydopamine treatment (LeviMontalcini et al, 1975). There are several members of the FGF family that may be of neurotrophic relevance (Delli-Bovi et al., 1988; Basilico et al., 1989). Polyclonal antibodies to basic fibroblast growth factor (bFGF) have been shown to react to a higher molecular weight molecule which appears to cause survival of ciliary neurons in culture (Grothe et al., 1990). bFGF is found in adrenal medullary chromaffin granules (Westermann et al., 1990) and gel foams soaked in bFGF can rescue adrenal preganglionic neurons after electrolytic destruction of the adrenal medulla (Blottner et al, 1989b). Acidic fibroblast growth factor (aFGF) enhances nerve regeneration in sciatic nerve (Cordeiro et al, 1989) and retinal ganglion cell processes (Lipton et al, 1988) and stimulates adrenal chromaffin cells to proliferate and extend neurites, but fails to cause long-term survival (Claude et al., 1988). In addition, aFGF causes the differentiation of ciliary neurons in chick embryos but fails to reverse naturally occurring cell death (Hill et al., 1991). CiUary neurotrophic factor (CNTF) can rescue adrenal preganglionic neurons after electrolytic destruction of the adrenal medulla (Blottner et al., 1989a) and facial neurons after axotomy (Sendtner et al., 1990). While the FGFs and CNTF are good candidates as ciliary neurotrophic factors (Watters et al., 1989) specific retrograde axonal transport of FGF (Hendry and Belford, 1991) or CNTF (Smet et al., 1991) by ciliary neurons have not been demonstrated. bFGF was also not retrogradely transported in the adult rat sciatic nerve, or from iris to trigeminal or SCG (Ferguson et al., 1990). Most molecules can be shown to have a low level of non-specific retrograde transport; for instance horseradish peroxidase, FGF, bovine serum albumin, cytochrome C are all only transported to a minor degree. This is presumably as a result of pinocytosis of extracellular fluid in the region of the nerve terminal (Heuser and Reese, 1973) into coated vesicles (Sacks and Saito, 1969). This uptake is dependent upon activity of the nerve

Retrograde factors in peripheral nerves

(Holtzman et al, 1971; Teichberg et al, 1975), unlike that of the receptor-mediated uptake of NGF, where alterations in the transport of NGF were not seen after alterations in the firing pattern of sympathetic neurons (Stockel et al., 1978; Lees etal., 1981). From the preceding discussion it can be concluded that there are many putative neurotrophic molecules that do not themselves have a high capacity retrograde transport and there may be two potential types of neurotrophic molecule with either a short-lived second message (Fig. 3A) or a long Hved-message (Fig. 3B). One class, like NGF, has a labile second messenger and thus requires the transport of the neurotrophin-receptor complex in order to be able to generate the second messenger near enough to the neuronal nucleus to be effective. The other class must generate a stable second messenger at the nerve terminal and it is the transport of this that results in the specificity of the target/nerve communication. There are three ways in which a long-lived message could be generated. Firstly, activation of the receptor might render it permanently active and this activated receptor could reach the cell body. Secondly, there could be the transport of a stable, activated second messenger complex. Thirdly, the normal transport of some inhibitory or cell death factor might be prevented. 5.3.2. Transport of permanently modified receptor One mechanism by which the signal might be transduced from the nerve terminal is by the permanent activation of a receptor. In platelets, the thrombin receptor is cleaved and this results in its permanent activation. It is interesting that thrombin itself has been shown to have direct effects on central nervous system neurons (Mazzoni and Kenigsberg, 1991). It is possible that proteolytically active molecules secreted by the target tissue could act to modify receptors on the nerve terminal and these may be internalized either alone, or together with other internalized receptor-factor complexes and be transported back to the cell body. Thus there is a mechanism for several signals to act in concert and provide a complex integration of target derived signals to be received by the nucleus.

161

LA. Hendry

Labile Second Messenger

B

Stable Second Messenger

Nerve Terminal

Nerve Terminal

»NGF Receptors

o o

Growth Factor Receptors

NGF Growth Factor NGF bound to Receptors leads to clustering

Retrograde Axonal Transport of NGF bound to Receptors in Coated Vesicles

_ Internalization of NGF into vesicles

Long lived second messenger undergoing Retrograde Axonal Transport

Growth Factor - Receptor Complex may generate stable or labile second messengers

Labile second messenger

Labile second messenger continues to be generated in cell body

Retrograde Axonal Transport of Stable Second Messenger

Vesicle containing NGF-Receptor complex

Interaction with Nucleus

Fig. 3. Two possibilities for the types of neurotrophic factors. (A) Neurotrophic factor with labile second messenger requires the retrograde axonal transport of a receptor/factor complex to enable the generation of active second messenger at the cell body. (B) Neurotrophic factor which generates a stable second messenger which is active after the long time required for its transport to the cell body does not require the transport of the factor itself.

5.3.3. Transport of second messenger One of the major arguments against aFGF or bFGF being neurotrophic factors is the lack of an appropriate secretory sequence (Esch et al., 1985a,b), making regulated release by the target tissue unlikely. If FGF is not a target-derived neurotrophic factor then one of the other members of the FGF gene-related family may be (Huebner et al., 1988; Zhan et al., 1988; Brookes et al., 1989). In this case, the protein would have to share an antigenic epitope with aFGF and may itself be retrogradely transported. If in fact aFGF is an essential neurotrophic factor for ciliary neurons, then the retrograde message must be conveyed by a

mechanism other than the retrograde transport of FGF itself. It is not essential, however, for the neurotrophic factor itself to be retrogradely transported but, clearly, some signal must reach the neuronal cell body. A likely alternative is the retrograde axonal transport of one of the second messengers. The most likely candidates for a stable second message are proteins that have undergone some covalent change. G-protein translocation. It has been shown that in platelets there is a translocation of G-protein from the cell membrane to the cytoskeleton (Crouch et al., 1989). It has also been shown that activation of

162

Balb/c 3T3 fibroblasts results in translocation of the a subunit of Gj from the membrane to the nuclear chromatin (Crouch, 1991). A similar phenomenon could result in the retrograde axonal transport of the neurotrophic message for FGF. There is an accumulation of G^a (Hendry and Crouch, 1991) and G^ (Crouch et al., 1991; Hendry and Crouch, 1992) on the proximal and distal sides of a ligature placed on the mouse sciatic nerve, demonstrating the anterograde transport of these second messenger molecules down to the nerve terminal as well as transport back to the cell body from the terminals. It has been shown, using antibodies for the immunohistochemical localization for Gz^, that it is not only retrogradely transported in axons (Crouch et al., 1994) but also translocates to the neuronal nucleus (Hendry et al., 1995a). Furthermore, the levels seen in the nuclear compartment decline alter axotomy or ligation of the axon in mice, suggesting it is the retrogradely transported G^ that is accumulating in the nucleus after activation at the nerve terminal (Hendry e al., 1995b) on the proximal and distal sides of a ligature placed on the mouse sciatic nerve, demonstrating that there is not only transport of these second messenger molecules to the nerve terminal but also transport back to the cell body from the terminals. In this case it has to be assumed that the nerve terminals are normally being activated by endogenous neurotrophic factors and that retrograde axonal transport of the GTP-binding proteins may be a representation of the stable message generated by this activation. Protein phosphorylation. Many neurotrophic factors act through tyrosine kinase receptors which seem to generate cascades of phosphorylation reactions. Any of these may, generate an appropriate stable second message to transport back to the cell body. Furthermore, phosphorylation reactions can provide a mechanism for the interaction of growth factors at the terminal to generate specific alterations in potential retrograde messengers (Greengard, 1987). These phosphorylation reactions can alter proteins such that they can generate new complexes via binding mechanisms involving phosphorylated tyrosines and the src homology

Retrograde factors in peripheral

nerves

domain 2 regions SH2 and SH3 of intracellular molecules that are thought to be important in signal transduction (Coughhn et al., 1989; Cantley et al., 1991; Klippel et al, 1992). For example, the platelet-derived growth factor receptor is a tyrosine kinase that, on activation, can form complexes with phospholipase-C y-I, GTPase activating protein and phosphatidylinositol 3-kinase via distinct phosphotyrosine-containing sequences (Fantl et al., 1992). Thus it is possible that at the nerve terminal, there can be a complex interaction between receptors and the intracellular transduction molecules, such that very specific complexes generating a unique message may be generated and this second messenger unit can be conveyed to the cell body via retrograde axonal transport. There is good evidence that there is accumulation of tyrosine phosphorylated proteins on the distal side of a ligated sciatic nerve demonstrating the retrograde transport of phosphorylated proteins (Johanson et al., 1995a,b). Acylation or deacylation of proteins. Alterations in the hydrophobicity of molecules may come about by addition or cleavage of a lipid component, resulting in a molecule suitable for retrograde transport. No matter what the signal is, it must be transported either in vesicles or as a soluble molecule, either alone or in a complex. The evidence for the retrograde transport of coated vesicles is good but there does not seem to be any evidence for the retrograde transport of isolated proteins. Thus there may be two types of protein alteration, either to add a lipid component and enhance the chances of becoming associated with coated vesicles or to remove an acyl group to promote the protein leaving the plasma membrane. The 87 kDa protein phosphorylated by protein kinase C seems to be the same as a 68 kDa protein myristoylated in response to lipopolysaccaride in macrophages (Aderem et al., 1988). This myristoylation could allow targeting of the protein to the membrane, where it is more readily phosphorylated by protein kinase C and thus would prime the cells for a subsequent activation. The 87 kDa protein has been described in brain, where it is phosphorylated in response to phorbol esters and depolarization

LA. Hendry

(Wang et al., 1988, 1989). This protein appears to be myristoylated in murine frontal cortex cells (Aderem et al., 1988). GAP-13 appears in dorsal root ganglion cells and in the dorsal horn in the spinal cord of neonatal rats following peripheral nerve injury (Woolf et al., 1990). Myristoylated alanine-rich C kinase substrate (MARCKS) (Stumpo et al., 1989) is a 87 kDa protein which is a major specific substrate for protein kinase C in rat brain (Ouimet et al., 1990). MARCKS is translocated after phosphorylation in isolated nerve terminals (Wang et al., 1989), an event associated with its release from membranes into the cytosol (Narayanan and Narayanan, 1981). Thus there are many acylation/deacylation reactions occurring in the nerve terminal which may generate a stable, modified protein and result in its becoming available for retrograde transport. 5.3.4. Cessation of retrograde transport of death molecule It has been shown that the low affinity p75 NGF receptor can mediate neuronal cell death (Barrett and Bartlett, 1994) and NGF can prevent this. There may be molecules taken up at the nerve terminals that could promote the death of the neuron. Interaction with the appropriate target could prevent the uptake of this substance and thus prevent neuronal death. 5.4. Physical mechanisms 5.4.1. Electrical Although this type of mechanism has not yet been explored in the literature, it should still be considered when analysing the many different strategies that could be used to signal to the cell from its target. At the time of synaptogenesis there is an intimate communication between neuron and target, often with the formation of gap junctions. The possibility for electrical coupling exists and the generation of retrograde electrical activity could effect neuronal survival or even neuronal phenotype.

163

the effect of good matching of cell-cell recognition molecules or the result of high concentrations of growth factors. The earliest experiments on NGF demonstrated that sensory axons, growing into a tumour synthesising large amounts of NGF, ended blindly, apparently ceasing elongation in the absence of any target contact (Levi-Montalcini and Hamburger, 1951). When a neuron reaches its appropriate target there is also a cessation of axon elongation and growth. Whatever the mechanism of this cessation of growth, there could be a buildup of proteins normally transported away from the cell body. In this way, factors necessary for neuronal survival may accumulate. 6. Conclusions It is likely that more than one mechanism is involved in the formation of correct neuronal connections. There must, however, be some feedback from the target to the innervating neuron to uphold and strengthen correct connections. There is little evidence to support the direct effect of a neurotrophic factor on the neuronal nucleus and there must be the mediation of a second messenger. It is proposed here that there are two classes of neurotrophic factor. The first, represented by NGF, has a labile second messenger which requires the transport of a factor-receptor complex capable of continuous generation of the second messenger to the cell body. The second class does not require retrograde transport of the factor, as it can generate a stable second messenger that survives the longtime for transport up the axon. It is probable that the message that is transported via the axon will be a complex resulting from the activation of multiple receptors at the nerve terminal, rather than a single molecule. The result of the transport of the complex will be to give the cell detailed information as to the state of its terminal region and the nature of the target tissue and target cell that it has contacted. References

5.4.2. Cessation ofneurite outgrowth It is not clear what the signal is that results in the cessation of neurite outgrowth, but it could be

Aderem, A.A., Albert, K.A., Keum, M.M., Wang, J.K., Greengard, P. and Cohn, Z.A. (1988) Stimulus-dependent

164 myristoylation of a major substrate for protein kinase C. Nature 332: 362-364. Adler, J.E. and Black, LB. (1985) Sympathetic neuron density differentially regulates transmitter phenotypic expression in culture. Proc.Natl. Acad. Sci. USA 82: 4296-4300. Andres, R.Y., Jeng, I. and Bradshaw, R.A. (1977) Nerve growth factor receptors: identification of distinct classes in plasma membranes and nuclei of embryonic dorsal root neurons. Proc. Natl. Acad. Sci. USA 74: 2785-2789. Baker, R.F., Corner, M.A. and Veltman, W.A.M. (1978) Topography of cuteaneous mechanoceptive neurons in dorsal root ganglia of skin-grafted frogs. /. Physiol. (London) 284: 181-192. Banner, L.R. and Patterson, P.H. (1994) Major changes in the expression of the mRNAs for cholinergic differentiation factor/leukemia inhibitory factor and its receptor after injury to adult peripheral nerves and ganglia. Proc. Natl. Acad. Sci. USA 91: 7109-7113. Barrett, G.L. and Bartlett, P.F. (1994) The p75 nerve growth factor receptor mediates survival or death depending on the stage of sensory neuron development. Proc. Natl. Acad. Sci. USA 91: 6501-6505. Basilico, C, Newman, K.M., Curatola, A.M., Talarico, D., Mansukhani, A., Velcich, A. and Delli-Bovi, P. (1989) Expression and activation of the K-fgf oncogene. Ann. N. Y. Acad. Sci. 567: 95-103. Black, LB. and Geen, S.C. (1973) Trans-synaptic regulation of adrenergic neuron development: inhibition by ganglionic blockade. Brain Res. 63: 291-302. Black, LB., Hendry, LA. and Iversen, L.L. (1971) Transsynaptic regulation of growth and development of adrenergic neurons in a mouse sympathetic ganglion. Brain Res. 34: 229-240. Black, LB., Hendry, LA. and Iversen, L.L. (1972) Effects of surgical decentralization and nerve growth factor on the maturation of adrenergic neurons in a mouse sympathetic ganglion. J. Neurochem. 19: 1367-1377. Blottner, D., Bruggemann, W. and Unsicker, K. (1989a) Ciliary neurotrophic factor supports target-derived preganglionic sympathetic spinal cord neurons. Neuro sci. Lett. 105:316-320. Blottner, D., Westermann, R., Grothe, C, Bohlen, P. and Unsicker, K. (1989b) Basic fibroblast growth factor in the adrenal gland. Eur. J. Neurosci. 1: 471^78. Bohn, M.C., Kessler, J.A., Golightly, L. and Black, LB. (1983) Appearance of enkephalin-immunoreactivity in rat adrenal medulla following treatment with nicotinic antagonists or reserpine. Cell Tissue Res. 231: 469-479. Bohn, M.C., Kessler, J.A., Adler, J.E., Markey, K., Goldstein, M. and Black, LB. (1984) Simultaneous expression of the SP-peptidergic and noradrenergic phenotypes in rat sympathetic neurons. Brain Res. 298: 378-381. Bradshaw, R.A. (1983) Nerve growth factor and related hormones. In: G. Litwack (Ed.), Biochemical Actions of Hormones, Academic Press, New York, pp. 91-114. Brookes, S., Smith, R., Thurlow, J., Dickson, C. and Peters, G.

Retrograde factors in peripheral nerves (1989) The mouse homologue of hst/k-FGF: sequence, genome organization and location relative to int-2. Nucleic Acids Res. 17:4037^045. Calissano, P. and Cozzari, C. (1974) Interaction of nerve growth factor with the mouse-brain neurotubule protein(s). Proc. Natl. Acad Sci. USA 71: 2131-2135. Cantley, L.C., Auger, K.R., Carpenter, C, Duckworth, B., Graziani, A., Kapeller, R. and Soltoff, S. (1991) Oncogenes and signal transduction. Cell 64: 281-302. Caroni, P. and Grandes, P. (1990) Nerve sprouting in innervated adult skeletal muscle induced by exposure to elevated levels of insulin-like growth factors. J. Cell Biol. 110: 1307-1317. Chang, J.Y., Martin, D.P. and Johnson. E.M. (1990) Interferon suppresses sympathetic neuronal cell death caused by nerve growth factor deprivation. J. Neurochem. 55: 436445. Claude, P., Parada, I.M., Gordon, K.A., D'Amore, P.A. and Wagner, J.A. (1988) Acidic fibroblast growth factor stimulates adrenal chromaffin cells to proliferate and to extend neurites, but is not a long-term survival factor. Neuron 1: 783-790. Cordeiro, P.G., Seckel, B.R., Lipton, S.A., D'Amore, P.A., Wagner, J. and Madison, R. (1989) Acidic fibroblast growth factor enhances peripheral nerve regeneration in vivo. Plast. Reconstr. Surg. 83: 1013-1019. Coughlin, S.R., Escobedo, J.A. and Williams, L.T. (1989) Role of phosphatidylinositol kinase in PDGF receptor signal transduction. Science 243: 1191-1194. Cragg, B.G. (1970) What is the signal for chromatolysis? Brain Res. 23: 1-21. Crouch, M.F. (1991) Growth factor-induced cell division is paralleled by translocation of Gj^ to the nucleus. FASEB J. 5: 200-206. Crouch, M.F. and Hendry, LA. (1991) Co-activation of insulin-like growth factor-I receptors and protein kinase C results in parasympathetic neuronal survival. J. Neurosci. Res. 2S: 115-120. Crouch, M.F., Heydon, K., Garnaut, S., Milburn, P.J. and Hendry, LA. (1994) Retrograde axonal transport of the asubunits of the GTP-binding protein G^ in mouse sciatic nerve: a novel pathway for signal transduction in neurones. Eur. J. Neurosci. 6: 626-631. Crouch, M.F., Winegar, D.A. and Lapetina, E.G. (1989) Epinephrine induces changes in the subcellular distribution of the inhibitory GTP-binding protein Gj alpha-2 and a 38-kDa phosphorylated protein in the human platelet. Proc. Natl. Acad Sci. USA 86: 1776-1780. Curtis, R. and DiStefano, P.S. (1994) Neurotrophic factors, retrograde axonal transport and cell signalling. Trends Cell Biol. 4: 383-386. Delli-Bovi, P., Curatola, A.M., Newman, K.M., Sato, Y. and Moscatelli, D. (1988) Processing secretion and biological properties of a novel growth factor of the fibroblast growth factor family with oncogenic potential. Mol. Cell. Biol. 8: 2933-2941.

LA. Hendry Dohrmann, U., Edgar, D. and Thoenen, H. (1987) Distinct neurotrophic factors from skeletal muscle and the central nervous system interact synergistically to support the survival of cultured embryonic spinal motor neurons. Dev. Biol. 124: 145-152. Eccleston, P.A., Gunton, D.J. and Silberberg, D.H. (1985) Requirements for brain cell attachment, survival and growth in serum-free medium: effects of extracellular matrix, epidermal growth factor and fibroblast growth factor. Dev. Neurosci. 7: 308-322. Esch, F., Baird, A., Ling, N., Ueno, N., Hill, F., Denoroy, L., Klepper, R., Gospodarowicz, D., Bohlen, P. and Guillemin, R. (1985a) Primary structure of bovine pituitary basic fibroblast growth factor (FGF) and comparison with the amino-terminal sequence of bovine brain acidic FGF. Proc. Natl. Acad. Set. USA 82: 6507-6511. Esch, F., Ueno, N., Baird, A., Hill, F., Denoroy, L., Ling, N., Gospodarowicz, D. and Guillemin, R. (1985b) Primary structure of bovine brain acidic fibroblast growth factor (FGF). Biochem. Biophys. Res. Commun. 133: 554-562. Fantl, W.J., Escobedo, J.A., Martin, G.A., Turck, C.W., Del Rosario, M., McCormick, F. and Williams, L.T. (1992) Distinct phosphotyrosines on a growth factor receptor bind to specific molecules that mediate different signaling pathways. Ce//69: 413-423. Ferguson, I.A., Schweitzer, J.B. and Johnson, E.M. (1990) Basic fibroblast growth factor: receptor-mediated internalization, metabolism, and anterograde axonal transport in retinal ganglion cells. J. Neurosci. 10: 2176-2189. Fukada, K. (1985) Purification and partial characterization of a cholinergic neuronal differentiation factor. Proc. Natl. Acad. Sci. USA 82: 8795-8799. Goodman, R., Oesch, F. and Thoenen, H. (1974) Changes in enzyme patterns produced by high potassium concentration and dibutyryl cyclic AMP in organ cultures of sympathetic ganglia. J. Neurochem. 23: 369-378. Greengard, P. (1987) Neuronal phosphoproteins. Mediators of signal transduction. Mol. Neurobiol. 1: 81-119. Griersmith, B.T. and Mark, R.F. (1982) Behavioral and electrophysiological study of cutaneous trigeminal nerves in axolotls. II. The effects of cross-anastomosis of nerves. Brain Res. ISA'. 2^1-^0\. Grothe, C., Zachmann, K., Unsicker, K. and Westermann, R. (1990) High molecular weight forms of basic fibroblast growth factor recognized by a new anti-bFGF antibody. F£55Lerr. 260:35-38. Gundersen, R.W. and Barrett, J.N. ( 1980) Characterization of the turning response of dorsal root neurites toward nerve growth factor. J. Cell Biol. 87: 546-554. Hamburger, V. (1934) The effects of wing bud extirpation on the development of the central nervous system in chick embryos. J. Exp. Zool. 68: 4 4 9 ^ 9 4 . Hamburger, V. (1939) Motor and sensory hyperplasia following limb bud transplantation in chick embryos. Physiol. Zool. 13: 268-284. Hamburger, V. (1975) Cell death in the development of the

165 lateral motor column of the chick embryo. J. Comp. Neurol. 160: 535-546. Hamburger, V. and Levi-Montalcini, R. (1949) Proliferation, differentiation and degeneration in the spinal ganglia of the chick embryo under normal and experimental conditions. J. Exp. Zool. 111:457-502. Hansson, H.A., Dahlin, L.B., Danielsen, N., Fryklund, L., Nachemson, A.K., Polleryd, P., Rozell, B., Skottner, A., Stemme, S. and Lundborg, G. (1986) Evidence indicating trophic importance of IGF-I in regenerating peripheral nerves. Acm Physiol. Scand. 126: 609-614. Hansson, H.A., Rozell, B. and Skottner, A. (1987) Rapid axoplasmic transport of insulin-like growth factor I in the sciatic nerve of adult rats. Cell Tissue Res. 247: 241-247. Hefti, F., Gnahn, H., Schwab, M.E. and Thoenen, H. (1982) Induction of tyrosine hydroxylase by nerve growth factor and by elevated K"*" concentrations in cultures of dissociated sympathetic neurons. J. Neurosci. 2: 1554-1566. Hendry, LA. (1973) Trans-synaptic regulation of tyrosine hydroxylase activity in a developing mouse sympathetic ganglion: effect of nerve growth factor (NGF), NGFantiserum and pempidine. Brain Res. 56: 313-320. Hendry, LA. (1975a) The response of adrenergic neurons to axotomy and nerve growth factor. Brain Res. 94: 87-97. Hendry, LA. (1975b) The effects of axotomy on the development of the rat superior cervical ganglion. Brain Res. 90: 235-244. Hendry, LA. (1976) Control in the development of the vertebrate sympathetic nervous system. In: S. Ehrenpris and I.J. Kopin (Eds.), Reviews of Neuroscience, Raven Press, New York, pp. 149-194. Hendry, LA. (1977a) The effect of the retrograde axonal transport of nerve growth factor on the morphology of adrenergic neurons. Brain Res. 134: 213-223. Hendry, LA. (1977b) Cell division in the developing sympathetic nervous system. J. Neurocytol. 6: 299-309. Hendry, LA. and Belford, D.A. (1991) Lack of retrograde axonal transport of the heparin-binding growth factors by chick ciliary neurons. Int. J. Dev. Neurosci. 9: 243-250. Hendry, LA. and Campbell, J. (1976) Morphometric analysis of rat superior cervical ganglion after axotomy and nerve growth factor treatment. J. Neurocytol. 5: 351-360. Hendry, LA. and Crouch, M.F. (1991) Retrograde axonal transport of the GTP-binding protein Gj^: a potential neurotrophic intra-axonal messenger. Neurosci. Lett. 133: 2 9 32. Hendry, LA. and Crouch, M.F. (1992) Synergy, retrograde transport and cell death. In: S.E. Loughlin and J.H. Fallon (Eds.), Neurotrophic Factors, Academic Press, New York, pp. 51-88. Hendry, LA. and Hill, C.E. (1980) Retrograde axonal transport of target tissue-derived macromolecules. Nature 287: 647-649. Hendry, LA., Stockel, K., Thoenen, H. and Iversen, L.L. (1974) The retrograde axonal transport of nerve growth factor. Brain Res. 68: 103-121.

166 Hendry, I.A., Hill, C.E. and Walters, DJ. (1986) Long-term retention of fast blue in sympathetic neurons after axotomy and regeneration-demonstration of incorrect reconnections. Brain Res. 376: 292-298. Hendry, LA., Johanson, S.O. and Heydon, K., (1995a) Nuclear translocation of the alpha subunit of the GTP-binding protein G^ in dorsal root ganglia after retrograde axonal transport. Proc. Aust. Neurosci. Soc. 6: 103. Hendry, LA., Johanson, S.O. and Heydon, K. (1995b) Retrograde axonal transport of the alpha subunit of the GTPbinding protein G^ to the nucleus of sensory neurons. Brain Res. 700: 157-163. Hendry, LA., Messina, A. and Bell, C. (1992) Neonatal nerve growth factor treatment alters the preganglionic innervation pattern of rat superior cervical ganglion. Neurosci. Lett. 148: 117-120. Henschen, A., Kessler, J.A., Seiger, A. and Olson, L. (1986) Enkephalin and capsaicin-resistant substance P-like immunoreactivities in intra-ocular grafts of different fetal spinal cord areas. Acta Physiol. Scand. 128: 175-185. Heumann, R., Schwab, M.E. and Thoenen, H. (1981) A second messenger required for nerve growth factor biological activity? Nature 292: 838-840. Heumann, R., Schwab, M.E., Merkl, R. and Thoenen, H. (1984) Nerve growth factor-mediated induction of choline acetyltransferase in PC 12 cells: evaluation of the site of action of nerve growth factor and the involvement of lysosomal degradation products of nerve growth factor. /. Neurosci. 4: 3039-3050. Heuser, J.E. and Reese, T.S. (1973) Evidence for recycling of synaptic vesicle membrane during transmitter release at the frog neuromuscular junctions. J. Cell Biol. 57: 315344. Hill, C.E. and Hendry, LA. (1979) The influence of preganglionic nerves on the superior cervical ganglion of the rat. Neurosci. Lett. 13: 133-139. Hill, C.E. and Vidovic, M. (1992) Projections of sympathetic neurons to targets during development in vivo. In: LA. Hendry and C.E. Hill (Eds.), Development, Regeneration and Plasticity of the Autonomic Nervous System, Ch. 4, Harwood Academic Publishers, London, pp. 179-230. Hill, C.E., Hirst, G.D. and Van Helden, D.F. (1983) Development of sympathetic innervation to proximal and distal arteries of the rat mesentery. J. Physiol. (London) 338: 129147. Hill, C.E., Belford, D.A., Godovac-Zimmermann, J. and Hendry, LA. (1991) Class 1 heparin binding growth factor promotes the differentiation but not the survival of ciliary neurons in vivo. Dev. Brain Res. 63: 13-19. Hill, M.A. and Bennett, M.R. (1983) Cholinergic growth factor from skeletal muscle elevated following denervation. Neurosci. Lett. 35: 31-35. Holtzman, E., Freeman, A.R. and Kashner, L.A. (1971) Stimulation-dependent alterations in peroxidase uptake at lobster neuromuscular junctions. Science 173: 733-736. Huebner, K., Ferrari, A.C., Delli Bovi, P., Croce, CM. and

Retrograde factors in peripheral

nerves

Basilico, C. (1988) The FGF-related oncogene, K-FGF, maps to human chromosome region l l q l 3 , possibly near int-2. Oncogene Res. 3: 263-270. Ip, N.Y., Nye, S.H., Boulton, T.G., Davis, S., Taga, T., Li, Y., Birren, S.J., Yasukawa, K., Kishimoto, T., Anderson, D.J., Stahl, N. and Yancopoulos, G.D. (1992) CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component gpl30. C^//69: 1121-1132. Ishida, I. and Deguchi, T. (1983) Effect of depolarizing agents on choline acetyltransferase and acetylcholinesterase activities in primary cell cultures of spine cord. J. Neurosci. 3: 1818-1823. Iversen, L.L., Stockel, K. and Thoenen, H. (1975) Autoradiographic studies of the retrograde axonal transport of nerve growth factor in mouse sympathetic neurons. Brain Res. 88: 37-43. Janig, W. and Koltzenburg, M. (1991) Plasticity of sympathetic reflex organization following cross-union of inappropriate nerves in the adult cat. J. Physiol. (London) 436: 309-323. Johnson, E.M., Taniuchi, M., Clark, H.B., Springer, J.E., Koh, S., Tayrien, M.W. and Loy, R. (1987) Demonstration of the retrograde transport of nerve growth factor receptor in the peripheral and central nervous system. /. Neurosci. 1: 9 2 3 929. Johanson, S.O., Crouch, M.F. and Hendry, LA. (1995a) Retrograde transport of signalling proteins in rat sciatic nerve. Proc. Aust. Neurosci. Soc. 6: 103. Johanson, S., Crouch, M.W. and Hendry, LA. (1995b) Retrograde axonal transport of signal transduction proteins in rat sciatic nerve. Brain Res. 690: 55-63. Jonakait, G.M. and Schotland, S. (1990) Conditioned medium from activated splenocytes increases substance P in sympathetic ganglia. /. Neurosci. Res. 26: 24-30. Kanakis, S.J., Hill, C.E., Hendry, LA. and Watters, D.J. (1985) Sympathetic neuronal survival factors change after denervation. Brain Res. 352: 197-202. Kanje, M., Skottner, A., Lundborg, G. and Sjoberg, J. (1991) Does insulin-like growth factor I (IGF-I) trigger the cell body reaction in the rat sciatic nerve? Brain Res. 563: 2 8 5 287. Kanje, M., Skottner, A., Sjoberg, J. and Lundborg, G. (1989) Insulin-like growth factor I (IGF-I) stimulates regeneration of the rat sciatic nerve. Brain Res. 486: 396-398. Kaplan, D.R., Martin-Zanca, D. and Parada, L.F. (1991) Tyrosine phosphorylation and tyrosine kinase activity of the trk proto-oncogene product induced by NGF. Nature 350: 158-160. Keen, P. and McLean, W.G. (1974) Effect of dibutyryl-cyclic AMP and dexamethasone on noradrenaline synthesis in isolated superior cervical ganglia. J. Neurochem. 22: 5-10. Kessler, J.A. (1984a) Non-neuronal cell conditioned medium stimulates peptidergic expression in sympathetic and sensory neurons in vitro. Dev. Biol. 106: 61-69. Kessler, J.A. (1984b) Environmental co-regulation of sub-

LA. Hendry stance P, somatostatin and neurotransmitter synthesizing enzymes in cultured sympathetic neurons. Brain Res. 321: 155-159. Kessler, J.A. (1985a) Differential regulation of peptide and catecholamine characters in cultured sympathetic neurons. Neuroscience 15: 827-839. Kessler, J.A. (1985b) Parasympathetic, sympathetic, and sensory interactions in the iris: nerve growth factor regulates cholinergic ciliary ganglion innervation in vivo. J. Neurosci. 5: 2119-2125. Kessler, J.A. (1986) Differential regulation of cholinergic and peptidergic development in the rat striatum in culture. Dev. Biol. 113:77-89. Kessler, J.A. and Black, I.B. (1982) Regulation of substance P in adult rat sympathetic ganglia. Brain Res. 234: 182-187. Kessler, J.A., Adler, J.E., Bohn, M.C. and Black, I.B. (1981) Substance P in principal sympathetic neurons: regulation by impulse activity. Science 214: 335-336. Kessler, J.A., Adler J.E., Bell, W.O. and Black, I.B. (1983) Substance P and somatostatin metabolism in sympathetic and special sensory ganglia in vitro. Neuroscience 9: 309318. Kessler, J.A., Adler, J.E., Jonakait, G.M. and Black, I.B. (1984) Target organ regulation of substance P in sympathetic neurons in culture. Dev. Biol. 103: 71-79. Kessler, J.A., Conn, G. and Hatchers, V.B. (1986) Isolated plasma membranes regulate neurotransmitter expression and facilitate effects of a soluble brain cholinergic factor. Proc. Natl. Acad. Sci. USA 83: 3528-3532. Klippel, A., Escobedo, J.A., Fantl, W.J. and Williams, L.T. (1992) The C-terminal SH2 domain of p85 accounts for the high affinity and specificity of the binding of phosphatidylinositol 3-kinase to phosphorylated platelet-derived growth factor )3 receptor. Mol. Cell. Biol. 12: 1451-1459. Korsching, S. and Thoenen, H. (1985) Treatment with 6hydroxydopamine and colchicine decreases nerve growth factor levels in sympathetic ganglia and increases them in the corresponding target tissues. J. Neurosci. 5: 1058-1061. Landis, S.C. (1988) Neurotransmitter plasticity in sympathetic neurons and its regulation by environmental factors in vitro and in vivo. In: A. Bjorklund, T. Hokfelt and C. Owman (Eds.), Handbook of Chemical Neuroanatomy, Vol. 6. The Peripheral Nervous System, Ch. 3, Elsevier, Amsterdam, pp. 65-115. Landis, S.C. and Keefe, D. (1983) Evidence for neurotransmitter plasticity in vivo: developmental changes in properties of cholinergic sympathetic neurons. Dev. Biol. 98: 349372. Landmesser, L.T. (1980) The generation of neuromuscular specificity. Annu. Rev. Neurosci. 3: 279-302. Langley, J.N. (1895) Note on regeneration of preganglionic fibres of the sympathetic. J. Physiol (London) 18: 280-284. Langley, J.N. (1897) On the regeneration of preganglionic and postganglionic visceral nerve fibres. J. Physiol. (London) 22: 215-230. Langley, J.N. (1900) On axon reflexes in the pre-ganglionic

167 fibres of the sympathetic system. /. Physiol. (London) 25: 364-398. Lees, J.M., Adler, J.E. and Black, I.B. (1990) Regulation of neurotransmitter expression by a membrane-derived factor. Exp. Neurol. 108: 109-113. Lees, G., Chubb, I., Freeman, C , Geffen, L. and Rush, R.A. (1981) Effect of nerve activity on transport of nerve growth factor and dopamine ^^-hydroxylase antibodies in sympathetic neurons. Brain Res. 214: 186-189. Lefebvre, P.P., Van De Water, T.R., Represa, J., Liu, W., Bernd, P., Modlin, S., Moonen, G. and Mayer, M.B. (1991) Temporal pattern of nerve growth factor (NGF) binding in vivo and the in vitro effects of NGF on cultures of developing auditory and vestibular neurons. Acta Otolaryngol. (Stockholm) ni'.'iOA-'in. Levi-Montalcini, R. (1965) Growth regulation of sympathetic nerve cells. Arch. Ital. Biol. 103: 832-846. Levi-Montalcini, R. (1966) The nerve growth factor: its mode of action on sensory and sympathetic nerve cells. Harvey Lect. 60: 217-259. Levi-Montalcini, R. and Angeletti, P.U. (1968) Nerve growth factor. Physiol. Rev. 48: 534-569. Levi-Montalcini, R. and Hamburger, V. (1951) Selective growth stimulating effects of mouse sarcoma on the sensory and sympathetic nervous system of the chick embryo. J. Exp. Zool\l6: 321-362. Levi-Montalcini, R., Aloe, L., Mugnaini, E., Oesch, F. and Teioenen, H. (1975) Nerve growth factor induces volume increase and enhances tyrosine hydroxylase synthesis in chemically axotomized sympathetic ganglia of newborn rats. Proc. Natl. Acad. Sci. USA 72: 595-599. Lin, L.F., Mismer, D., Lile, J.D., Armes, L.G., Butler, E.T., Vannice, J.L. and Collins, F. (1989) Purification, cloning, and expression of ciliary neurotrophic factor (CNTF). Science 246: 1023-1025. Lindsay, R.M., Wiegand, S.J., Altar, C.A. and DiStefano, P.S. (1994) Neurotrophic factors: from molecule to man. Trends Neurosci. 17: 182-190. Linnemann, D. and Bock, E. (1986) Developmental study of detergent solubility and polypeptide composition of the neural cell adhesion molecule. Dev. Neurosci. 8: 24-30. Lipton, S.A., Wagner, J.A., Madison, R.D. and D'Amore, P.A. (1988) Acidic fibroblast growth factor enhances regeneration of processes by postnatal mammalian retinal ganglion cells in culture. Proc. Natl. Acad. Sci. USA 85: 2388-2392. Mackay, A.V. and Iversen, L.L. (1972) Increased tyrosine hydroxylase activity of sympathetic ganglia cultured in the presence of dibutyryl cyclic AMP. Brain Res. 48: 424-426. Maderdrut, J.L., Oppenheim, R.W. and Prevette, D. (1988) Enhancement of naturally occurring cell death in the sympathetic and parasympathetic ganglia of the chicken embryo following blockade of ganglionic transmission. Brain Res. 444: 189-194. Manthorpe, M., Skaper, S.D., Williams, L.R. and Varon, S. (1986) Purification of adult rat sciatic nerve ciliary neuronotrophic factor. Brain Res. 367: 282-286.

168 Mark, R.F. (1980) Synaptic repression at neuromuscular junctions. Physiol. Rev. 60: 355-395. Marotte, L.R. (1990) Development of retinotopy in projections from the eye to the of the dorsal lateral geniculate nucleus and superior colliculus of the wallaby (Macropus eugenii). J. Comp. Neurol. 293: 524-539. Martin, D.P., Schmidt, R.E., DiStefano, P.S., Lowry, O.H., Carter, J.G. and Johnson, E.M. (1988) Inhibitors of protein synthesis and RNA synthesis prevent neuronal death caused by nerve growth factor, deprivation. J. Cell. Biol. 106: 829844. Martinou, J.C, Bierer, F., Van Thai, A.L. and Weber, M.J. (1989) Influence of the culture substratum on the expression of choline acetyltransferase activity in purified motoneurons from rat embryos. Dev. Brain Res. 47: 251-262. Matthews, M.R. and Nelson, V.H. (1975) Detachment of structurally intact nerve endings from chromatolytic neurons of the rat superior cervical ganglion during the depression of synaptic transmission induced by postganghonic axotomy. J. Physiol. (London) 145: 91-135. Mazzoni, E. and Kenigsberg, R.L. (1991) Thrombin indirectly affects cholinergic cell expression in primary septal cell cultures in a manner distinct from nerve growth factor. Neuroscience 45: 195-204. Narayanan, Y. and Narayanan, C.H. (1981) Ultrastructural and histochemical observations in the developing iris musculature in the chick. J. Embryol. Exp. Morphol. 62: 117127. Nasi, S., Cirillo, D., Naldini, L., Marchisio, P.C. and Calissano, P. (1982) Microtubules and microfilaments in fixed and permeabilized cells are selectively decorated by nerve growth factor. Proc. Natl. Acad. Sci. USA 79: 820824. Nawa, H. and Patterson, P.H. (1990) Section and partial characterization of neuropeptide-inducing factors in heart cell conditioned medium. Neuron 4: 269-277. Nawa, H. and Sah, D.W.Y. (1990) Different biological activities in conditioned media control the expression of a variety of neuropeptides in cultured sympathetic neurons. Neuron 4: 279-287. Nja, A. and Purves, D. (1978) Specificity of initial synaptic contacts made on guinea-pig superior cervical ganglion cells during regeneration of the cervical sympathetic trunk. J. Physiol. (London) 281: 45-62. Noakes, P.G., Bennett, M.R. and Davey, D.F. (1983) Growth of segmental nerves to the developing rat diaphragm: absence of pioneer axons. /. Comp. Neurol. 218: 365-377. Noakes, P.G., Bennett, M.R. and Stratford, J. (1988) Migration of Schwann cells and axons into developing chick forelimb muscles following removal of either the neural tube or the neural crest. J. Comp. Neurol. 277: 214-233. Okado, N. and Oppenheim, R.W. (1984) Cell death of motoneurons in the chick embryo spinal cord. IX. The loss of motoneurons following removal of afferent inputs. J. Neurosci. 4: 1639-1652. Oppenheim, R.W. (1981) Cell death of motoneurons in the

Retrograde factors in peripheral nerves chick embryo spinal cord. V. Evidence on the role of cell death and neuromuscular function in the formation of specific peripheral connections. J. Neurosci. 1: 141-151. Oppenheim, R.W. (1984) Cell death of motoneurons in the chick embryo spinal cord. VIII. Motoneurons prevented from dying in the embryo persist after hatching. Dev. Biol. 101:35-39. Oppenheim, R.W. and Nunez, R. (1982) Electrical stimulation of hindlimb increases neuronal cell death in chick embryo. Nature 295: 57-59. Oppenheim, R.W., Maderdrut, J.L. and Wells, D.J. (1982a) Cell death of motoneurons in the chick embryo spinal cord. VI. Reduction of naturally occurring cell death in the thoracolumbar column of Terni by nerve growth factor. J. Comp. Neurol. 210: 174-189. Oppenheim, R.W., Maderdrut, J.L. and Wells, D.J. (1982b) Reduction of naturally-occurring cell death in the thoracolumbar preganglionic cell column of the chick embryo by nerve growth factor and hemicholinium-3. Brain Res. 255: 134-139. Oppenheim, R.W., Cole, T. and Prevette, D. (1989) Early regional variations in motoneuron numbers arise by differential proliferation in the chick embryo spinal cord. Dev. Biol. 133: 468^74. Oppenheim, R.W., Prevette, D., Tytell, M. and Homma, S. (1990) Naturally occurring and induced neuronal death in the chick embryo in vivo requires protein and RNA synthesis: evidence for the role of cell death genes. Dev. Biol. 138:104-113. Ouimet, C.C, Wang, J.K., Walaas, S.I., Albert, K.A. and Greengard, P. (1990) Localization of the MARCKS (87 kDa) protein, a major specific substrate for protein kinase C, in rat brain. J. Neurosci. 10: 1683-1698. Paravicini, U., Stockel, K. and Thoenen, H. (1975) Biological importance of retrograde axonal transport of nerve growth factor in adrenergic neurons. Brain Res. 84: 279-991. Patterson, P.H. and Chun, L.L.Y. (1974) The influence of non-neuronal cells on catecholamine and acetylcholine synthesis and accumulation in cultures of dissociated sympathetic neurons. Proc. Natl. Acad. Sci. USA 71: 36073610. Patterson, P.H. and Chun, L.L.Y. (1977a) The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. I. Effects of conditioned medium. Dev. Biol. 56: 263-280. Patterson, P.H. and Chun, L.L.Y. (1977b) The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. II. Developmental aspects. Dev. Biol. 60: 413-4SI. Pilar, G., Landmesser, L. and Burstein, L. (1980) Competition for survival among developing ciliary ganglion cells. J. Neurophysiol. 43: 233-254. Potter, D.D., Landis, S.C, Matsumoto, S.G. and Furshpan, E.J. (1986) Synaptic functions in rat sympathetic neurons in microcultures. II. Adrenergic, cholinergic dual status and plasticity. J. Neurosci. 6: 1080-1098.

I.A. Hendry Prestige, M.C. (1974) Axon and cell numbers in the developing nervous system. Br. Med. Bull. 30: 107-111. Prestige, M.C. (1976) Evidence that at least some of the motor nerve cells that die during development have first made peripheral connections. 7. Comp. Neurol. 170: 123-133. Purves, D. (1975) Functional and structural changes in mammalian sympathetic neurons following interruption of their axons. J. Physiol. (London) 252: 429-463. Purves, D. (1988) Body and Brain. A Trophic Theory of Neural Connections. Harvard University Press, Cambridge. Rao, M.S., Sun, Y., Escary, J.L., Perreau, J., Tresser, S., Patterson, P.H., Zigmond, R.E., Brulet, P. and Landis, S.C. (1993) Leukemia inhibitory factor mediates an injury response but not a target-directed developmental transmitter switch in sympathetic neurons. Neuron 11:1175-1185. Rogelj, S., Klagsbrun, M., Atzmon, R., Kurokawa, M., Haimovitz. A., Fuks, Z. and Vlodavsky, I. (1989) Basic fibroblast growth factor is an extracellular matrix component required for supporting the proliferation of vascular endothelial cells and the differentiation of PC 12 cells. J. Cell. Biol. 109:823-831. Rohrer, H., Otten, U. and Thoenen, H. (1978) On the role of RNA synthesis in the selective induction of tyrosine hydroxylase by nerve growth factor. Brain Res. 159: 436-439. Rohrer, H., Schiifer, T., Korsching, S. and Thoenen, H. (1982) Internalization of nerve growth factor by pheochromocytoma PC 12 cells: absence of transfer to the nucleus. J. Neurosci. 2: 687-697. Romagnano, M.A., Newton, B.W., Pierre, A.H. and Hamill, R.W. (1989) Postnatal ontogeny of enkephalin fibers in spine sympathetic nuclei. Dev. Brain Res. 46: 263-279. Rush, R.A., Saltis , J., Smet, P.J. and Williams, R. (1999) Survival factors for autonomic neurons. In: I.A. Hendry and C.E. Hill (Eds.), Development, Regeneration, and Plasticity of the Autonomic Nervous System, Harwood Academic Publishers, London, pp. 305-363. Saadat, S., Sendtner, M. and Rohrer, H. (1989) Ciliary neurotrophic factor induces cholinergic differentiation of rat sympathetic neurons in culture. /. Cell Biol. 108: 18071816. Schotzinger, R.J. and Landis, S.C. (1988) Cholinergic phenotype developed by noradrenergic sympathetic neurons after innervation of a novel cholinergic target in vivo. Nature 335: 637-639. Schotzinger, R.J. and Landis, S.C. (1990) Acquisition of cholinergic and peptidergic properties by sympathetic innervation of rat sweat glands requires interaction with normal target. Neuron 5: 91-100. Schwab, M.E. (1977) Ultrastructural localization of a nerve growth factor-horseradish peroxidase (NGF-HRP) coupling product after retrograde axonal transport in adrenergic neurons. Brain Res. 130: 190-196. Schwab, M.E. and Thoenen, H. (1977) Selective transsynaptic migration of tetanus toxin after retrograde axonal transport in peripheral sympathetic nerves: a comparison with nerve growth factor. Brain Res. 122: 4 5 9 ^ 7 4 .

169 Schwarting, G.A., Tischler, A.S. and Donahue, S.R. (1990) Fucosylation of glycolipids in PC 12 cells is dependent on the sequence of nerve growth factor treatment and adenylate cyclase activation. Dev. Neurosci. 12: 159-171. Seeley, P.J. and Greene, L.A. (1983) Short-latency local actions of nerve growth factor at the growth cone. Proc. Natl. Acad. Sci. USA 80: 2789-2793. Seeley, P.J., Keith, C.H., Shelanski, M.L. and Greene. L.A. (1983) Pressure microinjection of nerve growth factor and anti-nerve growth factor into the nucleus and cytoplasm: lack of effects on neurite outgrowth from pheochromocytoma cells. J. Neurosci. 3: 1488-1494. Sendtner, M., Kreutzberg, G.W. and Thoenen, H. (1990) Ciliary neurotrophic factor prevents the degeneration of motor neurons after axotomy. Nature 345: 4 4 0 ^ 4 1 . Sjoberg, J. and Kanje, M. (1989) Insulin-like growth factor (IGF-1) as a stimulator of regeneration in the freeze-injured rat sciatic nerve. Brain Res. 485: 102-108. Smet, P.J., Abrahamson, I.K., Ressom, R.E. and Rush, R.A (1991) A ciliary neuronotrophic factor from peripheral nerve and smooth muscle which is not retrogradely transported, Neurochem. Res. 16: 613-620. Sperry, R.W. (1951) Regulative factors in the orderly growth of neural circuits. Growth 10: 63-87. Stevens, L.M. and Landis, S.C. (1990) Target influences on transmitter choice by sympathetic neurons developing in the anterior chamber of the eye. Dev. Biol. 137: 109192. Stockel, K., Dumas, M. and Thoenen, H. (1978) Uptake and subsequent retrograde axonal transport of nerve growth factor (NGF) are not influenced by neuronal activity. Neurosci. Lett. 10: 61-64. Stockli, K.A., Lottspeich, F., Sendtner, M., Masiakowski, P., Carroll, P., Gotz, R., Lindholm, D. and Thoenen, H. (1989) Molecular cloning, expression and regional distribution of rat ciliary neurotrophic factor. Nature 342: 920-923. Stumpo, D.J., Graff, J.M., Albert, K.A., Greengard, P. and Blackshear, P.J. (1989) Nucleotide sequence of a cDNA for the bovine myristoylated alanine-rich C kinase substrate (MARCKS). Nucleic Acids Res. 17: 3987-3988. Teichberg, S., Holtzman, E., Grain, S.M. and Peterson, E. (1975) Circulation and turnover of synaptic vesicle membrane in cultured fetal mammalian spinal cord neurons. J. Cell Biol. 61: 215-230. Thoenen, H. and Barde, Y.A. (1980) Physiology of nerve growth factor. Physiol. Rev . 60: 1284-1335. Thoenen, H., Korsching, S., Heumann, R. and Acheson, A. (1985) Nerve growth factor. Ciba Found. Symp. 116: 113128. Tosney, K.W. and Landmesser, L.T. (1985a) Specificity of early motoneuron growth cone outgrowth in the chick embryo. J. Neurosci. 5: 2336-2344. Tosney, K.W. and Landmesser, L.T. (1985b) Development of the major pathways for neurite outgrowth in the chick hindlimb. Dev. Biol. 109: 193-214. Walicke, P.A., Campenot, R.B. and Patterson, P.H. (1977)

170 Determination of transmitter function by neuronal activity. Proc. Natl Acad. Sci. USA 74: 5767-5771. Wang, J.K., Walaas, S.I. and Greengard, P. (1988) Protein phosphorylation in nerve terminals: comparison of calcium calmodulin-dependent and calcium diacylglycerol-dependent systems. J. Neurosci. 8: 281-288. Wang, J.K., Walaas, S.I., Sihra, T.S., Aderem, A. and Greengard. P. (1989) Phosphorylation and associated translocation of the 87-kDa protein, a major protein kinase C substrate in isolated nerve terminals. Proc. Natl. Acad. Sci. USA 86: 2253-2256. Watters, D.J., Belford, D.A., Hill, C.E. and Hendry, LA. (1989) Monoclonal antibody that inhibits biological activity of a mammalian ciliary neurotrophic factor. J. Neurosci. Res. 22: 60-64. Weiss, P. (1942) Lid-closure reflex from eyes transplanted to atypical locations in Tritorus torosus: evidence of a peripheral origin of sensory specificity. J. Comp. Neurol. 77: 131169. Westermann, R., Johannsen, M., Unsicker, K. and Grothe, C. (1990) Basic fibroblast growth factor (bFGF) immunoreactivity is present in chromaffin granules. J. Neurochem. 55: 285-292. Woolf, C.J., Reynolds, M.L., Molander, C , O'Brien, C ,

Retrograde factors in peripheral nerves Lindsay, R.M. and Benowitz, L.I. (1990) The growthassociated protein GAP-43 appears in dorsal root ganglion cells and in the dorsal horn of the rat spinal cord following peripheral nerve injury. Neuroscience 34: 465-478. Yamamori, T., Fukada, K., Aebersold, K., Korsching, S., Fann, M.-J. and Patterson, P.H. (1989) The cholinergic neuronal differentiation factor from heart cells is identical to leukemia inhibitory factor. Science 246: 1412-1416. Yan, Q., Snider, W.D., Pinzone, J.J. and Johnson, E.M. (1988) Retrograde transport of nerve growth factor (NGF) in motoneurons of developing rats: assessment of potential neurotrophic effects. Neuron 1: 335-343. Yin, Q.W., Johnson, J., Prevette, D. and Oppenheim, R.W. (1994) Cell death of spinal motoneurons in the chick embryo following deafferentation: rescue effects of tissue extracts, soluble proteins, and neurotrophic agents. J. Neurosci. 14: 7629-7640. Zacks, S.I. and Saito, A. (1969) Uptake of exogenous horseradish peroxidase by coated vesicles in mouse neuromuscular junctions. 7. Histochem. Cytochem. 17: 161-170. Zhan, X., Bates, B., Hu, X.G. and Goldfarb, M. (1988) The human FGF-5 oncogene encodes a novel protein related to fibroblast growth factors. Mol. Cell. Biol. 8: 3487-3495.

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 7

The regulation of nerve growth factor synthesis and deUvery to peripheral neurons R.A. Rush, R. Mayo and C. Zettler Department of Human Physiology and Centre for Neuroscience, GPO Box 2100, Adelaide 5001, South Australia, Australia

1. Introduction Nerve growth factor (NGF) is a potent protein identified more than 40 years ago, which exerts profound effects on components of the peripheral and central nervous systems. There are several excellent reviews covering many aspects of the extensive work that has allowed NGF to gain a prominent place in the study of the nervous system, most of which have addressed issues related to the known actions of NGF on sensitive neurons (Bradshaw, 1978, 1983; Greene and Shooter, 1980; Thoenen and Barde, 1980; Levi-Montalcini, 1987; Rush et al., 1992; Schecterson and Bothwell, 1992). This review will be restricted to an analysis of experiments that have provided information concerning the regulation of NGF synthesis, with special emphasis on peripheral tissues. The mechanisms involved in the regulation of NGF synthesis have received attention only recently, not because of disinterest by workers in the field, but primarily because the necessary techniques have not been available. Technical issues frequently have hindered the study of NGF and continue to prevent resolution of several important questions regarding its function, despite the explosion of data being generated by a rapidly expanding list of researchers. Much of this frustration can be traced to the extremely low concentration of NGF and its mRNA in relevant tissues, forcing researchers to seek new methods for their measurement and then pushing these to their limits. Nevertheless, we feel that it is an appropriate time to take stock of the current knowledge in the area of NGF regulation

and to discuss the issues requiring resolution in the immediate future. 2. NGF and the neurotrophins Molecular cloning has revealed that NGF is a member of a unique gene family, named the neurotrophins (NTs), whose other members include brain-derived neurotrophic factor (BDNF; Barde et al., 1982; Leibrock et al, 1989), NT-3, (Emfors et al., 1990; Hohn et al., 1990; Maisonpierre et al., 1990) and NT-4/5 (Berkemeier et al., 1991; Hallbook et al., 1991; Ip et al, 1992). Although these proteins share considerable amino acid homology (approximately 50%), much of which is clustered around six absolutely conserved cysteine residues (Leibrock et al., 1989), they display unique, as well as overlapping, biological activities. Examination of the biological activities of all but NGF have been almost entirely restricted so far to survival responses in various neuronal populations. For example, while cultured sympathetic neurons are dependent on NGF for survival (LeviMontalcini, 1987; Thoenen et al., 1987) and are refractory to BDNF (Barde et al, 1982; Lindsay et al., 1985; Davies et al., 1986a), they exhibit weak responsiveness to NT-3 (Emfors et al., 1990; Maisonpierre et al., 1990) and NT-4/5 (Berkemeier et al., 1991). In contrast, NGF, BDNF and NT-3 support distinct subpopulations of cultured neuralcrest-derived sensory neurons (Lindsay et al., 1985; Acheson et al., 1987; Thoenen et al, 1987; Barde, 1989; Knusel et al., 1991; Hory-Lee et al., 1993), while BDNF and NT-3 are active also on

172

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

subsets of NGF-unresponsive sensory neurons originating from the ectodermal placodes (Lindsay and Rohrer, 1985; Lindsay et al., 1985; Ernfors et al., 1990; Hohn et al, 1990; Rosenthal et al, 1990). Furthermore, NGF does not support the survival of embryonic motor neurons, whereas BDNF. NT-3 and NT-4/5 are all effective to varying degrees (Oppenheim et al., 1992; Sendtner et al., 1992; Yan et al., 1992; Henderson et al., 1993). A clear delineation in the actions of the NTs has also been demonstrated with cultured retinal ganglion cells (Rodriguez-Tebar et al., 1989; Snider and Johnson, 1989; Thanos and Von Boxberg, 1989; Thanos et al., 1989), developing dopaminergic neurons of the substantia nigra and ventral mesencephalon (Hyman et al., 1991; Knusel et al., 1991) and septal cholinergic neurons (Alderson et al, 1990; Knusel et al., 1991), which are solely responsive to BDNF. Consistent with these studies in vitro is the ability of NGF and BDNF supplementation to selectively support divergent neuronal populations in vivo (LeviMontalcini, 1987; Hofer and Barde, 1988). Detailed coverage of the properties and actions of a variety of neurotrophins other than NGF may be found in other chapters of this book. 3. Physiology of NGF 3.1. A brief overview In addition to an ability to keep neurons alive, NGF has many other actions on sensitive neurons, ranging from immediate alterations in the cell membrane to long-term control of metabolic function. While these effects have been well documented (see Bradshaw, 1978; Thoenen and Barde, 1980; Levi-Montalcini, 1987), little is known of the metabolic effects of the newly described NTs. NGF supplementation in vivo results in striking alterations in sympathetic and sensory neuron size (Aloe et al., 1975; Hendry and Campbell, 1976; Hendry, 1977; Levi-Montalcini, 1987; Zettler et al., 1991), peripheral axonal branching (Diamond et al., 1987) and dendritic arborization (Snider, 1988; Ruit et al., 1990). Significant effects on neurotransmitter, neuropeptide and neurofilament pro-

duction have also been described. Of particular importance to this current review is the regulation of enzymes involved in catecholamine synthesis (Thoenen et al., 1971; Stockel et al., 1974; Goedert et al, 1978; Otten et al., 1978; Kessler and Black, 1980a; Komblum and Johnson, 1982) and neuropeptides, such as substance P, somatostatin, calcitonin gene-related peptide and cholecystokinin (Kessler and Black, 1980b, 1981; Otten and Lorenz, 1982; Otten et al., 1982; Hayashi et al., 1985; Lindsay etal., 1989). NGF affects neurite outgrowth in a variety of neurons and has a potent chemotactic action both in vitro (Letoumeau, 1978; Gundersen and Barrett, 1979; Campenot, 1982) and in vivo (LeviMontalcini, 1983; Diamond et al., 1987; Zettler et al., 1991; Hoyle et al, 1993). Thus, its supplementation has been shown not only to alter the pattern and density of responsive sympathetic neurons, but also to exert indirect effects on preganglionic cholinergic and sensory fibres innervating sympathetic ganglia (Nja and Purves, 1978; Purves and Lichtman, 1978; Oppenheim et al., 1982; Schafer et al., 1983; Hendry et al., 1992). While it is generally agreed that endogenous NGF does not guide growing axons to their peripheral effector tissues (Davies and Lumsden, 1983; Davies et al, 1987), the expression of NGF by effector cells may be important for terminal arborization and the establishment of functional synaptic connections (Diamond et al., 1987; Hoyle et al., 1993). 3.2. Results from antibody administration While the administration of exogenous NGF has potent effects on responsive neurons, a physiological role for the factor is best demonstrated by the neutralization of endogenous NGF. This has been achieved by the administration of antibodies raised against NGF (anti-NGF), which generally exert opposite effects to those of NGF supplementation. Treatment of neonates with antiNGF results in depression of anabolic activity, neurotransmitter synthesis and neurite outgrowth and eventual destruction of most sympathetic neurons (Cohen 1960; Klingman and Klingman, 1972; Goedert et al., 1978; Levi-Montalcini, 1983),

R.A. Rush et al.

while prenatal exposure in utero leads to the destruction of certain spinal sensory neurons (see Johnson et al., 1986). However, recent studies indicate some antibodies to NGF cross-react with NT-3 (Negro et al., 1993; Murphy et al., 1993) and BDNF (Ruit et al., 1992; Murphy et al., 1993). Recognizing this limitation, Ruit et al. (1992) reexamined the effects of NGF antibodies on sensory neuron function in utero using antibodies that failed to cross-react with recombinant human NT3 or recombinant human BDNF. Their results have verified the NGF dependence of approximately 70% of dorsal root ganglion (DRG) neurons, almost all being of a small diameter and projecting to laminae I and II of the dorsal horn. While the most obvious effect of NGF deprivation is cell death, the mechanism by which this occurs is unknown. The existence of a suicide, or apoptotic, program has been shown and suggested as the target for NGF (Miller et al., 1991). Apoptotic cell death in neurons may require the presence of a functional low-affinity NGF receptor (Rabizadeh etal., 1993). In the mature animal, anti-NGF administration generally fails to cause neuronal death, but does result in altered metabolism of responsive neurons (Goedert et al., 1978; Schwartz et al., 1982; Lindsay and Harmar, 1989). Nevertheless, extensive neuronal death has been reported for sympathetic neurons when active immunization is used to achieve continuous, high circulatory levels of antibodies (Johnson et al., 1982). These responses are opposite to those seen with NGF administration. For example, neuropeptides, which are upregulated by NGF, are down-regulated by antiNGF (see, for example, Otten and Lorenz, 1982; Otten et al., 1982; Schwartz et al., 1982; Lindsay and Harmer, 1989). In addition, more subtle changes, including the retraction of preganglionic synapses from the postganglionic cell body, have been observed (Nja and Purves, 1978; Gorin and Johnson, 1979). These effects of anti-NGF indicate a continuing requirement of the mature neuron for the factor, but the mechanism that allows the maturing neuron to become more resistant to antiNGF is not known. Alterations in calcium metabolism (Franklin and Johnson, 1992; Larmet et

173

al., 1992), suppression of endonuclease activity (Batistatou and Greene, 1991) and the involvement of the proto-oncogenes, ras (Borasio et al., 1993) and bcl-2 (Garcia et al., 1992), are some pathways that appear to be involved. 4. NGF in disease 4.1. An overview Abnormal NGF levels have been described in association with several diseases in both animals and humans, suggesting that disordered control of NGF synthesis may result in neuronal pathologies. Study of these diseases, therefore, may lead to the identification of agents capable of regulating NGF synthesis (see Stewart and Appel, 1988). For example, peripheral neuropathies associated with streptozotocin-induced diabetes have been linked to reduced NGF levels, which can be reversed completely by pancreatic islet cell transplants or partially by administration of catechols (Hellweg and Hartung, 1990; Hellweg et al., 1991; Hanaoka et al., 1992). Altered corticosterone and vitamin D3 levels in this experimental animal may influence NGF synthesis also (Neveu et al., 1992b). Peripheral nerve damage resulting from the use of chemotherapeutic agents can be reversed partially by the simultaneous administration of NGF, suggesting that these agents reduce the availability of NGF to these neurons (Apfel et al., 1991, 1992; Figliomeni et al., 1992). Within the CNS, decreased levels of NGF have been found in the brains of aged rats (Larkfors et al., 1987) and genetically ataxic mice (Matsui et al., 1990). Identification of NGF-responsive cholinergic neurons in the basal forebrain (Honegger and Lenoir, 1982; Gnahn et al., 1983; Hefti et al., 1985; Hefti, 1986) has led to the proposal that degenerative changes in these neurons, which coincide with the onset of Alzheimer's disease, may be related to abnormal NGF expression (Hefti and Weiner, 1986; Thoenen et al., 1991). However, initial studies have found normal mRNA^^^ levels in autopsied brains (Goedert et al., 1986). Decreased concentrations of NGF have been found in the blood of patients suffering from a variety of

174

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

neurodegenerative diseases (Lorigados et al., 1992). A recent finding that NGF can restore catalase, superoxide dismutase and glutathione peroxidase activity in aged brains and, thus, protect sensitive neurons from free radical damage, suggests decreased NGF levels within the CNS may underlie some of the pathophysiological changes associated with disease and senescence (Nistico et al, 1992). While a decreased production of NT (or its receptors; see Snider and Johnson, 1989) may underlie some neurodegenerative diseases, increased production would also be expected to have dramatic effects on neuronal function. Abnormal development and growth of neural-crest derivatives in the autosomal dominant disorder of neurotibromatosis has been correlated with elevated levels of NGF in human sera (Siggers et al., 1975; Fabricant et al, 1979; Fabricant and Todaro, 1981). Similarly, increased 'nerve-growth-stimulating activity' was found in a patient with an intestinal ganglioneuromatosis, in which there is a massive proliferation of the intestinal myenteric plexus and associated extrinsic nerve fibres (DeSchryverKecskemeti et al., 1983). Augmented NGF synthesis also has been implicated in the pathogenesis of rheumatoid arthritis (Aloe et al., 1992) and essential hypertension (see Head, 1989). 4.2, The spontaneously hypertensive rat One animal model that is receiving increasing attention is the spontaneously hypertensive rat (SHR). This is because an abnormal NGF production in the SHR appears responsible for the vascular hyperinnervation implicated in the elevated blood pressure. The vascular system contributes significantly to the innervation territory of NGF-responsive sympathetic neurons (Hill et al., 1985; Lee et al., 1987, 1992; Isaacson et al., 1990; Zettler et al., 1991; Zettler and Rush, 1993), and several studies provide persuasive evidence for the presence of abnormal sympathetic nerves in blood vessels of the SHR (Head, 1989). This abnormality takes the form of an augmented sympathetic innervation (Lee et al., 1983a,b; Scott and Pang, 1983; Lee

and Saito, 1984; Cassis et al., 1985) and can be experimentally induced in normotensive genetic controls by chronic NGF supplementation (Zettler et al., 1991; Lee et al., 1992). However, NGF treatment alone is insufficient to induce hypertension, implicating the involvement of additional factors (Zettler et al., 1991; Lee et al., 1992). An enhanced vascular production and accumulation of NGF has been demonstrated recently in the SHR (Donohue et al., 1989; Falckh et al., 1992a,b; Ueyama et al., 1992; Zettler and Rush, 1993). Comparison of NGF levels in mesenteric arteries of the SHR and its normotensive control indicates levels become elevated from 15 days of age and coincide with the initiation of sympathetic transmission (Hill et al., 1983) and abnormal smooth muscle cell growth (Zettler and Rush, 1993). Although the precise mechanisms leading to the overexpression of NGF in this animal have yet to be determined, indirect evidence suggests a relationship between abnormal vascular growth, the production of NGF and the activity of the renin-angiotensin system. This evidence is derived from the above and following observations. Smooth muscle cells appear to be the primary site of vascular NGF synthesis in vivo (Scarisbrick et al., 1993). Angiotensin II (A II) has been shown to enhance both NGF gene expression (Creedon and Tuttle, 1991) and mitosis (Lyall et al., 1988) in cultured smooth muscle cells. Plasma and vascular levels of renin are elevated in the SHR during postnatal development (Naruse and Inagami, 1982; Harrap and Doyle, 1986) while postnatal administration of angiotensin converting enzyme inhibitors prevents the development of vascular abnormalities (Clozel et al., 1989) and hypertension (Harrap and Lever, 1989). Moreover, genetic linkage studies have implicated abnormalities in the genes encoding angiotensin converting enzyme in the pathophysiology of hypertension (Hilbert et al., 1991). In addition, the low affinity NT receptor (p75), which is abundantly expressed by vascular smooth muscle cells at the time of differentiation and innervation (Schecterson and Both well, 1992), is also implicated by genetic analysis (Hilbert et al., 1991; Nemoto et al., 1994).

RA. Rush et al.

Finally, the postnatal administration of NGF antiserum to SHR prevents the development of vascular abnormalities and hypertension (Folkow et al., 1972; Lee et al., 1987). These observations are consistent with the hypothesis that vascular NGF gene expression in vivo may be regulated by A II, and that an abnormal interaction between these two systems may manifest itself as high blood pressure. That elevated NGF alone cannot result in hypertension has been shown by the chronic administration of NGF to neonatal rats, which results in the hyperinnervation of resistance vessels, but a normal blood pressure (Zettler et al, 1991; Lee etal, 1992).

175

stimulate NGF synthesis within the appropriate cells in vivo (Lindholm et al., 1990a; Spranger et al., 1990; Hanaoka et al., 1992; Neveu et al., 1992b; Kaechi et al., 1993). However, the development of therapeutic agents to improve impaired neuronal function via stimulation of endogenous NGF synthesis requires answers to a number of fundamental questions. These include the identification of cell types synthesising NGF, the factors that initiate and regulate gene expression and NGF synthesis and secretion for each cell type, and the mechanisms by which these become disturbed in disease. 5. The NGF gene and transcriptional control

4.3. Therapeutic implications 5.1. Organization of the gene Taken together, these findings suggest that NGF (and other NTs) are likely to play an important role in disease. This has led to the optimistic view that NGF therapy may be employed to halt, or even reverse, some pathological processes. Several approaches can be used to increase NGF levels in vivo. Supplementation of the endogenous levels can be achieved either by direct administration or by introduction of genetically modified cells or viruses capable of producing NGF in vivo (Bartus et al., 1982; Kirino, 1982; Williams et al., 1986; Fischer et al., 1987; Kromer, 1987; Gage et al., 1988; Hagg et al., 1988, 1990; Montero and Hefti, 1988; Shigeno et al., 1991; Federoff et al, 1992). However, drug therapy aimed at reducing or enhancing endogenous NGF gene expression is possible and, indeed, has a number of advantages over direct NGF administration. In particular, since NGF appears not to act in an endocrine fashion, but as a paracrine hormone, the consequences of high circulatory levels of NGF require consideration. Some evidence of abnormal nerve growth has been observed as a result of elevated blood levels of NGF (Levi-Montalcini and Hamburger, 1951; Levi-Montalcini, 1987; Zettler etal., 1991). Current knowledge of the factors regulating NGF production in vivo is poor, so initial clinical trials have utilized NT administration (Olson et al, 1991, 1992). Nevertheless, the search has begun for compounds with the ability to selectively

Although the 'nerve growth-promoting protein' was first purified from the mouse salivary gland by Cohen (1960), it took more than a decade to determine the amino acid sequence of NGF (Angeletti et al., 1973) and another 10 years to achieve the first full gene sequence analysis (Scott et al., 1983; Ullrich et al., 1983). Following the sequence analysis of a portion of the NGF gene (Scott et al., 1983; Ullrich et al., 1983), Selby et al. (1987) reported the presence of four types of cDNA clones with different sequences preceding the region coding for NGF protein. The existence of four transcripts suggested a complex gene structure. Further analysis of the clones derived from a mouse genomic library (Selby et al., 1987) indicated the existence of at least four exons (lA and IB, II, III A and IIIB, IV) separated by introns, including one of 30 kb, and a possible transcription unit of more than 45 kb (see Fig. 1; Selby et al., 1987). Exons lA, IIIA and IIIB are non-coding and correspond to the different 5' sequences present in the clones derived from the cDNA library. The region coding for the NGF protein is located exclusively in exon IV, although exons lA and II both have additional in-frame translation initiation codons (AUG). On the basis of this NGF gene structure, the size of all alternative splice forms is predicted as approximately 1.3 kb. How-

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

176 142 bp

-

167 I 33 ^ •§-

>32kb

kb

127

I

^6 kb

161 124 ^

1

894

-

^ NGF

ATG

B

SMG Relative Abundance 100

10

Fig. 1. Diagrammatic representation of the predicted NGF transcripts in relation to the gene. The gene is shown at the top with exons as boxes and introns as lines. The size of exons (in base pairs) is shown above the boxes. The size of the introns is also shown. Mature NGF is stippled. The four identified NGF transcripts are depicted below in order of decreasing submaxillary gland (SMG) abundance (A, B, D and then C). The thick lines represent sequences formed in the mature RNA, and the thin lines represent regions that are removed by splicing of a primary transcript. The presumed sites for initiation of translation (ATG) are indicated. Reprinted from Selby et al. (1987), with permission of the authors and the copyright holder, The American Society for Microbiology, Washington, DC.

ever, Northern analysis data from various tissues has revealed mRNA forms additional to the predicted 1.3 kb. The origin of these larger mRNA species of 1.7 and 1.9 kb is yet to be explained (Shelton and Reichardt, 1986a,b; Whittemore et al., 1986, 1987; Whittemore and Seiger, 1987; Falckh et al., 1992a,b). Furthermore, the existence of multiple transcripts of 1.3 kb means that a signal of this size on Northern analysis may represent a mixed population of NGF transcripts. The results of primer extension and SI nuclease mapping experiments indicate that the 5' exons are differentially expressed. Transcript A (see Fig. 1) is highly expressed in mouse submandibular gland and placenta, whereas in all peripheral organs and throughout most brain regions, transcript B predominates, and transcript C is preferentially expressed in the cerebral cortex and heart. Interestingly, the amount of transcript D is relatively high in the submandibular gland (1% of mRNA^^^), which would suggest that it is likely to represent an uncharacterized mRNA^^^ rather than a par-

tially spliced intermediate of the primary NGF transcript (Selby et al., 1987). Differential expression can be explained either by specific tissue usage of different promoters or the presence of regulatory sequences {ciselements) within a unique promoter recognized by tissue-specific sets of trans-?iC\m% factors. The latter concept is commonly accepted for NGF gene regulation. However, a mechanism where alternate promoters generate complex spatio-temporal patterns of a gene product is well characterized for BDNF (Timmusk et al., 1993). Therefore, the possibility cannot be ruled out that this mechanism is important in NGF gene regulation. A common feature with all mRNA^^^ transcripts is the 5' untranslated sequences. The regulatory role, if any, of these sequences is unknown, but it has been demonstrated for several protooncogenes that non-coding 5' sequences on the mRNA impair translation (Kozak, 1988). Transcripts A and C have the potential to encode for different forms of NGF precursor due to the pres-

R.A. Rush et al.

ence of the additional in-frame AUG codons. Studies in vitro have demonstrated that the AUG codons of exons II and IV initiate translation (Edwards et al., 1988a). However, whether the AUG codon in exon lA is functional remains to be established. Regardless of the initiation site, all NGF protein products undergo proteolytic cleavage at the N- and C-termini, resulting in an identical, mature, biologically active protein of 118 amino acids (Edwards et al., 1988a). 5.2. NGF promoter To date, only the region immediately upstream to exon IB has been characterized structurally, exhibiting transcription-promoting activities. Within this region, several a^'-regulatory elements have been identified; two putative TATA-boxes (at positions -28 and -49 bp relative to the major cap site), two putative CAAT-boxes (at -379 and -546 bp, respectively) and a high content of G and C residues indicative of putative SPl sites (Zheng and Heinrich, 1988). The functional significance of these DNA elements in NGF gene regulation remains to be established. It has been demonstrated that intronic DNA flanking 5 kb upstream to exon IB contains all the elements required for appropriate tissue-specific expression of the NGF gene. A construct, where this particular gene fragment was fused to the human growth hormone gene as reporter, was introduced into transgenic mice. The transgenic mRNA reflected the tissue distribution of that predicted for mRNANGF (Alexander et al., 1989). Furthermore, studies in vitro have located several regions within 1 kb adjacent to exon IB that affect basal transcription in L929 cells (D'Mello and Heinrich, 1991). These include a proximal activator region (up to -300 bp), shown by sequence analysis to contain several SPl-like elements that are known to increase the number of transcription-initiation complexes, and a transcriptional suppressor recognition site (at -250 to -500 bp). Further experiments with different cell types will determine whether the suppression is a unique or a more general phenomenon. In addition to the identification of the promoter-

Ill

like motifs in the 5' flanking sequences, an API binding site for the transcription factors, c-fos and c-jun, was located at the junction of exon IB and the second intron (Fig. 1; D'Mello and Heinrich, 1991). To test the regulatory function of this intronic API consensus sequence on NGF promoter activity, primary cultures of skin fibroblasts were cotransfected with a c-fos expressing vector and chloramphenicol-acetyltransferase reporter constructs containing various fragments of the NGF IB promoter. Only the constructs with a native API binding site exhibited c-fos inducibility (Hengerer et al., 1990). A similar relationship between c-fos induction and an increase in mRNA^^^ also has been reported to exist in vivo. Lesion of the sciatic nerve is shown to cause a rapid increase in mRNA^->"'' and mRNA^-^*^' that is followed subsequently by an increase in mRNA^^^ (Hengerer et al., 1990). Increased mRNA^cF in the hippocampus in response to kainic acid is preceded by an increase in mRNA^^^^ (Zafra et al., 1990; Ballarin et al., 1991). Thus, induction of the c-fos and c-jun genes and their subsequent interaction with the API site may represent a final common pathway in the activation of the NGF gene by a large variety of stimuli, such as phorbol 12-myristate 13-acetate cAMP, serum, interleukin-1 (IL-1), 12-0-tetradecanoyl phorbol13-acetate, brain seizure activity and membrane depolarization. Regulation of the NGF gene by steroids, such as vitamin D3 (Wion et al., 1992) and dexamethasone (Lindholm et al., 1990a), is likely to occur via pathways independent of the cfos and c-jun mechanism. Members of the steroid hormone receptor superfamily regulate by binding to specific DNA motifs. A 162 bp sequence, which confers glucocorticoid-responsiveness, has been identified in the NGF promoter region (Lindholm et al., 1990a). Since other molecules, such as retinoic acid, are known to regulate the NGF gene, it is likely that more regulatory motifs will be found. One intriguing possibility is an involvement of hox genes, known to be important in developmental control and active at the time of NGF gene induction (Allen et al., 1991). Table 1 summarizes the effects of the activation or inhibition of alternate second messenger sys-

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

178 TABLE 1

Putative intracellular messengers involved in NGF synthesis and secretion regulation Cell types

Protein kinase C

cAMP

Fibroblasts

^1

^1

tx42

->2

cGMP

lonomycin

tx34 Schwann cells

4. to 70%'*

Ts-s''

Hippocampal culture Astrocytes

-pe

->^

^4

_^1,7,8,9

T x 10^ C6 glioma cells Dissociated iris cells

1^10,11,13

•1^12

T by 50%^2

->, No change; T, increase; T, decrease. ^Yoshida and Gage (1992); ^Wion et al. (1990); ^Wion et al. (1992); '^Matsuoka et al. (1991); ^Yamamoto et al. (1993); ^Friedman et al. (1992); '^Spranger et al. (1990); ^Neveu et al. (1992a); ^Furukawa et al. (1993); l^Mocchetti et al. (1989); ^^Schwartz (1988); ^^Hellweg et al. (1988); ^^Toso et al. (1988).

terns on the expression of mRNA^^^. While some of the differences reported are Hkely to be due to the variation in culture conditions, such as the presence or absence of serum, others remain to be explained. 6. Site of NGF synthesis The potential for NGF synthesis has been identified biochemically in many tissues by the presence of mRNA^GF^ 5m- precise knowledge of the cell types involved has generally escaped detection. The low levels of mRNA^^^ in tissues of adult animals have prevented its cellular localization by histological methods, despite the essential nature of this information for the accurate description of the physiological role played by NGF. During development and following experimental manipulations, however, the higher concentrations of mRNA^^p have allowed preliminary in situ hybridization studies to reveal its presence in some tissues (Wilson et al., 1986; Bandtlow et al., 1987; Davies et al., 1987; Rush et al., 1989; Wheeler and Bothwell, 1992; Scarisbrick et al., 1993). The biochemical and histological evidence gained so far has indicated that mRNA^GF synthesis is not restricted to cells receiving a direct innervation, but also includes epithelial and mesenchymal cells of the skin and Schwann cells and fibroblasts associ-

ated with peripheral nerve. In effector tissues of sympathetic and sensory neurons, a peak in the concentration of NGF and mRNA^GF is present around the time of innervation (Davies et al., 1987; Clegg et al., 1989; Wyatt et al.. 1990; Ueyama et al., 1992; Wheeler and Bothwell, 1992; Scarisbrick et al., 1993; Zettler and Rush, 1993). These preliminary studies need to be supported by further histological evidence, so that the site of NGF synthesis can be determined in both normal and pathological conditions. The generation of antibodies recognizing the precursor NGF, but not the native molecule, provides an alternate method to address this issue (Dicou et al., 1986). Ideally, these analyses of mRNA^^^ or precursor should be coupled with analysis of the native NGF molecule to demonstrate that synthesis is complete. 7. Agents regulating NGF production 7.7. Cytokines Following peripheral nerve trauma, mRNA^^^ levels increase and remain elevated for many days. In culture, however, the initial rise in mRNA^^^, which appears due to cell trauma and the action of poorly described agents in serum, is not sustained. This simple fact led Thoenen and co-workers to

R.A. Rush et al.

uncover the role of macrophages in the long-term elevation of mRNA^^^ by Schwann cells and fibroblasts in a damaged peripheral nerve (Bandtlow et al., 1987; Heumann et al., 1987a,b, 1989; Lindholm et al., 1987). Macrophages are present also in developing nerve (StoU and Muller, 1986), and their presence correlates with an elevation of mRNA^^^ in the sciatic nerve of neonatal rats (Heumann et al., 1987b; but see Edwards, 1993). The involvement of macrophages in the regulation of NGF synthesis led to a search through the many factors secreted by these cells for the agents mediating the NGF response. Lindholm et al. (1987) identified IL-1 as a possible mediator of mRNA^^^ induction in cultured explants of sciatic nerve. Subsequent experimentation identified fibroblasts as the target of the secreted IL-1 (Lindholm et al., 1988). Fibroblasts, isolated from adult rat sciatic nerve, responded to IL-1 challenge with a 15-fold increase in mRNA^^^, which was shown to result predominantly from mRNA stabilization rather than an enhanced transcription rate. Interestingly, Schwann cells that represent about 90% of the cell population of the mature sciatic nerve remained totally unresponsive to IL-1 stimulation (Matsuoka et al., 1991), despite the up-regulation following nerve damage (Bandtlow et al., 1987). What is responsible for the increased mRNA^^^ within Schwann cells of peripheral nerve during development and following trauma is yet to be established. However, it is most likely to be due to a macrophage-derived agent acting via a forskolinsensitive, cAMP-mediated, mechanism (Matsuoka etal., 1991). A number of other cytokines have been shown to regulate mRNA^^^ synthesis, and a summary of the effects reported for cultured cells is presented in Table 2. Cooperative regulation of NGF synthesis and secretion by cytokines has been observed in cultured astrocytes and fibroblasts (Yoshida and Gage, 1992). 7.2. Steroids Glucocorticoid hormones stimulate (Yamamoto, 1985) or inhibit (Akerblom et al., 1988) the expression of many genes with diverse physiological

179

consequences. As described in Section 5.2, the NGF gene has a glucocorticoid response element in the promoter region, so it is not surprising that the glucocorticoids produce marked changes in the synthesis of NGF. This has been studied best in vitro, where addition of glucocorticoids to the medium leads to a rapid decrease in mRNA^^^ levels in both non-neuronal cells and intact organs (Wion et al., 1986; Siminoski et al., 1987; Hellweg et al, 1988; Houlgatte et al., 1989: Lindholm et al., 1990a; Neveu et al., 1991). The decline in mRNA^^p is accompanied by a decreased secretion of NGF into the medium (Hellweg et al., 1988; Houlgatte et al., 1989) . This powerful suppressive action of glucocorticoids is further demonstrated by the action of dexamethasone in blocking the induction of RNA^^^" by IL-1 in cultured fibroblasts and mixed hippocampal cultures (Lindholm et al., 1988, 1990a; Friedman et al., 1990). The action of glucocorticoid on NGF expression has also been examined in vivo. Although dexamethasone does not affect the low basal levels of mRNA^^^ in the intact sciatic nerve, it is effective in blocking completely the lesion-induced increase in mRNANGF (Lindholm et al., 1990a). Furthermore, mRNA'^^^ levels in lesioned sciatic nerve are higher in adrenalectomized rats than in control animals (Lindholm et al., 1990a). In contrast, within the brain, steroids increase mRNA^^^ concentrations (Barbany and Persson, 1992; Lindholm et al., 1992a,b), and adrenalectomized animals display a reduced concentration of NGF and mRNA^^f^ in cerebral cortex and hippocampus (Aloe, 1989; Barbany and Persson, 1992). There is some evidence that the induction of the NGF gene by glucocorticoid hormones within the CNS is confined to neuronal populations, while production of NGF by glia is suppressed (Lindholm et al., 1992b). There is a sex difference in the concentration of NGF in tissues of adult mice, males having higher levels in the submandibular gland, serum, brain, spinal cord and adrenal glands (Pantazis and Jenson, 1985; Aloe et al., 1986; Katoh-Semba et al., 1989). A reduction of NGF in the brain, spinal cord and submandibular gland results from castra-

180

The regulation of nerve growth factor synthesis and delivery to peripheral

neurons

TABLE 2 Effects of cytokines and growth factors on NGF or mRNA^^^ in various tissues and cells Cytokines or growth factors

Tissue/cell types

Effect on NGF or mRNA

IL-1

Schwann cells Fibroblasts (of adult sciatic nerve)

No change in mRNA^GFl T X 30mRNNGF2 T X ISmRNA^GFl

Fibroblasts (prenatal skin) Sciatic nerve explant Astrocytes

t X 2.5 NGF^ T X 14 mRNANGF4 T X 5 mRNANGF5 T X 3 NGF^ T 60% NGF'7

IL-2 IL-3 IL-4 IL-5 IL-6 TGF-a

EGF

aFGF

bFGF

Mixed hippocampal cell culture

T X 3 NGF^ T X 6 NGF^ Tx3mRNANGF9

Cortical neurons Hippocampus in vivo Astrocytes Astrocytes Astrocytes Astrocytes Astrocytes Fibroblasts Astrocytes Schwann cells Fibroblasts Astrocytes

No change in NGF^^ T X 5 mRNANGF5 No change in NGF^'^^ or mRNA^GFH No change in NGF or mRNA^'^ ^ T x 3-fold in NGF^i T x 4-fold in NGF^i i x 2 0 % NGF^ No change in NGF or mRNANGF8,ll T X 3NGF3 T X 7 mRNA^GFS i 8 0 % mRNA^GFl T X 2 NGF^ Tx50mRNANGF12

Hippocampus in vivo Schwann cells Fibroblasts Astrocytes

Fibroblasts Sciatic nerve explants Astrocytes Schwann cells Fibroblasts Sciatic nerve explants Astrocytes

IGF PDGF

Hippocampus in vivo Schwann cells Astrocytes Schwann cells Fibroblasts Sciatic nerve explants Astrocytes

T X 2 NGF^ Tx3mRNANGFl2 No change in mRNA^GFi TX1.5NGF3

T X 7 mRNANGF5 i 40% NGF^ T 50% NGF^ T X 4 NGF^'S T X 2 mRNANGF4 t X 3 NGF3'8

No change in mRNA^FGi T X 3 NGF^ T X 2 NGF^ T X 2 mRNANGF4 T 13% NGF^ T X 2.5 NGF^ T X 7 mRNANGF5 T X 2.5 mRNA^GFS No change in mRNA^GFl ix27%NGFll No change in mRNA^GFl T X 3 NGF3

T X 2 mRNANGF4 i 20% NGF^

181

R.A. Rush et al. TABLE 2 (continued)

Cytokines or growth factors

Tissue/cell types

Effect on NGF or mRNA

TNF

Fibroblasts Sciatic nerve explants Astrocytes

T X 3 NGF^ T X 2 mRNANGF4 No change in NGF^''^ T X 3 NGF^

Abbreviation: TGF, transforming growth factor; EGF, epidermal growth factor; aPGF, acidic fibroblast growth factor; bFGF, basic fibroblast growth factor; IGF, insuHn-like growth factor; PDGF platelet-derived growth factor; TNF, tumour necrosis factor. ^Matsuoka et al. (1991); ^Lindlholm et al. (1988); ^Yoshida and Gage (1992); '^Lindholm et al. (1987); ^Spranger et al. (1990); ^CarmanDrzan et al. (1991); '^Gadient et al. (1990); ^Yoshida and Gage (1991); ^Friedman et al. (1990); ^^Vige et al. (1992); ^^Awatsufi et al. (1993); l^Lindholm et al. (1990b).

tion, but this can be reversed by continuous infusion of testosterone (Katoh-Semba et al., 1990). Infusion of testosterone into adult female rats only minimally increases levels of NGF in the brain and spinal cord, whereas the same treatment, when preceded by a single dose of testosterone at 5 days of age, causes a significant increase in NGF levels (Katoh-Semba et al., 1990). These results suggest that testosterone may influence endogenous NGF concentrations in the brain and spinal cord, but further experiments are required to uncover any direct effects. A sex difference in NGF levels in the mouse superior cervical ganglia and heart also has been detected (Korsching and Thoenen, 1988), but appears due to contamination from higher circulatory levels of NGF in the male as a result of the extraordinary levels in submandibular gland. 7.3. Vitamin DT, The steroid derivative 1,25-dihydroxyvitamin D (vit D3), a metabolically active form of vitamin D, was shown to act as an inducer of the NGF gene in murine fibroblasts cultured in a serum-free medium (Wion et al., 1991). This effect could be detected as early as 3 h after the addition of vit D3 and resulted in an increase in both cellular mRNA^*^^ and secreted NGF protein. Further experiments in the same culture system (Jehan et al., 1991) showed that a similar effect could be demonstrated using the vit D3 analogue, MC903. As this analogue is 100 000 times less potent on cal-

cium metabolism, it has potential therapeutic value. The molecular mechanism by which vit D3 enhances the pool of mRNA^^^ is likely to be mediated through specific receptors and the AP-I region of the NGF gene, similar to that shown for the osteocalcin gene (Schule et al., 1990). Interestingly, these receptors have been described in spinal cord, sensory ganglia and brain, but their possible role in NGF regulation has not been investigated (Stumpf et al., 1982, 1988; Stumpf and O'Brien, 1987). 7.4. Thyroid hormone Nuclear run-on assays have demonstrated directly a significant transcriptional effect of thyroid hormone on the NGF gene (Black et al., 1992). Thyroid hormone administration to immature mice significantly increases mRNA^^^ levels within the submandibular gland (Black et al., 1992). That thyroid hormone might have an important role in the regulation of neuronal function has been indicated by the demonstration of an interaction with NGF in the ontogeny of several neuronal structures in the CNS (Clos and Legrand, 1990; Legrand and Clos, 1991). In addition, nuclear thyroid hormone receptors are present within rat DRG neurons from the last week of embryonic life, but, in contrast, peripheral glia within the DRG and peripheral nerve display thyroid hormone receptor immunoreactivity for a few weeks only from the end of embryonic life. Expression of these receptors in Schwann cells is reactivated following ax-

182

The regulation of nerve growth factor synthesis and delivery to peripheral

neurons

onal degeneration and suppressed by axonal regeneration (Barakat-Walter et al., 1993). These preliminary studies suggest that the investigation of possible actions of thyroid hormones on NGF expression may prove a fruitful area of future research.

in vivo may result, in part, from the exposure of cells to serum components following tissue damage (Shelton and Reichardt, 1986b; Heumann et al., 1987a,b).

7.5. Serum

Primary cultures of cortical neurons and astrocytes synthesize and secrete NGF, but their coculture leads to a reduced NGF level in the culture medium, primarily by suppression of glial secretion (Vige et al., 1992). This inhibition can be reversed by decreasing the number of neurons in the culture. Furthermore, IL-1 induction of astrocytic NGF synthesis (Carman-Krzan et al., 1991; Carman-Krzan and Wise, 1993) is also inhibited by the presence of neurons. The agent(s) responsible for this inhibition could not be found in medium conditioned by neurons nor in a crude membrane or soluble fraction prepared from the cultured neurons. Thus, the authors hypothesized that direct glial-neuronal contact may be required for the inhibition to occur (Vige et al., 1992). The importance of cell-to-cell contact was also indicated by experiments showing that the concentration of mRNA^^P present during active glial growth is 8-fold higher than that from high-density cultures (Lu et al., 1991). The inhibition of NGF synthesis in a low-density glial culture could be mimicked by addition of glial membranes to the culture. The importance of this phenomenon for the regulation of NGF synthesis in vivo has not been investigated yet.

Serum is a powerful stimulator of NGF synthesis in both cell and organ cultures. For example, the concentration of NGF in the iris and peripheral nerve increases rapidly and by many-fold when placed into culture (Ebendal et al., 1980; Barth et al., 1984; Rush, 1984), and this has been correlated with a concomitant rise in mRNA^^^ (Heumann and Thoenen, 1986; Shelton and Reichardt, 1986b; Lindholm et al., 1990a). The response is due partially to factors present in serum (Lindholm et al., 1990a). The amount of NGF secreted by cultured irides also was increased several-fold in serumcontaining medium compared with serum-free medium (Houlgatte et al., 1989). Both large and small molecular weight components of serum are involved in this response, since the amount of NGF secreted into the medium from cultured sciatic nerve is reduced only by 40-70% when a dialysate of serum is used in place of whole serum (Ferguson et al., 1989). In cultured fibroblasts, mRNA^^^ concentration increases 4-fold under the influence of serum (Wion et al., 1985; Houlgatte et al., 1986, 1989), and preliminary studies have provided some hmited description of the physical characteristics of the active agent(s) (Houlgatte et al., 1989). Augmented expression of mRNA^^^ by cultured astrocytes and enhanced release of NGF protein into culture medium under the influence of foetal calf serum was reported by Furukawa et al. (1987) and Sprangeretal. (1990). The mechanism of this induction is likely to involve the interaction of serum factors with membrane receptors, the recruitment of signal transducing G-proteins and the induction and posttranslational modification of c-fos proteins (Wion et al., 1992). The induction of mRNA^GF in tissues

7.6. Co-culture

7.7. Depolarization High concentration of extracellular potassium (50 mM), which results in cellular depolarization, decreases NGF production by 30% in explanted rat iris (Hellweg et al, 1988). In contrast, elevated potassium levels increase the concentration of mRNA^^^ in cultured hippocampal neurons (Zafra et al., 1990). Depolarization produced by electrolytic lesions or systemic intraventricular administration of glutamate receptor agonists produces a rapid rise in mRNA^^^ in vivo (Gall and Isackson, 1989; Zafra et al., 1990; Gall et al., 1991).

RA. Rush et al.

7.8. Other 7.8.1. Methyl mercury Pre- and postnatal exposure of rats to this heavy metal results in a 50% increase in NGF concentration in the hippocampus of young rats, but a decrease of 30% in the septum (Larkfors et al., 1991). These authors speculated that these changes reflect a heavy metal-mediated interruption of the retrograde transport of NGF from the hippocampus to the basal forebrain. However, other changes, such as an up-regulation of NGF synthesis in the hippocampus as a response to heavy metal-induced damage, cannot be excluded. 7.8.2. Propentofyllin The xanthine derivative, propentofyllin, upregulates NGF synthesis 10-fold in cultured astrocytes (Shinoda et al., 1990). This was of particular interest, as the drug reverses learning and memory impairment in aged rats (Goto et al., 1987) and protects against delayed, ischaemically induced neuronal death in the hippocampus. Shinoda and co-workers (Shinoda et al., 1990) therefore speculated that the beneficial effects of propentofyllin, in part, could be due to an induction of NGF synthesis in astrocytes. 8. Regulation of NGF expression To detail the regulation of NGF synthesis, determination of both NGF and mRNA'^^^ levels is ideally required. However, due to the technical demands of both assays, few studies have provided such information. Thus, the following summary is based primarily on studies that have examined either the alterations to NGF or mRNA'^^*' concentrations. In the future, an additional level of sophistication will be required to describe the availability of newly synthesized NGF to sensitive neurons. 8.1. Central nervous system Most early studies examining the role of NGF were restricted to analysis within the peripheral nervous system. During the past decade, however.

183

much more attention has been given to the role of NGF in the CNS. This work will not be reviewed extensively here, but is discussed briefly to highlight those experiments that are relevant to peripheral regulation. It should be emphasized that while there are many similarities in the actions of NGF within the peripheral and central nervous systems, much of the evidence favours the view that the regulation of NGF synthesis in the two systems is quite distinct. The reasons for this conclusion have been reviewed recently (Zafra et al., 1990; Thoenenetal., 1991). That NGF-sensitive neurons exist within the CNS can be summarized by the following observations. Exogenous NGF is selectively acquired and transported by cholinergic forebrain neurons with terminals in the hippocampus (Schwab et al., 1979), cortex (Seller and Schwab, 1984) and olfactory bulb (Altar and Bakhit, 1991); findings consistent with the regional distribution of endogenous protein (Korsching et al., 1986; Shelton and Reichardt, 1986a; Conner et al., 1992). NGF increases the content of choline acetyltransferase in cultured forebrain neurons from embryonic and newborn rats (Honegger and Lenoir, 1982; Gnahn et al., 1983; Hefti et al., 1985). Lesion-induced changes in cholinergic neurons can be offset by NGF supplementation, which, in addition to reestablishing neurotransmitter synthesis, also promotes cholinergic fibre reinnervation (Hefti, 1986; WiUiams et al., 1986; Kromer, 1987; Gage et al., 1988; Hagg et al., 1988, 1989, 1990; Ernfors et al., 1989). Finally, developmental changes in hippocampal choline acetyltransferase levels mirror those of endogenous NGF (Auburger et al., 1987). The ability to quantify mRNA^^^ biochemically and semiquantitatively, by histochemical techniques, has led to a number of studies examining the factors involved in their regulation within the brain. Most workers have concentrated on the hippocampus, as this area contains the highest concentration of each mRNA^^^, overcoming some of the technical difficulties associated with the detection and quantification of the very low levels seen in other regions. Initial experiments have used procedures, such as induction of seizures, that alter electrical activity, since it is well known that ab-

184

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

normal electrical activity is associated with nerve fibre sprouting (Sutula et al., 1988, 1989; Babb, 1991). Seizures elicit a large and rapid increase in levels of mRNA^^P across much of the forebrain. This was shown in studies using focal electrolytic lesions of the dentate gyrus, which induced 10 h of intermittent limbic seizures (Gall and Isackson, 1989). Within 4 h of the onset of the seizures, mRNA^^p was increased 10-fold. Induction of mRNA^^p in the hippocampus also follows seizures induced by kainic acid, with the increased mRNA^^p restricted to the granule cells (Gall et al., 1991; Zafra et al., 1991). Induction of mRNA^^^ following limbic seizures is not restricted to the hippocampus, but is widespread across the neocortex, lateral and cortical amygdaloid nuclei and olfactory telencephalon (Gall and Isackson, 1989). However, the difference in the time course of the induction suggests a non-uniform mechanism in the various neuronal populations. For example, seizure activity triggered by hilus lesions radiates to the neocortex. In contrast to the hippocampal induction that is seen 4.5 h after the first seizure, neocortical mRNA^^^ remains normal during the most intense seizure activity, rising only hours after the seizures have terminated (Gall et al., 1991). These authors speculated that the hippocampal response represents transcriptional changes induced by neurotransmitter activity, whereas the late cortical increase may be mediated by trophic interactions. While initial experiments focussed on the effects of recurrent seizures, it is now evident that even brief bursts of activity are sufficient to cause large changes in NGF (and other NT) expression. A single electrical stimulation lasting 25-40 s produced a marked elevation of mRNA^^P as early as 30 min after the stimulation (Ernfors et al., 1991). Adrenalectomy, which decreases basal cortical mRNA^^^ concentration (Barbany and Persson, 1992), has no influence on the expression induced by limbic seizures (see Gall and Lauterborn, 1992). The mechanism involved in mRNA'^^^ induction has been investigated also for kainic acidinduced limbic seizures (Zafra et al., 1990). These

workers demonstrated that treatment of cultured rat hippocampal neurons with kainic acid results in an 8-fold rise of mRNA^^^. Kynurenic acid, a broadspectrum glutamate receptor antagonist, and CNQX, a competitive inhibitor of non-A^-methylD-aspartate (NMDA) receptors, both completely blocked the mRNA^^^ induction. In contrast, MK801, a specific blocker of NMDA receptors, was ineffective. Moreover, NMDA itself did not alter NT message levels. Thus, it can be concluded that kainic acid (in vitro) acts directly via its own, non-NMDA receptors to raise mRNA^^^ levels. To evaluate the pathophysiological significance of the above results, rats were treated with kainic acid, resulting in an increase of mRNA^^^ in the hippocampus (Zafra et al., 1990). This induction also was found to occur following electrical stimulation of the angular bundle supplying the efferent innervation to the dentate gyrus (Gwag et al., 1993). There was a corresponding rise in mRNA^^p levels in the cortex 2 h after the hippocampal induction, suggesting an indirect pathway of activation for this brain region (Zafra et al., 1990). In accordance with the in vitro findings, the NMDA antagonists MK801 and ketamine did not prevent the hippocampal induction, despite the suppression of seizures. Benzodiazepine, a glutamate receptor agonist, completely blocked the mRNA^^p induction, indicating that the inhibitory y-amino butyric acid inputs probably act in opposition to the excitatory glutamate transmitter. This clear regulation of mRNA^^^ by neurotransmitters in the CNS is in contrast to the situation in the periphery. That mRNA^^P induction is independent of seizure activity has been shown by direct intraventricular infusion of NMDA, which results in an increased expression exclusively in the dentate gyrus granule cells (Gwag et al., 1993). Ischaemic and hypoglycaemic insults, which do not induce seizures, also raise mRNA^^^ levels in the hippocampus (Lindvall et al., 1992). Physiologically relevant stimuli have been used to examine the regulation of mRNA^^^ in the brain. Zafra et al. (1991) used glutamate receptor antagonists and y-amino butyric acid agonists to determine the effects of reduced hippocampal mRNA^^^ levels on the NGF protein levels in the

R.A. Rush et al.

neurons projecting into this area. NGF levels were reduced in the septum, suggesting a diminished availability of the factor to the dependent neurons. Thus, the functional state of the hippocampal neurons regulates the concentration of mRNA^^^ present and, therefore, the availability of NGF to dependent neurons (see review by Lindholm et al., 1994). 8.2. Peripheral nervous system 8.2.1. In development The function of the vertebrate nervous system depends critically on the intricate network of neuronal connections generated and refined during development. This patterning emerges from precise and coordinated interactions between developing neurons and their cellular environment and culminates in the widespread loss of redundant neurons during a period of naturally occurring cell death (Hamburger and Oppenheim, 1982; Cowan et al., 1984; Oppenheim, 1991). Of particular importance in this rationalization process are the NTs, whose successful acquisition by developing neurons is essential for neuronal survival (LeviMontalcini, 1987; Thoenen et al., 1987; Hofer and Barde, 1988; Barde, 1989). Studies examining tissue NGF and mRNA'^^f' concentrations during development indicate that NGF synthesis is under powerful regulatory influences. For example, within effector tissues, the concentration of NGF and its mRNA is highest during a short developmental period correlating with the time of innervation (Davies et al., 1987; Ebendal and Persson, 1988; Korsching and Thoenen, 1988; Clegg et al., 1989; Falckh et al., 1992a,b; Ueyama et al., 1992; Zettler and Rush, 1993). High serum levels of NGF in the neonate suggest that an overflow of NGF escapes capture by the immature nerves (Murase et al., 1992). During this critical period, large changes in concentrations of both NGF and mRNA^^^ occur over a few days, and NGF begins to accumulate in the nerve cell bodies, following retrograde transport away from the effector tissues (Davies et al., 1987; Korsching and Thoenen, 1988). The initiation of NGF synthesis has been exam-

185

ined in detail within the skin during the period of sensory innervation. NGF synthesis begins just prior to the arrival of the first sensory fibres (Davies et al., 1987), indicating the process is independent of the innervation. This conclusion is supported by neural crest ablation experiments demonstrating that the absence of nerve fibres does not prevent the appearance of mRNA^^^ (Rohrer et al., 1988). Shortly after the arrival of nerve fibres into the developing skin, there is a rapid reduction in the level of NGF and a lesser, but significant, decrease in mRNA'^^f' concentration (Davies et al., 1987). This decline in NGF synthetic capacity might indicate a down-regulation by the maturing peripheral innervation. However, the decrease in mRNA^^^ concentration occurs even in the absence of a sensory innervation (Rohrer et al., 1988). Thus, this finding argues against a direct feedback control of NGF synthesis by the innervation and begs the question of what mechanisms are involved in this rapid rise and fall of NGF synthesis. Few in situ hybridization studies have been performed to complement the biochemical studies, since the very low levels of mRNA^^^ make its detection difficult. Nevertheless, mRNA^^^expressing cells have been found in regions innervated by NGF-responsive neurons (Wilson et al., 1986; Davies et al., 1987; Rush et al, 1989; Wheeler and Both well, 1992; Scarisbrick et al., 1993), but also within tissues apparently devoid of sensitive innervation. Thus, the smooth muscle cells of the thoracic aorta in the neonate (Scarisbrick et al., 1993) and skeletal muscle fibres in the embryo (Wilson, Vahaviolos and Rush, unpublished observations) both produce high levels of mRNA^^p. Furthermore, Schwann cells and fibroblasts within developing peripheral nerve (Bandtlow et al, 1987) and fibroblasts within skeletal muscle (Wilson, Vahaviolos and Rush, unpublished observations) also are active in NGF synthesis. None of these cells receive a direct innervation by NGF-sensitive neurons, indicating that the signal for synaptogenesis is not NGF alone. Interestingly, even in those effector tissues innervated by NGF-sensitive neurons, such as the skin and vasculature, many cells (epithelial and

186

The regulation of nerve growth factor synthesis and delivery to peripheral

smooth muscle fibres, respectively) receive no direct neuronal contact. Despite these limitations, the findings, taken together, provide strong evidence that the differing spatial and temporal patterns of NGF expression within peripheral effector tissues exert a key role in the establishment and maintenance of mature neuronal populations in vivo. Moreover, the ability of NGF supplementation to dramatically increase the size of responsive neural centres, and the extent by which they innervate effector tissues, argues strongly that individual effectors regulate the type and density of innervating neurons through the synthesis and release of limiting quantities of NGF. 8.2.2. In maturity In the mature animal, serum levels of NGF are very low (Murase et al., 1992), as might be predicted by its postulated action as a paracrine hormone (see Thoenen and Barde, 1980; Bradshaw, 1983; Levi-Montalcini, 1987). Although the ability of anti-NGF treatment to destroy sensitive neurons decreases with age (Goedert et al., 1978; Johnson et al., 1982; Schwartz et al., 1982) and cultured neurons show a decreasing dependence on NGF with age (Chun and Patterson, 1977; Greene, 1977a,b; Davies et al., 1986a,b; Lindsay, 1988; Snider and Johnson, 1989), the availability of NGF from denervated tissue or by administration of exogenous factor results in sprouting of nerve terminals, thus indicating the sensitivity of mature neurons to altered local concentrations (LeviMontalcini, 1983; Diamond et al., 1987; Zettler et al., 1991). Furthermore, mature neurons respond to disconnection from their effector tissues, via nerve transection or blockade of axonal transport, with reductions in neurotransmitter synthesis, atrophy and significant degrees of cell death (Wooten et al., 1977; Hefti, 1986; Williams et al., 1986; Kromer, 1987; Laiwand et al., 1987; Yawo, 1987; Villegas-Perez et al., 1988; Snider and Thanedar, 1989) suggesting that trophic factors continue to exert important controls on their cellular metabolism (Hendry et al, 1974; Schwab et al, 1979; Thoenen and Barde, 1980; Johnson et al, 1982; Seller and Schwab, 1984; Altar and Bakhit, 1991).

neurons

Indeed, chronic NGF deprivation in adult animals results in a significant destruction of responsive sympathetic neurons (Gorin and Johnson, 1979; Johnson et al., 1982). NGF and its mRNA are present in numerous peripheral tissue extracts innervated by sympathetic fibres, e.g. vas deferens, submandibular gland, heart, iris, spleen and blood vessels (Heumann et al., 1984; Shelton and Reichardt, 1984; Bandtlow et al., 1987; Zettler and Rush, 1993). Evidence has been presented for a correlation between the density of mature sympathetic innervation and the concentration of both NGF and mRNA^^^, although this has not been established for more than a few tissues and does not take into account the presence of sensory neurons (Korsching and Thoenene, 1983; Shelton and Reichardt, 1984; Harper and Davies, 1990). Nevertheless, this finding alone implies a close reciprocal regulation of NGF synthesis and neuronal metabolism by nerve and NGF, respectively. However, no clear evidence for an influence of nerve on NGF synthesis in peripheral effector tissues has been found (see Section 9). 9. Peripheral denervation and NGF synthesis 9.1. Pharmacological denervation Several studies have demonstrated that chemical sympathectomy increases the concentration of NGF 2 to 5-fold in effector tissues such as the submandibular iris and heart (Ebendal et al., 1980; Korsching and Thoenen, 1985; Whittemore et al., 1987). These changes however, are independent of mRNANGF concentrations (Clegg et al., 1989), with the exception of a small, but significant, decrease in the iris (Shelton and Reichardt, 1986b). It has been concluded, therefore, that the increased NGF concentration is due solely to the absence of a retrograde transport mechanism that normally removes the factor from the tissue following its secretion (Korsching and Thoenen, 1985). However, since most peripheral tissues receive both sympathetic and seinsory innervation, the remaining sensory nerves may continue to influence the

R.A. Rush et al.

production of NGF in the effector tissue. NGF concentrations have been shown to increase in sensory ganglia following loss of sympathetic nerves, suggesting that the remaining nerve fibres have access to these greater NGF concentrations (Korsching and Thoenen, 1985). Thus, the findings from surgical denervation should provide a better indication of neural control of NGF production. 9.2. Surgical denervation The effect of surgical interruption of the nerve supply has been examined in several peripheral effector tissues, and the results are summarized below. 9.2.7. Iris The iris receives a dense sympathetic, sensory and parasympathetic innervation. This tissue has been used extensively in the study of NGF. The work of Ebendal et al. (1980), Rush (1984) and Barth et al. (1984) has shown that the concentration of NGF within the iris is markedly increased by culture. Later, it was shown that culture induces an increase in mRNA^^^ (Heumann and Thoenen, 1986; Shelton and Reichardt, 1986b; Wilson et al., 1986; Bandtlow et al., 1987). These studies provided the impetus for an examination of the effects of denervation in situ. Progressive denervation of the iris followed by estimation of the mRNA^^^ concentrations was performed by Shelton and Reichardt (1986b), who drew several conclusions. Firstly sympathetic denervation alone does not cause the elevation of mRNA^^^ content of the denervated iris. Secondly, sympathetic and sensory denervation together results in raised mRNA^^^ levels. However, there was no significant difference in mRNA^^^ content between the ipsilateral and contralateral irides, suggesting that the trauma associated with this procedure, rather than the loss of nerve terminals, was responsible for the induction. Finally, isolation and culture of the iris increased the mRNA^^^ concentration by as much as 6-20-fold (Heumann and Thoenen, 1986; Shelton and Reichardt, 1986b). However, as discussed in Section 7.5, it is likely that serum com-

187

ponents in the culture medium, rather than isolation of the tissue per se, were responsible for much of this induction. To achieve a similar degree of denervation in situ, a retrobulbar section of the optic nerve and the long and short ciliary nerves was performed. This total eyeball denervation resulted in a significant increase in the ipsilateral iris mRNA^^^ production. Unavoidably, the procedure also led to significant trauma to the eye, and since mRNA^^^ synthesis could be induced by various insults to the anterior eye chamber without denervation, it is not possible to conclude whether the rise in mRNA^^^ was due to the denervation or the trauma. Nevertheless, it is interesting that trauma itself leads to increased mRNA^^^ and the mechanism involved deserves further investigation. 9.2.2, Ovary Immature rat ovary has been shown to express mRNA^^^ and to contain NGF protein levels comparable to those found in other sympathetically innervated tissues (Lara et al., 1990). Surgical denervation of this organ resulted in a 50% increase in NGF protein levels, but unaltered mRNA^^p levels. However, only limited measurements of mRNA^^^ levels were performed in this preliminary study. 9.2.3. Skeletal muscle NT production in skeletal muscle is sensitive to denervation. mRNA^^^ appears early in muscle development, decreases in concentration in maturity, but increases several-fold following nerve lesion (Hulst and Bennett, 1986). Although the methodology used in this early study may have led to the inclusion of NGF homologues in the analysis, we have been able to verify by in situ hybridization the existence of mRNA^^^ in developing skeletal muscle fibres and fibroblasts (Wilson, Vahaviolos and Rush, unpublished observations). Up-regulation of mRNA^^^^ and down-regulation of mRNA^^^^ also result from denervation (KoHatsos et al, 1993). This effect of denervation on NT mRNA levels in skeletal muscle in the absence of similar regulation for other tissues is intriguing.

188

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

9.2.4. Bloodvessels Recently, our laboratory has utilized an improved extraction procedure for NGF (Zettler et al., 1995) to examine the effect of surgical denervation on its concentration within the mesenteric artery. The high levels of NGF present were reduced by denervation to approximately 20% of control levels. By assaying NGF within the adventitia and media separately, it was possible to demonstrate that the loss of NGF was associated only with the loss of nerve fibres within the adventitia. Furthermore, the discrepancy with the studies described above could be explained by the ability of the new extraction procedure to free large amounts of NGF bound to tissue components (such as trkA receptors) prior to assay (Liu, Reid, Bridges and Rush, unpublished observations). 9.3. Nerve lesion Sciatic nerve transection results in an increased mRNA^^^ concentration and elevated NGF synthesis distal to the lesion (Abrahamson et al., 1987; Heumann et al., 1987a,b). This production is biphasic, with the initial peak of mRNA^^^ occurring at 6 h in segments both distal and proximal to the lesion, and appears to be a result of damage to the cells in the immediate vicinity in response to agents from blood or released from the damaged axons (Heumann et al., 1987a,b). Both fibroblasts and Schwann cells have been shown to produce mRNA^^^ at the site of injury (Bandtlow et al., 1987) and to up-regulate NGF synthesis in response to noradrenaline (Furukawa et al., 1986a). The second peak of mRNA^^^, which is not apparent until day 3, occurs predominantly distal to the lesion and probably is due to the release of IL-1 and other agents by infiltrating macrophages (see Heumann et al., 1989). This increased mRNA^^p synthesis persists until the nerve has regenerated (Heumann et al., 1987a,b). Since mRNA'^^p induction was detected at all sites distal to the trauma, it might be predicted that the fibroblasts and Schwann cells accompanying the nerve within its effector tissue would also respond with increased expression. Thus, the reports of unchanged mRNA^^^ in surgically denervated

sympathetic and sensory effector tissues are puzzling. 10. Do neurotransmitters regulate NGF synthesis in peripheral effector tissues? Since the available evidence favours a role for neurotransmitters in the regulation of mRNA^^^ synthesis within the CNS, their role in the regulation of peripheral NGF deserves particular consideration. To date, the evidence argues against the involvement of the nerve supply and, therefore, either neurotransmitters or specific trophic agents released from nerve terminals. However, this lack of effect appears to indicate that no homeostatic, feedback control mechanism operates to regulate NGF synthesis. This issue is discussed in the following sections. It is important to note, however, that recent evidence has demonstrated the importance of motor innervation for the regulation of mRNA'^'^ levels within sketelal muscle (Funakoshi et al., 1995). 10.1. In vitro A number of studies have investigated the effects of peripheral neurotransmitters and peptides on the regulation of NGF production and secretion by isolated organs or cells. For example, Hellweg et al. (1988) examined the influence of many agents on the release of NGF into the medium of cultured irides. Addition of neuropeptide Y, substance P, vasopressin, somatostatin, vasoactive intestinal polypeptide, neurotensin, serotonin and histamine all failed to change NGF levels. However, as these experiments were carried out in serum that may have resulted in maximal stimulation of the NGF gene, only inhibitory effects may have been detectable. Indeed, noradrenaline and dopamine were shown to be inhibitory. The effect of noradrenaline on NGF release was correlated with a reduction in mRNA^^^ levels within smooth muscle fibres and fibroblasts from the dissociated iris. Although the results of this experiment support the idea of an inhibitory regulation of NGF synthesis by sympathetic neurotransmitters, they involved the use of high concen-

RA. Rush et al.

trations of catecholamines and have not been confirmed by other laboratories. For example, noradrenaline at lower concentrations has been shown to be a potent stimulator (20-fold increase) of NGF production in fibroblastic (Furukawa et al., 1986a) and astrocytic cell lines (Furukawa et al., 1987, 1990; Schwartz, 1988; Mocchetti et al., 1989), as well as in cultured sciatic nerve segments (Ikegami et al., 1990). However, in low serum concentrations (0.5%), noradrenaline is without effect on mRNA^GF (Spranger et al, 1990). Schwann cells (Matsuoka et al., 1991) are also capable of increasing NGF production under the influence of catecholamines. The reported culture conditions and noradrenaline concentrations used in these experiments were dissimilar, and, thus, may account for the observed differences. These differences need to be reconciled. Despite these contrasting results, it is most unlikely that the effect of noradrenaline on mRNA^^p production is mediated through adrenergic receptors. This was concluded because the effects could neither be reproduced by cAMP analogues nor blocked by a- or ^-adrenoreceptor antagonists. Moreover, Furukawa and colleagues (Furukawa et al., 1986b, 1990) have postulated on the basis of analogue studies that the catechol ring is the essential moiety for the regulatory function. More recently, these workers have speculated on the existence of 'catechol' receptors to explain the catecholamine action and describe an unpublished observation of c-jun- and c-/(9^-mRNA induction by various catechol compounds (Furukawa et al., 1993). These findings appear to be in contrast to other studies, which described induction of mRNA^^^ by y3-adrenergic agonists, via induction of cAMP, in astrocytes and Schwann cell cultures (Schwartz, 1988; Mocchetti et al., 1989; Matsuoka etal., 1991). 70.2. In vivo Despite the many regulatory agents described to date, little is known of the control of NGF synthesis in vivo. Catecholamines have been shown to increase the concentration of NGF in vivo in two experimental paradigms. Firstly, 4-methylcatechol-

189

amine (a non-amine catechol compound) treatment can partially reverse the streptozotocin-induced loss of NGF within sciatic nerve (Hanaoka et al., 1992). Secondly, administration of 4-methylcatecholamine increases NGF levels in adult rat heart and submandibular gland (Kaechi et al., 1993). These results provide strong evidence for a role of catecholamines in regulating NGF concentrations in vivo, despite the lack of an obvious effect of denervation on mRNA^^P levels and the creation of a positive feedback loop (see Section 11). Elevated NGF concentrations have been seen in two conditions involving altered muscle morphology. Hypertrophy and hyperplasia of bladder smooth muscle cells induced by increased intravesicular pressure elevates NGF content 2-fold (Steers et al., 1991), while, within the mesenteric artery of the SHR, a rise in NGF concentration correlates with the onset of smooth muscle hyperplasia (Zettler and Rush, 1993). Since abnormalities also occur within the innervation of both these pathologies, it is unclear whether the altered NGF concentrations result from changes in muscle or nerve function. 11. Is there a negative feedback control of NGF synthesis? NGF is a potent regulator of sympathetic and sensory neuronal function. There is strong evidence for its role in controlling neuronal survival, the density of nerve terminal arborization, the concentration of neurotransmitters and neuropeptides, and a host of other metabolic functions. Some evidence also indicates an indirect control of sympathetic preganglionic function (Oppenheim et al., 1982). Such a powerful regulator of neuronal function, itself, must be under strict controls. Control at the gene level by mechanisms external to the neuroneffector tissue axis certainly has been shown, as outlined in Section 7, but functional control implies the existence of a negative feedback loop. Attempts to identify this feedback control have been confusing, since denervation appears to have little effect on NGF gene expression, whereas most studies examining the action of catecholamines have reported raised NGF concentrations. A

190

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

Stimulatory effect of catecholamines on NGF synthesis would lead to an unstable positive feedback mechanism, since higher NGF concentrations would drive catecholamine synthesis. An attempt to resolve this issue was made by Ueyama et al. (1991), who investigated whether NGF could directly affect noradrenaline release from peripheral sympathetic nerve endings of the rat mesenteric artery. In isolated intestinal loop preparations, NGF produced a dose-dependent inhibition of noradrenaline overflow evoked by electrical stimulation. Reduction in overflow also occurred in the presence of a neuronal uptake blocker. NGF did not affect the pressor response to exogenous noradrenaline. Therefore, NGF appears capable of acting as an inhibitory modulator of noradrenaline release from sympathetic nerve endings. With this notable exception, the available evidence argues in favour of a control of NGF synthesis that is independent of the nerve supply in the peripheral nervous system. How can this be? Where is the necessary negative feedback control in such a system? If regulation of NGF synthesis is not controlled directly via the innervation, indirect mediators may be responsible. For example, A II, which is known to be a potent stimulator of NGF synthesis within vascular smooth muscle and a modulator of noradrenaline release from nerve terminals, itself is regulated, in part, by the sympathetic innervation of the kidney. Alternatively, control of the availability of NGF may be affected by the innervation, and this is discussed in Section 12. There remain many issues that are poorly understood. For example, what determines the level of NGF synthesis in various peripheral tissues and why do Schwann cells and fibroblasts associated with nerve axons increase mRNA^^^ concentrations within peripheral nerve, but not apparently within effector tissues, in response to axon degeneration? Is the regulation of NGF synthesis in the mature animal the same as that controlling synthesis in development? Progress in NGF research has been plagued by the limitations of techniques necessary for quantification and localization of the low concentrations present in normal tissues. Many of the most important questions begging for

answers require further advances in methodology before their resolution is possible. 12. Is NGF secretion regulated? A crucial question regarding the regulation of NGF concerns its availability to sensitive nerve terminals, since its local concentration is thought to control the density of innervation directly (Korsching and Thoenen, 1983; Shelton and Reichardt, 1984; Donohue et al., 1989; Creedon and Tuttle, 1991; Steers et al., 1991; Zettler and Rush, 1993). While the developmental regulation of neuronal survival and the establishment of functional synaptic connections may be accounted for by a constitutive release of NGF (Ebendal et al., 1980; Thoenen and Barde, 1980; Edwards, 1993) following limited synthesis by neighbouring cells (see Edwards et al., 1989; Hoyle et al., 1993), such a mechanism should result in a homogeneous distribution of responsive nerve fibres throughout an effector tissue. This is certainly not the ease for peripheral and central resistance vessels in which smooth muscle cells, the putative source of NGF (Searisbrick et al., 1993), maintain a dense NGFsensitive sympathetic innervation only at their adventitial medial boundary (see, for example, Burnstoek, 1975; Lee et al., 1987, 1992; Zettler et al., 1991; Zettler and Rush, 1993). The uniform synthesis and constitutive release of NGF by vascular smooth muscle cells should result in its subsequent accumulation in the extracellular space and the attraction of an aberrant nerve supply penetrating deeply into the medial smooth muscle cell layer, as seen after exogenous NGF supplementation (Levi-Montalcini and Hamburger, 1951; Zettler et al., 1991; Lee et al., 1992). That this does not occur argues for an alternative mechanism of delivery. One possible alternative is the regulation of the availability of NGF via the control of its release. Edwards et al. (1988b) have investigated the modulation of NGF secretion in several mammalian cell lines after transfection with vaccinia virus expression vector rendered them competent to synthesize and secrete NGF. Transfection of L929 fibroblasts led to the expression of high levels of

191

R.A. Rush et al.

mRNA^^^, translation and rapid secretion of biologically active NGF. Comparison of extracellular and intracellular levels of NGF in this cell line revealed a ratio characteristic of proteins synthesized and released constitutively (see Moore et al., 1983). In contrast, transfection of cells with a regulated pathway of secretion, such as hamster insulinoma (HIT) cells and mouse anterior pituitary (AtT-20) cells, resulted in the intracellular storage of NGF (i.e. equal amounts of intracellular to extracellular NGF), which could be released after exposure to 8-bromo-cAMP, a general stimulator of secretory vesicle mobilization. Using cultured cortical neurons, Vige et al. (1992) observed a high extracellular to intracellular ratio of NGF (i.e. that of a cell with a regulated secretory pathway), similar to that of HIT and AtT-20 cells. Moreover, the ability to immunohistochemically label NGF in the cell bodies of cultured neurons and oligodendrocytes as opposed to cultured astrocytes supports the notion that many cell types also have the capacity to store NGF (Gonzalez et al., 1990). The concept that the control of the secretion of stored NGF is achieved by nerve terminals is supported by several studies. Thus, cultured vascular smooth muscle cells increase NGF secretion in response to a-adrenergic agonists, whereas ^adrenergic agonists inhibit secretion (Creedon and Tuttle, 1991). These authors also found a dissociation between altered mRNA levels and NGF secretion. A direct demonstration of a neurotransmitter effect on NGF release in vivo was provided by White and colleagues (White et al., 1987), who found that substance P, but not calcitonin gene-related peptide, could inhibit the release of locally synthesized NGF from a neuroma on the saphenous nerve. The unrestricted availability of NGF to sympathetic neurons, via the provision of a genetically engineered intraneuronal supply, leads to a significant reduction in the innervation density of peripheral effector tissues (Hoyle et al., 1993). Thus, the local regulation of NGF release via specific secretagogues may provide a mechanism by which functional synaptic connections are reinforced during development and altered during

functional plasticity during adaptation to injury and in disease. 13. Summary and conclusions We have examined in this review the factors known to be important for the regulation of NGF gene expression. Few studies have estimated simultaneously the concentration of NGF and mRNA^^^, so it is not yet possible to review the regulation of NGF synthesis with regard to independent transcriptional and translational components. That a few studies have demonstrated a dissociation of mRNA^^^ from NGF protein levels emphasizes the need for such combined analyses (Creedon and Tuttle, 1991; Hashimoto et al., 1992). Nevertheless, there is already sufficient information available to conclude that NGF synthesis is tightly controlled and that many agents are capable of increasing and decreasing this production. The lack of obvious regulation of mRNA^^^ by the nerves innervating each effector tissue is intriguing. If NGF synthesis is not regulated by the nerves it influences so profoundly, what mechanism exists to provide a feedback control for its action? We have argued that the supply of NGF to nerve terminals is likely to be influenced by these nerves. Since the available evidence favours the view that regulation appears not to be at the level of mRNA'^^^ expression, we have proposed that secretion, and, therefore, the availability of native NGF, might be under neuronal control. Clearly, there is still much to unravel in the physiology of NGF. Acknowledgements We wish to thank Kaia Palm, Andrew Rakowski and Jim Vahaviolos for valuable discussions and critical reading of the manuscript. Much of the work presented from our own laboratory was supported by the National Health and Medical Research Council of Australia (NH and MRC) and the National Heart Foundation of Australia (NHF). Raya Mayo is a recipient of an NH and MRC Postgraduate Medical Research Scholarship and Chris Zettler is supported by the NHF.

192

The regulation of nerve growth factor synthesis and delivery to peripheral

References Abrahamson, I.K., Ferguson, LA. and Rush, R.A. (1987) Endogenous chicken nerve growth factor from sheath cells is transported in regenerating nerve. Dev. Biol. 124: 551-556. Acheson, A., Barde, Y.A. and Thoenen, H. (1987) High K"^mediated survival of spinal sensory neurons depends on developmental stage. Exp. Cell Res. 170: 56-63. Akerblom, I.E., Slater, E.P., Beato, M., Baxter, J.D. and Mellon, P.L. (1988) Negative regulation by glucocorticoids through interference with a cAMP responsive enhancer. Science 241: 350-353. Alderson, R.F., Alterman, A.L., Barde, Y.A. and Lindsay, R.M. (1990) Brain-derived neurotrophic factor increases survival and differentiated functions of rat septal cholinergic neurons in culture. Neuron 5: 297-306. Alexander, J.M., Hsu, D., Penchuk, L. and Heinrich, G. (1989) Cell-specific and developmental regulation of a nerve growth factor-human growth hormone fusion gene in transgenic mice. Neuron 3: 133-139. Allen, J.D., Lints, T., Jenkins. N.A., Copeland, N.G., Strasser, A., Harvey, R.P. and Adams, J.M. (1991) Novel murine homeo box gene on chromosome 1 expressed in specific hematopoietic lineages and during embryogenesis. Genes Dev. 5: 509-520. Aloe, L. (1989) Adrenalectomy decreases nerve growth factor in young adult rat hippocampus. Proc. Natl. Acad. Sci. USA 86: 5636-5640. Aloe, L., Mugnaini, E. and Levi-Montalcini, R. (1975) Light and electron microscopic studies on the excessive growth of sympathetic ganglia in rats injected daily from birth with 6OHDA and NGF. Arch. Ital. Biol. 113: 326-353. Aloe, L., Alleva, E., Bohm, A. and Levi-Montalcini, R. (1986) Aggressive behaviour induces release of nerve growth factor from mouse salivary gland in the bloodstream. Proc. Natl. Acad. Sci. USA 83: 6184-6187. Aloe, L., Tuveri, M.A., Carcassi, U. and Levi-Montalcini, R. (1992) Nerve growth factor in the synovial fluid of patients with chronic arthritis. Arthritis Rheum. 35: 351-355. Altar, C.A. and Bakhit, C. (1991) Receptor-mediated transport of human recombinant nerve growth factor from olfactory bulb to forebrain cholinergic nuclei. Brain Res. 541: 8 2 88. Angeletti, R.H., Hermodson, M.A. and Bradshaw, R.A. (1973) Amino acid sequences of 2.5S nerve growth factor. II. Isolation and characterization of the thermolytic and peptic peptides and the complete structure. Biochemistry 12: 100115. Apfel, S.C., Arezzo, J.C and Kessler, J.A. (1991) Nerve growth factor prevents toxic neuropathy in mice. Ann. Neurol. 29: 87-89. Apfel, S.C, Arezzo, J.C., Lipson, L. and Kessler, J.A. (1992) Nerve growth factor prevents experimental cisplatin neuropathy. Ann. Neurol. 31: 76-80. Auburger, G., Heumann, R., Hellweg, R., Korsching, S. and Thoenen, H. (1987) Developmental changes of nerve

neurons

growth factor and its mRNA in the rat hippocampus: comparison with choline acetyltransferase. Dev. Biol. 120: 3 2 2 328. Awatsuji, H., Furukawa, Y., Hirota, M., Murakami, S., Nii, S., Furukawa, S. and Hayashi, K. (1993) Interleukin-4 and -5 as modulators of nerve growth factor synthesis-secretion in astrocytes. J. Neurosci. Res. 34: 539-545. Babb, T.L. (1991) Bilateral pathological damage in temporal lobe epilepsy. Can. J. Neurol. Sci. 18: 645s-648s. Ballarin, M., Ernfors, P., Linderfors, N. and Persson, H. (1991) Hippocampal damage and kainic acid injection induce a rapid increase in mRNA for BDNF and NGF in the rat brain. Exp. Neurol. 114: 3 5 ^ 3 . Bandtlow, C.E., Heumann, R., Schwab, M.E. and Thoenen, H. (1987) Cellular localization of nerve growth factor synthesis by in situ hybridization. EMBO J. 6: 891-899. Barakat-Walter, I., Due, C. and Puymirat, J. (1993) Changes in nuclear 3,5,3'-triiodothyronine receptor expression in rat dorsal root ganglia and sciatic nerve during development: comparison with regeneration. Eur. J. Neurosci. 5: 3 1 9 326. Barbany, G. and Persson, H. (1992) Regulation of neurotrophin mRNA expression in the rat brain by glucocorticoids. Eur. J. Neurosci. 4: 396-403. Barde, Y.A. (1989) Trophic factors and neuronal survival. Neuron 2: 1525-1534. Barde, Y.A., Edgar, D. and Thoenen, H. (1982) Purification of a new neurotrophic factor from mammalian brain. EMBO J. 1:549-553. Barth, E.M., Korsching, S. and Thoenen, H. (1984) Regulation of nerve growth factor synthesis and release in organ cultures of rat iris. J. Cell Biol. 99: 839-843. Bartus, R.T., Dean, R.L. and Beer, B. (1982) Neuropeptide effects on memory in aged monkeys. Neurobiol. Aging 3: 61-68. Batistatou, A. and Greene, L.A. (1991) Aurintricarboxylic acid rescues PC 12 cells and sympathetic neurons from cell death caused by nerve growth factor deprivation: correlation with suppression of endonuclease activity. J. Cell Biol. 115:461-471. Berkemeier, L.R., Winslow, J.W., Kaplan, D.R., Nikolics, K., Goeddel, D.V. and Rosenthal, A. (1991) Neurotrophin-5: a novel neurotrophic factor that activates trk and trkB. Neuron 7: 857-866. Black, M.A., Pope, L., Lefebvre, F., Lefebvre, Y. and Walker, P. (1992) Thyroid hormones precociously increase nerve growth factor gene expression in the submandibular gland of neonatal mice. Endocrinology 130: 2083-2090. Borasio, G.D., Markus, A., Wittinghofer, A., Barde, Y.A. and Heumann, R. (1993) Involvement of ras p21 in neurotrophin-induced response of sensory, but not sympathetic neurons. /. Cell Biol. 121: 665-672. Bradshaw, R.A. (1978) Nerve growth factor. Annu. Rev. Biochem. 47: 191-216. Bradshaw, R.A. (1983) Nerve growth factor and related hormones. In: G. Litwack (Ed.), Biochemical Actions of

R.A. Rush et al. Hormones, Vol. X, Academic Press, New York, pp. 9 1 115. Burnstock, G. (1975) Innervation of vascular smooth muscle: histochemistry and electronmicroscopy. Clin. Exp. Pharmacol. Physiol. 2: 7s-20s. Campenot, R.B. (1982) Development of sympathetic neurons in compartmentalized cultures. IL control of neurite growth by nerve growth factor. Dev. Biol. 93: 1-12. Carman-Krzan, M. and Wise, B.C. (1993) Arachidonic acid lipoxygenation may mediate interleukin-1 stimulation of nerve growth factor secretion in astroglial cultures. J. Neurosci. Res. 34: 225-232. Carman-Krzan, M., Vige, X. and Wise, B.C. (1991) Regulation by interleukin-1 of nerve growth factor secretion and nerve growth factor mRNA expression in rat primary astroglial cultures. /. Neurochem. 56: 636-643. Cassis, L.A., Stitzel, R.E. and Head, R.J. (1985) Hypernoradrenergic innervation of the caudal artery of the spontaneously hypertensive rat. An influence upon neuroeffector mechanisms. /. Pharmacol. Exp. Ther. 234: 792-803. Chun, L.L.Y. and Patterson, P.H. (1977) Role of nerve growth factor in the development of rat sympathetic neurons in vitro: affect on acetylcholine production. J. Cell Biol. 75: 712-718. Clegg, D.O., Large, T.H., Bodary, S. and Reichardt, L.F. (1989) Regulation of nerve growth factor mRNA levels in developing rat heart ventricle is not altered by sympathectomy. Dev. Biol. 134: 30-37. Clos, J. and Legrand, C. (1990) An interaction between thyroid hormone and nerve growth factor promotes the development of hippocampus, olfactory bulbs and cerebellum: a comparative biochemical study of normal and hypothyroid rats. Growth Factors 3: 205-220. Clozel, J.P., Kuhn, H. and Hefti, F. (1989) Decreases of vascular hypertrophy in four different types of arteries in spontaneously hypertensive rats. Am. J. Med. 87: 93s-95s. Cohen, S. (1960) Purification of a nerve growth promoting protein from the mouse salivary gland and its neurocytotoxic anti-serum. Proc. Natl. Acad. Sci. USA 46: 302-311. Conner, J.M., Muir, D., Varon, S., Hagg, T. and Manthorpe, M. (1992) The localization of nerve growth factor-like immunoreactivity in the adult rat basal forebrain and hippocampal formation. J. Comp. Neurol. 319: 454-462. Cowan, W.M., Fawcett, J.W., O'Leary, D.D. and Stanfield, B.B. (1984) Regressive events in neurogenesis. Science 225: 1258-1265. Creedon, D. and Tuttle, J.B. (1991) Nerve growth factor synthesis in vascular smooth muscle. Hypertension 18: 30-741. Davies, A. and Lumsden, A. (1983) Influence of NGF on developing dorso-medial and ventro-lateral neurons of chick and mouse trigeminal ganglia. Int. J. Dev. Neurosci. 1: 171-177. Davies, A.M., Thoenen, H. and Barde, Y.A. (1986a) The response of chick sensory neurons to brain-derived neurotrophic factor. J. Neurosci. 6: 1897-1904. Davies, A.M., Thoenen, H. and Barde, Y.A. (1986b) Different

193 factors from the central nervous system and periphery regulate the survival of sensory neurones. Nature 319: 497499. Davies, A.M., Bandtlow, C, Heumann, R., Korsching, S., Rohrer, H. and Thoenen, H. (1987) Timing and site of nerve growth factor synthesis in developing skin in relation to innervation and expression of the receptor. Nature 326: 353-357. DeSchryver-Kecskemeti, K., Clouse, R.E., Goldstein, M.N., Gersell, D. and O'Neal, L. (1983) Intestinal ganglioneuromatosis: a manifestation of overproduction of nerve growth factor? N. Engl. J. Med. 308: 635-639. Diamond, J., Coughlin, M., Maclntyre, L., Holmes, M. and Visheau, B. (1987) Evidence that endogenous betanerve growth factor is responsible for the collateral sprouting, but not the regeneration of nociceptive axons in adult rats. Proc. Natl. Acad. Sci. USA 84: 6596-6600. Dicou, E., Lee, J. and Brachet, P. (1986) Synthesis of nerve growth factor mRNA and precursor protein in the thyroid and parathyroid glands of the rat. Proc. Natl. Acad. Sci. USA 83:7084-7088. D'Mello, S.R. and Heinrich, G. (1991) Structural and functional identification of regulatory regions and cis elements surrounding the nerve growth factor gene promoter. Mol. Brain Res. 11:255-264. Donohue, S.J., Head, R.J. and Stitzel, R.E. (1989) Elevated nerve growth factor levels in young spontaneously hypertensive and Wistar-Koyoto rats. Hypertension 14: 421-426. Ebendal, T. and Persson, H. (1988) Detection of nerve growth factor mRNA in the developing chicken embryo. Development 102: 101-106. Ebendal. T., Olson, L., Seiger, A. and Hedlund, K.O. (1980) Nerve growth factors in the rat iris. Nature 286: 25-28. Edwards, R.H. (1993) Regulation of nerve growth factor expression. In: S.E. Loughlin and J.H. Fallon (Eds.), Neurotrophic Factors, Academic Press, New York, pp. 181207. Edwards, R.H., Selby, M.J., Garcia, P.D. and Rutter, W.J. (1988a) Processing of the native nerve growth factor precursor to form biologically active nerve growth factor. J. Biol. Chem. 263: 6S10-6S\5. Edwards, R.H., Selby, M.J., Mobley, W.C, Weinrich, S.L., Hruby, D.E. and Rutter, W.J. (1988b) Processing and secretion of nerve growth factor: expression in mammalian cells with a vaccinia virus vector. Mol. Cell. Biol. 8: 24562464. Edwards, R.H., Rutter, W.J. and Hanahan, D. (1989) Directed expression of NGF to pancreatic y3-cells in transgenic mice leads to selective hyperinnervation of the islets. Cell 58: 161-170. Ernfors, P., Ebendal, T., Olson, L., Mouton, P., Stromberg, I. and Persson, H. (1989) A cell line producing recombinant nerve growth factor evokes growth response in intrinsic and grafted central cholinergic neurons. Proc. Natl. Acad. Sci. USA 86: 4756^760. Ernfors, P., Ibanez, C.F., Ebendal, T., Olson, L. and Persson,

194

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

H. (1990) Molecular cloning and neurotrophic activities of a protein with structural similarities to nerve growth factor: developmental and topographical expression in the brain. Proc. Natl. Acad. Sci. USA 87: 5454-5458. Ernfors, P., Bengzon, J., Kokaia, Z., Persson, H. and Lindvall, O. (1991) Increased levels of messenger RNAs for neurotrophic factors in the brain during kindling epileptogenesis. Neuron 7: 165-176. Fabricant, R.N. and Todaro, G.J. (1981) Increased serum levels of nerve growth factor in von Recklinghausen's disease. Acta Neurol. 38: 401-405. Fabricant, R.N., Todaro, G.J. and Eldridge, R. (1979) Increased levels of nerve-growth-factor cross-reacting protein in 'central' neurofibromatosis. Lancet 6: 4-7. Falckh, P.H., Harkin, L.A. and Head, R.J. (1992a) Resistance vessel gene expression of nerve growth factor is elevated in young spontaneously hypertensive rats. J. Hypertension 10: 913-918. Falckh, P.H., Harkin, L.A. and Head, R.J. (1992b) Nerve growth factor mRNA content parallels altered sympathetic innervation in the spontaneously hypertensive rat. Clin. Exp. Pharmacol. Physiol. 19: 541-545. Federoff, H.J., Geschwind, M.D., Geller, A.I. and Kessler, J.A. (1992) Expression of nerve growth factor in vivo from a defective herpes simplex virus 1 vector prevents effects of axotomy on sympathetic ganglia. Proc Natl. Acad. Sci. USA 89: 1636-1640. Ferguson, I.A., Williams, R. and Rush, R.A. (1989) Chicken NGF and non-NGF trophic factor synthesis and release by sciatic nerves in vitro. J. Neurosci. Res. 22: 408-417. Figliomeni, B., Bacci, B., Panozzo, C, Fogarolo, F., Triban, C. and Fiori, M.G. (1992) Experimental diabetic neuropathy. Effect of ganglioside treatment on axonal transport of cytoskeletal proteins. Diabetes 41: 866-871. Fischer, W., Wictorin, K., Bjorklund, A., Williams, L.R., Varon, S. and Gage, F.H. (1987) Amelioration of cholinergic neuron atrophy and spatial memory impairment in aged rats by nerve growth factor. Nature 329: 65-68. Folkow, B., Hallback, M., Lundgrow, Y. and Weiss, L. (1972) The effects of immunosympathectomy on blood pressure and vascular reactivity in normal and spontaneously hypertensive rats. Acta Physiol. Scand. 84: 512-523. Franklin, J.L. and Johnson, E.M. (1992) Suppression of programmed neuronal death by sustained elevation of cytoplasmic calcium. Trenda Neurosci. 15: 501-508. Friedman, W.J., Larkfors, L., Ayer-Lelievre, C, Ebendal, T., Olson, L. and Persson, H. (1990) Regulation of ^-nerve growth factor expression by inflammatory mediators in hippocampal cultures. J. Neurosci. Res. 27: 374-382. Friedman, W.J., Altiok, N., Fredholm, B.B. and Persson, H. (1992) Mechanisms of nerve growth factor mRNA regulation by interleukin-1^ in hippocampal cultures: role of second messengers. J. Neurosci. Res. 33: 37^6. Funakoshi, H., Belluardo, N., Arenas, E., Yamamoto, Y., Casabona, A., Persson, H. and Ibafiez, C.F. (1995) Musclederived neurotrophin-4 as an activity-dependent trophic

signal for adult motor neurons. Science 268: 14951499. Furukawa, Y., Furukawa, S., Satoyoshi, E. and Hayashi, K. (1986a) Catecholamines induce and increase in nerve growth factor content in the medium of mouse L-M cells. J. Biol. Chem. 261: 6039-6047. Furukawa, Y., Furukawa, S., Ikeda, F., Satoyoshi, E. and Hayashi, K. (1986b) Aliphatic side chain of catecholamine potentiates the stimulatory effect of the catechol part on the synthesis of nerve growth factor. FEBS Lett. 208: 258-262. Furukawa, Y., Furukawa, S., Satoyoshi, E. and Hayashi, K. (1987) Regulation of nerve growth factor synthesis/ secretion by catecholamine in cultured mouse astroglial cells. Biochem. Biophys. Res. Commun. HI: 1048-1054. Furukawa, Y., Fukazawa, N., Miyama, Y., Hayashi, K. and Furukawa, S. (1990) Stimulatory effect of 4-alkylcatechols and their diacetylated derivatives on the synthesis of nerve growth factor. Biochem. Pharmacol. 40: 2337-2342. Furukawa, Y., Furukawa, S., Omae, F., Awatsuji, H. and Hayashi, K. (1993) Alkylcatechols regulate NGF gene expression in astroglial cells via both protein kinase C and cAMPindependent mechanisms. J. Neurosci. Res. 5: 522-529. Gadient, R.A., Cron, K.C. and Otten, U. (1990) Interleukin-1 and tumor necrosis factor-alpha synergystically stimulate nerve growth factor (NGF) release from cultured rat astrocytes. Neurosci. Lett. 117: 335-340. Gage, F.H., Armstrong, D.M., Williams, L.R. and Varon, S.V. (1988) Morphological responses of axotomized septal neurons to nerve growth factor. /. Comp. Neurol. 269: 147155. Gall, C. and Isackson, P.J. (1989) Limbic seizures increase neuronal production of messenger RNA for nerve growth factor. Science 245: 758-761. Gall, CM. and Lauterborn, J.C. (1992) The dentate gyrus: a model system for studies of neurotrophin regulation. Epilepsy Res. 1 (SuppL): 171-185. Gall, C, Murray, K. and Isackson, P. (1991) Kainic acidinduced seizures stimulate increased expression of nerve growth factor mRNA in rat hippocampus. Mol. Brain Res. 9: 113-123. Garcia, I., Martinou, I., Tsujimoto, Y. and Martinou, J.C. (1992) Prevention of programmed cell death of sympathetic neurons by the bcl-2 proto-oncogene. Science 258: 302304. Gnahn, H., Hefti, F., Heumann, R., Schwabb, M.E. and Thoenen, H. (1983) NGF-mediated increase of choline acetyltransferase (ChAT) in the neonatal rat forebrain evidence for a physiological role of NGF in the brain? Dev. Brain Res. 9: 45-52. Goedert, M., Otten, U. and Thoenen, H. (1978) Biochemical effects of antibodies against nerve growth factor on developing and differentiated sympathetic ganglia. Brain Res. 148: 264-268. Goedert, M., Fine, A., Hunt, S.P. and Ullrich, A. (1986) Nerve growth factor mRNA in peripheral and central rat tissues and in human central nervous system: lesion effects in the

R.A. Rush et al. rat brain and levels in Alzheimer's disease. Mol. Brain Res. 1:85-92. Gonzalez, D., Dees, W.L., Hiney, J.K., Ojeda, S.R. and Saneto, R.P. (1990) Expression of beta-nerve growth factor in cultured cells derived from the hypothalamus and cerebral cortex. Brain Res. 511: 249-258. Gorin, P.D. and Johnson, E.M. (1979) Experimental autoimmune model of NGF deprivation: effects on developing peripheral sympathetic and sensory neurons. Proc. Natl. Acad. Sci. USA 76: 5382-5386. Goto, M., Demura, N. and Sakaguchi, T. (1987) Effects of propentofylline on disorder of learning and memory in rodents. Jpn. J. Pharmacol. 45: 373-378. Greene, L.A. (1977a) Quantitative in vitro studies on the nerve growth factor (NGF) requirement of neurons: sensory neurons. Dev. Biol. 58: 106-113. Greene, L.A. (1977b) Quantitative in vitro studies on the nerve growth factor (NGF) requirements of neurons: sympathetic neurons. Dev. Biol. 58: 96-105. Greene, L.A. and Shooter, E.M. (1980) The nerve growth factor: biochemistry, synthesis and mechanism of action. Annu. Rev. Neurosci. 3: 353-402. Gundersen, R.W. and Barrett, J.N. (1979) Neuronal chemotaxis: chick dorsal-root axons turn toward high concentrations of nerve growth factor. Science 206: 10791080. Gwag, B.J., Sessler, F.M., Waterhouse, B.D. and Springer, J.E. (1993) Regulation of nerve growth factor mRNA in the hippocampal formation: effects of N-methyl-D-aspartate receptor activation. Exp. Neurol. 121: 160-171. Hagg, T., Manthorpe, M., Vahlsing, H.L. and Varon, S. (1988) Delayed treatment with nerve growth factor reverses the apparent loss of cholinergic neurons after acute brain damage. Exp. Neurol. 101: 303-312. Hagg, T., Fass-Holmes, B., Vahlsing, H.L., Manthorpe, M., Conner, J.M. and Varon, S. (1989) Nerve growth factor (NGF) reverses axotomy-induced decreases in choline acetyltransferase, NGF receptor and size of medial septum cholinergic neurons. Brain Res. 505: 29-38. Hagg, T., Vahlsing, H.L., Manthorpe, M. and Varon, S. (1990) Nerve growth factor infusion into the denervated adult rat hippocampus formation promotes its cholinergic reinnervation. J. Neurosci. 10: 3087-3092. Hallbook, F., Ibanez C.F. and Persson, H. (1991) Evolutionary studies of the nerve growth factor family reveal a novel member abundantly expressed in Xenopus ovary. Neuron 6: 845-858. Hamburger, V. and Oppenheim, R.W. (1982) Naturally occurring neuronal death in vertebrates. Neurosci. Comm. 1: 39-55. Hanaoka, Y., Ohi, T., Furukawa, S., Furukawa, Y., Hayashi, K. and Matsukura, S. (1992) Effect of 4-methylcatechol on sciatic nerve growth factor level and motor nerve conduction velocity in experimental diabetic neuropathic process in rats. Exp. Neurol. 115: 292-296. Harper, S. and Davies, A.M. (1990) NGF mRNA expression

195 in developing cutaneous epithelium related to innervation density. Development 110: 515-519. Harrap, S.B. and Doyle, A.E. (1986) Renal hemodynamics and exchangeable sodium in the spontaneously hypertensive rat. J. Hypertens. 4: 249s-252s. Harrap, S.B. and Lever, A.F. (1989) The long-term effects after ACE inhibitor treatment in young spontaneously hypertensive rats - clues to the pathogenesis of high blood pressure. In: G.A. MacGregor and P.S. Sever (Eds.), Current Advances in ACE Inhibition, Churchill Livingstone, Edinburgh, pp. 69-77. Hashimoto, Y., Omae, F. and Furukawa, S. (1992) Reduction in NGF protein level in rat dorsal hippocampus following administration of kainic acid. Neurosci. Lett. 140: 203205. Hayashi, M., Edgar, D. and Thoenen, H. (1985) Nerve growth factor changes the relative levels of neuropeptides in developing sensory and sympathetic ganglia of the chick embryo. Dev. Biol. 108: 49-55. Head, R.J. (1989) Hypernoradrenergic innervation: its relationship to functional and hyperplastic changes in the vasculature of the spontaneously hypertensive rat. Blood Vessels 26: 1-20. Hefti, F. (1986) Nerve growth factor promotes survival of septal cholinergic neurons after fimbrial transection. J. Neurosci. 6: 2155-2162. Hefti, F. and Weiner, W.J. (1986) Nerve growth factor and Alzheimer's disease. A/in. Neurol. 20: 275-281. Hefti, F., Hartikka, J., Eckenstein, F., Gnahn, H., Heumann, R. and Schwab, M. (1985) Nerve growth factor increases choline acetyltransferase but not survival or fibre outgrowth of cultured fetal septal cholinergic neurons. Neuroscience 14: 55-68. Hellweg, R. and Hartung, H.D. (1990) Endogenous levels of NGF are altered in experimental diabetes mellitus: a possible role for NGF in the pathogenesis of diabetic neuropathy. /. Neurosci. Res. 26: 258-267. Hellweg, R., Bandtlow, C.E., Heumanm, R. and Korsching, S. (1988) Nerve growth factor synthesis in cultured rat iris: modulated by endogenous transmitter substances. Exp. Cell Res. 179: 18-30. Hellweg, R., Wohrle, M., Hartung, H.D., Stracke, H., Hock, C. and Federlin, K. (1991) Diabetes mellitus-associated decrease in nerve growth factor levels is reversed by allogeneic pancreatic islet transplantation. Neurosci. Lett. 125: 14. Henderson, C.E., Camu, W., Mettling, C, Gouin, A., Poulsen, K., Karihaloo, M., Rullamas, J., Evans, T., McMahon, S.B., Armanini, M.P., Berkemeier, L., Phillips, H.S. and Rosenthal, A. (1993) Neurotrophins promote motor neuron survival and are present in embryonic limb bud. Nature 363: 266-270. Hendry, LA. (1977) The effect of the retrograde axonal transport of nerve growth factor on the morphology of adrenergic neurones. Brain Res. 134: 213-223. Hendry, LA. and Campbell, J. (1976) Morphometric analysis

196

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

of rat superior cervical ganglion after axotomy and nerve growth factor treatment. /. Neurocytol. 5: 351-360. Hendry, I.A., Stockel, K., Thoenen, H. and Iversen, L.L. (1974) The retrograde transport of nerve growth factor. Brain Res. 68: 103-121. Hendry, I.A., Messina, A. and Bell, C. (1992) Neonatal nerve growth factor treatment alters the preganglionic innervation pattern of rat superior cervical ganglion. Neurosci. Lett. 148: 117-120. Hengerer, B., Lindholm, D., Heumann, R., Ruther, U., Wagner, E.F. and Thoenen, H. (1990) Lesion-induced increase in nerve growth factor mRNA is mediated by c-fos. Proc. Natl. Acad. Sci. USA 87: 3899-3903. Heumann, R. and Thoenen, H. (1986) Comparison between the time course of changes in nerve growth factor and those of its messenger RNA in the cultured rat iris. J. Biol. Chem. 261:9246-9249. Heumann, R., Korsching, S., Scott, J. and Thoenen, H. (1984) Relationship between levels of nerve growth factor (NGF) and its messenger RNA in sympathetic ganglia and peripheral target tissues. EMBO J. 3: 3183-3189. Heumann, R., Korsching, S., Bandtlow, C. and Thoenen, H. (1987a) Changes of nerve growth factor synthesis in nonneuronal cells in response to sciatic nerve transection. J. Cell Biol. 104: 1623-1631. Heumann, R., Lindholm, D., Bandtlow, C, Meyer, M., Radeke, M.J., Misko, T.P., Shooter, E. and Thoenen, H. (1987b) Differential regulation of mRNA encoding nerve growth factor and its receptor in rat sciatic nerve during development, degeneration, and regeneration: role of macrophages. Proc. Natl. Acad.. Sci. USA 84: 8735-8739. Heumann, R., Lindholm, D., Bandtlow, C.E., Meyer, M., Rohrer, H. and Thoenen, H. (1989) Neuroimmune Networks. Physiology and Diseases. Alan R. Liss, New York. Hilbert, P., Lindpaintner, K., Beckmann, J.S., Serikawa, T., Soubrier, F., Dubay, C, Cartwright, P., De-Gouyon, B., Julier, C, Takahasi, S., Vincent, M., Ganten, D., Georges, M. and Lathrop, G.M. (1991) Chromosomal mapping of two genetic loci associated with blood-pressure regulation in hereditary hypertensive rats. Nature 353: 521-529. Hill, C.E., Hirst, G.D.S. and Van Helden, D.F. (1983) Development of sympathetic innervation to proximal and distal arteries of the rat mesentery. J. Physiol. 338: 129-147. Hill, C.E., Hendry, I.A., Ngu, M.C. and Van Helden, D.F. (1985) Subpopulations of sympathetic neurons differ in their sensitivity to nerve growth factor antiserum. Dev. Brain Res. 23: 121-130. Hofer, M.M. and Barde, Y.A. (1988) Brain-derived neurotrophic factor prevents neuronal death in vivo. Nature 331:261-262. Hohn, A., Leibrock, J., Bailey, K. and Barde, Y.A. (1990) Identification and characterization of a novel member of the nerve growth factor/brain derived neurotrophic factor family. A^amre 344: 339-341. Honegger, P. and Lenoir, D. (1982) Nerve growth factor

(NGF) stimulation of cholinergic telencephalic neurons in aggregating cell cultures. Dev. Brain Res. 3: 229-238. Hory-Lee, F., Russell, M., Lindsay, R.M. and Frank, E. (1993) Neurotrophin 3 supports the survival of developing muscle sensory neurons in culture. Proc. Natl. Acad. Sci. USA 90: 2613-2617. Houlgatte, R., Wion, D., Barrand, P., Dicou, E. and Brachet, P. (1986) Serum influences P-NGF gene expression in mouse L cells. In P. Calissano (Ed.), Molecular Aspects of Neurobiology, Springer, Berlin, pp. 4 0 ^ 2 . Houlgatte, R., Wion, D. and Brachet, P. (1989) Levels of nerve growth factor secreted by rat primary fibroblasts and iris transplants are influenced by serum and glucocorticoids. Dev. Brain Res. 47: 171-179. Hoyle, G.W., Mercer, E.H., Palmitier, R.D. and Brinster, R.L. (1993) Expression of NGF in sympathetic neurons leads to excessive axon outgrowth from ganglia but decreases terminal innervation within tissues. Neuron 10: 1019-1034. Hulst, J.R. and Bennett, M. (1986) Motoneurone survival factor produced by muscle increases in parallel with messenger RNA sequences homologous to ^NGF-cDNA. Dev. Brain Res. 25: 153-156. Hyman, C, Hofer, M., Barde, Y.A., Juhasz, M., Yancopoulos, G.D., Squinto, S.P. and Lindsay, R.M. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350: 230-232. Ikegami, R., Furukawa, Y., Kaechi, K., Hayashi, K. and Furukawa, S. (1990) Effects of catecholamines and 4methylcatechol on the synthesis and secretion of nerve growth factor by rat sciatic nerve segments in culture. Biomed. Res. 11:61-65. Ip, N.Y., Ibanez, C.F., Nye, S.H., McClain, J., Jones, P.F., Gies, D.R., Belluscio, L., Le Beau, M.M., Espinosa, R., Squinto, S.P., Persson, H. and Yancopoulos, G.D. (1992) Mammalian neurotrophin-4: structure, chromosomal localization, tissue distribution and receptor specificity. Proc. Natl. Acad. Sci. USA 89: 3060-3064. Isaacson, L.G., Saffran, B.N. and Crutcher, K.A. (1990) Intracerebral NGF infusion induces hyperinnervation of cerebral blood vessels. Neurobiol. Aging 11: 51-55. Jehan, F., Neveu, I., Barbot, N., Binderup, L., Brachet, P. and Wion, D. (1994) MC903, an analogue of 1,25dihydroxyvitamin D3, increases the synthesis of nerve growth factor. Eur. J. Pharmacol. 208: 189-191. Johnson, E.M., Gorin, P.D., Osborne, P.A., Rydel, R.E. and Pearson, J. (1982) Effects of autoimmune NGF deprivation in the adult rabbit and offspring. Brain Res. 240: 131-140. Johnson, E.M., Rich, J.K. and Yip, H.K. (1986) The role of NGF in sensory neurons in vivo. Trends Neurosci. 9: 3337. Kaechi, K., Furukawa, Y., Ikegami, R., Nakamura, N., Omae, F., Hashimoto, Y., Hayashi, K. and Furukawa, S. (1993) Pharmacological induction of physiologically active nerve growth factor in rat peripheral nervous system. J. Pharmacol. Exp. Ther. 264: 321-326. Katoh-Semba, R., Semba, R., Kashiwamata, S. and Kato, K.

RA. Rush et al. (1989) Sex-dependent and sex-independent distribution of the y3-subunit of nerve growth factor in the central nervous and peripheral tissues of mice. J. Neurochem. 52: 15591565. Katoh-Semba, R., Semba, R., Kashiwamata, S. and Kato, K. (1990) Influences of neonatal and adult exposures to testo sterone on the levels of the y3-subunit of nerve growth factor in the neural tissues of mice. Brain Res. 522: 112-117. Kessler, J.A. and Black, LB. (1980a) Nerve growth factor stimulates the development of substance P in sensory ganglia. Proc. Natl Acad. Sci. USA 11: 549-552. Kessler, J.A. and Black, LB. (1980b) The effects of nerve growth factor (NGF) and antiserum to NGF on the development of embryonic sympathetic neurons in vivo. Brain Res. 189: 157-168. Kessler, J.A. and Black, LB. (1981) Nerve growth factor stimulates the development of substance P in sensory ganglia. Proc. Natl. Acad. Sci. USA 11: 649-652. Kirino, T. (1982) Delayed neuronal death in the gerbil hippocampus following ischemia. Brain Res. 239: 57-69. Klingman, G.I. and Klingman, J.D. (1972) Effects of immunosympathectomy on biogenic amines and cholinesterases. In: E. Zaimis (Ed.), Nerve Growth Factor and its Antiserum, Athlone Press, London, pp. 125-145. Knusel, B., Winslow, J.W., Rosenthal, A., Burton, L.E., Seid, D.P., Nikolics, K. and Hefti, F. (1991) Promotion of central cholinergic and dopaminergic neuron differentiation by brain-derived neurotrophic factor but not neurotrophin 3. Proc. Natl. Acad. Sci. USA 88: 961-965. Koliatsos, v., Clatterbuck, R.E., Winslow, J.W., Cayoutte, H. and Price, D.L. (1993) Evidence that brain derived neurotrophic factor is a trophic factor for motor neurons in vivo. Neuron 10: 359-367. Kornblum, H.I. and Johnson, E.M. (1982) Time and dose dependencies of effects of nerve growth factor on sympathetic and sensory neurons in neonatal rats. Brain Res, 234: 41-51. Korsching, S. and Thoenen, H. (1983) Levels of nerve growth factor in sympathetic ganglia and corresponding target organs of the rat: correlation with density of sympathetic innervation. Proc. Natl. Acad. Sci. USA 80: 35133516. Korsching, S. and Thoenen, H. (1985) Nerve growth factor supply for sensory neurons: site of origin and competition with the sympathetic nervous system. Neurosci. Lett. 54: 201-205. Korsching, S. and Thoenen, H. (1988) Developmental changes of nerve growth factor levels in sympathetic ganglia and their target organs. Dev. Biol. 126: 40-^6. Korsching, S., Heumann, R., Thoenen, H. and Hefti, F. (1986) Cholinergic denervation of the rat hippocampus by fimbrial transection leads to a transient accumulation of nerve growth factor (NGF) without change in mRNA^^^ content. Neurosci. Lett. 66: 175-180. Kozak, M. (1988) A profusion of controls. J. Cell Biol. 107: 1-7.

197 Kromer, L.F. (1987) Nerve growth factor treatment after brain injury prevents neuronal death. Science 235: 214-216. Laiwand, R., Werman, R. and Yarom, Y. (1987) Time course and distribution of motor neuronal loss in the dorsal motor vagal nucleus of guinea pig after cervical vagotomy. J. Comp. Neurol. 256: 527-537. Lara, H.E.. Hill, D., Katz, K.H. and Ojeda, S.R. (1990) The gene encoding nerve growth factor is expressed in the immature rat ovary: effect of denervation and hormonal treatment. Endocrinology 126: 357-363. Larkfors, L., Ebendal, T., Whittemore, S.R., Persson, H., Hoffer, B. and Olson, L. (1987) Decreased level of nerve growth factor (NGF) in the aged rat brain. Mol. Brain Res. 3: 55-60. Larkfors, L., Oskarson, A., Sundberg, J. and Ebendal, T. (1991) Methylmercury induced alterations in the nerve growth factor level in the developing brain. Dev. Brain Res. 62:287-291. Larmet, Y., Dolphin, A.C. and Davies, A.M. (1992) Intracellular calcium regulates the survival of early sensory neurons before they become dependent on neurotrophic factors. Neuron 9:563-574. Lee, R.M.K.W., Forrest, J.B., Garfield, R.E. and Daniel, E.E. (1983a) Morphometric study of structural changes in the mesenteric blood vessels of spontaneously hypertensive rats. Blood Vessels 20: 57-71. Lee, R.M.K.W., Forrest, J.B., Garfield, R.E. and Daniel, E.E. (1983b) Comparison of blood vessel wall dimensions in normotensive and hypertensive rats by histometric and morphometric methods. Blood Vessels 20: 245-254. Lee, R.M.K.W., Triggle, G.R., Cheung, D.W.T. and Coughlin, M.D. (1987) Structural and functional consequence of neonatal sympathectomy on the blood vessels of spontaneously hypertensive rats. Hypertension 10: 328-338. Lee, R.M., Coughlin, M., Tsoporis, J., Kwan, C.Y., Guan, Y.Y. and Leenen, F.H. (1992) The effect of neonatal treatment of rats with nerve growth factor on the blood pressure and structure of the mesenteric arteries. Can. J. Physiol. Pharmacol. 10: 1154-1160. Lee, T.J. and Saito, A. (1984) Altered cerebral vessel innervation in the spontaneously hypertensive rat. Circ. Res. 55: 392^03. Legrand, C. and Clos, J. (1991) Biochemical, immunocytochemical and morphological evidence for an interaction between thyroid hormone and nerve growth factor in the developing cerebellum of normal and hypothyroid rats. Dev. Neurosci. 13: 382-396. Leibrock, J., Lottspeich, F., Hohn, A., Hofer, M., Hengerer, B., Masiakowski, P., Thoenen, H. and Barde, Y.A. (1989) Molecular cloning and expression of brain-derived neurotrophic factor. Nature 341: 149-152. Letourneau, P.C. (1978) Chemotactic response of nerve fiber elongation to NGF. Dev. Biol. 66: 183-196. Levi-Montalcini, R. (1983) The NGF-target cells interaction: a model system for the study of directed axonal growth and regeneration. In: B. Haber, J.R. Perez-Polo, G.A. Hashim

198

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

and A.M.G. Stella (Eds.), Nervous System Regeneration, Birth Defects, Vol. 19, Alan R. Liss, New York, pp. 3-22. Levi-Montalcini, R. (1987) The nerve growth factor 35 years later. Science 237: 1154-1162. Levi-Montalcini, R. and Hamburger, V. (1951) Selective growth stimulating effects of mouse sarcoma on the sensory and sympathetic nervous system of the chick embryo. J. Exp.Zool. 116:321-361. Lindholm, D., Heumann, R., Meyer, M. and Thoenen, H. (1987) Interleukin-1 regulates synthesis of nerve growth factor in non-neuronal cells of rat sciatic nerve. Nature 330: 658-659. Lindholm, D., Heumann, R., Hengerer, B. and Thoenen, H. (1988) Interleukin 1 increases stability and transcription of mRNA encoding nerve growth factor in cultured rat fibroblasts. J. Biol. Chem. 263: 16348-16351. Lindholm, D., Hengerer, B., Heumann, R., Carroll, P. and Thoenen, H. (1990a) Glucocorticoid hormones negatively regulate nerve growth factor expression in vivo and in cultured rat fibroblasts. Eur. J. Neurosci. 2: 795-801. Lindholm, D., Hengerer, B., Zafra, F. and Thoenen, H. (1990b) Transforming growth factor p\ stimulates expression of nerve growth factor in the rat CNS. Dev. Neurosci. 1:9-12. Lindholm, D., Castren, E., Hengerer, B., Zafra, P., Berninger, B. and Thoenen, H. (1992a) Differential regulation of nerve growth factor (NOP) synthesis in neurons and astrocytes by glucocorticoid hormones. Eur. J. Neurosci. 4: 404-410. Lindholm, D., Castren, E., Kiefer, R., Zafra, P. and Thoenen, H. (1992b) Transforming growth factor-yS in the rat brain: increase after injury and inhibition of astrocyte proliferation. J. Cell Biol. 117: 395-400. Lindholm, D., Castren, E., Berzaghi, M., Blochl, A. and Thoenen, H. (1994) Activity-dependent and hormonal regulation of neurotrophin mRNA levels in the brain implications for neuronal plasticity. J. Neurobiol. IS: 13621372. Lindsay, R.M. (1988) Nerve growth factors (NGP, BDNP) enhance axonal regeneration but are not required for survival of adult sensory neurons J. Neurosci. 8: 2394-2405. Lindsay, R.M. and Harmer, A.J. (1989) Nerve growth factor regulates expression of neuropeptide genes in adult sensory neurons. Nature 337: 362-364. Lindsay, R.M. and Rohrer, H. (1985) Placodal sensory neurons in culture: nodose ganglion neurons are unresponsive to NGP, lack NGP receptors but are supported by a liverderived neurotrophic factor. Dev. Biol. Ill: 30-48. Lindsay, R.M., Thoenen, H. and Barde, Y.A. (1985) Placode and neural crest-derived sensory neurons are responsive at early development stages to brain-derived neurotrophic factor. Dev. Biol. 112: 319-328. Lindsay, R.M., Lockett, C, Sternberg, J. and Winter, J. (1989) Neuropeptide expression in cultures of adult sensory neurons: modulation of substance P and calcitonin gene-related peptide levels by nerve growth factor. Neuroscience 33: 53-65.

Lindvall, O., Ernfors, P., Bengzon, J., Kokaia, Z., Smith, M.L., Siesjo, B.-K. and Persson, H. (1992) Differential regulation of mRNAs for nerve growth factor, brain-derived neurotrophic factor, and neurotrophin 3 in the adult rat brain following cerebral ischemia and hypoglycemic coma. Proc. Natl. Acad. Sci. USA 89: 648-652. Lorigados, L., Soderstrom, S. and Ebendal, T. (1992) Two site enzyme immunoassay for )3-NGP applied to human patient sera. J. Neurosci. Res. 32: 329-339. Lu, B., Lee, J.-M., Elliot, R., Dreyfus, C.P., Adler, J.E. and Black, LB. (1991) Regulation of NGP gene expression in CNS glia by cell-cell contact. Mol. Brain Res. 11: 359362. Lyall, P., Morton, J.J., Lever, A.P. and Cragoe, E.J. (1988) Angiotensin II activates Na-H exchange and stimulates growth in cultured vascular smooth muscle cells. Hypertension 4: 438s-441s. Maisonpierre, P.C, Belluscio, L., Squinto, S., Ip, N.Y., Purth, M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990) Neurotrophin-3: a neurotrophic factor related to NGP and BDNP. Science 241: 1446-1451. Matsui, K., Purukawa, S., Shibasaki, H. and Kikuchi, T. (1990) Reduction of nerve growth factor level in the brain of genetically ataxic mice (weaver, reeler). FEBS Lett. 276: 78-80. Matsuoka, I., Meyer, M. and Thoenen, H. (1991) Cell-typespecific regulation of nerve growth factor (NGP) synthesis in non-neuronal cells: comparison of Schwann cells with other cell types. J. Neurosci. 11: 3165-3177. Miller, P.D., Thazhumpal, C, Mathew, C. and Toma, J.G. (1991) Regulation of nerve growth factor receptor gene expression by nerve growth factor in the developing peripheral nervous system. J. Cell Biol. 112: 303-312. Mocchetti, I., De Bernardt, M.A., Szekely, A.M., Alho, H., Brooker, G. and Costa, M. (1989) Regulation of nerve growth factor biosynthesis by y3-adrenergic receptor activation in astrocytoma cells: a potential role of c-Pos protein. Proc. Natl. Acad. Sci. USA 86: 3891-3895. Montero, C.N. and Hefti, P. (1988) Rescue of lesioned septal cholinergic neurons by nerve growth factor: specificity and requirement for chronic treatment. J. Neurosci. 8: 29862999. Moore, H.P.H., Walker, M.D., Lee, P. and Kelly, R.B. (1983) Expressing a human proinsulin DNA in a mouse ACTHsecreting cell. Intracellular storage, proteolytic processing and secretion on stimulation. Cell 35: 531-538. Murase, K., Takeuchi, R., Iwata, E., Purukawa, Y., Purukawa, S. and Hayashi, K. (1992) Developmental changes in nerve growth factor level in rat serum. J. Neurosci. Res. 33: 282288. Murphy, R.A., Acheson, A., Hodges, R., Haskins, J., Richards, C, Reklow, E., Chlumecky, V., Barker, P.A., Alderson, R.P. and Lindsay, R.M. (1993) Immunological relationships of NGP, BDNP, and NT-3: recognition and functional inhibition by antibodies to NGP. J. Neurosci. 13: 2853-2862.

R.A. Rush et al. Naruse, M. and Inagami, T. (1982) Antibody-sensitive renin of adrenal and resistance vessels is markedly elevated in spontaneously hypertensive rats. Clin. Sci. 63: 187s-189s. Negro, A., Corsa, V., Skaper, S. and Callegaro, L. (1993) Nerve growth factor antibodies recognize neurotrophin-3. Neurochem. Res. 18: 705-709. Nemoto, K., Kageyama, H., Hagiwara, T., Tashiro, F., Tomita, T., Tomita, I., Hano, T., Nishio, I. and Ueyama, T. (1994) Mutation of low affinity nerve growth factor gene in spontaneously hypertensive and stroke-prone spontaneously hypertensive rats: one of the promising candidate genes for hypertension. Brain Res. 655: 267-270. Neveu, I., Barbot, N., Jehan, F., Wion, D. and Brachet, P. (1991) Antagonistic effects of dexamethasone and 1,25dihydroxyvitamin D3 on the synthesis of nerve growth factor. Mol. Cell. Endocrinol. 78: 1-6. Neveu, I., Jehan, F ., Houlgatte, R., Wion, D. and Brachet, P. (1992a) Activation of nerve growth factor synthesis in primary glial cells by phorbol 12-myristate 13-acetate: role of protein kinase C. Brain Res. 570: 316-322. Neveu, I., Jehan, F. and Wion, D. (1992b) Alteration in the levels of 1.25-(OH)2D3 and corticosterone found in experimental diabetes reduces nerve growth factor (NGF) gene expression in vitro. Life Sci. 50: 1769-1772. Nistico, G., Ciriolo, M.R., Fiskin, K., lannone, N., DeMartino, A. and Rotilio, G. (1992) NGF restores decrease in catalase activity and increases superoxide dismutase and glutathione peroxidase activity in the brain of aged rats. Free Radio: Biol. Med. 12: 177-181. Nja, A. and Purves, D. (1978) The effects of nerve growth factor and its antiserum on synapse in the superior cervical ganglion of the guinea pig. J. Physiol. 277: 53-75. Olson, L., Backlund, E.G., Ebendal, T., Freedman, R., Hamberger, B., Hansson, P., Hoffer, B., Lindblom, U., Meyerson, B., Stromberg, I., Sydow, O. and Seiger, A. (1991) Intraputaminal infusion of nerve growth factor to support adrenal medullary autografts in Parkinson's disease. One-year follow-up of first clinical trial. Arch. Neurol. 48: 373-381. Olson, L., Nordberg, A., Von-Holst, H., Backman, L., Ebendal, T., Alafuzoff, I., Amberla, K., Hartvig, P., Herlitz, A., Lilja, A., Lundquist, H., Langstrom, B., Meyerson, B., Viitanen, M., Winblad, B. and Seiger, A. (1992) Nerve growth factor affects ^^C-nicotine binding, blood flow, EEG, and verbal episodic memory in an Alzheimer patient (case report). J. Neural Transm. Park. Dis. Dement. Sect. 4: 79-95. Oppenheim, R.W. (1991) Cell death during development of the nervous system. Annu. Rev. Neurosci. 14: 453-501. Oppenheim, R.W., Maderdrut, J.L. and Wells, D.J. (1982) Reduction of naturally-occurring cell death in the thoracolumbar preganglionic cell column of the chick embryo by nerve growth factor and hemicholinium-3. Brain Res. 255: 134-139. Oppenheim, R.W., Qin-Wei, Y., Prevette, D. and Yan, Q. (1992) Brain-derived neurotrophic factor rescues developing avian motoneurons from cell death. Nature 360: 7 5 5 757.

199 Otten, U. and Lorenz, H.P. (1982) Nerve growth factor increases substance P, cholecystokinin and vasoactive intestinal polypeptide immunoreactivity in primary sensory neurones of new-born rats. Neurosci. Lett. 34: 153-158. Otten, U., Hatanaka, H. and Thoenen, H. (1978) Effect of NGF on general growth and differentiation of sympathetic neurons: induction of ornithine decarboxylase and tyrosine hydroxylase as respective examples. In: E. Usdin, I.J. Kopin and J. Barchas (Eds.), 4th International Catecholamine Symposium, Pacific Grove, CA, 1978, Pergamon Press, New York, pp. 115-117. Otten, U., Ruegg, U.T., Hill, R.C., Businger, F. and Peck, M. (1982) Correlation between substance P content of primary sensory neurones and pain sensitivity in rats exposed to antibodies to nerve growth factor. Eur. J. Pharmacol. 85: 351-354. Pantazis, N.J. and Jenson, R. (1985) The a and y subunits of 7S nerve growth factor are present in excess concentrations in the mouse submandibular gland. Mol. Cell Biochem. 69: 35-48. Purves, D. and Lichtman, J.W. (1978) Formation and maintenance of synaptic connections in autonomic ganglia. Physiol. Rev. 58: 821-862. Rabizadeh, S., Oh, J., Zhong, L.-T., Yang, J., Bitler, C M . , Butcher, L.L. and Bredesen, D.E. (1993) Induction of apoptosis by the low-affinity NGF receptor. Science 261: 3 4 5 348. Rodriguez-Tebar, A., Jeffrey, P.L., Thoenen, H. and Barde, Y.A. (1989) The survival of chick retinal ganglion cells in response to brain-derived neurotrophic factor depends on their embryonic age. Dev. Biol. 136: 296-303. Rohrer, H., Heumann, R. and Thoenen, H. (1988) The synthesis of nerve growth factor (NGF) in developing skin is independent of innervation. Dev. Biol. 128: 240-244. Rosenthal, A., Goeddel, D.V., Nguyen, T., Lewis, M., Shih, A., Laramee, G.R.. Nikolics, K. and Winslow, J.W. (1990) Primary structure and biological activity of a novel human neurotrophic factor. Neuron 4: 767-773. Ruit, K.G., Osborne, P.A., Schmidt, R.E., Johnson, E.M., Jr. and Snider, W.D. (1990) Nerve growth factor regulates sympathetic ganglion cell morphology and survival in the adult mouse. /. Neurosci. 10: 2412-2419. Ruit, K.G., Elliott, J.L., Osborne, P.A., Yan, Q. and Snider, W.D. (1992) Selective dependence of mammalian dorsal root ganglion neurons of nerve growth factor during embryonic development. Neuron 8: 573-587. Rush, R.A. (1984) Immunohistochemical localization of endogenous nerve growth factor. Nature 312: 364-367. Rush, R.A., Williams, R., Wilson, P.A., Vahaviolos, J. and McGowan, R. (1989) Histological identification of nerve growth factor and its mRNA. In: R.A. Rush (Ed.), Nerve Growth Factors, Wiley, Chichester, pp. 1 7 1 190. Rush, R.A., Saltis, J., Smet, P.J. and Williams, R. (1992) Survival factors for autonomic neurons. In: LA. Hendry and C.E. Hill (Eds.), Development, Regeneration and Plasticity

200

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

of the Autonomic Nervous System, Harwood, Berkshire, pp. 305-363. Scarisbrick, I.A., Jones, E.G. and Isackson, P.J. (1993) Coexpression of mRNAs for NGF, BDNF, and NT-3 in the cardiovascular system of the pre- and postnatal rat. J. Neurosci. 13: 875-893. Schafer, T., Scwab, M.E. and Thoenen, H. (1983) Increased formation of preganglionic synapses and axons due to a retrograde trans-synaptic action and NGF in rat sympathetic nervous system. J. Neurosci. 3: 1501-1510. Schecterson, L.C. and Bothwell, M. (1992) Novel roles for neurotrophins are suggested by BDNF and NT-3 mRNA expression in developing neurons. Neuron 9: 449463. Schule, R., Umesono, K., Mangelsdorf, D.J., Pike, J.W. and Evans, R.M. (1990) Jun-Fos and receptors for vitamin A and D recognise a common response element in cell. Cell 61:497-504. Schwab, M.E., Otten, U., Agid, Y. and Thoenen, H. (1979) Nerve growth factor (NGF) in the rat CNS: absence of specific retrograde axonal transport and tyrosine hydroxylase induction in locus coeruleus and substantia nigra. Brain Res. 168: 473-483. Schwartz, J.P. (1988) Stimulation of nerve growth factor mRNA content in C6 glioma cells by a ^-adrenergic receptor and by cyclic AMP. Glia 1: 282-285. Schwartz, J.P., Pearson, J. and Johnson, E.M. (1982) Effect of exposure to anti-NGF on sensory neurons of adult rats and guinea pigs. Brain Res. 244: 378-381. Scott, J., Selby, M.J., Urdea, M., Quiroga, M., Bell, G.I. and Rutter, W.J. (1983) Isolation and nucleotide sequence of a cDNA encoding the precursor of mouse nerve growth factor. Nature 302: 538-540. Scott, T.M. and Pang, S.C. (1983) The correlation between the development of sympathetic innervation and the development of medial hypertrophy in jejunal arteries in normotensive and hypertensive rats. J. Auton. Nerv. Syst. 8: 2532. Seiler, M. and Schwab, M.E. (1984) Specific retrograde transport of nerve growth factor (NGF) from neocortex to nucleus basalis in the rat. Brain Res. 300: 33-39. Selby, M.J., Edwards, R., Sharp, F. and Rutter, W.J. (1987) Mouse nerve growth factor gene: structure and expression. Mol. Cell. Biol. 1: 3057-3064. Sendtner, M., Holtmann, B., Kolbeck, R., Thoenen, H. and Barde, Y.A. (1992) Brain-derived neurotrophic factor prevents the death of motoneurons in newborn rats after nerve section. Nature 360: 757-759. Shelton, D.L. and Reichardt, L.F. (1984) Expression of the^nerve growth factor gene correlates with the density of sympathetic innervation in effector organs. Proc. Natl. Acad. Sci. USA SI: 7951-7955. Shelton, D.L. and Reichardt, L.F. (1986a) Studies on the expression of the )3-nerve growth factor (NGF) gene in the central nervous system: level and regional distribution of NGF mRNA suggests that NGF functions as a trophic factor

for several distinct populations of neurons. Proc. Natl. Acad. Sci. USA 83: 2714-2718. Shelton, D.L. and Reichardt, L.F. (1986b) Studies on the regulation of beta-nerve growth factor gene expression in the rat iris: the level of mRNA-encoding nerve growth factor is increased in irises placed in explant cultures in vitro, but not in irises deprived of sensory or sympathetic innervation in vitro. J. Cell Biol. 102: 19401948. Shigeno, T., Mima, T., Takakura, K., Graham, D.L, Kato, G., Hashimoto, Y. and Furukawa, S. (1991) Amelioration of delayed neuronal death in the hippocampus by nerve growth factor. J. Neurosci. 11: 2914-2919. Shinoda, I., Furukawa, Y. and Furukawa, S. (1990) Stimulation of nerve growth factor synthesis-secretion by propentofylline in cultured mouse astroglial cells. Biochem. Pharmacol. 39:1813-1816. Siggers, D.C., Boyer, S.H. and Eldridge, R. (1975) Nerve growth factor in disseminated neurofibromatosis. N. Engl. J. Med. 292: 1194. Siminoski., K., Murphy, R.A., Rennert, P. and Heinrich, G. (1987) Cortisone, testosterone and aldosterone reduce levels of nerve growth factor messenger ribonucleic acid in L929 fibroblasts. Endocrinology 121: 1432-1437. Snider, W.D. (1988) Nerve growth factor enhances dendritic arborization of sympathetic ganglion cells in developing mammals. J. Neurosci. 8: 2628-2634. Snider, W.D. and Johnson, Jr., E.M. (1989) Neurotrophic molecules. Ann. Neurol. 26: 489-506. Snider, W.D. and Thanedar, S. (1989) Target dependence of hypoglossal motor neurons during development and in maturity. J. Comp. Neurol. 279: 489-498. Spranger, M., Lindholm, D., Bandtlow, C., Heumann, R., Gnahn, H., Naher-Noe, M. and Thoenen, H. (1990) Regulation of nerve growth factor (NGF) synthesis in the rat central nervous system: comparison between the effects of interleukin-1 and various growth factors in astrocyte cultures and in vivo. Eur. J. Neurosci. 2: 69-76. Steers, W.D., Kolbeck, S., Creedon, D. and Tuttle, J.B. (1991) Nerve growth factor in the urinary bladder of the adult regulates neuronal form and function. /. Clin. Invest. 88: 1709-1715. Stewart, S.S. and Appel, S.H. (1988) Trophic factors in neurologic disease. AnnM. Rev. Med. 39: 193-201. Stockel, K., Paravicini, U. and Thoenen, H. (1974) Specificity of the retrograde axonal transport of nerve growth factor. Brain Res. 76:413^21. Stoll, G. and Muller, H.W. (1986) Macrophages in the peripheral nervous system and astroglia in the central nervous system of rat commonly express apolipoprotein E during development but differ in their response to injury. Neurosci. Lett. 72: 233-238. Stumpf, W.E. and O'Brien, L.P. (1987) l,25(OH)2D3 sites of action in the brain: an autoradiographic study. Histochemistry 87:393-406. Stumpf, W.E., Mandhabananda, S. and Clark, S.A. (1982)

R.A. Rush et al. Brain target sites for 1,25 dihydroxyvitamin D3. Science 215: 1403-1405. Stumpf, W.E., Clark, S.A., O'Brien, L.P. and Reid, F.A. (1988) l,25(OH)2D3 sites of action in spinal cord and sensory ganglion. Anat. Embryol. Ill: 307-310. Sutula, T., He, X.X., Cavazos, J. and Scott, G. (1988) Synaptic reorganization in the hippocampus induced by abnormal functional activity. Science 239: 1147-1150. Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L. (1989) Mossy fibre synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol. 26: 3 2 1 330. Thanos, S. and Von Boxberg, Y. (1989) Factors influencing regeneration of retinal ganglion cell axons in adult mammals. Metab. Brain Dis. 4: 67-72. Thanos, S., Bahr, M., Barde, Y.A. and Vaneslow, J. (1989) Survival and axonal elongation of adult rat retinal ganglion cells. In vitro effects of lesioned sciatic nerve and brainderived neurotrophic factor. Eur. J. Neurosci. 1: 19-26. Thoenen, H. and Barde, Y.A. (1980) Physiology of nerve growth factor. Physiol. Rev. 60: 1284-1335. Thoenen, H., Angeletti, P.U., Levi-Montalcini, R. and Kettler, R. (1971) Selective induction by nerve growth factor of tyrosine hydroxylase and dopamine ^^-hydroxylase in the rat superior cervical ganglia. Proc. Natl. Acad. Sci. USA 68: 1598-1602. Thoenen, H., Barde, Y.A., Davies, A.M. and Johnson, J.E. (1987) Neurotrophic factors and neuronal death. CIBA Fndn. Symp. 126: 82-95. Thoenen, H., Zafra, P., Hengerer, B. and Lindholm, D. (1991) The synthesis of nerve growth factor and brain derived neurotrophic factor in hippocampal and cortical neurons is regulated by specific transmitter systems. Ann. NY. Acad. Sci. 640: 86-90. Timmusk, T., Palm, K., Metsis, M., Reintam, T., Paalme, V. and Saarma, M. (1993) Multiple promoters direct tissuespecific expression of the rat BDNF gene. Neuron 10: 4 7 5 489. Toso, R.D., De Bernardi, M.A., Brooker, G.. Costa, E. and Mocchetti, I. (1988) Beta-adrenergic and prostaglandin receptor activation increases nerve growth factor mRNA content in C6-2B rat astrocytoma cells. J. Pharmacol. Exp. Ther. 246: 1190-1193. Ueyama, T., Hano, T., Hamada, M., Nishio, I. and Masuyama, Y. (1991) New role of nerve growth factor - an inhibitory neuromodulator of adrenergic transmission. Brain Res. 559: 293-296. Ueyama, T., Hamada, M., Hano, T., Nishio, I., Masuyama, Y. and Furukawa, S. (1992) Nerve growth factor is increased in spontaneously hypertensive rats. J. Hypertension 10: 215-219. Ullrich, A., Gray, A., Berman, C. and Dull, T.J. (1983) Human p-nerve growth factor gene sequence highly homologous to that of mouse. Nature 303: 821-825. Vige, X., Tang, B. and Wise, B.C. (1992) Cortical neurons inhibit basal and interleukin-1-stimulated astroglial cell se-

201 cretion of nerve growth factor. Brain Res. 591: 3 4 5 350. Villegas-Perez. M.P., Vidal-Sanz, M., Bray, G.M. and Aguayo, A.J. (1988) Influences of peripheral nerve grafts on the survival and regrowth of axotomised retinal ganglion cells in adult rats. /. Neurosci. 8: 265-280. Wheeler, E.F. and Bothwell, M. (1992) Spatiotemporal patterns of expression of NGF and NGF receptor in rat embryos suggests functional roles in tissue morphogenesis and myogenesis. J. Neurosci. 12: 930-945. White, D.M., Ehrhard, P., Hardung, M., Meyer, D.K., Zimmermann, M. and Otten, U. (1987) Substance P modulates the release of locally synthesized nerve growth factor from rat saphenous nerve neuroma. Naunyn-Schmiedeberg's Arch. Pharmacol. 336: 587-590. Whittemore, S.R. and Seiger, A. (1987) The expression, localization and functional significance of beta-nerve growth factor in the central nervous system. Brain Res. 434: 4 3 9 464. Whittemore, S.R., Ebendal, T., Larkfors, L., Olsen, L., Seiger, A., Stromberg, I. and Persson, H. (1986) Development and regional expression of y3-nerve growth factor messenger RNA and protein in the rat central nervous system. Proc. Natl. Acad. Sci. USA 83: 817-821. Whittemore, S.R., Larkfors, L., Ebendal, T., Holets, V.R., Ericsson, A. and Persson, H. (1987) Increased y5-nerve growth factor messenger RNA and protein levels in neonatal rat hippocampus following specific cholinergic lesions. J. Neurosci. 7:244-251. Williams, L.R., Varon, S., Peterson, G.M., Wictorin, K., Fischer, W., Bjorklund, A. and Gage, F.H. (1986) Continuous infusion of nerve growth factor prevents basal forebrain neuronal death after fimbria-fornix transection. Proc. Natl. Acad. Sci. USA 83: 9231-9235. Wilson, P.A., Scott, J., Penschow, J., Coghlan, J. and Rush, R.A. (1986) Identification and quantification of mRNA for nerve growth factor in histological preparations. Neurosci. Utt. 64: 323-329. Wion, D., Barrad, P., Dicou, E., Scott, J. and Brachet, P. (1985) Serum and thyroid hormones T3 and T4 regulate NGF mRNA levels in mouse L cells. FEBS Lett. 189: 3 7 ^ 1 . Wion, D., Houlgatte, R. and Brachet, P. (1986) Dexamethasone rapidly reduces the expression of the y5-NGF gene in mouse L-929 cells. Exp. Cell Res. 162: 562-565. Wion, D., MacGrogan, D., Houlgatte, R. and Brachet, P. (1990) Phorbol 12-myristate 13-acetate (PMA) increases the expression of the nerve growth factor (NGF) gene in mouse L-929 fibroblasts. FEBS Lett. 262: 42-^4. Wion, D., MacGrogan, D., Neveu, I., Jehan, R., Houlgatte, R. and Brachet, P. (1991) 1,25-Dihydroxyvitamin D3 is a potent inducer of nerve growth factor synthesis. J. Neurosci. Res.lS: 110-114. Wion, D., Neveu, I., Jehan, F. and Brachet, P. (1992) Pertussis toxin provides evidence for two independent signalling pathways leading to the activation of the nerve growth factor gene. J. Neurosci. Res. 31: 294-300.

202

The regulation of nerve growth factor synthesis and delivery to peripheral neurons

Wooten, F.G., Weise, V.K. and Kopin, I.J. (1977) Norepinephrine synthesis contributes to the accumulation of catecholamines after ligation of rat sciatic nerves. Brain Res. 136: 174-177. Wyatt, S., Shooter, E.M. and Davies, A.M. (1990) Expression of the NGF receptor gene in sensory neurons and their cutaneous targets prior to and during innervation. Neuron 2: 421-^27. Yamamoto, K.R. (1985) Steroid receptor regulated transcription of specific genes and gene networks. Anna. Rev. Genet. 19: 209-215. Yamamoto, M., Sobue, G., Li, M., Arakawa, Y., Mistuma, T. and Kimata, K. (1993) Nerve growth factor (NGF), brain derived neurotrophic factor (BDNF) and low-affinity nerve growth factor (LNGFR) mRNA levels in cultured rat Schwann cells; differential time- and dose-dependent regulation by dAMP. Neurosci. Lett. 152: 37^0. Yan, Q., Elliott, J. and Snider, W.D. (1992) Brain-derived neurotrophic factor rescues spinal motor neurons from axotomy-induced cell death. Nature 360: 753-755. Yawo, H. (1987) Changes in the dendritic geometry of mouse superior cervical ganglion cells following postganglionic axotomy. J. Neurosci. 7: 3703-3711. Yoshida, K. and Gage, F.H. (1991) Fibroblast growth factors stimulate nerve growth factor synthesis and secretion by astrocytes. Brain Res. 538: 118-126. Yoshida, K. and Gage, F.H. (1992) Cooperative regulation of

nerve growth factor synthesis and secretion in fibroblasts and astrocytes by fibroblast growth factor and other cytokines. Brain Res. 569: 14-24. Zafra, F., Hengerer, B., Leibrock, J., Thoenen, H. and Lindholm, D. (1990) Activity dependent regulation of BNDF and NGF mRNAs in the rat hippocampus is mediated by non-NM DA glutamate receptors. EMBO J. 9: 3545-3550. Zafra, F., Castren, E., Thoenen, H. and Lindholm, D. (1991) Interplay between glutamate and gamma-aminobutyric acid transmitter systems in the physiological regulation of brainderived neurotrophic factor and nerve growth factor synthesis in hippocampal neurons. Proc. Natl. Acad. Sci. USA 88: 10037-10041. Zettler, C. and Rush, R.A. (1993) Elevated concentrations of nerve growth factor in heart and mesenteric arteries of spontaneously hypertensive rats. Brain Res. 614: 15-20. Zettler, C, Head, R.J. and Rush, R.A. (1991) Chronic nerve growth factor treatment of normotensive rats. Brain Res. 538:251-262. Zettler, C, Zhai, X.-F., Bridges, D. and Rush, R.A. (1995) Increased concentrations of nerve growth factor detected in tissue extracts with an improved extraction procedure. J. Neurosci. Res., in press. Zheng, M. and Heinrich, G. (1988) Structural and functional analysis of the promoter region of the nerve growth factor gene. Mol. Brain. Res. 3: 133-140.

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair 1996 Elsevier Science B.V.

CHAPTER 8

Brain-derived neurotrophic factor K.A. Bailey* Department of Pathology, University of Melbourne, Parkville, Victoria 3052, Australia

1. Introduction During the development of the vertebrate nervous system most, if not all, neuronal populations undergo a period of programmed cell death. During this period neurons depend on the availability of epigenetic factors for their survival (Barde, 1989). Nerve growth factor (NGF) was the first described neurotrophic factor that was able to support the survival of subpopulations of neurons both in vivo and in vitro (see Levi-Montalcini, 1987 for review). However, NGF does not support all neuronal populations, indicating that other neurotrophic factors must also exist. The purification and characterisation of brain-derived neurotrophic factor (BDNF), and its similarities to NGF, have led to the discovery of a family of neurotrophic factors now known as the neurotrophins. This review will attempt to cover what is known about the BDNF protein, its site of synthesis and its mechanism of action. 2. Characterisation of the BDNF protein Unlike NGF, which can be isolated in large quantities from the submandibular gland of male mice, BDNF is a protein of extremely low abundance. Its initial purification was only possible due to the availability of large amounts of starting material (pig brain) and the relative resistance of its biological activity to denaturing agents. A millionfold purification was necessary to obtain a relatively homogeneous protein (Barde et al., 1982; Present address: Department of Genetica and Developmental Biology, Monash University, Clayton, Victoria 3052, Australia.

Thoenen et al., 1983). Although the purified protein possesses many characteristics similar to those of NGF, it can be distinguished from NGF by its antigenic and functional properties. The molecular weight of purified pig BDNF as determined by SDS-gel electrophoresis is approximately 12 kDa and its isoelectric point about 10.2 (Barde et al., 1982, 1987), which are very similar to those of the monomeric NGF (Thoenen and Barde, 1980). In addition, BDNF can be characterised by its ability to support the survival of chick DRG neurons in vitro; however, in contrast to NGF, it does not support the survival of sympathetic neurons. Furthermore, anti-NGF antibodies fail to block its biological activity, indicating that the protein is clearly antigenically distinct from NGF (Thoenen etal., 1983). The subsequent cloning of the porcine, murine and human BDNF genes allowed the primary structure of the trophin to be determined (Leibrock et al., 1989; Hofer et al., 1990; Rosenthal et al., 1991). The gene encodes for a secreted protein, with the mature protein possessing 119 amino acids. The calculated molecular mass and isoelectric point of the recombinant form are very similar to those obtained by electrophoretic methods on the purified protein (Barde et al., 1982). A striking feature of the primary structure of mature BDNF is its similarity to NGF. There are 51 amino acids common to the various NGF's and to BDNF, including all six cysteine residues (Leibrock et al, 1989). This homology allowed for the design of degenerate primers to search for related proteins using the polymerase chain reaction which lead to the subsequent identification of neurotrophin-3 (NT-3) (Ernfors et al., 1990a; Hohn et al., 1990;

204

Maisonpierre et al., 1990b; Rosenthal et al., 1990) neurotrophin-4 (NT-4) (Hallbook et al., 1991; Ip et al., 1992), neurotrophin-5 (NT-5) (Berkemeier et al., 1991) and neurotrophin-6 (NT-6) (Gotz et al., 1994) (see chapter by Rocamora and Arenus). As stated earlier, BDNF exists as a protein of low abundance and its limited availability has hindered studies on its biophysical characterisation. Following the cloning of the BDNF gene and the availability of recombinant BDNF, however, it has been possible to further characterise the protein and compare the results to those known for NGF. As was found for NGF, BDNF exists primarily as a homodimer (Radziejewski et al., 1992; Narhi et al., 1993), as does NT-3. Like NGF, BNDF contains a substantial ^-sheet with little or no a-helix. Therefore, it is most likely that all the neurotrophins have a three-dimensional structure similar to that described for NGF (McDonald et al., 1991). Recently, it has been shown that all the neurotrophins can exist both in vivo and in vitro as heterodimers as well as homodimers. The BDNF/NT3 heterodimer can induce tyrosine phosphorylation of receptors, but appears to have a lower specific biological activity than either homodimer alone (Radziejewski and Robinson, 1993; Arakawa et al, 1994; Jungbluth et al., 1994; Philo et al., 1994). A heterodimer between NGF and BDNF can be formed, but it is not stable (Arakawa et al., 1990). 3. Neuronal populations responsive to BDNF 3. L Peripheral nervous system Within the peripheral nervous system, both NGF and BDNF are able to support the survival of dorsal root ganglion (DRG) sensory neurons. However the populations supported by each factor in vitro is distinct, as a combination of the two factors has an additive effect on survival numbers (Davies et al., 1986; Lindsay et al., 1985; Barde, 1989). In contrast to NGF, BDNF is able to support the survival of sensory neurons of neural placode origin. These neurons, which include those of the nodose ganglia, are unresponsive to NGF but show cell survival and neurite outgrowth in re-

Brain-derived neurotrophic factor

sponse to BDNF (Lindsay et al., 1985). Furthermore the responsiveness of specific sensory populations to BDNF is tightly regulated, with different populations of neurons becoming BDNFdependent at different stages of development (Vogel and Davies, 1991). However, no survival effects have been observed on sympathetic or ciliary ganglion neurons (Lindsay et al., 1985). Apart from the cell survival effects, BDNF appears to have a role in the maturation of sensory neurons from neural precursor cells. In cultures of neural crest cells, the addition of BDNF leads to an increase in the number of sensory neurons (SieberBlum, 1991). This is not due to an increase in the total number of cells but to a corresponding decrease in the proportion of undifferentiated cells. Likewise, in cultures of early sensory neurons that are not yet neurotrophin responsive, the addition of BDNF causes an acceleration of their maturation (Wright et al., 1992). Thus, BDNF may be involved with cell differentiation at early stages of development and with cell survival at later stages. Sensory neurons require BDNF in vivo as well as in vitro. During normal development, sensory neurons undergo a period of naturally occurring cell death. The treatment of embryos with exogenous BDNF during this period results in a substantial reduction in the number of dying neurons in both the DRG and the nodose ganglion (Hofer and Barde, 1988; Oppenheim et al., 1992). The opposite effect is observed if embryos are deprived of BDNF during this period. Gene targeting experiments have resulted in the generation of mice that lack either BDNF (Ernfors et al., 1994; Jones et al., 1994) or its receptor, trkE (Klein et al., 1993). These animals show extensive degeneration in sensory ganglia, indicating that BDNF is necessary for the normal development of these neuronal populations in vivo. 5.2. Central nervous system Initial studies using retinal cultures suggested that BDNF supported subpopulations of central nervous system (CNS) neurons (Johnson et al., 1986; Rodriguez-Tebar et al., 1989). Although BDNF had no effect on the overall numbers of

205

K.A. Bailey

retinal cells, it did stimulate survival and neurite outgrowth of the retinal ganglion population (Johnson et al., 1986). Further evidence of BDNF activity on CNS neurons was obtained from treating cultured septal cholinergic neurons with BDNF. It was already known that NGF supports the survival of these neurons in vitro and increases the levels of the cholinergic enzymes choline acetyltransferase (ChAT) and acetylcholinesterase (AChE) (Hartikka and Hefti, 1988; Hatanaka et al., 1988). BDNF had a similar effect on these neurons, with an increase in cell survival and an increase in ChAT and AChE levels (Alderson et al., 1990). A combination of the two factors had no additive effect on cell survival, indicating that they probably act on the same populations, although the combination did have an additive effect on ChAT activity. As expected from the wide distribution of BDNF mRNA within the CNS, BDNF has effects on a number of neuron populations. When added to cultures of cerebellar granule cells, BDNF enhances survival and stimulates neurite sprouting (Lindholm et al., 1993). However the effect is predominantly on early granule neurons, whereas more mature neurons respond only to NT-3. This suggests that BDNF may promote the initial development, while NT-3 promotes the subsequent maturation of these cells (Segal et al., 1992). In contrast to granule cells, BDNF has no effect on the survival of hippocampal neurons in vitro, but these cells respond to the factor with activation of signal transduction mechanisms and an increase in the number of cells that exhibit AChE (Ip et al., 1993; Ohsawa et al., 1993). Thus BDNF can have either a survival effect and/or a differentiation effect, depending on the target cell. Two other CNS neuronal populations that respond to BDNF are dopaminergic neurons and somatic motor neurons. These effects are of particular interest due to the degeneration of dopaminergic and motor neurons in Parkinson's disease and motor neuron disease respectively. Addition of BDNF to cultures of mesencephalic dopaminergic neurons results in enhanced survival and an increase in dopamine uptake (Hyman et al., 1991; Kniisel et al, 1991; Beck et al., 1993). Further-

more, BDNF reduces the cytotoxicity of the neurotoxic agent l-methyl-4-phenylpyridinium (MPP+) on these cells (Hyman et al., 1991). Cultured motor neurons respond to BDNF with increased cell survival and an increase in ChAT activity (Henderson et al., 1993; Kato and Lindsay, 1994). BDNF also supports motor neuron survival in vivo, as addition of exogenous BDNF to a developing chick embryo substantially reduces the naturally occurring neuronal cell death in this pool (Oppenheim et al., 1992). 4. The BDNF gene 4.1. Primary structure As mentioned in Section 2, the BDNF protein was initially purified from pig brain. A partial amino acid sequence was obtained from the Nterminus and fragments purified by protease cleavage. Using this sequence, Leibrock et al. (1989) synthesised oligonucleotide primers to amplify a pig genomic template, using the polymerase chain reaction (PCR). The product was sequenced and used to design further sets of primers to amplify the cDNA encoding for BDNF. The primary structure of pig BDNF contains an open reading frame coding for a protein of 252 amino acids (Leibrock et al., 1989). This is made up of a signal peptide of 18 amino acids, a pro- sequence of 115 amino acids and the mature protein of 119 amino acids. It is thought that the pro-sequence is involved in the correct folding of the mature protein. The BDNF protein displays an extreme degree of evolutionary conservation compared to other secreted factors, including NGF (Fig. 1). The 119 amino acids that constitute the mature BDNF are identical in pig, mouse, rat, monkey and human (Hofer et al., 1990; Isackson et al, 1991b; Maisonpierre et al., 1991, 1992; Oz^elik et al., 1991), while the chicken BDNF protein has only six amino acid differences compared to the mammalian version (Isackson et al., 1991b; Maisonpierre et al., 1992; Herzog et al., 1994). Even between such diverse species as Xenopus and Xiphophorus (fish) there is a high percentage of conserved residues (Isackson et al., 1991b; Gotz et al, 1992).

206 Mammalian^ Chick2 Fish^ Xenopus^

Brain-derived neurotrophic factor BDNF BDNF BDNF BDNF

H S D P A R R G E L S V C D S I S E W V T A A D K K T A V D T E . . . . S Q . . . . V I . N

Mammalian Chick Fish Xenopus

BDNF BDNF BDNF BDNF

M S G G T V T V L E K V P V S K G Q L K Q Y F Y E T K C . . . A P . . . Q . . . . M PN . . . A

Mammalain Chick Fish Xenopus

BDNF BDNF BDNF BDNF

M G Y T K E G C R G I D K R H W N S Q C R T T Q S Y V R A L K D YT . . . M E . . Y F

Mammalian Chick Fish Xenopus

BDNF BDNF BDNF BDNF

T M D S K K R I G W R F I R I D T S C . . . N K K V

N P

V C T L T I K R G R

Fig. 1. Comparison of the BDNF protein from different species. The amino acid sequences of BDNF from different species are aligned to the mammalian sequence. Sequence identities are indicated by a dot (•). The cysteine residues conserved between all the neurotrophins are indicated in bold. (References: 1, Hofer et al. (1990), Isackson et al. (1991b), Maisonpierre et al. (1991, 1992), Oz9elik et al. (1991); 2, Herzog et al. (1994), Isackson et al. (1991b), Maisonpierre et al. (1992); 3, Gotz et al. (1992); 4, Isackson etal. (1991b).

4.2. Promoter region The rat BDNF gene consists of five exons that span more than 40 kb of genomic DNA. The four 5' exons are non-coding (Ohara et al., 1992; Metsis et al., 1993; Timmusk et al., 1993, 1995; Nakayama et al., 1994) whereas the 3' exon encodes for the entire preproBDNF protein. It is thought that eight different mRNAs are generated from the four 5' exons and two alternative polyadenylation sites within the 3' exon. In addition, there is evidence to suggest that exon 2 has internal splicing donor sites that can give rise to three different transcripts (Ohara et al., 1992; Nakayama et al., 1994). The BDNF gene is separated into three regions: exons I and II, exons III and IV and exon V. Exons I and II are located at the most distal position and are separated by an intron spanning less than 1 kb. Exons III and IV are also separated from each other by an intron of less than 1 kb (Timmusk et al., 1993; Nakayama et al, 1994). However, the three regions are separated by introns spanning greater than 15 kb. The sequences upstream of the

four 5' exons all contain promoter regions which direct tissue specific expression of the various transcripts (Timmusk et al., 1993, 1994). Furthermore, there is differential regulation of the promoters by neuronal activity (Falkenberg et al., 1993; Metsis et al., 1993; Timmusk and Metsis, 1994) 5. BDNF receptors 5.7. Low-affinity receptor Studies with NGF identified the presence of two classes of receptors based on binding and kinetic measurements (Sutter et al., 1979). Similar results are obtained from binding studies with BDNF on embryonic sensory neurons. The highaffinity receptor has a dissociation constant (K^) of 1.7 X 10"^^ M whereas the K^ for the low-affinity receptor is 1.3 x 10"^ M (Rodriguez-Tebar and Barde, 1988). The presence of the high-affinity binding site is necessary for the functional response to the neurotrophin, whereas low-affinity receptors are also found on cells that do not re-

K.A. Bailey

spond to BDNF (Rodriguez-Tebar and Barde, 1988). It is now known that low-affinity binding is due to the low-affinity NGF receptor (pTS^GFR) whereas high-affinity binding involves the trkB tyrosine kinase receptor. Binding studies show that p75NGFR ^Qi^g as ^j^^ low-affinity receptor for NGF and NT-3 as well as BDNF (Rodriguez-Tebar et al., 1990, 1992). The gene coding for pTS^GFR has been cloned and encodes the sequence for a transmembrane protein with a relatively short cytoplasmic domain, containing none of the structural motifs known to function in signal transduction (Johnson et al„ 1986; Radeke et al., 1987). The receptor belongs to the tumour necrosis factor receptor superfamily. Members of this family include diverse cell surface proteins such as the Fas antigen, the T cell antigen OX40, and the B cell antigens CD30 and CD40, as well as TNFRI and TNFRII (Stamenkovic et al., 1989; Mallet et al., 1990; Smith et al, 1990; Itoh et al., 1991; Durkop et al., 1992). The family of proteins exhibits homology in the cysteine-rich repeats in the extracellular domain. Although the function of p75^^^^ remains unclear there are several putative roles for this receptor. It may facilitate signalling through the Trk receptors (Barker and Shooter, 1994; Verdi et al., 1994). It may also be involved in retrograde transport of neurotrophins (Johnson et al., 1987) or aid in their concentration (Taniuchi et al., 1988). So far, evidence for these possible roles for p75^G^^ has only been demonstrated for NGF. There is still considerable debate whether this receptor has any functional relevance for BDNF binding in vivo, despite its ability to bind BDNF in vitro. Studies from gene targeting experiments show that p75NGFR_j^yll mice have defects in NGF-responsive neurons but not in those neurons supported by the other neurotrophins (Lee et al, 1992, 1994a, 1994b; Daviesetal., 1994). 5.2. High-affinity receptor The high-affinity receptors for all neurotrophins involve the Trk family of receptor kinases. Three members of this family have been described in mammalian cells: trkA binds preferentially to NGF

207

but can bind NT-3; trkB is the receptor for BDNF and NT-4/5, although it can bind NT-3; trkC is the receptor for NT-3. The trkB gene encodes for multiple mRNA transcripts which encode for either full-length (gpMS^''^^) or truncated trkR receptors (gp95^'*^^) (Klein et al., 1990a, 1991; Barbacid et al., 1991; Middlemas et al., 1991). The truncated trkB receptors have the same extracellular and transmembrane domains as the full length receptor, but differ in their cytoplasmic tails and lack the tyrosine kinase catalytic domain (Barbacid et al., 1991). These truncated proteins are therefore unable to activate signal transduction. Furthermore, gpMS^'^^^ and gp95^'^^^ are expressed in functionally distinct structures within the CNS, indicating that they probably have distinct functions (Barbacid et al., 1991). Transcripts encoding for trk^ mRNA are expressed predominantly in the brain, although significant levels of expression are also observed in the lung, muscle and ovaries (Klein et al., 1989). During development, transcripts can be detected from very early stages of embryogenesis in cells of both neuroepithelial and neural crest origins (Klein et al., 1990b; Barbacid et al., 1991). Within the adult brain, trkB mRNA expression is observed in many different regions, including the cerebrum and the cerebellum (Klein et al, 1989, 1990b). Several studies have shown that rrfcB is a functional receptor for BDNF. Most of these studies have used NIH 3T3 fibroblasts or PCI2 cells expressing the rrfcB protein. The addition of BDNF to 3T3 fibroblast cells induces rapid phosphorylation of gpl45^^^B (Klein et al., 1991; Soppet et al, 1991). Similarly, expression of trkB in PCI2 cells results in a biological response of these cells to BDNF similar to that observed for normal PC 12 cells exposed to NGF, including differentiation of the cells and extensive neurite outgrowth (Squinto et al., 1991). Thus, BDNF is able to bind to trkB and mediate a signal transduction response. Furthermore, the lack of p75NGFR in NIH 3T3 cells indicates that BDNF can activate trk& in the absence of the low-affinity receptor, suggesting that this protein is not required for BDNF action on responsive cells (Glass et al., 1991). However the K^ for binding of BDNF to fritB in these fibroblasts is approximately 10~^ M, which

208

would normally suggest binding to a low-affinity receptor (Soppet et al., 1991). It is still unclear, therefore, whether gpHS''^^^ can act independently as the high-affinity receptor, or whether it acts in association with pTS^^^*, or some other as yet unidentified protein. 6. Expression of BDNF mRNA 6.1. Normal distribution Due to the very low amounts of BDNF protein within tissues and the absence of good antibodies, it has not been possible to determine BDNF levels within various tissues. All of the evidence documenting the regional distribution of BDNF has involved studies of mRNA synthesis. BDNF mRNA appears as two transcripts on Northem blots (Hofer et al., 1990; Maisonpierre et al, 1990a); these transcripts run at approximately 1.4 and 4.2 kb and are probably due to the use of two alternative polyadenylation signals (Timmusk et al, 1993). Initial studies suggested that BDNF mRNA expression was limited to the CNS (Leibrock et al., 1989); however, later studies found that it is also expressed, although at lower levels, in other tissues such as heart, lung and platelets (Maisonpierre et al., 1990a; Yamamoto and Gurney, 1990; Schecterson and Both well, 1992). Within the adult CNS the highest levels are found in the hippocampus, although most brain areas express BDNF mRNA (Emfors et al., 1990b; Hofer et al., 1990; Phillips et al., 1990; Wetmore et al., 1990). In situ hybridisation studies correlate well with the Northem blot analyses showing high levels of mRNA within the hippocampus and cerebral cortex and lower levels in most other brain regions. Further studies have localised the mRNA to neurons in and around the pyramidal layer of the hippocampus, as well as in the granule layer and in the hilus of the dentate gyrus (Emfors et al., 1990b; Hofer et al, 1990; Wetmore et al., 1990; Friedman et al., 1991). 6.2. Regulation of mRNA synthesis The high levels of BDNF mRNA in the hip-

Brain-derived neurotrophic factor

pocampus suggest that neuronal activity may be involved in the regulation of BDNF synthesis. In support of this theory, studies on cultured hippocampal neurons showed that depolarisation with high potassium results in an increase in the level of BDNF mRNA (Zafra et al., 1990). These results prompted analysis of the effects of a large range of transmitter substances on BDNF mRNA levels. Of all the substances tested, the glutamate receptor agonist kainate produced the largest increase in mRNA levels (Zafra et al., 1990). The upregulation of BDNF mRNA levels by kianic acid occurs both in vivo and in vitro (Zafra et al., 1990; Ballarin et al., 1991; Dugich-Djordjevic et al., 1992a; Wetmore et al., 1994). The greatest increases in mRNA levels are in the hippocampus, particularly the dentate gyrus, although significant increases of mRNA levels are also observed in the superficial layers of the cortex and in the piriform cortex (Dugich-Djordjevic et al., 1992a; Ernfors et al., 1991). In addition, BDNF protein immunoreactivity is increased. The protein displays a spatial and temporal course distinct from that seen for the expression of BDNF mRNA, suggesting a constitutive release of BDNF (Wetmore et al., 1994). As is the case with NGF, BDNF mRNA levels are upregulated not only by kianic acid but by various forms of insult to the brain. Kindling, induced by electrical stimulation or drugs, results in a dramatic increase in BDNF mRNA in the dentate gyrus, parietal cortex and piriform cortex, as well as in the pyramidal layer of the hippocampus and in the amygdaloid complex (Emfors et al., 1991; Bengzon et al, 1993; Humpel et al., 1993). Similarly, lesion-induced limbic seizures or transient forebrain ischemia upregulates BDNF mRNA in many regions, with the highest levels observed in the granule cell layer of the dentate gyms (Gall et al., 1991; Isackson et al., 1991a; Lindvall et al., 1992; Rocamora et al., 1992; Takeda et al., 1993). Even needle insertion or saline injection causes a rapid transient increase of BDNF mRNA in granular neurons of the dentate gyms and the piriform cortex (Ballarin etal., 1991). Glutamate is the major excitatory neurotransmitter in the mammalian brain and its effects are mediated through at least three receptor subtypes.

KA. Bailey

These receptors have been classified pharmacologically based on their selective ligands, NMDA, kainate and quasiqualate, although the last two receptors are often grouped together as nonNMDA receptors (Watkins et al., 1990). In the adult brain, the upregulation of BDNF mRNA resulting from kianic acid treatment or neuronal damage is mediated by non-NMDA glutamate receptors and the effect can be blocked by the administration of non-NMDA antagonists (Zafra et al., 1990, 1991; Ernfors et al., 1991; Lindvall et al., 1992; Wetmore et al., 1994). NMDA antagonists have no effect on the upregulation and basal levels of BDNF mRNA are not affected by NMDA (Zafra et al., 1990; Wetmore et al., 1994). In contrast to the effects of glutamate, activation of the y-aminobutyric acid (GABA)ergic system downregulates BDNF mRNA levels (Zafra et al., 1991). Thus, BDNF mRNA synthesis is regulated by a balance between glutamatergic and GABAergic systems. The upregulation of BDNF mRNA by kianic acid in the adult brain contrasts with results obtained in the developing brain. Administration of kianic acid during the early postnatal period induces seizures, but these seizures are not associated with elevated BDNF mRNA levels (DugichDjordjevic et al., 1992b). There does, however, appear to be a regulation of BDNF mRNA synthesis by the cholinergic system during development. Muscurinic agonists increase rat hippocampal BDNF mRNA levels in the early postnatal period as well as in the adult. This upregulation of mRNA can be inhibited by muscarinic antagonists and by MK-801, a non-competitive antagonist of NMDA receptors (Berzaghi et al., 1993; Knipper et al, 1994). Thus, the activity-dependent regulation of BDNF mRNA levels during early postnatal development is mediated through the cholinergic system, via NMDA receptors. The mRNA and protein levels of the BDNF receptor trkB are also increased following brain insults (Bengzon et al., 1993; Rumpel et al., 1993; Merlio et al., 1993). Increased trkB levels are observed in the granule cell layer of the dentate gyrus following a single stimulus-evoked seizure, whereas multiple stimulations lead to elevated

209

levels throughout the pyramidal layer of the hippocampus, the hilar region and the piriform cortex in addition to the dentate gyrus (Merlio et al., 1993). However, trkB mRNA encoding for the full-length receptor (containing the tyrosine kinase domain) is only elevated in the dentate gyrus. Elevation of trkB mRNA levels in other regions results in an increase in truncated receptors that lack the tyrosine kinase domain (Merho et al., 1993). 7. Experimental actions of BDNF on neurodegenerative disease models 7.7. Alzheimer's disease In Alzheimer's disease (AD), the basal forebrain cholinergic system undergoes severe degenerative changes with neuronal loss (Whitehouse et al., 1982) and it has been suggested that agents which maintain the function of these neurons may be used to slow the associated deterioration of cognitive function. NGF has been shown to support the survival of basal forebrain and septal cholinergic neurons both in vivo and in vitro (Hefti, 1986; Gage et al, 1988; KoHatsos et al., 1990). Furthermore, NGF is able protect these neurons against lesion-induced degenerative changes, suggesting that NGF may be useful as a therapeutic agent for AD. BDNF can also support the survival of cholinergic neurons; however, its effect on lesioned cholinergic neurons appears less pronounced than that of NGF (Morse et al, 1993; Dekker et al., 1994; Kohatsos et al., 1994; Skup et al., 1994). For example, one study found that BDNF prevents the lesion-induced death of p75NGFR positive cells but that these cells still lost ChAT activity (Widmer et al, 1993). Although NGF supports cholinergic neurons in degenerative models, there is no evidence to indicate that Alzheimer's disease is associated with a deficiency of NGF. By contrast, BDNF mRNA levels are known to be decreased in AD brains compared with age-matched controls (Phillips et al., 1991; Murray et al., 1994). Normal aging produces no changes in BDNF mRNA levels (Lapchak et al., 1993), indicating that the results obtained from AD brains are not due to age-related neurodegeneration. The decrease in BDNF mRNA

210

has been attributed to down-regulation of mRNA expression rather than to cell death, as the level of calcium/calmodulin-dependent protein kinase type n mRNA is increased in the same neuronal populations (Murray et al., 1994). Despite these observations, it remains unclear whether lack of BDNF contributes to the progression of cell death in AD or whether the protein could offer any therapeutic benefits in the treatment of AD. 7.2. Parkinson's disease Degeneration of the dopaminergic neurons of the substantia nigra is responsible for the majority of the behavioural deficits associated with Parkinson's disease. Thus, factors that support the survival of these dopaminergic neurons may offer putative therapeutic benefits in the treatment of the disease. BDNF supports survival of dopaminergic neurons in embryonic mesencephalic cultures (Hyman et al., 1991) and reduces the susceptibility of these cells to the neurotoxic effects of 6hydroxydopamine (6-OHDA) and MPP+ (Hyman et al., 1991; Beck et al., 1992; Spina et al, 1992; Skaper et al., 1993). Furthermore, the protective effects of BDNF are enhanced in the presence of the ganglioside GMl, although the gangUoside has no effect by itself (Fadda et al., 1993; Skaper et al., 1993). The exact mechanisms mediating the protective effect of BDNF remain unclear: however, BDNF may regulate protective mechanisms against oxidative stress. For example, the levels of oxidised glutathione are increased by 6-OHDA and this can be prevented by BDNF treatment, possibly as a result of increased activity of glutathione reductase (Spina et al., 1992). Stimulation of synthesis of a protein is compatible with the observation that the protective effect of BDNF needs several days to develop. Combined treatment of neurons with MPP+ and BDNF shows no attenuation of the toxic effects of MPP+, further confirming that the protective mechanisms invoked by BDNF involves biochemical or structural changes (Beck et al., 1992; Beck, 1994). Chronic infusion of 6-OHDA into one half of the rat neostriatum in vivo causes a partial dopaminergic lesion which results in rotation of the

Brain-derived neurotrophic factor

animal towards the lesioned hemisphere after treatment with the dopamine releasing drug amphetamine. Infusion of BDNF into the substantia nigra of these rats before and during the lesioning decreases the number of ipsiversive rotations and increases the number of contraversive rotations (Altar et al., 1994). Furthermore, there is decreased dopamine turnover in the infused substantia nigra compared to the contralateral side and increased 5HT turnover in the striatum in both hemispheres, consistent with a concomitant activation of the dopamine and 5-HT systems (Martin-Iverson et al., 1994; Schults et al, 1994). In addition, BDNF is able to increase nigral dopaminergic neuronal survival in vivo following MPP+ treatment, indicating that it protects against neurotoxicity both in vivo and in vitro (Frim et al., 1994). However, intraventricular administration of BDNF does not reduce the axotomy-induced degeneration of dopaminergic neurons that follows transection of the medial forebrain bundle (Kniisel et al., 1992). This suggests that there are differences in the ability of BDNF to support dopaminergic neuronal survival in vivo and in vitro. 7.3. Motorneurondiseases Although initial studies suggested that the neurotrophins were unable to support the survival of cultured motor neurons (Arakawa et al., 1990), there is now substantial evidence to show that several neurotrophins, including BDNF, support motor neuron survival in vivo and in vitro. BDNF prevents the death of cultured motor neurons (Henderson et al., 1993) and addition of exogenous BDNF to the developing embryo rescues motor neurons from naturally occurring cell death (Oppenheim et al., 1992). Furthermore, BDNF mRNA is present in the spinal cord and limb bud during development, while trkB mRNA is present in motor neurons (Henderson et al., 1993; Koliatsos et al., 1993; Yan et al., 1993). In addition, BDNF is retrogradely transported to motor neurons (DiStefano et al., 1992; Yan et al., 1993). These results suggest a role for BDNF in the normal development of motor neurons. BDNF also has a role in the prevention of in-

K.A. Bailey

duced motor neuron cell death following axotomy or deafferentation. During the embryonic and neonatal period, axotomy or deafferentation results in the loss of motor neurons: in the adult animal, by contrast, these procedures result in a loss of ChAT activity but do not affect cell numbers (Schmalbruch, 1984; Snider and Thanedar, 1989; Armstrong et al., 1991). Addition of exogenous BDNF prevents the induced motor neuron death in neonatal animals (Oppenheim et al., 1992; Sendtner et al., 1992; Yan et al., 1992, 1993; Koliatsos et al., 1993; Li et al, 1994; Chiu et al., 1994; Clatterbuck et al., 1994); however, its effect on ChAT activity remains unclear. Some studies have found no effect in reversing the reduction in ChAT activity, in either neonatal or adult animals (Yan et al., 1993; Clatterbuck et al., 1994). Other studies, however, have found a reduced loss of ChAT expression (Chiu et al, 1994; Friedman et al., 1995). Furthermore, at least one study has shown that BDNF increases ChAT activity in cultures of enriched motor neurons and that co-treatment with BDNF and NT-3 has an additive effect, suggesting that the effects are mediated through both trkB and trkC receptors (Wong et al., 1993; Kato and Lindsay, 1994). The above-mentioned studies indicate that BDNF supports motor neuron survival during development and following axotomy and may therefore be of therapeutic use in the treatment of degenerative diseases of motor neurons. This idea is further supported by a study showing that pTS^^^^ and trkB mRNA are upregulated in motor neurons from patients with amyotrophic lateral sclerosis (Seeburger et al., 1993). In addition, another study has shown that a combination of BDNF and CNTF is able to arrest the progression of motor neuron dysfunction in the wobbler mouse model (Mitsumoto et al., 1994). The addition of either BDNF or CNTF alone was only able to slow progression, suggesting that the factors synergise to promote neuronal survival or, alternatively, that they support different subpopulations of motor neurons. Despite all these studies, gene targeting experiments suggest that BDNF is not essential for motor neuron development although its receptor, trkB, is

211

essential. Thus, mice homozygous for a germline mutation in the trkB gene show deficiencies in the motor neuron population (Klein et al., 1993), while mice lacking the BDNF gene have normal motor neuron development (Emfors et al., 1994; Jones et al., 1994). Initially it was thought that this difference was due to NT-4 supporting motor neuron survival in the absence of BDNF, as NT-4 is also a ligand for the trkB receptor. However, it has recently been shown that motor neurons remain unaffected in mice which lack both BDNF and NT-4, suggesting that another ligand may act on trkB in vivo (Conover et al., 1995; Liu et al., 1995). 8. Conclusions From the studies described above, it is clear that BDNF supports the survival of a wide range of neurons both during development and in the mature animal. Gene targeting studies provide clear evidence that BDNF and its receptor trkB is essential for the normal development of subpopulations within the nervous system. The upregulation of BDNF mRNA following neuronal injury indicates that its synthesis can be regulated in response to extemal stimuli and suggests that BDNF plays a role in neuronal protection and maintenance. However, despite all these studies there is still much that remains to be leamed about the physiological actions of BDNF. The availability of high-affinity antibodies against BDNF would enable comparisons to be made between protein levels and mRNA levels. These types of studies are necessary to determine if an increase in BDNF mRNA correlates with an increase in the protein. In addition, further studies are required to determine the physiological role of the truncated trkB receptor and any interaction between the truncated and the full length receptor. Acknowledgements I wish to thank Stephanie Fuller and Victor Nurcombe for critical reading of the manuscript. K.A.Bailey is supported by grants from the Bethlehem-Griffiths Research Foundation and the Motor-Neurone Association of Australia.

212

References Alderson, R.F., Alterman, A.L., Barde, Y.-A. and Lindsay, R.M. (1990) Brain-derived neurotrophic factor increases survival and differentiated functions of rat septal cholinergic neurons in culture. Neuron 5: 297-306. Altar, C.A., Boylan, C.B., Fritsche, M., Jones, B.E., Jackson, C, Wiegand, S.J., Lindsay, R.M. and Hyman, C. (1994) Efficacy of brain-derived neurotrophic factor and neurotrophin-3 on neurochemical and behavioral deficits associated with partial nigrostriatal dopamine lesions. J. Neurochem. 63: 1021-1032. Arakawa, T., Haniu, M., Narhi, L.O., Miller, J.A., Talvenheimo, J., Philo, J.S., Chute, H.T., Matheson, C, Carnahan, J., Louis, J.C, Yan, Q., Welcher, A.A. and Rosenfeld, R. (1994) Formation of heterodimers from three neurotrophins, nerve growth factor, neurotrophin-3, and brainderived neurotrophic factor. J. Biol Chem. 269: 2783327839. Arakawa, Y., Sendtner, M. and Thoenen, H. (1990) Survival effect of ciliary neurotrophic factor (CNTF) on chick embryonic motor neurons. J. Neurosci. 10: 3507-3515. Armstrong, D.M., Brady, R., Hersh, L.B., Hayes, R.C. and Wiley, R.G. (1991) Expression of choline acetyl-transferase and nerve growth factor receptor within hypoglossal notoneurons following nerve injury. /. Comp. Neurol. 304: 596607. Ballarin, M., Ernfors, P., Lindefors, N. and Persson, H. (1991) Hippocampal damage and kianic acid injection induce a rapid increase in mRNA for BDNF and NGF in the rat brain. Exp. Neurol. 114: 35—43. Barbacid, M., Lamballe, F., Pulido, D. and Klein, R. (1991) The trk family of tyrosine protein kinase receptors. Biochim. Biophys. Acta 1072: 115-127. Barde, Y.-A. (1989) Trophic factors and neuronal survival. Neuron 2:1525-1534. Barde, Y.-A., Davies, A.M., Johnson, J.E., Lindsay, R.M. and Thoenen, H. (1987) Brain derived neurotrophic factor. Prog. Brain Res. 71: 185-189. Barde, Y.-A., Edgar, D. and Thoenen, H. (1982) Purification of a new neurotrophic factor from mammalian brain. EMBO J. 1:549-553. Barker, P.A. and Shooter, E.M. (1994) Disruption of NGF binding to the low affinity neurotrophin receptor p75LNTR reduces NGF binding to Trk A on PCI 2 cells. Neuron 13: 203-215. Beck, K.D. (1994) Functions of brain-derived neurotrophic factor, insulin-like growth factor-1 and basic fibroblast growth factor in the development and maintenance of dopaminergic neurons. Prog. Neurobiol. 44: 497-516. Beck, K.D., Knusel, B., Winslow, J.W., Rosenthal, A., Burton, L.E., Nikolics, K. and Hefti, F. (1992) Pretreatment of dopaminergic neurons in culture with brain-derived neurotrophic factor attenuates toxicity of l-methyl-4pyridinium. Neurode generation 1: 27-36. Beck, K.D., Knusel, B. and Hefti, F. (1993) The nature of the

Brain-derived neurotrophic factor trophic action of brain derived neurotrophic factor, des (13)-insulin like growth factor I and basic fibroblast growth factor on mesencephalic dopaminergic neurons developing in culture. Neuroscience 52: 855-866. Bengzon, J., Kokaia, Z., Ernfors, P., Kokaia, M., Leanza, G., Nilsson, O.G., Persson, H. and Lindvall, O. (1993) Regulation of neurotrophin and trkA, trkB and trkC tyrosine kinase receptor messenger RNA expression in kindling. Neuroscience 53: 433^46. Berkemeier, L.R., Winslow, J.W., Kaplan, D.R., Nikolics, K., Goeddel, D.V. and Rosenthal, A. (1991) Neurotrophin-5: a novel neurotrophic factor that activates trk and trkB. Neuron 7: 857-866. Berzaghi, M.D., Cooper, J., Castren, E., Zafra, F., Sofroniew, M., Thoenen, H. and Lindholm, D. (1993) Chohnergic regulation of brain derived neurotrophic factor (BDNF) and nerve growth factor (NGF) but not neurotrophin 3 (NT 3) messenger RNA levels in the developing rat hippocampus. J. Neurosci. 13:3818-3826. Chiu, A.Y., Chen, E.W. and Loera, S. (1994) Distinct neurotrophic responses of axotomized motor neurons to BDNF and CNTF in adult rats. NeuroReport 5: 693-696. Clatterbuck, R.E., Price, D.L. and Koliatsos, V.E. (1994) Further characterization of the effects of brain-derived neurotrophic factor and ciliary neurotrophic factor on axotomized neonatal and adult mammalian motor neurons. J. Comp. Neurol. 342: 45-56. Conover, J.C, Erickson, J.T., Katz, D.M., Bianchi, L.M., Poueymirou, W.T., McClain, J., Pan, L., Helgren, M., Ip, N.Y., Boland, P., Friedman, B., Wiegand, S., Vejsada, R., Kato, A.C., DeChiara, T.M. and Yancopoulos, G.D. (1995) Neuronal deficits, not involving motor neurons, in mice lacking BDNF and/or NT4. Nature 375: 235-238. Davies, A.M., Thoenen, H. and Barde, Y.-A. (1986) The response of chick sensory neurons to brain-derived neurotrophic factor. /. Neurosci. 6: 1897-1904. Davies, A.M., Lee, K.F. and Jaenisch, R. (1994) p75Deficient trigeminal sensory neurons have an altered response to NGF but not to other neurotrophins. Neuron 11: 565-574. Dekker, A.J., Fagan, A.M., Gage, F.H. and Thai, L.J. (1994) Effects of brain-derived neurotrophic factor and nerve growth factor on remaining neurons in the lesioned nucleus basalis magnocellularis. Brain Res. 639: 149-155. DiStefano, P.S., Friedman, B., Radziejewski, C, Alexander, C, Boland, P., Schick, CM., Lindsay, R.M. and Wiegand, S.J. (1992) The neurotrophins BDNF, NT-3, and NGF display distinct patterns of retrograde axonal transport in peripheral and central neurons. Neuron 8: 983-993. Dugich-Djordjevic, M.M., Tocco, G., Lapchak, P.A., Pasinetti, G.M., Najm, I., Baudry, M. and Hefti, F. (1992a) Regionally specific and rapid increases in brain-derived neurotrophic factor messenger RNA in the adult rat brain following seizures induced by systemic administration of kianic acid. Neuroscience 47: 303-315. Dugich-Djordjevic, M.M., Tocco, G., Willoughby, D.A.,

KA. Bailey Najm, I., Pasinetti, G., Thompson, R.F., Baudry, M., Lapchak, P.A. and Hefti, F. (1992b) BDNF mRNA expression in the developing rat brain following kianic acid-induced seizure activity. Neuron 8: 1127-1138. Diirkop, H., Latza, U., Hummel, M., Eitelbach, F., Seed, B. and Stein, H. (1992) Molecular cloning and expression of a new member of the nerve growth factor receptor family that is characteristic for Hodgkin's disease. Cell 68: 421-427. Ernfors, P., Ibafiez, C.F., Ebandal, T., Olson, L. and Persson, H. (1990a) Molecular cloning and neurotrophic activities of a protein with structural similarities to b-NGF: developmental and topographical expression in the brain. Proc. Natl Acad. Sci. USA 87: 5454-5458. Ernfors, P., Wetmore, C, Olson, L. and Persson, H. (1990b) Identification of cells in rat brain and peripheral tissues expressing mRNA for members of the nerve growth factor family. Neuron 5: 511-526. Ernfors, P., Bengzon, J., Kokaia, Z., Persson, H. and Lindvall, O. (1991) Increased levels of messenger RNAs for neurotrophic factors in the brain following kindling epileptogenesis. Neuron 7: 165-176. Ernfors, P., Lee, K.-F. and Jaenisch, R. (1994) Mice lacking brain-derived neurotrophic factor develop with sensory deficits. Nature 368: 147-150. Fadda, E., Negro, A., Facci, L. and Skaper, S.D. (1993) Ganglioside GMl cooperates with brain derived neurotrophic factor to protect dopaminergic neurons from 6 hydroxydopamine induced degeneration. Neurosci. Lett. 159: 147-150. Falkenberg, T., Metsis, M., Timmusk, T. and Lindefors, N. (1993) Entorhinal cortex regulation of multiple brainderived neurotrophic factor promoters in the rat hippocampus. Neuroscience 57: 891-896. Friedman, W.J., Olson, L. and Persson, H. (1991) Cells that express brain-derived neurotrophic factor mRNA in the developing postnatal rat brain. Eur. J. Neurosci. 3: 688-697. Friedman, B., Kleinfeld, D., Ip, N.Y., Verge, V.M.K., Moulton, R., Boland, P., Zlotchenko, E., Lindsay, R.M. and Liu, L.M. (1995) BDNF and NT-4/5 exert neurotrophic influences on injured adult spinal motor neurons. J. Neurosci. 15: 1044-1056. Frim, D.M., Uhler, T.A., Galpern, W.R., Beal, M.F., Breakefield, X.O. and Isacson, O. (1994) Implanted fibroblasts genetically engineered to produce brain derived neurotrophic factor prevent 1 methyl 4 phenylpyridinium toxicity to dopaminergic neurons in the rat. Proc. Natl. Acad. Sci. USA 91:5104-5108. Gage, F.H., Armstrong, D.M., Williams, L.R. and Varon, S. (1988) Morphologic response of axotomised septal neurons to nerve growth factor. J. Comp. Neurol. 269: 147-155. Gall, C., Lauterborn, J., Bundman, M., Murray, K. and Isackson, P. (1991) Seizures and the regulation of neurotrophic factor and neuropeptide gene expression in brain. Epil. Res. Suppl. 4: 225-245. Glass, D.J., Nye, S.H., Hantzopoulos, P., Macchi, M.J., Squinto, S.P., Goldfarb, M. and Yancopoulos, G.D. (1991)

213 TrkB mediates BDNF/NT-3-dependent survival and proliferation in fibroblasts lacking the low affinity NGF receptor. Cell 66: 405^13. Gotz, R., Raulf, F. and Schartl, M. (1992) Brain-derived neurotrophic factor is more highly conserved in structure and function than nerve growth factor during vertebrate evolution. J. Neurochem. 59: 432-442. Gotz, R., Koster, R., Winkler, C., Raulf, F., Lottspeich, F., Schartl, M. and Thoenen, H. (1994) Neurotrophin-6 is a new member of the nerve growth factor family. Nature 372:266-269. Hallbook, F., Ibafiez, C.F. and Persson, H. (1991) Evolutionary studies of the nerve growth factor family reveal a novel member abundantly expressed in Xenopus ovary. Neuron 6: 845-858. Hartikka, J. and Hefti, F. (1988) Development of septal cholinergic neurons in culture: plating density and glial cells modulate effects of NGF on survival, fibre growth, and expression of transmitter-specific enzymes. J. Neurosci. 8: 2967-2985. Hatanaka, H., Tsukui, H. and Nihonmatsu, I. (1988) Developmental change in the nerve growth factor action from induction of choline acetyltransferase to promotion of cell survival in cultured basal forebrain cholinergic neurons from postnatal rats. Dev. Brain Res. 39: 85-95. Hefti, F. (1986) Nerve growth factor (NGF) promotes survival of septal cholinergic neurons after fimbria transection. J. Neurosci. 6: 2155-2162. Henderson, C.E., Camu, W., Mettling, C, Gouin, A., Poulsen, K., Karihaloo, M., RuUamas, J., Evans, T., McMahon, S.B., Armanini, M.P., Berkemeier, L., Phillips, H.S. and Rosenthal, A. (1993) Neurotrophins promote motor neuron survival and are present in embryonic limb bud. Nature 363: 266-270. Herzog, K.-H., Bailey, K. and Barde, Y.-A. (1994) Expression of the BDNF gene in the developing visual system of the chick. Development 120: 1643-1649. Hofer, M., Pagliusi, S.R., Hohn, A., Leibrock, J. and Barde, Y.-A. (1990) Regional distribution of brain-derived neurotrophic factor mRNA in the adult mouse brain. EMBO J. 9: 2459-2464. Hofer, M.M. and Barde, Y.-A. (1988) Brain-derived neurotrophic factor prevents neuronal death in vivo. Nature 331:261-262. Hohn, A., Leibrock, J., Bailey, K. and Barde, Y.-A. (1990) Identification and characterization of a novel member of the nerve growth factor/brain-derived neurotrophic factor family. A^^rwr^ 344: 339-341. Humpel, C, Wetmore, C. and Olson, L. (1993) Regulation of brain derived neurotrophic factor messenger RNA and protein at the cellular level in pentylenetetrazol induced epileptic seizures. Neuroscience 53: 909-918. Hyman, C, Hofer, M., Barde, Y.-A., Juhasz, M.Y., G.D., Squinto, S.P. and Lindsay, R.M. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350: 230-232.

214 Ip, N.Y., Ibaiiez, C.F., Nye, S.H., McClain, J., Jones, P.F., Gies, D.R., Belluscio, L., LeBeau, M.M., Espinosa, R.I., Squinto, S.P., Persson, H. and Yancopoulos, G.D. (1992) Mammalian neurotrophin-4: structure, chromosomal localization, tissue distribution and receptor specificity. Proc. Natl Acad. Sci. USA 89: 3060-3064. Ip, N.Y., Li, Y.P., Yancopoulos, G.D. and Lindsay, R.M. (1993) Cultured hippocampal neurons show responses to BDNF, NT 3, NT 4, but not NGF. J. Neurosci. 13: 33943405. Isackson, P.J., Huntsman, M.M., Murray, K.D. and Gall, CM. (1991a) BDNF mRNA expression is increased in adult rat forebrain after limbic seizures: temporal patterns of induction distinct from NGF. Neuron 6: 937-948. Isackson, P.J., Towner, M.D. and Huntsman, M.M. (1991b) Comparison of mammalian, chicken and Xenopus brainderived neurotrophic factor coding sequences. Fed. Eur. Biochem. Soc. 285: 260-264. Itoh, N., Yonehara, S., Ishii, A., Yonehara, M., Mizushima, S.-L, Sameshima, M., Hase, A., Seto, Y. and Nagata, S. (1991) The polypeptide encoded by the cDNA for human cell surface antigen Fas can mediate apoptosis. Cell 66: 233-243. Johnson, D., Lanahan, A., Buck, C.R., Seghal, A., Morgan, C, Mercer, E., Bothwell, M. and Chao, M. (1986) Expression and structure of the human NGF receptor. Cell 47: 545554. Johnson, E.M., Taniuchi, M., Clark, H.B., Springer, J.E., Koh, S., Tayrien, M.W. and Loy, R. (1987) Demonstration of the retrograde transport of NGF receptor in the peripheral and central nervous system. J. Neurosci. 7: 923-929. Johnson, J.E., Barde, Y.-A., Schwab, M. and Thoenen, H. (1986) Brain-derived neurotrophic factor supports the survival of cultured rat ganglion cells. J. Neurosci. 6: 30313038. Jones, K.R., Farinas, I., Backus, C. and Reichardt, L.F. (1994) Targeted disruption of the BDNF gene perturbs brain and sensory neuron development but not motor neuron development. Cell 76: 989-999. Jungbluth, S., Bailey, K. and Barde, Y.-A. (1994) Purification and charaterisation of a brain-derived neurotrophic factor/neurotrophin-3 (BDNF/NT-3) heterodimer. Eur. J. Biochem. 221: 611-6S5. Kato, A.C. and Lindsay, R.M. (1994) Overlapping and additive effects of neurotrophins and CNTF on cultured human spinal cord neurons. Exp. Neurol. 130: 196-201. Klein, R., Parada, L.F., Coulier, F. and Barbacid, M. (1989) trkB, A novel tyrosine protein kinase receptor expressed during mouse neural development. EMBO J. 8: 37013709. Klein, R., Conway, D., Parada, L.F. and Barbacid, M. (1990a) The trkB tyrosine kinase gene codes for a second neurogenic receptor that lacks the catalytic kinase domain. Cell 61:647-656. Klein, R., Martin-Zanca, D., Barbacid, M. and Parada, L.F. (1990b) Expression of the tyrosine kinase receptor gene

Brain-derived neurotrophic factor trkB is confined to the murine embryonic and adult nervous system. Development 109: 845-850. Klein, R., Nanduri, V., Jing, S., Lamballe, F., Tapley, P., Bryant, S., Cordon-Cardo, C, Jones, K.R., Reichardt, L.F. and Barbacid, M. (1991) The trkB tyrosine protein kinase is a receptor for brain-derived neurotrophic factor and neurtrophin-3. Cell 66: 395^03. Klein, R., Smeyne, R.J., Wurst, W., Long, L.K., Auerbach, B.A., Joyner, A.L. and Barbacid, M. (1993) Targeted disruption of the trkB neurotrophin receptor gene results in nervous system lesions and neonatal death. Cell 75: 113122. Knipper, M., Berzaghi, M.P., Blochl, A., Breer, H., Thoenen, H. and Lindholm, D. (1994) Positive feedback between acetylcholine and the neurotrophins nerve growth factor and brain-derived neurotrophic factor in the rat hippocampus. Eur. J. Neurosci. 6: 668-671. Knusel, B., Winslow, J.W., Rosenthal, A., Burton, L.E., Seid, D.P., Nikolics, K. and Hefti, F. (1991) Promotion of central cholinergic and dopaminergic neuron differentiation by brain-derived neurotrophic factor but not neurotrophin 3. Proc. Natl. Acad. Sci. USA 88: 961-965. Knusel, B., Beck, K.D., Winslow, J.W., Rosenthal, A., Burton, L.E., Widmer, H.R., Nikolics, K. and Hefti, F. (1992) Brain-derived neurotrophic factor administration protects basal forebrain cholinergic but not nigral dopaminergic neurons from degenerative changes after axotomy in the adult rat brain. J. Neurosci. 12: 4391-4402. Koliatsos, V.E., Nauta, H.J.W., Clatterbuck, R.E., Holzman, D.M., Mobley, W.C. and Price, D.L. (1990) Mouse nerve growth factor prevents degeneration of axotomized basal forebrain cholinergic neurons in the monkey. J. Neurosci. 10:3801-3813. Koliatsos, V.E., Clatterbuck, R.E., Winslow, J.W., Cayouette, M.H. and Price, D.L. (1993) Evidence that brain-derived neurotrophic factor is a trophic factor for motor neurons in vivo. Neuron 10: 359-367. Koliatsos, V.E., Price, D.L., Gouras, G.K., Cayouette, M.H., Burton, L.E. and Winslow, J.W. (1994) Highly selective effects of nerve growth factor, brain-derived neurotrophic factor, and neurotrophin-3 on intact and injured basal forebrain magnocellular neurons. J. Comp. Neurol. 343: 247262. Lapchak, P.A., Araujo, D.M., Beck, K.D., Finch, C.E., Johnson, S.A. and Hefti, F. (1993) BDNF and trkB mRNA expression in the hippocampal formation of aging rats. Neurobiol. Aging 14: 121-126. Lee, K.F., Li, E., Huber, L.J., Landis, S.C, Sharpe, A.H., Chao, M.V. and Jaenisch, R. (1992) Targeted mutation of the gene encoding the low affinity NGF receptor p75 leads to deficits in the peripheral sensory nervous system. Cell 69: 737-749. Lee, K.F., Bachman, K., Landis, S. and Jaenisch, R. (1994a) Dependence of p75 for innervation of some sympathetic targets. Science 263: 1447-1449. Lee, K.F., Davies, A.M. and Jaenisch, R. (1994b) p75-

K.A. Bailey deficient embryonic dorsal root sensory and neonatal sympathetic neurons display a decreased sensitivity to NGF. Development 120: 1027-1033. Leibrock, J., Lottspeich, F., Hohn, A., Hofer, M., Hengerer, B., Masiakowski, P., Thoenen, H. and Barde, Y.-A. (1989) Molecular cloning and expression of brain-derived neurotrophic factor. Nature 341: 149-152. Levi-Montalcini, R. (1987) The nerve growth factor 35 years later. Science 237: 1154-1162. Li, L., Oppenheim, R.W., Lei, M. and Houenou, L.J. (1994) Neurotrophic agents prevent motoneuron death following sciatic nerve section in the neonatal mouse. J. NeurobioL 25: 759-766. Lindholm, D., Dechant, G., Heisenberg, C.P. and Thoenen, H. (1993) Brain derived neurotrophic factor is a survival factor for cultured rat cerebellar granule neurons and protects them against glutamate induced neurotoxicity. Eur. J. Neurosci. 5: 1455-1464. Lindsay, R.M., Thoenen, H. and Barde, Y.-A. (1985) Placode and neural crest-derived sensory neurons are responsive at early developmental stages to brain-derived neurotrophic factor. Dev. Biol 112: 319-328. Lindvall, O., Ernfors, P., Bengzon, J., Kokaia, Z., Smith, M.L., Siesjo, B.K. and Persson, H. (1992) Differential regulation of mRNAs for nerve growth factor, brain-derived neurotrophic factor, and neurotrophin 3 in the adult rat brain following cerebral ischemia and hypoglycemic coma. Proc. Natl Acad. Scl USA 89: 648-652. Liu, X., Ernfors, P., Wu, H. and Jaenisch, R. (1995) Sensory but not motor neuron deficits in mice lacking NT4 and BDNF. Nature 375: 238-241. Maisonpierre, P.C, Belluscio, L., Friedman, B., Alderson, R.F., Wiegand, S.J., Furth, M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990a) NT-3, BDNF, and NGF in the developing rat nervous system: parallel as well as reciprocal patterns of expression. Neuron 5: 501-509. Maisonpierre, P.C., Belluscio, L., Squinto, S., Ip, N.Y., Furth, M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990b) Neurotrophin-3: a neurotrophic factor related to NGF and BDNF. Science 247: 1446-1451. Maisonpierre, P.C, Le Beau, M.M., Espinosa, R.I., Ip, N.Y., Belluscio, L., de la Monte, S.M., Squinto, S., Furth, M.E. and Yancopoulos, G.D. (1991) Human and rat brainderived neurotrophic factor and neurotrophin-3: gene structures, distributions, and chromosomal localizations. Genomics 10: 558-568. Maisonpierre, P.C., Belluscio, L., Conover, J.C. and Yancopoulos, G.D. (1992) Gene sequences of chicken BDNF and NT-3. DNA seq. 3: 49-54. Mallet, S., Fossum, S. and Barclay, A.N. (1990) Characterization of the MRC OX40 antigen of activated CD4 positive T lymphocytes - a molecule related to the nerve growth factor receptor. EMBOJ. 9: 1063-1068. Martin-Iverson, M.T., Todd, K.G. and Altar, C.A. (1994) Brain-derived neurotrophic factor and neurotrophin-3 activate striatal dopamine and serotonin metabolism and related

215 behaviors - interactions with amphetamine. /. Neuroscl 14: 1262-1270. McDonald, N.Q., Lapatto, R., Murray-Rust, J., Gunning, J., Wlodawer, A. and Blundell, T.L. (1991) New protein fold revealed by a 2.3A resolution crystal structure of nerve growth factor. Nature 354: 411^14. Merlio, J.-P., Ernfors, P., Kokaia, Z., Middlemas, D.S., Bengzon, J., Kokaia, M., Smith, M.-L., Siesjo, B.K., Hunter, T., Lindvall, O. and Persson, H. (1993) Increased production of the TrkB protein tyrosine kinase receptor after brain insults. Neuron 10: 151-164. Metsis, M., Timmusk, T., Arenas, E. and Persson, H. (1993) Diferential usage of multiple brain-derived neurotrophic factor promoters in the rat brain following neuronal activation. Proc. Natl Acad. Scl USA 90: 8802-8806. Middlemas, D.S., Lindberg, R.A. and Hunter, T. (1991) trkB, A neural receptor protein- tyrosine kinase: evidence for a full-lenght and two truncated receptors. Mol Cell Biol 11: 143-153. Mitsumoto, H., Ikeda, K., Klinkosz, B., Cedarbaum, J.M., Wong, V. and Lindsay, R.M. (1994) Arrest of motor neuron disease in wobbler mice cotreated with CNTF and BDNF. Science 265: 1107-1110. Morse, J.K., Wiegand, S.J., Anderson, K., You, Y.M., Cai, N., Carnahan, J., Miller, J., DiStephano, P.S., Altar, C.A., Lindsay, R.M. and Alderson, R.F. (1993) Brain derived neurotrophic factor (BDNF) prevents the degeneration of medial septal cholinergic neurons following fimbria transection. J. Neuroscl 13: 4146-4156. Murray, K.D., Gall, CM., Jones, E.G. and Isackson, P.J. (1994) Differential regulation of brain-derived neurotrophic factor and type II calcium/calmodulin-dependent protein kinase messenger RNA expression in Alzheimer's disease. Neuro science 60: 37-48. Nakayama, M., Gahara, Y., Kitamura, T. and Ohara, O. (1994) Distinctive four promoters collectively direct expression of brain-derived neurotrophic factor gene. Mol. Brain Res. 21: 206-2n. Narhi, L.O., Rosenfeld, R., Talvenheimo, J., Prestrelski, S.J., Arakawa, T., Lary, J.W., Kolvenbach, C.G., Hecht, R., Boone, T., Miller, J.A. and Yphantis, D.A. (1993) Comparison of the biophysical characteristics of human brainderived neurotrophic factor, neurotrophin-3 and nerve growth factor. /. Biol Chem. 268: 13309-13317. Ohara, O., Gahara, Y., Teraoka, H. and Kitamura, T. (1992) A rat brain-derived neurotrophic factor-encoding gene generates multiple transcripts through alternative use of 5' exons and polyadenylation sites. Gene 121: 383-386. Ohsawa, F., Widmer, H.R., Knusel, B., Denton, T.L. and Hefti, F. (1993) Response of embryonic rat hippocampal neurons in culture to neurotrophin 3, brain derived neurotrophic factor and basic fibroblast growth factor. Neuroscience 57: 61-11. Oppenheim, R.W., Qin-Wei, Y., Prevette, D. and Yan, Q. (1992) Brain-derived neurotrophic factor rescues developing avian motor neurons from cell death. Nature 360: 755-757.

216 Ozgelik, T., Rosenthal, A. and Francke, U. (1991) Chromosomal mapping of brain-derived neurotrophic factor and neurotrophin-3 genes in man and mouse. Genomics 10: 569-575. Phillips, H.D., Hains, J.M., Laramee, G.R., Rosenthal, A. and Winslow, J.W. (1990) Widespread expression of BDNF but not NT-3 by target areas of basal forebrain cholinergic neurons. Science 250: 290-294. Phillips, U.S., Hains, J.M., Armanini, M., Laramee, G.R., Johnson, S.A. and Winslow, J.W. (1991) BDNF mRNA is decreased in the hippocampus of individuals with Alzheimer's disease. Neuron 7: 695-702. Philo, J., Talvenheimo, J., Wen, J., Rosenfeld, R., Welcher, A. and Arakawa, T. (1994) Interactions of neurotrophin-3 (NT3), brain-derived neurotrophic factor (BDNF), and the NT-3 BDNF heterodimer with the extracellular domains of the TrkB and TrkC receptors. J. Biol. Chem. 269: 2784027846. Radeke, M.J., Misko, T.P., Hsu, C, Herzenberg, L.A. and Shooter, E. (1987) Gene transfer and molecular cloning of the rat nerve growth factor receptor: a new class of receptors. Nature 325: 593-597. Radziejewski, C. and Robinson, R.C. (1993) Heterodimers of the neurotrophic factors: formation, isolation, and differential stability. Biochemistry 32: 13350-13356. Radziejewski, C, Robinson, R.C, DiStefano, P.S. and Taylor, J.W. (1992) Dimeric structure and conformational stability of brain-derived neurotrophic factor and neurotrophin-3. Biochemistry 31: 4431^436. Rocamora, N., Palacios, J.M. and Mengod, G. (1992) Limbic seizures induce a differential regulation of the expression of nerve growth factor, brain-derived neurotrophic factor and neurotrophin-3, in the rat hippocampus. Mol Brain Res. 13: 27-33. Rodriguez-Tebar, A. and Barde, Y.-A. (1988) Binding characteristics of brain-derived neurotrophic factor to its receptor on neurons from the chick embryo. J. Neurosci. 8: 33373342. Rodriguez-Tebar, A., Jeffery, P.L., Thoenen, H. and Barde, Y.-A. (1989) The survival of chick retinal ganglion cells in response to brain-derived neurotrophic factor depends on their embryonic age. Dev. Biol. 136: 296-303. Rodriguez-Tebar, A., Dechant, G. and Barde, Y.-A. (1990) Binding of brain-derived neurotrophic factor to the nerve growth factor receptor. Neuron 4: 487-492. Rodriguez-Tebar, A., Dechant, G., Gotz, R. and Barde, Y.-A. (1992) Binding of neurotrophin-3 to its neuronal receptors and interactions with nerve growth factor and brain-derived neurotrophic factor. EMBO J. 11: 917-922. Rosenthal, A., Goeddel, D.V., Nguyen, T., Lewis, M., Shih, A., Laramee, G.R., Nikolics, K. and Winslow, J.W. (1990) Primary structure and biological activity of a novel human neurotrophic factor. Neuron 4:161-113. Rosenthal, A., Goeddel, D.V., Nguyen, T., Martin, E., Burton, L.E., Shih, A., Laramee, G.R., Wurm, F., Mason, A., Nikolics, K. and Winslow, J.W. (1991) Primary structure and

Brain-derived neurotrophic factor biological activity of human brain-derived neurotrophic factor. Endocrinology 129: 1289-1294. Schecterson, L.C. and Bothwell, M. (1992) Novel roles for the neurotrophins are suggested by BDNF and NT-3 mRNA Expression in developing neurons. Neuron 9: 449-463. Schmalbruch, H. (1984) Motoneuron death after sciatic nerve section in newborn rats. J. Comp. Neurol. 224: 252-258. Schults, C.W., Matthews, R.T., Altar, C.A., Hill, L.R. and Langlais, P.J. (1994) Single intramesencephalic injection of brain derived neurotrophic factor induces persistent rotational asymmetry in rats. Exp. Neurol. 125: 183-194. Seeburger, J.L., Tarras, S., Natter, H. and Springer, J.E. (1993) Spinal cord motoneurons express p75^^^^ and pj45trkB niRNA in amyotrophic lateral sclerosis. Brain Res. 621: 111-115. Segal, R.A., Takahashi, H. and McKay, R.D.G. (1992) Changes in neurotrophin responsiveness during the development of cerebellar granule neurons. Neuron 9: 10411052. Sendtner, M., Holtmann, B., Kolbeck, R., Thoenen, H. and Barde, Y.-A. (1992) Brain- derived neurotrophic factor prevents the death of motoneurons in newborn rats after nerve section. Nature 360: 757-759. Sieber-Blum, M. (1991) Role of the neurotrophic factors BDNF and NGF in the commitment of pluripotent neural crest cells. Neuron 6: 949-955. Skaper, S.D., Negro, A., Facci, L. and Daltoso, R. (1993) Brain derived neurotrophic factor selectively rescues mesencephalic dopaminergic neurons from 2,4,5 trihydroxyphenylalanine induced injury. J. Neurosci. Res. 34: 478-487. Skup, M.H., Figueiredo, B.C. and Cuello, A.C. (1994) Intraventricular application of BDNF and NT 3 failed to protect nucleus basalis magnocellularis cholinergic neurons. NeuroReport 5: 1105-1109. Smith, C.A., Davies, T., Anderson, D., Solam, L., Beckmann, M.P., Jerzy, R., Dower, S.K., Cosman, D. and Goodwin, R.G. (1990) A receptor for tumor necrosis factor defines an unusual family of cellular and viral proteins. Science 248: 1019-1023. Snider, W.D. and Thanedar, S. (1989) Target deprivation of hypoglossal motoneurons during development and during maturity. J. Comp. Neurol. 279: 489-498. Soppet, D., Escandon, E., Maragos, J., Middlemas, D.S., Reid, S.W., Blair, J., Burton, L.E., Stanton, B.R., Kaplan, D.R., Hunter, T., Nikolics, K. and Parada, L.F. (1991) The neurotrophic factors brain-derived neurotrophic factor and NT3 are ligands for the trkB tyrosine kinase receptor. Cell 65: 895-903. Spina, M.B., Squinto, S.P., Miller, J., Lindsay, R.M. and Hyman, C. (1992) Brain-derived neurotrophic factor protects dopamine neurons against 6-hydroxydopamine and A^methyl-4-phenylpyridinium ion toxicity: involvement of the glutathione system. J. Neurochem. 59: 99-106. Squinto, S.P., Stitt, T.N., Aldrich, T.H., Davis, S., Bianco, S.M., Radziejewski, C, Glass, D., J., Masiakowski, P.,

K.A. Bailey Furth, M.E., Valenzuela, D.M., DiStefano, P.S. and Yancopoulos, G.D. (1991) TrkB encodes a functional receptor for brain-derived neurotrophic factor and neurotrophin-3 but not nerve growth factor. Cell 65: 885-893. Stamenkovic, I., Clark, E.A. and Seed, B. (1989) A Blymphocyte activation molecule related to the nerve growth factor receptor and induced by cytokines in carcinomas. EMBOJ. 8: 1403-1410. Sutter, A., Riopelle, R.J., Harris-Warrick, R.M. and Shotter, E.M. (1979) Nerve growth factor receptors. Characterization of two distinct classes of binding site on chick embryo sensory ganglion cells. J. Biol. Chem. 254: 5972-5982. Takeda, A., Onodera, H., Sugimoto, A., Kogure, K., Obinata, M. and Shibahara, S. (1993) Coordinated expression of messenger RNAs for nerve growth factor, brain derived neurotrophic factor and neurotrophin 3 in the rat hippocampus following transient forebrain ischemia. Neuroscience 55:23-31. Taniuchi, M., Clark, H.B., Schweitzer, J.B. and Johnson, E.M. (1988) Expression of nerve growth factor receptors by Schwann cells of axotomised peripheral nerves: ultrastructural location, suppression by axonal contact, and binding properties. J. Neurosci. 8: 664-681. Thoenen, H. and Barde, Y.-A. (1980) Physiology of nerve growth factor. Physiol. Rev. 60: 1284-1335. Thoenen, H., Korsching, S., Barde, Y.-A. and Edgar, D. (1983) Quantitation and purification of neurotrophic molecules. Cold Spring Harbor Symp. Quant. Biol. 48: 679-684. Timmusk, T. and Metsis, M. (1994) Regulation of BDNF promoters in the rat hippocampus. Neurochem. Int. 25: 1115. Timmusk, T., Palm, K., Metsis, M., Reintam, T., Paalme, V., Saarma, M. and Persson, H. (1993) Multiple promoters direct tissue-specific expression of the rat BDNF gene. Neuron 10: 475-489. Timmusk, T., Belluardo, N., Persson, H. and Metsis, M. (1994) Developmental regulation of brain-derived neurotrophic factor messenger RNAs transcribed from different promoters in the rat brain. Neuroscience 60: 287-291. Timmusk, T., Lendahl, U., Funakoshi, H., Arenas, E., Persson, H. and Metsis, M. (1995) Identification of brain-derived neurotrophic factor promoter regions mediating tissuespecific, axotomy-, and neuronal activity-induced expression in transgenic mice. J. Cell Biol. 128: 185-199. Verdi, J.M., Birren, S.J., Ibanez, C.F., Persson, H., Kaplan, D.R., Benedetti, M., Chao, M.V. and Anderson, D.J. (1994) p75LNGFR regulates Trk signal transduction and NGF- induces neuronal differentiation in MAH cells. Neuron 12: 733-745. Vogel, K.S. and Davies, A.M. (1991) The duration of neurotrophic factor dependence in early sensory neurons is

111 matched to the time course of target field innervation. Neuron 7: 819-830. Watkins, J.C, Krogsgaard, L.P. and Honore, T. (1990) Structure-activity relationships in the development of excitatory amino acid receptor agonists and competitive antagonists. Trends Pharmac. Sci. 11: 25-33. Wetmore, C, Ernfors, P., Persson, H. and Olson, L. (1990) Localization of brain-derived neurotrophic factor mRNA to neurons in the brain by in situ hybridization. Exp. Neurol. 109: 141-152. Wetmore, C, Olson, L. and Bean, A.J. (1994) Regulation of brain-derived neurotrophic factor (BDNF) expression and release from hippocampal neurons is mediated by nonNMDA type glutamate receptors. J. Neurosci. 14: 16881700. Whitehouse, P.J., Price, D.L., Struble, R.G., Clark, A.W., Coyle, J.T. and Delong, M.R. (1982) Alzheimer's disease and senile dementia: loss of neurons in the basal forebrain. Science 215: 1237-1239. Widmer, H.R., Knusel, B. and Hefti, F. (1993) BDNF protection of basal forebrain cholinergic neurons after axotomy complete protection of p75NGFR positive cells. NeuroReport 4: 363-366. Wong, v., Arriaga, R., Ip, N.Y. and Lindsay, R.M. (1993) The neurotrophins BDNF, NT-3 and NT-4/5, but not NGF, upregulate the cholinergic phenotype of developing motor neurons. Eur. J. Neurosci. 5: 466-474. Wright, E.M., Vogel, K.S. and Davies, A.M. (1992) Neurotrophic factors promote the maturation of developing sensory neurons before they become dependent on these factors for survival. Neuron 9: 139-150. Yamamoto, H. and Gurney, M.E. (1990) Human platelets contain brain-derived neurotrophic factor. /. Neurosci. 10: 3469-3478. Yan, Q., Elliot, J. and Snider, W.D. (1992) Brain-derived neurotrophic factor rescues spinal motor neurons from axotomy-induced cell death. Nature 360: 753-755. Yan, Q., Elliott, J.L., Matheson, C, Sun, J., Zhang, L., Mu, X., Rex, K.L. and Snider, W.D. (1993) Influences of neurotrophins on mammalian motoneurons in vivo. J. Neurobiol. 24: 1555-1577. Zafra, F., Hengerer, B., Leibrock, J., Thoenen, H. and Lindholm, D. (1990) Activity dependent regulation of BDNF and NGF mRNAs in the rat hippocampus is mediated by non-NMDA glutamate receptors. EMBO J. 9: 3545-3550. Zafra, F., Castren, E., Thoenen, H. and Lindholm, D. (1991) Interplay between glutamate and gamma-aminobutyric acid transmitter systems in the physiological regulation of brainderived neurotrophic factor and nerve growth factor synthesis in hippocampal neurons. Proc. Natl. Acad. Sci. USA 88: 10037-10041.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 9

Neurotrophin-3 and neurotrophin-4/5 N. Rocamora^ and E. Arenas^ ^Department of Animal and Vegetal Cell Biology, Faculty of Biology, University of Barcelona, 08028-Barcelona, Spain ^Laboratory of Molecular Neurobiology, Karolinska Institute, Stockholm S-17177, Sweden

1. Introduction: neurotrophins from historic and evolutionary perspectives Nerve growth factor (NGF) was for a long time the only identified molecule fulfilling the neurotrophic factor definition; a target-derived, retrogradelytransported molecule supporting neuron survival during developmental programmed cell death (Levi-Montalcini, 1987). However, the highly restricted specificity of NGF, which only supports a limited number of neuronal populations, together with the widespread developmental cell death in the vertebrate nervous system, stimulated the search for other neurotrophic factors exerting effects similar to NGF but on distinct neuronal populations. A second neurotrophic factor, brainderived neurotrophic factor (BDNF), was purified from porcine brain twenty-five years after NGF (Barde et al., 1982). Functional studies in vitro showed that NGF and BDNF support survival and differentiation of different, yet overlapping, populations of peripheral neural cells, and subsequent BDNF molecular cloning (Leibrock et al., 1989) uncovered the close sequence homology between these neurotrophic factors. The classical experimental approach to assess neurotrophic activity of a given factor utilized cultured and/or explanted ganglia of the peripheral nervous system (PNS), extracted during the time when programmed neuronal death takes place. By this approach, NGF and BDNF were both found to support neural crest-derived sensory neurons of dorsal root ganglion (DRG). However, NGF but not BDNF was found to be a trophic factor for sympathetic neurons, and BDNF but not NGF

supported sensory cells from placode-derived nodose ganglia (NG) (Lindsay et al., 1985; for a review see Barde, 1989). Northern blots and in situ hybridization data showed that BDNF mRNA is mainly present in the brain and suggested a putative trophic effect for neurotrophins not only on PNS neurons but also on intrinsic neurons of the central nervous system (CNS). In fact, both NGF and BDNF were found to support central cholinergic neurons in culture (Martinez et al., 1985; Hefti, 1986; Hartikka and Hefti, 1988; Hatanaka et al., 1988; Alderson et al, 1990; Knusel et al., 1991). However, only BDNF supported chick and rat retinal ganglion cells, dopaminergic neurons of the ventral mesencephalon, and GAB Aergic neurons of the basal forebrain and striatum (Lindsay et al., 1985; Alderson et al., 1990; Hyman et al., 1991, 1994; Knusel et al., 1991, 1994; Ventimigliaetal., 1995). Taking advantage of the high degree of homology between NGF and BDNF, two further neurotrophins, neurotrophin-3 (NT-3) and neurotrophin-4 (NT-4), were not purified from natural sources, but directly cloned. Sequence analysis shows that the four neurotrophins are secretory polypeptides of 118 (NGF), 119 (BDNF and NT-3), and 123-130 (NT-4, -4/5) aminoacids, sharing 5060% homology between them. They have conserved domains, including the six cysteines involved in the formation of three intrachain disulfide bonds, which determine their basic structure; and specific variable domains, such as the three exposed ^-hairpin loops, and the three consecutive reverse turns, responsible for their specifity (Ibanez et al, 1991, 1992, 1993b; Hag etal, 1994).

220

Gene families with four members are common in vertebrates, which has led to the hypothesis of that two genomic duplications occurred close to the origin and early radiation of the vertebrates (Holland et al., 1994). Thus, neurotrophins may have appeared by duplication of a primitive neurotrophin-like gene. Sequence comparison showed that, although NT-3 and NT-4/5 are the most recently identified neurotrophins, they are very distant at the evolutionary level. NT-3 and BDNF are the most conserved neurotrophins between the different vertebrates. NT-3 from human, rat, chicken and Xenopus are 100% identical at the amino acid level. The same is true for BDNF from human, rat and salmon (Hallbook et al., 1991). Moreover, NT-3 and BDNF are the only neurotrophins present in the ray, a cartilaginous fish (the most primitive vertebrate group). By contrast, NT4/5 is less conserved, with only a 65% amino acid identity between Xenopus NT-4 and mammalian NT-4/5, and thus the youngest in evolutionary terms. NGF has an intermediate situation. It is found in teleost fishes, together with NT-3 and BDNF (Gotz et al., 1992; for reviews see Ebendal, 1992, Barde, 1994). A fifth neurotrophin-like factor, NT-6, which has an heparin-binding domain, has been found in the teleost fish Xiphophorus (Gotz et al., 1994). However, NT-6 counterparts have not been found in any other vertebrates. Thus, at the moment, it may be considered as a characteristic neurotrophin-like factor for fishes. Moreover, its homology with NGF suggests that it may have appeared from a duplication of this gene, during the specific evolution of the fishes. Neurotrophins are involved in survival, differentiation, and maintenance of neural cells and their circuits. Their predominant (active) forms are homodimers of 13-14 kDa polypeptides generated from 30 kDa precursors, whose internalization and function depends on the presence of *high-affinity' receptors. NGF, BDNF, NT-3 and NT-4/5 bind with comparable, nanomolar, low-affinity to the membrane glycoprotein p75^N^^^ (Chao et al., 1986; Raddeke et al., 1987, Rodriguez-Tebar et al., 1990, 1992; Hallbook et al., 1991). The discovery of trk (tropomyosine receptor kinase) (MartmZanca et al., 1989) as a signal transduction, high-

Neurotrophin-3 and neurotrophin-4/5

affinity receptor for NGF (Kaplan et al., 1991a; for reviews see Chao, 1992; Barbacid, 1994), introduced a new view of the functionality of neurotrophins. This allowed demonstration of the specificity of the different neurotrophins with respect to their binding and to induction of autophosphorylation of a set of receptor tyrosine kinases (trkA, trkB and trkC, the so-called Trk' family). Thus, NGF is the preferred ligand of trkA (Kaplan et al., 1991a; Klein et al., 1991a), BDNF and NT-4 of trkB (Berkemeier et al., 1991; Klein et al., 1991b; Soppet et al, 1991; Squinto et al., 1991), and NT-3 of trkC (Lamballe et al., 1991). However, NT-3 was found to be a more promiscuous ligand and binds also trkB (Glass et al., 1991; Klein et al., 1991b; Soppet et al., 1991; Squinto et al., 1991) and, with less affinity, trkA (Cordon-Cardo et al., 1991). Finally, alternative splicing of trk mRNAs leading to different trk receptors could be an important mechanism to regulate different responses to neurotrophins. Differential splicing of trkB and C transcripts lead both to truncated receptors (lacking the kinase domain) and full length receptors (Middlemas et al., 1991; Lamballe et al., 1993; Tsouflas et al., 1993; Valenzuela et al., 1993). An alternative spliced form of trkA, with enhanced responses to NT-3 (Clary and Reichardt, 1994), is present both in neurons and in nonneuronal cells. This receptor signals neurons to stop proliferation and initiate differentiation, and signals non-neuronal cells to proliferate. Other alternative spliced forms, such as trkC with insertion in the tyrosine kinase domain, unlike the noninserted forms, can mediate proliferation in fibroblasts and neuronal differentiation in PC 12 cells (Tsouflas et al., 1993; Valenzuela et al., 1993) In summary, neurotrophin signal transduction has been shown to be dependent on the presence of the appropriate trk receptors on the responsive cells. In contrast, p75L^^^^' was found not to be necessary for neurotrophin signal transduction in vitro (Cordon-Cardo et al., 1991; Glass et al., 1991; Weskamp and Reichardt, 1991; Ibanez et al., 1992; Ip et al., 1993b). Its involvement as mediator of neurotrophin function in vivo is poorly understood, at the moment. However, it has been sug-

N. Rocamora and E. Arenas

gested to play a role in the modulation and potentiation of trk function (Verdi et al., 1994; Hantzopoulos et al., 1994). After these introductory considerations, this review will focus mainly on the analysis of expression and function of the two most recently identified neurotrophins: NT-3 and NT-4/5. Three main points will be considered: (i) patterns of mRNA expression of these neurotrophins and their receptors during development and in adults, (ii) regulation of their expression and (iii) functional aspects in vitro and in vivo. General considerations about their effects on PNS and CNS neurons and their functional plasticity, as well as possible applications of NT-3 and NT-4/5 in therapeutics, are discussed. 2. Neurotrophin-3 In contrast to the long delay between the cloning of NGF and BDNF, the sequence of a third neurotrophin was achieved independently by several groups only a few months after the publication of the BDNF sequence. Taking advantage of the high degree of homology between NGF and BDNF, mouse, rat and human neurotrophin-3 were cloned by a PCR strategy, from genomic DNA (Emfors et al, 1990a; Hohn et al, 1990; Jones and Richardt, 1990; Maisonpierre et al., 1990a; Rosenthal et al., 1990), or by a low-stringency screening approach using human NGF, as a probe (Kaisho et al., 1990). NT-3 showed 57 and 58% amino acid identities with NGF and BDNF, respectively. Cloning and sequencing the NT-3 gene opened up many possibilities, from design of molecular probes to the generation of mice null mutated for NT-3. With all these tools at hand, a lot of information on NT-3 function, expression, binding, regulation, etc., has appeared in the last years. Furthermore, the parallel advances in knowledge of neurotrophins and trks has been extremely fruitful in the understanding of the functions of neurotrophins. 2.7. Expression of NT-3 2.1.1. NT-3 mRNA in development Northern blot analysis showed that initial expression of neurotrophins, in the CNS, coincides

221

approximately with the onset of neurogenesis. However, there is a clear difference in regulation of the different neurotrophins during development. NT-3 mRNA was present in the developing rat brain from E16-E17 onwards, mainly in cingulateretrosplenial cortex, hippocampus, cerebellum and somatic motoneurons. During early postnatal development, NT-3 mRNA is the most abundant neurotrophin messenger in brain, peaking at postnatal day 4, followed by BDNF, which peaks around the second postnatal week and, later on, by NGF. NT-3 mRNA levels decrease approximately 15-fold thereafter, reaching adult levels at 3 weeks after birth (Ernfors et al., 1990a; Kaisho et al., 1990; Maisonpierre et al., 1990b). Thus, during development, NT-3 is expressed mainly in regions of the CNS where proliferation, migration and differentiation of neuronal precursors is ongoing. Expression in cortex. NT-3 mRNA was transiently expressed at high levels in the cingulate cortex, during the first 2 weeks after birth (Friedman et al., 1991), coinciding with the time of innervation by dopaminergic and noradrenergic neurons. Moreover, reciprocal rostro-caudal gradients of NGF and NT-3 mRNA in the postnatal cingulate and retrosplenial cortices correlated with thalamic innervation (Lauterborn et al., 1994). Particularly, cortical fields expressing high levels of NT-3 are preferentially innervated by anteroventral thalamic nuclei, which contain the high-affinity NT-3 receptor (trkC) (Lauterborn et al., 1994), suggesting a role for NT-3 in the establishment of thalamocortical synapses. Expression in hippocampus. NT-3 was expressed at El8, in scattered cells of the medial zone of the developing hippocampus, where granule cells originate. trkC and trkB were also expressed at the same time in this region, suggesting the possibihty of an autocrine function of NT-3 in the early differentiation and survival of hippocampal neurons (Ernfors et al., 1992b). The adult pattern of NT-3 mRNA was already present at birth. In addition, in postnatal hippocampus, NT-3 was expressed in hilar cells and in the CA3 pyramidal layer during the first postnatal week, decreasing and disappear-

222

ing thereafter (Friedman et al., 1991). NT-3 mRNA level decreased along a septotemporal gradient. It was present at PI-2 in the septal extreme but not in the temporal extreme, and by PI2 it was present throughout the hippocampus. As in the adult, scattered cells in the CA3 lamina and in the hilus also expressed NT-3 (Lauterbom et al., 1994). Expression in cerebellum. In the developing cerebellum, NT-3 has been found to be highly and transiently expressed in postmitotic migratory external granule cells (Rocamora et al., 1993a). A dense hybridization signal was present in lateral aspects of the cerebellum at PI. At P9, high levels of NT-3 mRNA were present in paraflocculus and vermis (Lauterbom et al., 1994). These results agree with a function of NT-3 in the early development of the cerebellar system, prior to synaptogenesis between granule and Purkinje cells. TrkC and p75 mRNAs were expressed in Purkinje cells at late embryonic-early postnatal ages, suggesting the possibility of orthograde transport of NT-3 (Ernfors et al., 1992b; Lindholm et al., 1993a; Rocamora etal., 1993a). Expression in the neuromuscular system. Spinal cord motoneurons showed high levels of NT-3 mRNA at El3-16, decreasing to a lower plateau by El8-19 which was maintained until PI. Then, from PI onwards, NT-3 mRNA levels decreased to below the limit of detection in the adult (Ernfors and Persson, 1991; Funakoshi et al., 1995). Transient expression of NT-3 was found in developing muscle. High levels of NT-3 were present in nearly every muscle cell from El 1.5 to E15.5, decreasing thereafter to a very weak signal by E17.5 (Schecterson and Both well, 1992). Copray and Brouwer (1994) showed that NT-3 was expressed in muscle spindles not only in developing but also in adult muscle, suggesting a role for NT-3 as a trophic factor for proprioceptive somatosensory neurons. Expression in the PNS. In the PNS, NT-3 mRNA was expressed in developing dorsal root ganglia (DRG) and possibly in sympathetic ganglia, but

Neurotrophin-3 and neurotrophin-4/5

not in trigeminal ganglia (Ernfors et al., 1992b; Mu et al., 1993; Elkabes et al., 1994). A transient expression of NT-3 was found in large neurons of the developing DRG from E13.5 to E17.5, prior to target field innervation which, together with the presence of trkC in these ganglia from E12.5, suggested a possible local transient function of NT-3 on sensory neurons (Elkabes et al., 1994). NT-3 was found to be expressed in mouse sympathetic ganglia from E14.5 to PI (Schecterson and Bothwell, 1992). However, no NT-3 mRNA has been found in sympathetic ganglia of rat at any of the ages analysed (Ernfors et al., 1992b). TrkC was expressed from El3 onwards. Later on, at El6, trkA mRNA appeared. In summary, expression of NT-3 during early development suggests other functions than the target-derived protective effect against programmed cell death for this neurotrophin. TrkC was expressed from early stages of neuronal development to adult ages. TrkC expression, precedes and correlate with the outgrowth of axons towards their peripheral targets (Lamballe et al., 1994). Moreover, coexpression of NT-3 and trkC in the same area may generate paracrine and/or autocrine loops that could be responsible, at least in part, for the proliferation and differentiation of neuronal precursors and newly formed neurons. 2.7.2. NTS mRNA in the adult In contrast to BDNF, which is mainly expressed in the CNS, NT-3 mRNA was found to be widely distributed in peripheral tissues of the adult rat, but not in the PNS (see Table 1). Northern blotting showed a transcript of 1.3-1.4 kb expressed at the highest level in kidney. Lower levels, comparable with those of NGF in these tissues, were found in spleen and heart. NT-3 mRNA was also present in liver, lung, gut, skeletal muscle, adrenal gland and ovary, where neither NGF nor BDNF were found in comparable amounts. In the CNS, NT-3 was found to be expressed predominantly in the hippocampus, and with lower levels in other brain areas including cerebellum (granule cells), cortex (mainly entorhinal), and medulla oblongata (Enfors et al., 1990a; Hohn et al., 1990; Jones and Reichardt, 1990; Kaisho et al., 1990; Maisonpierre

N. Rocamora and E. Arenas

223

TABLE 1 Expression of NT-3, NT-4/5, trkC, trkB and p75, in neuronal (CNS, PNS) and non-neuronal tissues of the adult NT-3

NT-4/5

trk-C

trk-B

p75

CNS Diencephalon Thalamus Hypothalamus Medial habenular N. Substantia nigra Superior Colliculus Raphe

++ + -

++ ++ " " " **

+ ++ + ++ + "'

++ ++ ++ ++ + '*

+ +

Basal ganglia and fore brain Medial septum Lateral septum Nucleus acumbens Caudate-putamen Ventral striatum Globus pallidus Island of Calleja

••

+ + •• •• +

++ ++ ++ ++ ++ + ++

+ ++ + + ++ ++ -

++ -

_ ++ + ++

_ ++ ++ ++

•

• ••

+ +

Neocortex Layer I Layer II-III Layer IV Layer V Layer VI-

-

+

Cingulate cortex

-

++

++

Hippocampal formation Taenia tecta Dentate granular layer Hilus Pyramidal layer Subiculum Presubiculum Parasubiculum Entorhinal cortex

+ ++ ++(CA2, mCAl) •• •• •• ++

••

+ ++ ++ + ++ + + + +

++ + ++ ++ ++ + ++ ++

Brainstem Cerebellum Purkinje cells Granule cells Pontine Nuclei

+ -

++ ++ ••

++ ++ +

++ + +

+

Olfactory system

••

+

++

++

++

Spinal cord

-

+

+

+

••

NT-3

NT-4

Trk-C

Trk-B

p75

++

_ -

-

••

224

Neurotrophin-3 and neurotrophin-4/5

TABLE 1 {continued) NT-3

NT-4/5

PNS (Ganglia) Dorsal root Small-size neurons Medium-size neurons Large-size neurons

-

Sympathetic

-

"

Cranial Trigeminal Vestibular Petrose Nodose

•• •• " ...

•• •• •• ••

Kidney Thymus Spleen Pituitary Heart Liver Testis Ovary Skin Intestine

trk-B

p75

++ ++

+ ++ ++ ++

+ +

+

ENS Remak's ganglion Non-neuronal tissues Muscle Lung

trk-C

+

++

++ +

+++ ++ ++

••

++ + + ++

++ ++

-

••

+ ++

+ ++ +

••

NT-3, Ernfors et al., 1990a; Maissonpierre et al., 1990a; Friedman et al, 1991; Gall et al., 1992. NT-4/5, Timmusk et al., 1993. Receptors, Koh et al., 1989; Merlio et al., 1992; Verge et al, 1992; Mu et al., 1993; Tessarollo et al., 1993; Altar et al., 1994c; Barbacid, 1994.

et al., 1990a; Rosenthal et al., 1990; for a review seeThoenen, 1991). The hippocampus is the main site of localization of NGF, BDNF and NT-3 in the adult CNS. However, in situ hybridization shows specific, mainly non overlapping, patterns of expression for each one. NT-3 mRNA is found in the CA2 and the most medial CAl regions of the pyramidal cell layer, and in granule cells of the dentate gyrus. NT-3 mRNA was also present in the induseum griseum and taenia tecta, both developmentally derived from the hippocampus (Lauterbom et al., 1994). Moreover, NT-3 mRNA levels were found

to increase from caudal to rostral hippocampus. By contrast, the NGF mRNA hybridization pattern showed highly labeled neurons in the dentate and hilar regions, and scattered neurons in the pyramidal cell layer (Gall and Isackson, 1989; Ernfors et al., 1990b); and BDNF mRNA in the CA3 region of the pyramidal cell layer and granule cell layer of the dentate gyrus. Thus, NGF, BDNF and NT-3 are differentially expressed in the various hippocampal regions, with the highest colocalization of the three neurotrophins in the granule cell layer of the dentate gyrus. More detailed analysis combining in situ hybridization for neurotrophins and

N. Rocamora and E. Arenas

immunocytochemistry for specific cell markers, such as the calcium-binding proteins parvalbumin, calretinin and calbindin (markers for different populations of hippocampal GABAergic interneurons), showed a specific expression of the neurotrophins in the different hippocampal cell types. NT-3 mRNA was mainly found in granule cells of the dentate gyrus and pyramidal cells of the CA2 and medial CAl regions, and also in few parvalbumin- and calretinin-positive interneurons. BDNF, in contrast, was not found to be expressed in parvalbumin-, calretinin- or calbindin-positive interneurons, and NGF was mainly expressed by parvalbumin-positive interneurons (Rocamora et al., 1996). 2.2. Regulation ofNT-S mRNA expression In the nervous system, neurotrophins are mainly expressed by neurons, suggesting that their expression may be regulated by neuronal activity. In support of this hypothesis, analysis of cultured hippocampal cells has shown that potassiuminduced depolarization and glutamate receptor stimulation induce upregulation of NGF and BDNF, whereas GABA induces downregulation (Zafra et al., 1990, 1991, 1992; Lu et al., 1991). These results, together with the pionering work of Gall and Isackson (1989) showing NGF mRNA upregulation in adult hippocampus soon after the induction of limbic seizures by an electrolytic lesion in the hilus, suggest a putative role of neurotrophins in the adult activitydependent plasticity involved in processes such as memory and learning. Further support to this hypothesis comes from the facts that hippocampus is both the area with highest expression of neurotrophins, and the structure with higher levels of plasticity in the adult brain. Moreover, upregulation of NGF and BDNF mRNAs occurs with different seizure- or activation-inducing paradigms (Ballarin et al., 1991; Gall et al., 1991; Isackson et al. 1991; Ernfors et al., 1992a; DugichDjordjevic et al., 1992; Rocamora et al., 1992, 1994; Berzaghi et al., 1993; Rumpel et al., 1993; Metsis et al., 1993; Timmusk et al., 1995), as well as after hypoglycemic coma or ischemia

225

(Lindvall et al., 1992; Takeda et al., 1993; Kokaia etal, 1994). In contrast to the upregulation of NGF and BDNF, a transient decrease of NT-3 mRNA was found after an electrolytic lesion of the hilus (Rocamora et al., 1992) and also following cerebral ischemia and hypoglycemic coma (Lindvall et al., 1992; Takeda et al. 1992, 1993; Kokaia et al., 1994). Decreased levels of NT-3 were also found after kindling (Bengzon et al., 1993), unilateral intrahippocampal injection of quinolinic acid (Rocamora et al., 1994) and intraperitoneal or intrahippocampal injection of kainic acid (Rocamora et al., unpublished). Moreover, the high-affinity receptors trkB and trkC were also upregulated in the same regions after kindling-induced seizures, mechanical injury, hypoglycemic coma and ischemia (Bengzon et al., 1993; Merlio et al., 1993; Mudo et al, 1993). Taken together, the above data shows that neurotrophins and their receptors are regulated after brain insults, in both seizure- and activity-inducing experimental paradigms. NGF and BDNF were upregulated and NT3 downregulated. trkB and C were upregulated in the same brain regions, suggesting a possible local function of neurotrophins protecting neurons against injury and/or inducing resprouting after damage. By contrast with these strong artificial stimuli, relatively brief stimuli have been used extensively to evoke long-term-potentiation (LTP) as a model for the neural basis of memory and learning. Contradictory results with respect to the NT-3 mRNA regulation have been found in different LTP paradigms. On the one hand, in vitro LTP at the synapse between Schaffer collaterals and CAl pyramidal neurons was found to induce upregulation of NT-3 mRNA expression in the potentiated CAl cells (Patterson et al., 1992). On the other hand, in vivo LTP at the perforant path-dentate granule cell synapse resulted in an upregulation of NGF and BDNF, together with a slight downregulation of NT-3, in granule cells (Castren et al., 1993). The difference in the potentiated synapse and/or the experimental model could explain these conflicting results. Excitotoxins induce both neuronal stimulation

226

and toxicity, depending on the dose. Unilateral, intrahippocampal injection of excitotoxins allows separate analysis of their local and distal effects. A long-lasting upregulation of NT-3 mRNA expression in the adult brain was found locally in the degenerating ipsilateral CAl and CA4 pyramidal layers, after unilateral injection of the NMDA agonist, quinolinic acid (Rocamora et al., 1993b). By contrast, in the contralateral side, transient downregulation of NT-3 together with upregulation of NGF and BDNF were found (Rocamora et al., 1993b). Thus, opposite effects on regulation of NT-3 with respect to BDNF were found as a result of both local and distal effects of quinolinic acid. Although the NT-3 mRNA hybridization signal followed exactly the timecourse of degeneration in the pyramidal cell layer, it is not clear whether the cells upregulating NT-3 were degenerating pyramidal neurons or reactive glial cells. Interestingly, upregulation of NT-3 mRNA expression has been found to be induced both, in astrocytes by 1,25-dihydroxyvitamin D3 (Neveu et al., 1994) and in neurons by glucocorticoids (Barbany and Persson, 1992) and thyroid hormone (Leingartner et al, 1994). Increased levels of NT-3 mRNA were found in the dentate gyrus and hippocampal CAl and CA2 regions after immobilization stress. By contrast, BDNF mRNA levels were decreased, mainly in dentate gyrus (Smith et al., 1995). The decreased BDNF mRNA persisted after adrenalectomy, while the decrease of NT-3 mRNA was prevented by adrenalectomy, suggesting a dependency upon glucocorticoids. Although hormonal regulation of NT-3 expression is suggested by the above results, analysis of 5' regulatory sequences of the human (Shintani et al., 1993), and mouse (Leingartner and Lindholm, 1994) NT-3 genes did not reveal the presence of any hormone-responsive elements. Two promoters with silencing elements, and two different transcripts, A and B, were found. The presence of negative regulatory elements in most upstream sequences has been suggested because of the downregulation of NT-3 expression induced by long 5' flanks (Leingartner and Lindholm, 1994). Both positive and negative influences have been shown to regulate NT-3 mRNA expression. In

Neurotrophin-3 and neurotrophin-4/5

cerebellar granule cells, T3 has been found to upregulate transcript B in vitro, but had no effect over transcript A (Leingartner et al., 1994). In vitro and in vivo experiments have shown that BDNF and NT-4/5 increased the levels of NT-3 mRNA, in PI hippocampus and in immature hippocampal cells in culture (Lindholm et al., 1994a), and that BDNF induced NT-3 upregulation in cerebellar granule cells (Leingartner et al., 1994). Neurotrophins have been found to be differentially regulated also after peripheral nerve injury. While BDNF and NT-4/5 mRNAs levels increased markedly in the distal segment of the sciatic nerve after transection, NT-3 mRNA decreased rapidly and returned to basal levels over two weeks. In the denervated gastrocnemius muscle, BDNF mRNA increased, NT-3 mRNA did not change, and NT4/5 decreased. Only small changes of neurotrophins and trk mRNAs levels were found in the spinal cord (Funakoshi et al., 1993). These results suggest that the different neurotrophins could cooperate by both target-derived and local effects in the regeneration of injured peripheral nerves. In summary, NT-3, a very primitive neurotrophin with respect to its sequence conservation, its presence in lower vertebrates and its promiscuous binding to the different components of the trk family, has been found to be regulated in a manner opposite to NGF and BDNF. While neuronal activity and sciatic nerve transection induce NGF and BDNF upregulation, NT-3 is downregulated. Conversely, excitotoxicity, glucocorticoids and stress upregulate NT-3 mRNA and downregulate BDNF mRNA. From an evolutionary perspective, this could mean that (i) NT-3 lacks a specific sequence responsible for activity-dependent upregulation in NGF and BDNF and/or (ii) NGF and BDNF lost a negative regulatory sequence present in the NT-3 gene. A more extensive analysis of the regulatory elements in neurotrophin genes may shed light on this issue. 2.3. Functional aspects Three lines of evidence defined NGF as a prototypic neurotrophic factor: (i) sympathetic and sensory neurons require NGF to survive when they

N. Rocamora and E, Arenas

are innervating their targets (for a review see LeviMontalcini, 1987), (ii) NGF is synthetized in sympathetic and sensory target fields and (iii) NGF is retrogradely transported from target fields to the cell bodies of innervating neurons (Stokel and Thoenen, 1975; Jonhson et al, 1978). Interestingly, the distribution patterns of neurotrophins and their receptors during development and in the adult have suggested additional roles for these factors, including effects on proliferation, migration and differentiation of neurons. In the CNS, according to the classical neurotrophic theory, neurotrophins are secreted from the soma and/or dendrites of the target neuron, taken up by the axon terminals of the afferent neuron and transported to the cell body. Thus, capacity for retrograde transport has been considered as indicative of neurotrophin responsiveness (DiStefano et al., 1992). However, anterograde transport of neurotrophins released at the axon terminal and acting on targets such as Schwann cells (Schecterson and Both well, 1992), or Purkinje cells (Lindholm et al., 1993a) has also been suggested. In support of this possibility, trkB and trkC immunoreactivities have been found on neuronal dendrites and cell bodies (Okazawa et al., 1993). However, it remains to be established whether this represents the normal trafficking of receptors, present in the cytoplasm before being transported and inserted in the membrane and/or after internalization, before signaling or catabolism in the cell body. Finally, the high degree of colocalization of trkB and trkC with their specific neurotrophin-ligands, BDNF and NT-3 respectively, especially in the hippocampus (Merlio et al., 1992; Kokaia et al., 1993; Miranda et al., 1993; Lamballe et al., 1994), suggests a local autocrine and/or paracrine function for these neurotrophins. Data on expression of NT-3 and its regulation suggest that neurotrophins are involved not only in development, but also in activity-dependent plasticity of the adult nervous system and in the protection of neurons against injury. These functional aspects are reviewed in the next sections. 23.1. Actions of NT-3 in vitro Data on NT-3 activity in vitro comes from the

211

analysis of the effects of the recombinant protein, blocking antibodies, or antisense oligonucleotides on the survival and/or differentiation of primary cell cultures and/or explanted ganglia. An intrinsic characteristic of the in vitro studies is that they are limited to proliferating immature cells, developing neurons and tumor immortalized cell lines. The oversimplification of this system is the major limitation when deriving physiological implications from the in vitro results. A closer parallel to in vivo conditions is organotypic culture, in which local circuitry is preserved. Other in vitro studies take advantage of neuronal and non-neuronal cell lines (PC 12 rat pheochromocytoma cells, NIH 3T3 fibroblasts), to study molecular and/or cellular mechanisms involved in neurotrophin signal transduction pathways. Peripheral nervous system. Because of the classical views on neurotrophic activity, most of the work on survival-promoting effects of neurotrophins has been performed soon after the neurons contact their targets, when apoptosis takes place. Results from these studies have shown that survival of neurons is promoted by a particular neurotrophin or combination of them. In contrast, sensory neurons isolated prior to target contact survive independently of neurotrophins (Davies and Lumsden, 1984; Vogel and Davies, 1991), but they require neurotrophins to promote their maturation (Wright et al., 1992). Neurotrophic activity has been traditionally determined by analysing the ability of a given factor to induce survival and/or differentiation of embryonic cultured cells and/or to induce neurite outgrowth from explanted embryonic ganglia of the peripheral nervous system. Results from these studies have shown that the different neurotrophins give specific as well as overlapping trophic support to PNS ganglia. All the known neurotrophins have been found to exert trophic actions on dorsal root ganglion (DRG) neurons. NT-3 has been shown to support 60% of the DRG neurons, while BDNF and NGF together support 100%, suggesting that all the cells responsive to NT-3 are also responsive to either BDNF and/or NGF (Lindsay et al., 1985, Maisonpierre et al., 1990a). More de-

228

tailed analysis showed distinct specificities of different neurotrophins. Each DRG is composed of several different neuron types. Large neurons (processing sensory propioceptive information) have been found to mainly express trkC, medium sized neurons (processing mechanoceptive information) trkB, and small neurons (processing sensory nociception, pain and thermal sensation) trkA (Mu et al., 1993). NT-3 promotes the survival of propioceptive sensory neurons in culture (Hohn et al., 1990) and their peripheral and central targets (muscle spindles and motoneurons, respectively) express NT-3 (Emfors and Persson, 1991; Schecterson and Both well, 1992; Henderson et al., 1993). Moreover, DRGs distributed along the spinal cord are different depending on the segmental level. Although NGF and BDNF support DRG cells from all segmental levels, NT-3 preferentially induces robust neurite outgrowth from cervical and lumbar DRG, which innervate limb muscles (Hory-Lee et al., 1993). In the nodose ganglion, both NT-3 and BDNF have been shown to have trophic effects, with the response to NT-3 being greater than that to BDNF. Finally, neurons from the sympathetic ganglia are supported by NGF and NT-3, with the response to NT-3 being substantially lower and delayed compared to that of NGF (Maisonpierre et al, 1990a). A switch in neurotrophin requirements has been found for neurons which depend only on NGF for survival (sympathetic, dorsomedial trigeminal, jugular, some DRG cells; Buj-Bello et al. (1994)). A wide range of these neurons show a switch from a transient early dependence on NT-3 and/or BDNF to a later dependence on NGF (for a review see Davies, 1994). For instance, up until ElO, survival of trigeminal ganglion neurons is dependent on BDNF and/or NT-3; later, when naturally programmed cell death takes place (E13-E18) these neurons require only NGF (Buchman and Davies, 1993). Similarly, the survival of NGFdependent, but not BDNF-dependent, cranial sensory neurons is promoted by different neurotrophins early in their development (Buj-Bello et al., 1994). NT-3 is a survival (trophic) factor for E14.5 cultured proliferating sympathetic precursors (Birren et al., 1993; DiCicco-Bloom et al., 1993)

Neurotrophin-3 and neurotrophin-4/5

and peripheral sensory neurons (Gaese et al., 1994). NT-3 promotes survival and cell cycle arrest of proliferating sympathetic precursors (Verdi and Anderson, 1994). Once arrested, these cells express trkA, which accelerates NGF-induced differentiation. Thereafter, NGF induces expression of p75 and the cells became dependent on NGF for survival (Hempstead et al., 1992; Verdi and Anderson, 1994). Other functions of NT-3 early in development include induction of proliferation and differentiation of neural crest progenitor cells (Kalcheim et al., 1992, Chalazonitis et al., 1994), maintain survival of proliferating sympathetic neuroblasts and peripheral sensory neurons (Birren et al., 1993; DiCicco-Bloom et al., 1993; Gaese et al., 1994), promote survival and proliferation of oligodendrocyte precursors (Barres et al., 1993), and regulate the differentiation of sympathetic neuroblasts (Verdi and Anderson, 1994) (Fig. 1). NT-3 has been suggested to be a musclederived trophic factor for motoneurons, because (i) NT-3 is expressed in skeletal muscle (Schecterson and Bothwell, 1992; Funakoshi et al., 1993; Henderson et al., 1993; Koliatsos et al, 1993), (ii) motoneurons express both trkB and trkC (Hebderson et al., 1993) and (iii) [i25i]NT-3 is retrogradely transported by motoneurons (DiStefano et al., 1992; Koliatsos et al., 1993). In agreement with this, NT-3 and also BDNF were found to enhance synaptic strength of the Xenopus neuromuscular synapse in vitro (Lohof et al, 1993), and to increase choline acetyltransferase activity (Wong et al., 1993) and survival of motomeurons (Henderson et al., 1993; Huges et al., 1993). In addition, the presence of trkB and trkC in chicken skeletal muscle has suggested a local function of NT-3 in muscle (Escandon et al., 1994). Central nervous system. Primary cultures have been extensively used to study the effects of neurotrophins in the developing CNS. The higher complexity of the system and its web-like connectivity makes difficult to evaluate the physiological implications of these results. According to the classical model of targetderived mode of action, neurotrophins would be

N. Rocamora and E. Arenas

229

NEUROGENESIS

PRECURSORS

PROLIFERATION/SURVUVAL DIFFERENTIATION

TARGET INNERVATION

EARLY NEURONS

MATURATION DIFFERENTIATION MIGRATION

PROGRAMMED CELL DEATH

SURVIVAL

NT-3

I

I

Fig. 1. Different functions of NT-3 during development of PNS.

expected to support afferent neurons. Basal forebrain, locus coeruleus, and raphe project to the hippocampus, the region with highest levels of neurotrophins. Furthermore, these regions have been shown to express trkB and C. In vitro studies performed on embryonic cultures showed that BDNF, NT-3 and NT-4/5 increased ChAT activity and survival of cholinergic neurons of the basal forebrain (Friedman et al., 1993) and survival of noradrenergic neurons in the locus coeruleus increased in the presence of NT-3 and NT-4/5 (Friedman et al., 1993) Moreover, BDNF and NT3 increased serotonin uptake in raphe nucleus, suggesting a role of these factors in the maturation of serotonergic neurons (Lindholm et al., 1994b). Although the entorhinal cortex is the main afferent input to the hippocampus, forming part of the main hippocampal trisynaptic circuit (entorhinal cortexdentate granule cells- CA3 pyramidal-CAl pyramidal), trophic support to the entorhinal cortex from neurotrophins has not been directly analysed. In addition to possible target-derived functions, the presence of neurotrophins in the hippocampus, together with trkB and trkC but not trkA receptors, and assessment of retrograde transport of neurotrophins (DiStefano et al., 1992) suggest a local

autocrine/paracrine function for BDNF, NT-3 and NT-4/5, but not NGF, in this brain area, in vitro studies performed in embryonic dissociated cultures and postnatal PO organotypic slices showed that pyramidal neurons of CAl to CA3, dentate granule and hilar cells are responsive to NT-3. Clear effects of NT-3, BDNF and NT-4 but not NGF, have been found on the expression of calbindin-D28k (Collazo et al., 1992, Ip et al., 1993b). In addition, NT-3 has been found to enhance neurite outgrowth and branching in a dosedependent manner and to accelerate neuronal polarization in hippocampal cultures from El8 rat embryos (Morfini et al., 1994). Recently, both NT3 and bFGF have been shown to act as differentiation factors for calbindin-expressing hippocampal neurons, while only bFGF had mitogenic activity on hippocampal precursors (Vicario-Abejon et al., 1995). In fact, NT-3 has been reported to antagonize the proliferative effects of bFGF and to enhance neuronal differentiation of cortical neurons during neurogenesis (Gosh and Greenberg, 1995). NT-3 also protects embryonic hippocampal and cortical neurons from energy-deprivation- and excitotoxin-induced neuronal cell degeneration

230

Neurotrophin-3 and neurotrophin-4/5

TABLE 2 Effects of NT-3 and NT-4/5 in vitro NT-3I

NT-4/52 D

PNS Dorsal root ganglion^ Propioceptive neurons*' Sympathetic ganglia^ Cranial ganglia Trigeminal^ Jugular^ Nodose^ ENS8 Remak's ganglion^ CNS Striatum^ GABAergic Calbindin positive Substantia NigraJ Dopaminergic GABAergic Raphe Serotonergic^ Locus coeruleus Noradrenergic* Septum Cholinergic"^ Hippocampus" calbindin positive AChE positive Cortical neurons^ GABAergic Retinal ganglion neuronsP Cerebellum'! Granule cells Purkinje cells Spinal cord*" Motorneurons^

+ + + + + +

+ + +

+<.(d) +

+ +

+ +

+ +

+ +

+ +

+

+ +

P p,se

+

+ + + + +

+p,se

+1 + + + + +

+ +

+ +

S, Survival; D, differentiation; ('^^differentiation of precursors, p: protection, se: synaptic strength. ^^Maisonpierre et al, 1990a; Rosenthal et al., 1990; Gaese et al., 1994. ^^Berkemeier et al.,1991; Ryden et al., 1995. ^'^Hory-Lee et al., 1993. '^^DiCicco-Bloom et al., 1993; Birren et al., 1993. ^^^Buchman and Davies, 1993. '^^Davies et al., 1993; Ibdiiez et al., 1993a. ^*'^*Davies, 1994. ^^Oavies et al., 1993. Navies et al., 1993; Ibanez et al., 1993a. sichalazonitis et al., 1994. ^'^-^^'^Ventimiglia et al., 1995. ^^Ardelt et al., 1994; Widmer and Hefti, 1994a. J^'^Hyman et al., 1994; Studer et al., 1995. J^Hynes et al., 1994. ^^^Lindholm et al., 1994b. **'¥riedman et al., 1993. "^^'^Friedman et al., 1993. "^'^Collazo et al., 1992; Ip et al., 1993. "*Morfini et al., 1994; Vicario-Abej6n et al., 1995; Cheng and Mattson,1994; Kang and Schuman, 1995. "^Ip et al., 1993a; LeBmann et al., 1994; Cheng et al., 1994. '^^Cheng and Mattson, 1994; Gosh and Greenberg, 1995. °^Cheng et al., 1994; Widmer and Hefti, 1994b. P^Cohen et al., 1994. ^l^'^Gao et al., 1995a. '"^'^Kato and Lindsay, 1994; Wong et al., 1993. ^*Lohof et al., 1993. ^I'^Wong et al., 1993; Henderson et al., 1993; Huges et al., 1993.

N. Rocamora and E. Arenas

(Cheng and Mattson, 1994). Moreover, in 40 day old hippocampal slice cultures, both BDNF and NT-3 produced a dramatic and sustained enhancement of synaptic strength at the Schaffer collateral-CAl synapses (Kang and Schuman, 1995). This enhancement was blocked by K252a, an inhibitor of receptor tyrosine kinases, suggesting the involvement of trk receptors. Another system in the brain, the basal ganglia, has also been subject of particularly intense research. BDNF, NT-3 and NT-4/5, but not NGF, have shown clear effects on differentiation of both dopaminergic and GABAergic neurons of the ventral mesencephalon. Moreover, NT-3 has also been shown to promote the survival of both populations of neurons (Hyman et al., 1994, Studer et al., 1995). In addition, the presence of trkB and trkC receptors in the striatum, together with the existence of receptor-mediated retrograde transport of BDNF and NT-3 from the striatum to the substantia nigra in adult rat brain (Wiegand et al., 1992) further supports a physiological significance of these effects. In the striatum, BDNF, NT-3 and NT-4 have been suggested to promote survival and differentiation of both calbindin-containing and GABAergic striatal neurons (Ventimiglia et al., 1995). Finally, the expression of BDNF and NT-3 mRNA in substantia nigra and ventral tegmental area (Gall et al., 1992), suggests both a local effect for these neurotrophins in the substantia nigra and a classical target-derived trophic effect on GABAergic striato-nigral projecting neurons. Dissociated and organotypic cultures of rat cerebellum showed that BDNF and NT-3 affect granule cells at different stages of differentiation. Early neurons of the external granule layer have been shown to respond to BDNF, enhancing their survival, and only more mature granule cells induce c-Fos in response to NT-3 (Segal et al., 1992; Lindholm et al., 1993b). In addition, BDNF and NT-4/5, but not NT-3, have been recently reported to act at later stages of cerebellar granule cell differentiation, promoting neural extension and survival (Gao et al., 1995a), despite mature granule cells have been shown to respond to NT-3 by inducing c-Fos (Segal et al., 1992). This induction of c-Fos could be related to the recent finding of a

231

role for NT-3 in altering the morphology of granule neurons (Segal et al., 1995). High levels of trkB were found in the differentiated granule cells (Gao et al., 1995a) and also in the neonate, while higher levels of trkC were found in the adult cerebellum (Segal et al., 1995). However, NT-3 seems to play a more prominent role on Purkinje cells. The high levels of NT3 mRNA in the early postnatal granule cells, together with the presence of trkC in Purkinje cells, suggested a role for NT-3 on these cells. In fact, NT-3 has been found to increase survival of cultured calbindin-positive Purkinje cells (Mount et al., 1994) and to induce hypertrophy and neurite sprouting of Purkinje cells (Lindholm et al., 1993a). The fact that thyroxine causes both upregulation of NT-3 in granule cells and enlargement of Purkinje cells (Lindholm et al., 1993a) and the dramatic increase in Purkinje cell numbers when co-cultured with granule cells (Baptista et al., 1994) suggest that the effects of thyroxine on Purkinje cells could be mediated by NT-3 release from granule cells. All these data support a role for anterogradely transported NT-3 in Purkinje cell maturation (Lindholm et al., 1993a). 2.3.2. Actions of NT-3 in vivo Information concerning the function of NT-3 in the nervous system in vivo comes from experiments based either on deprivation or administration of the factor. The deprivation studies have utilized mutant mice with gene-targeted deletions of NT-3 gene by homologous recombination or have employed blocking antibodies. These techniques have provided valuable insights into the function of endogenous NT-3. The second approach has been based on administration of exogenous NT-3 or on the grafting of cell lines engineered to express and secrete high levels of NT-3 in vivo. These studies have provided information concerning the actions of NT-3 both during development and in the adult. Furthermore, the effects of NT-3 on neurons degenerating in animal models of disease have suggested a potential therapeutic application of this factor in neurodegenerative disorders.

232

Role of neurotrophin-3 during development: deprivation studies. Administration of NT-3-blocking antibodies in vivo by injection of anti-NT-3 secreting hybridomas has been shown to reduce the proliferation of oligodendrocyte precursors in the optic nerve of the rat (Barres et al, 1994) and to reduce the number of neurons in the placodederived nodose ganglia and the neural crestderived DRG in chick embyos, during early development (Gaese et al., 1994). These experiments show that endogenous NT-3 is necessary at early stages of development for functions as diverse as promoting the proliferation of oligodendrocyte precursors and ensuring the survival of sensory neurons. Targeted deletions of the NT-3 gene (Emfors et al.,1994; Farinas et al., 1994) have demonstrated that NT-3 is not only required for the correct development of the nervous system, but is necessary for life. Homozygous mutant mice die between postnatal days 0 and 21, are smaller in size and show pronounced limb ataxia and fixation of limbs in extension. Detailed analysis of NT-3 knockout animals has demonstrated a severe loss of sensory and sympathetic neurons, without loss of motor neurons, suggesting that NT-3 is required for the development of both sensory and sympathetic neurons, but not motor neurons. Sensory neurons involved in specialized senses such as hearing and balance have also recently been shown to be affected in these animals (Emfors et al., 1995). There is not only a 34% loss of neurons in the vestibular ganglion and a 87% loss of neurons in the spiral ganglion, but absence of the central processes from the spiral ganglion to the cochlear nuclei in the brainstem. In the NT-3 knockout mice, most sensory neuron populations were reduced in number. These alterations affected several sensory modalities, including somatosensory information processed by the DRGs and by the trigeminal ganglion, interoceptive visceral information conveyed by petrosenodose ganglion cells and propioceptive information processed by DRG neurons and by the mesencephalic nucleus of the trigeminal nerve. During normal development, innervation of skeletal muscle by la and lb DRG neurons induces formation

Neurotrophin-3 and neurotrophin-4/5

of the proprioceptive muscle spindles and Golgi tendon organs, respectively. Interestingly, lack of NT-3 during development leads to the absence of propioceptive neurons from the DRGs, as evidenced by the absence of carbonic anhydrase- and parvalbumin-positive neurons. Consequently, both peripheral projections to the muscle proprioceptive sense organs and central projections to the layer IX motor neurons of the spinal cord are missing. Furthermore, the propioceptive sense organs normally induced by la neurons (muscle spindles) and by lb neurons (Golgi tendon organs) are absent. Thus, all components of the proprioceptive pathway, from sense organ to central afferent to gamma-motoneuron are absent. These alterations may explain the limb ataxia observed in these animals, although a possible contribution of defects in intrinsic populations of neurons of the central nervous system cannot be excluded. Analysis of NT-3 and trkC gene knockout mice have not yet identified any deficits in intrinsic CNS neurons (Emfors et al., 1994; Farinas et al., 1994; Klein et al., 1993), but the broad pattem of expression of trkC in the brain suggests that there may be subtle changes not apparent in gross routine histological examination. Administration of NT-3, Alhough [125I]NT-3 injected in the gastrocnemius muscle of rats has been reported to be retrogradely transported by spinal motoneurons (DiStefano et al., 1992; Koliatsos et al., 1993), the administration of NT-3 in newborn rats has been shown to only partially rescue (Sendtner et al., 1992) or not rescue (Koliatsos et al., 1993) facial motoneurons from death after axotomy. These results, together with the absence of a reduction in the number of facial motoneurons in the NT-3 knockout mice (Ernfors et al, 1994; Fariiias et al., 1994) suggest a minor role of NT-3 in the survival of motor neurons during development. Retrograde transport of NT-3 has also been reported (DiStefano et al., 1992; Wiegand et al, 1992), suggesting a role for NT-3 in vivo. Accordingly, in utero injections of NT-3 for 2 days have also shown to increase the soma area of large DRG neurons at embryonic day 16 (Zhang et al., 1994) and, in the adult, administration of NT-3

233

N. Rocamora and E. Arenas TABLE 3 Neuronal populations responsive to NT-3 administration in vivo

Development Motorneuron Facial Adult Locus coeruleus Noradrenergic neurons Substantia nigra Dopaminergic neurons Corticospinal neurons

Development Dorsal root ganglion Corticospinal tract Anterior cortex and striatum Hippocampus Adult Dorsal root ganglion Corticospinal tract Basal forebrain Cholinergic neurons Substantia nigra Dopaminergic neurons Raphe nucleus Serotonergic neurons

Striatum Cholinergic neurons

Survival

Lesion model

Reference

Partial No

Axotomy in newborn Axotomy in newborn

Sendtner et al., 1992 Koliastos et al., 1993

80%

6-OHDA in adult

Arenas and Persson, 1994

Yes Yes

Axotomy Axotomy

Hagg, 1994 Giehl and Tetzlaff, 1994

Function

Model

Reference

(+)NPY, somatostatm

Embryo Newborn Neonate

Zhang etal., 1994 Schnell et al., 1994 Carnahan and Nawa, 1995

and substance P (SP) ^•*'^Somatostanin and SP

Neonate

Carnahan and Nawa, 1995

^•^^Sensory conduction velocity ^"•"^Regenerative sprouting

Cisplatin Axotomy

Gaoetal., 1995b Schnell et al., 1994

^"^Atrophy ^"•"^Spatial memory

Aged Aged

Fischer et al., 1994 Fischer et al., 1994

^"^Nigral dopamine turnover ^"^Ipsil.turning behaviour

Partial DA lesion

Martin-Iverson et al., 1994 Altar etal., 1994b

Hypertrophy ^~^Axon collaterals (+>Collateral sproutmg

^^^Striatal serotonin turnover ^~^Food intake and body weight ^•^^Antinociception ^"^Atrophy

Martin-Iverson et al., 1994 Siuciak et al., 1994 Aged

Fischer etal., 1994

^"••^Increase/induce, ^ decrease.

reduces the impairment of sensory velocity conduction induced by cisplatin (Gao et al, 1995b). These results are consistent with the role of NT-3 on propioceptive neurons suggested by in vitro and knockout studies (see Section on: Role of neurotrophin-3 during development: deprivation studies). NT-3 has also been shown to modify axon sprouting during developemnt and in adult

hood. Interestingly, collateral sprouting from embryonic DRG neurons into the gray matter was inhibited by NT-3 (Zhang et al., 1994), while it stimulated sprouting from corticospinal neurons (Schnell et al., 1994). In this latter report, injection of NT-3 into the spinal cord of P2 rats was reported to induce premature collateral sprouting from the corticospinal tract at P5 while, in adult

234

rats, injection of NT-3 into the lesioned spinal cord increased the regenerative sprouting of transected corticospinal fibres. Altogether, these results suggest that NT-3 can act either as a selective inhibitor or as an inducer of fibre growth of different neuronal populations during development and regeneration. Retrograde transport studies of [^^^I]NT-3 have suggested that, in addition to DRG neurons and spinal cord motor neurons, other neuronal systems may be able to respond to NT-3, in the adult in vivo (DiStefano et al, 1992; Wiegand et al., 1992). NT-3 has been reported to prevent the atrophy of both basal forebrain and striatal cholinergic neurons which occurs during aging (Fischer et al., 1994). Furthermore, NT-3 was shown to improve the associated functional decline of spatial memory in aged rats. In the basal ganglia, NT-3 has been shown to prevent the death of axotomized nigral dopaminergic neurons (Hagg, 1994), while supranigral infusions of NT-3 increase both dopamine and serotonin metabolism in the striatum (Martin-Iverson et al., 1994). NT-3 has also been shown to decrease turning behavior induced by partial dopamine lesions (Altar et al., 1994b) and to induce analgesia, possibly via serotonergic mechanisms (Siuciak et al, 1994). NT-3 also prevents degeneration of the noradrenergic neurons of the locus coeruleus following lesions (Arenas and Persson, 1994) and corticospinal cells of layer V (Tetzlaff et al., 1994). Interestingly, neurons in the locus coeruleus die in most degenerative disorders affecting the brain (see Table 6). Thus, the survival-promoting effects of NT-3 in the locus coeruleus (Arenas and Persson, 1994), as well as that in the dopaminergic substantia nigral neurons (Hagg, 1994), suggest the possibility of important therapeutic applications. Finally, NT-3 has recently been shown to increase the levels of several neuropeptides, including somatostatin and substance P, in cortex, hippocampus and striatum of neoanatal rats (Camahan and Nawa, 1995). If this action persists in adult animals, it may offer the possibility of NT3 being used in Alzheimer's disease to rescue not only basal forebrain cholinergic neurons but also cortical peptidergic neurons.

Neurotrophin-3 and neurotrophin-4/5

3. Neurotrophin-4/5 Neurotrophin-4 (NT-4) was cloned from Xenopus and Viper genomic DNA, by virtue of its homology with NGF, BDNF and NT-3 (Hallbook et al., 1991). In parallel, and following the same strategy, neurotrophin-5 was cloned from human DNA (Berkemeier et al, 1991). Thereafter, a mammalian neurotrophin-4 gene and a pseudogene were cloned from human and rat, using specific sequences of Xenopus/viper NT-4 (Ip et al., 1992), and were found to be the same protein; this was therefore named NT-4/5. Comparison between Xenopus NT-4 and mammalian NT-4/5 showed that this fourth neurotrophin is the less conserved, with only a 65% amino acid identity between Xenopus and the mammalian counterparts (compared with 93% for NGF, 95% for BDNF, and 100% for NT-3). Sequence analysis of mammalian NT-4/5 discovered an insertion of seven amino acids with respect to Xenopus NT-4 and the other neurotrophins in the less conserved region among neurotrophins. Binding experiments showed that NT-4/5 shares the trkB receptor with BDNF and the low-affinity receptor P75LNGFR ^jt^ all the other neurotrophins (Berkemeier et al., 1991; Hallbook et al., 1991; Ip et al., 1992). NT-4/5 mRNA was detected in human in a limited number of peripheral tissues, as thymus, placenta, skeletal muscle and skin (Ip et al., 1992). In rat embryo, it was detected already at El2, and can be found in both head and body at El8. In the adult, NT-4/5 mRNA was found in thymus, muscle, lung, ovary, brain, heart, stomach and kidney (Berkemeier et al., 1991). Analysis of the distribution of other neurotrophins in the brain has mainly been performed by Northern and in situ' hybridization. However, the levels of NT-4/5 mRNA are practically undetectable by these methods. Timmusk et al. (1993), using a highly sensitive RNAse-protection assay have performed a detailed analysis of NT-4/5 mRNA expression, both during development and in the adult. NT-4/5 mRNA levels in thymus, thyroid, muscle, lung and ovary were found to be higher than in brain. In the brain, NT-4/5 mRNA was widely distributed, with similar levels, being highest in pons-medulla, hypo-

N. Rocamora and E. Arenas

thalamus, thalamus and cerebellum (see Table 1). Interestingly, the expression of NT-4/5 in the brain varies during development, with the highest levels at El3 (coinciding with the time of target field innervation), a decrease around birth and a subsequent increase once again. NT-4/5 mRNA expression in the CNS has not been found to be regulated by neuronal activity. However, its level in skeletal muscle depends on muscle activity (Funakoshi et al., 1995), and NT4/5 mRNA decreases in denervated gastrocnemius muscle (Funakoshi et al, 1993). These data suggest a role for NT-4/5 in activity-dependent neuromuscular performance (Funakoshi et al., 1995). 3.1. Actions of nt-4/5 in vitro The distribution of NT-4/5 mRNA in peripheral tissues, together with the presence of its high- and low-affinity receptors (trkB and p75, respectively) in peripheral ganglia, suggest a role for this neurotrophin as a target-derived trophic factor for neurons of the PNS. In agreement with this, mouse trigeminal and jugular sensory neurons are supported transiently by NF-4/5 at early stages of target field innervation, with a later switch in neurotrophin dependence to NGF (Davies et al., 1993). Thus, NT-4/5 as well as NT-3 and BDNF has trophic effects before the phase of naturally occurring cell death (Davies et al., 1993). NT-4/5 also elicits neurite outgrowth from explanted embryonic trigeminal ganglia and promotes survival of dissociated trigeminal ganglion neurons and BDNF-dependent nodose neurons during the phase of naturally occuring cell death (Ibaiiez et al., 1993a). NT-4/5 mRNA was expressed in the developing whisker follicles of the mouse (decreasing from El3 to E19), with a different cellular distribution to those of the other neurotrophins (Ibanez et al., 1995) and, in addition, all three trk receptors were present in the developing trigeminal ganglia. Together, these data suggesta role for all four neurotrophins as target-derived trophic factors in the development of the trigeminal ganglion-whisker connection (Ibanez et al., 1993a).

235

Although the role of p75^N^^^ in mediating neurotrophin function is not clear, functional analysis with mutant neurotrophins, deficient in lowaffinity binding regions, has revealed a role for p75LNGFR ij^ regulating biological responsiveness to NT-4/5. The ability of an NT-4/5 mutant (with disrupted ability to bind p75, but not trkB) to promote neurite outgrowth from nnr5PC12-TrkB cells, expressing both TrkB and p75, and to promote survival of El6 rat DRG neurons, were drastically reduced, by 15 and 30 fold, respectively, when compared to the wt NT-4/5 (Ryden et al., 1995). Interestingly, retrograde transport of NT4/5 to DRG has been shown to depend selectively on p75LNGFR and, in contrast to BDNF, NT-4/5 was transported to motor neurons after axotomy, which upregulates P75LNGFR (Curtis et al, 1993). Like all neurotrophins except NGF, NT-4/5 has been shown to support the survival of cultured motor neurons (Henderson et al, 1993; Huges et al, 1993). In addition, there is an increase in choline acetyltransferase in both fetal human and rat spinal cord neurons after treatment with NT-4/5, NT-3 or BDNF (Wong et al, 1993; Kato and Lindsay, 1994). NT-4/5 has survival and/or differentiation effects on a wide range of brain neurons. Survival effects have been shown on embryonic cholinergic neurons of the basal forebrain, noradrenergic neurons of the locus coeruleus (Friedman et al, 1993), and dopaminergic mesencephalic neurons (Hynes et al, 1994). In addition, NT-4/5 was also shown to promote differentiation effects on GABAergic cells of the ventral mesencephalon (Hyman et al, 1994) and striatum (Widmer and Hefti, 1994b). Morphological analysis of the structure of mesencephalic dopaminergic cells after treatment with different neurotrophins showed that NT-4/5 had the most important effect on the structural differentiation of these neurons in culture (Studer et al, 1995). Moreover, in the striatum, NT-4/5 not only promoted morphological and biochemical differentiation of GABAergic neurons, but also their survival, both in organotypic slice cultures (Ardlet et al, 1994), and in cultured dissociated neurons (Widmer and Hefti, 1994a; Ventimiglia et al, 1995).

236

In the hippocampus, NT-4/5 induced the differentiation of El 8 cultured neurons (Ip et al, 1993a), and interestingly, both NT-4/5 and BDNF have been shown to upregulate NT-3 mRNA expression (Lindholm et al., 1994a). BDNF and NT-4/5 were also both found to enhance glutamatergic synaptic transmission in cultured hippocampal neurons (LeBmann et al., 1994), similarly to what has been reported for BDNF and NT-3 at the neuromuscular junction (Lohof et al, 1993). In addition, NT-4/5 has a protective effect in hippocampal and cortical neurons against energy deprivation- and excitatory amino acid-induced injury (Cheng et al., 1994) and stimulates the differentiation of GABAergic cortical neurons in culture (Widmer and Hefti, 1994b). Finally, other reported functions of NT-4/5 include the induction of granular cell differentiation at late stages of cerebellar development (Gao et al., 1995a), and the promotion of survival and neurite outgrowth of retinal ganglion neurons in cultured explants from adult rats (Cohen et al, 1994). In summary, NT-4/5 is a survival/differentiation factor for a wide range of neurons. In some instances, its action is redundant; it is more potent than BDNF and it is generally synergistic with NT-3. The trophic effect of NT-4/5 is particularly potent for striatal, cortical and mesencephalic GABAergic cells, and for motoneurons. In order to answer whether these in vitro actions could be important from the therapeutic and/or physiological viewpoints, results from knockout animals, and from the administration of NT-4/5 in vivo will be discussed next. 5.2. Actions of NT-4/5 in vivo Information concerning the function of NT-4/5 in the nervous system in vivo comes either from targeted deletions of the NT-4/5 gene, or from the admininstration of the protein or the grafting of cell lines engineered to secrete high levels of NT4/5. 3,2.1. Role of NT-4/5 during development: NT-4/5 knockout Mice homozygous for the NT-4/5 mutation appear not to have any deficits compromising their

Neurotrophin-3 and neurotrophin-4/5

survival. Growth and righting responses of NT-4/5 - / - mice were similar to controls and they were able to mate and produce viable offspring (Conover et al., 1995; Liu et al., 1995). The main alteration found in the NT-4/5 - / - mice was a 60% cell loss in the nodose-petrosal complex, where visceral sensory information relays for the regulation of several autonomic functions, including respiration, heart rate and blood pressure. Given the fact that the viability of the animals was not impaired, these data suggested that the visceral afferents supported by NT4 innervate targets that are not critical for cardiopulmonary homeostasis or survival (Conover et al., 1995). In addition, a 50% loss of geniculate ganglion neurons has been found (Liu et al., 1995), but again with no evidence of important functional alterations. In summary, the fact that the viability of NT-4/5 - / - animals is not compromised, together with the absence of obvious alterations in routine histological examinations, suggests that NT-4/5 may have more subtle functions than anticipated in the nervous system. Further detailed analysis of different neuronal populations may uncover interesting aspects of the function of NT-4/5 and its differential signaling or targets compared to BDNF. One suggestion is that NT-4/5 may have a potential role in injury responses or in the long-term maintenance of the nervous system (Conover et al., 1995). 3.2.2. Administration of NT-4/5 in vivo One of the most recent findings on the role of NT-4/5 in the developing nervous system is its involvement in the formation of segregated patches of axons from the lateral geniculate nucleus in the primary visual cortex (Cabelli et al., 1995). In this study, infusion of NT-4/5 into the cat visual cortex from P28 to P42 disrupted the formation of ocular dominance patches in layer 4 and induced a robust, uniform innervation. These findings suggest that endogenous NT-4/5 may mediate the activity-dependent control of axonal branching during development of the CNS. Furthermore, in neonates NT-4/5, like NT-3, increases the levels of neuropeptide Y, somatostatin, and substance P in different brain regions, including cortex, hippocampus and striatum (Carnahan and

237

N. Rocamora and E. Arenas TABLE 4 Alterations in NT3 knockout mice compared to NT4/5 knockout % Neuronal loss

NT-3 (%)

NT-4/5 (%)

PNS Sympathetic Superior cervical ganglion

50a,b

No^

Sensory neurons Dorsal root ganglion Trigeminal ganglion Trigeminal mesencephalic nu. Geniculate ganglion Vestibular ganglion Spiral ganglion Superior-Jugular ganglia Nodose-Petrosal ganglia

63^'^^ 55a,b* 25^ 34^ 85b,e

No^'^ No^

No^ 40a,b

CNS Motor neurons Spinal cord Trigeminal Facial

No^^b No^ No^'b

Other CNS neurons Substantia nigra Locus coeruleus

No^

No^

No^

NT3 Other alterations Cochlear afferents from Spiral G. Propioceptive DRG neurons Muscle spindles la innervation of spinal cord layer IX Golgi tendon organs and lb afferents Pacinian corpuscles Cutaneous SP and CGRP afferents Joint capsule and tendon afferents Spinal cord diameter

No^ No'^'^

NT4

Absent^ Absent^'^ Absent^'b Absent^'b'f Absent^ Present^ Present^ Present^ Reduced^

^Ernfors et al.,1994; barillas et al., 1994; '^Conover et a l , 1995; ^Liu et a l , 1995; ^Ernfors et a l , 1995; ^Tessarollo et a l , 1994. Percentages show statistically significant cell loss. *Neurons smaller in size.

Nawa, 1995), suggesting the existence of redundant regulatory mechanisms for these processes. NT-4/5 has also been shown to be active on motor neurons both during development and in the adult. Developing facial motor neurons can be rescued after axotomy by application of NT-4/5 in a gelfoam next to the proximal nerve stump (Huges et al., 1993; Koliatsos et al., 1994) while, in axo-

tomized adult spinal motorneurons, levels of both choline acetyltransferase and p75LN^P^ immunostaining are increased by NT-4/5 (Friedman et al., 1995). Finally, application of NT-4/5 to the gastrocnemius muscle of the adult rat, by grafting a cell line expressing excess NT-4/5, has been shown to induce both nodal and terminal sprouting of motor neurons (Funakoshi et al., 1995). This

238

Neurotrophin-3 and neurotrophin-4/5

TABLE 5 Neuronal populations responsive to NT4/5 administration in vivo

Development Motorneuron Facial Retinal ganglion Adult Basal forebrain Cholinergic neurons Substantia nigra Dopaminergic neurons Locus coeruleus Noradrenergic neurons

Development Visual cortex Anterior cortex, hippocampus and striatum Adult Motorneuron Spinal Basal forebrain Cholinergic neurons? Substantia nigra Dopaminergic neurons GABAergic neurons Raphe nucleus Serotonergic neurons Striatum Medium sized neurons Peptidergic neurons

Survival (% of lesioned)

Model

Reference

60 40 Yes

Axotomy, newborn Axotomy, newborn Newborn

Koliatsos et al., 1994 Hugesetal., 1993 Cui and Harvey, 1994

Yes

Axotomy

Alderson et al., 1996

Yes

Axotomy

Hagg, 1994

No

6-OHDA

Arenas and Persson, 1994

Function

Model

Reference

Oc.dominance columns

Postnatal Neonate

Cabelli et al., 1995 Carnahan and Nava, 1995

Sprouting: nodal,terminal (+)ChAT and p75

Axotomy

Funakoshi et al., 1995 Friedman et al., 1995

^•^•^Spatial memory

Aged

Fisher etal., 1994

(+)NPY, somatostatm

and substance P

^"•"^DA metabolism in striatum ^••"^Contraversive turning beh. ^•^^GAD activity in colliculi

Altar etal., 1994a Altar etal., 1994a Altar et al., 1994a

^"*"^5-HT metabolism in striatui

Altar etal., 1994a

Normalized GAD, PPTA (+)SP, NKA, DynA

6-OHDA

Saueretal, 1995 Arenas, Akervo, Wong, Boylan, Persson, Lindsay, and Altar, unpublished

^••"^Increase/induce, ^ decrease.

finding, together with regulation of the levels of NT-4/5 in the muscle in an activity-dependent fashion, suggests that muscle-derived NT-4/5 may act as an activity-dependent neurotrophic signal for growth and remodeling of motor innervation in the adult (Funakoshi et al., 1995). In the adult brain, NT-4/5 has shown interesting effects on monoaminergic systems of the basal

ganglia projecting to the striatum. Importantly, NT-4/5 has recently been shown to prevent the death of axotomized nigral dopaminergic neurons (Hagg, 1994). In addition, both nigral dopaminergic and raphe serotoninergic neurons respond to supranigral infusions of NT-4/5 by increasing their terminal metaboUsm (Altar et al., 1994a). These effects on monoaminergic systems are further ex-

N. Rocamora and E. Arenas

emplified by the demonstration of an NT-4/5induced counterclockwise circling, which is sensitive to haloperidol and SCH-23390 (Altar et al., 1994a). In addition to these effects on afferent neurons to the striatum, the most abundant neuronal type in the striatum, the medium-sized neuron, has also been shown to respond to NT-4/5. After 6-hydroxydopamine lesions of nigral dopaminergic neurons, instrastriatal injections of NT4/5 normalize the expression of GAD67 and PPTA mRNA by decreasing and increasing their levels, respectively (Sauer et al., 1995). Furthermore, we have observed recently that supranigral infusion of NT-4/5 in non-lesioned animals increases peptide levels of substance P, neurokinin A and dynorphin A in the substantia nigra, suggesting an effect of NT-4/5 on striatonigral peptidergic neurons (Arenas, Akervo, Wong, Boylan, Persson, Lindsay and Altar, unpublished observations). Thus, increasing the monoaminergic input and the peptidergic output of the striatonigral system with NT4/5 may be a productive approach to restoring balance in the basal ganglia, particularly after insults affecting the dopaminergic system. Finally, NT-4/5 has been shown to improve both acquisition and retention of spatial memory in aged rats, without affecting age-related cholinergic neuron atrophy in the basal forebrain (Fischer et al., 1994). These data suggest that the improvement of spatial memory could have been mediated both by cholinergic and/or non-cholinergic systems (Fischer et al., 1994). Interestingly, it has been found very recently that basal forebrain cholinergic neurons are rescued from axotomyinduced cell death by NT-4/5 (Alderson et al., 1996), suggesting an important role of NT-4/5 in this population of neurons and its therapeutic potential in neurodegenerative disorders such as Alzheimer's disease. 4. Discussion 4.1. Concluding remarks A growing body of evidence underlines the differences between the prototypic neurotrophin, NGF, and other members of the family. At the

239

evolutionary level, BDNF and NT-4 are the most closely related neurotrophins and they and NT-3 are all relatively distinct from NGF. With respect to their function in the developing PNS, there is a clear switch from NT-3, BDNF or NT-4/5 to NGF in those neuronal populations which depend on NGF for survival. In the CNS, the distributions of neurotrophins and their receptors, together with their distinct patterns of retrograde axonal transport, in the CNS, differentiate NGF from the other neurotrophins. NGF acts as the classical targetderived retrogradely transported neurotrophic factor, preventing programmed cell death of neurons after they innervate their targets. BDNF, NT-3 and NT-4/5, by contrast, act as differentiation factors at earlier developmental stages, before neurons reach their targets, and with a more important component of local autocrine and/or paracrine function. A general characteristic of neurotrophin function in the CNS is that BDNF, NT-3 and NT-4/5, but not NGF, act together in a wide range of neuronal populations. The high level of colocalization of trkB and trkC (Kokaia et al., 1995), supports responsiveness of these cells conjointly to BDNF, NT-3 and NT-4. Functional analyses have shown both redundant and pleiotropic functions for the different neurotrophins. However, the presence of specific trk receptors, although necessary, seems not to be sufficient to confer neurotrophin responsiveness. The regulation of the ratio between full (transduction active kinase receptors) and truncated receptors (without kinase but capable of binding) may be one of the elements in determining the responsiveness of neurons, and the specific activity of the different neurotrophins (Korsching, 1993). Knockout studies show a very different situation for PNS and CNS, with respect to their dependence on neurotrophins for survival. Although a clear decrease in population size of neurons projecting outside the CNS was found, there was no clear decrease in the number of cells projecting within the CNS. Because CNS cells normally coexpress trkB and C, one possibility is that they could be maintained by more than one neurotrophin. Also, double BDNF/NT-4/5 knockouts did

240

Neurotrophin-3 and neurotrophin-4/5

TABLE 6 Possible indications of neurotrophins 3 and 4/5 in therapy NTS

NT4/5

Peripheral sensory neuropathy

Other factors NGF, BDNF

Amyotrophic lateral sclerosis Motor neurons

BDNF,CNTF,IGF-1, GDNF CNTF

Spinal cord injury Parkinson's disease Sub. nigra dopaminergic neurons

BDNF, GDNF, TGFySl, IGVpX aFGF, bFGF, TGF6,EGF,IGF-1,

Loc. ceruleus noradrenergic neurons Raphe serotonergic neurons Huntington's disease Striatal GABA neurons LC noradrenergic neurons Alzheimer's disease Cortical and hippocampal neurons Basal forebrain cholinergic neurons LC noradrenergic neurons Raphe serotonergic neurons Ischemia

+ ++

++

+ ++

IGF-2, CNTF, IL-6 GDNF BDNF, FGF-5

+ +

NGF, BDNF, GDNF GDNF

+ ++

BDNF NGF, BDNF, bFGF, IL-3, CNTF GDNF BDNF, FGF-5 NGF, BDNF, bFGF, TGF^-1,IGF-1

++, Indicates effects observed on the adult neurons in vivo. +, Indicates effects observed on developing neurons in vivo or in vitro. Factors in bold are in clinical trials for the indicated pathology.

not show any detectable decrease of CNS neuronal populations. Thus, it remains to be demonstrated that CNS neurons depend on neurotrophins for survival during development. Double trkB/trkC or triple BDNF/NT-3/NT-4/5 knockouts may help to clarify whether redundancy in the effects of neurotrophins may account for this observation. However, since the redundancy found for neurotrophins, seems also to be the case for neurotrophic factors of other families, as indicated for the number of factors active on a given neuronal population (Table 6), a negative result may also be expected. In the future, the use of altemative experimental approaches such as transgenic animals carrying constructions with inducible promoters, or conditional knockouts, will surely provide very valuable insights into the role of these factors in vivo.

4.2. Therapeutic interest in NT-3 and NT-4/5 Clinical interest in the use of neurotrophic factors comes from the early finding that these factors are able to induce axon growth and survival of neurons, a population of cells considered otherwise unable to regenerate. Neurotrophic factors have so far been suggested to be possible therapeutic agents in many disorders of both peripheral and central nervous systems, including acute traumatic insults and chronic degenerative processes. 4.2.1. Neurotrophin-3 In the brain, several neuronal populations have been the subjects of particularly intense research, because of their involvement in neurodegenerative diseases. Amongst them, there are mesencephalic

N. Rocamora and E. Arenas

dopaminergic neurons, degenerating in Parkinson's disease (Hirsch et al., 1988; Jellinger, 1991); medium spiny projection striatal neurons, degenerating in Huntington's disease (Graveland et al., 1985; Reiner et al., 1988); noradrenergic neurons of the locus coeruleus, degenerating in Parkinson's, Huntington's and Alzheimer's diseases (Jellinger et al., 1991; German et al., 1992) and serotonergic neurons of the raphe nucleus, also degenerating in Alzheimer's disease and Parkinsonism (Halliday et al, 1990). Interestingly, NT3 has been reported to be active on hippocampal and cortical neurons in vitro (Collazo et al., 1992; Ip et al., 1993a; Vicario-Abejon et al, 1995; Cheng et al., 1994), to induce the phenotype of nigral dopaminergic and raphe serotonergic neurons in vivo (Altar et al., 1994b; Martin-Iverson et al., 1994) and the phenotype of basal forebrain cholinergic neurons in vivo (Fischer et al., 1994), and to promote the survival of locus coeruleus noradrenergic neurons in vivo (Arenas and Persson, 1994). Although NT-3 is active on most of the populations involved in these neurodegenerative disorders, the biological spectra of activities of NT-3 has not been reported to cover other important aspects of these diseases, namely the death of basal forebrain cholinergic neurons or the reinnervation of targets. Current knowledge of the function of neurotrophic factors in these systems suggests that the multiple aspects of the phenotype and survival of any one particular neuronal type are governed by many different neurotrophic factors. Therefore, despite the broad spectrum of activities shown by NT-3 on intrinsic CNS neurons affected by neurodegenerative diseases, it is likely that a therapeutic approach to these diseases will involve the administration of multiple factors or of chimeric molecules with multiple neurotrophic activities (Ibanez, 1995). Finally, two more possible therapeutic applications have been suggested for NT-3 outside the brain, on the basis of the survival- and/or phenotypic-promoting effects of NT-3 on large DRG neurons (Gao et al., 1995b; Zhang et al., 1994) and the induction of sprouting from corticospinal neurons (Schnell et al., 1994). These are: (i) acute spinal cord injury and (ii) peripheral sensory neuro-

241

pathies secondary to genetic alterations, diabetes, alcoholism, AIDS or cytostatic treatment. 4.2,2, Neurotrophin-4/5 Many diseases have been suggested to possibly benefit from an NT-4/5-based therapy. NT4 has been reported to be active on hippocampal neurons, cortical neurons, septal cholinergic neurons and locus coeruleus noradrenergic neurons, all of which are severely affected in Alzheimer's disease; and midbrain dopaminergic neurons, which together with locus coeruleus noradrenergic neurons degenerate in Parkinson's disease. Other populations of neurons suggested to potentially benefit from an NT-4/5 based therapy include motor neurons, affected in amyotrophic lateral sclerosis, peripheral sensory neurons, affected by neuropathies and finally, many other different populations of neurons affected by general insults like ischemia or hypoglycemia. It is, however, important to note that most of these pathological conditions occur in the adult, restricting the possible therapeutic applications to those neurons responsive to NT-4/5 in vivo in adulthood. Two recent reports have indicated that both adult basal forebrain cholinergic neurons and substantia nigra dopaminergic neurons are protected from axotomy-induced cell death by NT-4/5 (Hagg, 1994; Alderson et al., 1996), suggesting the possibility of therapeutic use in Parkinson's and Alzheimer's diseases of NT-4/5, along with many other factors (see Table 6). Interestingly, the survival- promoting effect of NT-4/5 on nigral dopaminergic neurons, together with the increases in diverse metabolic parameters of dopaminergic, serotonergic and peptidergic neurons known to be altered in Parkinson's disease, suggest a possible benefit from NT-4/5 based therapy in that disease. As well, the effects of NT-4/5 in increasing survival of cholinergic and serotonergic neurons (Alderson et al., 1996) and in increasing the metabolism of serotonergic cells (Altar et al., 1994a) could be of interest in Alzheimer's therapy. However, adult neurons of the locus coeruleus are not rescued from 6-hydroxydopamine lesions by NT4/5 (Arenas and Persson, 1994) and adult neurons of the hippocampus and cortex, although protected

242

by NT-4/5 in vitro, have not been reported to respond to NT-4/5 administration in vivo. A third possible therapeutic application of NT4/5 could be to increase and/or accelerate reinnervation of muscle endplates in partially denervated muscles, by increasing both nodal and terminal sprouting (Funakoshi et al., 1995). In this case, the induction of P75LNGFR and ChAT by NT-4/5 (Friedman et al., 1995) would also be a benefit of NT-4/5 based therapy. In conclusion, the growing information on the function of NT-3 and NT-4/5 in the nervous system makes currently possible new therapeutic approaches based on neurotrophic factors. However, their relatively low selectivity and the chronic nature of some of the disorders, will probably make it necessary to provide local and long-lasting administration. The design of new methods for the focal and controllable administration of neurotrophic factors or their genes may allow, in the near future, the development of effective neurotrophin-based therapy. Acknowledgements The authors thank Marta Pascual for preparing the expression table and Drs. Jordi Alberch, Patrik Ernfors, Carlos Ibanez and Miles Trupp for critical reading of the manuscript. N.R. was supported by the Spanish Government and E.A. by the Swedish Medical Research Council (MFR). References Alderson, R.F., Alterman, A.L., Barde, Y.A. and Lindsay, R.M. (1990) Brain-derived neurotrophic factor increases, survival and differentiated functions of rat, septal cholinergic neurons in culture. Neuron 5: 297-306. Alderson, R.F., Wiegand, S.J., Cai, N., Anderson, K.A., Lindsay, R.M. and Alter, C.A. (1996) Neurotrophin-4/5 maintains the cholinergic phenotype of axotomized septal neurons. Eur. J. Neurosci. (in press). Altar, C.A., Boylan, C.B., Fritsche, M., Jackson, C, Hyman, C. and Lindsay, R.M. (1994a) The neurotrophins NT-4/5 and BDNF augment serotonin, dopamine, and GABAergic systems during behaviorally effective infusions to the substantia nigra. Exp. Neurol. 130: 31-34. Altar, C.A., Boylan, C.B., Fritsche, M., Jackson, C , Lindsay, R.M. and Hyman, C. (1994b) Efficacy of BDNF and NT-3 on neurochemical and behavioral deficits associated with

Neurotrophin-3 and neurotrophin-4/5 partial nigrostriatal dopamine lesions. J. Neurochem. 63: 1021-1032. Altar, C.A., Siucak, J.A., Wright, P., Ip, N.Y., Lindsay, R.M. and Wiegand, S.J. (1994) In situ hybridization of trkB and trkC receptor mRNA in rat forebrain and association with high-affinity binding of [I^^IJBDNF, {^'^H'\KY-AI5

and

[125i]Nj_3 ^^j. J j^eurosci. 6: 1389-1405. Ardelt, A.A., Haris, N.A. and Roth, K.A. (1994) Neurotrophin-4 selectively promotes survival of striatal neurons in organotypic slice culture. Brain Res. 647: 340-344. Arenas, E. and Persson, H. (1994) Neurotrophin-3 prevents the death of adult central noradrenergic neurons in vivo. A^^mre 367: 368-371. Ballarm, M., Ernfors, P., Lindefors, N. and Persson, H. (1991) Hippocampal damage and kainic acid injection induce a rapid increase in mRNA for BDNF and NGF in the rat brain. Exp. Neurol. 114: 35-43. Baptista, C.A., Hatten, M.E., Blezeski, R. and Mason, C.A. (1994) Cell-cell interactions influence survival and differentiation of purified Purkinje cells 'in vitro'. Neuron 12: 243-260. Barbany, G. and Persson, H. (1992) Regulation of neurotrophin mRNA expression in the rat brain by glucocorticoids. Eur. J. Neurosci. 4: 396-403. Barbacid, M. (1994) The Trk family of neurotrophin receptors. J. Neurobiol. 25: 1386-1403. Barde, Y.-A (1989) Trophic factors and neuronal survival. Neuron 2:1525-1534. Barde, Y.-A. (1994) Neurotrophic factors: an evolutionary perspective. 7. Neurobiol. 25: 1329-1333. Barde, Y.-A., Edgar, D. and Thoenen, H. (1982) Purification of a new neurotrophic factor from mammalian brain. EMBO J. 1:549-553. Barres, B.A., Raff, M.C., Gaese, F., Bartke, I., Dechant, G. and Barde, Y.-A. (1994) A crucial role for neurotrophin-3 in oligodendrocyte development. Nature 367: 371-375. Barres, B.A., Schmid, R., Sendtner, M. and Raff, M.C. (1993) Multiple extracellular signals are required for longterm oligodendrocyte survival. Development 118: 283295. Bengzon, J., Kokaia, Z., Ernfors, P., Kokaia, M., Leanza, G., Nilsson, O.G., Persson, H. and Lindvall, O. (1993) Regulation of neurotrophins and trkA, trkB and trkC tyrosine kinase receptor messenger RNA expression in kindling. Neuroscience 53: 433-446. Berkemeier, L.R., Winslow, J.W., Kaplan, D.R., Nikolics, K., Goedded, D.V. and Rosenthal, A. (1991) Neurotrophin-5: a novel neurotrophic factor that activates trk and trkB. Neuron 7: 857-866. Berzaghi, M., Cooper, J., Castren, E., Zafra, F., Sofroniew, M., Thoenen, H. and Lindholm, D. (1993) Cholinergic regulation of brain-derived neurotrophic factor (BDNF) and nerve growth factor (NGF) but not neurotrophin-3 (NT-3) mRNA levels in the developing rat hippocampus. J. Neurosci. 13:3818-3826. Birren, S.J., Lo, L. and Anderson, D.J. (1993) Sympathetic

N. Rocamora and E. Arenas neurons undergo a developmental switch in trophic dependence. Development 119: 597-610. Buchman, V.L. and Davies, A.M. (1993) Different neurotrophins are expressed and act in a developmental sequence to promote the survival of embryonic sensory neurons. Development \ IS: 9S9-1001. Buj-Bello, A., Pinon, L.G.P. and Davies, A.M. (1994) The survival of NGF dependent cranial sensory neurons is promoted by several different neurotrophins early in their development. Development 120: 1573-1580. Cabelli, R.J., Hohn, A. and Shatz, C.J. (1995) Inhibition of ocular dominance column formation by infusion of NT-4/5 or BDNF. Science 267: 1662-1666. Carnahan, J. and Nawa, H. (1995) Regulation of neuropeptide expression in the brain by neurotrophins. Mol. Neurobiol. 10: 135-149. Castren, E., Pitkanen, M., Sirvio, J., Parsadanian, A., Lindholm, D., Thoenen, H. and Riekkinen, P. (1993) The induction of LTP increases BDNF and NGF mRNAs but decreases NT-3 mRNA in the dentate gyrus. NeuroReport 4: 895-898. Chalazonitis, A., Rothman, T.P., Chen, J., Lamballe, F., Barbacid, M. and Gershon, M.D. (1994) Neurotrophin-3 induces neural crest-derived cells from fetal rat gut to develop in vitro as neurons or glia. J. Neurosci. 14: 6 5 7 1 6584. Chao, M.V. (1992) Neurotrophin receptors: a window into neuronal differentiation. Neuron 9: 583-593. Chao, M.V., Bothwell, M.A., Ross, A.H., Koprowski, H., Lanahan, A.A., Buck, C.R. and Sehgal, A. (1986) Gene transfer and molecular cloning of the human NGF receptor. Science Til: Sn-Sll. Cheng, B., Goodman, Y., Begley, J.G. and Mattson, M.P. (1994) Neurotrophin-4/5 protects hippocampal and cortical neurons against energy deprivation and excitatory amino acid-induced injury. Brain Res, 650: 331-338. Cheng, B. and Mattson, M.P. (1994) NT-3 and BDNF protect CNS neurons against metabolic/excitotoxic insults. Brain Res. 640: 56-67. Clary, D.O. and Reichardt, L.F. (1994) An alternatively spliced form of the nerve growth factor receptor TrkA confers an enhanced response to neurotrophin 3. Proc. Natl. Acad. Sci. USA 91: 11133-11137. Cohen, A., Bray, G.M. and Aguayo, A.J. (1994) Neurotrophin-4/5 increases adult rat retinal ganglion cell survival and neurite outgrowth 'in vitro'. J. Neurobiol. 25: 953-959. Collazo, D., Takahashi, H. and McKay, R. (1992) Cellular targets and trophic functions of neurotrophin-3 in developing rat hippocampus. Neuron 9: 643-656. Conover, J . C , Erickson, J.T., Katz, D.M., Bianchi, L.M., Poueymirou, W.T., McClain, J., Pan, L., Helgren, M., Ip, N.Y., Boland, P., Friedman, B., Wiegand, S., Vejsada, R., Kato, A.C., CeChiara, T.M. and Yancopoulos, G.D. (1995) Neuronal deficits, not involving motor neurons, in mice lacking BDNF and/or NT4. Nature 375: 235-238. Copray, J.C.V.M. and Brouwer, N. (1994) Selective expres-

243 sion of neurotrophin-3 messenger RNA in muscle spindles of the rat. Neuroscience 63: 1125-1135. Cordon-Cardo, C , Tapley, P., Jing, S., Nanduri, V., O'Rourke, E., Lamballe, F., Kovary, K., Klein, R., Jones, K.R., Reichardt, L.F. and Barbacid, M. (1991) The trk tyrosine protein kinase mediates the mitogenic properties of nerve growth factor and neurotrophin 3. Cell 66: 173-184. Cui, Q. and Harvey, A.R. (1994) NT-4/5 reduces naturally occurring retinal ganglion cell death in neonatal rats. NeuroReport 5: 1882-1884. Curtis, R., Lee, K.-F., Jaenisch, R., Huber, J., Chao, M.V., Lindsay, R.M. and DiStefano, P.S. (1993). Potential involvement of p75LNGFR in retrograde transport of neurotrophin-4. Soc. Neurosci. Abstr. 19: 1476. Davies, A.M. (1994) Switching neurotrophin dependence. Current Biology 4: 213-216. Davies, A.M., Horton, A., Burton, L.E., Schmeizer, C , Vandlen, R. and Rosenthal, A. (1993) Neurotrophin-4/5 is a mammalian-specific survival factor for distinct populations of sensory neurons. /. Neurosci. 13: 4 9 6 1 ^ 9 6 7 . Davies, A.M. and Lumsden, A.G.S. (1984) Relation of target encounter and neuronal death to nerve growth factor responsiveness in the developing mouse trigeminal ganglion. /. Comp. Neurol. 223: 124-137. DiCicco-Bloom, E., Friedman, W. and Black, LB. (1993) NT3 stimulates sympathetic neuroblast proliferation by promoting precursor survival. Neuron 11:1101-1 111. DiStefano, P., Friedman, B., Radziejewski, C , Alexander, C , Boland, P., Schick, C M . , Lindsay, R.M. and Wiegand, S.J. (1992) The neurotrophins BDNF, NT-3 and NGF display distinct patterns of retrograde axonal transport in peripheral and central neurons. Neuron 8: 983-993. Dugich-Djordjevic, M.M., Tocco, G., Lapchak, P.A., Pasinetti, G., Najm, L, Baudry, M. and Hefti, F. (1992) Regionally specific and rapid increases in brain-derived neurotrophic factor messenger RNA in the adult rat brain following seizures induced by systemic administration of kainic acid. Neuroscience 47: 303-315. Ebendal, T. (1992) Function and evolution in the NGF family and its receptors. J. Neurosci. Res. 32: 461-470. Elkabes, S., Dreyfus, C.F., Schaar, D.G. and Black, LB. (1994) Embryonic sensory development: local expression of NT-3 and target expression of NGF. /. Comp. Neurol. 341:204-213. Ernfors, P. and Persson, H. (1991) Developmentally regulated expression of HDNF/NT-3 messenger RNA in rat spinal cord motoneurons and expression of BDNF messenger RNA in embryonic dorsal root ganglion. Eur. J. Neurosci. 3:953-961. Ernfors, P., Ibanez, C. F., Ebendal, T., Olson, L. and Persson, H. (1990a) Molecular cloning and neurotrophic activities of a protein with structural similarities to nerve growth factor: developmental and topographical expression in the brain. Proc. Natl. Acad. Sci. USA 87: 5454-5458. Ernfors, P., Wetmore, C , Olson, L. and Persson, H. (1990b) Identification of cells in a rat brain and peripheral tissues

244

expressing mRNA for members of the nerve growth factor family. Neuron 5: 511-526. Ernfors, P., Benzgon, J., Kokaia, Z., Persson, H. and Lindvall, O. (1992a) Increased levels of messenger mRNAs for neurotrophic factors in the brain during kindling epileptogenesis. Neuron 7: 165-176. Ernfors, P., Merlio, J.-P. and Persson, H. (1992b) Cells expressing mRNA for neurotrophins and their receptors during embryonic rat development. Eur. J. Neurosci. 4: 11401158. Ernfors, P., Lee, K.-R, Kucera, J. and Jaenisch, R. (1994) Lack of neurotrophin-3 leads to deficiencies in the peripheral nervous system and loss of limb propioceptive afferents.CW/77: 503-512. Ernfors, P., Van De Waler, T., Loring, J. and Jaenisch, R. (1995) Complementary roles of BDNF and NT-3 in vestibular and auditory development. Neuron 14: 1-20. Escandon, E., Soppet, D., Roshental, A., Mendoza-Ramirez, J.-L. Szonyi, E., Burton, L.E., Henderson, C.E., Parada, L.F. and Nikolics, K. (1994) Regulation of neurotrophin receptor expression during embryonic and postnatal development. J. Neurosci. 14: 2054-2068. Farinas, L, Jones, K.R., Backus, C, Wang, X.-Y. and Reichardt, L.F. (1994) Severe sensory and sympathetic deficits in mice lacking neurotrophin-3. Nature 369: 658-661. Fischer, W., Sirevaag, A., Wiegand, S.J., Lindsay, R.M. and Bjorklund, A. (1994) Reversal of spatial memory impairments in aged rats by nerve growth factor and neurotrophins 3 and 4/5 but not by brain-derived neurotrophic factor. Proc. Natl. Acad. Sci. USA9\: 8607-8611. Friedman, W.J., Ernfors, P. and Persson, H. (1991) Transient and persistent expression of NT-3/HDNF messenger RNA in the rat brain during postnatal development. J. Neurosci. 11: 1577-1584. Friedman, W.J., Ibafiez, C.F., Halbook, F., Persson, H., Cain, L.D., Dreyfus, C.F. and Black, LB. (1993). Differential actions of neurotrophins in the Locus coeruleus and basal forebrain. Exp. Neurol 119: 72-78. Friedman, B., Kleinfeld, D., Ip, N.Y., Verge, V.M., Moulton, R., Boland, P., Zlotchenko, E., Lindsay, R.M. and Liu, L. (1995) BDNF and NT-4/5 exert neurotrophic influences on injured adult spinal motor neurons. J. Neurosci. 15: 10441056. Funakoshi, H, Belluardo, N., Arenas, E., Yamamoto, Y., Casabona, A., Persson, H. and Ibafiez, C.F. (1995) Musclederived neurotrophin-4 as an activity-dependent trophic signal for adult motor neurons. Science 268: 14951499. Funakoshi, H, Frisen, J., Barbany, G., Timmusk, T., Zachrisson, O., Verge, V.M.K. and Persson, H. (1993) Differential expression of mRNAs for neurotrophins and their receptors after axotomy of the sciatic nerve. /. Cell Biol. 123: 455465. Gaese, F., Kolbeck, R. and Barde, Y.-A. (1994) Sensory ganglia require neurotrophin-3 early in development. Development 120: 1613-1619.

Neurotrophin-3 and neurotrophin-4/5 Gall, CM., Murray, K. and Isackson, P.J. (1991) Kainic acidinduced seizures stimulate increased expresion of nerve growth factor mRNA in rat hippocampus. Mol Brain Res. 9: 113-123. Gall, CM., Gold, S.J., Isackson, P.J. and Seroogy, K.B. (1992) Brain-derived neurotrophic factor and neurotrophin 3 mRNAs are expressed in ventral midbrain regions containing dopaminergic neurons. Mol. Cell. Neurosci. 3: 5663. Gall, CM. and Isackson, P.M. (1989) Limbic seizures increase neuronal production of messenger RNA for nerve growth factor. Science 245: 758-761. Gao, W.-Q., Zheng, J.L. and Karihaloo, M. (1995a) Neurotrophin 4/5 and brain-derived neurotrophic factor act at later stages of cerebellar granule cell differentiation. J. Neurosci. 15: 2656-2667. Gao, W.-Q., Dybdal, N., Shinsky, N., Murnane, A., Schmelzer, C, Siegel, M., Keller, G., Hefti, F., Phillips, H.S. and Winslow, J.W. (1995b) Neurotrophin-3 reverses experimental cisplatin-induced peripheral sensory neuropathy. Ann. Neurol. 38: 30-37. German, D.C, Manaye, K.F., White, CL., Woodward, D.J., Mclntire, D.D., Smith, W.K., Kalaria, R.N. and Mann, D.M.A. (1992) Disease-specific patterns of locus coeruleus cell loss. Ann. Neurol. 32: 667-676. Glass, D.J., Nye, S.H., Hantzopoulos, P., Macchi, M.J., Squinto, S.P., Goldfarb, M. and Yancopoulos, G.D. (1991) TrkB mediates BDNF/NT-3-dependent survival and proliferation in fibroblasts lacking the low affinity NGF receptor. CW/66: 405-413. Gosh, A. and Greenberg, M.E. (1995) Distinct roles for bFGF and NT-3 in the regulation of cortical neurogenesis. Neuron 15: 89-103. Gotz, R., Raulf, F. and Schartl, M. (1992) Brain-derived neurotrophic factor is more highly conserved in structure and function than nerve growth factor during vertebrate evolution. J. Neurochem. 59: 432-442. Gotz, R., Koster, R., Winkler, Ch., Raulf, F., Lottspeich, F., Schartl, M. and Thoenen, H. (1994) Neurotrophin-6 is a new member of the nerve growth factor family. Nature 372: 266-269. Hagg, T. (1994) Neurotrophins (BDNF, NT-3, NT-4) prevent death of axotomized adult rat substantia nigra neurons. Soc. Neurosci. Abstr. 20: 441. Hallbook, F., Ibafiez, C.F. and Persson, H. (1991) Evolutionary studies of the nerve growth factor family reveal a novel member abundantly expressed in Xenopus ovary. Neuron 6: 845-858. Halliday, G.M., Li, Y.W., Blumbergs, P.C, Joh, T.H., Cotton, R.G.H., Howe, P.R.C, Blessing, W.W. and Geffen, L.B. (1990) Neuropathology of immunohistochemically identified brainstem neurons in Parkinson's disease. Ann. Neurol. 21: 373-385. Hantzopoulos, P.A., Suri, C, Glass, D.J., Goldfarb, M.P. and Yancopoulos, G.D. (1994) The low-affinity NGF receptor, p75, can collaborate with each of the trks to potentiate

N. Rocamora and E. Arenas functional responses to the neurotrophins. Neuron 13: 187201. Hartikka, J. and Hefti, F. (1988) Development of septal cholinergic neurons in culture: plating density and glial cells modulate effects of NGF on survival and expression of transmitter-specific enzymes. J. Neurosci. 8: 2967-2985. Hatanaka, H., Tsukui, H. and Nihonmatsu, I. (1988) Developmental change in the nerve growth factor action from induction of cholinergic acetyltransferase to promotion of cell survival in cultured basal forebrain neurons from postnatal rat. Dev. Brain Res. 39: 88-95. Hefti, F. (1986) Nerve growth factor promotes survival of septal cholinergic neurons after fimbrial transections. /. Neurosci. 6: 2155-2162. Hempstead, B.L., Martin-Zanca, D., Kaplan, D.R., Parada, L.F. and Chao, M.V. (1991) High-affinity NGF binding requires coexpression of the trk proto-oncogene and the lowaffinity NGF rceptor. Nature 350: 678-683. Henderson, C.E., Camu, W., Mettling, C , Gouin, A., Poulsen, K., Karihaloo, M., Rullama, J., Evans, T., McMahon, S.B., Armanini, M.P., Berkemeier, L., Phillips, H.S. and Rosenthal, A. (1993) Neurotrophins promote motor neuron survival and are present in embryonic limb. Nature 363: 266. Hirsch, E., Graybiel, A.M. and Agid, Y.A. (1988) Melanized dopaminergic neurons are differentially susceptible to degeneration in Parkinson's disease. Nature 334: 345-348. Hohn, A., Leibrock, J., Bailey, K. and Barde, Y.-A. (1990) Identification and characterization of a novel member of the nerve growth factor/brain-derived neurotrophic factor family. Namre 344: 339-341. Holland, P.W.H., Garcia-Fernandez, J., WilUams, N.A. and Sidow, A. (1994) Gene duplications at the origins of vertebrate development. Development SuppL: 125-133. Hory-Lee, F., Russell, M., Lindsay, R.M. and Frank, E. (1993) Neurotrophin-3 supports the survival of developing muscle sensory neurons in culture. Proc. Natl. Acad. Sci. USA 90: 2613-2617. Huges, R.A., Sendtner, M. and Thoenen, H. (1993) Members of several gene families influence survival of rat motoneurons in vitro and in vivo. J. Neurosci. Res. 36: 6 6 3 671. Humpel, C., Wetmore, C. and Olson, L. (1993) Regulation of brain-derived neurotrophic factor messenger RNA and protein at the cellular level in phenylene-tetrazol-induced epileptic seizures. Neuroscience 53: 909-918. Hyman, C., Hofer, M., Barde, Y.-A., Juhasz, M., Yancopoulos, G.D., Squinto, S.P. and Lindsay, R.M. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350: 230-232. Hyman, C , Juhasz, M., Jackson, C , Wright, P., Ip, N.Y. and Lindsay, R.M. (1994) Overlapping and distinct actions of the neurotrophins BDNF, NT-3 and NT-4/5 on cultured dopaminergic and GABAergic neurons of the ventral mesencephalon. J. Neurosci. 14: 335-347. Hynes, M.A., Poulsen, K., Armanini, M., Berkemeier, L., Phillips, H. and Rosenthal, A. (1994) Neurotrophin-4/5 is a

245 survival factor for embryonic midbrain dopaminergic neurons in enriched cultures. J. Neurosci. Res. 37: 144-154. Ibafiez, C.F. (1995) Neurotrophic factors: from structurefunction studies to designing effective therapeutics. Trends Biotechnol. 13: 217-227. Ibafiez, C.F., Ebendal, T. and Persson, H. (1991) Chimeric molecules with multiple neurotrophic activities reveal structural elements determining the specificities of NGF and BDNF. EMBO J. 10: 2105-2110. Ibafiez, C.F., Ebendal, T., Barbany, G., Murray-Rust, J., Blundell, T.L. and Persson, H. (1992) Disruption of the low-affinity receptor-binding site in nerve growth factor allows neuronal survival and differentiation by binding to the trk gene product. Cell 69: 329-341. Ibafiez, C.F., Ernfors, P., Timmusk, T., Ip, N.Y., Arenas, E., Yancopoulos, G.D. and Persson, H. (1993a) Neurotrophin-4 is a target-derived neurotrophic factor for neurons of the trigeminal ganglion. Development 117: 1345-1353. Ibafiez, C.F., Hag, L.L., Murray-Rust, J. and Persson, H. (1993b) An extended surface of binding to Trk tyrosine kinase receptors in NGF and BDNF allows the engineering of a multifunctional pan-neurotrophin. EMBO J. 12: 2 2 8 1 2293. Hag, L.L., Lonnerberg, P., Persson, H. and Ibafiez, C.F. (1994) Role of variable 6-hairpin loop in determining biological specificities in neurotrophin family. J. Biol. Chem. 269: 19941-19946. Ip, N.Y., Ibafiez, C.F., Nye, S.H., McClain, J., Jones, P.F., Gies, D.R., Belluscio, L., Le Beau, M.M., Espinosa, III, E., Squinto, S.P., Persson, H. and Yancopoulos, G.D. (1992) Mammalian neurotrophin-4: structure, chromosomal localization, tissue distribution and receptor specificity. Proc. Natl. Acad. Sci. USA 89: 3060-3064. Ip, N.Y., Li, Y., Yancopoulos, G.D. and Lindsay, R.M. (1993a) Cultured hippocampal neurons show responses to BDNF, NT-3 and NT-4, but not NGF. J. Neurosci. 13: 3394-3405. Ip, N.Y., Stitt, T.N., Tapley, P., Klein, R., Glass, D.J., Fandl, J., Greene, L.A., Barbacid, M. and Yancopoulos, G.D. (1993b) Similarities and differences in the way neurotrophins interact with the Trk receptors in neuronal and nonneuronal cells. Neuron 10: 137-149. Isackson, P.J., Huntsman, M.M., Murray, K.D. and Gall, C M . (1991) BDNF mRNA expression is increased in adult rat forebrain after limbic seizures: temporal patterns of induction distinct from NGF. Neuron 6: 937-948. Jellinger, K.A. (1991) Pathology of Parkinson's disease: changes other than the nigrostriatal pathway. Mol. Chem. Neuropathol. 14: 153-197. Jones, K.J. and Reichardt, L.F. (1990) Molecular cloning of a human gene that is member of the nerve growth factor family. Proc. Natl. Acad. Sci. USA 87: 8060-8064. Jonhson, E.J., Andres, R.Y. and Bradshaw, R.A. (1978) Characterization of the retrograde transport of nerve growth factor (NGF) using high specific activity: 1251: NGF. Brain Res. 150:319-331.

246 Kaisho, Y., Yoshimura, K. and Nakahama, K. (1990) Cloning and expression of a cDNA encoding a novel human neurotrophic factor. FEES Lett. 266: 187-191. Kalcheim, C, Carmeli, C. and Rosenthal, A. (1992) Neurotrophin-3 is a mitogen for cultured neural crest cells. Proc. Natl Acad. ScL USA 89: 1661-1665. Kang, H. and Schuman, E.M. (1995) Long-lasting neurotrophin-induced enhancement of synaptic transmission in the adult hippocampus. Science 267: 1658-1662. Kaplan, D.R., Hempstead, B., Martin-Zanca, D., Chao, M. and Parada, L.F. (1991a) The trk proto-oncogene product: a signal transducing receptor for nerve growth factor. Science 252: 554-558. Kato, A.C. and Lindsay, R.M. (1994) Overlapping and additive effects of neurotrophins and CNTF on cultured human spinal cord neurons. Exp. Neurol. 130: 196-201. Klein, R., Jing, S. Nanduri, V., O'Rourke, E. and Barbacid, M. (1991a) The trk proto-oncogene encodes a receptor for nerve growth factor. Cell 65: 189-197. Klein, R., Nanduri, V., Jing, S., Lamballe, F., Tapley, P., Bryant, S., Cordon-Cardo, C, Jones, K.R., Reichardt, L.F. and Barbacid, M. (1991b) The trkB tyrosine protein kinase is a receptor for brain-derived neurotrophic factor and neurotrophin 3. Cell 66: 395-404. Klein, R., Smeyne, R.J., Wurst, W., Long, L.K., Auerbach, B.A., Joyner, A.L. and Barbacid, M. (1993) Targeted disruption of the trkB neurotrophin receptor gene results in nervous system lesions and neonatal death. Cell 75: 113122. Knusel, B., Winslow, J. W., Rosenthal, A., Burton, L.E., Seid, D.P., Nikolics, K. and Hefti, F. (1991) Promotion of central cholinergic and dopaminergic neuron differentiation by brain-derived neurotrophic factor but not neurotrophin-3. Proc. Natl. Acad. ScL USA 88: 961-965. Knusel, B., Rabin, S.J., Hefti, F. and Kaplan, D.R. (1994) Regulated neurotrophin receptor responsiveness during neuronal migration and early differentiation. J. Neurosci. 14: 1542-1554. Koh, S., Oyler, G.A. and Higgins, G.A. (1989) Localization of nerve growth factor receptor messenger RNA and protein in the adult rat brain. Exp. Neurol. 106: 209-221. Kokaia, Z., Bengzon, J., Metsis, M., Kokaia, M. and Persson, H. (1993) Coexpression of neurotrophins and their receptors in neurons of the central nervous system. Proc. Natl. Acad. Sci. USA 99: 6711-6715. Kokaia, Z., Metsis, M., Kokaia, M., Bengzon, J., Elmer, E., Smith, M.-L., Timmusk, T., Siesjo, B.K., Persson, H. and Lindvall, O. (1994) Brain insults in rats induce increased expression of the BDNF gene through differential use of multiple promoters. Eur. J. Neurosci. 6: 587-596. Kokaia, Z., Metsis, M., Kokaia, M., Elmer, E. and Lindvall, O. (1995) Co-expression of TrkB and TrkC receptors in CNS neurones suggests regulation by multiple neurotrophins. NeuroReport 6: 769-772. Koliatsos, V.E., Clatterbuck, E., Winslow, J.W., Cayouette, M.H. and Price, D.L. (1993) Evidence that brain-derived

Neurotrophin-3 and neurotrophin-4/5 neurotrophic factor is a trophic factor for motor neurons in vivo. Neuron 10: 359-367. Koliatsos, V.E., Cayouette, M.H., Berkemeier, L.R., Clatterbuck, R.E., Price, D.L. and Rosenthal, A. (1994) Neurotrophin 4/5 is a trophic factor for mammalian facial motor neurons. Proc. Natl. Acad. Sci. USA 91: 3304-3308. Korsching, S. (1993) The neurotrophic factor concept: a reexamination. J. Neurosci. 13: 2739-2748. Lamballe, F., Klein, R. and Barbacid, M. (1991) TrkC, a new member of the trk family of tyrosine protein kinases, is a receptor for neurotrophin 3. Cell 66: 967-980. Lamballe, F., Tapley, P. and Barbacid, M. (1993) trkC encodes multiple neurotrophin-3 receptors with distinct biological properties and substrate specificities. EMBO J. 12: 3083-3094. Lamballe, F., Smeyne, R.J. and Barbacid, M. (1994) Developmental expression of trkC, the neurotrophin-3 receptor, in the mammalian nervous system. /. Neurosci. 14: 14-28. Lauterborn, J.C, Isackson, P.J. and Gall, CM. (1994) Cellular localization of NGF and NT-3 mRNAs in postnatal rat forebrain. Mol. Cell. Neurosci. 5: 46-62. LeBmann, V., Gottman, K. and Heumann, R. (1994) BDNF and NT-4/5 enhance glutamatergic synaptic transmission in cultured hippocampal neurones. NeuroReport 6: 21-25. Leibrock, J., Lottspeich, F., Hohn, A., Hofer, M., Hengerer, B., Masiakowski, P., Thoenen, H. and Barde, Y.-A. (1989) Molecular cloning and expression of brain-derived neurotrophic factor. Nature 341: 149-152. Leingartner, A. and Lindholm, D. (1994) Two promoters direct transcription of the mouse NT-3 gene. Eur. J. Neurosci. 6: 1149-1159. Leingartner, A., Heisenberg, C.-P., Kolbeck, R., Thoenen, H. and Lindholm, D. (1994) Brain-derived neurotrophic factor increases neurotrophin-3 expression in cerebellar granule neurons. J. Biol. Chem. 269: 828-830. Levi-Montalcini, R. (1987) The nerve growth factor: thirtyfive years later. EMBO J. 6: 1145-1154. Lindholm, D., Castren, E., Tsoulfas, P., Kolbeck, R., Berzaghi, M.P., Leingartner, A., Heisenberg, C.-P., Tesarollo, L., Parada, L.F. and Thoenen, H. (1993a) Neurotrophin-3 induced by tri-iodothyronine in cerebellar granule cells promotes purkinje cell differentiation. /. Cell Biol. 122: 443-^50. Lindholm, D., Dechant, G., Heisenberg, C.-P. and Thoenen, H. (1993b) Brain-derived neurotrophic factor is a survival factor for cultured rat cerebellar granule neurons and protects them against glutamate-induced neurotoxicity. Eur. J. Neurosci. 5: 1455-1464. Lindholm, D., Berzaghi, M.-P., Cooper, J., Thoenen, H. and Castren, E. (1994a) Brain-derived neurotrophic factor and neurotrophin-4 increase neurotrophin-3 expression in the rat hippocampus. Int. J. Dev. Neurosci. 12: 745-751. Lindholm, D., Hartikka, J., Berzaghi, M.P., Castren, E., Tzimagiorgis, G., Hughes, R.A. and Thoenen, H. (1994b) Fibroblast growth factor-5 promotes differentiation of cultured rat septal cholinergic and raphe serotonergic neurons:

N. Rocamora and E. Arenas comparison with the effects of neurotrophins. Eur. J. Neurosci. 6: 244-252. Lindsay, R.M., Thoenen, H. and Barde, Y.-A. (1985) Placode and neural crest-derived sensory neurons are responsive at early developmental stages to brain-derived neurotrophic factor. Dev. Biol 112: 319-328. Lindvall, O., Ernfors, P., Bengzon, J., Kokaia, Z., Smith, M.L., Siesjo, B.K. and Persson, H. (1992) Differential regulation of mRNAs for nerve growth factor, brain-derived neurotrophic factor and neurotrophin-3 in the adult rat brain following cerebral ischemia and hypoglycemic coma. Proc. Natl Acad. ScL USA 89: 648-652. Liu, X., Ernfors, P., Wu, H. and Jaenisch, R. (1995) Sensory but not motor neuron deficits in mice lacking NT4 and BDNF. Nature 375: 238-241. Lohof, A.M., Ip, N.Y. and Poo, M. (1993) Potentiation of developing neuromuscular synapses by the neurotrophins NT-3 and BDNF. Nature 363: 350-353. Lu, B., Yokoyama, M., Dreyfus, C.F. and Black, LB. (1991) Depolarizing stimuli regulate nerve growth factor gene expression in cultured hippocampal neurons. Proc. Natl Acad. Scl USA 88: 6289-6292. Maisonpierre, P.C., Belluscio, L., Squinto, S., Ip, N.Y., Furth, M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990a) Neurotrophin-3: a neurotrophic factor related to NGF and BDNF. Science 247: 1446-1451. Maisonpierre, P.C, Belluscio, L., Friedman, B., Alderson, R.F., Wiegand, S.J., Furth, M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990b) NT-3, BDNF and NGF in the development of the rat nervous system: parallel as well as reciprocal pattern of expression. Neuron 5: 501-509. Martin-Iverson, M.T., Todd, K.G. and Altar, C.A. (1994) Behavioral and in vivo neurochemical effects of the neurotrophic factors BDNF and NT-3 and their interactions with amphetamine. J. Neuroscl 14: 1262-1270. Martin-Zanca, D., Oskam, R., Mitra, G., Copeland, T. and Barbacid, M. (1989) Molecular and biochemical characterization of the human trk proto-oncogene. Mol Cell Biol 9: 24-33. Martinez, H.J., Dreyfus, C.F., Jonakait, G.M. and Black, I. (1985) Nerve growth factor promotes cholinergic development in brain striatal cultures. Proc. Natl Acad. Scl USA 82:7777-7781. Merlio, J.-P., Ernfors, P., Jaber, M. and Persson, H. (1992) Molecular cloning of rat trkC and distribution of cells expressing messenger RNAs for members of the trk family in the rat central nervous system. Neuroscience 51: 513-532. Merlio, J.-P, Ernfors, P., Kokaia, Z., Middlemas, D.S., Bengzon, J., Kokaia, M., Smith, M.-L., Siesjo, B.K., Hunter, T., Lindvall, O. and Persson, H. (1993) Increased production of the trkB protein tyrosine kinase receptor after brain insults. Neuron 10:151-164. Metsis, M., Timmusk, T., Arenas, E. and Persson, H. (1993) Differential usage of multiple brain-derived neurotrophic factor promoters in the rat brain following neuronal activation. Proc. Natl Acad. Scl USA 90: 8802-8806.

247

Middlemas, D.S., Lindberg, R.A. and Hunter, T. (1991) trkB a neural receptor protein-tyrosine kinase: evidence for a full length and two truncated receptors. Mol Cell Biol. 11: 143-153. Miranda, R.C., Sohrabji, F. and Toran-Allerand, D. (1993) Neuronal localization of mRNAs for neurotrophins and their receptors in the developing central nervous system suggests a potential for autocrine interactions. Proc. Natl Acad. Scl USA 90: 6439-6443. Morfini, G., DiTella, M.C., Feiguin, F., Carri, N. and Caceres, A. (1994) Neurotrophin-3 enhances neurite outgrowth in cultured hippocampal pyramidal neurons. J. Neuroscl Res. 39: 219-232. Mount, H.T.J., Dreyfus, C.F. and Black, LB. (1994) Neurotrophin-3 selectively increases cultured Purkinje cell survival. NeuroReport 5: 2497-2500. Mu, X., Silos-Santiago, I., Carroll. S.L. and Snider, W.D. (1993) Neurotrophin receptor genes are expressed in distinct patterns in developing dorsal root ganglia. J. Neuroscl 13:4029-4041. Mudo, G., Persson, H., Timmusk, T., Funakoshi, H., Bindoni, M. and Belluardo, N. (1993) Increased expression of trkB and trkC mRNAs in the rat forebrain after focal mechanical injury. Neuroscience 57: 901-912. Neveu, I., Naveilhan, P., Baudet, C, Brachet, P. and Metsis, M. (1994) 1,25-dihydroxyvitamin D3 regulates NT-3, NT-4 but not BDNF mRNA in astrocytes. NeuroReport 6: 124126. Okazawa, H., Nishiyama, K., Kamei, M., Washizaki, K., Murayama, S., Kwak, S. and Kanazawa, I. (1993) Trk immunoreactivity at neuronal dendrite and cell body. Biochem. Biophys. Res. Commun. 194: 683-690. Patterson, S.L., Grover, L.M., Schwartzkroin, P.A. and Bothwell, M. (1992) Neurotrophin expression in rat hippocampal slices: a stimulus paradigm inducing LTP in CAl evokes increases in BDNF and NT-3 mRNAs. Neuron 9: 1081-1088. Raddeke, M.J., Misko, T.P., Hsu, C, Herzenberg, L.A. and Shooter, E.M. (1987) Genetransfer and molecular cloning of the rat nerve growth factor. Nature 325: 593-597. Rocamora, N., Palacios, J.M. and Mengod, G. (1992) Limbic seizures induce a differential regulation of the expression of nerve growth factor, brain-derived neurotrophic factor and neurotrophin-3, in the rat hippocampus. Mol Brain Res. 13: 27-33. Rocamora, N., Garcia-Ladona, F.J., Palacios, J.M. and Mengod, G. (1993a) Differential expression of BDNF, NT3 and LNGRF during the postnatal development of the rat cerebellar system. Mol Brain Res. 17: 1-8. Rocamora, N., Massieu, L., Boddeke, H.W.G.M., Mengod, G. and Palacios, J.M. (1993b) Neuronal death and neurotrophin gene expression: long-lasting stimulation of neurotrophin-3 mRNA in the degenerating CAl and CA4 pyramidal cell layers. Neuroscience 53: 905-908. Rocamora, N., Massieu, L., Boddeke, H.W.G.M., Palacios, J.M. and Mengod, G. (1994) Differential regulation of the

248

expression of nerve growth factor, brain-derived neurotrophic factor and neurotrophin-3 mRNAs in adult rat brain after intrahippocampal injection of quinolinic acid. Mol Brain Res. 26: 89-98. Rocamora, N., Pascual, M., Acsady, L., de Lecea, L., Freund, T.F. and Soriano, E. (1996) Expression of NGF and NT3 mRNAs in hippocampal interneurons innervated by the GABAergic septohippocampal pathway. J. Neurosci. 16: 3991^004. Rodriguez-Tebar, A., Dechant, G. and Barde, Y.-A. (1990) Binding of the brain-derived neurotrophic factor to the nerve growth factor receptor. Neuron 4: 487-492. Rodriguez-Tebar, A., Dechant, G., Gotz, R. and Barde, Y.-A. (1992) Binding of neurotrophin-3 to its neuronal receptors and interactions with nerve growth factor and brain-derived neurotrophic factor. EMBO J. 11: 917-922. Rosenthal, A., Goedded, D.V., Nguyen, T., Lewis, M., Shih, A., Laramee, G.R., Niokolics, K. and Winslow, J.W. (1990) Primary structure and biological activity of a novel human neurotrophic factor. Neuron 4:161-113. Ryden, M., Murray-Rust, J., Glass, D., L.llag, L., Trupp, M., Yancopoulos, G.D., McDonald, N.Q. and Ibanez, C.F. (1995) Functional analysis of mutant neurotrophins deficient in low-affinity binding reveals a role for p75^^^^^ in NT-4 signaling. EM^O 7. 14: 1979-1990. Sauer, H., Wong, V. and Bjorklund, A. (1995) Brain-derived neurotrophic factor and neurotrophin-4/5 modify neurotransmitter-related gene expression in the 6hydroxydopamine-lesioned rat striatum. Neuroscience 65: 927-933. Schecterson, L. and Bothwell, M. (1992) Novel roles for neurotrophins are suggested by BDNF and NT-3 mRNA expression in developing neurons. Neuron 9: 449-463. Schnell, L., Schneider, R., Kolbeck, R., Barde, Y.-A. and Schwab, M.E. (1994) Neurotrophin-3 enhances sprouting of corticospinal tract during development and after adult spinal cord lesion. Nature 367: 170-173. Segal, R.A., Omeroy, S.L. and Stiles, CD. (1995) Axonal growth and fasciculation linked to differential expression of BDNF and NT-3 receptors in the developing cerebellar granule cells. J. Neurosci. 17: 4970-4981. Segal, R.A., Takahashi, H. and McKay, R.D.G. (1992) Changes in neurotrophin responsiveness during the development of cerebellar granule neurons. Neuron 9: 10411052. Sendtner, M., Holtmann, B., Kolbeck, R., Thoenen, H. and Barde, Y.-A. (1992) Brain-derived neurotrophic factor prevents the death of motoneurons in newborn rats after nerve section. Nature 360: 757-759. Shintani, A., Ono, Y., Kaisho, Y., Sasada, R. and Igarashi, K. (1993) Identification of the functional regulatory region of the neurotrophin-3 gene promoter. Mol. Brain Res. 17: 129-134. Siuciak, J.A., Altar, C.A., Wiegand, S.J. and Lindsay, R.M. (1994) Antinociceptive effect of brain-derived neurotrophic factor and neurotrophin-3. Brain Res. 633: 326-330.

Neurotrophin-3 and neurotrophin-4/5 Smith, M.A., Makino, S., Kvetnansky, R. and Post, R.M. (1995) Stress and glucocorticoids affect the expression of brain-derived neurotrophic factor and neurotrophin-3 in the hippocampus. J. Neurosci. 15: 1768-1777. Soppet, D., Escandon, E., Maragos, J., Middlemas, D.S., Reid, S.W., Blair, J., Burton, L.E., Stantin, B.R., Kaplan, D.R., Hunter, T., Nikolics, K. and Parada, L.F. (1991) The neurotrophic factors, brain-derived neurotrophic factor and neurotrophin-3 are ligands for the trkB tyrosine kinase receptor. CW/65: 895-903. Squinto, S.P., Stitt, T.N., Aldrich, T.H., Davis, S., Bianco, S.M., Radziejewski, C, Glass, D.J., Masiakowski, P., Furth, M.E., Valenzuela, D.M., DiStefano, P.S. and Yancopoulos, G.D. (1991) trkB encodes a functional receptor for brainderived neurotrophic factor and neurotrophin-3 but not nerve growth factor. Cell 65: 885-894. Stockel, K. and Thoenen, H. (1975) Retrograde axonal transport of nerve growth factor: specificity and biological importance. Brain Res. 85: 337-341. Studer, L., Spenger, C, Seiler, R.W., Altar, C.A., Lindsay, R.M. and Hyman, C. (1995) Comparison of the effects of the neurotrophins on the morphological structure of dopaminergic neurons in cultures of rat substantia nigra. Eur. J. Neurosci. 1: 223-233. Takeda, A., Onodera, H., Yamasaki, Y., Furukawa, K., Kogure, K., Obinata, M. and Shibahara, S. (1992) Decreased expression of neurotrophin-3 mRNA in the rat hippocampus following transient forebrain ischemia. Brain Res. 569: 177-180. Takeda, A., Onodera, H., Sugimoto, A., Kogure, K., Obinata, M. and Shibahara, S. (1993) Coordinated expression of messenger mRNAs for NGF, BDNF and NT3 in the rat hippocampus following transient forebrain ischemia. Neuroscience 55: 23-31. Tessarollo, L., Tsouflas, P., Martm-Zanca, D., Gilbert, D.J., Jenkins, N.A., Copeland, N.G. and Parada, L.F. (1993) trkC, a receptor for neurotrophin-3, is widely expressed in the developing nervous system and in non-neuronal tissues. Development 118: 463^75. Tessarollo, L., Vogel, K.S., Palko, M.E., Reid, S.W. and Parada, L.F. (1994) Targeted mutation in the neurotrophin3 gene results in loss of muscle sensory neurons. Proc. Natl. Acad. Sci. USA 91: 11844-11848. Tetzlaff, W., Kobayashi, N.R., Giehl, K.M., Tsui, B.J., Cassar, S.L. and Badard, A.M. (1994) Response of rubrospinal and corticospinal neurons to injury. Prog. Brain Res. 103: 271286. Thoenen, H. (1991) The changing scene of neurotrophic factors. Trends Neurosci. 14: 165-170. Timmusk, T., Belluardo, N., Metsis, M. and Persson, H. (1993) Widespread and developmentally regulated expression of neurotrophin-4 mRNA in rat brain and peripheral tissues. Eur. J. Neurosci. 5: 605-613. Timmusk, T., Lendahl, U., Funakoshi, H., Arenas, E., Persson, H. and Metsis, M. (1995) Identification of brain derived neurotrophic factor promoter regions mediat-

N. Rocamora and E. Arenas ing tissue-specific, axotomy- and neuronal activity-induced expression in transgenic mice. /. Cell. Biol. 128: 185199. Tsoulfas, P., Soppet, D., Escandon, E., Tessarollo, L., Mendoza-Ramirez, J.-L., Rosenthal, A., Nikolics, K. and Parada, L.F. (1993) The rat trkC locus encodes multiple neurogenic receptors that exhibit differential response to neurotrophin-3 in PC12 cells. Neuron 10: 975-990. Valenzuela, D.M., Maisonpierre, P.C., Glass, D.J., Rojas, E., Nunez, L., Kong, Y., Gies, D.R., Stitt, T.N., Ip, N.Y. and Yancopoulos, G.D. (1993) Alternative forms of rat TrkC with different functional capabilities. Neuron 10: 963-974. Ventimiglia, R., Mather, P.E., Jones, B.E. and Lindsay, R. (1995) The neurotrophins BDNF, NT-3 and NT-4/5 promote survival and morphological and biochemical differentiation of striatal neurons in vitro. Eur. J. Neurosci. 7: 2 1 3 222. Verdi, J.M. and Anderson, D.J. (1994) Neurotrophins regulate sequential changes in neurotrophin receptor expression by sympathetic neuroblasts. Neuron 13: 1359-1372. Verdi, J., Birren, S., Ibafiez, C , Benedetti, M., Chao, M.V. and Anderson, D.J. (1994) p75^^^^^ regulates trk signal transduction and NGF-induced neuronal differentiation in MAH cells. Neuron 12: 733-745. Verge, V.M.K., Merlio, J.-P., Grondin, J., Ernfors, P., Persson, H., Riopelle, R.J., Hokfelt, T. and Richardson, P.M. (1992) Colocalization of NGF binding sites trk mRNA and low-affinity NGF receptor mRNA in primary sensory neurons: responses to injury and infusion of NGF. J. Neurosci. 12:4011-4022. Vicario-Abejon, C., Johe, K.K., Hazel, T.G., Collazo, D. and McKay, R.D.G. (1995) Functions of basic-fibroblast growth factor and neurotrophins in the differentiation of hippocampal neurons. Neuron 15: 105-114. Vogel, K.S. and Davies, A.M. (1991) The durationof neurotrophic factor independence in early sensory neurons is matched to the time course of target field innervation. A^ewron 7: 819-830. Wescamp, G. and Reichardt, L.F. (1991) Evidence that biological activity of NGF is mediated through a novel subclass of high affinity receptors. Neuron 6: 649-663.

249 Widmer, H.R. and Hefti, H. (1994a) Neurotrophin-4/5 promotes survival and differentiation of rat striatal neurons developing in culture. Eur. J. Neurosci. 6: 1669-1679. Widmer, H.R. and Hefti, H. (1994b) Stimulation of GABAergic neuron differentiation by NT-4/5 in cultures of rat cerebral cortex. Dev. Brain Res. 80: 279-284. Wiegand, S.J., Anderson, K., Alexander, C , Criden, M., Altar, C.A., Lindsay, R.M. and DiStefano, P.S. (1992) Receptor binding and axonal transport of 1251-labeled neurotrophins in the basal ganglia and related brain regions. Movement Disord. 7: 63. Wong, v., Arriaga, R., Ip, N.Y. and Lindsay, R.M. (1993) the neurotrophins BDNF, NT-3 and NT-4/5, but not NGF, upregulate the cholinergic phenotype of developing motor neurons. Eur. J. Neurosci. 5: 466-474. Wright, E.M., Vogel, K.S. and Davies, A.M. (1992) Neurotrophic factors promote the maturation of developing sensory neurons before they become dependent on these factors for survival. Neuron 9: 139-150. Zafra, F., Hengerer, B., Leibrock, J., Thoenen, H. and Lindholm, D. (1990) Activity dependent regulation of BDNF and NGF mRNAs in the rat hippocampus is mediated by non-NMDA glutamate receptors. EMBO J. 9: 3545-3550, Zafra, F., Castren, E., Thoenen, H. and Lindholm, D. (1991) Interplay between glutamate and t-aminobutyric acid transmitter systems in the physiological regulation of brainderived neurotrophic factor and nerve growth factor synthesis in hippocampal neurons. Proc. Natl. Acad. Sci. USA 88: 10037-10041. Zafra, F., Lindholm, D., Castren, E., Hartikka, J. and Thoenen, H. (1992) Regulation of brain-derived neurotrophic factor and nerve growth factor mRNA in primary cultures of hippocampal neurons and astrocytes. J. Neurosci. 12: 4793-4799. Zhang, L., Schmidt, R.E., Yan, Q. and Snider, W.D. (1994) NGF and NT-3 have differing effects on the growth of dorsal root axons in developing mammalian spinal cord. J. Neurosci. 14: 5187.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 10

Centrally-active differentiation factors in the nervous system Lorraine lacovittti Institute of Neuroscience, Hahnemann University, Broad and Vine Streets, Philadelphia, PA 19102, USA

1. Introduction One of Nature's most intriguing paradoxes is that cells of identical genetic make-up give rise to organisms of immense complexity. This extraordinary diversity is accomplished during development by a process known as differentiation, whereby specific genes come to be expressed only in specific cell types. Possibly the most exquisite example of this occurs in the generation of the nervous system where over 50 biochemically distinct cell types have been identified. As in other systems, the selective expression of genes in the nervous system is, in large part, inherited in the genome of the neuron. However, in the last few decades, it has become increasingly clear that intrinsic and extrinsic, or epigenetic substances, called differentiation factors, also instruct neurons in their differentiative choices. 7.7. A working definition Over the years, the term differentiation factor has been ascribed to molecules which produce a variety of different biological activities in any number of cellular systems. In particular, systems such as the nervous and immune systems, which require the diversity of their component cells to achieve multiple functions, have been fertile areas of inquiry. Oftentimes, these independent lines of research have led to the discovery of the same factors, revealing a previously unsuspected degree of overlap in the two systems. However, in many cases, the same factors have subserved quite dif-

ferent functions. For example, substances such as the interleukins, first described for their classic actions as immune cytokines (for review, see Oppenheim et al., 1986), have recently been shown to influence the expression of neuropeptide transmitters in sympathetic neurons (Jonakait et al., 1990; Frieden and Kessler, 1991; Hart et al., 1991). Thus, the same molecules that promote proliferation or survival in one cellular (e.g. immune) system may act to drive differentiation in another (e.g. nervous) system. Moreover, within any one system, individual molecules may play more than one role, signaling at different sites or stages either proliferation, survival or differentiation. For the purposes of this review, it is therefore necessary to provide a working definition of the term differentiation factor. I will limit the discussion to those substances which, in addition to their potential functions elsewhere in the body, play a definitive role in the differentiation of the nervous system. In this context, the term will be used to denote only molecules with the ability to induce the novel expression of genes rather than those which merely amplify existing expression. In particular, this review will focus on substances involved in the activation of those genes encoding neurotransmitter biosynthetic enzymes since their expression is a requisite first step in the specification of biochemical phenotype. This restrictive definition excludes many molecules which have historically been considered differentiation factors (i.e. nerve growth factor) since they do not initiate, but merely modulate, expression of transmitterassociated genes.

252

2. A historical perspective: studies in the peripheral nervous system The existence of environmentally-derived differentiation factors was first postulated nearly a quarter of a century ago in a series of landmark studies on the peripheral nervous system (PNS) (for review see, LeDouarin, 1980). Using quail/chick chimeras, LeDouarin and co-workers demonstrated that the differentiative choices of neural crest cells, the progenitors of sympathetic and parasympthetic neurons, could be modified depending on their migration pathway and their site of final residence in the nervous system. This raised the distinct possibility that factors present in the environment during and after cell migration (presumably differentiation factors) governed the original choice of neurotransmitter substance in neural crest cells (Ziller et al., 1979; Ziller et al.,1987). Moreover, studies in tissue culture later showed that these early differentiative decisions were not irrevocable. Instead, the phenotype of neural crest derivatives (autonomic neurons) remained plastic with the potential for further refinement at later life stages. Thus, fully differentiated sympathetic neurons committed to a noradrenergic fate could be instructed to express differentiated traits of the cholinergic neurotransmitter system if exposed to the appropriate environmental agents in culture (Johnson et al., 1976; Patterson and Chun, 1977, Bunge et al., 1978). A number of substances, including; ciliary neurotrophic factor (Saadat et al., 1989; Rohrer, 1992), membrane-associated neurotransmitter stimulating factor (Wong and Kessler, 1987; Adler, 1989; Adler et al., 1989), leukemia inhibitory factor (Yamamori et al., 1989), human placental serum (lacovitti et al., 1981), chick embryo extract (lacovitti et al., 1981) and rat serum (Wollinsky and Patterson, 1983) have all possessed the remarkable ability to induce cholinergic function in sympathetic neurons. Moreover, many of these same substances/extracts were able to simultaneously modulate the levels of neuropeptides (Wong and Kessler, 1987; Ernsberger at al., 1989; Rao et al., 1992) and other classical transmitters (Sah and Matsumoto, 1987) in sympathetic neu-

Centrally-active differentiation factors in the nervous system

rons. Similarly, the transmitter synthetic machinery of parasympathetic neurons was also capable of modification under appropriate culture conditions, such that, cholinergic neurons from the chick ciliary ganglion could be made to express enzymes normally associated with an adrenergic phenotype (lacovitti et al., 1985; Teitelman et al., 1985). These pioneering tissue culture studies clearly documented that certain extracellular factors were capable of re-directing the biochemical choices of neurons previously committed to another neurotransmitter pathway. A comparable role for differentiation factors in vivo is supported by the work of Landis and colleagues (for review see, Landis, 1994). In these studies, a substance similar but not identical to leukemia inhibitory factor (Rao and Landis, 1990), has been implicated in the normal phenotypic trans-differentiation (from adrenergic to predominantly cholinergic) of sympathetic neurons innervating the sweat glands of the rat footpad. Interestingly, it was recently shown that the production of this target-derived cholinergic differentiation factor from sweat glands is dependent upon the presence of innervating neurons (Habecker et al., 1995). Thus, it appears that the phenotypic choices of peripheral neurons in vitro and in vivo are, at least in part, originally specified but may later be modified by the differentiation factors which they encounter locally. 3. Centrally-acting differentiation factors Although it has long been anticipated that factors analagous to those described in the PNS are also important in the differentiation of neurons in the central nervous system (CNS), only in the last year have molecules with this designated role been identified. Studies from this laboratory have indicated the existence of important differentiation factors for the CNS in a number of tissue extracts. This line of research originated with the observation that muscle and its constituents could dramatically alter the differentiative course of certain brain neurons. Thus, when neurons from the cerebral cortex, which do not normally manufacture catecholamine (CA) neurotransmitters, were grown in co-culture with heart, skeletal muscle.

L. lacovittti

vasculature or clonal muscle cells (L6 line), a significant number expressed tyrosine hydroxylase (TH), the rate-limiting enzyme in catecholamine biosynthesis (lacovitti et al., 1989). It was postulated that cytosolic agents present in muscle were responsible for the changes in neurotransmitter phenotype observed in these co-cultures. This was based primarily on two facts; first, that muscle could, to varying degrees, condition its medium and second, that the soluble fraction of muscle homogenate elicited a dose-dependent increase in TH activity and in the number of THimmunoreactive cortical neurons. The musclederived agents appeared to exert their effects in a highly specific manner; inducing TH but not other catecholamine biosynthetic enzymes. Dopamine-/?hydroxylase and phenylethanolamine-A^-methyl transferase were not induced by muscle extract, nor was there evidence of acquisition of catecholamine uptake capacity (lacovitti et al., 1989). 3.1. Mechanism of muscle action: muscle as a differentiation agent A number of different cellular/molecular mechanisms could have produced the striking and specific muscle-induced expression of TH in neurons, including increased cell division, improved survival or modified differentiation (i.e. via proliferation, trophic or differentiation factors present in muscle) (Fig. 1). Several lines of evidence supported the conclusion that the relevant muscle agents acted as differentiation factors. First, since those neurons which were responsive to the soluble fraction of muscle were postmitotic (lacovitti et al., 1989), increased TH expression could not have resulted from the selective proliferation of cells. Second, the fact that neurons grown in the presence of muscle or its soluble components did not exhibit improved viability (lacovitti et al., 1989) indicated that the effect was not due to trophic (survival) factors. Rather, it was believed that muscle contained putative differentiation factors which promoted the appearance of catecholaminergic traits in brain neurons that did not normally manufacture detectable levels of catecholamines. The substance(s) responsible for this

253

Striatal neurons

+ Muscle-derived

Fig. 1. Schematic depiction of several possible cellualr mechanisms that might account for the increase in the number of TH immunoreactive neurons after incubation overnight with a soluble muscle-derived substance(s)

complex bioactivity was (were) broadly termed muscle-derived differentiation factor/s (MDF), pending molecular identification. 3.2. Studies on muscle-derived differentiation factor 3.2.1. Muscle-derived differentiation factorresponsive non-catecholaminergic neurons MDF's hallmark effect was the striking and rapid induction of the TH gene in noncatecholaminergic neurons. Thus, exposure overnight to MDF resulted in de novo transcription of TH messenger RNA (mRNA) and translation of TH protein in specific subsets of brain neurons (lacovitti et al., 1989). Although this effect was first observed in neurons of the cerebral cortex (Fig. 2), it was not a unique feature of these neurons since populations from a wide variety of brain regions, including the cerebellum, collicular plate and striatum were also MDF-responsive. Of particular note was the striatum which contained the

254

Centrally-active differentiation factors in the nervous system

Fig. 2. Effect of MDF on TH expression in cultures of El 3 cerebral cortical cells. Photomicrographs are shown of cultures grow either on control media (a) or in the presence of MDF (b) overnight before being processed for the immunocytochemical localization of TH.

greatest proportion of MDF-responsive cells, with as many as 80% of its cells capable of expressing the normally quiescent TH gene (lacovitti, 1991). Since over 90% of striatal neurons use yaminobutyric acid (GABA) for neurotransmission in vivo (Mugnaini and Oertel, 1985) and 100% of cultured striatal neurons contain the GABAsynthetic enzyme glutamic acid decarboxylase (Max, Bossio and lacovitti, unpublished observations), it was reasoned that TH was being expressed in GABA-producing cells. These findings indicate that differentiation factors modify the biochemical status of striatal neurons not by redirecting their differentiation pathway from GABAergic to catecholaminergic but, rather, by fostering the synthesis of both transmitters simultaneously. Interestingly, not all brain regions contained MDF competent neurons; TH was not expressed in hippocampal neurons after MDF treatment

(lacovitti, 1991). Likewise, peripheral noncatecholaminergic neurons from the sensory ganglia were irresponsive to MDF. Thus, MDF's actions were confined to the CNS and, there, only to specific sub-populations of neurons from certain brain regions. 3.2.2. Time course of muscle-derived differentiation factor efficacy Further suggestive of MDF's role as a differentiation factor was the fact that it was effective only during precise periods in the development of noncatecholaminergic neurons (lacovitti, 1991). The window of MDF opportunity varied from brain region to brain region, but corresponded in all cases to the few days following peak withdrawal from mitosis, when neurons normally choose a differentiative course. If competent cells were explanted to culture during those few days and treated with MDF overnight, TH was expressed.

L. lacovittti

After close of the critical period in vivo, their phenotype became fixed, making further modification by incubation in vitro with MDF impossible. The timing of the critical period did not appear to depend on interactions with other developing cells in situ but, instead, was inherent in the neurons. This was suggested by the finding that neurons grown in culture and thus isolated from target and other influences were not MDF-responsive if initial treatment was delayed several days. Thus, the initiation of TH expression required exposure of competent neurons to MDF during a critical window in their development. The perpetuation of TH expression in culture required MDF's continued presence there, indicative of its probable depletion or degradation over time in vitro (lacovitti, 1991). Quite surprisingly, however, a single priming exposure to MDF during the critical period, allowed neurons at later stages of life to re-express TH if re-challenged with the muscle extract. It was postulated that early interactions of the genome with differentiation factors like those found in muscle may establish a biochemical memory which makes possible transmitter phenotypic plasticity later in life (lacovitti, 1991). 3.3. Identification of dopamine differentiation factors 3.3.1. Muscle-derived agents Purification studies were initiated in an attempt to identify those agents in muscle responsible for the induction of TH in cultured mouse striatal neurons. These studies revealed that bioactivity in vitro required not one reagent but the obligate interaction of two substances present in muscle, neither of which alone possessed intrinsic TH-inducing ability. One requisite component was subsequently identified as acidic fibroblast growth factor (aFGF), the other an as yet unidentified molecule(s) of < lOkDa molecular weight (Du et al, 1994). Thus, muscle-derived aFGF, if incubated in the presence but not the absence of the < 10 kDa fraction of muscle, induced a dose-dependent increase in the number of striatal neurons which novelly expressed TH. This expression was blocked by prior incubation and precipitation of the factor

255

with polyclonal antibodies to aFGF. Similarly, commercial preparations of native bovine and human recombinant aFGF were potent inducers of TH when co-incubated with the < 10 kDa molecule. Heparin, a known activator of aFGF, did not potentiate the growth factor's TH-inducing activity, indicating that the unidentified muscle agent was not heparin (Du et al., 1994). These data thus support the conclusion that MDF activity requires the synergistic participation of at least two partner molecules present in muscle extract: aFGF and an unidentified <10kDa molecule(s) which is not heparin. 3.3.2. Agents found in brain extract Although exposure in culture to these substances triggered novel expression of the normally quiescent TH gene in non-catecholaminergic neurons, it is not yet clear whether aFGF and its partner substance are also responsible for signaling the routine expression of TH in catecholaminergic neurons. Supporting this possibility, however, is the co-incidental appearance of TH (Specht et al, 1981) simultaneously with aFGF in differentiating catecholaminergic neurons of the developing brainstem (Ferrari et al., 1989; Engele and Bohn, 1991; Fu et al., 1991; Schnurch and Risau, 1991; Wilcox and Unnerstall, 1991; Nurcombe et al., 1993). The proposition that catecholaminergic neurons possess or have local access to all of the agents necessary for their own biochemical differentiation, including aFGF and its partner molecule, was tested by evaluating the TH-inducing capacity of an extract of catecholaminergic neurons derived from adult substantia nigra (SN) tissue. Interestingly, SN extract, similarly to muscle extract, was capable of inducing TH expression in cultured striatal neurons. Since SN neurons, in addition to their supply of endogenous aFGF (Engele and Bohn, 1991; Schnurch and Risau; 1991; Wilcox and Unnerstall, 1991; Nurcombe et al., 1993), also contain high concentrations of catecholamine neurotransmitters, it was conceivable that both were involved in TH induction. When tested, the combined treatment of aFGF with transmitters, specifically CA transmitters, resulted in a striking induction in TH in striatal neurons. Dopamine (DA) was

256

particularly effective as an aFGF partner, initiating the expression of TH in the majority of cultured striatal neurons (Du and lacovitti, 1995). However, dose-response studies revealed that other catecholamine-containing compounds, such as L-dopa, norepinephrine and epinephrine, were also effective, albeit at higher concentrations. In contrast to the effects of catecholamines in combination with aFGF, incubation of cultures of striatal neurons with aFGF plus various concentrations of monomine precursors or their metabolites or other neurotransmitter compounds, caused no TH induction. Along with aFGF, a number of other growth factors have also been implicated in the regulation of dopaminergic traits in the developing brain (Knusel et al., 1990; Hyman et al, 1991; Knusel et al., 1991; Beck et al, 1992; Lewis et al, 1993; Louis et al., 1993; Magal et al., 1993; Sauer et al, 1993; Zhou et al., 1994). Studies designed to examine that issue revealed that both basic fibroblast growth factor (bFGF) and brain-derived neurotrophic factor (BDNF) were also capable of moderate levels of TH induction in striatal neurons (Du et al., 1995). As with aFGF, the activities of bFGF and of BDNF depended on their coupling with a catecholamine partner molecule. In comparing these growth factors with aFGF, the similarity in their time courses of action suggested that these agents may be working through analogous cellular mechanisms. However, their dose-response curves indicate that aFGF is a more potent and more efficatious inducer of TH than either bFGF or BDNF. In contrast, regardless of the partner molecule used, other growth factors such as glial-derived neurotrophic factor (GDNF), FGF-6, FGF-7, epidermal growth factor (EGF), nerve growth factor (NGF), ciliary neurotrophic factor (CNTF), leukemia inhibitory factor (LIF), transforming growth factor p (TGF^), and interleukin 1 (ILl) did not elicit significant levels of TH induction (Du and lacovitti, 1995; Du et al, 1995), although they may be involved in the enzyme's subsequent modulation (Lewis et al., 1993; Louis et al., 1993; Magal et al., 1993). Thus, only certain specific growth factors, including aFGF, bFGF and/or BDNF, working coordinately with partner molecules, can initiate TH gene expression in vitro.

Centrally-active differentiation factors in the nervous system

3.33. Mechanisms mediating TH induction by growth factors and catecholamines An understanding of the mechanisms responsible for bringing about TH expression in cultured striatal neurons is only now beginning to emerge. It is presumed that aFGF, bFGF and BDNF exert their effects by binding to one the high-affinity tyrosine kinase-linked membrane receptors. Although it is still unproven, this notion is supported by the recent finding that growth factor-induced TH expression is inhibited by pretreatment with the tyrosine kinase inhibitors, tyrphostin B46 or genistein (Du, Stull and lacovitti, unpublished observations), catecholamine partners, on the other hand, may mediate their cellular actions through several routes. First, membrane receptors for catecholamines, including DA receptors (Fremont et al., 1983; Maus et al., 1989) as well as a (Weiss et al., 1987) and^S (Van Vliet et al., 1991) adrenoceptors present on cultured striatal neurons might bring about TH induction via activation of second messenger systems. Indeed, SKF 38393, a D1/D5 receptor agonist, successfully substitutes for DA, inducing TH expression in the majority of those striatal neurons co-treated with aFGF (Du and lacovitti, 1995). Surprisingly, however, induction is not blocked by treatment with the Dl receptor antagonists SCH 23390, apomorphine or haloperidol. Similarly, the induction produced by the beta adrenoceptor agonists NE and isoproterenol is not prevented by the specific beta antagonist propranolol (Du and lacovitti, 1995). Agonists and antagonists at other DA receptor (D2, D3, D4) or at alpha-adrenoceptors neither stimulated nor inhibited TH expression in striatal neurons. In addition to exerting their action through classic receptor-mediated mechanisms, catecholamines may also gain direct intracellular access after uptake via a neurotransporter system. In general, striatal neurons contain uptake sites for GABA (Radian et al., 1990) but not DA. However, neither GABA uptake blockers (nipecotic acid; 1amino-1-cyclohexane carboxylic acid) nor DA uptake blockers (mazindol, nomifensine, GYKI 52895) prevent induction by aFGF and DA (Du and lacovitti, 1995). Paradoxically, mazindol mimicked the effects of DA, producing 40% TH

L. lacovittti

induction when co-incubated with aFGF. Although it is possible that by blocking re-uptake of DA, mazindol increased the extracellular levels of transmitter sufficiently to partner aFGF, its additive effects when DA was also added to the incubation mixture suggests that the mechanisms of action of the two reagents may be different. On the surface, the molecules which successfully partner aFGF appear quite dissimilar (D1/D5 agonists, beta-adrenoceptor agonists, dopamine uptake inhibitors). However, several distinguishing biochemical features are shared by effective partners; all contain an amine group separated from a catechol nucleus by two carbons. Despite the fact that adult striatal neurons do not contain the DA neurotransporter, these traits are virtually identical to those required for optimal uptake of DA (Meiergerd and Schenk, 1994). It is conceivable that, unlike adult neurons, embryonic striatal cells contain the DA transporter in addition to the usual GABA transporter (Radian et al., 1990). The coexistence of GABA and DA transporters on the same cell has been reported previously (Bonanno and Raiteri, 1987). Another possibility is that embryonic striatal neurons contain a new neurotransporter or a less discriminating form of the GABA transporter which also recognizes DA and other related catechols. Alternatively, since transport blockers do not inhibit TH induction, DA may utilize a totally novel mechanism, such as lipid peroxidation, to increase gene transcription and stimulate biochemical differentiation in striatal neurons, as has been observed in a variety of developing tissues (Allen, 1991). 3.3.4. Relevance to catecholaminergic differentiation Whether any or all of the aforementioned substances actually play a role in directing the choice of a dopaminergic phenotype in vivo remains an open quetsion. Certainly, FGF and BDNF are found locally in the brainstem (Ferrari et al., 1989; Engele and Bohn, 1991; Fu et al, 1991; Schnurch and Risau; 1991; Wilcox and Unnerstall, 1991; Nurcombe et al., 1993; Seroogy and Gall, 1993; Seroogy et al., 1994) during the period when catecholaminergic neurons first differentiate

257

(Specht et al., 1981). Regardless of the growth factor involved, activation by partner molecules such as catecholamines appears to be a necessary step in TH gene induction. The requirement for endogenous catecholamines, however, poses an obvious paradox for neurons attempting to first initiate catecholamine synthesis. Consequently, bioenic amines must originate extraneuronally if they are to act as partners to aFGF during differentiation. Although this putative role for catecholamines in differentiation is indeed consistent with the widely held view that neurotransmitters mediate many critical functions in foetal development (Coyle, 1972; Lauder, 1988; Mattson, 1988; Meier et al., 1991), it is still possible that these are not the physiologically relevant partners but are merely mimicking their effects by activating common signalling pathways to those required for TH induction. 3.4. Regulation ofTH by dopamine differentiation factors in vivo and in vitro 3.4.1. Effects on cultured catecholaminergic neurons during development The effects of aFGF and DA on noncatecholaminergic neurons established their potential role as differentiation factors in the developing brain. Still open was the question of whether these factors might also serve in a regulatory capacity, modulating TH enzyme levels in those neurons which normally synthesize and use catecholamine neurotransmitters. This possibility was first suggested by our early studies on partially purified factors from muscle (MDF) (lacovitti et al., 1992). When neurons which normally express TH (dopamine neurons from the developing rat mesencephalon) were treated with MDF in culture, a striking increase in TH mRNA and TH activity was observed (lacovitti et al., 1992). These early experiments were partially repeated using pure reagents (aFGF and DA) (Stull and lacovitti, 1996). Although catecholamine transmitters have traditonally been considered feedback inhibitors of TH (Nagatsu et al., 1964; Spector et al, 1967; Weiner et al., 1972; Zigmond et al., 1989), these results suggest that, when combined with aFGF, catecholamines may instead act as feedback indue-

258

Centrally-active differentiation factors in the nervous system

ers of TH. The cross-talk between growth factors and neurotransmitters may be essential for proper regulation of TH at the transcriptional and translational levels, as has been observed in other systems (Manietal., 1994). 3A. 2, Effects in the intact adult rat brain Unfortunately, because it is not possible to isolate dopaminergic neurons for culture from embryos older than El8, the age-dependence of these effects could not be directly tested as it had been in non-catecholaminergic neurons. However, in our initial studies with MDF, in situ infusion into the adult intact rat (Fig. 3) produced none of the changes observed in developing neurons, with no detectable increments in TH activity, DA or 3,4dihydroxyphenylacteic acid (DOPAC) levels (Jin and lacovitti, 1995). In addition, the localization of

TH protein by immunocytochemical staining or of TH mRNA by in situ hybridization was unmodified by treatment. Finally, there were no observed differences in motor behavior in infused animals compared with untreated controls. These findings suggest that, like non-catecholaminergic neurons, catecholaminergic neurons only respond to exogenous DA differentiation factors during a brief developmental period when endogeous agents are likely to be at physiologically low levels. Once endogenous concentrations reach adult steady state, neurons may become refractory to their exogenous application. 3A.3. Effects on catecholaminergic neurons following damage Although DA differentiation factors have no apparent effect on the intact adult brain, their abil-

MDF or vehicle is osmotically pumped into the striatum of an adult rat.

Rats are sacrificed 7-14 days later and the brains analyzed.

\ \

Fig. 3. Schematic depiction of the rat model used to test MDF's effects in vivo after MDF infusion into the intact striatum.

259

L. lacovittti

6-OHDA is injected into the substantia nigra producing a partial lesion of dopamine fibers in the striatum. Four weeks later MDF or PBS + BSA is osmotically pumped into the damaged striatum.

Partial lesion Substantia nigra

Animals are allowed to recover and motor behaviour monitored prior to sacrifice.

Fig. 4. Schematic depiction of the model used to test MDF's effects in vivo after infusion into the denervated striatum of a unilaterally 6-hydroxydopamine-lesioned rat. PBS, phosphate buffered saline (pH = 7.2); BSA, bovine serum albumin.

ity to up-regulate TH during development suggests their possible efficacy under other conditions of reduced gene expression, such as after injury or disease. My research group therefore examined, in vivo or in vitro, whether neurotoxin-induced deficits rendered dopaminergic neurons responsive to these agents once again. If so, the possibiHty was raised that the administration of DA differentiation factors might be a useful therapy for amplifying transmitter synthesis in circumstances of known catecholamine compromise, as occurs in Parkinson's disease. This proposition was tested using a well-studied animal model of Parkinson's disease in which partial lesions of the nigrostriatal system are created by unilateral microinjection of the catecholaminergic-specific neurotoxin 6-hydroxydopamine (6-

OHDA) into the rat substantia nigra (Fig. 4). Toxin treatment resulted 2-4 weeks later in the loss of dopaminergic neurons and their target projections in the dorsolateral striatum on the side of the lesion. Post mortem neurochemical analysis revealed a concomitant depletion in TH activity, DA and DOPAC levels on the lesioned side. The imbalance in transmitter levels from one side of the brain to the other produces a characteristic motor syndrome, causing animals to circle robustly in the direction of the lesion if challenged with a DA agonist such as amphetamine. By monitoring motor behavior, this model makes it possible to study the efficacy of compounds like MDF, aFGF, etc. in reversing motor assymetry when they are directly infused into the damaged striatum via a mini-osmotic pump. Moreover, at the conclusion

260

of the experiment, brains can be processed for DA histochemistry and biochemistry so that post mortem measurements can be correlated with behavioral observations. A 2-week infusion of MDF or human recombinant aFGF and/or its muscle-derived partner substance into the striatum of unilaterally 6-OHDAlesioned rats caused a significant (48-100%) and long-lasting reduction in amphetamine-induced rotational asymmetry (Jin and lacovitti, 1995). In parallel with behavioral recovery, striatal TH activity, DA and DOPAC levels recovered in a dosedependent manner in all treated rats when compared to controls. The greatest increments were observed in rats infused with aFGF and its musclederived partner. Since the number of THimmunoreactive neurons and their striatal innervation were unmodified by treatment, it is unlikely that the observed changes resulted from the trophic properties of these substances. Instead, DA differentiation factors appear to mediate their effects by increasing dopaminergic neurochemical indices. Similarly, in mice made Parkinsonian by another DA-specific neurotoxin, l-methyl-4-phenyl1,2,3,6-tetrahydropyridine (MPTP), infusion of MDF or aFGF also resulted in a significant increase in striatal TH activity and DOPAC levels. Since the increases were not accompanied by a comparable change in DA levels, these findings were indicative of an elevation in DA synthesis (TH/DA) and turnover (DOPAC/DA) (Jin and lacovitti, 1996). Likewise, studies in culture revealed that treatment of MPP+-damaged dopaminergic neurons with aFGF and DA also reversed the toxin-induced loss in TH activity, returning TH to control levels (StuU and lacovitti, 1996). These in vivo and in vitro findings indicate that aFGF and its partner molecules, apart from their differentative properties, can also have important modulatory effects on the dopaminergic system of lesioned animals. In summary, studies in vivo, taken together with previous results in culture (lacovitti et al., 1992; Du et al., 1994), raise the hypothesis that aFGF together with its partner may initiate and subsequently regulate the levels of TH and ultimately neurotransmitter production in dopaminergic neu-

Centrally-active differentiation factors in the nervous system

rons. If this does occur, then it is reasonable that, as dopaminergic neurons begin to degenerate in Parkinson's disease, the intrinsic supply of their regulatory factors declines. Replacement of missing DA differentiation factors, therefore, may prove a useful way of restoring brain transmitter levels. Unlike current therapies for Parkinson's disease, these agents should selectively affect injured adult dopaminergic neurons, thereby minimizing the risk of side-effects caused by the unintentional involvement of uninjured catecholaminergic neurons. 4. Concluding remarks The studies of the past 25 years have clearly established the role of growth factors as critical determinants of biochemical phenotype in peripheral neurons (Saadat and Thoenen, 1986; Wong and Kessler, 1987; Adler, 1989; Yamamori et al., 1989; Rohrer, 1992; Howard and Gershon, 1993; Rao and Landis, 1993). More recently, studies on the brain suggest that the induction of neurotransmitter-specific genes may involve more complex mechanisms, requiring the obligatory interactions of multiple signal molecules (Du et al., 1994). Although growth factors play a prominent role, their cooperation with auxiliary agents is essential. Thus, exposure in culture both to a specific growth factor (aFGF, bFGF or BDNF) and an additional partner molecule is necessary to trigger novel expression of the normally quiescent TH gene in non-catecholaminergic neurons of the striatum (Du et al., 1994, 1995; Du and lacovitti, 1995). Although the required participation of a partner molecule may be valuable for restricting the effects of ubiquitous growth factors, the availability of more than one partner could provide the versatility needed to induce the TH gene under a variety of different conditions. Conceivably, these different partner molecules might achieve their effects via different signalling pathways, leading to activation of one or more of the many important transcription binding sites already identified on the TH gene (Lewis and Chikaraishi, 1987; Jones et al., 1988; Kemmler et al., 1989; Carroll et al, 1991; Fung et al., 1992; Kim et al, 1993). Consis-

261

L. lacovittti

tent with this view is the recent finding that tissuespecific expression of TH indeed requires an interaction between two consensus binding sites on the gene (Yoon and Chikaraishi, 1992). In addition to mechanistic differences, several other important distinctions between peripheral and central differentiative processes have been observed. First, in peripheral neurons, the biochemical phenotype appears to be labile for protracted periods, extending even into postnatal life. If DA differentiation factors are behaving as prototypic centrally-acting differentiation factors, then brain neurons may be receptive to these agents only during precise developmental periods (lacovitti, 1991). In fact, central neurons appear to have only several days after their withdrawal from mitosis to establish their differentiative pathway. Once this window in time is closed, the chemical phenotype apparently becomes fixed, with future adjustment possible only under certain circumstances. In the case of DA differentiation factors, several opportunities for phenotypic plasticity have been identified. First, provided that brain neurons have been transiently exposed to the factors during the critical preriod, then TH can be re-expressed upon re-challenge at a later time (lacovitti et al., 1989). Second, as has been observed in the PNS (Landis, 1994), if neurons become damaged, as with a neurotoxin, then they are again receptive to the influence of differentiation factors (Jin and lacovitti, 1995, 1996; Stull and lacovitti, 1996). This latter trait suggests that manipulation of the genes responsible for neurotransmitter production with highly specific factors may represent a new therapeutic approach to treat damaged, diseased or aged transmitter systems. Indeed, in vivo studies with DA differentiation factors have suggested that up-regulation of TH may lessen the neurotransmitter shortages associated with Parkinsonism (Jin and lacovitti, 1995; 1996). Conversely, identification of substances able to antagonize the actions of differentiation factors may be helpful in diseases characterized by an overproduction of transmitter. Although, to date, only DA differentiation factors have been identified, the multitiude of transmitter types in the brain suggests that many other factors await discovery.

Acknowledgements This work was supported by NIH NS24204 and the Amgen Corp. References Adler, J.E. (1989) Neuronal aggregation and neurotransmitter regulation: partial purification and characterization of a membrane-derived factor. Int. J. Dev. Neurosci. 7: 533538. Adler, J.E., Leonard, S.S. and Black, LB. (1989) Partial purification and characterization of a membrane-derived factor regulating neurotransmitter phenotypic expression. Proc. Natl Acad. Sci. USA 86: 1080-1083. Allen, R.G. (1991) Oxygen-reactive species and antioxidant responses during development: the metabolic pardox of cellular differentiation. Proc. Soc. Exp. Biol. Med. 196: 117-129. Beck, K.D., Knusel, B., Winslow, J.W., Rosenthal, A., Burton, L.E., Nikolics, K. and Hefti, F. (1992) Pretreatment of dopaminergic neurons in culture with brain-derived neurotrophic factor attenuates toxicity of l-methyl-4phenylpyridinium. Neurode generation 1: 27-36. Bonanno, G. and Raiteri, M. (1987) Coexistence of carriers for dopamine and GABA uptake on a same nerve terminal in the rat brain. Br. J. Pharmacol. 91: 237-243. Bunge, R.P., Johnson, M. and Ross, CD. (1978) Nature and nurture in the development of the autonomic neurons. Science 199: 1409-1415. Carroll, J.M., Evinger, M.J., Goodman, H.M. and Joh, T.H. (1991) Differential and coordinate regulation of TH and PNMT MRNAs in chromaffin cell cultures by second messenger system activation and steroid treatment. J. Mol. Neurosci. 3: 75-83. Coyle, J.T. (1972) Tyrosine hydroxylase in rat-brain-cofactor requirements, regional and subcellular distribution. Biochem. Pharmacol 21: 1935-11944. Du, X. and lacovitti, L. (1995) Synergy between growth factors and neurotransmitters required for catecholamine differentiation in brain neurons. J. Neurosci 15: 5420-5427. Du, X., Stull, N.D. and lacovitti, L. (1994) Novel expression of the tyrosine hydroxylase gene requires both acidic fibroblast growth factor and an activator. J. Neurosci 14: 7688-7694. Du, X., Stull, N.D. and lacovitti, L. (1995) Brain-derived neurotrophic factor works coordinately with partner molecules to initiate tyrosine hydroxylase expression in striatal neurons. Brain Res. 680: 229-233. Engele, J. and Bohn, M.C. (1991) The neurotrophic effects of fibroblast growth factors on dopaminergic neurons in vitro are mediated by mesencephalic glia. J. Neurosci 11: 30703078. Ferrari, G., Minozzi, M.C, Toffano, G., Leon, A. and Skaper,

262 S.D. (1989) Basic fibroblast growth factor promotes the survival and development of mesencephalic neurons in culture. Dev. Biol. 133: 140-147. Frieden, M. and Kessler, J.A. (1991) Cytokine regulation of substance P expression in sympathetic neurons. Proc. Natl. Acad. Sci. USA 88: 3200-3203. Fu, Y.-A., Spirito, P., Yu, Z.-X., Biro, S., Sasse, J., Lei, J., Ferrans, V.J., Epstein, S.E. and Casscells, W. (1991) Acidic fibroblast growth factor in the developing rat embryo. Cell Biol. 114: 1261-1273. Fung, B.P., Yoon, S.O. and Chikaraishi, D.M. (1992) Sequences that direct the rat tyrosine hydroxylase gene expression. J. Neurochem. 58: 2044-2052. Habecker, B.A., Tressler, S.J., Rao, M.S. and Landis, S.C. (1995) Production of sweat gland cholinergic differentiation factor depends on innervation. Dev. Biol. 167: 307316. Hart, R.P., Shadiack, A. and Jonakait, G.M. (1991) Substance P gene expression is regulated by interleukin-1 in cultured sympathetic ganglia. J. Neurosci. Res. 29: 282-291. Howard, M.J. and Gershon, M.D. (1993) Role of growth factors in catecholaminergic expression by neural crest cells: in vitro effects of transforming growth factor betaj. Dev. Dynam. 196: 1-10. Hyman, C, Hofer, M., Barde, Y.A., Juhasz, M.Y., Squinto, G.D. and Lindsay, R.M. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350: 230-235. lacovitti, L. (1991) Effects of a novel differentiation factor on the development of catecholamine traits in noncatecholamine neurons from various regions of the rat brain: studies in tissue culture. J. Neurosci. 11: 2403-2409. lacovitti, L., Joh, T.H., Park, D.H. and Bunge, R.P. (1981) Dual expressions of neurotransmitter synthesis in cultured autonomic neurons. J. Neurosci. 1: 685-690. lacovitti, L., Joh, T.H., Albert, V.R., Park, D.H., Reis, D.J. and Teitelman, G. (1985) Partial expression of catecholaminergic traits in cholinergic chick ciliary ganglia: studies in vivo and in vitro. Dev. Biol. 110: 402-412. lacovitti, L., Evinger, M.J., Joh, T.H. and Reis, D.J. (1989) A muscle-derived factor induces expression of a catecholamine phenotype in cultured rat cerebral cortex. J. Neurosci. 9: 3529-3537. lacovitti, L., Evinger, M.J. and Stull, N.D. (1992) Musclederived differentiation factor increases expression of the tyrosine hydroxylase gene and enzyme activity in cultured dopamine neurons from the rat midbrain. Mol. Brain Res. 16: 215-222. Jin, B.K. and lacovitti, L. (1995) Dopamine differentiation factors produce partial motor recovery in 6hydroxydopamine lesioned rats. Neurobiol. Dis. 2: 1-12. Jin, B.K. and lacovitti, L. (1996) Dopamine differentiation factors increase striatal dopaminergic function in 1-methyl4phenyl-l,2,3,6tetrahydropyridine (MPTP) - lesioned mice. J. Neurosci. Res. 43: 331-334. Johnson, M., Ross, D., Meyers, M., Rees, R., Bunge, E., Wak-

Centrally-active differentiation factors in the nervous system shull, E. and Burton, H. (1976) Synaptic vesicle cytochemistry changes when cultured sympathetic neurons develop cholinergic interactions. Nature 262: 308-310. Jonakait, G.M., Schotland, S. and Hart, R.P. (1990) Interleukin-1 specifically increases substance P in injured sympthetic ganglia. In: M.S. O'DCorisio and A. Panerai (Eds.), Neuropeptides and Immunopeptides: Messengers in a Neuroimmune Axis. New York Academy of Sciences, New York, pp. 222-230. Jones, N.C., Rigby, P.W.J, and Ziff, E.B. (1988) Trans-acting protein factor and the regulation of eukaryotic transcription: lessons from studies on DNA tumor viruses. Genes Dev. 2: 267-281. Kemmler, I., Schreiber, E., Muller, M.M., Matthias, P. and Schaffner, W. (1989) Octamer transcription factor bind to two different sequence motifs of the immunoglobulin heavy chain promotor. EMBO J. 8: 2001-2008. Kim, K.T., Park, D.H. and Joh, T.H. (1993) Parallel upregulation of catecholamine biosynthetic enzymes by dexamethasone in PC 12 cells. J. Neurochem. 60: 946-951. Knusel, B., Michel, P.P., Schwaber, J.S. and Hefti, F. (1990) Selective and nonselective stimulation of central cholinergic and dopaminergic development in vitro by nerve growth factor, basic fibroblast growth factor, epidermal growth factor, insulin and the insulin-like growth factors I and II. /. Neurosci. 10: 558-570. Knusel, B., Winslow, J.W., Rosenthal A., Burton, L.E., Seid, D.P., Nikolics, K. and Hefti, F. (1991) Promotion of central cholinergic and dopaminergic neuron differentiation by brain-derived neurotrophic factor but not neurotrophin 3. Proc. Natl. Acad. Sci. USA 88: 961-965. Landis, S.C. (1994) Development of sympathetic neuroneurotransmitter plasticity and differentiation factors. Prog. Brain Res. 100: 19-23. Lauder, J. (1988) Neurotransmitters as morphogens. Prog. Brain Res. 73: 365-387. Le Douarin, N.M. (1980) The ontogeny of the neural crest in avian embryo chimaeras. Nature 286: 663-669. Lewis, S.E., Rao, M.S., Symes, A.J., Dauer, W.T., Fink, J.S., Landis, S.C. and Hyman, S.E. (1993) Coordinate regulation of choline acetyltransferase, tyrosine hydroxylase, and neuropeptide mRNAs by ciliary neurotrophic factor and leukemia inhibitory factor in cultured sympathetic neurons. J. Neurochem. 63: 429-438. Lewis, E.J. and Chikaraishi, D.M. (1987) Regulated expression of the tyrosine hydroxylase gene by epidermal growth factor. Mol. Cell Biol. 7: 3332-3336. Louis, J.-C, Magal, E., Burnham, P. and Varon, S. (1993) Cooperative effects of ciliary neurotrophic factor and norepinephrine on tyrosine hydroxylase expression in cultured rat locus coeruleus neurons. Dev. Biol. 155: 1-13. Magal, E., Burnham, P., Varon, S., and Louis, J.-C. (1993) Convergent regulation by ciliary neurotrophic factor and dopamine of tyrosine hydroxylase expression in cultures of rat substantia nigra. Neuroscience 52: 867-881. Mani, S.K., Allen, J.M.C., Clark, J.H., Blaustein, J.D. and

L. lacovittti O'Malley, B.W. (1994) Convergent pathways for steroid hormone- and neurotransmitter-induced rat sexual behavior. Science 265: 1246-1249. Mattson, M. (1988) Neurotransmitters in the regulation of neuronal cytoarchitecture. Brain Res. Rev. 13: 179-212. Maus, M.P., Bertrand, S., Drouva, R., Rasolonjanahary, C, Kordon, J., Glowinski, J., Fremont, J. and Enjalbert, A. (1989) Differential modulation of D] and D2 dopaminesensitive adenylate cyclases by 17y3-estradiol in cultured striatal neurons and anterior pituitary cells. J. Neurochem. 52:410-418. Meier, E., Hertz, L. and Schousboe, A. (1991) Neurotransmitters as developmental signals. Neurochem. Int. 19: 1-15. Meiergerd, S.M. and Schenk, J.O. (1994) Striatal transporter for dopamine: catechol structure-activity studies and susceptibility to chemical modification. J. Neurochem. 62: 998-1008. Mugnaini, E. and Oertel, W.H. (1985) An atlas of the distribution of gabaergic neurons and terminals in the rat CNS as revealed by GAD immunohistochemistry. In: A. Bjorklund and T. Hokfelt (Eds.), Handbook of Chemical Neuroanatomy. GABA and Neuropeptides in the CNS, Part I, Vol. 4. Elsevier, Amsterdam, pp. 436-595. Nagatsu, T., Levitt, M. and Udenfriend, S. (1964) Tyrosine hydroxylase: the initial step in norepinephrine biosynthesis. /. Biol. Chem. 239: 2910-29117. Nurcombe, V., Ford, M.D., Wildschut, J.A. and Bartlett, P.P. (1993) Developmental regulation of neural response of FGF-1 and FGF-2 by heparan sulfate proteoglycan. Science 260: 103-106. Oppenheim, J.J., Kovacs, E.J., Matsushima, K. and Durum, S.K. (1986) There is more than one interleukin 1. Immunol. Today 7: 45-56. Patterson, P.H. and Chun, L.L. (1977) The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. I. Effects of conditioned medium. Dev. Biol. 56: 263-280. Premont, J., Dauguet-de Montety, M.-C, Herbet, A., Glowinski, J., Bockaert, J. and Prochiantz, A. (1983) Biogenic amines and adenosine-sensitive adenylate cyclases in primary cultures of striatal neurons. Brain Res. 285: 53-61. Radian, R., Ottersen, O.P., Storm-Mathisen, J., Castel, M. and Kanner, B.I. (1990) Immunocytochemical localization of the GABA transporter in rat brain. /. Neurosci. 10: 13191330. Rao, M.S. and Landis, S.C. (1990) Characterization of a target-derived neuronal cholinergic differentiation factor. Neuron 5: S99-910. Rao, M.S., Tyrrell, S., Landis, S.C. and Patterson, P.H. (1992) Effects of ciliary neurotrophic factor (CNTF) and depolarization on neuropeptide expression in cultured sympathetic neurons. Dev. Biol. 150: 281-293. Rao, M.S. and Landis, S.C. (1993) Cell interactions that determine sympathetic neuron transmitter phenotype and the neurokines that mediate them. J. Neurobiol. 24: 215-232. Rohrer, H. (1992) Cholinergic neuronal differentiation fac-

263 tors: evidence for the presence of both CNTF-like and nonCNTF-like factors in developing rat footpad. Development 114:689-698. Saadat, S., Sendtner, M. and Rohrer, H. (1989) Ciliary neurotrophic factor induces cholinergic differentiation of rat sympathetic neurons in culture. J. Cell Biol. 108: 18071816. Saadat, S. and Thoenen, H. (1986) Selective induction of tyrosine hydroxylase by cell-cell contact in bovine adrenal chromaffin cells is mimicked by plasma membranes. J. Cell Biol. 103: 1991-1997. Sah, W.Y.and Matsumoto, S.G. (1987) Evidence for serotonin synthesis, uptake, and release in dissociated rat sympathetic neurons in culture. J. Neurosci. 7: 391-399. Sauer, H., Fischer, W., Nikkhah, G., Wiegand, S.J., Brundin, P., Lindsay, R.M. and Bjorklund, A. (1993) Brain-derived neurotrophic factor enhances function rather than survival of intrastriatal dopamine cell-rich grafts. Brain Res. 626: 37^4. Schntirch, H. and Risaaau, W. (1991) Differentiating and mature neurons express the acidic fibroblast grwoth factor gene during chick neural development. Development 111: 1143-1154. Seroogy, K.B. and Gall, CM. (1993) Expression of neurotrophins by midbrain dopaminergic neurons. Exp. Neurol. 124: 119-128. Seroogy, K.B., Lundgre, K.H., Tran, T.M.D., Guthrie, K.M., lasackson, P.J. and Gall, CM. (1994) Dopaminergic neurons in rat ventral midbrain express brain-derived neurotrophic factor and neurotrophin-3 mRNAs. J. Comp. Neurol. 342: 321-334. Specht, L.A., Pickel, V.M., Joh, T.H. and Reis, D.J. (1981) Light microscopic immunocytochemical localization of tyrosine hydroxylase in prenatal rat brain. I. Early ontogeny. J. Comp. Neurol. 199: 233-253. Spector, S., Gordon, R., Sjoerdsma, A. and Udenfriend, S. (1967) End-product inhibition of tyrosine hydroxylase as a possible mechanism for regulation of norepinephrine synthesis. Mol. Pharmacol. 23: 549-555. StuU, N.D. and lacovitti, L. (1996) Acidic fibroblast growth factor and catecholamines (CA) together up-regulate tyrosine hydroxylase activity in developing and/or damaged dopamine neurons. /. Neurochem. 69: 1519-1525. Teitelman, G., Joh, T.H., Grayson, L., Reis, D.J. and lacovitti, L. (1985) Cholinergic neurons of the chick ciliary gangha express adrenergic traits in vivo and in vitro. J. Neurosci. 5: 29-39. Van Vliet, B.J., Ruuls, S.R., Drukarch, B., Mulder, A.H. and Schloffelmeer, A.N.M. (1991) ^-adrenoceptor-sensitive adenylate cyclase is inhibited by activation of ^-opiod receptors in the rat striatum. Eur. J. Pharmacol. 195: 295300. Weiner, N., Cloutier, G., Bjur, R. and Pfeffer, R.I. (1972) Modification of norepinephrine synthesis in intact tissue by drugs and during short-term adrenergic nerve stimulation. Pharmacol. Rev. 24: 203-221.

264 Wilcox, B. and Unnerstall, J. (1991) Expression of acidic fibroblast growth factor mRNA in the developing and adult rat brain. Neuron 6: 397-409. Weiss, S., Kemp, D.E., Lenox, R.H. and Ellis, J. (1987) Alpha2-adrenergic receptors mediate inhibition of cyclic AMP production in neurons in primary culture. Brain Res. 414: 390-394. Wollinksy, E. and Patterson, P.H. (1983) Tyrosine hydroxylase activity decreases with induction of cholinergic properties in cultured sympathetic neurons. J. Neurosci. 7: 1495-1500. Wong, V. and Kessler, J.A. (1987) Solubilization of a membrane factor that stimulates levels of substance P and choline acetyltransferase in sympathetic neurons. Proc. Natl. Acad. Sci. USA 84: 8726-8729. Yamamori, T., Fukada, K., Aebersold, R., Korsching, S., Fann, M.J. and Patterson, P.H. (1989) The cholinergic neuronal differentiation factor from heart cells is identical to leukemia inhibitory factor. Science 246: 1412-1416. Yoon, S.O. and Chikaraishi, D.M. (1992) Tissue-specific

Centrally-active differentiation factors in the nervous system transcription of the rat tyrosine hydroxylase gene requires synergy between and AP-1 motif and an overlapping E boxcontaining dyad. Neuron 9: 55-67. Zhou, J., Bradford, H.F. and Stern, G.M. (1994) The stimulatory effect of brain-derived neurotrophic factor on dopaminergic phenotype expression of embryonic rat cortical neurons in vitro. Dev. Brain Res. 81: 318-324. Zigmond, R.E., Schwarzschild, M.A. and Rittenhouse, A.R. (1989) Acute regulation of tyrosine hydroxylase by nerve activity and by neurotransmitters via phosphorylation. Annu. Rev. Neurosci. 12: 415-461. Ziller, C, Smith, J., Fauquet, M. and Le Douarin, N.M. (1979) Environmentally directed nerve cell differentiation: in vivo and in vitro studies. Prog. Brain Res. 51: 60-74. Ziller, C, Fauquet, M., Kalcheim, C, Smith, J. and Le Douarin, N.M. (1987) Cell lineages in peripheral nervous system ontogeny: medium-induced modulation of neuronal phenotypic expression in neural crest cell cultures. Dev. Biol. 120: 101-111.

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 11

Leukemia inhibitory factor and phenotypic specialization Tetsuo Yamamori National Institute for Basic Biology, 38 Nishigonaka, Myodaijicho, Okazaki, 444, Japan

1. Introduction

2. LIF as a pleiotropic factor in culture systems

LIF (leukemia inhibitory factor) is a cytokine which inhibits the growth of mouse leukemia cells (Ml cells) and induces their differentiation into macrophages (Gearing et al., 1987; Hilton et al., 1988). It was originally thought that LIF only affected hematopoietic development. However, soon after cDNA cloning, LIF was found to exhibit remarkably diverse biological activities. In particular, it was unexpected that LIF proved to be the active component underlying activity of the following two factors: (1) cholinergic differentiation factor (CDF). This was demonstrated originally to be secreted by cultured heart cells and to induce a change in the phenotype of cultured sympathetic neurons from adrenergic to cholinergic (see Yamamori et al., 1989). (2) Differentiation-inhibitory activity (DIA)/differentiation-retarding factor (DRF). This prevents embryonic carcinoma (EC) cells or embryonic stem (ES) cells from differentiating into various cell types (Smith et al., 1988; Williams etal., 1988). The pleiotropy of LIF raises two questions. First, why does LIF possess such a variety of activities in a wide range of cell types? Second, how is its specificity controlled in vivo ? This review describes recent progress in research on LIF, emphasizing the above two questions. I will first discuss the pleiotropic effects exerted by LIF in vitro. Second, I will discuss the structures of LIF and its receptors. Third, the interaction between LIF and its receptors and the consequent intracellular signaling pathways will be described. Finally, I will discuss the physiological functions of LIF in vivo.

2.1. D (differentiation) factor/LIF In the late 1960s, Ichikawa established a spontaneous myeloid leukemia cell line derived from a highly leukemia-prone strain of mice (SL strain). These cells can be induced to differentiate into cytologically defined macrophages by adding a conditioned medium (CM) obtained from mouse secondary embryos, adult spleen cells, peritoneal macrophages or rat granulocytes (Ichikawa, 1969). The differentiation-inducing activity in CM was purified and found to consist of two different activities, eventually identified as D (differentiation) factor/LIF and IL (interleukin)-6, depending on the subtype of Ml cells used for the assays (Tomida et al., 1984; Gearing et al., 1987; Abe et al, 1989; Gough and Williams, 1989; Sachs, 1990). In the 1980s, three groups developed methods to purify the factors showing differentiationinducing activities from CM. Hozumi's group established a M1-T22 cell line by subcloning Ml cells. Using this cell line for assay, they purified a component which they named D (differentiation) factor (Tomida et al., 1984) from CM of a mouse L cell line. On the other hand, Metcalf's group performed their purification using CM obtained from Krebs II ascites cells as the starting material. Using the M1-T22 cell line for the assay, they purified a factor which they named LIF (Gearing et al., 1987; Hilton et al., 1988). Sachs's group used the same starting material as Metcalf's group; however, using a different subclone of M1-T22,

266

cell line 11, for assays, they purified a factor designated as MGI (macrophage-granulocyte inducer)-2A (Shabo et al., 1988). The cDNA of LIF was isolated and sequenced by Gearing et al. (1987). This success in molecular cloning enabled the differentiation of previously reported similar activities. D factor and LIF were found to be identical by comparison of the amino acid sequence of a peptide fragment of D factor with that deduced from the nucleotide sequences of LIF, whereas MGI-2A was found to be identical to IL-6 based on immunoreactivity and amino acid sequences (Shabo et al., 1988; Sachs, 1990). Nucleotide sequencing of LIF further revealed its pleiotropic nature, which was attributed to different factors (Gough and Williams, 1989; Gough et al., 1989; Kurzrock et al., 1991). Whether LIF is identical to the previously termed human leukocyte inhibitory factor (also designated as 'LIF') remains to be determined, although their isoelectric points, molecular weights and Concanavalin A inducibilities are very similar (Rocklin, 1975; Hilton et al., 1988). In addition, human 'leukocyte inhibitory factor' has been reported to stimulate neutrophil-endothelial cell adhesion, suggesting that it plays a role in proinflammation (Schainberg et al., 1988). In this regard the activity of human 'leukocyte inhibitory factor' also resembles LIF. 2.2. CDF (cholinergic differentiation factor) The phenotype of postmitotic neurons is influenced by 'environmental factors' produced by surrounding cells. One of the best-characterized factors in this regard is CDF (cholinergic differentiation factor). CDF is secreted by cultured heart cells and induces a change in the phenotype of cultured sympathetic neurons from adrenergic to cholinergic. This activity has been studied extensively by Patterson and his colleagues since the early 1970s (Patterson and Chun, 1974; Patterson and Chun, 1977; Patterson, 1978). The purification of CDF was very difficult, due to the requirement of a long assay time and to the presence of only small amounts of CDF in CM. However, CDF was finally purified by 10^-fold to homogeneity from

Leukemia inhibitory factor and phenotypic specialization

crude CM of cultured heart cells in 1985 (Weber, 1981; Fukada, 1985). The amino acid sequence of a peptide fragment of CDF revealed its very close relationship to LIF. Although there was some discrepancy between the nucleotide sequence of the rat LIF cDNA and the amino acid sequence of the rat CDF, this was most likely caused by technical difficulty due to limited amounts of CDF available. Indeed, recombinant LIF expressed in bacteria (Escherichia coli) showed the same activity as CDF (Yamamori et al., 1989). Therefore, we now designate the factor with the cholinergic differentiation activity as CDF/LIF. 2.3. DIA (differentiation inhibitory activity)/DRF (differentiation retarding factor) Another aspect of LIF was revealed by studies on the activity that prevents differentiation of totipotent ES cells derived from normal blastocysts (Williams et al., 1988; Smith et al., 1988; Gough et al, 1989; Smith et al, 1992). It is known that to maintain the totipotency of ES cells, they must be grown on a feeder layer of fibroblasts. The activity secreted by the feeder layer fibroblasts, termed DIA (differentiation inhibitory activity) or DRF (differentiation retarding factor), has been purified, and its characteristics suggest that DIA/DRF may be identical to LIF. Indeed, not only does LIF exhibit the same activities as DIA and DRF in vitro, but also ES cells cultured with LIF can retain their totipotency. When the ES cells treated with LIF are injected into mouse blastocysts and then introduced into pseudo-pregnant females, they contribute extensively to the development of all of the somatic tissues (Williams et al., 1988; Gough et al., 1989). 2.4. Regulation of bone cell function Bone metaboHsm is regulated by local growth factors and by cytokines (Canalis et al., 1989). D factor (LIF) has been found to stimulate bone resorption (Abe et al., 1989) and regulate bone metabolism (Reid et al., 1990). Furthermore, both recombinant human (rh) and recombinant mouse

T. Yamamori

(rm) LIFs stimulate the release of "^^Ca from prelabeled calvaria in a dose-dependent manner. In a murine osteoblast-like cell line (MC3T3E1), rhLIF suppresses cell proliferation and enhances expression of osteopontin mRNA, while levels of alkaline phosphatase activity and type-I collagen mRNA decrease (Noda et al, 1990). Moreover, it has been reported that rh LIP is mitogenic for human bone-derived osteoblast-like cells (Evans et al., 1994), probably by acting on the osteoprogenitor cells (Evans et al., 1994). Details of the role of LIP in bone marrow stromal cell differentiation will be discussed later. 2.5. Other pleiotropic effects ofLIF After recombinant LIP became available, reports on the pleiotropic effects of LIP increased in number and, within a surprisingly short period of time, a variety of activities in vitro were attributed to LIP (see also reviews of Burgess, 1989; Kurzrock et al., 1991; Metcalf, 1991; Metcalf 1992). In particular, LIP appears to be identical with HSP-III (hepatocyte-stimulating factor III), which is known to stimulate the synthesis of acute phase plasma proteins (Baumann and Wong, 1989), and with MLPLI (melanoma-derived lipoprotein lipase-inhibitor) which is produced in a human melanoma cell line, SEKI. LIP inhibits lipoprotein lipase activity in adipocytes and possibly causes cachexia (Mori et al., 1989). Although some researchers are still cautious to accept some of these activities as in vivo functions, a substantial amount of data has now become available to enable evaluation of recent findings. To better understand the pleiotropic nature of LIP, structural analysis of LIP receptors and studies of its signal transduction are essential. 3. Structure of LIF

3.1. Primary structure of LIF The active agent inducing differentiation of Ml cells into macrophages was purified as a protein

267

from CM of Krebs II ascites cells, and the amino acid sequences of 12 peptides derived from the purified protein were determined. Using oligonucleotide probes designed from the partial amino acid sequences. Gearing et al. (1987) screened a library of cDNA clones constructed from the mRNA isolated from Conconavalin A-stimulated LB3 cells (a T cell line), and determined the complete primary structure of LIP cDNA. The primary structure of CDP was determined, and shown to be identical to that of LIP (Yamamori et al., 1989). The original report by Gearing et al. (1987) implied that the N-terminal amino acid residue is proline. However, later reports suggested the presence of a serine residue at the N-terminal (Gascan et al., 1989). The amino acid sequence homology between murine and rat LIPs is 92%, and that between rat and human LIPs is 82%. The LIP signal peptide is composed of 23 amino acid residues in mice and 24 in both rats and humans (Yamamori et al., 1989). Mouse and human LIP genes are composed of three exons (Gough et al., 1988; Stahl et al., 1990). The human LIP gene has been mapped to 22ql2 or 22qll.2- > qlB.l, distal to the Ewing sarcoma break point on 22q (Budarf et al., 1989; Sutherland et al., 1989) and suggested to be localized near a putative tumor suppressor locus associated with meningiomas (Pergolizzi and Erster, 1994). Moreover, the genes for oncostatin M (OSM) and LIP are also tightly linked (Rose et al., 1993). The murine LIP gene has been reported to be located on proximal chromosome 11 (Bottorff and Stone, 1992). Another form of LIP associated with the extra cellular matrix (M-LIP) arises from alternative splicing within exon 1 (Rathjen et al, 1990b). In the 5'-flanking region of the murine LIP gene three sequence elements have been found to control the constitutive action of the LIP promoter. The region essential for transcription contains the major startsite of transcription (+1), a TATA-box (-31) and 72 additional base pairs from -32 to -103, while a negative regulatory element has been identified between positions -360 and -249. An additional positive control element appears to be located between positions -860 and -661 (Stahl and Gough, 1993).

268

3.2. Secondary and tertiary structure ofCDF/LIF Nucleotide sequences does not immediately reveal the level of homology of various factors to LIF. However, similarity of effects of IL-6 and LIF on myeloid leukemia cell lines suggest that they may share the same receptors and the same signal transduction pathways (Gough and Williams, 1989; Gough et al, 1989). The similar activities of CNTF (ciliary neurotrophic factor) and CDF/LIF in induction of neurotransmitter changes in sympathetic neurons supports this hypothesis (Saadat et al, 1989; Yamamori et al.,1989; Yamamori, 1992). Bazan (1991) was the first to point out that CDF/LIF has a similar structure to those of CNTF, GCSF, IL-6, OSM and growth hormone (GRH), which contain four a-helices (Abdel-Meguid et al, 1987). Rose and Bruce (1991) also pointed out similarities among these cytokines. In addition to iH NMR studies (Smith et al., 1994), X-ray crystallography at 2.0 A resolution (Robinson et al, 1994) confirmed the four a-helical structure of LIF. Two regions located in the fourth helix and the preceding loop are suggested to be important in the interaction of LIF with its receptors (Robinson etal., 1994).

Leukemia inhibitory factor and phenotypic specialization

different IL-6 receptor subunits (Taga et al, 1989; Hibi et al., 1990). One of these subunits was designated gp80 (glycosylated protein of molecular weight 80kDa) and the other as gpl30. gp80 functions as a binding site for IL-6 whereas gpl30 transfers into the cytoplasm the signal generated by the interaction between IL-6 and its receptors (Stahl and Yancopoulos, 1993; Kishimoto et al., 1994; Rose-John and Heinrich, 1994; Zhang et al., 1994). Based on nucleotide sequence analyses of the cDNAs of LIF and CNTF receptors, two groups have reported that both receptors belong to the IL6 receptor family (Davis et al., 1991; Gearing et al., 1991; Hall and Rao, 1992; Patterson, 1992). It soon became clear that six subunits of gp80, gpl30, NKSFp40 (natural killer cell stimulatory factor), IL-6-R and GCSF-R together form an IL-6 receptor family (Fig. 1). One common feature of the cytokine receptor family, including the IL-6 family, is that its members contain immunoglobulin (Ig)-like domains, of which the structures are similar to that of Ig, in the extracellular site of each subunit (Bazan 1991). Near the C-terminal end of the second domain ( Cterminal domain), there is a cytokine receptor family consensus sequence, W(Trp)-S(Ser)-X(a variable residue)-W-S.

4. Receptors for LIF: IL-6 receptor family 4.2. Receptor interaction Nucleotide sequence analysis of cloned receptors for the above-described cytokines revealed that they belong to the same family, namely the IL-6 receptor family (Yamasaki et al., 1988; Taga et al., 1989; Davis et al., 1991; Gearing et al., 1991). 4.1. IL-6 receptor family Kishimoto and his colleagues have identified a soluble factor, i.e. the B cell-stimulating factor, which stimulates proliferation of B cells. Later, this factor was designated IL-6 (Kishimoto, 1989). Molecular cloning and nucleotide sequencing of IL-6 were completed and recombinant IL-6 became available. This enabled cloning of the receptors for IL-6 (see reviews by Kishimoto, 1989; Taga and Kishimoto, 1992), and identified two

The cytokine receptor subunits can be classified into at least two groups. The a-type subunits are important for binding of the ligands of the cytokines, whereas the )8-type subunits are important for transduction of extracellular signals to the cytoplasmic site. The a-type subunits include NKSFp40, CNTF-R and IL-6-R, whereas the /8type subunits include gpl30, GCSF-R and LIF-R. It should be noted that LIF-R also functions as a binding site for LIF (Gearing et al., 1991, Gearing et al., 1992), and that, therefore, it cannot be simply classified as a^S-type subunit. The mode of interaction between different members of the IL-6 family and their receptors has been investigated (Taga and Kishimoto, 1992; Stahl and Yancopulous, 1993; see Fig. 2). The IL-

T. Yamamori

269

Extracellular Domains

linker (Ig-like)

(IL-1 2 soluble form) Ig-like

cytokine receptor domain

Cytoplasmic Domains

gp130

GCSF-R

LIF-R

Fig. 1. Structures of members of the IL-6 receptor family are schematically shown. Modified from Yamamori and Sarai (1994) with addition of recent findings obtained by Hilton et al. (1994) and Stahl et al. (1995).

6 receptor complex consists of one gp80 and two gpl30s as described above. The LIF receptor complex consists of LIF-R and gpl30, whereas the CNTF receptor complex consists of the heterotrimer of gpl30, LIF-R and CNTF-R(a) (Ip et al., 1992; Davis et al., 1993; Stahl and Yancopoulous,

1993; Stahl and Yancopoulos, 1994). The complex of gpl30 and LIF-R also serves as the active receptor complex for OSM (Gearing and Bruce, 1992; Gearing et al, 1992; Liu et al., 1992). Recently, a new member of the IL-6 family, a binding subunit (a-chain) of the murine IL-11 receptor, has

Leukemia inhibitory factor and phenotypic specialization

270

CDF/LIF, OSM

IL-6,IL-11?

CNTF, OSM?

LIF-R CNTF-R(a) or OSM-R?

gp80orIL-llR?

gpl30 gpl30

gpl30

gpl30

Membrane Fig. 2. The interaction between ligands (cytokines) and receptors of the IL-6 family is schematically shown. Only cytokine domains and fibronectin type III domains are shown. References: Taga and Kishimoto (1992); Stahl and Yancopoulos (1993).

been cloned (Hilton et al., 1994). In order to form a functional receptor complex, IL-11 requires both gpl30 and its unique a-chain of an IL-11 binding subunit (Yang, 1993; Hilton et al., 1994; Yin et al, 1994). In addition to the membrane-integrated form, cDNAs of LIF-R which do not encode transmembrane or cytoplasmic domains have been identified. These mRNA species may encode soluble forms of LIF-R (Tomida et al., 1994). A soluble form of LIF-R is produced in mouse liver during pregnancy (Tomida et al., 1993), and also found in normal mouse serum (Layton et al., 1992; RoseJohn and Heinrich, 1994). LIF and OSM seem to directly interact with a soluble form of gpl30 whereas IL-6 does not (Modrell et al., 1994). Although the physiological significance of these soluble receptors remains largely unknown, it has been found that soluble IL-6-R triggers osteoclast formation in vitro , possibly via binding with IL-6 (Tamuraetal., 1993). The scheme described above explains the essential mode of interaction between the members

of the IL-6 family and their receptors. However, there are still unsolved issues. Firstly, we may not yet have identified all members of the IL-6 family. At present, six members of the family are known (IL-6, CDF/LIF, OSM, CNTF, IL-11, GCSF) are known, and IL-12 (NKSF; complex of p35 and p40) is likely to be a seventh member. As to the receptors, six members (IL-6-R, CNTF-R, gpl30, GCSF-R, LIF-R, NRI/IL-11-R) are thus far known. Are there more members of the family? At least one other binding (OSM-R) subunit appears to exist although it has not yet been isolated (Baumann et al., 1993; Fourcin et al., 1994). LIFR and CNTF-R can form high-affinity receptor complexes with CNTF; however, these complexes generate only inefficient signal transduction (Gearing et al., 1994). Since CNTF exhibits activity on the cells that do not express any significant level of CNTF-a, this raises a possibility that there may exist another binding subunit for CNTF besides CNTF-a (Baumann et al., 1993). Second, the current scheme may be oversimplified. As pointed out by Thoma et al. (1994), if all

T. Yamamori

possible combinations of three receptor subunits differ in specificity and affinity, complexity of the receptor complex would be much higher than discussed above. In fact, CNTF, LIF or OSM does not exhibit exactly the same biological activities in the same cells, even though all sets of receptor subunits (gpl30, LIF-R and CNTF-a) are expressed (Piquet-Pellorce et al., 1994). Third, while human (h) LIF competes with murine (m) LIF for binding to hLIF receptors, mLIF does not efficiently compete with hLIF for binding to hLIF receptors (Layton et al., 1994b). Further analyses suggested that mLIF may bind to only one site of both mLIF-R and hLIF-R, whereas hLIF binds to two sites of mLIF-R (Layton et al. 1994a,b). The above specificity of mLIF can be converted to that of hLIF by replacement of six amino acid residues (Layton et al., 1994b). The biological significance of this difference in the binding sites between hLIF and mLIF is not yet understood. 4.3. Signal transduction 4.3.1. Cytoplasmic consensus sequences of signal transduction The extracellular sites of receptors are important in determining the binding specificity of ligands, while the cytoplasmic sites play essential roles in signal transduction. Dimerization of two signal transduction subunits, either heterodimeric or homodimeric, is necessary for signal transduction. Two gpl30 subunits or two GCSF-R subunits form homodimers, whereas gpl30 and LIF-R form heterodimers. There are consensus amino acid sequences in the cytoplasmic region of cytokine receptors. A critical cytoplasmic region of 60 amino acids is found proximal to the transmembrane domain (Murakami et al., 1991). Within this region there are two consensus sequences termed 'box-T and 'box-2\ whose sequences are conserved among signal transducers of the cytokine receptor family. The importance of the cytoplasmic region near the transmembrane domain has been shown in studies on signal transduction subunits of IL-2-R, GCSFR and EPO (erythropoietin)-R (Hatakeyama et al.,

271

1989; D'Andrea et al., 1991; Fukunaga et al., 1991). 4.3.2. Kinases involved in the signalling pathway Study of the kinases which can be activated by the IL-6 family suggested that their signal transduction system may differ from those of growth factors. It also suggested that some kind of tyrosine phosphorylation is involved in the process (Hoffman-Liebermann and Liebermann, 1991a; Lord et al., 1991; Michishita et al., 1991; Murakami etal., 1991). A new type of tyrosine kinase named Janus kinase (Jak) appears to be involved in the signal transduction process of IL-3 (Wilks, 1989; Firmbach-Kraft et al, 1990; Harpur et al., 1992). One of the Jak kinases, Jak2, is shown to be specifically phosphorylated via activation of EPO-R and to consequently bind to EPO-R (Miura et al., 1991; Witthuhn et al., 1994). In addition, Jak2 activation is induced via receptors for growth hormone and IL-3 (Miura et al., 1991; Harpur et al., 1992; Velazquez et al., 1992; Argestsinger et al., 1993; Silvennoinen et al., 1993; Witthuhn et al, 1993). Activation of Jakl/Jak2 is induced by IFN-y, whereas that of Jakl/Tyk2 is induced by IFN-a, ^ (Velazquez et al., 1992; Muller et al., 1993). Jak3 is involved in signaling via interleukin 2 and 4 in lymphoid and myeloid cells (Witthuhn et al., 1994). Finally, three members of the Jak-Tyk family (Jakl, Jak2 and Tyk2) are activated by the IL-6 family (Bonni et al, 1993; Lutticken et al., 1993; Stahl et al., 1994; Yin et al., 1994). Phosphorylation and internalization of gpl30 appear to occur after activation of Jak2 kinase by IL-6 in hepatocytes (Wang and Fuller, 1994). Association of gpl30 with a transcription factor, acute-phase response factor (APRF), or a protein immunologically related to the p91 subunit of the interferon-stimulated gene factor-32 (ISGF-32) has also been reported (Lutticken et al., 1993). Besides Statl (p91), other members of the Stat family, Stat2, Stat3 and Stat4, have been identified (Darnell et al., 1994; Zhong et al., 1994b). Members of the IL-6 family preferentially activate Stat3 (Wegenka et al., 1993; Wegenka et al., 1994; Zhong et al., 1994b), whereas IFN-y activates

272

Statl and IFN-a activates Statl, Stat2 and Stat3 (Schindler et al, 1992; Shuai et al, 1992; Shuai et al., 1993; Darnell et al, 1994; Stahl et al, 1995). Stahl et al. (1995) proposed a sequence of phosphorylation events, including activation of StatB, in response to its interaction with the members of the IL-6 receptor family. According to their model, activation of the cytoplasmic pathway involves the following four steps: (1) dimerization of the receptors activates phosphorylation of Jak kinases, (2) phosphorylated Jak kinases phosphorylate Stat3 and a tyrosine residue of a YXXQ consensus sequence in the box-3 region identified by Baumann

Fig. B. The signal pathway of the cells treated by the IL-6 family including LIF is schematically shown.

Leukemia inhibitory factor and phenotypic

specialization

TABLE 1 Jaks and Stats induced by cytokines IFN-a^ IFN-y

Epo IL-3 IL-6* CNTF

LIP OSM^

Jakl, Tyk2 Jakl, Jak2 Jak2 Jak2

Stats

Jakl, Jak2, Tyk2 Jakl, Jak2, Tyk2 Jakl,Jak2,Tyk2 Jakl,Jak2,Tyk2

Stat3 StatB StatB StatB

Statl, Stat2,Stat3 Statl

?

^The degree of phosphorylation of each JAK-TYK kinases is reported to vary among cell lines. References: Darnell et al. (1994), Zhong et al. (1994a,b), Mui et al. (1994), Schinlder et al.(1994), Stahl etal. (1995).

et al. (1994), (3) the phosphorylated sites of the YXXQ sequence and phosphorylated Stat3 proteins bind to each other and (4) subsequently, StatB proteins dissociate, dimerize and translocate into the nucleus (Stahl et al., 1995). In ES cells, functional and biochemical association with gpl30 of Hck, one of the Srk family of protein tyrosine kinases (Ernst et al, 1994) has been suggested to occur, although its exact role in tyrosine phosphorylation remains unknown. The overall picture of how these tyrosine kinases interact with each other has not been delineated. Analogous to the interaction of members of the IL-6 receptor family with those of the IL-6 family, specific interaction of several kinases might determine the specificity of the signaling (see Table 1). Although the signal pathways of members of the IL-6 receptor family are very similar, there may be also some differences (Piquet-Pellorce et al., 1994; Thoma et al., 1994). Fifteen different proteins have been found to be tyrosine phosphorylated upon CNTF treatment (Boulton et al., 1994). Although five of them are unidentified, 10 are known proteins, i.e. LIF-R, gpl30, PLCy, PP120 src substrate, P110-PI3kinase, STAT91 (91 kDa component of the ISGF3 complex), PTPID (a tyrosine phosphatase), SHC, and ERKl/2 (MAP kinases), suggesting sequential activation of phospholipase C, IP3 and MAP kinases.

273

T. Yamamori

There might be another pathway distinct from that mediated by Stat3. Such a pathway may be responsible for IL-6-mediated cell proliferation (Stahl et al., 1995), because a truncated gpl30, the amino acid sequence of which possesses box-1 and box-2 but not the YXXQ sequence, still can induce cell proliferation (Hatakeyama et al., 1989; Stahl et al, 1995). Following treatment of cells with LIF or CNTF, both PKC-dependent and PKC-P21^^^independent pathways seem to be activated (Johnson and Nathanson, 1994; Schiemann and Nathanson, 1994; Schwarzschild et al., 1994). Interestingly, the signal transduction pathway whose activation induces neurotransmitter and neuropeptide synthesis in a neuroblastoma cell line is suggested to be PKC-P21'•^^-independent (Johnson and Nathanson, 1994). However, there are reports which do not agree wholly with these findings and which suggest the involvement of a PKC-'^^^dependent pathway activated by LIF and CNTF in neuroblastoma and cultured sympathetic neurons (Lewis et al, 1993; Schwarzschild et al., 1994). A common signal transduction pathway may be activated by a variety of cytokines and growth factors, including the IL-6 family, IFN-a, -yS, -y, IL-3, IL-5, IL-10, GM-CSF, PDGF and EGF, EPO and growth hormone (Fu and Zhang 1993; Lutticken et al., 1993; Muller et al., 1993; Larner et al., 1993; Ruff-Jamison et al., 1993; Silvennoinen et al., 1993; Stahl and Yancopoulos, 1993; Hirano etal., 1994). S6 protein kinase (PP90rsk; Chen and Blenis, 1990), a possible target of H7 (Ser/Thr kinase inhibitor), is reported to be activated by the IL-6 receptor family (Yin and Yang, 1994). H7 inhibits the kinase activity of PP90rsk but not that of MAPK in vitro. Since H7 inhibits the activation of primary response genes, PP90rsk is possibly involved in the transcriptional activation. This S6mediated pathway activated by cytokines may eventually be transferred into the nucleus to activate the CRE sequence, which binds to proteins of the CREB/ATF family. The second regulatory sequence known to control the transcription of the genes responding to IL6 is JEBS (a putative ETS binding site), or named differently (Fig. 4), which binds to a nuclear fac-

tor, APRF/Stat3, or to very similar factors (Hirano et al., 1994). The Stat protein complex also binds to the JEBS sequence (Hirano et al., 1994). Phosphorylation of Stat proteins and APRF has been reported to be induced by IL-6 and CNTF (Lutticken et al., 1993; Sadowski et al., 1993). Stat proteins are phosphorylated variably by IFN-y, IL6 or CNTF. For example, Stat p91 was reported to be activated by IFN-y and CNTF but not by IL-6 (Lutticken et al., 1993; Sadowski et al., 1993). However, cross-immunoreactivity among the Stat family members may obscure the species of Stats involved in each signal transduction. All reports do not agree at the moment in this regard. The third type of factor is NF (nuclear factor)IL6 (Akira et al., 1990) which binds to a consensus sequence of T(T/G)NNGNAAT. The NF-IL6 gene is induced by IL-6 (Baumann et al., 1992; Baumann et al., 1993; Yuan et al., 1994) and, in turn, NF-IL6 proteins enhance expression of the genes containing the consensus sequence. 4.33. Genes regulated by the IL-6 family In Ml myeloblastic leukemia cells, the addition of either LIF or IL-6 induces the immediate early response genes jun-B, c-jun, jun-D, ICAM and MyDWG (myeloid differentiation primary response). On the other hand, My788 mRNA in Ml

T C C !|T T CT| I G G A A fr T C — I

rat a2M, IL-6RE

CAGj T A A QT G G A A lAGT —^ G G -TT T C

qG G A A l A G C

hICAM-l,pIRE

C A T | T T CG|

dG G A A A T C

hIRF-l,IFN-7RE

G A TIT T A

4 G T A A '^ ^^

hGBP, GAS

qCGAA^

CNTF-RE (core)

T T C

T T N CNNNAA AGTTTCNNTTTCNCA'

IFN-y -activated site (consensus) IFN-a-RE (consensus)

Fig. 4. Sequence comparison of IL-6, IFN-y, CNTFresponsive elements. Modified from Caldenhoven et al. (1994) and Bonni et al. (1993). Abbreviations: rat a2M, the rat a2 macroglobulin; hICAM-I, human ICAM-1; hIRF-1, human interferon regulatory factor-1; hGBP, human GBP gene; CNTF-RE, CNTF-responsive elements and IFN-yactivated site are also shown (Darnell et al, 1994).

274

cells is stabilized by the addition of LIF or IL-6 (Lord et al., 1991). In sympathetic neurons, c-fos and jun-B are induced within 30 min upon the addition of CDF/LIF (Yamamori, 1991a). jun-B Induction by IL-6 has also been observed in other cell types (Oritani et al., 1992; Nakajima et al., 1993; Hirano et al., 1994; Kahn and De Vellis, 1994). There are no distinguishable differences in the level of the induction of immediate early response genes in Ml cells treated with LIF and with IL-6 (Lord et al., 1991). However, differential regulation of the SCL (stem-cell/leukemia) gene by IL-6, LIF and OSM has been observed during myeloid leukemia cell differentiation. Numbers and affinity of the receptors for LIF, OSM and IL-6 in Ml cells are comparable. Thus, the differential regulation in SCL gene expression observed several days after the addition of the members of the IL-6 family may be due to differences in the signaling pathways of the receptors, the functions of which are apparently indistinguishable in immediate early gene responses (Tanigawa et al., 1993). In Ml cells, LIF suppresses c-myc and c-myb expression, which is tightly linked to the terminal differentiation of the cells that are induced by LIF or IL-6 (Hoffman-Liebermann and Liebermann, 1991b) in that the constitutive c-myc expression blocks the terminal differentiation of myeloid leukemia cells. jun-B is induced within 2 h after the addition of LIF or IL-6, whereas c-myb and c-myc are induced at 2 and 12 h, respectively (HoffmanLiebermann and Liebermann, 1991a; Kasukabe et al., 1994). Some genes which show late expression after addition of CDF/LIF have been identified. For example, in cultured sympathetic neurons CDF/LIF induces transcription of not only the acetylcholine gene, but also the genes coding for substance P, somatostatin, cholecystokinin, enkephalin and vasoactive intestinal polypeptide (Nawa and Patterson, 1990; Fann and Patterson, 1994). The transcription of muscarinic acetylcholine receptor genes and substance P receptor genes is also induced (Ludlam et al, 1994). On the other hand, CDF/LIF suppresses catecholamine synthesis. In osteoblast-like cells, LIF suppresses alkaline

Leukemia inhibitory factor and phenotypic specialization

phosphatase and collagen type I synthesis, whereas it stimulates transcription of osteopontin. 5. Roles of CDF/LIF in vivo The LIF receptor complex consists of gpl30 and LIF-R. Since the heterodimer complex of gpl30 and LIF-R functions as a receptor for OSM as well, the two factors exhibit in general very similar activities in vitro. If a cell also expresses the CNTF-R(a) receptor, CNTF should induce similar effects to those induced by LIF and OSM. Thus, the activities in vitro exhibited by the members of the IL-6 family overlap. In fact, this is one of the reasons why we see such multiple in vitro activities of LIF, as described in Section 2. In order to determine the precise activities of CDF/LIF in vivo, it is essential to satisfy the following criteria. (1) Sufficient amounts of a factor and its receptors should be present in the tissue. (2) The administration of the factor should have a specific effect on a specific type of cells in the tissue. (3) Depletion of the factor should cause a specific effect. The above criteria are similar to those designated for studies on target-derived neurotrophic factor (Purves, 1988). However, the redundancy of the activities of the members of the IL-6 family makes the situation more complex. To overcome this problem, two approaches have been taken by using gene manipulation in mice. One approach involves the overexpression of LIF (Metcalf and Gearing, 1989). Cells of a murine hematopoietic cell line which are multiply infected with a retroviral construct containing LIF cDNA, secrete high levels of LIF. Injection of these cells into mice causes them to exhibit 'fatal syndromes' within 12-70 days of injection. Irradiated animals injected with FD/LIF develop these fatal syndromes as early as 10 days after injection, whereas non-irradiated animals injected with FD/LIF develop the syndromes only about 30 or more days after injection. In both cases, the increased level of LIF causes serious abnormalities (Table 2). It is difficult to evaluate whether the abnormalities observed in these 'fatal syndromes' are due to

T. Yamamori

275

TABLE 2 Effects of overexpression Tissue lesion ,

Control

Excess LIF

Bone Excess osteoblasts Excess bone formation Bone resorption

-

+++ +++ ++

Calcification Heart Skeletal muscle

+

-

++^ +^

Liver Calcification Hemopoiesis Necrosis Fibrosis

-

+/-^ ++ +^ ++

Pancreas Edema Acinar degeneration Infiltration

-

++ ++ +^

Adrenal cortex Brown degeneration

-

-

Ovary^ Follicles Corpora lutea

4.0 ±3.1 9.6 ±4.7

7.2 ±5.2 5.1 ±4.8

The data are cited from Metcalf and Gearing (1989). LIFproducing cells were injected to mice (Excess LIF) and the degrees of symptoms were compared to those control animals. A group of irradiated mice were also injected with LIFproducing cells. -, no animals showed the specific phenotype indicated; +/-, 100% of the animals showed the phenotype. Total number of each group is about 20 (range 7-29). ^The number of mice that show the phenotype was lower in irradiated mice than in non-irradiated mice. ^Mean numbers ± standard deviation of ovarian follicles and corpora lutea in the organ section.

a primary or a secondary effect of LIF. Taking the in vitro data into account, however, most of these abnormalities seem to be directly caused by the overexpression of LIF. LIF is generally considered to be expressed at very low levels and is often expressed only at limited sites within a tissue (Murray et al., 1990; Yamamori, 1991b; Robertson et al., 1993). Since the mRNAs of gpl30 and LIF-

R are expressed in a wide variety of cells (Ip et al., 1993), the presence of high levels of LIF might be expected to generate strong signals in most cells. However, apparently not all cells respond. Metcalf and Gearing (1989) reported that the most consistent abnormalities observed in FD/LIF mice were an increase in number of osteoblasts in the bone marrow with excess new bone formation, often accompanied by increased bone resorption and calcium deposition in muscle tissue. The other tissues that exhibit abnormalities were those containing cells that respond to LIF in vitro. The second approach using genetic manipulation is to 'knock out' the LIF gene by genetargeting methods. At least two clear defects have been identified: (1) lif" female mice are infertile. In normal development, the expression of LIF in the uterine wall significantly increases upon blactocyst implantation (Croy et al., 1991). In lif- mice, absence of LIF causes failure of implantation. Blastocysts from lif- mice can be implanted into lif+ mice. Hence, failure of implantation appears to be due to a defect in the uterine wall. In lif" mice, the failure of implantation can be 'cured' by LIF injection (Stewart et al, 1992). (2) In lif- mice, the number of myeloid-producing stem cells in spleen and bone marrow is much lower than that in normal mice (Escary et al., 1993). Despite the above two clear examples, difficulties are encountered in the evaluating the exact roles of LIF in vivo. Metcalf (1993) pointed out that 'the ability of the deletion or inactivation approach to obtain answers to the redundancy problem may in fact be a more difficult task than was envisaged by the exponents of this approach. Deletion of a growth factor gene may well induce no obvious abnormality until a particular challenge situation is applied, when a deficiency then becomes evident. Worse, it may become necessary to perform the observation at a quite restricted time after deletion. It has recently been reported that the dramatic defects in op/op mice resolve spontaneously as the animals age. This suggests that, although compensating mechanisms may require some time to become operative, genuine defects may be missed if analysis is delayed too long'. Nonetheless, with these reservations, the last Sec-

276

tion of this chapter will attempt to clarify how pleiotropic activities of CDF/LIF are regulated in vivo, with emphasis on embryonic development, haematopoiesis, bone metabolism and the nervous system. 5.7. Embryonic system A fertilized egg develops into a hatched blastocyte in vitro when cultured in a protein-free medium at a slower growth rate than that observed in vivo. Growth factors and cytokines appear to accelerate the zygotic development in both autocrine and paracrine manners. CDF/LIF is one of the factors produced by preimplantation embryos (Conquet and Brulet, 1990; Murray et al., 1990). The addition of hLIF to day-5 parthenogenetic bovine morulae produced in vitro stimulates their further development to blastocysts (Fukui et al., 1994). So far, studies of this process remain descriptive (Adamson, 1993) and hence the role of CDF/LIF is not well understood (see reviews by Lee, 1992 and Adamson, 1993). CDF/LIF is also expressed in postimplantation egg cylinders (Conquet and Brulet, 1990; Rathjen et al., 1990a), in the pregnant uterus (Bhatt et al., 1991) and in the preimplantation uterus (Shen and Leder, 1992; Charnock-Jones et al, 1994). CDF/LIF mRNA levels in the uterus increase at the time of implantation in several mammalian species (Anegon et al., 1994; Charnock-Jones et al., 1994; Yang et al., 1994) and the infertility of female lif" mice can probably be attributed to the failure of implantation. Although the exact role of CDF/LIF in implantation remains unknown, one obvious possibility is an enhancement of adhesiveness of the preimplanted blastocyte to the uterus wall (see the articles of Cross et al., 1994; De Felici and Pesce, 1994; Edwards, 1994). 5.2. Hematopoietic system The overexpression of LIF induced by injection of FD/LIF cells causes the development of fatal syndromes in mice, as previously described at the beginning of Section 5. The direct injection of recombinant LIF into mice causes toxic effects as

Leukemia inhibitory factor and phenotypic specialization

well. LIF injection results in an increase in the number of bone marrow cells, particularly marrow lymphocytes and a moderate increase in spleen weight, with a reduction in number of spleen lymphocytes and an increase in the number of erythroid cells (Metcalf et al., 1990). Moreover, an increase in the number of megakaryocytes is observed in the marrow and spleen, with an associated 10-fold rise in megakaryocyte progenitor cells in the spleen (Metcalf et al., 1990). These effects observed with excess LIF in vivo are consistent with those observed in culture systems. However, most of these effects of LIF are unlikely to occur in normal animals in vivo, because the sites where LIF is expressed are generally very limited. Also, redundancy of effects among members of the IL-6 family may mask the specific effects of LIF in vivo (Metcalf, 1993). The most straightforward approach to clarify the role of the LIF gene in vivo is to knock it out. In lif' mice, a dramatic decrease in the number of several stem cell types was observed in the spleen (Escary et al., 1993). The numbers of progenitors in the spleen of heterozygous mice are intermediate between those in wild-type and homozygous mutant animals. The total nucleated cell numbers in the spleen and the bone marrow are normal in lif mice and the levels of circulating mature red and white blood cells and blood platelets are also normal in homozygous mutant mice. The stem cell population in lif mice can be restored to its normal level by adding exogenous LIF, suggesting that the ability of lif~ stem cells to differentiate is not affected but that the LIF-producing microenvironment has been changed. Apparently the lymphoid population is unaffected, although the thymic Tcell responsiveness is decreased, presumably due to the defective microenvironment (Escary et al., 1993). In the thymus of transgenic mice in which LIF is overexpressed in T cells, cortical CD4+CD8'*" lymphocytes are lost, while numerous B cell follicles develop (Shen et al, 1994). Moreover, peripheral lymph nodes contain a vastly expanded CD4+CD8+ lymphocyte population, and the thymic epithelium is profoundly disorganized. This is consistent with the idea that disruption of stroma-

277

T. Yamamori

lymphocyte interaction is responsible for many observed defects and for the apparent interconversion of cell populations between the thymus and the lymph node in the transgenic mice (Shen et al., 1994). In fact, LIF mRNA expression can be regulated in murine bone marrow stromal cells by IL-1, IL-4, TNF-a and the cAMP analogue 8bromoadenosine 3':5'-monophosphate (SBrcAMP) (Derigs and Boswell, 1993; Wetzler et al, 1994). Thus, in normal stroma-lymphocyte interaction, expression of LIF is likely to be maintained at appropriate levels via cytokine networks. It is beyond the scope of this article to present numerous hypotheses regarding more precise roles of LIF as a hematopoietic regulator. In brief, however, at least 6 factors are involved in the development of small maturing neutrophilic granulocytic colonies, and at least seven factors in formation of megakaryocyte colonies (Metcalf, 1993). Members of the IL-6 family are included in both these groups of cytokines. IL-6 may act particularly on mature megakaryocyte precursors (Ishibashi, 1989; Metcalf, 1993). Thus, even if LIF has a role in hematopoiesis, neither analyzing lif~ mice nor the effects of LIF overproduction would reveal the exact roles of LIF in these processes. If IL-6-R is expressed in addition to gpl30 and LIFR, the effects of IL-6 on normal haematopoiesis might explain the effects of LIF overexpression in the mouse. By contrast, even if LIF is indeed involved in hematopoiesis, the absence of LIF in lif mice may be compensated by IL-6, which is known to be involved in all steps of neutrophil and megakaryocyte production in vitro (Metcalf, 1993). However, it is worth noting that the apparent interconversion of cell populations between thymus and lymph nodes is a unique characteristic of LIF transgenic mice and is not observed in IL-6 transgenic mice (Suematsu et al., 1989, 1992; Shen et al., 1994). Combinations of more than two cytokines, growth factors or other compounds result in synergistic effects (Metcalf, 1993; Ip et al., 1994). In the presence of IL-3, for example, LIF, IL-6, or GCSF stimulates formation of very primitive blast colonies (Leary et al., 1990) or suppresses cell proliferation of human myeloid leukemia cells (Maekawa et al, 1990). One might

obtain informative results about more precise roles of each member of the IL-6 family if one could target each gene of the IL-6 receptor family in a tissue-specific manner (Gu et al., 1993, 1994). 5.3. Bone metabolism Of the total amount of calcium in the human body, most (about 1 kg) is stored in the bones, whereas only 1 g is found in the circulatory system. Cellular Ca"^"*" level is strictly controlled by hormones, cytokines and growth factors, and exerts an effect on bone formation and resorption by osteoblasts and osteoclasts, respectively (Raisz, 1988; Canalis et al., 1988, 1989; Suda et al., 1992). As Raisz noted in 1993, 'probably the most important advance in our understanding of bone biology during the last 20 years has been the identification of the role of cytokines in bone remodeling'. Indeed, a few cytokines, namely IL-1, TNF(2, TNF-)8 and colony stimulating factor (CSF), have been found to play clear roles in bone remodeling (Mundy, 1989; Canalis et al., 1991). Roles of LIF in bone formation and resorption have also been reported, although its effects are not as distinct as those of the above four cytokines. D-factor (LIF) exhibits bone resorption activity (Abe et al, 1989; Reid et al., 1990) and stimulates ^^Ca release from neonatal mouse calvaria in vitro and the synthesis of DNA and protein as well (Reid et al., 1990). On the other hand, Lowe et al. (1991) reported that LIF induced DNA synthesis and cell proliferation in isolated foetal rat osteoblasts, whereas it inhibited DNA synthesis and cell proliferation in an osteogenic sarcoma cell line, UMR-106. LIF also inhibits basal bone resorption in long-term cultures of rat fetal bone (Lorenzo et al., 1990). rhLIF is mitogenic for human bone-derived osteoblast-like cells in vitro (Evans et al., 1994). Following treatment with rhLIF, ^H-thymidine-labelled cell nuclei are colocalized with alkaline phosphatase (AP) activity, suggesting that the mitogenic effect of rhLIF occurs in cells of the osteogenic lineage. LIF has been reported to stimulate formation of osteoclast-like multinucleated cells when co-

278

cultured with mouse bone marrow cells (Abe et al., 1989). However, LIF was found to bind to osteoblasts, but not to multinucleated osteoclasts (Allan et al., 1990). The clonal MMR 106-06 rat osteogenic sarcoma cells also apparently express LIF receptors. In addition, both calvarial osteoblasts and osteoblast-like cells express low levels of LIF transcripts. These lines of evidence suggest that LIF acts directly on the osteoblast lineage including preosteoblastic cells (Rodan et al., 1990; Hakeda et al., 1991). In fact, the members of the IL-6 family whose receptors share gpl30 induce differentiation of BMS2 (the murine stromal cell line) cells into adipocytic and osteoblastic phenotypes (Gimble et al., 1994). 5.4. Nervous system 5.4.1. Phenotypic change of neurotransmitter by CDF/LIF CDF/LIF was originally described as a neurotransmitter switching factor in cultured sympathetic neurons (Patterson, 1978; Yamamori et al., 1989), where it induced expression of choline acetyltransferase and of neuropeptides characteristic of cholinergic neurons. Very similar effects in induction of phenotypic change are exerted by all members of the IL-6 family whose receptors contain gpl30 and LIF-R (Fann and Patterson 1994). On the other hand, IL-6 stimulates synthesis of choline acetyltransferase but not neuropeptides in sympathetic neurons (Oh and O'Malley, 1994). CDF/LIF seems likely to be capable of a similar phenotypic switching effect in vivo as, in a transgenic mouse in which CDF/LIF is expressed in pancreatic islets under the control of the insulin promoter, the neurotransmitter phenotype of the pancreatic sympathetic innervation is changed from adrenergic to cholinergic (Bamber et al, 1994). In normal developing rats, a minor population of sympathetic neurons is cholinergic. These cholinergic neurons, which predominantly supply the sweat glands of the foot pads, are originally adrenergic and undergo a postmitotic change in phenotype. As a model system of target-derived phenotypic determination, the phenotypic switch of these

Leukemia inhibitory factor and phenotypic specialization

sympathetic sudomotor neurons has been extensively studied by Landis and her colleagues (Landis, 1988; Rao and Landis, 1993). It is intriguing to postulate that CDF/LIF is responsible for this phenotypic conversion and it is therefore of interest that CDF/LIF is expressed at a significant level in the rat foot pad, whereas it cannot be detected in most other tissues innervated by sympathetic neurons (Yamamori, 1991b). However, two lines of evidence suggest that CDF/LIF does not act as a cholinergic factor in vivo. Firstly, in lif mice the sympathetic neurons that innervate the sweat glands are cholinergic as in normal mice (Rao et al., 1993a) and, in cntf mice, the sympathetic innervation to the sweat gland is also normal (Masu et al., 1993), suggesting that neither CNTF nor CDF/LIF alone is involved in the phenotypic conversion. Furthermore, CDF/LIF transcripts appear to be normally expressed in the foot pad of Tabby mutant mice which lack sweat glands, suggesting that the mRNA of CDF/LIF is expressed in other cellular compartments than the sweat gland (Rao et al., 1993a). It has been reported that extracts of rat foot pads contain an additional protein termed SGF (sweat gland factor). SGF seems to be distinct from known members of the IL-6 family, and its identity remains to be determined (Leung et al., 1992; Rao etal., 1992;Rohrer, 1992). 5.4.2. CDF/LIF as a neurotrophic factor in vitro Murphy et al. (1991) were the first to report that CDF/LIF exerts neurotrophic effects in developing sensory neurons (DRG). CDF/LIF also exerts neurotrophic effects on spinal motoneurons and nodose ganglia neurons (Martinou et al., 1992; Henderson et al., 1993; Thaler et al., 1994). Furthermore, both CNTF and CDF/LIF stimulate survival of spinal motoneurons in vivo (Oppenheim et al., 1991), and up-regulate the activity of choline acetyltransferase in cultured rat motoneurons (Michikawa et al., 1992; Wong et al., 1993; Cervini et al., 1994). Because CNTF is known to be a neurotrophic factor (Adler et al., 1979) and CNTF has a close relationship with CDF/LIF (Davis, et al., 1991), it is not surprising

T. Yamamori

to find that CDF/LIF also possesses neurotrophic activity. However, the mechanism behind this activity may be complex. For example, the responsiveness of rat sympathetic neurons to CDF/LIF and to CNTF changes during development (Kessler et al., 1993; Kotzbauer et al, 1994), and CDF/LIF may sometimes drive developing sympathetic neurons toward their death, at least in vitro (Kessler etal., 1993). The functions of CDF/LIF in vivo appear to differ from those of CNTF. As already described, the tissue distributions of CNTF and CDF/LIF are different. Among the peripheral tissues of rat innervated by sympathetic neurons, detectable amounts of CDF/LIF are expressed only in the foot pad, while CNTF is widely distributed. In the central nervous system, CDF/LIF is restricted to some visual neurons (Yamamori, 1991b), whereas CNTF is predominantly expressed in astrocytes of the optic nerve and olfactory bulb, and at moderate levels in other areas (Stockli et al., 1991). Conversely, a specific receptor subunit for CNTF (CNTF-a) is expressed only in the nervous system and muscle, whereas gpl30 and LIF-R are expressed in most tissues (Ip et al., 1993). Among distinct populations of cholinergic neurons in the rat brain (Mesulam et al., 1983), CNTF and CDF/LIF may exert neurotrophic effects on those that project to peripheral targets (motoneurons of the spinal cord and cranial motor nuclei), but do not affect neurons that project internally (neurons of the basal forebrain, striatum, and penduculopontine nucleus) (Arakawa et al, 1990; Martinou et al., 1992; Henderson et al., 1993; Wong et al., 1993; Zum and Werren, 1994). CNTF has also been suggested to be a 'lesion factor' (Thoenen, 1991). Rescue of lesioned facial motoneurons, septal neurons, thalamic neurons, and dopaminergic mesencephalic neurons has been reported after exogenous application of CNTF in the rat and the chicken (Sendtner et al., 1990; Hagg et al., 1992; Clatterbuck et al., 1993). CNTF also prevents degeneration of motoneurons in cnlf' mice with progressive motor neuropathy (Sendtner et al., 1992), and disruption of the CNTF gene results in motor neuron degeneration in the mutant mice (Masu et al., 1993). Preliminary data of the

279

gene targeting of cntf-R(a) reveal several developmental defects in cntf-R{a)- mice (Stahl and Yancopoulos, 1994). Curiously, however, a null mutation caused by an alternative splicing in the human CNTF gene does not cause any apparent functional defects (Takahashi et al., 1994). This discrepancy between the mouse and the human should be kept in mind. Recent studies suggest that CDF/LIF also acts as a repair factor, but in a different manner from that of CNTF. For example, levels of transcripts encoding muscarinic acetylcholine receptors and substance P receptors are regulated in cultured sympathetic neurons differentially by CDF/LIF and CNTF (Ludlam et al., 1994). 5.4,3. CDF/LIF as a repair factor in vivo It has been reported that expression of some neuropeptides such as vasoactive intestinal peptide and substance P are induced following explantation of ganglia into organ culture (Adler and Black, 1984; Freidin and Kessler, 1991; Zigmond et al., 1992) and after axotomy in situ (Hyatt-Sachs et al., 1993; Rao et al., 1993b; Mahendra et al., 1994; Sun et al., 1994). On the other hand, CDF/LIF has been shown to induce several neuropeptides including substance P and vasoactive intestinal peptide (Nawa and Patterson, 1990; Nawa et al., 1991; Freiden and Kessler, 1991; Mulderry, 1994; Lewis et al., 1994). Rao et al. (1993a) found that the induction of imunorecativity for vasoactive intestinal peptide is almost abolished in ganglia explanted from lif' mice. Furthermore, the induction of immunoreactivities for vasoactive intestinal peptide and neurokinin A following axotomy of SCO neurons is significantly suppressed in lif- mice (Rao et al., 1993a). Interestingly, antiserum against CDF/LIF blocks induction of vasoactive intestinal peptide in normal axotomized rats, but antiserum against CNTF does not (Sun et al., 1994). After peripheral nerve injury, CDF/LIF is found to be subjected to retrograde axonal transport. The expression of CDF/LIF also increases near damaged tissues (Banner and Patterson, 1994; Curtis et al., 1994). There is no direct evidence as to the type of cells that produce CDF/LIF in vivo, but Schwann-like

280

cells have been demonstrated to produce it in explanted organ culture (Banner and Patterson, 1994). IL-1 has been shown to induce CDF/LIF in cultured non-neuronal cells of sympathetic ganglia (Shadiack et al, 1993). IL-1, TGF-)3 and TNF stimulate CDF/LIF production in other nonneuronal cells, such as lung fibroblasts and cultured human astrocytes (Aloisi et al., 1994; Elias et al., 1994), with IL-1 and TGF-^ appearing to interact in a synergistic fashion via a PKC-dependent pathway (Elias et al., 1994). By applying CDF/LIF to the terminal axons of cultured sympathetic neurons, long-range intracellular signaling to the cell bodies can be achieved (Ure et al., 1992; Ure and Campenot, 1994). Although the mechanism by which CDF/LIF is retrogradely transported is not understood, at least some CDF/LIF is transported into the cell bodies in an intact form, rapidly degraded there, and then released into the extracellular space (Ure and Campenot, 1994). Application of CDF/LIF to lesioned rat motor neurons significantly reduces neuronal death in vivo, although CDF/LIF has a weaker effect on motoneuron survival in vitro than NT3 or BDNF have (Hughes et al., 1993; Cheema et al, 1994b). CDF/LIF also prevents the death of axotomized sensory neurons in the DRG of the neonatal rat (Cheema et al., 1994a). These data strongly suggest that CDF/LIF may act as a repair factor in the peripheral nervous system. It has recently been suggested that CDF/LIF may play a similar role as a lesion factor in the brain (Patterson, 1994). It seems unlikely that a similar role exists for CNTF. Both the distributions of the two factors in the nervous system (see Section 5.4.2.) and their regulation (Friedman et al., 1992; Rabinovsky et al., 1992; Sendtner et al., 1992; Seniuk et al., 1992) are different. Furthermore, CDF/LIF is dramatically induced after injury, while the expression of CNTF is suppressed (Seniuk et al., 1992). 5A.4. Effects of CDF/LIF on glial cells Raff et al. (1983) identified a glial progenitor (0-2A progenitor) cell which can differentiate into

Leukemia inhibitory factor and phenotypic

specialization

either an oligodendrocyte or an astrocyte. In the presence of platelet-derived growth factor (PDGF) and basic fibroblast growth factor (bFGF), 0-2A progenitor cells do not undergo differentiation but continue to divide (Bogler, et al., 1993). Both CNTF and CDF/LIF can induce differentiation of 0-2A progenitor cells into astrocytes in vitro (Hughes et al., 1988) and Mayer et al. (1994) found that this induction occurs only when these proteins are applied together with extracellular matrix derived from cultures of endothelial cells. After more than a decade, it was concluded that 0-2A progenitor cells do not differentiate into astrocytes during normal development (see Franklin and Blakemore, 1995). Although CNTF seems to be the key factor involved in the control of astrocyte differentiation in vitro, in vivo studies strongly suggest that they are actually derived from distinct precursor cells (Skoff, 1990; Fulton et al., 1992; Franklin and Blakemore, 1995). A possible function of CNTF and CDF/LIF may be to exert astrocyte-inducing activity in pathological states involving damage to oligodendrocytes and astrocytes (see Franklin and Blakemore, 1995), or to play complementary roles in repair of the nervous system. In any case, the effects of CDF/LIF in glial cell functions in vivo remain to be clarified, as well as those of CNTF in glial development. 5.5. Other systems 5.5.1. A defense factor against injury The role of CDF/LIF after injury to body systems in general may be analogous to that discused in the previous Section with respect to the nervous system. Overproduction of CDF/LIF causes cachexia (Metcalf and Gearing, 1989). OSM possesses a potent mitogenic activity on AIDSKaposi's sarcoma-derived cells whereas CDF/LIF stimulates HIV replication in mononuclear phagocytes (Miles et al., 1992; Broor et al., 1994). On the other hand, the presence of CDF/LIF at a subthreshold level may be advantageous for defense against invasive stimuli, endotoxic shock, hyperoxia, and radiation injury (Alexander et al., 1992;

T. Yamamori

1994; Tsan et al., 1992; Wong et al., 1992; Moran et al., 1994; Wesselingh et al., 1994). Sera from the patients with hematological disorders induce suppression of growth and differentiation of myeloid leukemia cells. These effects are presumably mediated by members of the IL-6 family including CDF/LIF (Kanatani et al., 1993). In addition, biologically active CDF/LIF is constitutively produced and secreted by cultured bone marrow stromal cells of patients with acute and chronic myelogenous leukemia, myelodysplastic syndrome, and hairy cell leukemia. In these patients, CDF/LIF might play a role to suppress leukemia. Synergistic effects of CDF/LIF and TNF against radiation-induced injury have also been reported (Wong et al., 1992; see also Alexander et al., 1994). CDF/LIF inhibits the development of experimentally induced atherosclerosis and hence may play a role in the pathogenesis of human disease (Gillett et al., 1993; Moran et al., 1994). In denervated rat skeletal muscle, synthesis of the three CNTF receptor subunits is up-regulated (Helgren et al., 1994) and both CNTF or CDF/LIF exert myotrophic effects, attenuating the morphological and functional changes caused by the denervation (Barnard et al., 1994; Helgren et al., 1994). 5.5.2. CDF/LIF in inflammatory reactions CDF/LIF levels are elevated in various inflammatory body fluids (Lotz et al., 1992; Waring et al., 1992; Hamilton et al., 1993; Alexander et al., 1994; Waring et al., 1994a,b). CDF/LIF is produced in articular tissues, and is also detected in synovial joint fluid in patients with rheumatoid arthritis (Lotz et al., 1992). Acute phase proteins are induced by the members of the IL-6 family (Richards et al., 1992; Mayer et al., 1993) and these proteins may play a role in the pathogenesis of rheumatoid arthritis (Blackburn, 1994). CDF/LIF is also present during acute rejection in the urine of kidney graft patients (Taupin et al., 1992). Cyclosporine, a potent suppressive agent of T lymphocyte activation, inhibits expression of the CDF/LIF gene (Bentouimou et al., 1993). This suggests that CDF/LIF and other members of the

281

IL-6 family may play roles in the inflammatory response, in coordination with other cytokines, growth factors and compounds such as prostaglandins (Fattori et al., 1994). 5.5.3. Cachexia It has been reported that cachexia develops in mice engrafted with LIF-overproducing cells (FD/LIF) (Metcalf and Gearing, 1989). Purification of the proteins produced by a human melanoma cell line, SEKI, which suppresses the expression of lipoprotein lipase in 3T3-L1 adipocytes, revealed its identity as CDF/LIF (Mori et al, 1989). CDF/LIF may affect the expression of lipoprotein lipase in 3T3-L1 cells, primarily at a post-transcriptional level (Marshall et al., 1994). Mori et al. (1991) further reported that CDF/LIF is constitutively expressed in nude mice bearing grafted melanoma cells. Mice which received melanoma cells producing little CDF/LIF do not develop cachexia at all, whereas those which received cells expressing high levels of CDF/LIF developed severe cachexia. Surgical removal of the tumors eliminated the syndrome (Mori et al., 1991). Not only the melanoma cell line but also a variety of other tumor cell lines produce CDF/LIF constitutively (Gascan et al., 1990). Thus, CDF/LIF produced by tumors may be a direct cause of cachexia. 6. Conclusion CDF/LIF, identified only relatively recently, has become one of the most extensively studied cytokines and tumed out to play a remarkable varietyof roles in living organisms. Recent advances in molecular and cellular biological studies of CDF/LIF reveal that this factor has at least three important specific functions in vivo. First, it is essential for blastocyst implantation. Second, it is required for self-renewal of a group of myeloid stem cells. Third, it probably functions as a repair factor following nerve injury and in some host defence systems. It remains, however, unknown in what manner CDF/LIF, as a member of the IL-6 family, interacts with other cytokines and growth factors to fulfil these and other complex physiological functions.

282

Acknowledgments I thank Dr. Motoy Kuno for comments on the in vivo role of CDF/LIF and Ms. Noriko Ishikawa for her help in preparation of this manuscript. References Abdel-Meguid, S.S., Shieh, H.S., Smith, W.W., Dayringer, H.E., Violand, B.N. and Bentl, L.A. (1987) Threedimensional structure of a genetically engineered variant of porcine growth hormone. Proc. Natl Acad. Sci. USA 84: 6434-6437. Abe, E., Tanaka, H., Ishimi, Y., Miyaura, C, Hayashi, T., Nagasawa, H., Tomida, M., Yamaguchi, Y., Hozumi, M. and Suda, T. (1986) Differentiation-inducing factor purified from conditioned medium of mitogen-treated spleen cell cultures stimulates bone resorption. Proc. Natl. Acad. Sci. USA 83: 5958-5962. Abe, E., Ishimi, Y., Takahashi, N., Akatsu, T., Ozawa, H., Yamana, H., Yoshiki, S. and Suda, T. (1988) A differentiation-inducing factor produced by the osteoblastic cell hne MC3T3-E1 stimulates bone resorption by promoting osteoclast formation. J. Bone Min. Res. 3: 635-645. Abe, T., Murakami, M., Sato, T., Kajiki, M., Mitsuhara, O. and Kodaira, R. (1989) Macrophage differentiation inducing factor from human monocytic cells is equivalent to murine leukemia inhibitory factor. J. Biol. Chem. 264: 8941-8945. Adamson, E.D. (1993) Activities of growth factors in preimplantation embryos. J. Cell. Biochem. 53: 280-287. Adler, J. and Black, I. (1984) Plasticity of substance P in mature and aged sympathetic neurons in culture. Science 225: 1499-1500. Adler, R., Landa, K.B., Manthorpe, M. and Varon, S. (1979) Cholinergic neuronotrophic factors: intraocular distribution of soluble trophic activity for ciliary neurons. Science 204: 1434-1436. Akira, S., Isshiki, H., Sugita, T., Tanabe, O., Kinoshita, S., Nishio, Y., Nakajima, T., Hirano, T. and Kishimoto, T. (1990) A nuclear factor for IL-6 expression (NF-IL-6) is a member of a C/EBP family. EMBO J. 9: 1897-1906. Alexander, H.R., Wong, G.G.H., Doherty, G.M., Venzon, D.J., Fraker, D.L. and Norton, J.A. (1992) Differentiation factor/leukemia inhibitory factor protection against lethal endotoxemia in mice: synergistic effect with interleukin 1 and tumor necrosis factor. J. Exp. Med. 175: 1139-1142. Alexander, H.R., Billingsley, K.G., Block, M.I. and Fraker, D.L. (1994) D-factor/leukemia inhibitory factor: evidence for its role as a mediator in acute and chronic inflammatory disease. Cytokine 6: 589-596. Allan, E.H., Hilton, D.J., Brown, M.A., Evely, R.S., Yumita, S., Metcalf, D., Gough, N.M., Ng, K.W., Nicola, N.A. and Martin, T.J. (1990) Osteoblasts display receptors for and re-

Leukemia inhibitory factor and phenotypic specialization sponses to leukemia- inhibitory factor. J. Cell. Physiol. 145: 110-119. Aloisi, F., Rosa, S., Testa, U., Bonsi, P., Russo, G., Peschle, C. and Levi, G. (1994) Regulation of leukemia inhibitory factor synthesis in cultured human astrocytes. J. Immunol. 152: 5022-5031. Anegon, I., Cuturi, M.C., Godard, A., Moreau, M., Terqui, M., Martinat- Botte, F. and Soulillou, J.P. (1994) Presence of leukaemia inhibitory factor and interleukin 6 in porcine uterine secretions prior to conceptus attachment. Cytokine 6: 493^99. Arakawa, Y., Sendtner, M. and Thoenen, H. (1990) Survival effect of ciliary neurotrophic factor (CNTF) on chick embryonic neurons in culture: comparison with other neurotrophic factors and cytokines. J. Neurosci. 10: 35073515. Argestsinger, L.S., Campbell, G.S., Yang, X., Witthuhn, B.A., Silvennoinen, O., Ihle, J.N. and Carter-Su, C. (1993) Identification of JAK2 as a growth hormone receptor-associated tyrosine kinase. Cell 74: 237-244. Bamber, B.A., Masters, B.A., Hoyle, G.W., Brinster, R.L. and Palmiter, R.D. (1994) Leukemia inhibitory factor induces neurotransmitter switching in transgenic mice. Proc. Natl. Acad. Sci. USA 91: 7839-7843. Banner, L.R. and Patterson, P.H. (1994) Major changes in the expression of the mRNAs for cholinergic differentiation factor/leukemia inhibitory factor and its receptor after injury to adult peripheral nerves and ganglia. Proc. Natl. Acad. Sci. USA 91: 7109-7113. Barnard, W., Bower, J., Brown, M.A., Murphy, M. and Austin, L. (1994) Leukemia inhibitory factor (LIF) infusion stimulates skeletal muscle regeneration after injury: injured muscle expresses LIF mRNA. J. Neurol. Sci. 123: 108113. Baumann, H. and Wong, G.G. (1989) Hepatocyte-stimulating factor III shares structural and functional identity with leukemia-inhibitory factor././mmMno/. 143: 1163-1167. Baumann, H., Morella, K.K., Campos, S.P., Cao, Z. and Jahreis, G.P. (1992) Role of CAAT-enhancer binding protein isoforms in the cytokine regulation of acute-phase plasma protein genes. /. Biol. Chem. 267: 19744-19751. Baumann, H., Ziegler, S.F., Mosley, B., Morella, K.K., Pajovie, S. and Gearing, D.P. (1993) Reconstitution of the response to leukemia inhibitory factor, oncostatin M, and ciliary neurotrophic factor in hepatoma cells. J. Biol. Chem. 268: 8414-8417. Baumann, H., Symes, A.J., Comeau, M.R., Morella, K.K., Wang, Y.P., Friend, D., Ziegler, S.F., Fink, J.S. and Gearing, D.P. (1994) Multiple regions within the cytoplasmic domains of the leukemia inhibitory factor receptor and gpl30 cooperate in hepatic and neuronal cells. Mol. Cell. Biol. 14: 138-146. Bazan, J.F. (1991) Neuropoietic cytokines in the hematopoietic fold. Neuron 7: 197-208. Bentouimou, N., Moreau, J.F., Payrat, M.A., Soulillou, J.P. and Hallet, M.M. (1993) The effects of cyclosporine on

T. Yamamori HILDA/LIF gene expression in human T cells. Transplantation S5\ 163-167. Bhatt, H., Brunei, L.J. and Stewart, C.L. (1991) Uterine expression of leukemia inhibitory factor coincides with the onset of blastocyst implantation. Proc. Natl. Acad. Sci. USA 88: 11408-11412. Blackburn, Jr., W.D. (1994) Validity of acute phase proteins as markers of disease activity. J. Rheumatol. 21 (suppl. 42): 9-13. Bogler, O., Wren, D., Barnett, S.C, Land, H. and Noble, M. (1993) Cooperation between two growth factors promotes extended self- renewal and inhibits differentiation of oligodendrocyte- type-2 astrocyte (0-2A) per progenitor cells. Proc. Natl. Acad. Sci. USA 87: 6368-6372. Bonni, A., Frank, D.A., Schindler, C. and Greenberg, M.E. (1993) Characterization of a pathway for ciliary neurotrophic factor signaling to the nucleus. Science 262: 1575-1579. Bottorff, D. and Stone, J.C. (1992) The murine leukemia inhibition factor gene (Lif) is located on proximal chromosome 11, not chromosome 13. Mamm. Genome 3: 681684. Boulton, T.G., Stahl, N. and Yancopoulos, G.D. (1994) Ciliary neurotrophic factor/leukemia inhibitory factor/interleukin- 6/oncostatin-M family of cytokines includes tyrosine phosphorylation of a common set of proteins overlapping those induced by other cytokines and growth factors. J. Biol Chem. 269: 11648-11655. Broor, S., Kusari, A.B., Zhang, B., Seth, P., Richman, D.D., Carson, D.A., Wachsman, W. and Lotz, M. (1994) Stimulation of HIV replication in mononuclear phagocytes by leukemia inhibitory factor. J. Acq. Imm. Defic. Syndrome 1: 647-654. Budarf, M., Emanuel, B.S., Mohandas, T., Goeddel, D.V. and Lowe, D.G. (1989) Human differentiation-stimulating factor (leukemia inhibitory factor, human interleukin DA) gene maps distal to the Ewing sarcoma breakpoint on 22q. Cytogenet. Cell Genet. 52: 19-22. Burgess, A.W. (1989) Life after LIF? BioEssays 10: 166-169. Caldenhoven, E., Coffer, P., Yuan, J., Van De Stolpe, A., Horn, F., Kruijer, W. and Van Der Saag, P.T. (1994) Stimulation of the human intercellular adhesion molecule-1 promoter by interleukin-6 and interferon-y involves binding of distinct factors to a palindromic response element. J. Biol. Chem. 269: 21146-21154. Canalis, E., McCarthy, T. and Centrella, M. (1988) Growth factors and the regulation of bone remodeling. J. Clin. Invest. S\: 211-2Sl. Canalis, E., McCarthy, T. and Centrella, M. (1989) The regulation of bone formation by local growth factors. Bone Min. Res. 6: 27-57. Canalis, E., McCarthy, T.L. and Centrella, M. (1991) Growth factors and cytokines in bone cell metabolism. Annu. Rev. Med. 42: 17-24. Cervini, R., Berrard, S., Bejanin, S. and Mallet, J. (1994) Regulation by CDF/LIF and retinoic acid of multiple ChAT

283 mRNAs produced from distinct promoters. NeuroReport 5: 1346-1348. Charnock-Jones, D.S., Sharkey, A.M., Fenwick, P. and Smith, S.K. (1994) Leukaemia inhibitory factor mRNA concentration peaks in human endometrium at the time of implantation and the blastocyst contains mRNA for the receptor at this time. J. Reprod. Fertil. 101: 421-426. Cheema, S.S., Richards, L., Murphy, M. and Bartlett, P.F. (1994a) Leukemia inhibitory factor prevents the death of axotomised sensory neurons in the dorsal root ganglia of the neonatal rat. /. Neurosci. Res. 37: 213-218. Cheema, S.S., Richards, L.J., Murphy, M. and Bartlett, P.F. (1994b) Leukaemia inhibitory factor rescues motoneurones from axotomy-induced cell death. NeuroReport 5: 989-992. Chen, R.H. and Blenis, J. (1990) Identification of Xenopus S6 protein kinase homologs (pp90rsk) in somatic cells: phosphorylation and activation during initiation of cell proliferation. Mol. Cell. Biol. 10: 3204-3215. Clatterbuck, R.E., Price, D.L. and Koliatsos, V.E. (1993) Ciliary neurotrophic factor prevents retrograde neuronal death in the adult central nervous system. Proc. Natl. Acad. Sci. USA 90: 2222-2226. Conquet, F. and Brulet, P. (1990) Developmental expression of myeloid leukemia inhibitory factor gene in preimplantation blastocysts and in extraembryonic tissue of mouse embryos. Mol. Cell. Biol. 10: 3801-3805. Cross, J.C, Werb, Z. and Fisher, S.J. (1994) Implantation and the placenta: key pieces of the development puzzle. Science 266: 1508-1518. Croy, B.A., Guilbert, L.J., Brown, M.A., Gough, N.M., Stinchcomb, D.T., Reed, N. and Wegmann, T.G. (1991) Characterization of cytokine production by the metrial gland and granulated metrial gland cells. /. Reprod. Immunol. 19: 149-166. Curtis, R., Scherer, S.S., Somogyi, R., Adryan, K.M., Ip, N.Y., Zhu, Y., Lindsay, R.M. and DiStefano, P.S. (1994) Retrograde axonal transport of LIF is increased by peripheral nerve injury: correlation with increased LIF expression in distal nerve. Neuron 12: 191-204. D'Andrea, A.D., Yoshimura, A., Youssoufian, H., Zon, L.I., Koo, J.W. and Lodish, H.F. (1991) The cytoplasmic region of the erythropoietin receptor contains nonoverlapping positive and negative growth-regulatory domains. Mol. Cell. Biol. 11: 1980-1987. Darnell, Jr., J.E., Kerr, I.M. and Stark, G.R. (1994) Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins. Science 264: 1415-1421. Davis, S., Aldrich, T.H., Valenzuela, D.M., Wong, V.V., Furth, M.E., Squinto, S.P. and Yancopoulos, D. (1991) The receptor for ciliary neurotrophic factor. Science 253: 5963. Davis, S., Aldrich, T.H., Stahl, N., Pan, L., Taga, T., Kishimoto, T., Ip, N.Y. and Yancopoulos, G.D. (1993) LIFy3 and gpl30 as heterodimerizing signal transducers of the tripartite CNTF receptor. Science 259: 1736-1739.

284

De Felici, M. and Pesce, M. (1994) Interactions between migratory primordial germ cells and cellular substrates in the mouse. In: Germline Development, Ciba Foundation Symposium 182, Wiley, Chichester, pp. 140-153. Derigs, H.G. and Boswell, H.S. (1993) LIP mRNA expression is transcriptionally regulated in murine bone marrow stromal cells. Leukemia 7: 630-634. Edwards, R.G. (1994) Implantation, interception and contraception. Hum. Reprod. 9 (suppl. 2): 73-87. Elias, J.A., Zheng, T., Whiting, N.L., Marcovici, A. and Trow, T.K. (1994) Cytokine-cytokine synergy and protein kinase C in the regulation of lung fibroblast leukemia inhibitory factor. Am. J. Physiol. 266: L426-L435. Ernst, M., Gearing, D.P. and Dunn, A.R. (1994) Functional and biochemical association of Hck with the LIF/IL-6 receptor signal transducing subunit gpl30 in embryonic stem cells. EMBO J. 13: 1574-1584. Escary, J.-L., Perreau, J., Dumenil, D., Ezine, S. and Brulet, P. (1993) Leukaemia inhibitory factor is necessary for maintenance of haematopoietic stem cells and thymocyte stimulation. Nature 363: 361-364. Evans, D.B., Gerber, B. and Feyen, J.H. M. (1994) Recombinant human leukemia inhibitory factor is mitogenic for human bone-derived osteoblast-like cells. Biochem. Biophys. Res. Commun. 199: 220-226. Fann, M.J. and Patterson, P.H. (1994) Neuropoietic cytokines and activin A differentially regulate the phenotype of cultured sympathetic neurons. Proc. Natl. Acad. Sci. USA 91: 43-47. Fattori, E., Cappelletti, M., Costa, P., Sellitto, C, Cantoni, L., Carelli, M., Faggioni, R., Fantuzzi, G., Ghezzi, P. and Poh, V. (1994) Defective inflammatory response in interleukin 6-deficient mice. /. Exp. Med. 180: 1243-1250. Firmbach-Kraft, I., Byers, M., Shows, T., Dalla-Fevera, R. and Krolewski, J.J. (1990) Tyk2, protoype of a novel class of non-receptor tyrosine kinase genes. Oncogene 5: 13291336. Fourcin, M., Chevalier, S., Lebrun, J.J., Kelly, P., Pouplard, A., Wijdense, J. and Gascan, H. (1994) Involvement of gpl30/ interleukin-6 receptor transducing component in interleukin- 11 receptor. Eur. J. Immunol. 24: 277-280. Franklin, R.J.M. and Blakemore, W.F. (1995) Glial-cell transplantation and plasticity in the 0-2A lineage - implications for CNS repair. Trends Neurosci. 18: 151-156. Freidin, M. and Kessler, J. (1991) Cytokine regulation of substance P expression in sympathetic neurons. Proc. Natl. Acad. Sci. USA 88: 3200-3203. Friedman, B., Scherer, S.S., Rudge, J.S., Helgren, M., Morrisey, D., McClain, J., Wang, D.-Y., Wiegand, S.J., Furth, M.E., Lindsay, R.M. and Ip, N.Y. (1992) Regulation of ciliary neurotrophic factor expression in myelin-related Schwann cells in vivo. Neuron 9: 295-305. Fu, X.Y. and Zhang, J.J. (1993) Transcription factor p91 interacts with the epidermal growth factor receptor and mediates activation of the c-fos gene promoter. Cell 74: 11351145.

Leukemia inhibitory factor and p he no typic specialization Fukada, K. (1985) Purification and partial characterization of a cholinergic differentiation factor. Proc. Natl. Acad. Sci. USA 82: 8795-8799 Fukui, Y., Saito, T., Miyamoto, A., Yamashita, H. and Okamoto, Y. (1994) Effect of human inhibitory factor on in vitro development of parthenogenetic bovine morulae. Theriogenology 42: 1133-1139. Fukunaga, R., Ishizaka-Ikeda, E., Pan, C.-X., Seto, Y. and Nagata, S. (1991) Functional domains of the granulocyte colony- stimulating factor receptor. EMBO J. 10: 28552865. Fulton, B.P., Burne, J.F. and Raff, M.C. (1992) Visualization of 0-2A progenitor cells in developing and adult rat optic nerve by quisqualate-stimulated cobalt uptake. J. Neurosci. 12:4816-4833. Gascan, H., Godard, A., Ferenz, C, Naulet, J., Praloran, V., Peyrat, M.-A., Hewick, R., Jacques, Y., Moreau, J.-F. and Soulillou, J.-P. (1989) Characterization and NH2-terminal amino acid sequence of natural human interleukin for DA cells: leukemia inhibitory factor. J. Biol. Chem. 264: 21509-21515. Gascan, H., Anegon, I., Praloran, V., Naulet, J., Godard, A., Soulillou, J.-P. and Jacques, Y. (1990) Constitutive production of human interleukin for DA cells/leukemia inhibitory factor by human tumor cell lines derived from various tissues. /. Immunol. 144: 2592-2598. Gearing, D.P. and Bruce, A.G. (1992) Oncostatin M binds the high- affinity leukemia inhibitory factor receptor. New Biol. 4: 61-65. Gearing, D.P., Gough, N.M., King, J.A., Hilton, D.J., Nicola, N.A., Simpson, R.J., Nice, E.C., Kelso, A. and Metcalf, D. (1987) Molecular cloning and expression of cDNA encoding a murine myeloid leukaemia inhibitory factor (LIF). EM50 7. 6:3995^002. Gearing, D.P., Thut, C.J., VandeBos, T., Gimpel, S.D., Delaney, P.B., King, J., Price, V., Cosman, D. and Beckmann, M.P. (1991) Leukemia inhibitory factor receptor is structurally related to the IL-6 signal transducer, gpl30. EMBO J. 10: 2839-2848. Gearing, D.P., Comeau, M.R., Friend, D.J., Gimpel., S.D., Thut, C.J., McGourty, J., Brasher, K.K., King, J.A., Gillis, S., Mosley, B., Ziegler, S.F. and Cosman, D. (1992) The IL6 signal transducer, gpl30: an oncostatin M receptor and affinity converter for the LIF receptor. Science 255: 14341437. Gearing, D.P., Ziegler, S.F., Comeau, M.R., Friend, D., Thoma, B., Cosman, D., Park, L. and Mosley B. (1994) Proliferative responses and binding properties of hematopoietic cells transfected with low-affinity receptors for leukemia inhibitory factor, oncostatin M, and ciliary neurotrophic factor. Proc. Natl. Acad. Sci. USA 91: 1119-1123. Gillett, N.A., Lowe, D., Lu, L., Chan, C. and Ferrara, N. (1993) Leukemia inhibitory factor expression in human carotid plaques: possible mechanism for inhibition of large vessel endothelial regrowth. Growth Factors 9: 301-305. Gimble, J.M., Wanker, F., Wang, C.S., Bass, H., Wu, X.Y.,

T. Yamamori Kelly, K., Yancopoulos, G.D. and Hill, M.R. (1994) Regulation of bone marrow stromal cell differentiation by cytokines whose receptors share the gpl30 protein. J. Cell. Biochem. 54: 122-133. Gough, N.M. and Williams, R.L. (1989) The pleiotropic actions of leukemia inhibitory factor. Cancer Cells 1: 77-80. Gough, N.M., Gearing, D.P., King, J.A., Willson, T.A., Hilton, D.J., Nicola, N.A. and Metcalf, D. (1988) Molecular cloning and expression of the human homologue of the murine gene encoding myeloid leukemia-inhibitory factor. Proc. Natl Acad. Sci. USA 85: 2623-2627. Gough, N.M., Williams, R.L., Hilton, D.J., Pease, S., Willson, T.A., Stahl, J., Gearing, D.P., Nicola, N.A. and Metcalf, D. (1989) LIF: a molecule with divergent actions on myeloid leukaemic cells and embryonic stem cells. Reprod. Fertil. Dev. 1:281-288. Gu, H., Zou, Y.R. and Rajewsky, K. (1993) Independent control of immunoglobulin switch recombination at individual switch regions evidenced through Cre-loxP-mediated gene targeting. Cell 73: 1155-1164. Gu, H., Marth, J.D., Orban, P.C., Mossmann, H. and Rajewsky, K. (1994) Deletion of a DNA polymerase ^ gene segment in T cells using cell type-specific gene targeting. Science 265: 103-106. Hagg, T., Quon, D., Higaki, J. and Varon, S. (1992) Ciliary neurotrophic factor prevents neuronal degeneration and promotes low-affinity NGF receptor expression in the adult rat CNS. Neuron 8: 145-158. Hakeda, Y., Sudo, T., Ishizuka, S., Tanaka, K., Higashino, K., Kusuda, M., Kodama, H. and Kumegawa, M. (1991) Murine recombinant leukemia inhibitory factor modulates inhibitory effect of 1,25 dihydroxyvitamin D3 on alkaline phosphatase activity in MC3T3-E1 cells. Biochem. Biophys. Res. Commun. 175: 577-582. Hall, A.K. and Rao, S.M. (1992) Cytokines and neurokines: related ligand and related receptors. Trends Neuro sci. 15: 35-37. Hamilton, J.A., Waring, P.M. and Filonzi, E.L. (1993) Induction of leukemia inhibitory factor in human synovial fibroblasts by IL-1 and tumor necrosis factor-a. J. Immunol. 150: 1496-1502. Harpur, A.G., Andres, A.C., Ziemiecki, A., Aston, R.R. and Wilks, A.F. (1992) JAK2, a third member of the JAK family of protein tyrosine kinases. Oncogene 7: 1347-1353. Hatakeyama, M., Mori, H., Doi, T. and Taniguchi, T. (1989) A restricted cytoplasmic region of IL-2 receptor p chain is essential for growth signal transduction but not for ligand binding and internalization. Cell 59: 837-845. Helgren, M.E., Squinto, S.P., Davis, H.L., Parry, D.J., Boulton, T.G., Heck, C.S., Zhu, Y., Yancopoulos, G.D., Lindsay, R.M. and DiStefano, P.S. (1994) Trophic effect of ciliary neurotrophic factor on denervated skeletal muscle. Cell 76: 493-504. Henderson, C.E., Camu, W., Mettling, C , Gouin, A., Poulsen, K., Karihaloo, M., Rullamas, J., Evans, T., McMahon, S.B., Armanimi, M.P., Berkemeier, L., Phillips, H.S. and Rosen-

285 thal, A. (1993) Neurotrophins promote motor neuron survival and are present in embryonic limb bud. Nature 363: 266-270. Hibi, M., Murakami, M., Saito, M., Hirano, T. and Taga, T. (1990) Molecular cloning and expression of an IL-6 signal transducer, gpl30. Cell 63: 1149-1157. Hilton, D.J., Nicola, N.A. and Metcalf, D. (1988) Purification of a murine leukemia inhibitory factor from krebs ascites cells. Ana/. Biochem. 173: 359-367. Hilton, D.J., Hilton, A.A., Raicevic, A., Raker, S., Harrisonsmith, M., Gough, N.M., Begley, C.G., Metcalf, D., Nicola, N.A. and Willson, T.A. (1994) Cloning of a murine IL-U receptor a-chain; requirement for gpl30 for highaffinity binding and signal transduction. EMBO J. 13: 4765-4775. Hirano, T., Matsuda, T. and Nakajima, K. (1994) Signal transduction through gpl30 that is shared among the receptors for the interleukin 6-related cytokine subfamily. Stem Cells 12: 262-277. Hoffman-Liebermann, B. and Liebermann, D.A. (1991a) Suppression of c-myc and c-myb is tightly linked to terminal differentiation induced by IL-6 or LIF and not growth inhibition in myeloid leukemia cells. Oncogene 6: 9 0 3 909. Hoffman-Liebermann, B. and Liebermann, D.A. (1991b) Interleukin-6 and leukemia inhibitory factor-induced terminal differentiation of myeloid leukemia cells is blocked at an intermediate stage by constitutive c-myc. Mol. Cell. Biol. 11:2375-2381. Hughes, S.M., Lillien, L.E., Raff, M.C., Rohrer, H. and Sendtner, M. (1988) Ciliary neurotrophic factor induces type-2 astrocyte differentiation in culture. Nature 335: 7 0 73. Hughes, R.A., Sendtner, M. and Thoenen, H. (1993) Members of several gene families influence survival of rat motoneurons in vitro and in vivo. J. Neurosci. Res. 36: 663-671. Hyatt-Sachs, H., Schreiber, R., Bennett, T. and Zigmond, R. (1993) Phenotypic plasticity in adult sympathetic ganglia in vivo: effects of deafferentation and axotomy on the expression of vasoactive intestinal peptide. J. Neurosci. 13: 16421653. Ichikawa, Y. (1969) Differentiation of a cell line of myeloid leukemia. J. Cell. Physiol. 74: 223-234. Ip, N.Y., Nye, S.H., Boulton, T.G., Davis, S., Taga, T., Li, Y., Birren, S.J., Yasukawa, K., Kishimoto, T., Anderson, D.J., Stahl, N. and Yancopoulos, G.D. (1992) CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component gpl30. C^//69: 1121-1132. Ip, N.Y., McClain, J., Barrezueta, N.X., Aldrich, T.H., Pan, L., Li, Y., Wiegand, S.J., Friedman, B., Davis, S. and Yancopoulos, G.D. (1993) The a component of the CNTF receptor is required for signaling and defines potential CNTF targets in the adult and during development. Neuron 10: 89-102. Ip, N.Y., Boulton, T.G., Li, Y., Verdi, J.M., Birren, S.J., An-

286 derson, D.J. and Yancopoulos, G.D. (1994) CNTF, FGF, and NGF collaborate to drive the terminal differentiation of MAH cells into postmitotic neurons. Neuron 13: 443^55. Ishibashi, T., Kimura, H., Uchida, T., Kariyone, S., Friese, P. and Burstein, S.A. (1989) Human interleukin 6 is a direct promotor of maturation of megakaryocytes in vitro. Proc. Natl. Acad. Sci. USA 86: 5953-5957. Johnson, J.A. and Nathanson, N.M. (1994) Differential requirements for p21ras and protein kinase C in the regulation of neuronal gene expression by nerve growth factor and neurokines. y. Biol. Chem. 269: 18856-18863 Kahn, M.A. and De Vellis, J. (1994) Regulation of an oligodendrocyte progenitor cell line by the interleukin-6 family of cytokines. Glia 12: 87-98. Kanatani, Y., Honma, Y., Tsuchimochi, T., Okabekado, J., Nagata, N., Motoyoshi, K. and Hozumi, M. (1993) Induction of differentiation and growth suppression of myeloid leukemia cells by sera of patients with hematological disorders. Am/cancer/?^5. 13:1905-1912. Kasukabe, T., Okabekado, J., Hozumi, M. and Honma, Y. (1994) Inhibition by interleukin 4 of leukemia inhibitory factor-induced, interleukin 6-induced, and dexamethasoneinduced differentiation of mouse myeloid leukemia cells role of c-myc and jun^ proto-oncogenes. Cancer Res. 54: 592-597. Kessler, J.A., Ludlam, W.H., Freidin, M.M., Hall, D.H., Michaelson, M.D., Spray, D.C., Dougherty, M. and Batter, D.K. (1993) Cytokine- induced programmed death of cultured sympathetic neurons. Neuron 11: 1123-1132. Kishimoto, T. (1989) The biology of interleukin-6. Blood 74: 1-10. Kishimoto, T., Taga, T. and Akira, S. (1994) Cytokine signal transduction. Cell 76: 253-262. Kotzbauer, P.T., Lampe, P.A., Estus, S., Milbrandt, J. and Johnson, E.M. (1994) Postnatal development of survival responsiveness in rat sympathetic neurons to leukemia inhibitory factor and ciliary neurotrophic factor. Neuron 12: 763773. Kurzrock, R., Estrov, Z., Wetzler, M., Gutterman, J.U. and Talpaz, M. (1991) LIF: not just a leukemia inhibitory factor. Endocrine Rev. 12: 208-217. Landis, S.C. (1988) Neurotransmitter plasticity in sympathetic neurons and its regulation by environmental factors in vitro and in vivo. In: A. Bjorklund, T. Hokfelt and C. Owman (Eds.), Handbook of Chemical Neuroanatomy, Vol. 6, Peripheral Nervous System, Elsevier, Amsterdam, pp. 65-115. Larner, A.C., David, M., Feldman, G.M., Igarashi, K., Hackett, R.H., Webb, D.S.A., Sweitzer, S.M., Petricoin, III, E.F and Finbloom, D.S. (1993) Tyrosine phosphorylation of DNA binding proteins by multiple cytokines. Science 261: 1730-1733. Layton, M.J., Cross, B.A., Metcalf, D.A., Ward, L.D., Simpson, R.J., Nicola, N.A. (1992) A major binding protein for leukemia inhibitory factor in normal mouse serum - identification as a soluble form of the cellular receptor. Proc. Natl. Acad. Sci. USA 89: 8616-8620.

Leukemia inhibitory factor and phenotypic specialization Layton, M.J., Lock, P., Metcalf, D. and Nicola, N.A. (1994a) Cross-species receptor binding characteristics of human and mouse leukemia inhibitory factor suggest a complex binding interaction. J. Biol. Chem. 269: 17048-17055. Layton, M.J., Owczarek, CM., Metcalf, D., Lock, P.A., Wilson, T.A., Gough, N.M., Hilton, D.J. and Nicola, N.A. (1994b) Complex binding of leukemia inhibitory factor to its membrane- expressed and soluble receptors. Proc. Soc. Exp. Biol. Med. 206: 295-298. Layton, M.J., Owczrek, CM., Metcalf, D., Clark, R.L., Smith, D.K., Treutlein, H.R. and Nicola, N.A. (1994b) Conversion of the biological specificity of murine to human leukemia inhibitory factor by replacing 6 amino acid residues. J. Biol. Chem. 47: 29891-29896. Leary, A.G., Wong, G.G., Clark, S.C, Smith, A.G. and Ogawa, M. (1990) Leukemia inhibitory factor differentiation-inhibiting activity/human interleukin for DA cells augments proliferation of human hematopoietic stem cells. Blood 15: 1960-1964. Lee, F.D. (1992) The role of interleukin-6 in development. Dev.Biol. 151:331-338. Leung, D.W., Parent, A.S., Cachianes, G., Esch, F., Coulombe, G., Nikolics, K., Eckenstein, F.P. and Nishi, R. (1992) Cloning, expression during development, and evidence for release of a trophic factor for ciliary ganglion neurons. Neuron 8: 1045-1053. Lewis, S.E., Schwarzschild, M.A., Dauer, W.T., Hamill, L.K., Symes, A.J., Fink, J.S. and Hyman, S.E. (1993) Leukemia inhibitory factor (LIF), ciliary neurotrophic factor (CNTF) and oncostain M (OSM) activate p21ras in a neuroblastoma cell line. Soc. NeuroscL Abstr., 19 (2): 1306. Lewis, S.E., Rao, M.S., Symes, A.J., Dauer, W.T., Fink, J.S., Landis, S.C. and Hyman, S.E. (1994) Coordinate regulation of choline acetyltransferase, tyrosine hydroxylase, and neuropeptide mRNAs by ciliary neurotrophic factor and leukemia inhibitory factor in cultured sympathetic neurons. J. Neurochem. 63: 429-438. Liu, J.W., Modrell, B., Aruffo, A., Marken, J.S., Taga, T., Yasukawa, K., Murakami, M., Kishimoto, T. and Shoyab, M. (1992) Interleukin-6 signal transducer gpl30 mediates oncostatin-M signaling. J. Biol. Chem. 267: 1676316766. Lord, K.A., Abdollahi, A., Thomas, S.M., DeMarco, M., Brugge, J.S., Hoffman-Liebermann, B. and Liebermann, D.A. (1991) Leukemia inhibitory factor and interleukin-6 trigger the same immediate early response, including tyrosine phosphorylation, upon induction of myeloid leukemia differentiation. Mol. Cell. Biol. 11: 4371-4379. Lorenzo, J.A., Sousa, S.L. and Leahy, CL. (1990) Leukemia inhibitory factor (LIF) inhibits basal bone resorption in fetal rat long bone cultures. Cytokine 2: 266-271. Lotz, M., Moats, T. and Villiger, P.M. (1992) Leukemia inhibitory factor is expressed in cartilage and synovium and can contribute to the pathogenesis of arthritis. J. Clin. Invest. 90: 888-896. Lowe, C, Cornish, J., Gallon, K., Martin, T.J. and Reid, I.R.

T. Yamamori (1991) Regulation of osteoblast proliferation by leukemia inhibitory factor. J. Bone Min. Res. 6: 1277-1283. Ludlam, W.H., Zang, Z., McCarson, K.E., Krause, J.E., Spray, D.C. and Kessler, J.A. (1994) mRNAs encoding muscarinic and substance P receptors in cultured sympathetic neurons are differentially regulated by LIF or CNTF. Dev. Biol. 164: 528-539. Liitticken, C, Wegenka U.M., Yuan, J., Buschmann, J., Schindler, C, Ziemiecki, A., Harpur, A.G., Wilks, A.F., Yasukawa, K., Taga, T., Kishimoto, T., Barbieri, G., Pellegrini, S., Sendtner, M., Heinrich, P.C. and Horn, F. (1993) Association of transcription factor APRF and protein kinase Jakl with interleukin-6 signal transducer gpl30. Science 263: 89-92. Maekawa, T., Metcalf, D. and Gearing, D.P. (1990) Enhanced suppression of human myeloid leukemic cell lines by combinations of IL-6, LIF, GM-CSF and G-CSF. Int. J. Cancer 45: 353-358. Mahendra, Y.S., Rao, S., Zigmond, R.E. and Landis, S.C. (1994) Regulation of vasoactive intestinal peptide expression in sympathetic neurons in culture and after axotomy: the role of cholinergic differentiation factor/leukemia inhibitory factor. J. Neurobiol. 25: 415^30. Marshall, M.K., Doerrler, W., Feingold, K.R. and Grunfeld, C. (1994) Leukemia inhibitory factor induces changes in lipid metabolism in cultured adipocytes. Endocrinology 135: 141-147. Martinou, J.C, Martinou, I. and Kato, A.C. (1992) Cholinergic differentiation factor (CDF/LIF) promotes survival of isolated rat embryonic motoneurons in vitro. Neuron 8: 737-744. Masu, Y., Wolf, E., Holtmann, M., Sendtner, M., Brem, G. and Thoenen, H. (1993) Disruption of the CNTF gene results in motor neuron degeneration. Nature 365: 27-32. Mayer, P., Geissler, K., Ward, M. and Metcalf, D. (1993) Recombinant human leukemia inhibitory factor induces acute-phase proteins and raises the blood platelet counts in nonhuman primates. Blood 81: 3226-3233. Mayer, M., Bhakoo, K. and Noble, M. (1994) CiUary neurotrophic factor and leukemia inhibitory factor promote the generation, maturation and survival of oligodendrocytes in vitro. Development 120: 143-153. Mesulam, M.M., Mufson, E.J., Wainer, B.H. and Levey, A.I. (1983) Central cholinergic pathways in the rat: an overview based on an alternative nomenclature (Chl-Ch6). Neuroscience 10: 1185-1201. Metcalf, D. (1991) The leukemia inhibitory factor (LIF). Int. J. Cell Cloning 9: 95-108. Metcalf, D. (1992) Leukemia inhibitory factor. A puzzling polyfunctional regulator. Growth Factors 7: 169-173. Metcalf, D. (1993) Hematopoietic regulators - redundancy or subtlety? Blood 82: 3515-3523. Metcalf, D. and Gearing, D.P. (1989) Fatal syndrome in mice engrafted with cells producing high levels of the leukemia inhibitory factor. Proc. Natl. Acad. Sci. USA 86: 59485952.

287

Metcalf, D., Nicola, N.A. and Gearing, D.P. (1990) Effects of injected leukemia inhibitory factor on hematopoietic and other tissues in mice. Blood 76: 50-56. Michikawa, M., Kikuchi, S. and Kim, S.U. (1992) Leukemia inhibitory factor (LIF) mediated increase of choline acetyltransferase activity in mouse spinal cord neurons in culture. Neurosci. Lett. 140: 75-77. Michishita, M., Satoh, M., Yamaguchi, M., Hirayoshi, K., Okumma, M. and Nagata, K. (1991) Phosphorylation of the stress protein hsp27 is an early event in murine myelomonocytic leukemic cell differentiation induced by leukemia inhibitory factor/D-factor. Biochem. Biophys. Res. Commun. 176: 979-984 Miles, S.A., Martinez-Maza, O., Rezai, A., Magpantay, L., Kishimoto, T., Nakamura, S., Radka, S.F. and Linsley, P.S. (1992) Oncostatin M is a potent mitogen for AIDS-kaposi's sarcoma-derived cells. Science 255: 1432-1434. Miura, O., D'Andrea, A., Kabat, D. and Ihle, J.N. (1991) Induction of tyrosine phosphorylation by the erythropoietin receptor correlates with mitogenesis. Mol. Cell. Biol. 11: 4895^902. Modrell, B., Liu, J., Miller, H. and Shoyab, M. (1994) LIF and OM directly interact with a soluble form of gpl30, the IL-6 receptor signal transduction subunit. Growth Factors 11: 81-91. Moran, C.S., Campbell, J.H., Simmons, D.L. and Campbell, G.R. (1994) Human leukemia inhibitory factor inhibits development of experimental atherosclerosis. Arterioscler. Thromb. 14: 1356-1363. Mori, M., Yamaguchi, K. and Abe, K. (1989) Purification of a lipoprotein lipase-inhibiting protein produced by a melanoma cell line associated with cancer cachexia. Biochem. Biophys. Res. Commun. 160: 1085-1092. Mori, M., Yamaguchi, K., Honda, S., Nagasaki, K., Ueda, M., Abe, O. and Abe, K. (1991) Cancer cachexia syndrome developed in nude mice bearing melanoma cells producing leukemia inhibitory factor. Cancer Res. 51: 6656-6659. Mui, A.L. and Miyajima, A. (1994) Interleukin-3 and granulocyte- macrophage colony-stimulating factor receptor signal transduction. Proc. Soc. Exp. Biol. Med. 206: 284288. Mulderry, P.K. (1994) Neuropeptide expression by newborn and adult rat sensory neurons in culture - effects of nerve growth factor and other neurotrophic factors. Neuroscience 59: 673-688. Muller, M., Briscoe, J., Laxton, C, Guschin, D., Ziemiecki, A., Silvennoinen, O., Harpur, A.G., Barbieri, G., Witthuhn, B.A., Schindler, C, Pellegrini, S., Wilks, A.F., Ihle, J.N., Stark, G.R. and Kerr, I.M. (1993) The protein tyrosine kinase Jakl complements defects in interferon-a/y? and -y signal transduction. Nature 366: 129-135. Mundy, G.R. (1989) Local factors in bone remodeling. Recent Prog. Horm. Res. 45: 507-531. Murakami, M., Narazaki, M., Hibi, M., Yawata, H., Yasukawa, K., Hamaguchi, M., Taga, T. and Kishimoto, T. (1991) Critical cytoplasmic region of the IL-6 signal trans-

288 ducer, gpl30, is conserved in the cytokine receptor family. Proc. Natl. Acad. Sci. USA 88: 11349-11353. Murphy, M., Reid., K., Hilton, D.J. and Bartlett, P.F. (1991) Generation of sensory neurons is stimulated by leukemia inhibitory factor. Proc. Natl. Acad. Sci. USA 88: 34983501. Murray, R., Lee, F. and Chiu, C.-P. (1990) The genes for leukemia inhibitory factor and interleukin-6 are expressed in mouse blastocysts prior to the onset of hemopoiesis. Mol. Cell. Biol. 10: 4953-4956. Nakajima, K., Kusafuka, T., Takeda, T., Fujitani, Y., Nakae, K. and Hirano, T. (1993) Identification of a novel interleukin 6 responsive element containing an Ets-binding site and a CRE-like site in the jun B promoter. Mol. Cell. Biol. 13:3027-3041. Nawa, H. and Patterson, P.H. (1990) Separation and partial characterization of neuropeptide-inducing factors in heart cell conditioned medium. Neuron 4: 269-277. Nawa, H., Yamamori, T., Le, T. and Patterson, P. H. (1991) The generation of neuronal diversity: analogies and homologies with hematopoiesis. Cold Spring Harbor Symp. Quant. Biol. 55: 247-253. Noda, M., Vogel, R.L., Hasson, D.M. and Rodan, G.A. (1990) Leukemia inhibitory factor suppresses proliferation, alkaline phosphatase activity, and type I collagen messenger ribonucleic acid level and enhances osteopontin mRNA level in murine osteoblast-like (MC3T3E1) cells. Endocrinology 127: 185-190. Oh, Y.J. and O'Malley, K.L. (1994) IL-6 increases choline acetyltransferase but not neuropeptide transcripts in sympathetic neurons. NeuroReport 5: 937-940. Oppenheim, R.W., Prevette, D., Yin, Q.W., Collins, F. and MacDonald, J. (1991) Control of embryonic motoneuron survival in vivo by ciliary neurotrophic factor. Science 251: 1616-1618. Oritani, K., Kaisho, T., Nakajima, K. and Hirano, T. (1992) Retinoic acid inhibits interleukin 6-induced macrophage differentiation and apoptosis in a murine hematopoietic cell line, Y6. Blood 80: 2298-2305. Patterson, P.H. (1978) Environmental determination of autonomic neurotransmitter functions. Annu. Rev. Neurosci. 1: 1-17. Patterson, P.H. (1992) The emerging neuropoietic cytokine family: first CDF/LIF, CNTF and IL-6; next ONC, MGF, GCSF? Curr. Opin. Neurobiol. 2: 94-97. Patterson, P.H. (1994) Leukemia inhibitory factor, a cytokine at the interface between neurobiology and immunology. Proc. Natl. Acad. Sci. USA 91: 7833-7835. Patterson, P.H. and Chun, L.L.Y. (1974) The influence of non-neuronal cells on catecholamine and acetylcholine synthesis and accumulation in cultures of dissociated sympathetic neurons. Proc. Natl. Acad. Sci. USA 71: 36073610. Patterson, P.H. and Chun, L.L.Y. (1977) The induction of acetylcholine synthesis in primary cultures of dissociated rat sympathetic neurons. Dev. Biol. 56: 2563-2580.

Leukemia inhibitory factor and phenotypic

specialization

Pergolizzi, R.G. and Erster, S.H. (1994) Analysis of chromosome 22 loci in meningioma. Alterations in the leukemia inhibitory factor (LIF) locus. Mol. Chem. Neuropathol. 21: 189-217. Piquet-Pellorce, C , Grey, L., Mereau, A. and Heath, J.K. (1994) Are LIF and related cytokines functionally equivalent? Exp. Cell Res. 213: 340-347. Purves, D. (1988) Body and Brain: A Trophic Theory of Neural Connections, Harvard University Press, Cambridge, MA. Rabinovsky, E.D., Smith, G.M., Browder, D.P., Shine, H.D. and McManaman, J.L. (1992) Peripheral nerve injury down- regulates CNTF expression in adult rat sciatic nerves. J. Neurosci. Res. 31: 188-192. Raff, M.C., Miller, R.H. and Noble, M. (1983) A glial progenitor cell that develops in vitro into an astrocyte or an oligodendrocyte depending on culture medium. Nature 303: 390-396. Raisz, L.G. (1988) Local and systemic factors in the pathogenesis of osteoporosis. N. Engl. J. Med. 318: 818-828. Raisz, L.G. (1993) Bone cell biology: new approaches and unanswered questions. J. Bone Min. Res. 8: S 4 5 7 ^ 6 5 . Rao, M.S. and Landis, S.C. (1993) Cell interactions that determine sympathetic neuron transmitter phenotype and the neurokines that mediate them. J. Neurobiol. 24: 215-232. Rao, M.S., Patterson, P.H. and Landis, S.C. (1992) Multiple cholinergic differentiation factors are present in footpad extracts: comparison with known cholinergic factors. Development 1X6'. 12>\-1AA. Rao, M.S., Sun, Y., Escary, J.L., Perreau, J., Tresser, S., Patterson, P.H., Zigmond, R.E., Brulet, P. and Landis, S.C. (1993a) Leukemia inhibitory factor mediates an injury response but not a target- directed developmental transmitter switch in sympathetic neurons. Neuron 11:1175-1185. Rao, M., Sun, Y., Vaidyanathan, U., Landis, S. and Zigmond, R. (1993b) Regulation of substance P is similar to that of vasoactive intestinal peptide after axotomy or explantation of the rat superior cervical ganglion. J. Neurobiol. 24: 5 7 1 580. Rathjen, P.D., Nichols, J., Toth, S., Edwards, D. R., Heath, J.K. and Smith, A.G. (1990a) Developmentally programmed induction of differentiation inhibiting activity and the control of stem cell populations. Genes Dev. 4: 23082318. Rathjen, P.D., Toth, S., Willis, A., Heath, J.K. and Smith, A.G. (1990b) Differentiation inhibiting activity is produced in matrix-associated and diffusible forms that are generated by alternate promoter usage. Cell 62: 1105-1114. Reid, I.R., Lowe, C , Cornish, J., Skinner, S.J.M., Hilton, D.J., Willson, T.A., Gearing, D.P. and Martin, T.J. (1990) Leukemia inhibitory factor: a novel bone-active cytokine. Endocrinology 126:1416-1420. Richards, C D . , Brown, T.J., Shoyab, M., Baumann, H. and Gauldie, J. (1992) Recombinant oncostatin M stimulates the production of acute-phase proteins in HepG2 cells and rat primary hepatocytes in vitro. J. Immunol. 148: 1731-1736.

T. Yamamori Robertson, M., Chambers, I., Rathjen, P., Nichols, J. and Smith, A. (1993) Expression of alternative forms of differentiation inhibiting activity (DIA/LIF) during murine embryogenesis and in neonatal and adult tissues. Dev. Genet. 14: 165-173. Robinson, R.C., Grey, L.M., Staunton, D., Vankelecom, H., Vernallis, A.B., Moreau, J.-F., Stuart, D.I., Heath, J.K. and Jones, E.Y. (1994) The crystal structure and biological function of leukemia inhibitory factor: implications for receptor binding. Cellll: 1101-1116. Rocklin, R.E. (1975) Partial characterization of leukocyte inhibitory factor by concanavalin A-stimulated human lymphocytes (LIF Con A). J. Immunol. 114: 1161-1165. Rodan, S.B., Wesolowski, G., Hilton, D.J., Nicola, N.A. and Rodan, G.A. (1990) Leukemia inhibitory factor binds with high affinity to preosteoblastic RCT-1 cells and potentiates the retinoic acid induction of alkaline phosphatase. Endocrinology 127: 1602-1608. Rohrer, H. (1992) Cholinergic neuronal differentiation factors: evidence for the presence of both CNTF-like and nonCNTF-like factors in developing rat footpad. Development 114:689-698. Rose, T.M. and Bruce, A.G. (1991) Oncostatin M is a member of a cytokine family that includes leukemia-inhibitory factor, granulocyte colony-stimulating factor and interleukin-6. Proc. Natl. Acad. Sci. USA 88: 8641-8645.. Rose, T.M., Lagrou, M.J., Fransson, I., Werelius, B., Delattre, O., Thomas, G., De Jong, P.J., Todaro, G.J. and Dumanski, J.P. (1993) The genes for oncostatin M (OSM) and leukemia inhibitory factor (LIF) are tightly linked on human chromosome 22. Genomics 17: 136-140. Rose-John, S. and Heinrich, P.C. (1994) Soluble receptors for cytokines and growth factors: generation and biological function. BiochemJ. 300: 281-290. Ruff-Jamison, S., Chen, K. and Cohen, S. (1993) Induction by EGF and interferon-y of tyrosine phosphorylated DNA binding proteins in mouse liver nuclei. Science 261: 17331739. Saadat, S., Sendtner, M. and Rohrer, H. (1989) Ciliary neurotrophic factor induces cholinergic differentiation of rat sympathetic neurons in culture. J. Cell Biol. 108: 1807-1816. Sachs, L. (1990) The control of growth and differentiation in normal and leukemic blood cells. Cancer 65: 2196-2206. Sadowski, H.B., Shuai, K., Darnell, Jr., J.E. and Gilman, M.Z. (1993) A common nuclear signal transduction pathway activated by growth factor and cytokine receptors. Science 261: 1739-1744. Schainberg, H., Borish, L., King, M., Rocklin, R.E. and Rosenwasser, L.J. (1988) Leukocyte inhibitory factor stimulates neutrophil-endothelial cell adhesion. J. Immunol. 141: 3055-3060. Schiemann, W.P. and Nathanson, N.M. (1994) Involvement of protein kinase C during activation of the mitogen-activated protein kinase cascade by leukemia inhibitory factor - evidence for participation of multiple signalling pathways. J. Biol. Chem. 269: 6376-6382.

289 Schindler, C, Shuai, K., Prezioso, V.R. and Darnell, Jr., J.E. (1992) Interferon-dependent tyrosine phosphorylation of a latent cytoplasmic transcription factor. Science 151: 809813. Schwarzschild, M.A., Dauer, W.T., Lewis, S.E., Hamill, L.K., Fink, J.S. and Hyman, S.E. (1994) Leukemia inhibitory factor and ciliary neurotrophic factor increase activated ras in a neuroblastoma cell line and in sympathetic cultures. J. Neurochem. 63: 1246-1254. Sendtner, M., Kreutzberg, G.W. and Thoenen, H. (1990) Ciliary neurotrophic factor prevents the degeneration of motoneurons after axotomy. Nature 345: 440^41. Sendtner, M., Stockli, K.A. and Thoenen, H. (1992) Synthesis and localization of ciliary neurotrophic factor in the sciatic nerve of the adult rat after lesion and during regeneration. J. Cell Biol. 118: 139-148. Seniuk, N., Altares, M., Dunn, R. and Richardson, P.M. (1992) Decreased synthesis of ciliary neurotrophic factor in degenerating peripheral nerves. Brain Res. 572: 300302. Shabo, Y., Lotem, J., Rubinstein, M., Revel, M., Clark, S.C, Wolf, S.F., Kamen, R. and Sachs, I. (1988) The myeloid blood differentiation-inducing protein MGI-2A is interleukin-6. Blood 12: 2070-2073. Shadiack, A.M., Hart, R.P., Carlson, CD. and Jonakait, G.M. (1993) Interleukin-1 induces substance P in sympathetic ganglia through the induction of leukemia inhibitory factor (LIF). /. Neurosci. 13: 2601-2609. Shen, M.M. and Leder, P. (1992) Leukemia inhibitory factor is expressed by the preimplantation uterus and selectively blocks primitive ectoderm formation in vitro. Proc. Natl. Acad. Sci. USA 89: 8240-8244. Shen, M.M., Skoda, R.C., Cardiff, R.D., Campos-Torres, J., Leder, P. and Ornitz, D.M. (1994) Expression of LIF in transgenic mice results in altered thymic epithelium and apparent interconversion of thymic and lymph node morphologies. EMfiO 7. 13:1375-1385. Shuai, K., Schindler, C, Prezioso, V.R. and Darnell, Jr., J.E. (1992) Activation of transcription by IFN-y. Tyrosine phosphorylation of a 91-kD DNA binding protein. Science 258: 1808-1812. Shuai, K., Stark, G.R., Kerr, I.M. and Darnell, Jr., J.E. (1993) A single phosphotyrosine residue of Stat91 required for gene activation by interferon-/. Science 261: 1744-1746. Silvennoinen, O., Witthuhn, B., Quelle, F.W., Cleveland, J.L., Yi, T. and Ihle, J.N. (1993) Structure of the murine Jak2 protein-tyrosine kinase and its role in interleukin 3 signal transduction. Proc. Natl. Acad. Sci. USA 90: 8429-8433. Skoff, R.P. (1990) Gliogenesis in rat optic nerve: astrocytes are generated in a single wave before oligodendrocytes. Dev. Biol. 139: 149-168. Smith, A.G., Heath, J.K., Donaldson, D.D., Wong, G.G., Moreau, J., Stahl, M. and Rogers, D. (1988) Inhibition of pluripotential embryonic stem cell differentiation by purified polypeptides. Nature 336: 688-690. Smith, A.G., Nichols, J., Robertson, M. and Rathjen, P.D.

290 (1992) Differentiation inhibiting activity (DIA/LIF) and mouse development. Dev. Biol. 151: 339-351. Smith, D.K., Treutlein, H.R., Maurer, T., Owczarek, CM., Layton, M.J., Nicola, N.A. and Norton, R.S. (1994) Homology modelling and IHNMR studies of human leukaemia inhibitory factor. FEBS Lett. 350: 275-280. Stahl, J. and Gough, N.M. (1993) Delineation of positive and negative control elements within the promoter region of the murine leukemia inhibitory factor (LIF) gene. Cytokine 5: 386-393. Stahl, N. and Yancopoulos, G.D. (1993) The alphas, betas, and kinases of cytokine receptor complexes. Cell 74: 587590. Stahl, N. and Yancopoulos, G.D. (1994) The tripartite CNTF receptor complex: activation and signaling involves components shared with other cytokines. J. Neurohiol. 25: 1454^1466. Stahl, J., Gearing, D.P., Willson, T.A., Brown, M.A., King, J.A. and Gough, N.M. (1990) Structural organization of the genes for murine and human leukemia inhibitory factor. J. Biol. Chem. 265: 8833-8841. Stahl, N., Boulton, T.G., Farruggella, T., Ip, N.Y., Davis, S., Witthuhn, B.A., Quelle, F.W., Silvennoinen, O., Barbieri, G., Pellegrini, S., Ihle, J.N. and Yancopoulos, G.D. (1994) Association and activation of Jak/Tyk kinases by CNTF/LIF/OSM/IL-6 y3 receptor components. Science 263: 92-95. Stahl, N., Farruggella, T.J., Boulton, T.G., Zhong, Z., Darnell, Jr., J.E. and Yancopoulos, G.D. (1995) Choice of STATs and other substrates specified by modular tyrosine-based motifs in cytokine receptors. Science 267: 1349-1353. Stewart, C.L., Kaspar, P., Brunet, L.J., Bhatt, H., Gadi, I., Kontgen, F. and Abbondanzo, S.J. (1992) Blastocyst implantation depends on maternal expression of leukaemia inhibitory factor. Nature 359: 76-79. Stockli, K.A., Lillien, L.E., Naher-Noe, M., Breitfeld, G., Hughes, R.A., Raff, M.C., Thoenen, H. and Sendtner, M. (1991) Regional distribution, developmental changes, and cellular localization of CNTF-mRNA and protein in the rat brain. J. Cell Biol 115: 447-459. Suda, T., Takahashi, N. and Martin, T.J. (1992) Modulation of osteoclast differentiation. Endocrine Rev. 13: 66-80. Suematsu, S., Matsuda, T., Aozasa, K., Akira, S., Nakano, N., Ohno, S., Miyazaki, J., Yamamura, K., Hirano, T. and Kishimoto, T. (1989) IgGl plasmacytosis in interleukin 6 transgenic mice. Proc. Natl. Acad. Sci, USA 86: 7547-7551. Suematsu, S., Matsusaka, T., Matsuda, T., Ohno, S., Miyazaki, J., Yamamura, K., Hirano, T. and Kishimoto, T. (1992) Generation of plasmacytomas with the chromosomal translocation t(12;15) in interleukin 6 transgenic mice. Proc. Natl. Acad. Sci. USA 89: 232-235. Sun, Y., Rao, M.S., Zigmond, R.E. and Landis, S.C. (1994) Regulation of vasoactive intestinal peptide expression in sympathetic neurons in culture and after axotomy - the role of cholinergic differentiation factor/leukemia inhibitory factor. J. Neurobiol. 25: 415^30.

Leukemia inhibitory factor and phenotypic specialization Sutherland, G.R., Baker, E., Hyland, V.J., Callen, D.F., Stahl, J. and Gough N.M. (1989) The gene for human leukemia inhibitory factor (LIF) maps to 22ql2. Leukemia 3: 9-13. Taga, T., Hibi, M., Hirata, H., Yawata, S., Natsuka, K., Yasukawa, T., Totsuka, K., Yamasaki, T., Hirano, T. and Kishimoto, T. (1989) Interleukin-6 receptor and a unique mechanism of its signal transduction. Cold Spring Harb. Symp. Quant. Biol. 54 (Part 2): 713-722. Taga, T. and Kishimoto, T. (1992) Cytokine receptors and signal transduction. FASEB J. 6: 3387-3396. Takahashi, R., Yokoji, H., Misawa, H., Hayashi, M., Hu, J. and Deguchi, T (1994) A null mutation in the human CNTF gene is not causally related to neurological diseases. Nature Genet. 7: 79-84. Tamura, T., Udagawa, N., Takahashi, N., Miyaura, C, Tanaka, S., Yamada, Y., Koishihara, Y., Ohsugi, Y., Kumaki, K., Taga, T., Kishimoto, T. and Suda, T. (1993) Soluble interleukin-6 receptor triggers osteoclast formation by interleukin-6. Proc. Natl. Acad. Sci. USA 90: 1192411928. Tanigawa, T., Elwood, N., Metcalf, D., Cary, D., DeLuca, E., Nicola, N.A. and Begley, C.G. (1993) The SLC gene product is regulated by and differentially regulates cytokine responses during myeloid leukemic cell differentiation. Proc. Natl. Acad. Sci. USA 90: 7864-7868. Taupin, J.-L., Morel, D., Moreau, J.-F., Gualde, N., Potaux, L. and Bezian, J.-H. (1992) HILDA/LIF urinary excretion during acute kidney rejection. Transplantation 53: 65. Thaler, CD., Suhr, L., Ip, N. and Katz, D.M. (1994) Leukemia inhibitory factor and neurotrophins support overlapping populations of rat nodose sensory neurons in culture. Dev. Biol 161:338-344. Thoenen, H. (1991) The changing scene of neurotrophic factors. Trends NeuroscL 14: 165-170. Thoma, B., Bird, T.A., Friend, D.J., Gearing, D.P. and Dower, S.K. (1994) Oncostatin M and leukemia inhibitory factor trigger overlapping and different signals through partially shared receptor complexes. J. Biol Chem. 269: 62156222. Tomida, M., Yamamoto-Yamaguchi, Y. and Hozumi, M. (1984) Purification of a factor inducing differentiation of mouse myeloid leukemic Ml cells from conditioned medium of mouse fibroblast L929 cells. J. Biol Chem. 259: 10978-10982. Tomida, M., Yamamoto-Yamaguchi, Y. and Hozumi, M. (1993) Pregnancy associated increase in mRNA for soluble D-factor/ LIF receptor in mouse liver. FEBS Lett. 334: 193197. Tomida, M., Yamamoto-Yamaguchi, Y. and Hozumi, M. (1994) Three different cDNAs encoding mouse Dfactor/LiF receptor, y. Biochem. 115: 557-562. Tsan, M.F., White, J.E. and Wong, G.H. (1992) D-factor and growth hormone enhance tumor necrosis factor-induced increase of Mn superoxide dismutase mRNA and oxygen tolerance. Cytokine A\ 101-105. Ure, D.R. and Campenot R.B. (1994) Leukemia inhibitory

T. Yamamori factor and nerve growth factor are retrogradely transported and processed by cultured rat sympathetic neurons. Dev. Biol. 162: 339-347. Ure, D.R., Campenot, R.B. and Acheson, A. (1992) Cholinergic differentiation of rat sympathetic neurons in culture: Effects of factors applied to distal neurites. Dev. Biol. 154: 388-395. Velazquez, L., Fellous, M., Stark, G.R. and Pellegrini, S. (1992) A protein tyrosine kinase in the interferon a/B signaling pathway. Cell 70: 313-322. Wang, Y. and Fuller, G.M. (1994) Phosphorylation and internalization of gpl30 occur after IL-6 activation of Jak2 kinase in hepatocytes. Mol. Biol. Cell 5: 819-829. Waring, P., Wycherley, K., Gary, D., Nicola, N. and Metcalf, D. (1992) Leukemia inhibitory factor levels are elevated in septic shock and various inflammatory body fluids. J. Clin. Invest. 90: 2031-2037. Waring, P.M., Waring, L.J. and Metcalf, D. (1994a) Circulating leukemia inhibitory factor levels correlate with disease severity in meningococcemia. /. Infect. Dis. 170: 12241228. Waring, P.H., Romero, R., Laham, N., Gomez, R. and Rice, G.R. (1994b) Leukemia inhibitory factor: association with intraamniotic infection. Am. J. Obstet. Gynecol. 171: 13351341. Weber, M.J. (1981) A diffusible factor responsible for the determination of cholinergic functions in cultured sympathetic neurons: partial purification and characterization. /. Biol. Chem. 256: 3447-3453. Wegenka, U.M., Buschmann, J., Liitticken, C., Heinrich, P.C. and Horn, F. (1993) Acute-phase response factor, a nuclear factor binding to acute-phase response elements, is rapidly activated by interleukin-6 at the posttranslational level. Mol Cell. Biol 13: 276-288. Wegenka, U.M., Liitticken, C, Buschmann, J., Yuan, J., Lottspeich, F., Muller-Esterl, W., Schindler, C, Roeb, E., Heinrich, P.C. and Horn, F. (1994) The interleukin-6activated acute-phase response factor is antigenically and functionally related to members of the signal transducer and activator of transcription (STAT) family. Mol Cell Biol 14:3186-3196. Wesselingh, S.L., Levine, B., Fox, R.J., Choi, S. and Griffin D.E. (1994) Intracerebral cytokine messenger RNA expression during fatal and nonfatal alphavirus encephalitis suggests a predominant type 2 T cell response. J. Immunol 152: 1289-1297. Wetzler, M., Estrov, Z., Talpaz, M., Kim, K.J., Alphonso, M., Srinivasan, R. and Kurzrock, R. (1994) Leukemia inhibitory factor in long-term adherent layer cultures - increased levels of bioactive protein in leukemia and modulation by IL4, IL-1 beta, and TNF-alpha. Cancer Res. 54: 1837-1842. Wilks, A.F. (1989) Two putative protein-tyrosine kinases identified by application to the polymerase chain reaction. Proc. Natl Acad. ScL USA 86: 1603-1607. Williams, R.L., Hilton, D.J., Pease, S., Willson, T.A., Stewart, C.L., Gearing, D.P., Wagner, E.F., Metcalf, D., Nicola,

291 N.A. and Gough, N.M. (1988) Myeloid leukaemia inhibitory factor maintains the developmental potential of embryonic stem cells. Nature 336: 684-687. Witthuhn, B.A., Quelle, F.W., Silvennoinen, O., Yi, T., Tang, B., Miura, O. and Ihle, J.N. (1993) JAK2 associates with the erythropoietin receptor and is tyrosine phosphorylated and activated following stimulation with erythropoietin. Cell 74: 227-236. Witthuhn, B.A., Silvennoinen, O., Miura, K., Lai, K.S., Cwik, C, Liu, E.T. and Ihle, J.N. (1994) Involvement of the Jak-3 Janus kinase in signaling by interleukin 2 and 4 in lymphoid and myeloid cells. Nature 370: 153-157. Wong, G.H.W., Neta, R. and Goeddel, D.V. (1992) Protective roles of MnSOD, TNF-a, TNF-/3 and D-factor (LIF) in radiation injury. In: K. Steglitz (Ed.), Eicosanoids and Other Bioactive Lipids in Cancer, Inflammation and Radiation Injury, 2nd International Conference, Berlin, pp. 353358. Wong, v., Arriaga, R., Ip, N.Y. and Lindsay, M. (1993) The neurotrophins BDNF, NT-3 and NT-4/5, but not NGF, upregulate the cholinergic phenotype of developing motor neurons. Eur. J. NeuroscL 5: 466^74. Yamamori, T. (1991a) CDF/LIF selectively increases c-fos and jun-B transcripts in sympathetic neurons. NeuroReport 2: 173-176. Yamamori, T. (1991b) Localization of CDF/LIF mRNA in the rat brain and peripheral tissues. Proc. Natl Acad. ScL USA 88:7298-7302. Yamamori, T. (1992) Molecular mechanisms for generation of neural diversity and specificity: roles of polypeptide factors in development of postmitotic neurons. NeuroscL Res. 12: 545-548. Yamamori, T., Fukada, K., Aebersold, R., Korshing, S., Fann, M.J. and Patterson, P.H. (1989) The cholinergic neuronal differentiation factor from heart cells is identical to leukemia inhibitory factor. Science 246: 1412-1416. Yamasaki, K., Taga, T., Hirata, Y., Yawata, H., Kawanishi, Y., Seed, B., Taniguchi, T., Hirano, T. and Kishimoto, T. (1988) Cloning and expression of the human interleukin-6 (BSF-2/IFNb2) receptor. Science 241: 825-828. Yang, Y.-C. (1993) Interleukin 11: an overview. Stem Cells 11:474-486. Yang, Z.-M., Le, S.-P., Chen, D.-B. and Harper, M.J.K. (1994) Temporal and spatial expression of leukemia inhibitory factor in rabbit uterus during early pregnancy. Mol Reprod. Dev. 38: 148-152. Yin, T.G. and Yang, Y.C. (1994) Mitogen-activated protein kinases and ribosomal-S6 protein kinases are involved in signaling pathways shared by interleukin-U, interleukin-6, leukemia inhibitory factor, and oncostatin M in mouse 3T3Ll cells. J. Biol Chem. 269: 3731-3738. Yin, T., Yasukawa, K., Taga, T., Kishimoto, T. and Yang, Y.C. (1994) Identification of a 130-kilodalton tyrosinephosphorylated protein induced by interleukin-41 as JAK2 tyrosine kinase, which associates with gpl30 signal transducer. Exp. Hematol 22: 467-472.

292 Yuan, J., Wegenka, U.M., Liitticken, C, Buschmann, J., Decker, T., Schindler, C, Heinrich, P.C. and Horn, F. (1994) The signalling pathways of interleukin-6 and gamma interferon converge by the activation of different transcription factors which bind to common responsive DNA elements. Mol Cell Biol 14: 1657-1668. Zhang, X.G., Gu, J.J., Lu, Z.Y., Yasukawa, K., Yancopoulos, G.D., Turner, K., Shoyab, M., Taga, T., Kishimoto, T., Bataille, R. and Klein, B. (1994) Ciliary neurotrophic factor, interleukin 11, leukemia inhibitory factor, and oncostatin M are growth factors for human myeloma cell lines using the interleukin 6 signal transducer gpl30. J. Exp. Med. 179: 1337-1342. Zhong, Z., Wen, Z. and Darnell, Jr., J.E. (1994a) Stat3: a Stat family member activated by tyrosine phosphorylation in re-

Leukemia inhibitory factor and phenotypic specialization sponse to epidermal growth factor and interleukin-6. Science 264: 95-98. Zhong, Z., Wen, Z. and Darnell, Jr., J.E. (1994b) Stat3 and Stat4: members of the family of signal transducers and activators of transcription. Proc. Natl Acad. ScL USA 91: 4806-4810. Zigmond, R., Hyatt-Sachs, H., Baldwin, C, Qu, X., Sun, Y., McKeon, T., Schreiber, R. and Vaidyanathan, U. (1992) Phenotypic plasticity in adult sympathetic neurons: changes in neuropeptide expression in organ culture. Proc. Natl Acad. ScL USA 89: 1507-1511. Zurn, A.D. and Werren, F. (1994) Development of CNS cholinergic neurons in vitro: selective effects of CNTF and LIF on neurons from mesencephalic cranial motor nuclei. Dev. Biol 163: 309-315.

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 12

Ciliary neurotrophic factor P.M. Richardson and M.C. Subang Division of Neurosurgery, McGill University and Montreal General Hospital, 1650 Cedar Avenue, Montreal, Canada H3G 1A4

1. Historical notes Ciliary neurotrophic factor (CNTF), now known to have a wide spectrum of target neurons, was discovered and named for its actions on parasympathetic neurons of cihary gangUa from chick embryos. Chick cihary gangha were chosen for study by developmental biologists because they are surgically accessible and contain several thousand neurons, as compared to several hundred for rodent ciliary ganglia. During the second week of development of the chick embryo, approximately one-half of the ciliary parasympathetic neurons that innervate the iris, choroid, and ciliary body die (Landmesser and Pilar, 1974). It is generally believed that ciliary neurons die during embryogenesis in competition for limiting supplies of growth factors from glial and target cells. In search of such trophic agents, culture systems were devised in the 1970s to study parasympathetic neurons from ciliary ganglia of chick embryos, either within explanted ganglia, or after dissociation into individual neurons (Helfand et al., 1976; Nishi and Berg, 1977; Ebendal et al., 1978). Neurons from Eg-Eio ciliary ganglia (removed at the eighth to tenth embryonic day) die within 24 h in culture under basal conditions but survive and extend neurites in the presence of co-explants, conditioned medium, or tissue extracts from several sources. Supportive agents from chick heart, bovine heart, ciliary body and iris were soon characterized as proteins with molecular weights of 20 000-50 000 (Ebendal et al., 1979; Bonyhady et al., 1980; Manthorpe et al., 1986) active on sensory, sympathetic and parasympathetic neurons

from embryonic chick. In early purification studies, the tissues chosen as sources of CTNF were ocular tissues and heart muscle, which are normally innervated by parasympathetic nerves. Subsequently, ciliary neurotrophic bioactivity was discovered to be even more concentrated in peripheral mixed nerves (Richardson and Ebendal, 1982; Williams et al., 1984). By the mid-1980s, substantial purifications had been reported from chick ocular tissues and rat peripheral nerves of acidic proteins of slightly more than 20 kDa with ciliary neurotrophic activity (Barbin et al., 1984; Manthorpe et al., 1986). The use of bioassay to trace bioactive fractions during purification was facilitated by the stability of CNTF even after sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). In 1989, deduced amino acid sequences were reported for rabbit and rat CNTF after purification and partial amino acid sequencing of sciatic nerve proteins, followed by cloning and sequencing of cDNA clones (Lin et al., 1989; Stockli et al., 1989). With this sequence information, it became possible to synthesize milligram quantities of bioactive CNTF in bacterial expression systems (Masiakowski et al., 1991; Negro et al., 1991; Gupta et al, 1992) for biological studies and raising of antibodies and to prepare DNA and RNA probes for analysis of CNTF synthesis. 2. Structure of CNTF CNTF cDNAs have been cloned and sequenced for three mammalian species: rat, rabbit, and human (Lin et al., 1989; Stockh et al., 1989; Lam et al., 1991; Masiakowski et al., 1991; Negro et al..

294

1991). Rabbit CNTF has 199 amino acids; human and rat CNTF have 200 amino acids. Amino-acid sequences of CNTF in the three species are approximately 85% identical. Partial amino acid sequencing was obtained for a ciliary neurotrophic factor from chick sciatic nerve (Eckenstein et al., 1990) and the sequence data was used to clone GPA cDNA from an E15 chick eye library (Leung et al., 1992). The corresponding protein has 195 amino acids and shows 50% homology to sequences for rat, human, and rabbit CNTF. The lack of closer homology could represent an interspecies difference or, perhaps, the existence of a family of CNTFs. However, no other members of a possible CNTF family have yet been detected in mammals. CNTF, like acidic and basic fibroblast growth factor, interleukin (IL)-l, and platelet-derived growth factors, lacks a hydrophobic signal peptide which mediates conventional release of proteins from their cells of synthesis. Accordingly, when mammalian CNTF cDNA is inserted into an expression vector and expressed in mammalian cells, no CNTF bioactivity is secreted into the culture medium (Lin et al, 1989; Stockli et al., 1989). In contrast, 50% of recombinant chick CNTF is released from mammalian cells, despite the absence of a signal peptide in the chick CNTF sequence (Leung etal., 1992). In its primary sequence, CNTF does not bear strong homology with any other known protein. Initially, computer-based 'profile analysis' was used to predict that CNTF, leukemia inhibitory factor (LIF), oncostatin M, IL-6, and granulocyte colony-stimulating factor all have a common tertiary framework, incorporating four a-helices found in growth hormone (Bazan, 1991). This structure was confirmed by X-ray crystallography for LIF (Robinson et al, 1994) and for CNTF (McDonald et al., 1995). Multi-wavelength anomalous difraction (MAD) data reveal that CNTF is a dimer with anti-parallel arrangement different from that of other dimeric helical cytokines. Twleve amino acid residues scattered among the four helices and located within the bundle core are highly conserved among the cytokines which use gpl30 as a signal transducing sub-unit. (McDonald et al., 1995). By mutagenesis, a cluster of three arginine residues

Ciliary neurotrophic factor

(Panayotatus et al., 1995) and a glutamine residue within helix D (Saggio et al., 1995) have been implicated separately in binding of CNTF to the CNTFa receptor. Substitution of two amino acids in or near helix D decreases biological activity (Panayotatus et al., 1995). The binding site of CNTF for gpl30 includes residues in helix A (Panayotatus et al., 1995) and probably in other helixes as for IL-6 (Savino et al., 1994). Truncations at N- or C-termini of CNTF which do not affect the putative helices are not detrimental to CNTF bioactivity (Their et al., 1995). In fact, the crystallized molecule is a human CNTF lacking the C-terminal 13 residues that has similar activity to the untruncated form (McDonald et al., 1995). Similarities or predicted similarities in tertiary structures of CNTF, LIF, oncostatin-M, IL-6, and IL-11 are reflected by overlap in biological actions and in components of their receptor complexes (Gearing et al, 1992; Ip et al., 1992b). The CNTF gene has been mapped to human chromosome 11 (Lam et al., 1991) and mouse chromosome 19 (Kaupmann et al., 1991). Human and rat CNTF genes contain a single 1 kb intron within the coding domain (Lam et al., 1991; Carroll et al., 1993). CNTF transcripts are 1.2 and 4.3 kb in rat and rabbit, respectively. Approximately 0.5 kb of sequence 5' to the translation start site of the human and rat CNTF genes have been obtained and found to contain a G-rich region, AP1 binding site and TATA box (Carroll et al., 1993). However, the upstream promoter region of CNTF has not been analyzed in sufficient detail to determine which response elements and transcriptional factors regulate CNTF transcription in Schwann cells and astrocytes. 3. Sources and synthesis of CNTF In contrast to the widespread distribution of CNTF bioactivity in peripheral organs of rat and chicken (Ebendal, 1987), CNTF immunoreactivity appears to be restricted to non-neuronal cells of the peripheral nervous system (PNS) and central nervous system (CNS) (Stockli et al., 1991; Dobrea et al., 1992; Friedman et al., 1992; Rohrer, 1992; Sendtner et al., 1992b). The most probable explanation

P.M. Richardson and M.C. Subang

for this lack of correlation is that FGFs, which mimic CNTF in their actions on ciliary neurons (Unsicker et al., 1987), are present in many tissues. For example, one-half of the ciliary neurotrophic bioactivity in rat sciatic nerves (Eckenstein et al., 1991; Gupta et al., 1992) and most of the activity in bovine heart (Hill et al., 1991) are attributable to acidic fibroblast growth factor (aFGF). The concentration of CNTF in peripheral nerves has been estimated to be in the order of 5 nmol/kg (Gupta et al., 1992), 100-fold higher than concentrations of nerve growth factor (NGF) or other neurotrophins in nervous tissue and one thousandfold higher than would be necessary for halfmaximal biological activity if CNTF were homogeneously distributed. The supraphysiological concentration of CNTF may reflect the fact that CNTF is trapped in its cells of synthesis and so is not readily released into the extracellular space to reach putative neuronal receptors. In peripheral nerves, the synthesis of CNTF, like that of myelin proteins, appears to be regulated by direct or indirect signals from axons. During development, CNTF mRNA and protein are first detected during the first postnatal week (Stockli et al., 1989; Dobrea et al., 1992). In peripheral nerves of adult rats, CNTF is synthesized by some but not all Schwann cells (Dobrea et al., 1992; Friedman et al., 1992; Rende et al., 1992; Sendtner et al., 1992b) and is more abundant in myelinated than unmyelinated nerves (Dobrea et al., 1992; Friedman et al., 1992). Following peripheral nerve transection, CNTF immunoreactivity, mRNA, and bioactivity are substantially reduced in the distal nerve stump (Friedman et al., 1992; Rao et al., 1992b; Sendtner et al., 1992b; Seniuk et al., 1992). During axonal regeneration and remyelination after nerve crush, the concentration of CNTF mRNA returns towards normal. The fluctuations of CNTF mRNA concentrations in normal, degenerating, and regenerating nerves are very similar to changes in mRNAs for myelin proteins and suggest that direct or indirect signals from axons act on the CNTF gene. However, the molecular signals from neurons and other cells that influence CNTF and myelin protein synthesis in Schwann cells have not yet been defined.

295

CNTF protein and mRNA are also present in the CNS, although at lower concentrations than in the PNS. Highest concentrations of CNTF mRNA within the CNS are found in the optic nerve and olfactory bulb, where CNTF immunoreactivity is restricted to GFAP-immunopositive astrocytes (Stockli et al., 1991). The concentration of CNTF mRNA is apparently below the threshold of detection in brains of prenatal rats (Stockli et al., 1991) and increases rapidly during the second postnatal week. As estimated by concentrations of CNTF mRNA, the rate of synthesis of CNTF is modest in the normal mature brain, but is up-regulated by mechanical injury (Ip et al., 1992c; Asada et al., 1995). These observations lend some support to the suggestion that CNTF is an injury factor in the CNS. Astrocytes cultured from neonatal rat brain constitutively express CNTF mRNA at concentrations similar to that present in the olfactory bulb in vivo (Nagao et al., 1995). Removal of serum from the culture medium does not significantly affect the levels of CNTF mRNA in astrocytes (Rudge et al., 1992). Treatment with forskolin, other cAMPlinked agonists and members of the fibroblast growth factor family reduce CNTF mRNA (Carroll et al., 1993; Nagao et al., 1995) while gamma-interferon increases it (Carroll et al., 1993). The rat CNTF mRNA is quite stable, with a half-life of 6-7 h (Carroll et al., 1993; Nagao et al., 1995). In comparison to the accumulated knowledge regarding the synthesis of CNTF, understanding of the release of CNTF is more limited. As noted previously, CNTF lacks a signal peptide, is poorly secreted and usually remains within its cell of synthesis. However, a fraction of cytoplasmic CNTF must be released from synthesizing cells if CNTF is to reach receptors on target cells. Observations regarding the release of basic fibroblast growth factor (bFGF) and IL-1 despite the absence of signal peptides may be pertinent to the release of CNTF. Some bFGF is localized by immunohistochemistry to the extracellular matrix (Gonzalez et al., 1990), whereas CNTF immunoreactivity is strictly intracellular (Sendtner et al., 1992b). The existence of a secreted FGF species, FGF-5 (Zhan et al., 1988; Hughes et al., 1993) raises the possi-

296

bility that a secreted analogue of CNTF remains to be discovered. However, even if a secretable member of the CNTF family is discovered, the function of poorly secreted classical CNTF would still require explanation. As much as 50% of the lL'l/3 synthesized by activated monocytes is released into the medium, where it is readily detected by conventional methods (Hazuda et al., 1988). JL-ip is released from monocytes through a process closely linked with cleavage of the mature 17 kDa IL-1^ from a 31 kDa precursor by a recently characterized protease (Cerretti et al., 1992; Thornberry et al., 1992). The mechanism of secretion of IL-ly8 differs from the conventional transport and secretion in being enhanced, rather than blocked, by brefeldin-A and momensin (Rubartelli et al., 1990; Thornberry et al., 1992). This novel pathway of secretion of IL-1)3 has been speculated to depend on the multigene resistance glycoprotein, to avoid the oxidizing milieu of the endoplasmic reticulum, and to segregate ligand and receptor, thereby preventing autocrine stimulation. Finally, it has been noted that FGFs are released from cells subjected to sublethal mechanical or thermal stimulus (McNeil et al., 1989; Jackson et al., 1992), and it was suggested that this release from injured cells might serve to accelerate wound repair. As noted earlier in this Section, release of a small fraction of the total CNTF into the extracellular spaces would provide adequate concentration for receptor activation. However, the mechanism by which any CNTF is released remains enigmatic. 4. Actions of CNTF 4. L Survival ofPNS neurons In vitro, CNTF promotes the survival of at least some members of all classes of peripheral neurons: parasympathetic, sympathetic, sensory, and motor. To prove that CNTF can act on neurons directly rather than indirectly through non-neuronal cells, individual parasympathetic ciliary neurons were plated in microwells; survival at 1 day was increased from 15 to 76% by CNTF (Unsicker et al., 1987). The demonstration of high-affinity CNTF

Ciliary neurotrophic factor

binding sites on chick ciliary neurons (Richardson et al., 1993) is further evidence for a direct neuronal response to CNTF. The actions of CNTF on dorsal root ganglia neurons and sympathetic neurons are influenced by the age and species of the donor embryos. CNTF acts on dorsal root ganglion neurons from mice embryos and Ejo but not Eg chick embryos (Barbin et al., 1984; Manthorpe et al., 1986): it has little survival effect on sympathetic neurons from newborn rats (Saadat et al., 1989) but does promote survival of sympathetic neurons from postnatal rats (Kotzbauer et al., 1994) and Ejo chick embryos (Manthorpe et al., 1986). CNTF supports the survival of spinal motoneurons in vitro. When dissociated cells from Eg chick spinal cord were enriched for motoneurons by density centrifugation (Arakawa et al., 1990), CNTF increased survival twelvefold and the combination of FGF and CNTF increased survival twentyfold. The combination of CNTF plus brain derived neurotrophic factor (BDNF) or NT-3 also was more effective than any single agent in promoting motoneurons survival in vitro (Kato and Lindsay, 1994). On the other hand, when motoneurons were purified by 'panning' with a monoclonal antibody, no increase in survival was afforded by CNTF (Bloch-Gallego et al., 1991). One possible explanation of this latter negative result is that an insulin-like growth factor or some other constituent of serum is necessary together with CNTF for motoneuron survival. Delivery of exogenous CNTF has been shown to reduce the death of motoneurons during development, after axotomy, or in mutant mice. Daily application of 2-20/^g on the chorio-allantoic membrane from day 6 to 10 of chick embryo development prevented approximately one-half of the normal developmental death of motoneurons in the spinal cord (Oppenheim et al., 1991). Ironically, in the same animals, CNTF failed to reduce developmental death of neurons in ciliary ganglia. CNTF also had no effect on the numbers of neurons in sympathetic ganglia or dorsal root ganglia. Retrograde neuronal death in the facial nerve following axotomy in newborn rats was reduced significantly by application of a pledget soaked in 5/ig CNTF to the stump of the transected nerve

P.M. Richardson and M.C. Subang

(Sendtner et al, 1990): counts of surviving motoneurons were increased from 20 to 80% of normal. The efficacy of this neuronal protection was questioned in another laboratory (Clatterbuck et al., 1994) where it was noted that the biological halflife of CNTF delivered in this way was much less than 24 h. Also, 5 jug CNTF applied to the nerve stump failed to abrogate the death of 90% of hypoglossal neurons 1 week after nerve transection in rats 1 week old (Grothe and Unsicker, 1992) although the combination of CNTF and NGF did rescue some axotomized hypoglossal neurons. In facial motor nuclei of rats axotomized at 2 weeks of age, neuronal survival was increased from 25 to 50% by continuous administration of CNTF over a narrow therapeutic range (Zhang et al., 1995). In this experimental preparation, CNTF mitigated neuronal death if delivered intrathecally but not extrathecally and was uniformly lethal if administered at a slightly higher dose. Protection by CNTF of axotomized preganglionic sympathetic neurons also has been reported (Blottner et al., 1989). As a third example of trophic effects of CNTF on motoneurons in vivo, peritoneal implantation of a CNTF-secreting cell line in mutant pmn/pmn mice mitigated the deterioration of motor function, loss of facial motoneurons, and loss of phrenic nerve axons that normally occur during the second 3 weeks of life (Sendtner et al., 1992a). Intermittent systemic administration of CNTF also slowed motoneuron degeneration in wobbler mice (Mitsumoto et al., 1994a; Ikeda et al., 1995), the effect being additive to the effects of BDNF (Mitsumoto et al., 1994b). The cellular distribution of CNTF mRNA in astrocytes and Schwann cells plus the responses to CNTF of motoneurons in vitro and in vivo suggest that CNTF might be a neurotrophic factor during development, after injury or disease, or for homeostasis. The hypothesis that CNTF has a developmental neurotrophic function is not supported by the difficulty in detecting CNTF mRNA in prenatal animals (Stockli et al., 1991) although much lower concentrations of CNTF mRNA than are found in the mature nervous system might yield sufficient protein for biological activity. The hypothesis that CNTF is an 'injury factor' in the PNS

297

is not supported by observations that its synthesis is decreased after nerve injury, whereas LIF and IL-6 are induced (Aloisi et al., 1994; Curtis et al., 1994; Murphy et al., 1995). Nevertheless, administration of exogenous CNTF has been shown to accelerate axonal regeneration and/or reinnervation of injured motoneurons (Sahenk et al., 1994; Ulenkate et al., 1994) and to counteract developmental elimination of multiple neuromuscular synapses (EngHsh and Schwartz, 1995). Some of these actions of CNTF may be on muscle cells rather than purely on neurons (Helgren et al., 1994). The hypothesis that CNTF maintains motoneurons in mature animals is supported by the finding of accelerated motoneuronal death during aging of mice after deletion of the CNTF gene by homologous recombination (Masu et al., 1993). Decreased CNTF immunoreactivity has been detected in the ventral horn of the spinal cord of patients with amyotrophic lateral sclerosis (Anand et al., 1995) although no clinical evidence of motoneuron dysfunction has been observed in patients with mutations of the CNTF gene (Takahashi et al., 1994). Because of actions of CNTF on animal motoneurons in vitro and in vivo, clinical trials were undertaken in patients with amyotrophic lateral sclerosis (Barinaga, 1994; but see also Section 4.6). CNTF is involved in the generation as well as the survival of PNS neurons. In chick embryos at an early stage of development, CNTF inhibits the proliferation of neuronal precursor cells in sympathetic ganglia (Ernsberger et al., 1989). In arresting the proliferation of precursor cells and promoting the generation of neurons, CNTF again resembles LIF, which has these two actions on neuronal precursor cells in the neural crest (Murphy et al., 1991) and the spinal cord (Richards et al., 1992). 4.2. Survival of CNS neurons CNTF augments the survival in vitro of neurons from several populations in the embryonic CNS. When hippocampal cells from Eig rat brains were maintained in culture for 1 week, CNTF increased the survival of cholinergic GABAergic, and cal-

298

bindin-immunopositive cells (Ip et al., 1991). Because these effects were abrogated when addition of CNTF was delayed for 3 days, it was concluded that CNTF promotes survival rather than merely inducing phenotypic differentiation. When retinal ganglion cells from Eio chick embryos were purified by panning on an antibody against Thy-1 and cultured at low density in the absence of serum for 24 h (Lehwalder et al., 1989), survival was increased by CNTF from 3 to 49%. CNTF also maintains neurons from the embryonic cerebellum (Larkfors et al., 1994), cortex (Magal et al., 1991a), brainstem (Magal et al., 1991a), and spinal cord (Magal et al., 1991b). In vivo, CNTF can prevent cell death of several classes of CNS neurons after axotomy in adult rats. Intraventricular infusion of CNTF in rats with fimbria-fornix transection protected neurons in the medial septal region from death, atrophy, and down-regulation of low-affinity NGF receptor but not from down-regulation of choline acetyl transferase (ChAT) (Hagg et al, 1992). Intracerebral infusion of CNTF reduced neuronal death in the thalamus following cortical lesions (Clatterbuck et al., 1993) and in the substantia nigra following nigrostrial tractotomy (Hagg and Varon, 1993), while intravitreal injections of CNTF prolonged the survival of axotomized retinal ganglion cells (Mey and Thanos, 1993). 4.3. Differentiation of neurons The best characterized differentiating action of CNTF is the induction of a cholinergic phenotype in sympathetic neurons from neonatal rats. Under selected conditions, including the presence of NGF to keep sympathetic neurons alive, CNTF increases ChAT activity 100-fold and decreases tyrosine hydroxylase activity (Saadat et al., 1989). In similar cultures, CNTF up-regulates vasoactive intestinal peptide, substance P and somatostatin, but down-regulates neuropeptide Y (Rao et al., 1992b) and muscarinic receptors (Ludlam and Kessler, 1993; Ludlam et al., 1994). For these peptides and neurotransmitter enzymes in rat sympathetic neurons, the actions of CNTF are identical to those of LIF (Yamamori et al., 1989; Nawa et

Ciliary neurotrophic factor

al., 1990), GPA (Heller et al., 1995) and oncostatin M (Fann and Patterson, 1994). In contrast to LIF, CNTF does not alter concentrations of ChAT, vasoactive intestinal peptide (VIP), substance P and somatostatin in sensory neurons from neonatal rats (Rao et al., 1992b). Also, depolarization of sympathetic neurons has different influences on the inductive effects of LIF and CNTF (Rao et al., 1992b). CNTF remains a candidate for the 'switch factor' responsible for the developmental conversion from adrenergic to cholinergic phenotype of a subpopulation of sympathetic neurons. In rats, sympathetic fibres innervating sweat gland acquire cholinergic properties during the second and third postnatal week (Leblanc and Landis, 1986). Evidence that target tissues regulate neurotransmitter properties of sympathetic neurons in vivo was obtained from experiments in which glabrous skin from rat paws was grafted to a thoracic area normally covered by hairy skin (Schotzinger and Landis, 1988). These grafts became innervated by sympathetic fibres, which transiently contained catecholamines and subsequently acquired ChAT activity. The ChAT-inducing activity in sweat glands is, in turn, induced by noradrenergic sympathetic fibres (Habecker and Landis, 1994). Historically, a long search for the agent mediating this neurotransmitter plasticity appeared to have ended when a cholinergic differentiation factor purified from heart-conditioned medium (Yamamori et al., 1989) proved to be identical to LIF, previously characterized for its actions in the hematopoietic system (Gearing et al., 1987). In transgenic mice, where the LIF gene is regulated by the insulin promoter, sympathetic neurons projecting to the pancreas express the ChAT gene (Bamber et al., 1994). However, cholinergic sympathetic neurons are present in mice with deletions of the LIF gene (Rao et al., 1993). CNTF, like LIF, stimulates a cholinergic phenotype of sympathetic neurons in vitro (Nawa and Patterson, 1990; Fann and Patterson, 1994). Extracts from the foot pads (which contain sweat glands) of 21 day old rats contain activity that induces ChAT and suppresses catecholamines in sympathetic neurons in vitro and is immunoprecipitated by antibodies to CNTF,

P.M. Richardson and M.C. Subang

but is not blocked by antibodies to LIF (Rao et al., 1992a; Rohrer, 1992). Like CNTF, the ChATinducing agent present in footpads is decreased following nerve transection (Rohrer, 1992) and has a molecular weight of 22-26 kDa and pi of 5.0 (Rao et al., 1992a). Like CNTF and LIF, swaet gland extracts induce substance P, enkephalin, VIP and ChAT mRNAs in sympathetic neurons. However, CNTF mRNA and CNTF immunoreactivity are localized in Schwann cells ensheathing nerve fibres in the skin, but not in sweat glands (Rao et al., 1992a; Rohrer, 1992). Also, sympathetic nerves with cholinergic phenotype are present in mice with null mutation of the CNTF gene (Masu et al., 1993). The mRNA for activin ySs, a member of the transforming growth factor-^ superfamily, is present in sweat glands, but activins do not have the same actions on sympathetic neurons as sweat gland extracts (Fann and Patterson, 1995). In summary, the molecular nature of the * switch factor' is believed to be similar, but not identical, to CNTF (see also chapter by Yamamori). The regulatory actions of CNTF on CNS neurons have not been examined as extensively as those on PNS neurons, but may include enhanced expression of the genes for the low-affinity NGF receptor (Magal et al., 1991a; Magal et al., 1991b) and for tyrosine hydroxylase (Louis et al., 1993a). CNTF induces morphological as well as molecular differentiation of neurons. Embryonic mammalian and avian neurons cultured in the presence of CNTF manifest neurofilamentcontaining processes, although it cannot be ascertained whether CNTF actively stimulates neurite outgrowth or is merely permissive by virtue of its effect on neuronal survival. For several classes of molluscan neurons, however, the situation is clearer. These neurons survive without neurite extension in the absence of CNTF, but rapidly acquire neurites when CNTF is added to the culture medium (Bulloch et al., 1992). 4.4. Actions on neuronal cell lines To facilitate studies of CNTF signal transduction, cell lines have been screened for morphologi-

299

cal or molecular responses to CNTF. To date, five cell lines have been reported to survive and/or differentiate in response to CNTF. The MAH cell line, derived by immortalization of rat sympathetic adrenal precursor cells with w-myc (Birren and Anderson, 1990), responds to CNTF or LIF by an arrest of proliferation and doubling of ChAT activity (Ip et al., 1992a). MAH cells differentiate into postmitotic NGF-responsive neurons under the cooperative influence of CNTF and FGF (Ip et al., 1994). Three human neuroblastoma cell lines differentiate when CNTF is added to the culture medium. In NBLF neuroblastoma cells, CNTF increases concentrations of vasoactive intestinal peptide, somatostatin, and calcitonin gene-related peptide mRNAs (Symes et al., 1993); in LAN-2 neuroblastoma cells, CNTF increases neurite length, ChAT activity, and acetyl choline synthesis (Lawrance et al., 1995) and, in SK-N-SH neuroblastoma cells, CNTF increases voltage-gated potassium channel activity (Lesser and Lo, 1995). Finally, CNTF together with insulin supports the survival of PI9 murine embryonal carcinoma cells in serum-free medium and stimulates the outgrowth of neurofilament-containing processes (Gupta etal., 1993). These cell lines have proved to be helpful in investigating the intracellular signal transduction pathways that mediate the actions of CNTF on survival or differentiation. 4.5. Actions on glial cells CNTF acts on progenitor glial cells from neonatal rat optic nerves, which can be driven to develop into either oligodendrocytes or type 2 astrocytes by CNTF, depending upon the culture conditions. In serum-free medium, CNTF combines with unknown components of the extracellular matrix to induce astrocytic differentiation of 0-2A cells (Hughes et al., 1988). Under different conditions or with selection of 0-2Aoligodendrocyte (0-2A) subtypes, CNTF and LIF promote the generation and maturation of oligodendrocytes (Mayer et al., 1994; Card et al., 1995). It is suggested that CNTF can initiate but not maintain differentiation of O2A cells into type 2 astrocytes. In vivo, injection of

300

CNTF into the brain increases the number of cells with immunoreactivity for glial fibrillary acidic protein (Winter et al., 1995). A second action of CNTF on CNS glial cells is to prevent the death of oligodendrocytes. In a cell line of 0-2A progenitor cells, CNTF reduces cell death due to serum-free medium or exposure to tumour necrosis factor (Louis et al, 1993b). CNTF, like insulin-like growth factor-1, supports the survival of oligodendrocytes in vitro (Barres et al., 1993; Mayer et al., 1994). Insulin together with CNTF and NT-3 support oligodendrocyte survival better than any one or two of these factors alone. In the developing optic nerves, the majority of dead cells are newly formed oligodendrocytes. When 293 cells transfected with cDNA encoding a secreted form of CNTF were injected into the subarachnoid space of neonatal rats, the average number of dead optic nerve cells 4 days post-implantation was reduced by about 80% (Barres et al., 1993). A LIF-like factor present in astrocyte-conditioned medium supports the survival of oligodendrocytes in culture better than any single identified trophic factor (Card et al., 1995a). 4.6. Other actions Several actions of CNTF reflect its similarity to LIF and IL-6. In vitro, CNTF maintains embryonic stem cells in an undifferentiated state (Conover et al., 1993) and induces the synthesis of acute phase proteins in hepatocytes or hepatoma cells (Schooltink et al., 1992). In vivo, also, CNTF has some of the same actions as LIF and IL-6. Delivered systemically, it is pyrogenic (Shapiro et al., 1993), induces cachexia and death (Henderson et al., 1994; Zhang et al., 1995) and stimulates synthesis of acute phase proteins (Dittrich et al., 1994). Clinical trials of CNTF as a putative therapeutic agent in amyotrophic lateral sclerosis were stopped because of weight loss, fever and activation of herpes simplex infections (Barinaga, 1994). The increase in toxicity of CNTF associated with intrathecal delivery (Zhang et al., 1995) is presumptive evidence that some of its toxicity may be due to effects within the CNS.

Ciliary neurotrophic factor

5. CNTF receptors and signal transduction 5.1. CNTF receptors In binding studies with cell lines or rat sympathetic neurons, [^^^IJCNTF binds to its receptor with a dissociation equilibrium constant of <10pmol (Wong et al., 1995). In cross-linking studies performed after incubation of PI9 embryonal carcinoma cells (Gupta et al., 1993) or Ewing sarcoma cells (Stahl et al., 1993) with [^^qjCNTF, three binding proteins have been detected. These three components of the CNTF receptor complex have now been identified: one is used uniquely by CNTF and the other two are implicated in LIF signalling also. The CNTFa receptor, consisting of 372 amino acids and with a weak sequence similarity to the IL-6 receptor, was discovered by screening of a cDNA expression library from human neuroblastoma cells (Davis et al., 1991). It binds CNTF at low affinity and is anchored to the cell surface by glycosylphosphatidyl inositol linkage, rather than by a transmembrane domain. Hematopoietic cells, which normally respond to LIF but not CNTF, show rapid phosphorylation responses to CNTF following transfection of the CNTFa receptor gene or addition of a soluble form of recombinant CNTFa receptor (Taga et al., 1992; Davis et al., 1993a; Ip et al., 1993). However, because the CNTFa receptor lacks any cytoplasmic domain, it can transduce CNTF actions only by acting in combination with other cell membrane molecules. By in situ hybridization in rats, CNTFa receptor mRNA has been detected in all classes of PNS neurons and in CNS neurons of the cortex, hippocampus, thalamus, substantia nigra, cerebellum and brainstem (Murakami et al., 1991). CNTFa receptor mRNA appears in the nervous system early during development. The distribution of the CNTFa receptor, more restricted than that of other members of the CNTF receptor complex, is thought to define the physiological targets for CNTF. Two other putative components of the CNTF receptor complex were identified by studies of tyrosine phosphorylation in MAH cells stimulated

P.M. Richardson and M.C. Subang

by CNTF (Ip et al., 1992b). One membraneassociated protein phosphorylated on tyrosine residues is similar or identical to gpl30, discovered through its participation in IL-6 signal transduction (Taga et al., 1989; Hibi et al., 1990) and later found to confer high-affinity binding for LIF and oncostatin-M in cells already containing the LIF^S receptor (Gearing et al., 1992). A protein in Ewing sarcoma cells that is tyrosine-phosphorylated upon CNTF or LIF stimulation can be immunoprecipitated by antibodies to gpl30 (Ip et al., 1992b) and antibodies to gpl30 block CNTF-induced and LIFinduced tyrosine phosphorylations in MAH cells (Ip et al., 1992b) and erythroleukemia cells (Taga etal., 1992). The third subunit of the CNTF receptor complex is the LIF^ receptor, whose gene has been cloned and sequenced (Gearing et al., 1991). Evidence implicating the LIF^ receptor is 3-fold. (1) A high-molecular weight protein in Ewing sarcoma cells that is phosphorylated on tyrosine residues following LIF or CNTF stimulation can be immunoprecipitated by antibodies to a LIF^ receptor peptide (Stahl et al., 1993). (2) Binding of [^^^I]CNTF to the high-molecular weight highaffinity binding protein on PI9 cells is competed by LIF (Gupta et al., 1993). (3) More directly, transfection of the LIF^ receptor gene together with genes for CNTFa receptor and gpl30, confers responsiveness to CNTF in COS cells, hematopoietic cells and hepatoma cells (Baumann et al., 1993; Davis et al., 1993b; Gearing et al., 1994). CNTF signalling appears to be initiated by heterodimerization of the LIF/3 receptor and gpl30, triggered in some manner by binding of CNTF to the CNTFa receptor (Davis et al., 1993b). Deletion of the LIFyS gene by homologous recombination has much more severe consequences than does deletion of LIF or CNTF genes (Ware etal., 1995). 5.2. CNTF signal transduction Several molecular events now have been discovered that mediate changes in gene expression after dimerization of gpl30 with itself or with LIF^ receptor. Three members of the Jak (Janus

301

kinase) family of tyrosine kinases, Jakl, Jak2, and Tyk2, are associated physically with gpl30 and LEF^ receptor (Narazaki et al., 1994; Stahl et al., 1994). The same cytokine activates different combinations of Jaks in different cell types (Stahl et al., 1994). Following activation of these nonreceptor tyrosine kinases when the associated receptors bind CNTF (Lutticken et al., 1994; Stahl et al., 1994), tyrosine residues in the cytoplasmic domain of the receptors are phosphorylated. These phosphorylated tyrosine residues on gpl30 and LIF/3R bind to signal transducer and activation of transcription factor (STAT3) also known as acutephase response factor (APRF) which is phosphorylated and translocated from cytoplasm to the cell nucleus. There, it acts as a transcription factor, binding to consensus sequences in regulatory elements of responsive genes (Darnell et al., 1994; Kishimoto et al., 1994) such as the vasoactive intestinal peptide gene (Bonni et al., 1993). The specificity of interaction between receptor and STAT molecules is governed by consensus sequences surrounding tyrosine residues in the receptor (Stahl et al., 1995) and within SH2 domains of the STATs (Heim et al., 1995). The duration of activation of Jaks appears to be truncated by activation of protein tyrosine phosphatases (PTPs) that are recruited to phosphorylated tyrosine residues in the receptors (Klingmiiller et al., 1995; Stahl et al., 1995). PTPID is tyrosine-phosphorylated following CNTF stimulation (Boulton et al., 1994). Jaks and STATs are involved in signal transduction of several classes of receptors that lack their own tyrosine kinases (Taniguchi, 1995). The C/EBP)8 family of transcription factors, known for their involvement in IL-6 signal transduction (Kishimoto et al., 1994), participate together with STAT proteins in mediating CNTFinduced signal transduction (Symes et al., 1995). The existence of such combinatorial actions provides an explanation for differences in the actions of cytokines that activate common signaltransducing systems. CNTF and other neuropoietic cytokines activate several kinases that also are activated by receptorlinked tyrosine kinases. The mitogen-activated protein kinases ERKl and ERK2 and their target

302

ribosomal S6 protein kinases are activated by all five of these cytokines (Daeipour et al., 1993; Boulton et al., 1994; Yin and Yang, 1994). Raf-1, generally considered to precede ERKl and ERK2 in the signalling cascade, is also activated by CNTF but evidence for the involvement of the even earlier ras protein is conflicting (Borasio et al., 1993; Boulton et al., 1994; Schwarzschild et al., 1994). Intracellular injection of anti-ras antibodies interferes with neurite-promoting activities of neurotrophins but not CNTF (Borasio et al., 1993). SHC, GRB2, PLC-y, the llOkDa subunit of PIP3 kinase and the src-like factor Hck (Ernst et al., 1994) are five other signal transducing factors that are tyrosine phosphorylated by CNTF (Boulton et al., 1994). Insulin receptor substrate-1 is tyrosine-phosphorylated by LIE (Argetsinger et al., 1995). The immediate early response genes activated by CNTF, similar to those activated by IL-6 (Lord et al., 1991; Nakajima and Wall, 1991), include cfos, c-jun, and tis-W (Squinto et al, 1990; Ip et al., 1992b). It has been suggested that protein kinase C is required for the stimulation by CNTF of a cholinergic phenotype in sympathetic neurons (Kalberg et al., 1993), although the protein kinase inhibitor H7 does not block the ability of CNTF to induce tis-\ 1 in MAH cells (Ip et al., 1992b). The multitude of signal transduction pathways activated by the neuropoietic cytokines help to explain their pleiotrophic actions. References Aloisi, R, Rosa, S., Testa, U., Bonsi, P., Russo, G., Peschle, C. and Levi, G. (1994) Regulation of leukemia inhibitory factor synthesis in cultured human astrocytes. /. Immunol 152:5022-5031. Anand, P., Parrett, A., Martin, J., Zeman, S., Foley, P., Swash, M., Leigh, P.N., Cedarbaum, J.M., Lindsay, R.M., Williams-Chestnut, R.E. and Sinicropi, D.V. (1995) Regional changes of ciliary neurotrophic factor and nerve growth factor levels in post mortem spinal cord and cerebral cortex from patients with motor disease. Nature Med. 1: 168-172. Arakawa, Y., Sendtner, M. and Thoenen, H. (1990) Survival effect of ciliary neurotrophic factor (CNTF) on chick embryonic motoneurons in culture: comparison with other neurotrophic factors and cytokines. J. Neurosci, 10: 35073515. Argetsinger, L. S., Hsu, G. W., Myers, M. G.,Jr., Billestrup,

Ciliary neurotrophic factor N., White, M. F. and Carter-Su, C. (1995) Growth hormone, interferon-, and leukemia inhibitory factor promoted tyrosyl phosphorylation of insulin receptor substrate-1. J. Biol. Chem. 270: 14685-14692. Asada, H., Ip, N. Y., Pan, L., Razack, N., Parfitt, M. M and Plunkett, R. J. (1995) Time course of ciliary neurotrophic factor mRNA expression is coincident with the presence of protoplasmic astrocytes in traumatized rat striatum. J. Neurosci. Res. 40: 22-30. Bamber, B. A., Masters, B. A., Hoyle, G. W., Brinster, R. L. and Palmiter, R. D. (1994) Leukemia inhibitory factor induces neurotransmitter switching in transgenic mice. Proc. Natn. Acad. Sci. USA 91: 7839-7843. Barbin, G., Manthorpe, M. and Varon, S. (1984) Purification of the chick eye ciliary neurotrophic factor. J. Neurochem. 43:1468-1478. Barinaga, M. (1994) Neurotrophic factors enter the clinic. Science 264:112-11 A. Barres, B.A., Schmid, R., Sendtner, M. and Raff, M.C. (1993) Multiple extracellular signals are required for long-term oligodendrocyte survival. Development 118: 283-295. Baumann, H., Ziegler, S.F., Mosley, B., Morella, K.K., Pajovic, S. and Gearing, D.P. (1993) Reconstitution of the response to leukemia inhibitory factor, oncostatin M, and ciliary neurotrophic factor in hepatoma cells. J. Biol. Chem. 268: 8414-8417. Bazan, J.F.(1991) Neuropoietic cytokines in the hematopoietic fold. Neuron 7: 197-208. Birren, S.J. and Anderson, D.J. (1990) A v-myc-immortalized sympathoadrenal progenitor cell line in which neuronal differentiation is initiated by FGF but not NGF. Neuron 4: 189-201. Bloch-Gallego, E., Huchet, M., El M'hamdi, H., Xie, F.-K., Tanaka, H. and Henderson, C.E. (1991) Survival in vitro of motoneurons identified or purified by novel antibody-based methods is selectively enhanced by muscle-derived factors. Development H I : 221-232. Blottner, D., Bruggemann, W. and Unsicker, K. (1989) Ciliary neurotrophic factor supports target-deprived preganglionic sympathetic spinal cord neurons. Neurosci. Lett. 1105: 316320. Bonni, A., Frank, D.A., Schindler, C. and Greenberg, M.E. (1993) Characterization of a pathway for ciliary neurotrophic factor signaling to the nucleus. Science 262: 1575-1579. Bonyhady, R.E., Hendry, I.A., Hill, C.E. and McLennan, I.S. (1980) Characterization of a cardiac muscle factor required for the survival of cultured parasympathetic neurons. Neurosci. Lett. 18: 197-201. Borasio, G.D., Markus, A., Wittinghofer, A., Barde, Y.-A. and Heumann, R. (1993) Involvement of ras p21 in neurotrophin-induced response of sensory, but not sympathetic neurons. J. Cell Biol. 121: 665-672. Boulton, T.G., Stahl, N. and Yancopoulos, G.D. (1994) Ciliary neurotrophic factor/leukemia inhibitory factor/ interleukin 6/oncostatin M family of cytokines induces ty-

P.M. Richardson and M.C. Subang rosine phosphorylation of a common set of proteins overlapping those induced by other cytokines and growth fact o r s . / Biol. Chem. 269: 11648-11655. Bulloch, A.G.M., Syed, N.I. and Richardson, P.M. (1992) Ciliary neurotrophic factor (CNTF) induces sprouting by adult lymnaea neurons in vitro. Soc. Neurosci. Abstr. 18: 1095. Carroll, P., Sendtner, M., Meyer, M. and Thoenen, H. (1993) Rat ciliary neurotrophic factor (CNTF): gene structure and regulation of mRNA levels in glial cell cultures. Glia 9: 176-187. Cerretti, D.P., Kozlosky, C.J., Mosley, B., Nelson, N., Van Ness, K., Greenstreet, T.A., March, C.J., Kronheim, S.R., Druck, T., Cannizzaro, L.A., Huebner, K. and Black, R.A. (1992) Molecular cloning of the interuleukin-lbeta converting enzyme. Science 256: 97-100. Clatterbuck, R.E., Price, D.L. and Koliatsos, V.E. (1993) Ciliary neurotrophic factor prevents retrograde neuronsl death in the adult central nervous system. Proc. Natl. Acad. Sci. USA 90: 2222-2226. Clatterbuck, R.E., Price, D.L. and Koliatsos, V.E. (1994) Further characterization of the effects of brain-derived neurotrophic factor and ciliary neurotrophic factor on axotomized neonatal and adult mammalian motor neurons. J. Comp. Neurol. 342: 45-56. Conover, J . C , Ip, N.Y., Poueymirou, W.T., Bates, B., Goldfarb, M.P., DeChiara, T.M. and Yancopoulos, G.D. (1993) Ciliary neurotrophic factor maintains the pluripotentiality of embryonic stem cells. Development 119: 559-565. Curtis, R., Scherer, S.S., Somogyi, R., Adryan, K.M., Ip, N.Y., Zhu, Y., Lindsay, R.M. and DiStefano, P.S. (1994) Retrograde axonal transport of LIF is increased by peripheral nerve injury: correlation with increased LIF expression in distal nerve. Neuron 12: 191-204. Daeipour, M., Kumar, G., Amaral, M.C. and Nel, A.E. (1993) Recombinant IL-6 activates p42 and p44 mitogen-activated protein kinases in the IL-6 responsive B cell line, AF-10. J. Immunol. 150: 4743-4753. Darnell, J.E., Jr., Kerr, I.M. and Stark, G.R. (1994) Jak-STAT pathways and transcriptional activation in response to IFNs and other extacellular signaling proteins. Science 264: 1415-1421. Davis, S., Aldrich, T.H., Valenzuela, D.M., Wong, V., Furth, M., Squinto, S.P. and Yancopoulos, G.D. (1991) The receptor for ciliary neurotrophic factor. Science 253: 59-63. Davis, S., Aldrich, T.H., Ip, N.Y., Stahl, N., Scherer, S., Farruggella, T., DiStefano, P.S., Curtis, R., Panayotatos, N., Gascan, H., Chevalier, S. and Yancopoulos, G.D. (1993a) Released form of CNTF receptor alpha component as a soluble mediator of CNTF responses. Science 259: 17361739. Davis, S., Aldrich, T.H., Stahl, N., Pan, L., Taga, T., Kishimoto, T., Ip, N.Y. and Yancopoulos, G.D. (1993b) LIFRy? and gpl30 as heterodimerizing signal transducers of the tripartite CNTF receptor. Science 260: 1805-1808. Dittrich, F., Thoenen, H. and Sendtner, M. (1994) Ciliary

303 neurotrophic factor: pharmacokinetics and acute-phase response in rat. Ann. Neurol. 35: 151-163. Dobrea, G.M., Unnerstall, J.R. and Rao, M.S. (1992) The expression of CNTF message and immunoreactivity in the central and peripheral nervous system of the rat. Dev. Brain Res. 66: 209-219. Ebendal, T. (1987) Comparative screening for ciliary neurotrophic activity in organs of the rat and chicken. J. Neurosci. Res. 17: 19-24. Ebendal, T., Belew, M., Jacobson, C.-O. and Porath, J. (1979) Neurite outgrowth elicited by embryonic chick heart: partial purification of the active factor. Neurosci. Lett. 14: 91-95. Ebendal, T., Jordell-Kylberg, A. and Soderstrom, S. (1978) Stimulation by tissue explants on nerve fibre outgrowth in culture. In: C.-O. Jacobson and T. Ebendal (Eds.), Formshaping Movements in Neurogenesis, Almqvist & Wiksell, Stockholm, pp. 235-243. Eckenstein, F. P., Esch, F., Holbert, T., Blacher, R. W. and Nishi, R. (1990) Purification and characterization of a trophic factor for embryonic peripheral neurons: comparison with fibroblast growth factors. Neuron 4: 623-631. Eckenstein, F.P., Shipley, G.D. and Nishi, R. (1991) Acidic and basic fibroblast growth factors in the nervous system: distribution and differential alteration of levels after injury of central versus peripheral nerve. J. Neurosci. 11: 4 1 2 419. English, A.W. and Schwartz, G. (1995) Both basic fibroblast growth factor and ciliary neurotrophic factor promote the retention of polyneuronal innervation of developing skeletal muscle fibers. Dev. Biol. 169: 57-64. Ernsberger, U., Sendtner, M. and Rohrer, H. (1989) Proliferation and differentiation of embryonic chick sympathetic neurons: effects of ciliary neurotrophic factor. Neuron 2: 1275-1284. Ernst, M., Gearing, D.P. and Dunn, A.R. (1994) Functional and biochemical association of Hck with the LIF/IL-6 receptor signal transducing subunit gpl30 in embryonic stem cells. EMBOJ. 13: 1574-1584. Fann, M.-J. and Patterson, P.H. (1994) Neuropoietic cytokines and activin A differentially regulate the phenotype of cultured sympathetic neurons. Proc. Natl. Acad. Sci. USA 91: 43-47. Fann, M.-J. and Patterson, P.H. (1995) Activins as candidate cholinergic differentiation factors in vivo. Int. J. Dev. Neurosci. 13: 317-330. Friedman, B., Scherer, S.S., Rudge, J.S., McClain, J., Wang, D., Wiegand, S.J., Furth, M.E., Lindsay, R.M. and Ip, N.Y. (1992) Regulation of ciliary neurotrophic factor expression in myelin-related Schwann cells in vivo. Neuron 9: 2 9 5 305. Card, A.L., Burrell, M.R., Pfeiffer, S.E., Rudge, J.S. and Williams, W.C. (1995a) Astroglial control of oligodendrocyte survival mediated by PDGF and leukemia inhibitory factorlike protein. Development 121: 2187-2197. Card, A.L., Williams, W.C. and Burrell, M.R. (1995b) Oligodendroblasts distinguished from 0-2A glial progenitors

304

by surface phenotype (04'^GalC ) and response to cytokines using signal transducer LIFRy3. Dev. Biol. 167: 596608. Gearing, D.P., Gough, N.M., King, J.A., Hilton, D.J., Nicola, N.A., Simpson, R.J., Nice, E.G., Kelso, A. and Metcalf, D. (1987) Molecular cloning and expression of cDNA encoding a murine myeloid leukemia inhibitory factor (LIF). £ M ^ 0 / . 6:3995^002. Gearing, D.P., Thut, C.J., VandenBos, T., Gimpel, S.D., Delaney, P.B., King, J., Price, V., Cosman, D. and Beckmann, M.P. (1991) Leukemia inhibitory factor receptor is structurally related to the IL-6 signal transducer, gpl30. EMBO J. 10: 2839-2848. Gearing, D.P., Comeau, M.R., Friend, D.J., Gimpel, S.D., Thut, C.J., McGourty, J., Brasher, K.K., King, J.A., Gillis, S., Mosley, B., Ziegler, S. F. and Cosman, D. (1992) The IL-6 signal transducer, gpl30: an oncostatin M receptor and affinity converter for the LIF receptor. Science 255: 14341437. Gearing, D.P., Ziegler, S.F., Comeau, M.R., Friend, D., Thoma, B., Cosman, D., Park, L. and Mosley, B. (1994) Proliferative responses and binding properties of hematopoietic cells transfected with low-affinity receptors for leukemia inhibitory factor, oncostatin M, and ciliary neurotrophic factor. Proc. Natl. Acad. Sci. USA 91: 1119-1123. Gonzalez, A.-M., Buscaglia, M., Ong, M. and Baird, A. (1990) Distribution of basic fibroblast growth factor in the 18-day rat retus: localization in the basement membranes of diverse tissues. J. Cell Biol. 110: 753-765. Grothe, C. and Unsicker, K. (1992) Basic fibroblast growth factor in the hypoglossal system: specific retrograde transport, trophic, and lesion-related responses. J. Neurosci. Res. 32: 317-328. Gupta, S.K., Altares, M., Benoit, R., Riopelle, R.J., Dunn, R.J. and Richardson, P.M. (1992) Preparation and biological properties of native and recombinant ciliary neurotrophic factor. J. Neurobiol. 23: 481-490. Gupta, S.K., Haggarty, A.J., Carbonetto, S., Riopelle, R.J., Richardson, P.M. and Dunn, R.J. (1993) Trophic actions of ciliary neurotrophic factor on murine embryonal carcinoma cells. Eur. J. Neurosci. 5: 977-985. Habecker, B.A. and Landis, S.C. (1994) Noradrenergic regulation of cholinergic differentiation. Science 264: 16021604. Hagg, T., Quon, D., Higaki, J. and Varon, S. (1992) Ciliary neurotrophic factor prevents neuronal degeneration and promotes low affinity NGF receptor expression in the adult rat CNS. Neuron 8: 145-158. Hagg, T. and Varon, S. (1993) Ciliary neurotrophic factor prevents degeneration of adult rat substantia nigra dopaminergic neurons in vivo. Proc. Natl. Acad. Sci. USA 90: 6315-6319. Hazuda, D.J., Lee, J.C. and Young, P.R. (1988) The kinetics of interleukin 1 secretion from activated monocytes. J. Biol. Chem. 263: 8473-8479. Heim, M.H., Kerr, I.M., Stark, G.R. and Darnell, J.E., Jr.

Ciliary neurotrophic factor (1995) Contribution of STAT SH2 groups to specific interferon signaling by the Jak-STAT pathway. Science 267: 1347-1349. Helfand, S.L., Smith, G.A. and Wessells, N.K. (1976) Survival and development in culture of dissociated parasympathetic neurons from ciliary ganglia. Dev. Biol. 50: 541547. Helgren, M.E., Squinto, S.P., Davis, H.L., Parry, D.J., Boulton, T.G., Heck, C.S., Zhu, Y., Yancopoulos, G.D., Lindsay, R.M. and DiStefano, P.S. (1994) Trophic effect of ciliary neurotrophic factor on denervated skeletal muscle. Cell 76: 493-504. Henderson, J.T., Seniuk, N.A., Richardson, P.M., Gauldie, J. and Roder, J.C. (1994) Systemic administration of CNTF induces cachexia in rodents. J. Clin. Invest. 93: 2632-2638. Hibi, M., Murakami, M., Saito, M., Hirano, T., Taga, T. and Kishimoto, T. (1990) Molecular cloning and expression of an IL-6 signal transducer, gpl30. Cell 63: 1149-1157. Hill, C.E., Belford, D.A., Godovac-Zimmermann, J. and Hendry, LA. (1991) Class I heparin binding growth factor promotes the differentiation but not the survival of ciliary neurones in vivo. Dev. Brain Res. 63: 13-19. Hughes, S.M., Lillien, L.E., Raff, M.C., Rohrer, H. and Sendtner, M. (1988) Ciliary neurotrophic factor induces type-2 astrocyte differentiation in culture. Nature 335: 7073. Hughes, R.A., Sendtner, M., Goldfarb, M., Lindholm, D. and Thoenen, H. (1993) Evidence that fibroblast growth factor 5 is a major muscle-derived survival factor for cultured spinal motoneurons. Neuron 10: 369-377. Ikeda, K., Wong, V., Holmlund, T.H., Greene, T., Cedarbaum, J.M., Lindsay, R.M. and Mitsumoto, H. (1995) Histometric effects of ciliary neurotrophic factor in wobbler mouse motor neuron disease. Ann. Neurol. 37: 47-54. Ip, N.Y., Li, Y., van de Stadt, I., Panayotatos, N., Alderson, R.F. and Lindsay, R.M. (1991) Ciliary neurotrophic factor enhances neuronal survival in embryonic rat hippocampal cultures. J. Neurosci. 11: 3124-3134. Ip, N.Y., Nye, S.H., Boulton, T.G., Davis, S., Taga, T., Li, Y., Birren, S., Yasukawa, K., Kishimoto, T., Anderson, D.J., Stahl, N. and Yancopoulos, G.D. (1992a) CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component gpl30.Ce//69: 1121-1132. Ip, N.Y., Nye, S.H., Boulton, T.G., Davis, S., Taga, T., Li, Y., Birren, S.J., Yasukawa, K., Kishimoto, T., Anderson, D.J., Stahl, N. and Yancopoulos, G.D. (1992b) CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component gpl30.Ce//69: 1121-1132. Ip, N.Y., Wiegand, S.J., Morse, J. and Rudge, J.S. (1992c) Injury-induced regulation of ciliary neurotrophic factor mRNA in the adult rat brain. Eur. J. Neurosci. 5: 25-33. Ip, N.Y., McClain, J., Barrezueta, N.X., Aldrich, T.H., Pan, L., Li, Y., Wiegand, S.J., Friedman, B., Davis, S. and Yancopoulos, G.D. (1993) The alpha component of the CNTF

P.M. Richardson and M.C. Subang receptor is required for signaling and defines potential CNTF targets in the adult and during development. Neuron 10: 89-102. Ip, N.Y., Boulton, T.G., Li, Y., Verdi, J.M., Birren, S.J., Anderson, D.J. and Yancopoulos, G.D. (1994) CNTF, FGF, and NGF collaborate to drive the terminal differentiation of MAH cells into postmitotic neurons. Neuron 13: 443^55. Jackson, A., Friedman, S., Zhan, X., Engleka, K.A., Forough, R. and Maciag, T. (1992) Heat shock induces the release of fibroblast growth factor 1 from NIH 3T3 cells. Proc. Natl. Acad. Sci. USA 89: 10691-10695. Kalberg, C, Yung, S.Y. and Kessler, J.A. (1993) The cholinergic stimulating effects of ciliary neurotrophic factor and leukemia inhibitory factor are mediated by protein kinase C. J. Neurochem. 60: 145-153. Kato, A.C. and Lindsay, R.M. (1994) Overlapping and additive effects of neurotrophins and CNTF on cultured human spinal cord neurons. Exp. Neurol. 130: 196-201. Kaupmann, K., Sendtner, M., Stockli, K.A. and Jockusch, H. (1991) The gene for ciliary neurotrophic factor (CNTF) maps to murine chromosome 19 and its expression is not affected in the hereditary motoneuron disease 'wobbler' of the mouse. Eur. J. Neurosci. 3: 1182-1186. Kishimoto, T., Taga, T. and Akira, S. (1994) Cytokine signal transduction. Cell 76: 253-262. Klingmiiller, U., Lorenz, U., Cantley, L.C., Neel, E.G. and Lodish, H.F. (1995) Specific recruitment of SH-PTPl to the erythropoietin receptor causes inactivation of JAK2 and termination of proliferative signals. Cell 80: 729-738. Kotzbauer, P.T., Lampe, P.A., Estus, S., Milbrandt, J. and Johnson, E.M., Jr. (1994) Postnatal development of survival responsiveness in rat sympathetic neurons to leukemia inhibitory factor and ciliary neurotrophic factor. Neuron 12: 763-773. Lam, A., Fuller, F., Miller, J., Kloss, J., Manthorpe, M., Varon, S. and Cordell, B. (1991) Sequence and structural organization of the human gene encoding ciliary neurotrophic factor. Gene 102: 271-276. Landmesser, L. and Pilar, G. (1974) Synaptic transmission and cell death during normal ganglionic development. /. Physiol. 241:737-749. Lawrance, G., Rylett, R.J., Richardson, P.M., Dunn, R.J., Dow, K.E. and Riopelle, R.J. (1995) Differentiation effects of ciliary neurotrophic factor on human neuroblastoma cells. /. Neurochem. 64: 1483-1490. Larkfors, L., Lindsay, R.M. and Alderson, R.F. (1994) Ciliary neurotrophic factor enhances the survival of purkinje cells in vitro. Eur. J. Neurosci. 6: 1015-1025. Leblanc, G. and Landis, S. (1986) Development of choline acetyltransferase (CAT) in the sympathetic innervation of rat sweat glands. /. Neurosci. 6: 260-265. Lehwalder, D., Jeffrey, P.L. and Unsicker, K. (1989) Survival of purified embryonic chick retinal ganglion cells in the presence of neurotrophic factors. J. Neurosci. Res. 24: 329337. Lesser, S.S. and Lo, D.C. (1995) Regulation of voltage-gated

305 ion channels by NGF and ciliary neurotrophic factor in SKN-SH neuroblastoma cells. J, Neurosci. 15: 253-261. Leung, D.W., Parent, A.S., Cachianes, G., Esch, F., Coulombe, J.N., Nikolics, K., Eckenstein, F.P. and Nishi, R. (1992) Cloning, expression during development, and evidence for release of a trophic factor for ciliary ganglion neurons. Neuron 8: 1045-1053. Lin, L.-F., Drzislav, M., Lile, J.D., Armes, L.G., Butler, E.T., Vannice, J.L. and Collins, F. (1989) Purification cloning and expression of ciliary neurotrophic factor (CNTF). Science 246: 1023-1025. Lord, K.A., Abdollahi, A., Thomas, S.M., DeMarco, M., Brugge, J.S., Hoffman-Liebermann, B. and Liebermann, D.A. (1991) Leukemia inhibitory factor and interleukin-6 trigger the same immediate early response, including tyrosine phosphorylation, upon induction of myeloid leukemia differentiation. Mol. Cell. Biol. 11: 4371-4379. Louis, J.-C, Magal, E., Burnham, P. and Varon, S. (1993a) Cooperative effects of ciliary neurotrophic factor and norepinephrine on tyrosine hydroxylase expression in cultured rat locus coeruleus neurons. Dev. Biol. 155: 1-13. Louis, J.-C, Magal, E., Takayama, S. and Varon, S. (1993b) CNTF protection of oligodendrocytes against natural and tumor necrosis factor-induced death. Science 259: 689-692. Ludlam, W.H. and Kessler, J.A. (1993) Leukemia inhibitory factor and ciliary neurotrophic factor regulate expression of muscarinic receptors in cultured sympathetic neurons. Dev. Biol. 155: 497-506. Ludlam, W.H., Zang, Z., McCarson, K.E., Krause, J.E., Spray, D.C. and Kessler, J.A. (1994) mRNAs encoding muscarinic and substance P receptors in cultured sympathetic neurons are differentially regulated by LIF or CNTF. Dev. Biol. 164: 528-539. Lutticken, C, Wegenka, U.M., Yuan, J., Buschmann, J., Schindler, C, Ziemiecki, A., Harpur, A.G., Wilks, A.F., Yasukawa, K., Taga, T., Kishimoto, T., Barbieri, G., Pellegrini, S., Sendtner, M., Heinrich, P.C. and Horn, F. (1994) Association of transcription factor APRF and protein kinase JAKl with the interleukin-6 signal transducer gpl30. Science 263: 89-92. Magal, E., Burnham, P. and Varon, S. (1991a) Effect of CNTF on low-affinity NGF receptor expression by cultured neurons from different rat brain regions. J. Neurosci. Res. 30: 560-566. Magal, E., Burnham, P. and Varon, S. (1991b) Effects of ciliary neuronotrophic factor on rat spinal cord neurons in vitro: survival and expression of choline acetyltransferase and low-affinity nerve growth factor receptors. Dev. Brain Res. 63: 141-150. Manthorpe, M., Skaper, S.D., Williams, L.R. and Varon, S. (1986) Purification of adult rat sciatic nerve ciliary neuronotrophic factor. Brain Res. 367: 282-286. Masiakowski, P., Liu, H., Radziejewski, C, Lottspeich, F., Oberthuer, W., Wong, V., Lindsay, R.M., Furth, M.E. and Panayotatos, N. (1991) Recombinant human and rat ciliary neurotrophic factors. /. Neurochem. 57: 1003-1012.

306 Masu, Y., Wolf, E., Holtmann, B., Sendtner, M., Brem, G. and Thoenen, H. (1993) Disruption of the CNTF gene results in motor neuron degeneration. Nature 365: 27-32. Mayer, M., Bhakoo, K. and Noble, M. (1994) Ciliary neurotrophic factor and leukemia inhibitory factor promote the generation, maturation and survival of oligodendrocytes in vitro. Development 120: 143-153. McDonald, N.Q., Panayotatos, N. and Hendrickson, W.A. (1995) Crystal structure of dimeric human ciliary neurotrophic factor determined by MAD phasing. EMBO J. 14: 2689-2699. McNeil, P.L., Muthukrishnan, L., Warder, E. and D'Amore, P.A. (1989) Growth factors are released by mechanically wounded endothelial cells. /. Cell Biol. 109: 811-822. Mey, J. and Thanos, S. (1993) Intravitreal injections of neurotrophic factors support the survival of axotomized retinal ganglion cells in adult rats in vivo. Brain Res. 602: 304317. Mitsumoto, H., Ikeda, K., Holmlund, T., Greene, T., Cedarbaum, J.M., Wong, V. and Lindsay, R.M. (1994a) The effects of ciliary neurotrophic factor on motor dysfunction in wobbler mouse motor neuron disease. Ann. Neurol. 36: 142-148. Mitsumoto, H., Ikeda, K., Klinkosz, B., Cedarbaum, J.M., Wong, V. and Lindsay, R.M. (1994b) Arrest of motor neuron disease in wobbler mice cotreated with CNTF and BDNF. Science 265: 1107-1110. Murakami, M., Narazaki, M., Hibi, M., Yawata, H., Yasukawa, K., Hamaguchi, M., Taga, T. and Kishimoto, T. (1991) Critical cytoplasmic region of the interleukin 6 signal transducer gpl30 is conserved in the cytokine receptor family. Proc. Natl. Acad. Sci. USA 88: 1134911353. Murphy, M., Reid, K., Hilton, D.J. and Bartlett, P.F. (1991) Generation of sensory neurons is stimulated by leukemia inhibitory factor. Proc. Natl. Acad. Sci. USA 88: 34983501. Murphy, P.G., Grondin, J., Altares, M. and Richardson, P.M. (1995) Induction of interleukin-6 in axotomized sensory neurons. J. Neurosci. 15: 5130-5138. Nagao, H., Matsuoka, I. and Kurihara, K. (1995) Effects of adenyl cyclase-linked neuropeptides on the expression of ciliary neurotrophic factor-mRNA in cultured astrocytes. FEBS Lett. ^62:15-19. Nakajima, K. and Wall, R. (1991) Interleukin-6 signals activating junB and TISll gene transcription in a B-cell hybridoma. Mol. Cell. Biol. 11: 1409-1418. Narazaki, M., Witthuhn, B.A., Yoshida, K., Silvennoinen, O., Yasukawa, K., Ihle, J.N., Kishimoto, T. and Taga, T. (1994) Activation of JAK2 kinase mediated by the interleukin 6 signal transducer gpl30. Proc. Natl. Acad. Sci. USA 91: 2285-2289. Nawa, H., Yamamori, T., Le, T. and Patterson, P.H. (1990) Generation of neuronal diversity: analogies and homologies with hematopoiesis. Cold Spring Harbor Symp. Quant. Biol. LV: 247-253.

Ciliary neurotrophic factor Nawa, H. and Patterson, P.H. (1990) Separation and partial characterization of neuropeptide-inducing factors in heart cell conditioned medium. Neuron 4: 269-277. Negro, A., Corona, G., Bigon, E., Martini, I., Grandi, C, Skaper, S.D. and Callegaro, L. (1991) Synthesis, purification, and characterization of human ciliary neurotrophic factors from E. coli. J. Neurosci. Res. 29: 251-260. Nishi, R. and Berg, D.K. (1977) Dissociated ciliary ganglion neurons in vitro: survival and synapse formation. Proc. Natl. Acad Sci. USA 14: 5171-5175. Oppenheim, R.W., Prevette, D., Qin-Wei, Y., Collins, F. and MacDonald, J. (1991) Control of embryonic motoneuron survival in vivo by ciliary neurotrophic factor. Science 251: 1616-1618. Panayotatos, N., Radziejewska, E., Acheson, A., Somogyi, R., Thadani, A., Hendrickson, W.A. and McDonald, N.Q. (1995) Localization of functional receptor epitopes on the structure of ciliary neurotrophic factor indicates a conserved, function-related epitope topography among helical cytokines. J. Biol. Chem. 270: 14007-14014. Rao, M.S., Patterson, P.H. and Landis, S.C. (1992a) Multiple cholinergic differentiation factors are present in footpad extracts: comparison with known cholinergic factors. Development 116:731-744. Rao, M.S., Tyrrell, S., Landis, S.C. and Patterson, P.H. (1992b) Effects of ciliary neurotrophic factor (CNTF) and depolarization on neuropeptide expression in cultured sympathetic neurons. Dev. Biol. 150: 281-293. Rao, M.S., Sun, Y., Escary, J.L., Perreau, J., Tresser, S., Patterson, P.H., Zigmond, R.E., Brulet, P. and Landis, S.C. (1993) Leukemia inhibitory factor mediates an injury response but not a target-directed developmental transmitter switch in sympathetic neurons. Neuron 11: 1175-1185. Rende, M., Muir, D., Ruoslahti, E., Hagg, T., Varon, S. and Manthorpe, M. (1992) Immunolocalization of ciliary neuronotrophic factor in adult rat sciatic nerve. Glia 5: 2532. Richards, L.J., Kilpatrick, T.J., Bartlett, P.F. and Murphy, M. (1992) Leukemia inhibitory factor promotes the neuronal development of spinal cord precursors from the neural tube. J. Neurosci. Res. 33: 476-484. Richardson, P. M. and Ebendal, T. (1982) Nerve growth activities in rat peripheral nerve. Brain Res. 246: 57-64. Richardson, P.M., Gupta, S.K., Carbonetto, S., Dunn, R.J. and Riopelle, R.J. (1993) Glial-neuronal interactions exemplified by the synthesis and actions of ciliary neurotrophic factor. In: S. Fedoroff and G. Burkholder (Eds.), Biology and Pathology of Astrocyte-Neuron Interactions, Plenum Press, New York, pp. 173-179. Robinson, R.C., Grey, L.M., Staunton, D., Vankelecom, H., Vernallis, A.B., Moreau, J.-F., Stuart, D.I., Heath, J.K. and Jones, E.Y. (1994) The crystal structure and biological function of leukemia inhibitory factor: implications for receptor binding. Cell 11: 1101-1116. Rohrer, H. (1992) Cholinergic neuronal differentiation factors: evidence for the presence of both CNTF-like and non-

P.M. Richardson andM.C. Subang CNTF-like factors in developing rat footpad. Development 114:689-698. Rubartelli, A., Cozzolino, F., Talio, M. and Sitia, R. (1990) A novel secretory pathway for interleukin-lbeta, a protein lacking a signal sequence. EMBO J. 9: 1503-1510. Rudge, J.S., Alderson, R.F., Pasnikowski, E., McClain, J., Ip, N.Y. and Lindsay, R.M. (1992) Expression of ciliary neurotrophic factor and the neurotrophins - nerve growth factor, brain-derived neurotrophic factor and neurotrophin 3 in cultured rat hippocampal astrocytes. Eur. J. Neurosci. 4: 459-471. Saadat, S., Sendtner, M. and Rohrer, H. (1989) Ciliary neurotrophic factor induces cholinergic differentiation of rat sympathetic neurons in culture. J. Cell Biol. 108: 18071816. Saggio, I., Gloaguen, I., Poiana, G. and Laufer, R. (1995) CNTF variants with increased biological potency and receptor selectivity define a functional site of receptor interaction. EMBO J. 14: 3045-3054. Sahenk, Z., Seharaseyon, J. and Mendell, J.R. (1994) CNTF potentiates peripheral nerve regeneration. Brain Res. 655: 246-250. Savino, R., Ciapponi, L., Lahm, A., Demartis, A., Cabibbo, A., Toniatti, C, Delmastro, P., Altamura, S. and Ciliberto, G. (1994) Rational design of a receptor super-antagonist of human interleukin-6. EMBO J. 13: 5863-5870. Schooltink, H., Stoyan, T., Roeb, E., Heinrich, P.C. and RoseJohn, S. (1992) Ciliary neurotrophic factor induces actephase protein expression in hepatocytes. EEBS Lett. 314: 280-284. Schotzinger, R.J. and Landis, S.C. (1988) Cholinergic phenotype developed by noradrenergic sympathetic neurons after innervation of a novel cholinergic target in vivo. Nature 335: 637-639. Schwarzschild, M.A., Dauer, W.T., Lewis, S.E., Hamill, L.K., Fink, J.S. and Hyman, S.E. (1994) Leukemia inhibitory factor and ciliaryneurotrophic factor increase activated ras in a neuroblastoma cell line and in sympathetic neuron cultures. J. Neurochem. 63: 1246-1254. Sendtner, M., Kreutzberg, G. and Thoenen, H. (1990) Ciliary neurotrophic factor prevents the degeneration of motoneurons after axotomy. Nature 345: 440-441. Sendtner, M., Schmalbruch, H., Stockli, K.A., Carroll, P., Kreutzberg, G.W. and Thoenen, H. (1992a) Ciliary neurotrophic factor prevents degeneration of motor neurons in mouse mutant progressive motor neuronopathy. Nature 358:502-504. Sendtner, M., Stockli, K.A. and Thoenen, H. (1992b) Synthesis and localization of ciliary neurotrophic factor in the sciatic nerve of the adult rat after lesion and during development. 7. Cell Biol. 118: 139-148. Seniuk, N., Altares, M., Dunn, R. and Richardson, P.M. (1992) Decreased synthesis of ciliary neurotrophic factor in degenerating peripheral nerves. Brain Res. 572: 300-302. Shapiro, L., Zhang, X.-X., Rupp, R.G., Wolff, S.M. and Dinarello, C.A. (1993) Ciliary neurotrophic factor is an en-

307 dogenous pyrogen. Proc. Natl. Acad. Sci. USA 90: 86148618. Squinto, S.P., Aldrich, T.H., Lindsay, R.M., Morrissey, D.M., Panayotatos, N., Bianco, S.M., Furth, M.E. and Yancopoulos, G.D. (1990) Identification of functional receptors for ciliary neurotropic factor on neuronal cell lines and primary neurons. Neuron 5: 757-766. Stahl, N., Davis, S., Wong, V., Taga, T., Kishimoto, T., Ip, N.Y. and Yancopoulos, G.D. (1993) Cross-linking identifies leukemia inhibitory factor-binding protein as a ciliary neurotrophic factor receptor component. J. Biol. Chem. 268:7628-7631. Stahl, N., Boulton, T.G., Farruggella, T., Ip, N.Y., Davis, S., Witthuhn, B.A., Quelle, F.W., Silvennoinen, O., Barbieri, G., Pellegrini, S., Ihle, J.N. and Yancopoulos, G.D. (1994) Association and activation of Jak-Tyk kinases by CNTFLIF-OSM-IL-6 beta receptor components. Science 263: 9295. Stahl, N., Farruggella, T.J., Boulton, T.G., Zhong, Z., Darnell, J.E., Jr. and Yancopoulos, G.D. (1995) Choice of STATs and other substrates by modular tyrosine-based motifs in cytokine receptors. Science 267: 1349-1353. Stockli, K.A., Lotttspeich, F., Sendtner, M., Masiakowski, P., Carroll, P., Gotz, R., Lindholm, D. and Thoenen, H. (1989) Molecular cloning expression and regional distribution of rat ciliary neurotrophic factor. Nature 342: 920-923. Stockh, K.A., Lillien, L.E., Naher-Noe, M., Breitfeld, G., Hughes, R.A., Raff, M.C., Thoenen, H. and Sendtner, M. (1991) Regional distribution developmental changes and cellular localization of CNTF-mRNA and protein in the rat brain. /. Cell Biol. 115: 447^59. Symes, A.J., Rao, M.S., Lewis, S.E., Landis, S.C, Hyman, S.E. and Fink, J.S. (1993) Ciliary neurotrophic factor coordinately activates transcription of neuropeptide genes in a neuroblastoma cell line. Proc. Natl. Acad. Sci. USA 90: 572-576. Symes, A.J., Rajan, P., Corpus, L. and Fink, J.S. (1995) C/EBP-related sites in addition to a Stat site are necessary for ciliary neurotrophic factor-leukemia inhibitory factordependent transcriptional activation by the vasoactive intestinal peptide cytokine response element. J. Biol. Chem. 270: 8068-8075. Taga, T., Hibi, M., Hirata, Y., Yamasaki, K., Matsuda, T., Hirano, T. and Kishimoto, T. (1989) Interleukin-6 triggers the association of its receptor with a possible signal transducer, gpl30. Cell 58: 573-581. Taga, T., Narazaki, M., Yasukawa, K., Saito, T., Miki, D., Hamaguchi, M., Davis, S., Shoyab, M., Yancopoulos, G.D. and Kishimoto, T. (1992) Functional inhibition of hematopoietic and neurotrophic cytokines by blocking the interleukin 6 signal transducer gpl30. Proc. Natl. Acad. Sci. USAS9: 10998-11001. Takahashi, R., Yokoji, H., Misawa, H., Hyashi, M., Hu, J. and Deguchi, T. (1994) A null mutation in the human CNTF gene is not causally related to neurological diseases. Nature Genet. 7: 79-84.

308 Taniguchi, T. (1995) Cytokine signalling through nonreceptor protein tyrosine kinases. Science 268: 251-255. Thier, M., Simon, R., Kriittgen, A., Rose-John, S., Heinrich, P.C, Schroder, J.M. and Weis, J. (1995) Site-directed mutagenesis of human CNTF: functional analysis of recombinant variants. J. Neurosci. Res. 40: 826-835. Thornberry, N.A., Bull, H.G., Calaycay, J.R., Chapman, K.T., Howard, A.D., Kostura, M.J., Miller, D.K., Molineaux, S.M., Schmidt, J.A. and Tocci, M.J. (1992) A novel heterodimeric cysteine protease is required for interleukinIbeta processing in monocytes. Nature 356: 768-774. Ulenkate, H.J.L., Kaal, E.C.A., Gispen, W.-H. and Jennekens, F.G.I. (1994) Ciliary neurotrophic factor improves muscle fibre reinnervation after facial nerve crush in young rats. Acta Neuropathol 88: 558-564. Unsicker, K., Reichert-Preibsch, H., Schmidt, R., Pettmann, B., Labourdette, G. and Sensenbrenner, M. (1987) Astroglial and fibroblast growth factors have neurotrophic functions for cultured peripheral and central nervous system neurons. Proc. Natl Acad. Sci. USA 84: 5459-5463. Ware, C.B., Horowitz, M.C., Renshaw, B.R., Hunt, J.S., Liggitt, D., Koblar, S.A., Gliniak, B.C., McKenna, H.J., Papayannopoulou, T., Thoma, B., Cheng, L., Donovan, P.J., Peschon, J.J., Bartlett, P.P., Willis, C.R., Wright, B.D., Carpenter, M.K., Davison, B.L. and Gearing, D.P. (1995) Targeted disruption of the low-affinity leukemia inhibitory factor receptor gene causes placental, skeletal, neural and metabolic defects and results in perinatal death. Development \2\: 1283-1299.

Ciliary neurotrophic factor Williams, L.R., Manthorpe, M., Barbin, G., Nieto-Sampedro, M., Cotman, C.W. and Varon, S. (1984) High ciliary neurotrophic activity in rat peripheral nerve. Int. J. Dev. Neurosci. 2: 177-180. Winter, C.G., Saotome, Y., Levison, S.W. and Hirsh, D. (1995) A role for ciliary neurotrophic factor as an inducer of reactive gliosis, the glial response to central nervous system injury. Proc. Natl. Acad. Sci. USA 92: 5865-5869. Wong, v., Pearsall, D., Arriaga, R., Ip, N.Y., Stahl, N. and Lindsay, R.M. (1995) Binding characteristics of ciliary neurotrophic factor to sympathetic neurons and neuronal cell lines. J. Biol. Chem. 270: 313-318. Yamamori, T., Fukada, K., Aebersold, R., Korsching, S., Fann, M.-J. and Patterson, P.H. (1989) The cholinergic neuronal differentiation factor from heart cells is identical to leukemia inhibitory factor. Science 246: 1412-1416. Yin, T. and Yang, Y.-C. (1994) Mitogen-activated protein kinases and ribosomal S6 protein kinases are involved in signaling pathways shared by interleukin-11, interleukin-6, leukemia inhibitory factor, and oncostatin M in mouse 3T3Ll cells. J. Biol. Chem. 269: 3731-3738. Zhan, X., Bates, B., Hu, X. and Goldfarb, M. (1988) The human FGF-5 oncogene encodes a novel protein related to fibroblast growth factors. Mol. Cell. Biol. 8: 3487-3495. Zhang, F., Richardson, P.M., Holland, D.P., Guo, Q. and Tatton, W.G. (1995) CNTF or (-)-deprenyl in immature rats: survival of axotomized facial motoneurons and weight loss. /. Neurosci. Res. 40: 564-570.

Section III Factors Implicated in Neuronal Support and Repair

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 13

Melanocortins as factors in somatic neuromuscular growth and regrowth F.L. Strand, K.A. Williams, S.E. Alves, F.J. Antonawich, T.S. Lee, S.J. Lee, J. Kume and L.A. Zuccarelli Biology Department and Center for Neural Science, New York University, Washington Square, New York, NY 10003, USA

1. History of effects of melanocortins on the neuromuscular system Experiments in the 1960s indicated that exposure to cold stress increased the amplitude of peripheral nerve action potentials and muscle contractions and, unexpectedly, that this effect was intensified by adrenalectomy, leading to the inference that stress-released adrenocorticotropin (ACTH) was acting directly on the neuromuscular system without the intervention of adrenocortical hormones (Strand et al., 1962). Further specific evidence for a direct effect of ACTH on the neuromuscular system was provided by studies on peripheral nerve regeneration: administration of ACTH accelerated the return of motor and sensory function after sciatic nerve crush (Strand and Kung, 1980). At this time, synthetic peptide fragments of the naturally occurring ACTH 1-39 became available and, as behavioral studies by de Wied showed, despite their potent effects on the central nervous system these fragments did not evoke adrenocortical secretion (De Wied and Jolles, 1982). This permitted the separation of peptide and steroid effects in the experimental protocol and led to an intensive study of the structure-activity relationships of the various peptide fragments of ACTH, referred to as the melanocortins, as related to their effects on peripheral nerve regeneration (Bijlsma et al., 1983a). There is consistent evidence that administration of melanocortins accelerates both sensory and motor nerve regeneration, improves the pattern of motor unit formation in reinnervated

skeletal muscle, returns the morphology of the endplate towards that of the undamaged organ and permits a more rapid return of motor function and sensory responses (see Strand et al., 1989, 1990b, 1993a; Bar et al., 1990; Van der Zee et al.,1991). Developmental studies are equally convincing. ACTH peptide fragments administered to pregnant rats accelerate the development of the neuromuscular system, affecting the morphology, electrophysiology, motor behavior and muscle metabolism of the resulting pups (Smith and Strand, 1981; Saint-Come et al., 1982; Acker et al., 1984, 1985; Saint-Come and Strand, 1985; Frischer et al., 1985, 1988; Frischer and Strand, 1988; Rose and Strand, 1988; Rose et al., 1988; Rose and Strand, 1990). Recent studies indicate that regeneration following injury to the developing neuromuscular system of neonates also is positively affected by the melanocortins (Zuccarelli and Strand, 1990, 1991, 1992). Yet, despite intensive research during the last decade, the mechanisms by which the melanocortins exert such basic neurotrophic actions remain elusive. This chapter will discuss the evidence for melanocortins as factors in neuromuscular growth and regrowth and outhne some of the theories suggested for their wide-ranging neurotrophic influence. 2. Melanocortins The peptide sequences of ACTH 1-39 which do not stimulate the adrenal cortex, yet are effective neurotrophic factors, are limited to the amino acid

312

Melanocortins as factors in somatic neuromuscular growth and re growth

$4mi\m Y-MSH 1-25

Intermediate Lobe

a4m

16K Fragment

CUP 18-39

ACW m M l Itf'^dgyhln

ACTH Biosynthetic Intermediate

POMC

ACTH

rSlgnah

Anterior J

Lobe I

ACTH Biosynthetic Intermediate 16K Fragment

ACIHI'^

Y-LPH

B-LPH 1-91

p-IJ>H p-LPH 1-91 B-LPH 1-91

JHtflSHt*2D

Fig. 1. Processing of proopiomelanocortin (POMC) the precursor for the melanocortins and opiates in the mammalian pituitary gland. Processing occurs by proteolytic cleavage at sites of paired amino acids, some of which are shown here as dark bands. In both the anterior and the intermediate lobes, POMC is processed into an ACTH biosynthetic intermediate and into /3-lipotropin (y^-LPH 1-91). In the anterior lobe, subsequent processing yields the two biologically important products, ACTH 1-39 and )3-MSH. In the intermediate lobe, ACTH 1-39 is further processed to yield a-MSH (ACTH 1-13) and corticotropin-like intermediate peptide (CLIP). y3-MSH is derived, via y-MSH, from the 16K fragment of POMC. This hormone varies in length in different species. Also in the intermediate lobe, fi-LPU is processed to produce pendorphin 1-31. Many other small peptide fragments, of uncertain biological properties, are also produced, (from Strand et al., 1989).

fragments contained within the sequence ACTH 1-13 (a-melanocyte-stimulating hormone, aMSH) and are collectively termed melanocortins. 2.1. Localization, synthesis and processing ACTH is derived from a large polypeptide precursor, proopiomelanocortin (POMC) that also includes in its amino acid sequence both MSH and the related opiate )8-lipotropin (fi-LPK), the latter giving rise to the biologically active ^-endorphin (Mains and Eipper, 1980). POMC, a 31 kDa polypeptide, is synthesized in the anterior and intermediate lobes of the pituitary gland, which is the most important source of this prohormone. Other notable sources of POMC synthesis in the brain include the amygdala and the arcuate nucleus of the hypothalamus (Watson and Akil, 1980; Civelli et al., 1982). In addition, the structure of the POMC molecule, the POMC gene and the mRNA sequence for POMC have been elucidated (Roberts

and Herbert, 1977a, 1977b; Nakanishi et al., 1979). y-Aminobutryic acid (GAB A) and GAB A receptor agonists can directly inhibit POMC neurons in some brain regions (Blasquez et al., 1994; Garcia etal., 1994). POMC is processed in the anterior lobe of the pituitary gland to ACTH 1-39 and yS-LPH, which then may be cleaved further to desacetyl a-MSH and y3-endorphin, respectively (Fig. 1). However, most a-MSH (ACTH 1-13) is produced in the intermediate lobe of the pituitary, first by cleavage from ACTH 1-39 to desacetyl-MSH, then by subsequent acetylation to a-MSH (Eipper and Mains, 1980). In humans, who lack an intermediate lobe, POMC is processed in the anterior lobe. MSH is present in multiple forms within the central nervous system and structural differences between these molecules result in differential behavioral and physiological responses (De Wied and Jolles, 1982; Beckwith et al., 1989). There is as yet no evidence for the physiological processing of

F.L. Strand et al.

ACTH or MSH to the smaller melanocortins such as ACTH 4-9 or ACTH 4-10, although immunoreactivity to ACTH 4-10 has been demonstrated in the septal area and perimeter of the third ventricle, including the median eminence (Lee et al., 1992). Very limited amounts of immunoreactive aMSH are found in spinal motoneurons and motor endplates of skeletal muscle in normal, healthy adult animals but this is increased considerably in young rodents, in adults with inherited muscular dystrophy or motoneurone disease, or following nerve injury (Haynes and Smith, 1985; Hughes and Smith, 1989b). 2.2. Receptors Immunocytochemically identified binding sites for both ACTH and MSH are found throughout the brain and certain areas of the spinal cord but it is only recently that specific receptors for the melanocortins been characterized. A family of five homologous receptors for melanocortins (MC) has been cloned. The MCI receptor, which is the aMSH receptor, is expressed in melanocytes and melanoma tissue. The MC2 receptor is characteristic of the adrenal cortex and binds ACTH (Mountjoy et al., 1992). The MC3 and MC4 are brain receptors, localized to the hypothalamus and regions of the brain associated with behavioral effects of the melanocortins (Adan et al., 1994). Ganz et al. (1993) describe an MC3 receptor that recognizes all peptides with the heptapeptide melanocortin core equally and is distributed in a variety of tissues including the brain, placenta and gut. The MC5 receptor has been found in muscle, spleen and the pars tuberalis of the brain (Tatro 1990; Mountjoy et al., 1992: Roselli-Reyfuss et al., 1993). The corticosteroid secretory regions of the adrenal gland express the mRNA for the ACTH receptor but not for the MSH receptor. In murine melanoma, on the other hand, it is the mRNA for MSH that is expressed. As melanocortins in the nervous system and immune system differ from each other and from receptors on adrenocortical cells and melanocytes (De Wied and Jolles, 1982; Cannon et al., 1986; Tatro and Reichlin, 1987; Ta-

313

tro, 1990), it is clear that melanocortin receptors are tissue- and function-specific. In quantitative studies using a radiolabelled MSH tracer, MSH was shown to bind most intensely in the septal area, septohypothalamic nucleus, the bed nucleus of the stria terminalis and the medial preoptic area. Moderate binding was seen in hypothalamic structures including the median eminence, the ventromedial, dorsomedial, arcuate and paraventricular nuclei, and the lateral hypothalamus. These structures are profusely innervated by immunoreactive a-MSH- and ACTHcontaining fibers (Eskay et al., 1979; O'Donohue et al, 1979; Guy et al., 1980; Joseph, 1990). The median eminence is also well supplied with MSH receptors, supporting a role for this structure by which peripherally administered neuropeptides may penetrate to exert their central effects (Banks andKastin, 1988). Binding sites for ACTH 4-10 and the ACTH 4 9 analog Org 2766 have been localized in limited regions of the septum, caudate-putamen, preoptic area, hypothalamus, amygdala and hippocampus (Verhoef et al., 1977; Rees et al., 1980). Org 2766 binds to Schwann cells in vitro (Dyer et al., 1992, 1995) and to axonal sprouts or glia in the dorsal spinal cord, but only during the initial stages of sciatic nerve regeneration (Dekker and Tonnaer, 1989). Recent studies have suggested the specificity of MC receptors extends to the peptide fragments insofar as ACTH 4-9-NH2 is the core sequence able to activate both the MC3 and MC4 receptors, but that ACTH 4-9 more strongly activates the MC4 receptor whereas y-MSH displays selectivity for the MC3 receptor. It also appears that the excessive grooming response elicited by MC peptides is mediated by MC4 receptors. In addition, the potent ACTH 4-9 analog ORG 2766 does not activate, antagonize or bind to MC3 or MC4 receptors, indicating that additional MC receptors probably exist (Adan et al., 1994; Vandertop et al, 1994). Specific binding sites for [^^^IJACTH, as shown by autoradiography, have been located on the surface of muscle fibers of developing, regenerating and dystrophic murine skeletal muscles. Normal adult mouse muscle has very few ACTH receptors

314

Melanocortins as factors in somatic neuromuscular growth and re growth

(Hughes and Smith, 1988, 1989a, 1991). These observations fit well with the comparable restriction of beneficial melanocortin activity to the developing, regenerating or metabolically disturbed nerve and muscle. Melanocortins have little discernible influence upon the adult, healthy neuromuscular system.

1991). Acetylation of a-MSH has powerful effects on its biological properties (Eberle, 1988). These peptides have different potencies depending upon the parameter studied, i.e. electrophysiological, morphological, functional or biochemical, and whether the model is growth or regrowth (see reviews by Strand et al., 1989, 1991; Bar et al., 1990).

2.3. Structure-activity correlation

2.4. Mechanisms of action

As processed from POMC, ACTH 1-39 possesses full biological activity, stimulating the adrenal cortex and exerting potent neurotrophic influences upon components of both the central and peripheral nervous systems. ACTH 1-24 is as potent as ACTH 1-39, the sequence 25-39 apparently being limited to bestowing improved binding characteristics. a-MSH (ACTH 1-13) retains both neurotrophic and melanotropic attributes but has lost adrenocortico-stimulating activity. Structure-activity studies on ACTH peptide fragments by Bijlsma et al. (1981b) showed most neurotrophic potency to be in three ACTH-like peptides: ACTH 4-10, a trisubstituted analog of ACTH 4-9, Org 2766, and a-MSH. The ACTH 4 10 analog BIM 22015 has similar neurotrophic characteristics (Atella et al., 1992; Strand et al., 1993b). The structures of these peptides are shown in Table 1. The remarkable structural specificity of melanocortins is exemplified in a study which shows that a-MSH, [A^-leu^ D-Phe^l-a-MSH, desacetyl-a-MSH and Org 2766 are equally effective when tested in the foot withdrawal reflex, whereas ACTH 7-16 is without effect and ACTH/MSH 11-13 inhibits this response (Van der Zee et al.

Despite dependable evidence of the effectiveness of the melanocortins during periods of growth and regrowth of neural tissue, convincing proof of specific second messenger coupling, or other modes of action, is still lacking. In non-neural tissues, such as the adrenal cortex, ACTH binds to G proteins that activate adenylate cyclase and thus elevate cyclic AMP (cAMP) levels. Similarly, MSH acts through cAMP in melanoma cells and in the secretory cells of the lacrimal gland (Leiba et al, 1990; Salomon, 1990). Much of this evidence has been gathered from cloned murine and human receptors (Mountjoy et al., 1992). By raising the base level of cAMP and thus increasing the sensitivity of the assay, Florijn et al. (1992) have demonstrated the presence of an ACTH and/or MSH receptor coupled to adenylate cyclase in the brain. The striatum also selects cAMP as a second messenger (Wiegant et al., 1979, 1981; Horijn et al., 1991) but the melanocortins preferentially interact with striatal dopamine D2 receptors (Florijn et al., 1992). The action of melanocortins may involve cholinergic transmission. Acetylcholine synthesis in brain slices is increased by ACTH 1-39 and cur-

TABLE 1 Structure of related ACTH peptides

1 a-MSH CH3CO-SER ACTH 4-10 ACTH 4-9 Org 2766 BIM 22015

-TYR

-SER -MET -MET -MET (O2) -MET D-ALA

-GLU -GLU -GLU -GLU -GLN

-HIS -HIS -HIS -HIS -TYR

PHE PHE PHE PHE -PHE

-ARG -ARG -ARG -D-LYS -ARG

-TRP -TRP -TRP -PHE -TRP

10

11

12

13

-GLY -GLY

-LYS

-PRO

-VAL

-OH -GLY

-NH2

315

F.L. Strand et al.

tailed by hypophysectomy (Torda and Wolff, 1952). ACTH stimulates choline uptake by brain synaptosomes and accelerates choline release in certain brain areas (hippocampus, parietal cortex, medulla-pons, olfactory tracts and hypothalamus). Choline uptake, however, is inhibited by ACTH in synaptosomes from the anterior thalamus and cerebellum (Veals, 1979; Veals and Strand, 1979). The short melanocortin analogues Org 2766 and another ACTH 4-9 analog Ebiratide (Hoe 427) increase acetylcholine release in the striatum, hippocampus and frontal cortex of the rat (Geiger et al., 1987). The melanocortins may act as neuromodulators, affecting the release of neurotransmitters such as acetylcholine, dopamine and norepinephrine (Fiixe et al., 1973, van Loon, 1973; Versteeg and Wurtman, 1975; Keim et al., 1977, 1978; Keim, 1978; Strand and Smith, 1986; Versteeg et al., 1986; De Graan et al., 1990). In almost all cases, melanocortin effects on transmission and modulation are Ca^'^-dependent. An example of this dependence is shown in Fig. 2. We have evidence from other studies, involving perinatal administration of ACTH, of interaction between this stress hormone and the serotonergic and dopaminergic systems of the developing brain (Strand et al., 1990a; Segarra et al., 1991; Alves et al., 1993). Since both the dopaminergic and cholinergic systems may be manipulated by glutamine, an interaction between ACTH and the NMDA glutaminergic receptor has been suggested as another possible mechanism of action. Balanced glutaminergic neurotransmission is a necessary criterion to avoid neuronal damage and it also is involved in agerelated neuronal degeneration (Maragos et al., 1987; Wenk et al., 1991). Since ACTH analogs offer beneficial effects in lesion and aging studies associated with the glutaminergic system, ACTH may be acting to modulate NMDA receptor activation (Spruijt, 1992a, 1992b) Another postulated mechanism for melanocortins involves the phosphorylation of the growth-associated protein B-50 (GAP-43). B-50 is associated with both the growth and regrowth of neurites (Skene and Willard, 1981) and the growth of the presynaptic terminals of developing neuromuscular junctions (Verhaagen et al., 1988). There

CZl CONTROL 777\ ACTH (0.01

^lq/m\)

60 50 +

5

i

30 +

i

^i

Qi

^

20-110 0 mM

0.5 mM

1.0 mM

2.5 mM

5.0 mM

CALCIUM CONCENTRATIONS * p < 0.01

Fig. 2. Effect of ACTH on spontaneous radioactivity release from [^^C]choline prelabelled synaptosomes at various calcium concentrations. Prelabelled synaptosomes (P2 fraction, 0.1 ml) were incubated for 20 min. In all samples Ca^"*" was added in concentrations as indicated at the start of incubation. Each value represents the mean ± SEM of five determinations.*? < 0.01 (from Veals, 1979).

is a correlation between three factors that leads to the hypothesis that melanocortins are associated with B-50 levels; first, ACTH 1-24 has an inhibitory action on B-50 phosphorylation (Zwiers et al., 1976; Gispen et al., 1979; Aloyo et al., 1988); second, there is an immediate rise of B-50 to peak levels immediately after nerve injury (Van der Zee et al., 1989); and third, the time of greatest sensitivity to the beneficial effects of melanocortins is at the time of injury or immediately thereafter (Bijlsma et al., 1981a; Saint-Come and Strand, 1985, 1988). 2.5. Melanocortin administration 2.5.1. Timing and duration In working with the melanocortins, we have observed that they are most effective when administration commences at the time of, or very shortly after, nerve injury. The best pattern of administration appears to be one injection every 48 h for a period of 8 days following the nerve lesion (Verhaagen et al., 1986; Saint-Come and Strand, 1988). We have found that prolonged melanocortin treatment, beyond 2 weeks, may have deleterious effects on nerve and muscle (Saint-Come and

316

Melanocortins as factors in somatic neuromuscular growth and re growth

TABLE 2 Role of duration of peptide treatment in improving recovery from sciatic nerve crush 1-3 h

2 Days

8 Days

14 Days

21 Days

+*

++ +*

++ + 0*

+++ + 0

+ + 0

Data from electophysiology, sprouting, functional recovery and morphology. +++, Greatest effect; ++, moderate effect; +, slight effect; 0, no effect. Treatment commenced on the hour or day indicated (*) and was continued to 21 days. Reprinted from strand et al. (1989), with permission of the copyright holder, Pergamon Press Ltd., Oxford.

Strand, 1988; Strand et al., 1993a). Data are compiled in Table 2 to show that the effectiveness of the melanocortins reaches a peak between 8 and 14 days after peroneal nerve injury; after 21 days there is little difference between peptide-treated and saline-treated animals, probably due to the very rapid and efficient regrowth of crushed nerves in the rat. 2.5.2. Mode of administration There appears to be very little difference in effectiveness, whether the peptides are given intraperitoneally (Strand and Kung, 1980; Saint-Come et al., 1982; Saint-Come and Strand, 1985) or subcutaneously (Bijlsma et al., 1981a, 1981b; 1983a, 1983b). Microspheres or minipumps filled with the peptides are as effective as subcutaneous injections given twice daily (Dekker et al., 1987; Van der Zee et al., 1988). Local application of melanocortins to the site of injury through the use of the microporous Accurel® polypropylene tube is also effective, serving a double function in the case of nerve section, as the tube acts as a guide for the regrowth of the transected axons (Edwards et al., 1986). 2.5.3. Dosage The effective dosage of the melanocortins depends greatly upon the specific melanocortin, the parameter studied, and the physiological state i.e.

healthy or diseased, during growth or regrowth. Critical periods within periods of growth and regrowth exist, so that responses vary during early or late gestation, early or late regeneration. To compHcate this even further, the effective dosage of neuropeptides usually follows an inverted Ushaped curve. A higher dosage often has an opposite or inhibitory effect. Table 3 summarizes data showing the differential effectiveness of some melanocortins on the electrophysiological and sensorimotor responses, and of nerve terminal branching, in rats subjected to nerve crush. In general, a range of 0.1-10//g/kg per 48 h is appropriate for a dose-response investigation. One should be aware, however, in comparing data that some investigators express doses as yWg/kg whereas others express them as/^g/rat. 3. Neuromuscular growth 3.1. Normal growth of the neuromuscular system Innervation of embryonic muscle fibers occurs TABLE 3 Role of dosage in improving recovery from sciatic nerve crush Dosage (ag/kg/48 h) 0.10 intaperitoneal Dosage (Mg/kg/48 h) 0.75 subcutaneous

0.1

10.0

7.5

75.0

ACTH 1-24 ACTH 4-10 Org 2766* a-MSH (ACTH 1-13;) ACTH 11-24

++ ++ +++ ++ 0

0 ++ ++ ++ 0

0 0 0 0 0

Data from nerve termical branching Dosage (ug/kg/48hi.p.) ACTH (4-10) Org 2766

0.01

0.1

1.0

10.0

nd ++

0 +

+ 0

++

-

Data from electrophysiological parameters, sensorimotor recovery, and number of growing axons. +++, Greatest effect; ++, moderate effect; +, slight effect; 0, no effect; *ACTH 4-9 analog; nd, no data.

F.L. Strand et al.

317

Agrin appears to direct the aggregation of AChRs at both developing and regenerating neuromuscular junctions (see review by Hall and Sanes, 1993). The immature muscle fibers appear to be differentiated into fiber types prior to innervation although the final determination of muscle fiber type is the culmination of genetic cues (Brooke et al., 1971), hormonal environment and motor nerve stimulation (Buller et al., 1960; Brown and Lunn, 1988). The muscle fibers are organized into motor units, each innervated by one motor neuron (Fig. 3) and, in the adult, all the muscle fibers of a motor unit are of the same type. At this stage in neuromuscular development, polyneuronal innervation is still prevalent, with the endplate of the muscle fiber being innervated by more than one axon terminal (Redfern, 1970; Van Essen, 1982).

Fig. 3. A motor unit of the peroneal nerve is seen branching and innervating several EDL muscle fibers. 220x (from Frischer and Strand, 1988).

around gestation day (G) 16 or 17 in the rat. Although these motor axons bear no noticeable morphological relationship to each other, they soon form patterns that become the recognizable adult motor nerves (Hollyday, 1980; Bennett, 1981) and primitive synaptic contacts develop even prior to the formation of discrete muscles (Landmesser and Morris, 1975). The neuromuscular system of the rat does not complete its morphological and physiological development until the end of the second postnatal week. 3.1.1. During gestation By the time of innervation at G16, the acetylcholine receptors (AChRs) which had been scattered along the length of the immature muscle fibers, cluster at the areas which will become the neuromuscular junction (Anderson and Cohen, 1977), a phenomenon that is at least partially dependent upon the release of agrin from the approaching axon terminal (McMahan et al., 1980).

3.1.2. Neonatal development The most important developmental changes that occur in the first 2 weeks of postnatal life involve both the nerve and the muscle. Polyneuronal innervation is eliminated, so that each mature muscle fiber receives input from only one motor nerve terminal, and the surviving nerve terminal branches profusely within the endplate, which itself becomes folded more deeply into the muscle (Bevan and Steinbach, 1977; Steinbach, 1981, Fahim et al., 1983; Frischer and Strand, 1988). Biochemical maturation of muscle fiber types affects myosin isoforms (Pette and Staron, 1988) as well as the sarcoplasmic reticulum, permitting genetically programmed fast muscle fibers to attain their destiny, as seen by their physiological characteristics of faster, stronger contractions (Close, 1964). 3.2. Melanocortin influences on neuromuscular growth Perinatal administration of melanocortins, whether during gestation or in the susceptible first 2 weeks of postnatal life, accelerates all aspects of neuromuscular growth and function, so that the neonate attains both structural and functional neuromuscular maturation considerably sooner than its untreated littermates. By 3 or 4 weeks after birth, there is no longer any distinction between the pep-

318

Melanocortins as factors in somatic neuromuscular growth and re growth O — O SALINE • - - • ACTH 4 - 1 0 A A Org 2766

il

DAYS

Fig. 4. Increase in nerve terminal branching within EDL endplates following melanocortin administration. This increase is significant at 14 days following administration of ACTH 4-10 (10/ig/kg per 24 h) or Org 2766 (0.01 ^g/24 h) from day of birth to the day prior to the experiment. This effect is not seen at 7 days of age.

tide-treated and untreated rat pups, indicating that maturation has been accelerated but only to the normal physiological end-point. 3,2.1. Morphology Both ACTH 4-10 and Org 2766 effectively stimulate nerve terminal branching, as visualized by light and scanning electron microscopy, when

administered to rat neonates from the day of birth for 1, 2 or 3 weeks. There is an age-related increase in nerve terminal branching both prior to nerve entrance into the immature muscle (Frischer and Strand, 1988) and within the endplate, which is most evident at 2 weeks of age (Fig. 4). There is a considerable difference in tissue responsiveness to ACTH peptides: at 7 days of age, muscle fiber diameter and endplate area are actually reduced by ACTH 4-10 treatment and are subsequently unresponsive to this neuropeptide (Table 4). Org 2766, however, does not evoke these early deleterious effects on muscle, stimulating both endplate perimeter and nerve terminal branching at 7 days of age and encouraging nerve terminal branching still more impressively at 2 weeks of age. The effective dose for ACTH 4-10 is lO/^g/kg per 24 h s.c. whereas Org 2766 evokes the same stimulatory effect at a much lower dose, i.e. 0.01 /^g/kg per 24 h. When the ACTH 4-9 analog is administered at the higher dosage (lO^g/kg) it has a marked inhibitory action on nerve sprouting (Frischer et al., 1985, 1988; Frischer and Strand, 1988), demonstrating clearly the inverted U-shaped dose-response curve for melanocortins discussed in Section 2.5.3.

TABLE 4 Endplate measurements from EDL muscles of rats treated with ACTH/MSH 4-10 (lO^ug/kg), Org 2766 (0.01 /Wg/kg) or saline daily s.c. from day of birth Age 7 Days Saline ACTH 4-10 Org 2766^ 14 Days Saline ACTH 4-10 Org 2766^ 21 Days Saline ACTH 4-10 Org 2766^

Muscle fiber diameter (um)

End-plate area

End-plate perimeter (am)

32.1 ±0.7 29.4 ± 0.5^ 32.9 ± 0.9

1010.8 ±42.1 907.2 ± 27.7^ 1084.0 ±40.6

126.1 ±2.7 122.3 ±2.2 137.7 ±2.5^

38.2 ± 0.9 38.2 ± 0.9 38.5 ±0.9

1556.8 ±82.9 1631.9 ±63.6 1792.9 ± 104.8

160.5 ±4.0 159.0 ±2.9 166.9 ±4.8

55.1 ±1.2 53.9 ±1.4 52.3 ±1.1

2721.8 ±108.6 2686.8 ±101.8 2942.6 ±101.5

209.3 ± 4.2 211.1 ±4.1 220.4 ± 4.2

Values are means ± SE. n = 40 per subgroup. ^/'<0.05, ^P<0.01, ^P<0.01 versus saline for same age group, ^ACTH 4-9 analog.

F.L. Strand etal.

The time sequence for susceptibility to melanocortin exposure during neuromuscular growth corresponds with the critical period for neuromuscular junction maturation and ceases when polyneuronal innervation is eliminated. It also correlates with the electrophysiological and behavioral evidence of accelerated functional maturity, as discussed in the following paragraphs.

319

IHIIIfcilii ltiiiiA*i*il

GD 3-12

"^^

Innervation (GD 16)

GD 13-21

"^^

3.2.2. Electrophysiology When rat pups are exposed in utero to ACTH 4-10 during G 3-12, and tested at 2 weeks of age, a considerable maturation of muscle contractile responses is seen. Twitch amplitude and speed of contraction are increased, while rise time is shortened in the fast extensor digitorum longus (EDL) muscle when stimulated either directly, or indirectly through its nerve. After this critical period of sensitivity to melanocortin administration during the early phase of gestation, muscle permanently loses its responsiveness to melanocortins, whether they are given during the latter part of gestation (G 13-21), perinatally, or during adulthood (Gonzalez and Strand, 1981; Rose and Strand, 1988). These critical periods are summarized in Fig. 5. Mature muscle only becomes susceptible to neuropeptide treatment following denervation or myopathy, as discussed later in this review. Interestingly, nerve responsiveness is attained and culminates during the first 2 weeks of postnatal life (Smith and Strand, 1981; Saint-Come et al., 1982; Acker et al., 1985; Saint-Come and Strand, 1985), presumably due to the simultaneous structural maturation of the neuromuscular junction and enhanced nerve terminal branching. Once the nerve has matured, it loses its responsiveness to melanocortins, to regain it only if subjected to nerve injury or neuropathy. Thus both tissue specificity and chronological age are important determinants of melanocortin effectiveness.

nates, whereas 15 day old pups show both a shift in the normal development of hyperactivity from 15 days to 13 days (Acker et al., 1985) and, as shown in Fig. 6, a distinct hyperactivity at this time, with both total slight activity and total very active behavior increasing, and a corresponding decrease in immobility (Rose et al., 1988).

3.2.3. Motor behavior Behavioral studies of neonatal activity demonstrate again the age-dependent susceptibility to melanocortins. Daily administration of ACTH 4 10 (10/^g/kg s.c.) from the day of birth does not affect the spontaneous activity of 7 day old neo-

3.2.4. Growth of the serotonergic system Although not considered a part of the neuromuscular system, central serotonin (5-HT) neurons are intimately involved in the control of certain motor behaviors, including both female and male sexual reflexes in the rat. ACTH peptides have

PD 1-16 PD 18-Adult Muscle Atrophy after Nerve Trauma

Fig. 5. Time-based sensitivity to ACTH peptides. ACTH peptides affect muscle and nerve differentially, dependent on time of administration. Administration early (G 3-12) affects muscle but not nerve; later administration (G 13-21) affects the nerve positively but causes muscle atrophy. Postnatal treatment improves nerve only during neonatal days 1-16 but neither nerve nor muscle respond to peptide treatment as adults unless trauma ensues. All melanocortins improve nerve regeneration but only some prevent muscle atrophy.

320

Melanocortins as factors in somatic neuromuscular growth and regrowth Saline Day 7

Postnatal ACTH (4-10) Day 7

39%

42% 50%

55%

8%

Postnatal ACTH (4-10) Day 15

Saline Day 15

7% 38%

39%

31% 62%

DTI HTVA

• TSA

24% Fig. 6. Postnatal administration of ACTH 4-10 increases active behavior in neonates. Both TV A and TSA are significantly increased in 15 day old rat pups by melanocortin treatment (lOjug/kg s.c. daily from day of birth to the day prior to the experiment). This effect is not seen at 7 days of age. There is a marked decrease in total inactivity (TI) in peptide treated 15 day old rats, (from Rose and Strand, 1988).

been shown to alter the developing 5-HT system, both in culture (Azmitia and deKloet, 1987) and in vivo (Davila-Garcia et al, 1988; King et al., 1991; Alves et al., 1993). ACTH peptides appear to act as a trophic factor to these neurons, promoting neurite outgrowth and therefore increasing fiber density in the terminal fields. The extent and duration of these changes depend upon the ontogenetic period during which the 5-HT neurons are exposed to ACTH. We found that prenatal ACTH transiently increased serotonin (5-HT) innervation of the brainstem, as measured by high-affinity specific uptake of 5-HT, in 7 day old rats but this increase is reversed by 21 days of age (Fig. 7). In contrast, postnatal administration of ACTH significantly increases 5-HT uptake in the brainstem and the hippocampus at 7 days and 21 days (Fig. 8) (King et al., 1991) and in the hypothalamus at 7 days and at adulthood (80-90 days) (Alves et al, 1993). This increase in hypothalamic 5-HT innervation following early postnatal ACTH admini-

stration is associated with some deficits in the lordosis response, a female sexual reflex, among virgin female rats. Once again, the critical period of sensitivity to melanocortins is illustrated, although there may be an additional factor to be considered, i.e. that the postnatal serotonergic system may be more vulnerable to changes in its hormonal environment when the placenta no longer acts as a selective barrier. 4. Neuromuscular regrowth in adults 4.1. Normal neuromuscular regrowth While regeneration, or regrowth, is often viewed as a resurrection of growth processes, there are several important differences between these two phenomena. Sprouting during ontogeny occurs in a pristine environment, unsullied by prior innervation. The growing axons innervate the embry-

321

F.L. Strand et al. I I prenatal saline VTA prenatal ACTH 1 - 3 9

uj ?,

700

g t 600 %l

500

g E

400

g S

300

LJJ

O

^ E

200

^

100 0

^

a

7 day 21 day BRAINSTEM

Ii

7 day 21 day HIPPOCAMPUS

* p < 0.05 ** p < 0.01

Fig. 7. Serotonin uptake in the brainstem and hippocampus of 7 and 21 day old rats. Changes in serotonin high-affinity uptake (fmol/mg wet weight) in the brainstem and hippocampus of 7 and 21 day old neonates treated prenatally with ACTH 139 or saline (/i = 4 per subgroup). V < 0.05, **P < 0.01.

onic myotubes to form the original endplates and motor units. The situation differs considerably after nerve damage in adults for, although the original endplates are re-innervated, there is no evidence to suggest that the original nerves match up with the same endplates they had initially penetrated. The newly reformed motor units are considerably larger and fewer in number than before, and the terrain they occupy is constricted and compact (Kugelberg, 1981). Untreated, reinnervated motor units regain their contractile strength but are very susceptible to fatigue and lack fine control (Saint-Come and Strand, 1985, 1988). The denervated endplate loses the complex of nerve terminals that had branched within it, and is seen as an empty structure. These pre-existing, empty endplates then are re-innervated by one or more growing axons, reconstructing the developmental stage of polyneuronal innervation. These changes are illustrated in Fig. 9. Physiologically, early reinnervation is marked by polyphasic nerve and muscle potentials, due to the polyneuronal innervation. Muscle contractions are weak and slow, with a prolonged rate of rise to peak amplitude, all characteristics of the immature neuromuscular system. With time, more complete reinnervation occurs, and many of the contractile

parameters return to normal as extra nerve terminals are eliminated and muscle atrophy is reversed. However, due to the large size of the motor units, fine control of muscle responses usually is not reestablished. Muscle atrophy is the typical accompaniment to denervation, the degree of atrophy usually being dependent upon the length of time the muscle has been separated from its nerve and on the distance from the muscle at which nerve injury has occurred. These two factors are, of course , often interrelated but the length of the distal nerve stump has additional significance in that it apparently contains neurotrophic factors that maintain muscle health. In our work, we have sometimes used sciatic nerve crush, which requires 30-40 days for reinnervation to occur, but more often we have resorted to peroneal nerve crush close to the point of muscle innervation. This model permits reinnervation within about 9 days and is accompanied by a lesser degree of muscle atrophy than the sciatic crush model. Nerve crush has the additional advantage over nerve section that crush injury destroys all axonal fibers but leaves the connective tissue sheath intact. Consequently, most of the regrowing axons find their way without detour into the correct muscle, but not necessarily into the

I I postnatal saline 77P\ postnatal ACTH 1-39

^

ii^.gi 1500

I

m

7 day 21 day BRAINSTEM

I

m

I

Ml ^

7 day 21 day HIPPOCAMPUS

* p < 0.05 *• p < 0.01

Fig. 8. Serotonin uptake in the brainstem and hippocampus of 7 and 21 day old rats. Changes in serotonin high-affinity uptake (fmol/mg wet weight) in the brainstem and hippocampus of 7 and 21 day old neonates treated postnatally with ACTH 1-39 or saline (« = 4 per subgroup). *P < 0.05, **P < 0.01.

322

Melanocortins as factors in somatic neuromuscular growth and regrowth

F.L. Strand et al.

original muscle fibers of that muscle (de Medinaceli and Rawling, 1987). The age and species (Lieberman, 1971) and the gender (Kume-Kick and Strand, 1994) of the animal are among other factors that influence the extent and speed of nerve regrowth. While axonal transport and protein synthesis are essential for nerve regeneration (see review by Grafstein and Forman, 1980), the overall rate of fast transport is not affected by sciatic axotomy (Griffin et al, 1976; Bisby, 1978; CresciteUi et al., 1989). Despite this, there is an increase in the amount of small proteins and polypeptides, known as *growthassociated-peptides' (GAPs), destined to be incorporated in the regrowing plasma membrane and cytoskeleton (Skene and Willard, 1981; Bisby, 1988). One of these GAPs is B-50, also known as GAP-43, which is discussed in Section 2.4. The administration of selected members of the neurotrophin family as well as other welldefined neurotrophic factors can prevent the death of lumbar motoneurons following axotomy. Brain derived neurotrophic factor (BDNF), neurotrophin-3 (NT3), insulin-like growth factor (IGF), and to a lesser degree NGF, NT4-5 and ciliary neurotrophic factor (CNTF), can rescue these cells from axotomy-induced death (Li et al., 1994). There are several performance tests for the return of motor function. One of the most often used is the Sciatic Function Index (SFI) derived by de Medinaceli et al. (1982) from a series of measurements of the walking gait of the rat. This index has been modified by Bain et al. (1989) for analysis of foot prints after peroneal nerve injury, and is known as the Peroneal Function Index (PFI). We and others have found that the quantitative analysis of toespread and of print length is more reliable indicator of recovery from peroneal nerve crush than is the PFI (Hare et al., 1993: Strand et al., 1993b).

323

Return of sensory function can be evaluated through sensory reflexes involving foot or tail withdrawal from noxious stimuli, or by changes in the conduction velocity of sensory nerves (De Koning et al., 1986; Gerristen van der Hoop et al., 1988; Van der Zee et al., 1988). 4.2, Melanocortin influences on neuronal and neuromuscular regrowth Provided that the right dosage and pattern of administration of melanocortins are adhered to, there is little doubt that ACTH 4-10, Org 2766, BIM 22015 and a-MSH all improve the rate and quality with which regrowing nerves innervate their muscle target. However, there are distinct differences in the effectiveness of the specific melanocortins on the process of nerve regrowth and the degree of muscle atrophy. 4.2.1. Morphology Nerve crush severely reduces endplate area, perimeter and nerve terminal branching within the endplate. The administration of ACTH 1-39 increases the size of motor endplates in adrenalectomized rats (Strand and Kung, 1980) and this neurotrophic attribute is retained in all four of the ACTH peptide fragments tested. We have found that a-MSH is the most potent of the melanocortins, vigorously increasing nerve terminal branching, endplate area and perimeter, as measured by quantitative image analysis. ACTH 4-10 and BIM 22015 are effective in increasing nerve terminal branching and arborization within the endplate but do not alter other endplate parameters (Table 5). 4.2.2, Electrophysiology Nerve crush destroys all muscle responses to indirect stimulation until reinnervation occurs around 9 days after peroneal nerve crush. We test

Fig. 9. Nerve terminal branching within endplates of EDL muscle 14 days after peroneal nerve surgery, (a) Shamdenervated + saline, 200 x; (b) nerve crush, saline treated, 200x; (c) nerve crush, saline treated, 600 x; (d) nerve crush + BIM 22015 0.1/ig/kg600x.

324

Melanocortins as factors in somatic neuromuscular growth and re growth

TABLE 5 Effects of admistration of peptides (lOyUg/kg/48 h) on endplate parameters from EDL muscles 15 days after crush lesion Parameters

Saline

ACTH4-10

a-MSH

Nerve terminal branching (um) Perimeter Area (um-^)

256.6 ± 4.3 232.0 ± 6.0 3501.2 ±200.3

336.8 ± 12.5^ 234.8 ± 5.9 3572.0 ±188.5

369.3 ± 17.5^ 255.7 ± 4.7^ 4091.9 ±152.4^

n = 36; mean ± SE. ^P<0.001; V < 0.01; ^P<0.001 versus sham-crushed saline.

the responses of the muscle through the injured nerve at 5, 7 and 9 days after the lesion to detect any acceleration of the regrowth process, and the consequent improvement in muscle performance. All of the melanocortins improve tetanic ampli-

tude, rate of rise to peak of tetanus, contraction time and resistance to fatigue of indirectly stimulated muscle (Fig. 10). There appears to be little difference in the efficacy of different melanocortins on these parameters but more detailed experi-

Denervated + Saline

0.00

0.40

0.80

1.20

1.60

0.80

2.00

Denervated + ACTH 4 - 1 0

0.00

0.40

0.80

2.00

1.20

Time (sec)

Time (sec)

1.20

Time (sec)

1.60

2.00

Denervated + BIM 22015

0.00

0.40

0.80

1.20

1.60

2.00

Time (sec)

Fig. 10. Representative traces showing peak amplitude and decline from peak amplitude of EDL muscle during peroneal nerve stimulation with 300 pulse-pair units of supramaximal strength 11 days after peroneal nerve crush, (a) Intact muscle; (b) denervated + saline; (c) denervated + ACTH/MSH 4-10 (40^g/kg per 48 h); (d) denervated + BIM 22015 (40/^g/kg per 48 h). Note the difference in scale of grams pull from the intact muscle. Peak from end function; UNKELSCOPE. (from Strand et al., 1993b).

325

F.L. Strand et ah TABLE 6 Maximum rate of development of tension (dP/di) of EDL muscle in rsponse to increasing numbers of pulse pairs 11 days after peroneal nerve crush and treatment with ACTH peptides Intact

Saline

ACTH 4-10

BIM 22015

Mean ± SE

Mean ± SE

Mean ± SE

Mean ± SE

741 ± 72 1450 ± 224 1902 ±307 2024 ± 340 2551 ±217 2721 ±280 2626 ± 288

334 ± 54 641 ± 198 966 ±189 1007 ± 185 977 ± 285 888 ± 247 851 ±249

395 ± 75 818 ±273 1037 ±276 1163 ±357 1352 ±357 1305 ±277^ 1302 ±221^

544 ± 44^'b 1022 ±205^ 1157 ±241 1420 ±338^ 1771 ±295^ 1715 ±394^ 1587 ± 350^

Number of pulse pairs

1 2 3 4 10 20 40

Crush denervated

Stimulated indirectly at 2x threshold; /i = 5; mean±SE. All crush denervated values are significantly lower than intact values (P<0.05). In denervated muscle, tension develops more rapidly following BIM 22015 treatment than in the saline controls or in rats treated with ACTH 4-10. Dosage was lOyUg/kg of peptide administered immediately following sugery and every 48 h therafter. ^P<0.05 versus denervated saline controls; "P<0.05 versus denervated ACTH 4-10 treated rats.

merits indicate that ACTH 4-10 and Org 2766 promote greater endurance and resistance to fatigue than is seen in muscles treated with BIM 22015 (Strand et al., 1993b). On the other hand, BIM 22015-treated muscles respond better when challenged with an increased demand, such as an increased number of pulse pairs as stimuli (Table 6). A clearer distinction between the melanocortins appears when muscle atrophy is appraised. Peroneal nerve crush results in a marked decrease in muscle fiber diameter measured 5 and 7 days after the lesion. Treatment with BIM 22015 (lO^g/kg per 48 h), but not with a lower dosage, prevents muscle atrophy, the peptide-treated muscles retaining 92% of their precrush diameter. At this same dosage, ACTH 4-10 does not alleviate denervation atrophy (Fig. 11). This interesting example of structural changes in the melanocortins resulting in differences in neurotrophic and myotrophic potencies may be attributed to modifications of BIM 22015 at the N-terminus (D-Ala-Gln-Tyr-Phe) being responsible for the positive myotrophic effects. Thus the C-terminal fragment (Phe-Arg-Try-Gly) becomes accountable for the attenuating effect. Wolterink et al. (1991) have described a similar division of behavioral potencies in Org 2766.

4.2.3. Motor and sensory function Melanocortin treatment accelerates the return of the normal SFI following sciatic nerve crush, indicating that the walking gait is approaching the

Saline

BIM 22015

BIM 22015 10.0^g/kg

ACTH 4-10 10.0(ig/lcg

Fig. 11. EDL muscle fiber diameter 7 days after peroneal nerve crush Saline treatment (black column); BIM 22015 0.1/ig/kg (open column); BIM 22015 10|ng/kg (hatched column); or lOyMg/kg ACTH (cross-hatched column) every 48 h from time of nerve injury. Data are expressed as percent of muscle fiber diameter of sham nerve crush group (100%). n = 240 fibers per group, six animals per subgroup. Bars represent SEM. P < 0.01 versus other denervated groups. BIM 22015 lO^g/kg is not significantly different from sham nerve crush. All other nerve-crush groups differ significantly from sham nerve crush (P < 0.05) (from Strand et al., 1993b).

326

Melanocortins as factors in somatic neuromuscular growth and regrowth

Standard values. This technique has been extensively used to evaluate structure-activity studies by Gispen and his colleagues (Bijlsma et al., 1981b). The effectiveness of melanocortin therapy is limited, however, by the duration of the treatment as we found that continuing melanocortin administration for 21 days following nerve injury delayed or prevented the reappearance of the normal SFI. The most effective duration of treatment was for the initial 8 days immediately following nerve injury. In a short-term study in which ACTH 4-10 and a-MSH (lOjug/kg per 48 h) were administered for only 8 days but the animals were tested over 21 days, the peptides were equally effective, with significant positive effects 14 days after nerve crush. In the long-term experiment, during which the rats received peptide treatment for 20 of the 21 days following nerve crush, only a-MSH was effective and this action was significant only at 18 days after the lesion (Strand et al, 1992, 1993c). Another assessment of return of normal motor function is through the use of several parameters of the PFI: toespread and print area as measured from paw prints obtained when the animal walks up an inclined path lined with white paper. The paws are dipped in non-toxic ink and the resulting prints analyzed by an image analysis system. This technique permitted us to evaluate three ACTH 4 10 analogs, BIM 22028, 22029 and 22030 as well as a-MSH (Strand et al, 1993c). In this study we used higher dosages than previously (40 ju g/kg per 48 h) and were able to distinguish between the three BIM analogs, in that only BIM 22029 effectively improved both print area and toespread. Interestingly, the usually potent a-MSH was ineffective at this high dosage, indicating again that elevated dosages may have a negative effect on certain parameters. Using the foot withdrawal reflex as an indicator of return of sensory function following nerve crush. Van der Zee et al. (1991) compared the competence of various melanocortins: a-MSH, (A^leu^, D-Phe7)-a-MSH, desacetyl-a-MSH and the ACTH 4-9 analog Org 2766. They found no differences in their effectiveness in eliciting the reflex. Other putative neurotrophic factors such as

gangliosides and nimodipine also were effective in this test but isaxonone and thyrotropin-releasing hormone were not, although the latter two factors accelerate maturation and neurite outgrowth in cultured fetal neurons (Hugelin et al., 1979; Azmitia, 1989). In addition, as hypophysectomy did not affect the time needed for recovery, the authors conclude that endogenous pituitary melanocortins were unlikely to be involved in normal nerve regrowth. However, as shown by Crescitelli et al. (1989), hypophysectomy slows the rate of axonal outgrowth and exogenous melanocortins do not improve this parameter. Thus differences in outgrowth of the injured sciatic nerve complicate the interpretation of the experiments by Van der Zee and her colleagues and, in a later paper from the same laboratory, Plantinga et al. (1995) demonstrate that the administration of an a-MSH anatagonist further slows the delayed nerve recovery following hypophysectomy. These investigators interpret this as evidence for the involvement of an endogenous a-MSH-like peptide in peripheral nerve. In experiments on an animal model of experimental allergic neuritis, a demyelinating disease, treatment with Org 2766 protected the myelinated nerve fibers of the sural nerve from degeneration (Duckers et al., 1994), an action similar to that observed for lesions of the substantia nigra (see Section 6) and for sensory and glial cells in culture (see Section 7). Not only peripheral sensorimotor neuropathies are beneficially affected by the ACTH 4-9 analog Org 2766 but autonomic neuropathies, such as the crushed parasympathetic fibers of the oculomotor nerve are responsive to treatment with this neuropeptide, especially in the initial stages of regeneration. However, if the oculomotor nerve is sectioned, systemic treatment with Org 2766 improves neither the rate nor quality of functional recovery as determined by pupil diameter (Vandertop et al., 1994). 5. Neuromuscular regrowth in neonates The age at which nerve injury is sustained plays a significant role in the extent to which full recovery

F.L. Strand et al.

327

TABLE 7 EDL muscle fiber diameter following nreve crush at 2 days of age and peptide treatment

Diameter

Uncrushed

Saline crush

Z-MSH crush

ACTH 4-10 crush

38.15 ±0.89

32.49 ± 0.73*

40.41 ± 0.49

38.47 ±0.61

Dosage of peptides was 10 /ig/kg/48 h for 8 days starting at time of nerve crush. Fiber diameter measured 2 weeks after nerve crush. *P < ANOVA. Saline versus all other groups.

is attained. In humans, recovery is generally better in children than in adults (Sutherland, 1978) and this appears to be true for other species such as rodents (Black and Lasek, 1979; Pestronk et al., 1980). However, the early postnatal period is more vulnerable and 2 week old animals subjected to nerve crush recover poorly in comparison to older neonates and adults (Bharali and Lisney, 1990; Connold et al., 1992). The administration of selected members of the neurotrophin family as well as other well-defined neurotrophic factors can prevent the death of lumbar motoneurons following axotomy. Our interest, of course, is to compare the effectiveness of melanocortins on nerve regrowth in animals of different ages. In a group of rats varying in age from 6-7 weeks to 20 months of age. Van der Zee et al. (1991) found that the younger the rat, the faster the recovery of sensorimotor function. As neonates are most sensitive to nerve injury, we developed a new model in which 2 day old rat pups are subjected to sciatic nerve crush and treatment with ACTH 4-10 or a-MSH (10/^g/kg per 48 h for 8 days). Their recovery is then monitored in terms of morphological, electrophysiological, behavioral and metabolic characteristics.

maximum by this time (Zuccarelli and Strand, 1990). Muscle fibers also atrophy following neonatal denervation. Treatment with ACTH 4-10 but not with a-MSH, exacerbates the muscle atrophy in 7 day old neonates. By 15 days, however, both melanocortins effect an increase in muscle fiber diameter, overcoming the denervation atrophy seen in the saline-treated controls. A comparison of muscle fiber diameters from 15 day old lesioned neonates treated with melanocortins shows them to be comparable in size to normal saline-treated pups of the same age (Table 7). 5.2. Metabolism During the normal maturation of fast muscle fibers there is a switch from oxidative to glycolytic metabolism. Seven and 15 days after nerve crush, glycolytic metabolism predominates in the EDL muscles of those nerve-crushed neonates treated with either ACTH 4-10 or a-MSH (Zuccarelli and Strand, 1991; Strand et al, 1993c). By 3 weeks of age, the saline-treated controls have caught up and glycolytic metabolism is comparable in all nerveinjured groups. Thus melanocortins appear to be capable of accelerating maturation even during the harsh period of nerve regrowth after injury.

5.7. Morphology 5.3. Electrophysiology In the 1 week old neonates, interior endplate branching is increased by ACTH 4-10 whereas aMSH treatment increases this parameter as well as endplate area and perimeter. By 2 weeks of age, both peptides increase all measurements except interior branching which apparently has reached its

Melanocortins accelerate reinnervation of the EDL muscle following sciatic nerve crush in 2 day old neonates. Few of the nerve-injured neonates treated with saline can respond to electrical stimulation of the nerve, proximal to the lesion, 33 days

328

Melanocortins as factors in somatic neuromuscular growth and regrowth

after nerve crush. Fifty percent of the lesioned neonates treated with a-MSH respond as soon as 13 days post-lesion, and by 19 days 93% of these pups react. ACTH 4-10 is less effective but nevertheless speeds up regrowth in that the neonates respond to nerve stimulation by 19 days postlesion. The improvement in regrowth is reflected in the improved electrophysiological parameters of the EDL muscle in the melanocortin-treated, injured neonates. Many of the contractile parameters of both the twitch and tetanus are in the normal range: force amplitude, contraction rate, halfrelaxation time, post-tetanic twitch amplitude, and rate of rise (Zuccarelli and Strand, 1992). Of considerable significance is that a-MSH also increases the number of motor units, presumably restoring fine motor control in the traumatized, developing neuromuscular system. 5.4. Motor behavior A test for motor behavior in the neonate consists of having the pup pull itself up on to a platform by its forelimbs, then raise the hindlimbs to ascend the platform Bregman and Kunkel-Bagden, 1988). Saline-treated, nerve-injured neonates require twice as long as uninjured pups to perform this task whereas the lesioned pups treated with ACTH 4-10 or a-MSH perform as effectively as the uncrushed controls (Zuccarelli and Strand, 1991; Fig. 12). 6. Melanocortins and central neurons controlling neuromuscular responses Regrowth of central neurons is difficult to demonstrate and much of the functional recovery from central lesions following melanocortin administration may be compensatory, (Isaacson and Poplawsky, 1983; Hannigan and Isaacson, 1985; Luneburg and Flohr, 1988) protective (Spruijt, 1992b) or due to the development of denervation sensitivity (Wolterink et al., 1990). Unilateral destruction of the substantia nigra in rats, using the neurotoxin 6-hydroxydopamine, results in deficits in motor behavior, which are improved by the

PLATFORM PULL-UP 15 DAYS POSTNATAL

1.500

0.000 NORM SHAM * ANOVA p < 0.05 saline vs all other groups

MSH

ACTH

SAL

Fig. 12. The platform pull-up test. Suspended pups are challenged to lift their hindlimbs to a platform. Saline-treated pups take almost twice as long to perform this task as do normal pups. a-MSH- and ACTH 4-10 (40/^g/kg per 48 h per 8 days) treated pups have times similar to those of normal and sham pups. *P < 0.05 by ANOVA and Newman-Keuls posthoc comparisons.

administration of Org 2766 (lO/^g/kg per 24 h) starting immediately after lesioning (Antonawich et al., 1992). We have found that this improvement appears to be due to an acceleration of denervation sensitivity and an increase in the intensity of dopamine immunoreactivity in the substantia nigra (Fig. 13), supporting the concept of melanocortin protection of threatened neurons. 7. Melanocortins and nerve cells in culture Trophic effects of melanocortins on neurons in vitro can be demonstrated by increased cell survival, accelerated neurite outgrowth and increased neurite length, the establishment of synaptic contacts and specificity of cell types as substrate. Van der Neut et al. (1988, 1990) using the marker B-50 as an indicator of neuronal regrowth, showed in both dissociated spinal cord cells and spinal slices that while a-MSH elevates B-50 levels, both Org 2766 and ACTH 4-10 are ineffective. When neurite outgrowth was the indicator, however, both aMSH and ACTH 4-10 have positive actions. Using the three-chamber tissue culture system (Campenot, 1977) we tested the neurotrophic properties of ACTH 4-10, Org 2766 and BIM 22015 on dissociated, mixed cells of the spinal

329

F.L. Strand et al.

SUBSTANTIA NIGRA immyr^ocylochemistry 1:1500 anti-TH lAccyratt)

1"""""^?*'' *

1 "-

^v'

X .' o ..

;:'•••«•-^-i'

^^

'^T'*-

^.

I^V;^ '

wL \ \ 5'^-\ :v^^\-, ./. ~\ ; . ,'\

i\,

^ ^ ,-^^* - » ." ^'""V ~

9h

.^*^

1 ^'i^3

12;

P>i W

k' *

" '

^ ^ ^ ^

^ }5*>^^ffl

•PI

;

DA¥ s ORG iim puptrnf) ^ w f tmm.

Fig. 13. Immunoreactivity of substantia nigra slices for tyrosine hydroxylase (TH). Thick rat brain slices (40yum) immunocytochemically stained for TH at 6 days after unilateral 6-hydroxydopamine administration (left hand panels) and treated with either saline or Org 2766 (lOjug/kg per 24 h). There is an enhanced immunoreactivity to anti-TH antibody 1:1500 (accurate) in Org 2766 treated animals (n = 4) compared to treatment with saline (n = 4).

cord and compared the effects of these melanocortins with that of nerve growth factor (NGF). We found that while both ACTH analogs (Org 2766 and BIM 22015) promoted neurite outgrowth, ACTH 4-10 did not promote neurites in this system and that this response was independent of NGF (Lee et al., 1993; Strand et al., 1993a). Using the conventional culture dishes, ACTH 4-10, BIM 22015 and Org 2766, stimulated neurite promotion and increased cell survival of the spinal cord cells. While the spinal cord cells perished by 10 days in vitro in the absence of NGF, the ACTH 4-10 and its analogs supported cell survival up to 14 days in vitro. ACTH peptide

fragments may exert their neurotrophic effects through different mechanisms. In the experiments utilizing the three-chambered system. BIM 22015 and Org 2766 were highly effective in inducing neurite outgrowth into the peptide-containing media of the side chambers, indicating a chemotropic effect. In contrast, ACTH 4-10 was completely ineffective on neurite outgrowth under these conditions but was perfectly capable of sustaining cells in the conventional culture dishes, implying an exclusive and direct action on the cell soma to sustain and maintain neurons. This supports the existence of anterograde trophic signalling as well as a local mode of action of ACTH peptides di-

330

Melanocortins as factors in somatic neuromuscular growth and re growth

rectly on the soma, sustaining the vitality of the neurons and subsequently stimulating neurite formation and growth. These neuron-supporting actions of the peptides may be mediated by subtypes of the melanocortin (MC) receptors which were recently cloned (Chhajlani and Wikberg, 1992; Mountjoy et al., 1992; Ganz et al, 1993; RoselliRehfuss et al., 1993). The presence of MC receptors in spinal cord has yet to be determined. 8. Conclusions Melanocortins, non-corticotropic fragments of ACTH, include a-MSH, ACTH 4-9, ACTH 4-10 and several synthetic analogs, all of which affect the growth and regrowth of peripheral nerve and the neuromuscular synapse, and some of which also affect denervated muscle itself. Perinatal administration of melanocortins accelerates neuromuscular maturation, as demonstrated by morphological, electrophysiological and behavioral parameters, but the susceptibility of muscle to melanocortin treatment ceases during gestation at the time that neural response to these peptides is initiated. The accelerated maturation induced by the melanocortins coincides with the critical first 2 weeks of postnatal life, after which time the normal, healthy neuromuscular system is impervious to these peptides and the difference between peptide-treated rats and saline-treated controls rapidly disappears. Serotonergic neuronal pathways that subserve motor behaviors, such as female and male sexual reflexes, also are affected by the melanocortins, which appear to act as trophic factors during development, increasing 5-HT neurite outgrowth and fiber density and subsequently adversely affecting adult sexual behavior in both sexes. Sensitivity to melanocortins again is clearly time-dependent, the age at which exposure occurs determining the extent and permanence of the resulting serotonergic changes. In animal models of nerve regrowth in adults following crush injury, the melanocortins improve both the initiation of sprouting from damaged axons and the pattern by which the newly growing nerve terminals reinnervate denervated muscle.

Peptide-treated regenerating nerves form smaller and more efficient motor units than those regrowing with only saline treatment. Peptide-treated, reinnervated endplates appear normal, with extensive interior nerve terminal branching. Electrophysiological parameters of indirectly stimulated muscle recovering from nerve crush are considerably improved by the melanocortins. In addition, tests of motor and sensory responses show that functional recovery is more rapid in melanocortintreated animals. In all these models of nerve regrowth, melanocortin treatment must commence as close to the time of injury as possible, implying that it is the initial stages of nerve sprouting that are primarily affected by these neuropeptides. In animal models of human neuropathies, including sensorimotor, demyelinating and autonomic neuropathies, melanocortin administration appears to exert a protective function, preventing further damage while not completely restoring normal structure or function. Nerve regrowth in neonatal rats appears to be even more sensitive to melanocortin administration than regrowth in the adult rat. Endplate morphology, muscle fiber diameter, muscle metabolism, motor unit size and electrophysiological parameters all respond positively to these peptides by 15 days of age, although there are variations in response during the first week of postnatal life. Melanocortins appear to protect damaged central neurons from neurotoxins, as demonstrated in a model employing unilateral destruction of the substantia nigra. Accelerated denervation sensitivity and increased dopamine immunoreactivity in the substantia nigra account for the melanocortininduced improvement in motor behavior of these injured rats. Similarly, dorsal root ganglia neurons subjected in vitro to cisplatin toxicity are protected to some degree by the ACTH 4-9 analog. This protection is specific in that it does not extend to Schwann cells or satellite cells, nor does it protect any of these cells from taxol toxicity. Throughout this review, it has been emphasised that there are differences in potency among the melanocortins, with a-MSH being the most consistently effective peptide, regardless of the parameter investigated. ACTH 4-10 and the ACTH 4-9

331

F.L. Strand et al.

analog Org 2766 vary in potency but their positive effects, whether during growth or regrowth, are purely neurotrophic. Prolonged administration of these peptides has deleterious effects, particularly on muscle fiber diameter. The ACTH 4-10 analog BIM 22015, however, while a less potent neurotrophic factor, prevents or retards denervation atrophy of muscle. These disparate attributes of the melanocortins are clearly displayed by the responses of neurons in tissue culture: when B-50 is used as a marker of neuronal regrowth, only aMSH is effective on dissociated spinal cord cells and spinal slices. In the 3-chamber tissue culture system, both Org 2766 and BIM 22015 promote neurite outgrowth whereas ACTH 4-10 is ineffective. Not only are there specific differences in effectiveness of the melanocortins, but the dosage, pattern of administration and duration of treatment are all important variables. Thus an understanding of the mechanism of action of the melanocortins becomes of supreme importance, both for our understanding of physiological regulation of neuronal growth and regrowth and for the design of new analogs appropriate for the treatment of specific neuromuscular disorders. Present evidence links melanocortins variably to cAMP second messenger systems, acetylcholine metabolism, and B-50 phosphorylation. Other neurotransmitter systems may be involved, such as the serotonergic and dopaminergic systems. While binding sites for ACTH and MSH are found throughout the brain and certain areas of the spinal cord, it is only recently that specific receptors for the melanocortins have been characterized. Neuronal melanocortin receptors differ from those found in the adrenal cortex and on melanoma cells, suggesting that there also may be differences in receptors in nerve and muscle that would account for the marked diversity in response to melanocortins characteristic of these two tissues. Melanocortin effectiveness may vary with the metabolic milieu, the stage of development of the target tissue, and critical interactions with other growth factors, including the sex steroids. In both growing and regrowing systems, it is clear that the melanocortins enhance growth processes, but con-

strain these dynamic, plastic systems to normal physiological limits. Melanocortins have little effect on the more stable, healthy, full-grown neuromuscular system. While emphasizing differences between neurotrophic neuropeptides, it is nevertheless obvious that many neuropeptides are multifunctional, and that one neuropeptide growth factor may effectively replace another, resulting in a fail-safe redundancy that protects vital physiological processes. Acknowledgements We are grateful for the support given to work in the authors' laboratory by Biomeasure, Inc., Organon B.V. and The Council for Tobacco Research. References Acker, G.R., Frischer, R.E. and Strand, F.L. (1984) ACTH modulation of the developing neuromuscular system as seen through three different perspectives. Ann, N. Y. Acad. ScL 435: 370-375. Acker, G.R., Berran, J.R. and Strand, F.L. (1985) ACTH neuromodulation of the developing motor system and neonatal learning in the rat. Peptides 6: 41-49. Adan, R.A., Cone, R.D., Burbach, J.P. and Gispen, W.H. (1994). Differential effects of melanocortin peptides on neural melanocortin receptors. Mol. Pharmacol. 46: 11821190. Aloyo, V.J., Zwiers, H., DeGraan, P.N.E. and Gispen, W.H. (1988) Phosphorylation of the neuronal protein kinase C and substrate B50 in vitro assay conditions alter sensitivity to ACTH. Neurochem. Res. 13: 343-348. Alves, S.E., Akbari, H., Azmitia, B.C. and Strand, F.L. (1993) Neonatal ACTH and corticosterone alter hypothalamic monoamine innervation and reproductive parameters in the female rat. Peptides 14: 381-386. Anderson, M.J. and Cohen, M.W. (1977) Nerve-induced and spontaneous redistribution of acetylcholine receptors on cultured muscle cells. J. Physiol. (London) 268: 757-773. Antonawich, F.J., Akbari, H.M., Azmitia, B.C. and Strand, F.L. (1992) The effects of an ACTH 4-9 analog, Org 2766, on 6-OHDA lesioning of the substantia nigra. Soc. Neuro-

sci.Abstr. 18:629. Atella, M.J., Hoffman, S.W., Pilotte, M.P. and Stein, D.G. (1992) Effects of BIM 22015, an analog of ACTH 4-10, on functional recovery after frontal cortex injury. Behav. Neural Biol. 51: 151-166. Azmitia, E.C. (1989) Nimodipine attenuates toxicity by MDMA, glutamate and caffeine on serotonergic neurons: evidence for a generic model of calcium toxicity. In: J. Tra-

332

Melanocortins as factors in somatic neuromuscular growth and re growth

ber and W.H. Gispen (Eds.), Nimodipine and Central Nervous Function, Schaattauer, Stuttgart, pp. 141-160. Azmitia, E.G. and de Kloet, E. (1987) AGTH neuropeptide stimulation of serotonergic neuronal maturation in tissue culture: modulation by hippocampal cells. Prog. Brain Res. 72:311-318. Bain, J.R., Mackinnon, S.E. and Hunter, D.A. (1989) Functional evaluation of complete sciatic, peroneal, and posterior tibial nerve lesions in the rat. Plast. Reconstr. Surg. 83: 129-136. Banks, W.A. and Kastin, A.J. (1988) Review: interactions between the blood-brain barrier and endogenous peptides: emerging clinical implications. Am. J. Med. Sci. 295: 459465. Bar, P.R.D., Schrama, L.H. and Gispen, W.H. (1990) Neurotropic effects of ACTH/MSH-like peptides in the peripheral nervous system. In: D. De Wied (Ed.), Neuropeptides: Basics and Perspectives, Elsevier, New York, pp. 175211. Beckwith, B.E., Tinius, T.P., Hruby, V.J., Al-Obeidi, R, Sawyer, T. and Affholter, J.A. (1989) The effects of structureconformation modifications of melanotropin analogs on learning and memory: D-amino acid substituted linear and cyclic analogs. Peptides 10: 361-368. Bennett, M.R. (1981) The development of neuromuscular synapses. Proc. Aust. Physiol. Pharmacol. Soc. 12: 41-58. Beven, S. and Steinbach, H. (1977) The distribution of abungarotoxin binding sites on mammalian skeletal muscle developing in vivo. J. Physiol. 267: 195-213. Bharali, L.A.M. and Lisney, S.J.W. (1990) Success of regeneration of peripheral nerve axons in rats after injury at different postnatal ages. J. Neurol. Sci. 100: 203-210. Bijlsma, W.A., Jennekens, F.G.I., Schotman, P. and Gispen, W.H. (1981a) Corticotropin (ACTH) like peptides stimulate peripheral nerve regeneration. In: M.W. Van Hof and H. Mohn (Eds.), Functional Recovery from Brain Damage, Elsevier, Amsterdam, pp. 411-416. Bijlsma, W.A., Jennekens, F.G.I., Schotman, P. and Gispen, W.H. (1981b) Effects of corticotropin (ACTH) on recovery of sensorimotor function in the rat: structure-activity study. Eur. J. Pharmacol. 76: 73-79. Bijlsma, W.A., Jennekens, F.G.I., Gispen, W.H. and De Wied, D. (1983a) The enhanced recovery of sensorimotor function in rats is related to the melanotropic moiety of ACTH/MSH neuropeptides. Eur. J. Pharmacol. 92: 231-236. Bijlsma, W.A., Jennekens, F.G.I., Schotman, P. and Gispen, W.H. (1983b) Stimulation by ACTH 4-10 of nerve fiber regeneration following sciatic nerve crush. Muscle Nerve 6: 104-112. Bisby, M.A. (1978) Fast axonal transport of labelled protein in sensory axons during peripheral regeneration. Exp. Neurol. 6\: 2S\-200. Bisby, M.A. (1988) Dependence of GAP43 (B50, Fl) transport on axonal regeneration in rat dorsal root ganglion neurons. Brain Res. 458: 157-166. Black, M.M. and Lasek, R.J. (1979) Slowing of the rate of

axonal regeneration during growth and maturation. Exp. Neurol. 63: 108-119. Blasquez, C, Jegou, S., Feuilloley, M. Rosier, A., Vandesande, F. and Vaudry, H. (1994). Visualization of gamma aminobutryric acid A receptors on proopiomelanocortinproducing neurons in the rat hypothalamus. Endocrinology 135: 2759-2764. Bregman, B.S. and Kunkel-Bagden, E. (1988) Methods of determining the recovery of motor function after spinal cord lesions and transplants in rats. In: M. Goldberger (Ed.), Criteria for Assessing Recovery of Function: Behavioral Methods, American Paralysis Association, Washington, DC, pp. 47-55. Brooke, M.H., Williamson, E. and Kaiser, K.K. (1971) The behavior of four fiber types in developing and reinnervated muscle. Arch. Neurol. 25: 360-366. Brown, M.C. and Lunn, E.R. (1988) Mechanism of interaction between motoneurons and muscles. In: Ciba Foundation Symposium 138 (Ed.), Plasticity of the Neuromuscular System, Wiley, New York, pp. 78-96. Buller, A.J., Eccles, J.C. and Eccles, R.M. (1960) Interactions between motoneurons and muscles in respect to the characteristic speeds of their responses. J. Physiol. 150: 399-416. Campenot, R.B. (1977) Local control of neurite outgrowth by nerve growth factor. Proc. Natl. Acad. Sci. USA 74: 45164519. Cannon, J.G., Tatro, J.B., Reichlin, S. and Dinarello, C.A. (1986) a-Melanocyte stimulating hormone inhibits immunostimulatory and inflammatory actions of interleukin 1. /. Immunol. 137: 2232-2236. Chhajlani, V. and Wikberg, J.E.S. (1992) Molecular cloning and expression of the human melanocyte stimulating hormone receptor cDNA. FEBS Lett. 309: 417-420. Civelli, O., Birnberg, N. and Herbert, E. (1982) Detection and quantitation of proopiomelanocortin mRNA in pituitary and brain tissues from different species. J. Biol. Chem. 257: 6783-6787. Close, R. (1964) Dynamic properties of fast and slow skeletal muscles of the rat during development. J. Physiol. 173: 7495. Connold, A.L., Fisher, T.J., Maudarbocus, S. and Vrbova, G. (1992) Response of developing fast muscles to partial denervation. Neuroscience 46: 981-988. Crescitelli, L., Strand, F.L. and Keim, K.L. (1989) Melanocortins and fast axonal transport in intact and regenerating sciatic nerve. Peptides 10: 883-886. Davila-Garcia, M.I., Hlibczuk, V., Akbari, H.M., Alves, S.E. and Azmitia, E.C. (1988) Postnatal effects of prenatal administration of neuropeptides on ^H-5-HT uptake by serotonergic neurons. Soc. Neurosci. Abstr. 14: 416. De Graan, P.N.E., Schotman, P. and Versteeg, D.H.G. (1990) Neural mechanism of action of neuropeptides: macromolecules and neurotransmitters. In: D. De Wied (Ed.), Neuropeptides: Basic and Perspectives, Elsevier, Amsterdam, pp. 175-211. De Koning, P., Brakkee, J.H. and Gispen, W.H. (1986) Meth-

F.L. Strand et al. ods for producing a reproducible crush in the sciatic nerve of the rat and the rapid and precise testing of return of sensory function: beneficial effects of melanocortins. J. Neurol. ScL 74: 237-246. de Medinaceli, L. and Rawling, R.R. (1987) Is it possible to predict the outcome of peripheral nerve injuries? A probability model based on prospects for regenerating neurites. BioSystems 20: 243-258. de Medinaceli, L., Freed, W.J. and Wyatt, R.J. (1982) An index of the functional condition of rat sciatic nerve based on measurements made from walking tracks. Exp. Neurol. 11: 634-643. De Wied, D. and Jolles, J. (1982) Neuropeptides derived from pro-opiocortin: behavioral, physiological, and neurochemical effects. Physiol. Rev. 62: 976-1059. Dekker, A.J.A.M. and Tonnaer, J.A.D.M. (1989) Binding of the neurotrophic peptide Org 2766 to rat spinal cord sections is affected by sciatic nerve crush. Brain Res. 411: 327-331. Dekker, A.J.A.M., Princen, M.M., De Nijs, H., De Leeds, L.G.J, and Broekkamp, C.L.E. (1987) Acceleration of recovery from sciatic nerve damage by ACTH (4-9) analog Org 2766: different routes of administration. Peptides 8: 1057-1059. Duckers, H.J., Verhaagen J., de Bruijn, E. and Gispen, W.H. (1994). Effective use of a neurotrophic ACTH 4-9 analog in the treatment of a peripheral demyelinating syndrome (experimental allergic neuritis). An intervention study. Brain 111: 365-374. Dyer, J.K., Philipsen, H.L.A., Tonnaer, J.A.D.M. and Haynes, L.W. (1992) An analysis of the binding specificity of the aMSH derivative Org 2766 in cultured rat Schwann cells. The Melanotropic Peptides, Rouen, France, PI-25: (Abstract). Dyer, J.K., Philipsen, H.L.A., Tonnaer, J.A.D.M., Hermkens, P.H.H. and Haynes, L.W. (1995). Melanocortin analogue Org 2766 binds to rat Schwann cells, upregulates NGF lowaffinity receptor p75, and releases neurotrophic activity. Peptides 16:515-522. Eberle, A.N. (1988) Structure-activity relationships of the melanotropins. In: The Melanotropins: Chemistry, Physiology and Mechanism of Action, Karger, Basel, pp. 333-379. Edwards, P.M., Kuiters, R.R.F., Boer, G.J. and Gispen, W.H. (1986) Recovery from peripheral nerve transection is accelerated by local application of MSH by means of microporous Accurel polypropylene tubes. J. Neurol. Sci. 74: 171-176. Eipper, B.A. and Mains, R.E. (1980) Structure and biosynthesis of pro-adrenocorticotropin/endorphin and related peptides. Endocrinol. Rev. 1: 1-27. Eskay, R.L., Giraud, P., Oliver, C. and Brownstein, M.J. (1979) Distribution of a-melanocyte stimulating hormone in rat brain: evidence that a-MSH-containing cells in the arcuate region send projections to extrahypothalamic areas. Brain Res. 178:55-67. Fahim, M.A., HoUey, J.A. and Robbins, N. (1983) Scanning

333

and light microscopic study of age changes at a neuromuscular junction in the mouse. J. Neurocytol. 12: 12-25. Rorijin, W.J., De Boer, T.H., Tonnaer, J.A.D.M., Van Nispen, J.W. and Versteeg, D.H.G. (1991) ACTH/MSH-like peptides inhibit the binding of dopaminergic ligands to the dopamine D2 receptors in vitro. Eur. J. Pharmacol. 207: 43. Rorijin, W.J., Gispen, W.H. and Versteeg, D.H.G. (1992) Functional characterization of a putative adenylate cyclasecoupled ACTH/MSH receptor in the brain. The Melanotropic Peptides, Rouen, France, PI-31: (Abstract) Frischer, R.E. and Strand, F.L. (1988) ACTH peptides stimulate motor nerve sprouting in development. Exp. Neurol. 100:531-541. Frischer, R.E., El Kawa, N.M. and Strand, F.L. (1985) ACTH peptides as organizers of neuronal patterns in development: maturation of the rat neuromuscular junction as seen by scanning electron microscopy. Peptides 6: 13-19. Frischer, R.E., King, J.A., Rose, K.J. and Strand, F.L. (1988) maturational changes in neonatal rat motor system with early postnatal administration of nicotine. Int. J. Dev. Neurosci. 6: 149-154. Fuxe, K., Hokfelt, T., Jonssen, G. and Lidbrink, P. (1973) Brain endocrine interaction: are some effects of ACTH and adrenocortical hormones on neuroendocrine regulation and behavior mediated via central catecholamine neurons?. In: K. Lissak (Ed.), Hormones and Brain Function, Plenum Press, New York, pp. 409^25. Ganz, I., Konda, Y., Tashiro, T., Shimoto, Y., Miwa, H., Munzert, G., Watson, S., Delvalle, J., and Yamada, T. (1993). Molecular cloning of a novel melanocortin receptor. J. Biol. Chem. 268: 8246-8250. Garcia D.E., Yebenes, E. and Pelletier, G. (1994). Negative regulation of proopiomelanocortin gene expression by GAB A A receptor activation in the rat arcuate nucleus. Peptides 15:615-618. Geiger, R., Gerhards, H.J., Wiemer, G. and Hock, F.J. (1987) Studies with peptides related to the sequence 4-9 of ACTH. In: D. Theodoropoulos (Ed.), Peptides, de Gruyer, Berlin, pp. 489. Gerritsen Van Der Hoop, R.P., De Koning, P., Boven, E., Neijt, J.P., Jennekens, F.G.I, and Gispen, W.H. (1988) Efficacy of the neuropeptide Org 2766 in the prevention and treatment of cisplatin-induced neurotoxicity in rats. Eur. J. Can. Clin. Oncol. 24: 637-624. Gispen, W.H., Zwiers, H., Wiegant, V.M., Schotman, P. and Wilson, J.E. (1979) The behaviorally active neuropeptide ACTH as neurohormone and neuromodulator: the role of cyclic nucleotides and membrane phosphoproteins. Adv. Exp. Med Biol. 116: 119-224. Gonzalez, E.R. and Strand, F.L. (1981) Neurotrophic action of ACTH/MSH 4-10 on neuromuscular function in hypophysectomized rats. Peptides 1: 107-113. Grafstein, B. and Forman, D.S. (1980) Intracellular transport in neurons. Physiol. Rev. 60: 1167-1283. Griffin, J.W., Drachman, D.B. and Price, D.L. (1976) Fast

334

Melanocortins as factors in somatic neuromuscular growth and re growth

axonal transport in motor nerve regeneration. J. NeurobioL 7: 355-370. Guy, J., Leclerc, R., Vaudry, H. and Pelletier, G. (1980) Identification of a second category of a-melanocyte stimulating hormone (a-MSH) neurons in the rat hypothalamus. Brain Res. 199: 135-146. Hall, Z.W. and Sanes, J.R. (1993) Synaptic structure and development: the neuromuscular junction. Cell 72: 99-121. Hannigan, J. and Isaacson, R. (1985) The effects of Org 2766 on the performance of sham, neocortical and hippocampallesioned rats in a food search task. Pharmacol. Biochem. Behav. 23: 1019-1027. Hare, G.M., Evans, P.J., MacKinnon, S.E., Best, T.J., Midha, R., Szalai, J.P. and Hunter, D.A. (1993). Walking track analysis: utilization of individual foot parameters. Ann. Plastic Surg. 30: 147-153. Haynes, L. and Smith, M.E. (1985) Presence of immunoreactive a-melanotropin and /^-endorphin in spinal motoneurones of the dystrophic mouse. Neurosci. Lett. 58: 13-18. Hoi, E.M., Mandys, V., Sodaar, P., Gispen, W.H. and Bar, P.R. (1994). Protection by an ACTH 4-9 analogue against the toxic effects of cisplatin and taxol on sensory neurons and glial cells in vitro. J. Neurosci. Res. 39: 178-185. Hollyday, M. (1980) Motoneuron histogenesis and the development of limb innervation. Curr. Top. Dev. NeurobioL 15: 181-215. Hugelin, A., Legrain, Y. and Bondoux-Jahan, M. (1979) Nerve growth promoting action of isaxonine in rat. Experientia 35: 626-627. Hughes, S. and Smith, M.E. (1988) Effect of nerve section on y^-endorphin and a-melanotropin immunoreactivity in motor nerves of normal and dystrophic mice. Neurosci. Lett. 92: 1-7. Hughes, S. and Smith, M.E. (1989a) Proopiomelanocortinderived peptides in transected and contralateral motor nerves of the rat. J. Chem. Neuroanat. 2: 227-237. Hughes, S. and Smith, M.E, (1989b) Proopiomelanocortinderived peptides in mice with motoneurone disease. Neurosci. Lett. 103: 169-173. Hughes, S. and Smith, M.E. (1991) Corticotropin receptors in murine skeletal muscles in vitro. J. Physiol. 435: 85. Isaacson, R. and Poplawsky, A. (1983) An ACTH 4-9 analog (Org 2766) speeds recovery from septal hyperemotionality in the rat. Behav. Neural Biol. 39: 52-59. Joseph, S.A. (1990) Immunoreactive adrenocorticotropin in rat brain: a neuroanatomical study using antiserum generated against synthetic ACTH 1-39. Am. J. Anat. 158: 533548. Keim, K. (1978) Effects of synthetic ACTH 1-24 and selected peptide fragments on rat heart. New York University PhD. thesis . Keim, K.L., Sigg, E.B. and Strand, F.L. (1977) ACTH 1-24 alters rat atrial strip contractility. Fed. Proc. 36: 364. Keim, K.L., Sigg, E.B. and Strand, F.L. (1978) ACTH 1-24 influences contractility in rat atrial strip. Fed. Proc. 37: 523.

King, J.A., Davila-Garcia, M., Azmitia, E.C. and Strand, F.L. (1991) Differential effects of prenatal and postnatal ACTH or nicotine exposure on 5-HT high affinity uptake in the neonatal rat brain. Int. J. Dev. Neurosci. 9: 281-286. Kugelberg, E. (1981) The motor unit: morphology and function. In: J.E. Desmedt (Ed.), Motor Unit Types, Recruitment and Plasticity in Health and Disease, Karger, Brussels, pp. 1-16. Kume, J., Baradarian, R. and Strand, F.L. (1992) Sex steroid and castration effects on the regenerating EDL muscle after peroneal nerve crush. Soc. Neurosci. Abstr. 18: 560.2. Kume-kick, J. and Strand, F.L. (1004). Sex hormones affect contractile and motor functional properties. Exp. Neurol. 128: 115-123. Landmesser, L.T. and Morris, D.G. (1975) The development of functional innervation on the hind limb of the chick embryo. J. Physiol. 249: 301-326. Lee, S.J., Aoki, C.J. and Strand, F.L. (1992) Immunocytochemical localization of ACTH 4-10 in rat brain. Soc. Neurosci. Abstr. 18: 121.2. Lee, S.J., Lee, T.S. and Strand, F.L. (1993) The effects of ACTH 4-10, BIM 22015 and Org 2766 on neurite outgrowth in three chamber system. Winter Neuropeptide Conf., Breckenridge, CO, (Abstract). Lee, S.J., Lee, T.S. and Strand, F.L. (1993) The effects of ACTH 4-10, BIM 22015 and Org 2766 on neurite outgrowth in three chamber system. Winter Neuropeptide Conf.y Breckenridge, CO, (Abstract). Li, L., Oppenheim R., Lei, M., and Houenou, L.J. (1994). Neurotrophic agents prevent motoneuron death following sciatic nerve section in the neonatal mouse. J. Neurol. 25: 759-766. Leiba, H., Garty, N.B., Schmidt-Sole, J., Piterman, O., Azrad, A. and Salomon, Y. (1990) The melanocortin receptor in the rat lacrimal gland: a model system for the study of MSH (melanocyte stimulating hormone) as a potential neurotransmitter. Eur. J. Pharmacol. 181: 71-82. Lieberman, A.R. (1971) The axon reaction: a review of the principal features of perikaryal responses to axon injury. Int. Rev. NeurobioL 14: 49-124. Luneburg, U. and Rohr, H. (1988) Effects of melanocortins on vestibular compensation. Prog. Brain Res. 76: 421429. Mains, R.E. and Eipper, B.A. (1980) Biosynthetic studies on ACTH, )3-endorphin and a-melanotropin in the rat. Ann. N.Y.Acad. ScL 343:94-110. Maragos, W.F., Greenamyre, T.J., Penney, J.B. and Young, A.B. (1987) Glutamate dysfunction in Alzheimer's disease: a hypothesis. Trends Neurosci 10: 65-68. McMahan, U.J., Edgington, D.R. and Krffler, D.P. (1980) Factors that influence regeneration of the neuromuscular junction. J. Exp. BioL 78: 176-198. Mountjoy, K.G., Robbins, L.S., Mortrud, M.T. and Cone, R.D. (1992) The cloning of a family of genes that encode the melanocortin receptors. Science 257:1248-1251. Nakanishi, S., Inoue, A. and Kita, T. (1979) Nucleotide se-

F.L. Strand et al. quence of cloned cDNA for bovine corticotropin-)3lipotropin precursor. Nature 278: 423-427. O'Donohue, T.L., Miller, R.L. and Jacobowitz, D.M. (1979) Identification, characterization and stereotaxic mapping of intraneuronal alpha-melanocyte stimulating hormone-like immunoreactive peptides in discrete areas of the rat brain. Brain Res. 176: 101-123. Pestronk, A., Drachman, D.B. and Griffin, J.W. (1980) Effects of age on nerve sprouting and regeneration. Exp. Neurol. 8: 676-689. Pette, D. and Staron, R.S. (1988) Molecular basis of the phenotypic characteristics of mammalian muscle fibres. In: Ciba Foundation Symposium 138 (Ed.), Plasticity of the Neuromuscular System, Wiley, New York, pp. 22-34. Plantinga, L.C., Verhaagen J., Edwards, P.M., Hall, M., Brakkee, J.H. and Gispen, W.H. (1995) Pharmacological evidence for the involvement of endogenous a-MSH-like peptides in peripheral nerve regeneration. Peptides 16: 319324. Redfern, P.A. (1970) Neuromuscular transmission in newborn rats. J. Physiol. 209: 701-709. Rees, H.D., Verhoef, J., Witter, A., Gispen, W.H. and DeWied, D. (1980) Autoradiographic studies with a behaviorally potent 3H-ACTH 4-9 analog in the brain after intraventricular injection in rats. Brain Res. Bull. 5: 509-514. Roberts, J.L. and Herbert, E. (1977a) Characterization of a common precursor to corticotrophin and beta-lipotrophin: cell-free synthesis of the precursor and identification of corticotrophin peptides in the molecule. Proc. Natl. Acad. Sci. 14: 4826^830. Roberts, J.L. and Herbert, E. (1977b) Characterization of a common precursor to corticotrophin and beta-lipotrophin peptide and their arrangement relative to corticotrophin in the precursor synthesized in a cell-free system. Proc. Natl. Acad. Sci. USA 74: 5300-5304. Rose, K.J. and Strand, F.L. (1988) Mammalian neuromuscular development accelerated with early but slowed with late gestational administration of ACTH peptide. Synapse 2: 200-204. Rose, K.J. and Strand, F.L. (1990) Accelerated neuromuscular development in the rat with prenatal exposure to nicotine. Int. J. Dev. Neurosci. 8: 565-573. Rose, K.J., Frischer, R.E., King, J.A. and Strand, F.L. (1988) Neonatal muscular parameters vary in susceptibility to postnatal ACTH/MSH 4-10 administration. Peptides 9: 147-171. Roselli-rehfuss, L., Mountjoy, K, Robbins, L.S., Mortrud, M.T., Low, M.J., Tatro, J.B., Entwhistle, M.L., Simerly, R.B. and Cone, R.D. (1993). Identification of a receptor for y-melanotropin and other proopiomelanocortin peptides in the hypothalamus and limbic system. Proc. Natl. Acad. Sci. USA 90: 8856-8860. Saint-Come, C, Acker, G.R. and Strand, F.L. (1982) Peptide influences on the development and regeneration of motor performance. Peptides 3: 439-449. Saint-Come, C. and Strand, F.L. (1985) ACTH/MSH 4-10

335 improves motor unit reorganization during peripheral nerve regeneration in the rat. Peptides 6: 77-83. Saint-Come, C. and Strand, F.L. (1988) ACTH 4-9 analogue (Org 2766) improves qualitative and quantitative aspects of motor nerve regeneration. Peptides 8: 215-221. Salomon, Y. (1990) Melanocortin receptors: targets for control by extracellular calcium. Mol. Cell. Endocrinol. 70: 39145. Segarra, A.C., Luine, V.N. and Strand, F.L. (1991) Sexual behavior of male rats is differentially affected by timing of perinatal ACTH administration. Physiol. Behav. 50: 689-697. Skene, J.H.P. and Willard, M. (1981) Changes in axonally transported proteins during axon regeneration in toad retinal ganglion cells. J. Cell Biol. 89: 86-95. Smith, CM. and Strand, F.L. (1981) Neuromuscular responses of the immature rat to ACTH/MSH 4-10. Peptides 2: 197-206. Spruijt, B.M. (1992a) The ACTH 4-9 analog Org 2766 modulates NMD A receptor activity. Soc. Neurosci. Abstr. 18: 897. Spruijt, B.M. (1992b) Effects of the ACTH 4-9 analog. Org 2766 on brain plasticity: modulation of excitatory neurotransmission. Psychoneuroendology 17: 315-325. Steinbach, J.H. (1981) Developmental changes in acetylcholine receptors aggregates at rat skeletal neuromuscular junctions. Dev. Biol. 84: 267-276. Strand, F.L. and Kung, T.T. (1980) ACTH accelerates recovery of neuromuscular function following crushing of peripheral nerve. Peptides 1: 135-138. Strand, F.L. and Smith, CM. (1986) LPH, ACTH, MSH, and motor systems (update of 1981). In: D. DeWied, W.H. Gispen and T.J.B. Van Wimersma Greidanus (Eds.), Neuropeptides and Behavior; CNS Effects of ACTH, MSH, and Opioid Peptides, Pergamon, Oxford, pp. 245-272. Strand, F.L., Friedebold, G. and Stoboy, H. (1962) Electrical characteristics of rat skeletal muscle following adrenalectomy. Acta Physiol. Pharmacol. Neerlandica 11: 213234. Strand, F.L., Rose, K.J., King, J.A., Segarra, A.C and Zuccarelli, L.A. (1989) ACTH modulation of nerve development and regeneration. Prog. Neurobiol. 33: 45-85. Strand, F.L., Segarra, A.C, Zuccarelli, L.A., Kume, J. and Rose, K.J. (1990a) Neuropeptides as neuronal growth regulating factors. Ann. N. Y. Acad. Sci. 579: 68-90. Strand, F.L., Segarra, A.C, Zuccarelli, L.A., Kume, J. and Rose, K.J. (1990b) Neuropeptides: fail-safe, redundant or special?. In: D. Gupta, H.a> Wollman and M.B. Ranke (Eds.), Neuroendocrinology: New Frontiers, Brain Research Promotion, London, Tubingen, pp. 19-28. Strand, F.L., Rose, K.J., Zuccarelli, L.A., Kume, J., Alves, S.E., Antonawich, F.J. et al. (1991) Neuropeptide hormones as neurotrophic factors. Physiol. Rev. 71: 1017-1046. Strand, F.L., Zuccarelli, L.A., Williams, K.A., Lee, T.S., Lee, S.J., Antonawich, F.J. et al. (1992) Melanotropins as growth factors. The Melanotropic Peptides, Rouen, France, September 6-9: 2 (Abstract).

336

Melanocortins as factors in somatic neuromuscular growth and re growth

Strand, F.L., Lee, S.J., Lee, T.S., Antonawich, F.J., Zuccarelli, L.A., Kume, J. and Williams, K.W. (1993a) Noncorticotropic ACTH peptides modulate nerve development and regeneration. Rev. Neurosci. 4: 321-364. Strand, F.L., Saint-Come, C, Lee, T.S., Lee, S.J., Kume, J.A. and Zuccarelli, L.A. (1993b) An ACTH/MSH 4-10 analog BIM 22015 has neurotrophic and myotrophic attributes during peripheral nerve regeneration. Peptides 14: 287296. Strand, F.L., Zuccarelli, L.A., Williams, K.A., Lee, T.S., Lee, S.J. and Antonawich, F.J. (1993c) Melanotropins as growth factors. In: A.N. Eberle and H. Vaudry (Eds.), The Melanotropic Peptides, Ann. N. Y. Acad. Sci. 680: 29-50. Sutherland, S. (1978) Nerves and Nerve Injuries, 2nd edn., Churchill-Livingston, London. Tatro, J.B. (1990) Melanotropin receptors in the brain are differentially distributed and recognize both corticotropin and a-melanocyte stimulating hormone. Brain Res. 536: 124-132. Tatro, J.B. and Reichlin, S. (1987) Specific receptors for alpha melanocyte-stimulating hormone are widely distributed in tissues of rodents. Endocrinology 121: 1900-1907. Torda, C. and Wolff, H.G. (1952) Effect of pituitary hormones, cortisone and adrenalectomy on some aspects of neuromuscular function and acetylcholine synthesis. Am. J. Physiol. 169: 140-149. Van der Neut, R., Bar, P.R., Sodaar, P. and Gispen, W.H. (1988) Trophic influences of alpha-MSH on ACTH 4-10 on neuronal outgrowth in vitro. Peptides 9: 1015-1020. Van der Neut, R., Oestreicher, A.B., Gispen, W.H. and Bar, P.R. (1990) The expression of B-50/GAP-43 during development of rat spinal neurones in culture is regulated by interneuronal contact. Neurosci. Lett. 109: 36-41. Vandertop, W.P., de Vries, W.B., Notermans, N.C., Tulleken, C.A.F. and Gispen, W.H. (1994). Experimentally induced autonomic neuropathy: beneficial effects of a systemic ACTH 4-9 analog on oculomotor nerve regeneration. Rest. Neurol. Neurosci. 7: 37-43. Van der Zee, C.E.E.M., Brakkee, J.H. and Gispen, W.H. (1988) a-MSH and Org 2766 in peripheral nerve regeneration: different routes of delivery. Eur. J. Pharmacol. 147: 351-357. Van der Zee, C.E.E.M., Nielander, H.B., Vos, J.P., Lopes de Silva, S., Verhaagen, J., Oestreichler, A.B. et al. (1989) Expression of growth-associated protein B-50 (GAP-43) in dorsal root ganglia and sciatic nerve during regenerative sprouting. J. Neurosci. 9: 3505-3512. Van der Zee, C.E.E.M., Brakkee, J.H. and Gispen, W.H. (1991) Putative neurotrophic factors and functional recovery from peripheral nerve damage in the rat. Br. J. Pharmacol. 103: 1041-1046. Van Essen, D.C. (1982) Neuromuscular synapse elimination: structural, functional and mechanistic aspects. In: N.C. Spitzer (Ed.), Neuronal Development, Plenum, New York, pp. 333-371. Van Loon, G.R. (1973) Brain catecholamines and ACTH

secretion. In: L. Martini and W.F. Ganong (Eds.), Neuroendocrinology, Oxford University Press, New York, pp. 209. Veals, J. (1979) Effects of adrenocorticotropic hormone on 14C-choline accumulation and ^"^C-radioactivity release by brain synaptosomes. New York University, New York, PhD. thesis. Veals, J. and Strand, F.L. (1979) Efects of adrenocorticotropic hormone on 14C-choline accumulation by brain synaptosomes. Physiologist (Abstract) 22: 75. Verhaagen, J., Edwards, P.M., Jennekens, F.G.I., Schotman, P. and Gispen, W.H. (1986) Alpha-melanocyte stimulating hormone stimulates the outgrowth of myelinated nerve fibers after peripheral nerve crush. Exp. Neurol. 92: 451-454. Verhaagen, J., Oestreicher, A.B., Edwards, P.M., Veldman, H., Jennekens, F.G.I, and Gispen, W.H. (1988) Light and electron microscopical study of phosphoprotein B-50 following denervation and reinnervation of the rat soleus muscle. J. Neurosci. 8: 1759. Verhoef, J., Palkovits, M. and Witter, A. (1977) Distribution of a behaviorally highly potent ACTH 4-9 analog in rat soleus muscle. Brain Res. 126: 89-104. Versteeg, D.H.G. and Wurtman, R. (1975) Effect of ACTH 4 10 on the rate of synthesis of [-^HJ-catecholamines in the brains of intact, hypophysectomized and adrenalectomized rats. Brain Res. 93: 552-557. Versteeg, D.H.G., DeCrom, M.P.G. and Mulder, A.H. (1986) ACTH (1-14) and a-MSH antagonize dopamine receptormediated inhibition of striatal dopamine and acetylcholine release. Life Sci. 38: 835-840. Watson, S.J. and Akil, H. (1980) Anatomical and functional studies of ACTH and lipotropin in the central nervous system. In: D. DeWied and P.A. Van Keep (Eds.), Hormones and the Brain, Lancaster, UK, pp. 73-86. Wenk, G.L., Walker, L.C., Price, B.L. and Cork, L.C. (1991) Loss of NMDA, but not GABA-A, binding in the brains of aged rats and monkeys. Neurobiol. Aging 12: 93-98. Wiegant, V.M., Dunn, P., Schotman, P. and Gispen, W.H. (1979) ACTH-like neurotropic peptides: possible regulators of rat brain cyclic AMP. Brain Res. 168: 565. Wiegant, V.M., Zwiers, H. and Gispen, W.H. (1981) Neuropeptides and brain cAMP and phosphoproteins. Pharmacol. Ther. 12: 463^90. Wolterink, G., Van Zanten, E., Kamsteeg, K., Radhakishun, F. and Van Ree, J. (1990) Functional recovery after destruction of dopamine systems in the nucleus accumbens of rats. II. Facilitation by the ACTH (4-9) analog Org 2766. Brain Res. 507: 101-108. Wolterink, G., van Ree, J.M., van Nispen, J.W. and de Wied, D. (1991) Structural modifications of the ACTH (4-9) analog Org 2766 yields peptides with high biological activity. L//e 5d. ^5: 155-161. Zuccarelli, L.A. and Strand, F.L. (1990) ACTH peptides increase endplate parameters during the regeneration of the developing peripheral nervous system. Soc. Neurosci. Abstr. 1:479.6. Zuccarelli, L.A. and Strand, F.L. (1991) ACTH/MSH peptides

F,L. Strand etal accelerate muscle maturation and functional recovery following trauma of the developing nervous system. Soc. Neurosci. Abstr. 1:95.10. Zuccarelli, L.A. and Strand, F.L. (1992) ACTH/MSH peptides promote neuromuscular efficiency following trauma of the

337 developing nervous system. Soc. Neurosci. Abstr. 1: 546.18. Zwiers, H., Veldhuis, D., Schotman, P. and Gispen, W.H. (1976) ACTH, cyclic nucleotides and brain protein phosphorylation in vitro. Neurochem. Res. 1: 669-677.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved. CHAPTER 14

Functions of fibroblast growth factors (FGFs) in the nervous system Sophie Bieger and Klaus Unsicker Institut fur Anatomic und Zellbiologie, Im Neuenheimer Feld 307, 69120 Heidelberg, Germany

1. Introduction Despite their name, the fibroblast growth factors (FGFs) are far more than just mitogens for fibroblasts. In the past five years, the list of functions attributed to the 'original' FGFs, FGF-1 (or acidic FGF) and -2 (or basic FGF) has grown significantly, as has the size of the FGF family. FGFs have been demonstrated to influence the growth and function of cells of the vascular, muscular, epithelial and nervous systems; and they are now thought to be involved in processes ranging from morphogenesis, tissue maintenance and repair to oncogenesis. This wide variety of effects has raised interest in the FGFs as potential therapeutic agents, for example, to improve wound healing, and has also spawned attempts to engineer FGFs in order to enhance their biological activity. Currently, nine FGFs are known to occur in vertebrates. Since the latest additions to the FGF family, FGF-8 and FGF-9, were discovered relatively recently (Tanaka et al., 1992; Miyamoto et al., 1993), there is every reason to expect that the growth of the FGF family is not yet at an end. Why should there be so many different FGFs? The tissue-specific expression of most of these factors during development (as well as in the adult) suggests one advantage of factor diversity: specialization of cellular targets and increasing complexity of effects. FGFs -3, -5, and -7, for example, are expressed in only few regions of the developing mouse embryo, and only for short times, consistent with their postulated roles in the induction of specific tissues (Wilkinson et al, 1988; Finch et al..

1989; Haub and Goldfarb, 1991). The FGF-4 gene is apparently transcribed only during development and not at all in normal adult cells and tissues (Yoshida et al., 1991), suggesting an even more specialized function for this protein. Perhaps the only exception to the rule of restricted expression is FGF-2, which is relatively ubiquitous in the adult organism, although its mRNA, like that of the other members of the family, is quite rare (Abraham et al., 1986). The observation that FGFs can have pronounced effects on cells of the nervous system both glial and neuronal - has prompted not only intensive research on FGFs in the nervous system but also the publication of a number of relevant reviews (Sensenbrenner, 1993; Unsicker et al., 1993; Baird, 1994a; Eckenstein, 1994). These, and a number of more general reviews on FGFs (Baird and Bohlen, 1990; Coulier et al., 1994), can be recommended to readers seeking further information. In the following, we will first give a brief overview of FGFs and their receptors before focusing on some of the important effects of FGFs on neural cells at both cell and tissue levels. 2. The FGF family 2.1. FGF genetics 2.1.1. Conservation and diversification At present, the FGF family contains 9 members with related sequences. The family relationship is evident even at the DNA level: all FGF genes (with the exception of the more complex FGF-8)

340

consist of three exons separated by two relatively large introns. This common gene architecture strongly suggests derivation from a common ancestral gene, and a possible phylogenetic tree has been constructed on the basis of amino acid sequence comparisons (Coulier et al., 1994). Despite this apparent diversification of FGFs, there is also a tendency for individual genes to be conserved in evolution. Comparison of analogous FGFs derived from rodents (rat or mouse) and humans generally reveals better than 90% similarity at the amino acid level, representing only one or two amino acid differences between the FGFs of these species. The finding of homologous genes in lower vertebrates (e.g. FGF-3 in Xenopus; Tannahill et al., 1992) provides further evidence for conservation of FGF genes during phylogeny. 2.1.2. Expression of FGF genes The expression of FGFs at both the RNA and protein level is generally very low in normal adult tissues; peaks of FGF expression often occur in brief phases of embryogenesis. In situ hybridization studies have provided evidence for tissuespecific patterns of FGF mRNA expression in early stages of embryonic development (see Section 4.5). Such spatial and temporal regulation of gene expression correlates well with some of the documented effects of these factors, such as induction of mesoderm (by FGF-2, -3, -4, and other FGFs; reviewed by Slack, 1994), or induction of specific primordia such as the limbs (by FGF-2 and -4; reviewed by Niswander et al., 1994; Olwin et al., 1994), the eye (by FGF-1 and -2; reviewed by McAvoy et al., 1991), and the inner ear (by FGF-3; Represa et al., 1991; Mansour, 1994). 2.1.3. Regulation of FGF expression Regulation of the expression of FGF genes is apparently necessary, as FGFs have transforming potential: overexpression of these growth factors is frequently associated with a transformed phenotype. This explains why the newer members of the FGF family were first identified in or purified from tumor cells. FGF-3, for example, is a protooncogene (the integration site for mouse mammary tumor virus); FGF-4 was cloned from human gastric

Functions offibroblast growth factors (FGFs) in the nervous system

cancer tissue, FGF-5 was identified as a product of a human oncogene; FGF-8 and -9 were identified in mammary carcinoma and glioblastoma cell lines, respectively. A number of regulatory mechanisms have been described for specific FGF genes. Some of the FGFs are apparently kept 'under control' by sequences adjacent to or inherent to the FGF genes themselves. The FGF-5 gene, for example, is overlapped at the 5' end by an open reading frame that reduces the frequency with which ribosomes initiate translation at the FGF gene proper (Goldfarb et al, 1991), while the (murine) FGF-4 gene contains enhancing elements (octamer binding motifs) that are active primarily during early development, when octamer binding proteins are expressed at high levels (Rizzino and Rosfjord, 1994). Other FGF transcripts possess 3' sequences thought to destabilize mRNA (e.g. FGF-9). Interestingly, FGF-2 and -6 antisense transripts have been documented; these too may be involved in regulating the availability of the FGF mRNA for transcription (Kimelman and Kirschner, 1989; Coulier et al., 1991; Borja et al., 1993). These examples represent only a few of the mechanisms that presumably contribute to keeping the FGF mRNAs at low levels. The responsiveness of various FGF genes to stimuli such as serum (FGF-2, 6; Maries et al., 1989; Powell and Klagsbrun, 1993), phorbol esters (FGF-2; Stachowiak et al., 1994), or steroids (FGF-8; Tanaka et al., 1992) indicates that the promoters of these genes contain classical regulatory elements such as serumresponsive elements, Spl, AP-1, and AP-2 binding sites, etc.; some of these have already been characterized. 2.2. FGF proteins 2.2.1. Common structural features In general, FGFs are proteins of moderate size, ranging from 155 to 268 amino acids (16-32 kDa, depending on glycosylation). The characteristic feature of the FGF family is a 'core' region of 120-130 amino acids in which the amino acids are relatively well-conserved (at least 30%) (Fig. 1). An invariant feature in this region are two cysteine

S. Bieger and K. Unsicker

150 200 amino acids

341

250

300

350

Fig. 1. Homologies in FGF amino acid sequences. FGFs have been aligned to indicate the most highly conserved 'core' region (shaded); gaps were inserted (dashed lines) to maximize the homology. Alternative initiation sites for FGF-2, -3 and -8 are represented by vertical bars.

residues spaced 67 amino acids apart; however, chemical and structural analyses of FGF-1 and -2 indicate that these cysteines do not form disulfide bonds (Thomas et al, 1991; Thompson and Fiddes, 1991) and may not be critical in formation of the tertiary structure (Ago et al., 1991; Eriksson et al., 1991, 1993). Also contained in the core region are sequences involved in heparin binding and in receptor binding. The sequences responsible for these important interactions have been investigated in detail for FGF-2 (Baird et al, 1988; Seno et al., 1990; Seddon et al, 1991; Presta et al., 1992; Eriksson et al., 1993). At the amino- and carboxyterminal ends of the ends of the FGF *core', the primary structures are widely divergent. These are the regions containing factor-specific characteristics such as hydrophobic signal sequences, glycosylation sites, alternative initiation sites, and presumed nuclear localization signals (see next Section).

1995). Alternative splicing has also been described for FGF-1 (Voulgaropoulu et al., 1994). Alternative initiation sites (upstream CUG codons in the mRNA transript) give rise to amino-terminally extended forms FGF-2 and -3 (Florkiewicz and Sommer, 1989; Prats et al., 1989; Acland et al., 1990). Interestingly, the amino-terminal extensions contain putative nuclear localization sequences that appear to be involved in transport of these proteins into the nucleus (Acland et al., 1990; Amalric et al., 1991; Bugler et al., 1991; Kiefer et al., 1994). FGF-1 has also been shown to enter the nucleus, but the amino acid sequence involved in this process remains to be elucidated (Cao et al., 1993). The specific function of 'nuclear' FGF is still unclear, but there is evidence that FGF-2 import into the nucleus is involved in the mitogenic response of endothelial cells to FGF (Bouche et al., 1987; Baldin et al., 1990). Such a nuclear mechanism of FGF action could supplement the 'classical' signal pathway by receptor-mediated mechanisms. A number of other important features of the various FGF proteins are summarized in Table 1. It should be noted that all FGFs may be secreted, but FGF-1, -2 and -9 are unusual in that they lack a 'classical' signal sequence. The mechanism by which these proteins exit the cell remain speculative at present; a recent study by Florkiewicz et al. (1995) suggests that a novel non-ER/Golgi pathway may be involved. All members of the FGF family with the exception of FGF-1 and -2 possess consensus sequences for addition of sugar residues and apparently may be glycosylated. However, glycosylation does not appear to be necessary for function in, for example, FGF-4 (Fuller-Pace et al., 1991) or FGF-5 (Goldfarb et al., 1991; Clements et al., 1993). 3. FGF receptors

2.2.2. Special features of FGF proteins Some of the FGFs can occur in multiple forms that result from alternative transcription or RNA processing. In this respect, FGF-8 is one of the most complex members of the family: at least 7 different-sized protein products can be generated by alternative splicing (Crossley and Martin,

3.1. High-affinity receptors It is hardly surprising that the expanding family of FGFs is matched by a growing family of FGF receptors (FGFRs). Since a number of recent articles have covered these molecules in detail

Functions offibroblastgrowth factors (FGFs) in the nervous system

342 TABLE 1 Selected features of FGFs Factor

Other names

Human mRNA transcript(kb)

Protein

Glycosylation

Signal peptide

Occurrence in adult brain

FGF-1

aFGF

4.6

No

bFGF

FGF-3

int-2

FGF-4

hst-1, kFGF

7.0, 3.7, others? 1.4(antisense) 2.9, 2.7, 1.8, 1.6 3.0, 1.7

No; but secreted No; but secreted Yes; but short; inefficient cleavage Yes; cleaved Yes; cleaved Yes; cleaved (?) Yes; cleaved Yes (?) (secreted) No; but secreterl

Yes

FGF-2

155aa 16kDa 155, 196, 210aa^ 18, 21.5, 22.5 kDa 245, 274aaa -28, -32 kDa 206aa 22 kDa 268aa -32 kDa 208 -20kD? 194aa 28 kDa 215, 268aa, others*' 28, 32kD 208aa 30

FGF-5

1.7

FGF-6 FGF-7

KGF

4.7, 3.9 0.85 (antisense) 5,2.4

FGF-8

AIGF

-1.0-1.2 4.3, 3.4, 2.7

FGF-9

No Yes Yes Yes Yes Yes Yes Yes

Yes No No Yes No No 7 Yes

Observed molecular weights vary due to glycosylation and cleavage of signal peptides. ^Alternative (CUG) inititation sites produce proteins of different sizes and with different fates (secretion vs. nuclear transport); see text. ^^Alternative splicing apparently yields proteins of different sizes.

(Coulier et al., 1992; Jaye et al., 1992; Partanen et al., 1992; Fantl et al., 1993; McKeehan and Kan, 1994), only a summary will be presented here. To date, four FGFRs have been identified. These receptors, which have affinities for various FGFs in the range of 10-200 pM, share a common structure illustrated in Fig. 2. The entire receptor

molecule, around 820 amino acids long, spans the membrane once. The extracellular region consists of a signal peptide and one to three immunoglobulin-like domains; characteristic features are an *acid box' of 4-8 acidic amino acids, located between the first and second immunoglobulin domain, as well as a number of glycosylation sites.

membrane 4aa "acid box"

insertion

-S - - S- VLJ-S - - S-L~/-S — S lg-1

ig-2

ig-3

Fig. 2. Typical structure of FGF receptors.

Tyrosine kinase

S. Bieger and K. Unsicker

Intracellularly, the receptor contains a tyrosine kinase domain split by a 14-amino acid insertion. This overall FGFR structure is reminiscent of that of other tyrosine-linked growth factor receptors, such as the receptors for insulin and insulin-like growth factors, the epithelial growth factor receptor, and the platelet-derived growth factor receptor, but differs in several important structural aspects (e.g. number of immunoglobulin loops, size of tyrosine kinase insert), so that the FGFRs are grouped as a separate family. Similarities exist, however, in the signal transduction mechanism: binding of FGF to its receptor induces receptor dimerization, which can occur between different receptor types. Dimerization is followed by transphosphorylation of the two receptor molecules; this in turn causes phosphorylation of various cellular proteins, notably phospholipase C-y (PLC-y). The ensuing intracellular changes - production of 1,2-diacylglycerol (Altin and Bradshaw, 1990), stimulation of protein kinase C, induction of immediate early genes (Ferhat et al, 1993; Simpson and Morris, 1994) - have been relatively well characterized with other tyrosine kinase-linked growth factor receptors and appear to be conserved in neuronal cells. 3.1.L Molecular biological aspects The FGFR types are most similar in the tyrosine kinase domain (75-92% homology), a region that is also homologous to the tyrosine kinase domain of other growth factor receptors. The immunoglobulin loops 2 and 3 are moderately conserved (-60-80%), while the immunoglobulin loop 1 is most divergent between receptor types. Comparison of the FGFRs isolated from different species suggests that there is selective pressure to conserve the structure of the FGFRs, like that of the FGFs themselves. 3.1.2, Variant FGFR forms Alternative splicing can give rise to a large number of variants of the 'standard' FGFR described above. Functionally, perhaps the most important variation is in the second half of the third immunoglobulin domain. Use of alternative exons generates receptors with affinities for different

343

FGFs, or truncated forms that contain neither the transmembrane nor the tyrosine kinase domain and thus cannot function in FGF signal transduction. Such soluble receptor forms, to date described for FGFR-1 and -2, are postulated to exert a regulatory function in FGF signalling. Potentially they could act as carriers, transporting FGFs to high affinity membrane-bound receptors; alternatively, they could act in a downregulatory fashion by competing with the signal-transducing receptors (Dionne et al., 1991). Other frequent splice variants (also described for FGFR-1 and -2) lack the first immunoglobulin domain or have variations in the intracellular tyrosine kinase domain. 3.1.3. Specificities Apparently, FGFRs are not highly selective in terms of ligand. Exhaustive data are not yet available regarding the binding of individual FGFs to each receptor type (FGFR-1 to -4), but available data suggest a high degree of redundancy. One region implicated in conferring ligand specificity is the second half of the third immunoglobulin domain, a region which undergoes alternative splicing (Vainikka et al., 1992; Cheon et al., 1994). FGFR-1, for example, binds FGF-1 and -2 with approximately equal affinity but, in an alternative splice form (see preceding Section), it preferentially binds FGF-1 and -4 (Werner et al., 1992). FGFR-3 similarly undergoes alternative splicing to yield a form specific for FGF-1 (Chellaiah et al., 1994). 3.2. 'Low-affinity receptors' Although the high-affinity receptors presumably play the critical signal-transducing role for FGFs, increasing attention is also being paid to the contribution made by lower-affinity 'receptors'. These are proteins that associate with FGFs through the heparin-binding domain: sulfated glycoproteins (heparan sulfate proteoglycans, HSPGs) such as syndecan, glypican, perlecan, and potentially other molecules of the extracellular matrix. These proteins have been proposed to function as 'regulators of FGF bioavailability' (Baird, 1994b), acting in a number of different ways:

344

Functions of fibroblast growth factors (FGFs) in the nervous system

Protection of FGFs from degradation by proteolysis: Binding of FGF-1 and -2 to HSPGs has been shown to make the growth factors more resistant to proteases, thus increasing their biological half-life (Damon et al., 1989; Sommer and Rifkin, 1989). Storage of FGFs: Low affinity receptors may fulfill a 'reservoir' function by attracting FGFs and immobilizing them in the vicinity of cells. FGFs associated in this way with basement membranes or extracellular matrix could be released in response to local proteolytic activity. This mechanism could be important in situations involving acute protease activity, such as the response to tissue trauma (see Ruoslahti and Yamaguchi, 1991; Vlodavskyetal., 1991). Interactions with high affinity receptors'. A number of studies have shown that low affinity receptors (A'd = 2-10nM) of the heparan sulfate proteoglycan (HSPG) type play an important role in signal transduction. Cells that cannot synthesize HSPGs or are treated with inhibitors of glycosaminoglycan synthesis, for example, show greatly decreased FGF binding and signal transduction (Yayon et al., 1991; Olwin and Rapraeger, 1992). Similarly, heparin and related polyanions such as protamine sulfate (Neufeld and Gospodarowicz, 1987), pentosan polysulfate (Wellstein et al, 1991), or hexadimethrine (Cook et al., 1995), which are soluble competitors for cell-associated HSPGs, strongly modulate the ability of specific FGFs to bind to high-affinity sites and initiate signaling. Low-affinity FGFRs appear to induce a conformational change in the FGF molecule, possibly dimerization, that enables binding to the high-affinity FGFRs. A current model suggests that FGF must form a ternary complex with both low-affinity and high-affinity receptors in order to initiate signaling (Nugent and Edelman, 1992; Kan et al., 1993; Spivak-Kroizman et al., 1994). Studies by Ishihara (1994) and by Guimond et al. (1993) indicate that the interaction of FGFs with low-affinity receptors exhibits some degree of specificity, as different members of the FGF family require distinct saccharide sequences for bind-

ing and signaling. This may explain some contradictory results regarding the necessity for heparin in FGF signaling (e.g. Rifkin et al., 1994). Internalization of FGF: Low-affinity FGF receptors as well as high-affinity receptors have been postulated to be involved in the internalization of FGFs (Roghani and Moscatelli, 1992; Zhan et al., 1992; Rusnati et al., 1993; Hawker and Granger, 1994). This internalization has been suggested to be important in the transport of FGF-1 and -2 to the nucleus and comcomitant mitogenic effects. However, since cells that do not respond to FGF by proliferating also internalize FGFs (GannounZaki et al., 1994), internalization of FGF may have other functions, such as the clearance of active growth factor from the cell surface. Interactions with other growth factors: FGFs can modulate the composition of the extracellular matrix (ECM) by up-or down-regulating genes for specific cellular receptors, ECM molecules, as well as matrix-degrading proteases (e.g. Tan et al., 1993); thus to some extent, FGFs can modulate their own functions. At the same time, other growth factors such as the transforming growth factor-betas (TGF-^Ss) also regulate the synthesis of ECM proteins or protease inhibitors. The cell surface may thus be considered an interface at which FGFs interact with other growth factors. In endothelial cells, for example, FGF-2 has been observed to regulate plasminogen activator activity and thus to regulate the activity of TGF-)8 (Flaumenhaft et al, 1992), while TGF-^ in turn may regulate the activity of FGF-2 on these cells by inducing increased synthesis of the protease inhibitor TIMP-2 (Murphy et al., 1993). In neural cells, interactions of FGFs with TGF-ySs (in glial cells; Ryken et al., 1992) and of FGFs with IGFs (in hypothalamic cell cultures; Pons and TorresAleman, 1992) have also been described. 4. Localization of FGFS and FGFRs in the nervous system The central nervous system is one of the richest sources of both FGF-1 and FGF-2. Calculations

S. Bieger and K. Unsicker

indicate that approximately 50 ng of FGF-2 is present in one gram of cerebral cortex, while spinal cord contains about 2500 biological units (approximately 250 ng) of FGF-1 per gram tissue (Eckenstein et al., 1991; Eckenstein, 1994). Thus, FGF-1 and FGF-2 by far exceed the amounts of nerve growth factor (NGF) in central nervous system regions, where it is highly expressed (Korsching et al., 1985). Although both proteins and mRNAs occur in abundance, their precise localization has not been fully and reUably mapped, in particular with regard to specific populations of neurons and glial cells. 4.1. Caveat Results describing the distribution of FGF-1 and FGF-2 proteins and their mRNAs in the central nervous system are consistent in some parts, but inconsistent in others. A study by Hanneken and Baird (1992) addressing the localization of FGF-2 protein in the eye strikingly exemplifies the dilemma: FGF immunoreactivity varies significantly depending on the antibodies used and the tissue fixation procedures employed. A comparable study has not yet been performed for the brain and spinal cord, but the same sources of variability are likely to be important. These observations emphasize the need to verify immunohistochemical data by other methods, such as in situ hybridization, bioassay, and Western blot. There is also disagreement with regard to FGF-1 and FGF-2 mRNA expression. Several groups have published Northern blots of embryonic rat brain RNA with multiple hybridization bands indicating the presence of FGF-2 mRNAs (Ernfors et al., 1990; Powell et al., 1991; Weise et al., 1993), while nuclease protection assays reveal little, if any FGF-2 mRNA at the time of birth (Riva and Mocchetti, 1991). The data summarized below should therefore be interpreted with caution, especially if the results are not confirmed by more than one method. 4.2. Localization of FGF-1 and -2: tissue levels 4.2.1. Central nervous system FGF-1 and FGF-2 are found in different ratios in different areas of the central nervous system and

345

peripheral nervous system (Eckenstein et al., 1991). The first study using a combination of methods (Westem blot, immuno- and bioassay) rather than a single technique (e.g. immunocytochemistry) revealed that sciatic nerve contained very high levels of FGF-1, but very little FGF-2. Transection caused FGF-1 to disappear from the distal, but not from the proximal stump, suggesting its presence and possible transport in axons. In the same study the optic nerve was found to contain both FGF-1 and -2. Transection induced FGF-2 in the distal nerve portion, indicating synthesis of FGF in glial (probably astroglial) cells. FGF-1. Initial studies on the regional distribution of FGF-1 mRNA had already suggested its significantly higher expression in the brainstem as compared to cerebrum and cerebellum (Sullivan and Storch, 1991). Early investigations (Pettmann et al., 1986; Janet et al., 1987) using antibodies that crossreacted with both FGF-1 and -2 indicated that FGFs are localized exclusively to neurons, in virtually all regions of the adult central nervous system. Later, in situ hybridization studies revealed that FGF-1 mRNA and protein occur in specific neuronal populations, notably motor neurons of brainstem and spinal cord, sensory neurons of spinal and trigeminal ganglia, and the trigeminal mesencephalic nucleus (Elde et al., 1991; Kresse et al, 1995). Other brainstem and cerebellar nuclei including substantia nigra, red nucleus, reticular, pontine, raphe, cochlear, and vestibular nuclei, as well as distinct nuclei of the brainstem auditory pathways, also contain substantial numbers of FGF-1-positive (mRNA and protein-expressing) neurons (Bean et al, 1991; Stock et al., 1992; Kresse et al., 1995; Luo et al., 1995). FGF-1 was consistently absent from non-neuronal cells. Within the prosencephalon, FGF-1 positive neurons are sparse in the cerebral cortex and hippocampus, but abundant in the septal area, nucleus basalis, and select nuclei of the thalamus (paraventricular, anterodorsal, lateral geniculate) and hypothalamus (lateroanterior). Other areas including pallidum and many thalamic and hypothalamic nuclei contain smaller but still substantial numbers of FGF-1 immunoreactive neurons (Stock et al, 1992).

346

FGF-2. The distribution of FGF-2-synthesizing and -storing neurons in the adult cerebral cortex and subcortical areas is restricted to a few loci. In situ hybridization identified the CA2 field of the hippocampus, layers 2 and 6 of the cingulate cortex, induseum griseum, and fasciola cinereum as the only sites of FGF-2 mRNA expression (Emoto et al, 1989). Colocalization of FGF mRNA and FGFimmunoreactivity was demonstrated within the CA2 region of the hippocampus, making the pyramidal neurons in this area the principal FGF-2-expressing neuron population of the telencephalon (GomezPinilla et al., 1992; Woodward et al., 1992). FGF-2 immunoreactivity has also been reported to occur in neurons of many brainstem sensory, motor, and relay nuclei (Grothe et al., 1991; Gomez-Pinilla et al, 1992; Grothe and Unsicker, 1992; Matsuda et al., 1992a; Chadi et al, 1993b; Grothe and Janet, 1995; Riedel et al., 1995; Fuxe, personal communication) as well as in Purkinje cells (Matsuda et al., 1992b) and spinal cord motoneurons (Otsuka et al, 1993; Hassan et al., 1994). No signs of FGF-2 mRNA have been detected in brainstem neurons by Luo et al. (1995). Although FGF-2-synthesizing and -immunoreactive neurons are found together with FGF-1-synthesizing neurons in the rat and monkey substantia nigra (Bean et al., 1991), a colocahzation of FGF-1 and FGF-2 in single neurons has not been demonstrated. Non-neuronal cells identified as astrocytes by co-staining for glial fibrillary acidic protein display strong, predominantly nuclear, but also cytoplasmic immunoreactivity for FGF-2 throughout the brain (Gomez-Pinilla et al., 1992; Woodward et al., 1992). This distributional pattern concurs with the presence of FGF-2 mRNA in the cerebral cortex, hippocampus, and hypothalamus (Emoto et al., 1989) as well as with the prominent bioactivity that can be detected in isolated cortical astrocytes, but not in isolated cortical neurons (Woodward et al., 1992). In summary, the available data suggest that in the central nervous system, FGF-2 is predominantly synthesized and stored by astrocytes; its localization in brainstem neurons (with the exception of the substantia nigra) requires further evidence from in situ hybridization. In contrast, FGF-

Functions of fibroblast growth factors (FGFs) in the nervous system

1 is present in many neuron populations throughout the central nervous system. 4.2.2. Sensory organs The eye has long been a preferred object in studies on FGF (Baird et al., 1985; Courty et al., 1985). FGF-1 is most abundantly expressed in rat photoreceptors, but also present in the other layers of the retina (Caruelle et al., 1989; Bugra et al., 1993). The rat neural retina also seems to express FGF-2 mRNA (Rakoczy et al., 1993), but the cellular localization is not entirely clear (see Gao and Hollyfield, 1992). Interestingly, pinealocytes of the pineal gland, which are phylogenetically related to retinal photoreceptors, are immunonegative for FGF-2 (Marin et al., 1994). In the auditory system, FGF-2 (immunoreactivity) has been localized to developing inner hair cells, spiral ganglion cells (Despres et al, 1991), and distinct neuron populations of auditory relay nuclei (Riedel et al., 1995; cf. Luo et al., 1995, however). LocaUzation of FGF-2 to the cochlea might be artefactual, since its mRNA was not detected there at any age (Luo et al., 1993). However, both developing hair cells and spiral ganglionic neurons have been shown to express FGF-1 mRNA (Luo et al., 1993, 1995), suggesting that FGF-1 may be involved in the establishment of cochlear innervation. 4.2.3. Peripheral nervous system In the peripheral nervous system, FGF synthesis and localization have been best studied in rat sensory ganglia. Most dorsal root ganglionic neurons, including all size classes, express high levels of FGF-1 mRNA (Hide et al., 1991; OeUig et al., 1995), whereas FGF-2 mRNA is restricted to a subpopulation of neurons, possibly the somatostatinand bombesin-immunoreactive neurons (Weise et al, 1992). FGF-1 and -2 expression in sympathetic, parasympathetic, and enteric ganglia has not been fully clarified. One study suggested that FGF-2 immunoreactivity was absent from the sympathetic superior cervical and parasympathetic otic ganglia (Weise et al., 1992). The neuroendocrine subdivision of the sympathetic nervous system, the adrenal chromaffin cells, express FGF-2 mRNA and protein (Grothe and Unsicker, 1989, 1990; Stachowiak et

S. Bieger and K. Unsicker

347

al., 1994; Bieger et al, 1995), possibly in the noradrenergic subpopulation only (Grothe and Unsicker, 1990).

which FGF is imported into and exported from these cells, as from other types of non-neural cells, remains to be clarified.

4.3, Subcellular localization ofFGF-1 and -2

4.4. Axonal transport of FGFS

One of the intriguing features of FGF-2 is the occurrence of distinct nuclear and cytosolic forms (see Section 2.2.2). Immuno-electron microscopic analysis has demonstrated that FGF-2-synthesizing astrocytes and neurons contain FGF both in the nucleus and the cytoplasm (Woodward et al., 1992). Particularly interesting is a report describing the colocalization of FGF-2 and the gap junction protein connexin 43 in astrocytes (Yamamoto et al., 1991), which has also been observed in cardiac myocytes (Doble and Kardami, 1995). This colocalization may indicate a role of FGF-2 in regulating astroglial coupling. Early electron microscopic studies using antibodies that crossreacted with both FGF-1 and FGF-2 showed immunolabeling of neurons throughout the cytosol and in large-, but not small-diameter processes (Janet et al., 1987). An electron microscopic study using monospecific antibodies to FGF-1 clearly revealed a distinct labeling on the inner side of the axonal membrane of myelinated and unmyelinated axons within peripheral nerves (Elde et al., 1991). Antibodies recognizing both nuclear and cytosolic forms of FGF-2 have been employed in ultrastructural and cell fractionation analyses of the neuron-like adrenal chromaffin cells (Stachowiak et al., 1994; Bieger et al., 1995). Interestingly, in these studies, 'Nuclear' (amino-terminally extended) and 'cytoplasmic' forms were found in both nuclear and cytosolic fractions. A substantial fraction of FGF-2 was located on the outer surface of the cell membrane, consistent with the welldocumented presence of FGF-2 in the extracellular matrix. Most of the cytoplasmic FGF-2 was found in endosome-like structures, while secretory granules, which had previously been reported to contain FGF-2 (Westermann et al., 1990), were devoid of label at the ultrastructural level. Accumulation of FGF in the nucleus or the cytoplasm could be triggered by activating either cAMP or protein kinase C pathways. However, the mechanism by

Neurons, like other cell types, can internalize FGFs. This observation raises the possibility that FGFs could function as retrograde messengers by axonal transport. Ferguson and co-workers (1990) were the first to investigate receptor-mediated internalization and transport of FGF-2. They found no evidence for retrograde transport of iodinated protein in sciatic, trigeminal, sympathetic, and optic nerves, but striking anterograde transport by retinal ganglion cells. Expanding these studies, Ferguson and Johnson (1991) reported selective uptake and subsequent retrograde transport of iodinated FGF-1 and FGF-2 after intracerebral injection in specific populations of neurons, including the nigrostriatal system. Although there is evidence against retrograde axonal transport of FGF in spinal cord motoneurons (Ferguson et al., 1990), another report indicates that the factor is retrogradely transported in hypoglossal motoneurons (Grothe and Unsicker, 1992). Whether this reflects distinct transport properties of different motoneuron populations remains to be shown. The significance of receptor-mediated internalization of FGFs, however, may not only reside in their possible axonal transport. There is evidence that FGFs can elicit immediate responses in the axon terminal, resulting in a local rise of second messenger components which may then constitute the retrograde axonal signal (Hendry, 1992; see also chapter by Hendry in this volume). 4.5. Developmental regulation of FGF-1 and -2 expression The earliest expression of FGF-2 (immunoreactivity) has been reported for the neuroepithelium of the ElO mouse embryo (Ford et al, 1994). Whether this is authentic FGF-2 is not clear, since FGF-2 mRNA levels measured by RNase protection assay are very low (less than 1-10% of adult levels) in all central nervous system regions of the newborn

348

rat (Riva and Mocchetti, 1991). Region-specific alterations in FGF-2 mRNA expression during development have been further investigated in a recent study by El-Husseini et al. (1994), who demonstrated that FGF mRNA levels increased during the first postnatal month in the occipital cortex and inferior coUiculus, while the highest expression at postnatal day 1 was found in the cerebellum. However, others have reported an earlier onset of FGF-2 expression in the brain (Emfors et al, 1990; Gonzalez et al., 1990; Weise et al., 1993). Since the initiation of FGF-2 expression during early development may vary between central nervous system neuron populations (see Riedel et al, 1995) and may be transient, the issue of when and where FGF-2 is expressed will only be solved by more thorough in situ hybridization studies. As for FGF-2, several studies support the hypothesis that FGF-1 may be more important for events in the adult than in the developing central and peripheral nervous systems (e.g. Ishikawa et al, 1991a; Oellig et al., 1995). Levels of FGF-1 quantified by two-site ELISA are very low in all areas of the neonatal rat brain, but increase significantly between three and seven weeks of age. Extremely high levels can be detected in the pons/medulla oblongata, while levels in several cortical areas, hippocampus, olfactory bulb, cerebellum and striatum are approximately 10-fold lower (see Oellig et al., 1995). The available data clearly indicate regionspecific differences. Somewhat in contrast to the above studies, an in situ hybridization study by Wilcox and Unnerstall (1991) demonstrated strong signals in the prenatal brain, particularly over the developing cortical plate. However, these authors also reported that labeling in the hippocampus became more intense with maturation, while labeling of cortical, nigral, and locus coeruleus neurons became significantly reduced in the adult brain. Dorsal root gangha show a prominent increase in FGF-1 during postnatal development, reaching approximately 0.1% of the total protein in ganghonic extracts in the adult rat (OelHg et al., 1995). 4,6. FGFs other than FGF-1 and FGF-2 Several members of the FGF family other than

Functions of fibroblast growth factors (FGFs) in the nervous system

FGF-1 and FGF-2 have been reported to be expressed in the nervous system and may have important functions, especially in the development of the nervous system. 4AL FGF-3(int-2) In the early embryo, FGF-3 is expressed beginning just before gastrulation, but is rarely found in adult tisssues (Tannahill et al., 1992). Particularly intriguing is the observation that rhombomeres 4 and 5, and later 6, in the hindbrain adjacent to the developing otocyst express high levels of FGF-3 mRNA (Wilkinson et al., 1988; Mahmood et al, 1995). During further development of the hindbrain, FGF-3 expression disappears from these rhombomeres and becomes restricted to rhombomere boundaries, where it is expressed until the morphological boundaries are lost. This localization is consistent with the documented role of FGF-3 as a signal for the induction of the otic vesicle (Represa et al., 1991), but may also point at boundary-associated functions. 4,62. FGFS There is widespread expression of FGF-5 in different tissues during embryonic development (Haub and Goldfarb, 1991). In the adult mouse and rat brain, FGF-5 mRNA can be localized to select neuron populations in the olfactory bulb, primary olfactory cortex, hippocampal formation, entorhinal cortex and neocortex (Haub et al., 1990; Gomez-Pinilla and Cotman, 1993). In the retina, pigment epithelial cells synthesize FGF-5 (Bost et al., 1992). The observation that FGF-5 is strongly expressed in developing skeletal muscle during the period of ontogenetic neuron death prompted an investigation of potential trophic effects of FGF-5 on motoneurons (Hughes et al., 1993a,b). FGF-5 was found to have pronounced effects on the survival of cultured embryonic chick motoneurons. The significance of this finding, as well as of the observation that FGF-5 affects the differentiation of cultured rat septal cholinergic and raphe serotonergic neurons, is however uncertain, since FGF-5 knockouts show normal central nervous system development (Hebert et al., 1994).

5. Bieger and K. Unsicker

4,63. FGF-7 FGF-7 mRNA expression in the developing mouse brain is restricted to a 24 h period around E14.5, when transcripts can be localized to three distinct sites within the ventricular layer of the forebrain, including the ganglionic eminence (Mason et al., 1994). FGF-7 does not seem to be expressed in the peripheral nervous system. It is not known whether FGF-7 has mitogenic functions or participates in the differentiation processes of neural progenitor cells. 4.6.4. FGF-8 FGF-8 mRNA is present from E9 to El2 of murine gestation as shown by RNase protection assay and in situ hybridization studies (Heikinheimo et al., 1994; Ohuchi et al., 1994; Crossley and Martin, 1995). FGF-8 is widely expressed in the telencephalon, diencephalon, and metencephalon, particularly in the midbrain-hindbrain junction, and in the developing pituitary gland. These observations suggest an important role of FGF-8 in early brain development. 4.6.5. FGF'9 Expression of the FGF-9 gene can be detected in the brain, and the pattern of its distribution has just been revealed (Tagashira et al., 1995). 4.7. Localization ofFGF receptors in the central and peripheral nervous systems The regulation of FGFR expression is likely to be important in determining FGF signaling, particularly since the various FGFs apparently have preferences for different FGF receptor subtypes (see Section 3). Several groups have shown that individual FGF receptor genes (FGFR-1 to -4 ) have different patterns of expression in embryonic and adult tissues, which may indicate specific roles for each receptor during development. The expression, developmental regulation, and spatial distribution of FGFR-1 and FGFR-2 have been most thoroughly studied (see Heuer et al., 1990; Reid et al., 1990; Wanaka et al., 1990; 1991; Peters et al., 1992; Asai et al., 1993; El-Husseini et al., 1994). Several reports have suggested that neu-

349

rons preferentially express FGFR-1 mRNA, while glia express FGFR-2 mRNA (Peters et al., 1992; Asai et al., 1993). This may be different for cultured neural cells: human foetal neurons (and microglia) have been reported to express all four FGFRs at different levels, whereas astrocytes express only FGFR-1 and FGFR-4 (Balaci et al., 1994). As far as localization of receptor protein is concerned, immunocytochemical results obtained with the commercially available antibodies to FGFR-1, which are not accompanied by mRNA localization data, should be interpreted with caution. FGFR-1 transcripts can already be detected in the germinal layers of the neuroepithelium in the chicken and mouse nervous system (Heuer et al., 1990; Reid et al., 1990). Expression ceases as cells migrate into the mantle layer and returns in maturing neuron populations. In the adult rat central nervous system, FGFR-1 mRNA is strongly expressed in hippocampus, cerebellum, brainstem, and spinal cord. By in situ hybridization, strong signals are detectable over the mitral cell layer of the olfactory bulb, hippocampal pyramidal and dentate gyrus granular neurons, in layers 5 and 6 of the cerebral cortex, parts of the amygdaloid complex, neurons of the hypothalamic median eminence, cerebellar Purkinje and granular cells, pontine nuclei, the pedunculopontine tegmental nucleus as well as motoneurons of brainstem and spinal cord (Heuer et al., 1990; Wanaka et al., 1990; Yazaki et al., 1994). These neuron populations are partially identical to the neuron populations exhibiting pronounced FGFR-1 immunoreactivity (Matsuo et al., 1994). In contrast to the strong FGFR-1 expression observed in the above studies, Bugra et al. (1994) found almost undetectable levels of FGFR-1 mRNA expression in the unlesioned rat hippocampus, but significant elevations following kainate-induced seizures. In accordance with this study, others have reported that FGFR-1 protein is hardly detectable in the adult rat brain (including the hippocampus), but is significantly upregulated following kaninate-induced seizures (Van der Wal et al., 1994). The immunoreactivity observed after lesioning is apparently associated with both astrocytes and neurons.

350

With regard to FGFR-2, the overall pattern of its expression in the central nervous system is most consistent with a glial expression (Peters et al., 1995). Thus, fiber tracts devoid of neuronal perikarya, such as the cerebellar peduncles, hippocampal fimbria and corpus callosum, are not labeled with FGFR-1-specific probes, but are moderately to heavily labeled with FGFR-2. There are apparent exceptions to the glial expression pattern. The cerebellar Purkinje cell layer, for example, labels strongly for FGFR-2, and dorsal root ganglionic neurons (the large and medium-sized subpopulations) show labeling not only for FGFR-1, but also for FGFR-2 mRNA (Oellig et al., 1995). In the retina, FGFR-2 protein has been localized to the synaptic and optic fiber layers, but these studies do not enable resolution of the precise cellular localization (axons versus glia; Torriglia and Blanquet, 1994; Torriglia et al, 1994). Schwann cells, the glia of the peripheral nervous system, show no detectable expression of either FGFR-1 or FGFR-2 (Oellig etal., 1995). FGFR-3 transcripts can be detected early in the germinal epithelium of the neural tube and are expressed diffusely in the postnatal and adult mouse brain, in cells with the morphological characteristics of glia (Peters et al., 1993; Yazaki et al., 1994). FGFR-3 is also expressed at high levels in differentiating hair cells of the cochlear duct, but not in other sensory epithelia (Peters et al., 1993). Dorsal root ganglionic neurons express neither FGFR-3 nor FGFR-4 (Oellig et al., 1995). FGFR4 mRNA has an extremely restricted distribution in the central nervous system: its expression is confined to the medial habenular nucleus (Itohetal., 1994). 5. Functions of FGFs in the nervous system The 'original' FGFs, FGF-1 and -2, which have been commercially available as purified or recombinant proteins for a number of years, are by far the best studied in terms of their actions on cells of the nervous system. Less is known about the newer FGF family members, which are only recently becoming characterized in terms of their expression and their tissue distribution patterns.

Functions of fibroblast growth factors (FGFs) in the nervous system

For this reason, the following sections will focus on the functions of FGF-1 and -2; information on other FGFs will be included where it is available. 5.1. Functions of FGF for neurons 5.LL Survival Promotion of neuronal survival is perhaps one of the most well-documented effects of FGF-2 (bFGF). In ng/ml concentrations, FGF-2 has been shown to reduce cell death in cultures of neurons isolated from a variety of embryonic brain regions (Table 2). The mechanism by which FGF-2 exerts this neurotrophic effect has been poorly investigated. A study by Schmidt and Kater (1993), indicating that the survival-promoting effect of FGF-2 is maximal in the presence of either laminin or depolarizing stimuli, suggests that additional signals are necessary. In at least some systems, indirect effects of FGF-2 on the proliferation or the function of non-neuronal (especially glial) cells may be important (Engele and Bohn, 1991). Neuron populations supported by FGF-1 (aFGF) overlap partially with those supported by FGF-2, but the two factors have significantly different potencies for different neuron types. Whereas the TABLE 2 Some neuron populations supported by FGF-2 in vitro Neuron source

References

Cerebral cortex

Morrison, 1986; Walicke, 1988 Ferrari et al., 1989; Engele and Bohn, 1991 Walicke, 1988; Grotheetal., 1989 Kushima et al., 1992 Walicke et al., 1986 Matsuda et al., 1990 W^alicke, 1988 Zhou and DiFiglia, 1993 V^alicke, 1988 Ishikawa et al., 1992 Unsicker et al., 1987; Sweetnam et al., 1991 Unsicker et al., 1987; Schubert et al., 1987

Mesencephalon Septum (embryonic) (adult) Hippocampus (embryonic) (postnatal) Striatum (embryonic) (postnatal) Thalamus Hypothalamus Spinal cord Ciliary ganalion

S. Bieger and K, Unsicker

majority of neuron populations respond to far lower concentrations of FGF-2 than of FGF-1, certain types of neurons, such as retinal ganglion cells (Lipton et al., 1988), appear to be much more sensitive to FGF-1. The only other FGF demonstrated to affect neuron survival in vitro (to date) is FGF-5, which specifically enhances the survival of cultured spinal motoneurons (Hughes et al., 1993a). It should be noted that the ability of FGFs to promote neuronal survival in vivo has been less well investigated. Although Dreyer et al. (1989) demonstrated that FGF-2 administration at an appropriate time can rescue neurons of the ciliary ganglion from ontogenetic cell death, other investigators have failed to observe in vivo rescue effects for FGF-1 (Hill et al., 1991) or FGF-1/ FGF2 (Oppenheim et al., 1992). In general, the observation that specific FGFs occur at an appropriate time and in an appropriate region during development is taken as evidence that they may play a role in the normal development of certain populations of neurons, but there is no rigorous proof that they can act as classical neurotrophic factors of the NGF type. 5.1.2. Proliferation Apart from being a 'maintenance' factor for cultured neurons, FGF-2 is also a mitogen for certain populations of neuronal precursors. Proliferation in response to FGF treatment has been observed in cultures of neuronal progenitors isolated from the cerebral hemispheres (Gensburger et al., 1987), hippocampus (Oshawa et al., 1993), mesencephalon (Mayer et al., 1993a) and spinal cord (Deloulme et al., 1991; Sweetnam et al., 1991), in immortalized neural progenitors (Kitchens et al., 1994) and in striatal progenitor cells that have been 'primed' with EGF (Vescovi et al., 1993; Gritti et al., 1995). However, in situ hybridization studies have failed to show the appearance of FGF-2 at appropriate times (i.e. periods during which neuronal proliferation occurs) in the developing brain, suggesting that an FGF other than FGF-2 could function as physiological mitogen (Eckenstein, 1994). FGF-8 may be a promising candidate (Crossley and Martin, 1995). In the pe-

351

ripheral nervous system, FGF-2 has been shown to influence the development of the sympathoadrenal lineage: FGF treatment drives the division as well as the differentiation of precursor cells into NGFdependent sympathetic neurons (Stemple et al., 1988; Birren and Anderson, 1990). 5.1.3. Differentiation One of the most striking effects of FGF-2 in vitro is the induction of neurite outgrowth from cultured neurons. This phenomenon has been observed for a wide variety of neurons of the central nervous system, including hippocampal (Walicke et al., 1986), cortical (Morrison et al, 1986; Kornblum et al., 1990), cerebellar granule (Hatten et al., 1988) and others (Walicke, 1988). Examples of peripheral nervous system neurons responding in this fashion to FGFs are neurons of the ciliary ganglion (Hill et al., 1991) and cells of the sympathoadrenal lineage (neurons and adrenal chromaffin cells; see Section 5.1.2). Outgrowth of neurites in response to FGF-2 treatment is only one example of the ability of FGFs to influence the neuronal phenotype. The literature abounds with reports of the effects of FGF-2 on specific functions and on phenotypic features of cultured neurons. In general, FGF-2 acts as a differentiative signal for cultured neurons, for example by stimulating enymes required for synthesis of neurotransmitters (Vaca et al., 1989; Knusel et al., 1990; Sweetnam et al., 1991; Zurn, 1992; Puchacz et al., 1993; Yokoyama et al., 1994) and by stimulating the expression of other markers of differentiation, such as acetylcholinesterase activity (Gensburger et al., 1992), neurotransmitter uptake (Ferrari et al., 1989, 1991) or, in the retina, expression of opsin (Hicks and Courtois, 1992). FGF-2 also influences the adhesive interactions of cultured neuronal cells such as neuroepithelial cells of the neural tube (Kinoshita et al., 1993) or the 'neuronal' PC-12 cell line (Schubert et al., 1987); these interactions are important in determining cell migration and morphology. FGF-1 generally exerts effects similar to FGF-2, i.e. differentiation and neurite outgrowth (e.g. Rydel and Greene, 1987; Du et al., 1994). Although FGF-1 is often found to be less potent than FGF-2 in vitro, it

352

is significantly more active for certain neuronal populations, e.g. retinal ganglion and ciliary neurons (Lipton et al., 1988; Hill et al., 1991). FGF-5 can induce choline acetyltransferase activity in septal neurons and serotonin uptake in raphe neurons (Lindholm et al., 1994b), indicating its potential function as neuronal 'differentiator'. A number of reports suggest that FGF-2 not only affects developmental differentiation of neurons but also stabilizes the mature neuronal phenotype. Examples are the ability of FGF-2 to maintain neurotransmitter stores in cultured neonatal chromaffin cells without inducing transmittersynthesizing enzymes (Unsicker and Westermann, 1992) and to stabilize acetylcholine receptors in cultured muscle cells, indicating a possible function at the neuromuscular junction (Dai and Peng, 1992). An interesting aspect of FGF-2 function is its reported ability to affect the expression and function of ion channels. Treatment of PCI2 cells with FGF-2 for several days induces increases in sodium channel density (Pollock et al., 1990) and increased calcium currents (Rane and Pollock, 1994). FGF-2 may act even more rapidly, as indicated by its enhancement of NMDA receptormediated calcium influx into hippocampal cells (Abe and Saito, 1992), its stimulation of light responses in photoreceptors (Schmidt et al., 1995) and its ability to enhance long-term potentiation in hippocampal slices after short-term treatments (20 min) (Terlau and Seifert, 1990), an effect also observed with FGF-1 (Sasaki et al, 1994b). Again, it should be noted that most of these effects of FGF have been observed in vitro and cannot be extrapolated to the in vivo situation, in which many other FGF-responsive cells and other factors interacting with FGF provide complications. Lewis et al. (1992), for example, have shown that intraocular injection of FGF induces proliferation of a variety of nonneural cells (glial, vascular, epithelial, etc.). Such a multiplicity of effects could certainly be expected in many other organs or systems exposed to FGFs. 5 J.4. Intracellular mechanisms of FGF action Many of the effects of FGFs mentioned above can be explained in terms of the activation of the

Functions of fibroblast growth factors (FGFs) in the nervous system

FGF receptors and stimulation of associated signal transduction pathways. Recently, Williams et al. (1994) have provided convincing evidence that FGF-2 induces neurite outgrowth from cultured cerebellar neurons by a cascade involving activation of phospholipase C-gamma, stimulation of the production of arachidonic acid and enhanced Ca^^ influx. In these studies, arachidonic acid itself could mimic the neuritogenic effect of FGF-2. Further mediators of the neurite growth response to FGF have been identified in the rat PCI2 pheochromocytoma cell line. p21 ras protein (Altin et al., 1991) and the microtubule-associated protein kinase (Sano and Kitajima, 1992) appear to play important roles in FGF signalling. Aoyagi et al. (1994) have demonstrated that, in cultured hippocampal neurons, FGF-2 stimulates neurite branching (bifurcation) rather than elongation. FGF-2 may induce neurotransmitter synthesis in PC 12 cells by regulation at the post-transcriptional level, i.e. by modulation of tyrosine hydoxylase activity through enzyme phosphorylation (see Haycock, 1993); but presumably regulation occurs also at the transcriptional level, through induction of the tyrosine hydroxylase gene, as indicated by studies with adrenal chromaffin cells (Puchacz et al., 1993) and cultured striatal neurons (Du et al., 1994). FGF-2, like NGF and other growth factors, can induce a number of other neuronal proteins such as MAPI, Thy-1 and phosphorylated NILE (Rydel and Greene, 1987). The frequent observation of nuclear FGF-2 immunoreactivity suggests that this growth factor may exert an effect directly at the nucleus. In fact, in vascular endothelial cells, FGF-2 is transported into the nucleolus at specific phases of the cell cycle, and apparently directly stimulates the transcription of ribosomal genes and enhances the activity of RNA polymerase I (Bouche et al., 1987; Baldin et al., 1990; Amalric et al, 1994). FGF-1 has similarly been shown to undergo cell cyclespecific nuclear import (Imamura et al, 1994). Although this transport to the nucleus has been thought to be mediated by low-affinity FGF receptors, the observation that the (high-affinity) FGF receptor type 1 is transported to the nucleus of NIH 3T3 cells after stimulation of proliferation

S. Bieger and K. Unsicker

(Prudovsky et al., 1994) complicates the 'classical' view of FGF receptor action. Further studies may reveal novel mechanisms of FGF action. Some FGF effects on neuronal phenotype could be interpreted as indirect effects on the extracellular matrix (rather than as effects on cellular structural proteins themselves). In the nervous system, such effects are likely mediated by glial and nonneural cells (see next Section). 5.2. FGF effects on glial cells 5.2.1. Proliferation Glial cells as well as neurons can respond to FGF-2. One of the most significant effects of FGF2 is its induction of glial cell proliferation, documented in vitro with astroglia (Sensenbrenner et al., 1987) as well as with ohgodendrocytes (Eccleston and Silberberg, 1985; Besnard et al., 1989), Schwann cells (Davis and Stroobant, 1990), and glial progenitors (McKinnon et al., 1990; Engele and Bohn, 1992). The potency of FGF-2 (and, in at least some cases, FGF-1) in stimulating glial proliferation suggests that in vivo, FGF could play an important role in processes involving glial proliferation - development, gliosis (e.g. in wound healing) and tumour formation. In fact, normal astrocytes synthesize and release FGF-2 (Araujo and Cotman, 1992a) and can be prevented from proliferating by antisense FGF-2 oligonucleotides (Gerdes et al., 1992), indicating that FGF may act in autocrine fashion. This autocrine function is supported further by the observation that many transformed glial cells express unusually high levels of FGF-2 (Stefanik et al., 1991; Takahashi et al., 1992) and also may be inhibited from proliferating by antisense FGF-2 oligonucleotides (Morrison, 1991; Murphy et al., 1992; Morrison et al., 1994). The mechanisms underlying the mitogenic response of glial cells to FGF have been studied in some detail; not surprisingly, activation of tyrosine kinase activity, followed by activation of phospholipase C and protein kinase C and leading to increased c-fos expression, are key initial steps (Radhakrishna and Almazan, 1994). In 02-A progenitor cells, FGF-2 has been shown to induce the

353

transcription of the SCIP protein, a transcription factor which binds to POU domains and thereby stimulates proliferation while inhibiting differentiation (Collarini et al., 1992). Synergistic interactions of FGF-2 with serum (Kniss and Burry, 1988), with retinoic acid (Laeng et al., 1994) and with other growth factors such as PDGF (McKinnon et al, 1990) have been described. In fact, the rapid induction of FGF-2 after stimulation of C6 glioma cells with serum or with phorbol ester (Powell and Klagsbrun, 1993) indicates that FGF-2 may be a general mediator of other proliferative signals. 5.2.2. Differentiation Along with mitosis, FGF-2 can stimulate alterations in glial cell morphology, notably by inducing structural proteins such as intermediate filaments, actin, and GFAP (Weibel et al., 1985, 1987; Perraud et al., 1988) and by inducing other glial markers, some of which are summarized in Table 2. Overall, the effects on glial morphology suggest that FGF can act as a 'maturation factor' for glial cells. Whether glial cells respond to FGF by maturation or by proliferation undoubtedly depends on the developmental stage as well as on the presence or absence of other growth factors, serum or stimuli such as extracellular ATP (Neary etal., 1994). 5.2.3. Lineage A key role for FGF-2 has been suggested in the development of the oligodendrocyte lineage. In vitro, oligodendrocyte (02A) progenitor cells maintained under serum-free conditions and in the absence of any added growth factor differentiate into mature oligodendrocytes. Treatment of 02A progenitors with FGF-2 inhibits this differentiation and induces cell proliferation, apparently by maintaining high levels of PDGF receptors (McKinnon et al., 1990). The effect of FGF-2 can be counteracted by retinoic acid (Laeng et al., 1994) as well as by factors released from 02A precursors (Fressinaud et al., 1993; Grinspan et al., 1993) or from astrocytes (Mayer et al., 1993), which are thought to enable oligodendrocyte differentiation in vivo. Mature oligodendrocytes re-

354

spond to FGF treatment with de-differentiation (Wolswijk and Noble, 1992; Fressinaud et al., 1995) and can be induced to proliferate by combined FGF-2/PDGF treatment, suggesting a means by which oligodendrocyte progenitor populations may be replenished. A recent report suggests that mature oligodendrocytes undergo apoptosis in the absence of FGF-2 (Yasuda et al, 1995). This indicates that the activity of FGF-2 must be highly regulated in the nervous system to allow establishment and maintenance of mature oligodendrocytes. In the development of the Schwann cell lineage, also, FGF-2 may play an important role. A number of studies (Stocker et al., 1991; Sherman et al., 1993; Eckenstein et al., 1994) have shown that FGF-2 induces neural crest cells (progenitors to Schwann and other cells) to differentiate into melanocytes. Mature Schwann cells respond to FGF-2 by proliferating (Davis and Stroobant, 1990; Peulve et al., 1994). Together, these observations show that, at least in vitro, FGF-2 can influence the differentiation of various glial cell types by acting either as a ^maturation' factor (astrocytes) or as a *dedifferentiation/ proliferation' factor (oligodendrocytes and Schwann cells). 5.2.4. Mechanisms of FGF action As discussed above, many of the effects of FGF-2 on glial cell morphology (differentiation) involve the direct induction of specific glial structural genes. Some of the effects may also be indirect; for example, they may involve modulation of the composition of the ECM. In cultured astroglial cells, for instance, FGF-2 has been demonstrated to induce plasminogen activator activity (Rogister et al., 1988). This could lead to alteration of local proteolytic activities and thereby influence the morphology of both glial and neuronal cells by altering adhesion of cells to the substratum and their interaction with other cells (e.g. aggregation). Such mechanisms may underlie the observed ability of FGF-2 to enhance the migration of astrocytes (Hou et al., 1995). The importance of the ECM in mediating or modulating the effects of FGF-2 on neuronal cells, particularly in early development, is only beginning to be investigated

Functions of fibroblast growth factors (FGFs) in the nervous system

(e.g. Schmidt and Kater, 1993; Chu and Tolkovsky, 1994). 5.5. 'Higher-level'FGF effects As discussed previously, certain FGFs are expressed at high levels during the development of the nervous system, suggesting their participation in the formation of specific neural structures. FGF3 has been implicated in the normal development of the hindbrain (Carpenter et al., 1993), FGF-7 in the development of the forebrain (Mason et al., 1994) and FGF-8 in the development of the rhombencephalon, telencephalon and metencephalon (see Section 4.6). Intriguing observations have also been made on potential functions of FGF-1 and -2 in the mature central nervous system. The observations that FGF-1 levels are elevated in the cerebrospinal fluid of rats after feeding (Hanai et al., 1989), and that intraventricular application of FGF-1 causes neurological changes associated with satiation (de Sainte Hilaire and Nicolaidis, 1992; Sasaki et al, 1994a), have been suggested as evidence for the involvement of this factor in regulation of feeding. FGF-2 can act as a vasodilator when applied to the pial surface of the brain (Regli et al., 1994; Rosenblatt et al., 1994), indicating that in the central nervous system (as in the periphery), FGF-2 is an important regulatory molecule in the vascular system. The beneficial effects of FGF-2 in ischemia and certain other brain lesions (see Section 6.3) are also undoubtedly mediated in large part by cells of the cerebral vasculature; FGF-2 is highly active in promoting angiogenesis in the brain, as indicated by studies performed by Cuevas et al. (1993). Angiogenic activity has also been ascribed to FGF-1 and -4 (e.g. Brustle et al, 1992; Walter et al., 1993; Yoshidaetal., 1994). 6. Expression and functions of FGFs in lesions and pathologies Levels of FGF-1 and FGF-2 have been reported by a number of investigators to undergo an impressive rise during postnatal brain development. This suggests important, but still poorly understood roles

S. Bieger and K, Unsicker

for the proteins in the adult nervous system. However, the still unclarified mechanism of their secretion (or possibly their failure to be secreted at all from intact cells in the central nervous system) has led to the proposition that the FGFs might serve as 'lesion' and 'disease' factors rather than fulfilling physiological functions in the unlesioned nervous system. In support of such functions, a substantial body of literature demonstrates that FGFs are regulated in peripheral nerve and central nervous system lesions as well as in various neurological disorders, and that they can act both directly and indirectly to facilitate neuronal survival and reconstitution of circuits (for reviews, see Unsicker et al., 1992a,b; 1993, Logan and Berry, 1993; Baird, 1994a). 6.1. Regulation ofFGF and FGF receptor mRNAs and proteins in response to experimental lesions 6.1.1. Central nervous system injuries Patients disabled by penetrating head and spinal cord injuries are estimated to constitute more than 1:1000 of the population of North America (Office of Technology Assessment, USA, 1990; see Logan and Berry 1993). Following the discovery that FGF-1 and FGF-2 have neurotrophic potential (Morrison et al, 1986; Walicke et al., 1986; Unsicker et al., 1987) and are widely distributed in the central nervous system (for review, see Unsicker et al., 1993), numerous studies therefore started to address the regulation of FGF expression in brain injuries and the possible involvement of FGFs in the response to injury. Cerebral cortical (cavity) lesions. In a defined cavity lesion (Finklestein et al., 1988; Frautschy et al., 1991; Logan et al., 1992a) FGF-2 mRNA starts to increase within 4 h after surgery. In areas adjacent and ipsilateral to the lesion peak, FGF-2 mRNA and protein levels peak about 1 week after surgery and decline thereafter. Increased expression could be localized to macrophages and microglia, apparently the earliest cell types to respond, as well as to neurons, astrocytes, ependymal and even vascular endothelial cells (which do

355

not show FGF-2 expression in the unlesioned brain). A focal increase in FGFR-1 mRNA does not consistently appear until 7 days after injury, implying that other FGFRs may be involved in the early lesion response. Reactive and hypertrophic astroglial cells constitute the majority of cells and are intensely FGF-2-immunoreactive between one and two weeks after lesioning. Staining is visible both within nuclei and within the cytoplasm, in contrast to 'normal' astrocytes, which exhibit predominantly nuclear staining. FGF-1 expression seems to follow a different time course, supporting the notion that its functions in the lesioned brain are distinct from those of FGF-2. Following cortical cavity lesioning, FGF-1 (quantified by ELISA) first became detectable in the cortical lesion fluid 10 days after surgery, and its level increased further until 30 days (Ishikawa et al., 1991b). In contrast, NGF levels increased much more rapidly, with a peak 16 h after lesioning. It is unclear how these data can be reconciled with a study reporting a rapid and drastic increase of FGF-1 in a cortical cavity within 1 h after after trauma (Nieto-Sampedro et al., 1988). Given these and other data that suggest roles for the two 'classic' FGFs in central nervous system lesions, there is an apparent need for experiments using locally applied antibodies or antisense oligonucleotides to interfere with endogenous FGFs and thereby elucidate their functional significance. Implantation of microdialysis probes. Microdialysis has become a powerful tool for monitoring extracellular levels of neurotransmitters and neuropeptides in the brain. In light of the widespread use of this method as a monitoring system for 'physiological' events, it is important to realize that implantation of a microdialysis probe into the hippocampus causes within 8 h a massive increase in FGF-2 mRNA in astrocytes in a relatively large area around the implant, including both the hippocampus and the overlying cerebral cortex (Rumpel et al., 1994). There is also an increase in astroglial FGF-2 immunoreactivity. Clearly, such effects have to be considered when evaluating microdialysis data.

356

Lesions of specific central nervous system pathways. Several investigations have shown that FGFs can respond to more subtle lesions than those involving damage of large brain areas. Relatively small lesions of the entorhinal cortex remove a major input to the dentate gyrus and cause massive sprouting of cholinergic septal afferents as well as other inputs in the dentate molecular layer. Following such a lesion, FGF-2 immunoreactivity is markedly increased in astrocytes and in the extracellular matrix of the dentate molecular layer in a time course that parallels the sprouting of cholinergic fibers (Gomez-Pinilla et al., 1992). This suggests that FGF-2 may play an important role in lesion-induced plasticity. A different pattern of FGF-2 induction is seen upon transection of the septo-hippocampal pathway. Interruption of the fimbria-fornix induces FGF-2 immunoreactivity in astrocytes within the area of injury, but not in the hippocampus, suggesting that FGF-2 may be involved in mediating plasticity in this system. The distance between lesion and putative effector sites in the hippocampus per se probably does not account for the differences in FGF-2 induction in the entorhinal-hippocampal and septo-hippocampal lesion paradigm; Kostyk and collaborators (1994) have shown that a crush lesion to the optic nerve causes an increase of FGF-2 over a significant distance both in retinal photoreceptors and in astrocytes of the optic tract. These results exemplify both the local and long-range roles that FGF-2 may play in mediating central nervous system regeneration and plasticity. Riedel et al. (1995) have employed the wellestablished selective toxicity of gentamycin on sensory cells in the inner ear of early postnatal rats to evalute subsequent changes in the pattern of FGF-2 immunoreactivity in auditory brainstem nuclei. Gentamycin treatment abolished FGF-2 immunoreactivity in one part of the medial geniculate body, but not in other auditory nuclei, suggesting that FGF-2 may be a mediator of (or at least a marker for) the acquisition of function in a distinct auditory relay nucleus. Spinal cord injuries. Spinal cord lesions illustrate well the distinct patterns of regulation, and there-

Functions of fibroblast growth factors (FGFs) in the nervous system

fore the likely distinct functions of FGF-1 and FGF-2 in central nervous system lesions. Moreover, these studies demonstrate the importance of investigating both mRNA and protein levels after lesions (Koshinaga et al., 1993; Follesa et al., 1994). Following a standardized incomplete thoracic contusive spinal cord injury, RNase protection assays revealed a rapid 3-fold increase of FGF-2 mRNA by 6 h at the injury site. No effect was seen in cervical and lumbar segments of the spinal cord. mRNA levels remained significantly increased at 1 and 7 days, whereas FGF-1 mRNA levels were unchanged (Follesa et al., 1994). Using immunocytochemical staining for FGF-1 and FGF-2 after complete destruction of the dorsal columns at T8, Koshinga et al. (1993) noted an increase in FGF-1 staining in ventral motoneurons two days post-lesion and enhanced staining in the dorsal fiber tracts about four segments above and below the lesion site. FGF-1 immunoreactivity disappeared from the nucleus gracilis, suggesting that FGF-1 is anterogradely transported in ascending sensory fibers. This pattern of FGF-1 immunoreactivity was maintained for 12 days postlesion. An increase in FGF-2 immunoreactivity was not noted until 5 days post-lesion at the edge of the lesion cavity and in the dorsal colums at T4/5, and was confined to astroglial nuclei and cytoplasm. In toxically demyelinated spinal cords, Tourbah and collaborators (1992) observed a different temporal pattern of FGF-1 expression (by immuncytochemistry). Following an initial decrease in staining, they found a dramatic increase during the initial period of remyelination. This could indicate a role of FGF-1 in remyelination. 6.1.2. Peripheral nerve lesions In transected sciatic nerves, Eckenstein et al. (1991) found a massive and irreversible loss of FGF activity from the distal stump within 3 7 days, whereas activity in the proximal stump decreased by about 50% during the first week after lesion and recovered to normal levels during the following 40 days. FGF-1 and -2 were distinguished in this study by analyzing nerve homogenates in terms of mitogenic activity and its

S. Bieger and K. Unsicker

heparin dependence. The total disappearance of heparin-dependent activity from the distal stump suggested that FGF-1 might be anterogradely transported, which is consistent with its localization in sciatic nerve axons. Piehl and collaborators (1993) have supplemented this study, providing data on FGF-1 expression in motoneurons in response to ventral rhizotomy and peripheral nerve axotomy. They found only marginally elevated levels of FGF-1 mRNA expression in lesioned motoneurons. However, FGF-1 immunoreactivity completely disappeared in ventral root axons after ventral rhizotomy, but not after peripheral axotomy, raising the possibility that FGF-1 might be released from damaged axon stumps. 6.1,3. Cerebral ischaemia A growing body of literature reports alterations in FGF-1 and FGF-2 mRNA and protein expression in animal models of brain ischemia. Following a 20 min period of forebrain ischemia in rat, strong induction of FGF-2 immunoreactivity and mRNA can be observed in astrocytes within all hippocampal areas, most notably in the CAl subfield, from 48 h until at least 30 days post-lesion (Takami et al., 1992). This increase is accompanied by an early (24 h after ischemia) and persistent induction of FGFR-1 mRNA in CAl astroglial cells (Takami et al, 1993). FGFR-2 expression is only slightly altered. Other brain areas, such as caudate putamen, temporal cortex, and corpus callosum also exhibit increased FGF-2 immunoreactivity in astrocytes (Kiyota et al., 1991). Observations after unilateral occlusion of the middle cerebral artery suggest that neurons as well as astrocytes may display enhanced immunoreactivity for FGF-1 (Hara et al., 1994) and FGF-2 (Kumon et al., 1993) as well as elevated FGFR-1 mRNA (Sakaguchi et al, 1994). Increased FGF-2 immunoreactivity seems to be associated with persistent c-fos expression at the infarct periphery and in regions of 'selective vulnerability' beginning 3 h post-infarction and lasting up to 1-2 weeks (Liu and Chen, 1994a). The spatial pattern of changes in FGF-2 immunoreactivity is followed by a period of intense cell proliferation, perhaps indicating that cell proliferation due to the infarct may

357

be induced and sustained by FGF-2 (Liu and Chen, 1994b). 6.2. Modulation ofFGF expression by neurotransmittersy neurotoxins, and peptide growth factors While surgical trauma and infarction models convincingly demonstrate alterations in the expression of FGFs and FGFRs in response to mass lesion events, they cannot provide information on the underlying mechanisms. Studies employing specific chemical lesions or stimuli such as neurotransmitter receptor agonists and peptide growth factors have made it possible to elucidate mechanisms of FGF and FGFR regulation under pathophysiological conditions. 6.2.1. In vitro studies Glutamate is the most important excitatory transmitter in the brain and is a crucial mediator of the deleterious consequences of brain trauma and infarction. Treatment of rat cortical astrocytes with glutamate (100/^M, 15 min) rapidly induces FGF2 and FGFR-1 mRNAs, reaching a maximal induction at 4 h (Pechan et al., 1993). FGF-2 can be released from astrocytes (Ferrara et al., 1988; Araujo and Cotman, 1992a), thereby affecting not only astrocyte proliferation and phenotypic characteristics (see Section 5.2), but also the expression of its own mRNA, which is increased about 5-fold (Araujo and Cotman, 1992a; Flanders et al., 1993). FGF has also been reported to induce TGF-^1 in cultured astrocytes (Lindholm et al., 1992), which acts synergistically with FGF-2 to inhibit astroglial proliferation (Flanders et al., 1993). Thus FGF, in conjunction with TGF-^, seems to act as an important determinant of astroglial function, especially in the injury response of the central nervous system (see also Eddleston and Mucke, 1993; Logan and Berry, 1993). Other cytokines that are important in the response of the central nervous system to trauma are the interleukins IL-1^ and IL-6 and ligands to EGF receptors, such as EGF, which have been shown to upregulate FGF-2 in cultured astroglial cells (Araujo and Cotman, 1992a). As will be reviewed below, FGF-2 has been demon-

358

strated to be associated with amyloid plaques in Alzheimer's disease. 6.2,2. In vivo studies Experimental and clinical neurologists have an interest in pharmacologically manipulating neurotrophic proteins endogenous to the CNS, with the aim of exploiting their beneficial potential for neural repair processes. Several transmitters, most notably glutamate, noradrenaline and acetylcholine, have been shown to be involved in regulating the expression of neurotrophins (Lindholm et al., 1994a). The potential regulation of FGF expression by transmitters has not been so well investigated. Clenbuterol, a lipophilic )8-adrenergic receptor agonists, was shown to induce and increase both the NGF and the FGF-2 mRNA content in rat cerebral cortex within 5 h. FGF-2, but not NGF, was also induced in hippocampus and cerebellum. Surprisingly, isoproterenol, which does not cross the blood-brain barrier, also elicited increases in NGF and FGF-2 mRNA expression within the cerebral cortex, indicating that activation of peripheral ^-adrenergic receptors might also play a role in the regulation of FGF-2 mRNA expression (Follesa and Mocchetti, 1993). Seizures induced by the glutamate agonist kainate cause rapid and widespread increases in mRNA levels for FGF-1, FGF-2 and FGFR-1 (Riva et al., 1992, 1994; Rumpel et al., 1993; Bugra et al., 1994). Upregulation of FGF-2 mRNA and increased FGF-2 immunoreactivity can be observed in astroglial cells throughout the brain. With regard to neurons, increased expression of FGF-1 and FGF-2 in the hippocampus is limited to neuron populations that are resistant to seizureinduced injury. Increased FGFR-1 mRNA levels occur in pyramidal neurons of all subfields and in dentate gyrus granule cells. Focal stainless-steel wire lesions in the hilar region bilaterally raise FGF-2 mRNA levels, primarily in astrocytes in hippocampus, neocortex, olfactory cortex, amygdala, and septum (Gall et al, 1994). Lesions with 6-hydroxydopamine (6-OHDA) and l-methyl-4-phenyl-l,2,3,6-tetrahydropyridine (MPTP) drastically impair the catecholaminergic and especially dopaminergic transmitter systems.

Functions offibroblast growth factors (FGFs) in the nervous system

6-OHDA injected into the substantia nigra (SN) induces a strong and sustained elevation of FGF-2 mRNA in astrocytes within the ipsilateral SN pars compacta and ventral tegmental area, but only a transient induction in the SN pars reticulata and neostriatum as well as in the ipsi- and contralateral hippocampus and neocortex (Chadi et al., 1994). In the SN, increases in astroglial FGF-2 mRNA levels are accompanied by a drastic increase in FGF-2 immunoreactivity in astroglial nuclei. MPTP systemically applied to mice causes a significant increase in the amounts of striatal FGF1 and -2 mRNA, as determined by semiquantitative PCR or Northern blot (Leonard et al., 1993; Rufer, personal communication). This increase in striatal FGF-2 is accompanied by enhanced immunoreactivity for the astroglial intermediate filament protein, glial fibrillary acidic protein, but not by altered numbers of FGF-2 immunoreactive astroglial cells (Wirth, personal communication), indicating that FGF-2 can regulate the astroglial cell phenotype not only in vitro (see Section 5.2), but also in vivo. Local application of FGF-2, but not cytochrome c, to the MPTP-lesioned striatum, causes a rapid and sustained disappearance of FGF-2 immunoreactivity from astroglial nuclei. Whether this change in intracellular location of FGF-2 is causally related to the marked neuroprotective effects of FGF-2 for toxically impaired nigrostriatal neurons is unclear. Few studies have addressed the possible regulation of FGFs and FGFRs by cytokines in vivo. Intraventricularly injected JL-Xp induces significant and widespread increases in FGF-2 mRNA, but not FGF-1 mRNA (Rivera et al., 1994). Increased hybridization signals peak at 8 h post-injection in most areas and seem to be mostly associated with astroglial cells rather than with neurons. 6.3. Beneficial ejfects of exogenous FGF in lesions and animal models of neurodegenerative disorders 6.3.1. Central nervous system pathway lesions and target destruction Sievers and collaborators (1987) were the first to study the rescuing potential of FGF-1 and FGF-

S. Bieger and K. Unsicker

2 for axotomized central nervous system neurons. Following transection of the optic nerve in adult rats and implantation of FGF-soaked gelfoam to the transection site, they observed a significant improvement (3-fold for FGF-2, 4-fold for FGF-1) of retinal ganglion cell survival after 30 days. A similar survival promoting effect of FGF-2 on retinal ganglion cells has also been observed for retinal explants from adult rats (Bahr et al., 1989). In another classic central nervous system lesion paradigm, transection of the fimbria-fornix, intraventricular infusion or local administration of FGF-2 causes significant protection of cholinergic neurons in the medial septum (Anderson et al., 1988; Otto et al., 1989). In rats with partial fimbrial transections, FGF-2 injected intraventricularly at 10 ng every second day partially prevented the lesioninduced deficit in hippocampal choline acetyltransferase activity (Araujo et al., 1993). At 1 /^g, by contrast, FGF-2 was toxic, causing an increased death rate and marked reductions in body weight. In other experiments, FGF-2 was demonstrated to prevent the death of entorhinal cortex neurons after transection of the perforant path (Cummings et al., 1992). While these results underscore the ability of FGF to protect specific populations of central neurons from axotomy-induced death, there is also evidence that FGF can ameliorate the morphological and functional deficits of neurons after destruction of their central or peripheral nervous target area (Blottner et al., 1989, 1990; Figueiredo et al., 1993). FGF-1 was shown to prevent degeneration of nucleus basialis cholinergic neurons following cortical devascularization, and to reduce lesion-induced impairment in a spatial memory task (Figueiredo et al., 1993). Similarly, FGF-2 prevented the degeneration of preganglionic sympathetic spinal cord neurons that innervate the adrenal medulla after destruction of their target organ (Blottner et al., 1989, 1990). Whether FGF acts directly on the affected neuron populations (via axonal transport), or indirectly through the activation of other growth and neurotrophic factors has not been elucidated. 6.3.2. Peripheral nerve lesions and regeneration FGF-2, like NGF, has a marked promoting ef-

359

fect on the survival of rat dorsal root ganglionic neurons after transection of the sciatic nerve (Otto et al., 1987). Moreover, both FGF-1 and FGF-2 can enhance peripheral nerve regeneration even across long nerve gaps (Aebisher et al., 1989; Cordeiro et al., 1989; Walter et al., 1993; Vergara et al, 1993). 6.3.3. Ischaemic, excitotoxic, metabolic, and oxidative insults A rapidly growing body of literature indicates that several growth factors including FGF-1 and FGF-2 can reduce neuronal injury resulting from ischaemic, excitotoxic, metabolic, and oxidative insults (see Unsicker et al., 1992b; Mattson et al., 1993b, Mattson and Scheff, 1994). Initial studies demonstrated that FGF-2 can protect cultured rat hippocampal neurons from glutamate toxicity (Mattson et al, 1989; Mattson and Rychlik, 1990). FGF-2 also protects cultured striatal neurons against excitotoxicity (Freese et al., 1992) and is effective in counteracting glucose-deprivationinduced injury of rat hippocampal and human cortical neurons (Cheng and Mattson, 1991, 1992). In vivo, FGF-2 protects against ischemic injury and NMD A neurotoxicity (Nozaki et al., 1993) and reduces infarct size after focal cerebral ischemia in adult rats (Koketsu et al., 1994). Other in vivo targets of FGF-2, and of FGF-1, include rat and thalamic neurons, which are rescued after cerebral infarction (Yamada et al., 1991) and gerbil brain neurons after ischemic injury (Mitani et al., 1992; Sasaki et al., 1992; Nakata et al., 1993). Free radical-mediated injury can be attenuated by several growth factors. Iron catalyzes the formation of hydroxyl radical. Its damaging effect on neurons (hippocampal and cortical) can be inhibited by pretreatment with FGF-2 (Zhang et al., 1993). FGF-2 has been shown to protect mesencephalic dopaminergic neurons against the free radical-mediated toxicity of MPP+ and MPTP both in vitro (Otto and Unsicker, 1993a) and in vivo (Otto and Unsicker, 1990, 1993b; Chadi et al., 1993a). The effects of FGF-2 incude partial compensation of loss of transmitters, induction of axonal sprouting, and marked improvement of motor performances. In vitro, these effects appear to be

360

indirect (Engele and Bohn, 1991) and can be suppressed by elimination of astrocytes from the cultures. In vivo, there is also recent evidence that FGF-2 causes a local increase in reactive astroglial cells within the striatum ispilateral to the administration site (Wirth, personal communication), suggesting that astrocytes may mediate at least some of the beneficial actions of FGF-2 in this lesion paradigm. The effect on astrocytes includes suppression of nuclear FGF-2 immunoreactivity and an increase in immunoreactivity for the gap junction protein connexin 43 (Wirth, personal communication), raising the possibility that FGF-2 localized in astrocytes may be part of the cytokine cascade mediating the indirect actions of exogenous FGF-2. Although details on the cellular and molecular mechanisms underlying the protective effects of FGF-2 are only beginning to emerge, recent evidence suggests that FGF-2 can directly influence calcium homeostasis and free radical metabolism, two pivotal systems implicated in neuronal injury. Thus, FGF-2 can prevent or suppress the elevation of Cd?^ in cultured neurons induced by excitatory amino acids (Mattson et al., 1989a), glucose deprivation (Cheng and Mattson 1991), treatment with calcium ionophores, or treatment with )8-amyloid (Mattson et al., 1993a,d). A number of mechanisms have been proposed to be involved. For example, FGF-2 may enhance a calcium extrusion or buffering mechanism, possibly by acting on calcium-binding proteins such as calbindin, which can remove or sequester cytoplasmic free calcium (Collazo et al, 1992). FGF-2 may also influence glutamate receptors more directly. This hypothesis is supported by the observation that FGF-2 suppresses the expression of a 71 kDa NMD A receptor protein which mediates Ca^+ influx and excitotoxicity in cultured hippocampal neurons (Mattson et al., 1993c). The second messenger systems involved in the protective actions of FGF-2 are still under investigation (see Boniece and Wagner, 1993; Maiese et al., 1993; Mattson and Scheff, 1994).

Functions offibroblastgrowth factors (FGFs) in the nervous system

tions, in which photoreceptors degenerate and disappear, are major causes of blindness. FGF-1 and -2, like a large number of other growth factors (e.g. brain-derived neurotrophic factor, ciliary neurotrophic factor, neurotrophin-3, insulin-like growth factor II; LaVail et al., 1992), have a remarkable ability to rescue photoreceptors in a rat strain with inherited retinal dystrophy (Faktorovich et al., 1990, 1992) or following exposure to constant light (LaVail et al., 1992). 6.4, FGFs in human pathologies Several studies have reported an association with and changes in FGF in several human pathologies, including Parkinson's and Alzheimer's disease. Thus, FGF-2 immunoreactivity can be detected in most dopaminergic neurons of the substantia nigra in aged human subjects, but is only seen in approximately 10% of the surviving dopaminergic neurons in post mortem brains of Parkinsonian patients (Tooyama et al., 1994). Increases in FGF-1 and -2 immunoreactivities have been observed in neurons, glial, and vascular cells in the brains of Alzheimer patients (GomezPinilla et al., 1990; Tooyama et al., 1991; Cummings et al., 1993). FGF-2 has been shown to colocalize with)8-APP (Imaizumi et al., 1993) and, in fact, is associated with or binds to senile plaques and neurofibrillary tangles (Kato et al., 1991; Siedlak et al., 1991; Cummings et al., 1993), raising the possibility that it may attract neurites into plaques. Further evidence for an association of FGF-2 with Alzheimer's disease comes from the observation that in vitro, FGF-2 can induce )8-APP mRNA in glial cell lines (Quon et al., 1990) as well as in astrocytes (Gray and Patel, 1993) and that )3-amyloid induces FGF-2 expression in cultured astrocytes (Araujo and Cotman, 1992b). Alterations in FGF-1 and FGF-2 staining patterns have also been demonstrated in the striatum and midbrain in Huntington's disease (Tooyama et al., 1993). 7. Therapeutic applications of FGFs

6.3A. Protective roles of FGF for photoreceptors Hereditary and lesion-induced retinal degenera-

Some of the most potent effects of FGF-1 and -2

S, Bieger and K. Unsicker

are on cells of the vasculature and epithelial cells, and these targets may be the most promising for therapeutic applications of FGFs. Both in vitro and in vivo experiments have demonstrated that topical application of FGFs can enhance processes of angiogenesis and wound healing (Davidson and Broadley, 1991; Fiddes et al., 1991). Treatment of the brain with growth factors is, of course, more difficult. Systemic treatments are impeded by the blood-brain-barrier, as well as by the rapid clearance of intravenously administered FGF (Hondermarck et al., 1990). Infusion of FGF-2 into the ventricles has been reported to be inefficient, as the growth factor is rapidly cleared or degraded (Gonzalez et al., 1991a, 1994). The most effective means of delivering FGF to the targeted neurons is apparently by local application, for instance by capsular implant. However, since FGFs have a high affinity for cell surface proteoglycans, the diffusional radius from the application site is relatively low. This can be ameliorated by coapplication of FGF with heparin or heparin sulfate, which not only extends the effective range of exogenous FGF (Chadi et al., 1993a) but also stabilizes them against oxidation, denaturation (Belford et al., 1993; Volkin et al, 1993a,b), and proteolysis (Damon et al., 1989; Sommer and Rifkin, 1989), thus enhancing the biological half-life and even potentiating activity (Gomez-Pinilla et al., 1995). Protection of FGFs from degradation is certainly an important consideration in any in vivo or long-term applications of growth factor. Attempts to improve the stability of FGF-1 and -2 include the design of altered FGFs that are less susceptible to oxidation or degradation (e.g. Seno et al., 1988; Ortega et al., 1991; Arakawa et al., 1993), the design of microspheres that allow sustained release of FGF (Edelman et al., 1991), and the optimization of carriers in which FGFs are applied (Tardieu et al., 1992; Tsai et al., 1993). A number of other synthetic polyanions that bind to the heparin-binding site of FGF (more appropriately perhaps termed the polyanion-binding site) serve not to protect FGF, but rather to antagonize its activity. The most familiar compound of this type is probably suramin, which prevents

361

interaction of FGFs, as well as a number of other growth factors, with their receptors. Suramin apparently acts both by inducing a conformational change in the FGF molecule and by inducing its aggregation (Middaugh et al., 1992). Many suramin relatives (Braddock et al., 1994) and other synthetic polyanions including modified heparins (Soulet et al., 1994) and even phosphorthioate oligodeoxynucleotides (Guvakova et al., 1995), as well as natural sulfated polysaccharides such as bacterial D-gluco-D-galactan sulfate (Nakayama et al., 1993) exert similar antagonistic effects with greater or lesser specificity. Potentially, such modulators of FGF activity could prove clinically useful, for example in treating solid tumors that depend on FGF. Although several reports have demonstrated the feasability of such treatments outside the nervous system, for example in the inhibition of angiogenesis and tumour growth (Taylor and Folkman, 1982; Wellstein et al., 1991), an initial attempt to use suramin systemically to treat gliosarcoma has not been so promising: not only does suramin show poor penetration into the brain, but it is also toxic (Olson et al., 1994). These results suggest that effective pharmacological manipulations of endogenous FGF activity in the brain are probably still a long way off. More promising results, perhaps, have been obtained with a quite different approach: conjugation of FGF to saporin, an intracellularly acting toxin (Lappi et al., 1989). This FGF-saporin 'mitotoxin', which is selectively taken up by cells expressing the FGF receptor, has been shown to exert specific effects when injected into the hippocampus (Gonzalez et al., 1991b). Although the application of exogenous FGFs to the brain, for example after chemical and physical lesions (see Section 6), has been shown to have positive effects on neuron survival or regeneration, it is unclear whether therapy with growth factors is a realistic approach in vivo, especially in light of the difficulty in targeting the factor to specific cells and in ensuring long-term deliverance. Improvements in neural transplantation techniques raise the possibility that implantation of growth factorsecreting cells may be a feasible therapeutic strategy at some time in the future.

362

References Abe, K. and Saito, H. (1992) Selective enhancement by basic fibroblast growth factor of NMDA receptor-mediated increase of intracellular Ca2+ concentration in hippocampal neurons. Brain Res. 595: 128-132. Abraham, J.A., Whang , J.L., Tumolo, A., and Fiddes, J.C. (1986) Human basic fibroblast growth factor: nucleotide sequence, genomic organization, and expression in mammalian cells. Cold Spring Harbor Symp. Quant. Biol. 51: 657-668. Acland, P., Dixon, M., Peters, G. and Dickson, C. (1990) Subcellular fate of the int-2 oncoprotein is determined by choice of initiation codon. Nature 343: 662-665. Aebischer, P., Salessiotis, A.N. and Winn, S.R. (1989) Basic fibroblast growth factor released from synthetic guidance channels facilitates peripheral nerve regeneration across long nerve gaps. J. Neurosci. Res. 23: 282-289. Ago, H., Kitagawa, Y., Fujishima, A., Matsuura, Y. and Katsube, Y. (1991) Crystal structure of basic fibroblast growth factor at 1.6 A resolution. J. Biochem. 110: 360-363. Altin, J.G. and Bradshaw, R.A. (1990) Production of 1,2diacylglycerol in PC 12 cells by nerve growth factor and basic fibroblast growth factor. J. Neurochem. 54: 16661676. Altin, J.G., Wetts, R. and Bradshaw, R.A. (1991) Microinjection of a p21ras antibody into PC 12 cells inhibits neurite outgrowth induced by nerve growth factor and basic fibroblast growth factor. Growth Factors 4: 145-155. Amalric, F., Baldin, V., Bosc-Bierne, I., Bugler, B., Couderc, B., Guyader, M., Patry, V., Prats, H., Roman, A.M. and Bouche, G. (1991) Nuclear translocation of basic fibroblast growth factor. Ann. N. Y. Acad. Sci. 638: 127-138. Amalric, F., Bouche, G., Bonnet, H., Brethenou, P., Roman, A.M., Truchet, I. and Quarto, N. (1994) FGF-2 in the nucleus: translocation process and targets. Biochem. Pharmacol. 41: 111-115. Anderson, K.J., Dam, D., Lee, S. and Cotman, C.W. (1988).Basic fibroblast growth factor prevents death of lesioned cholinergic neurons in vivo. Nature 332, 360-361. Aoyagi, A., Nishikawa, K., Saito, H. and Abe, K. (1994) Characterization of basic fibroblast growth factor-mediated acceleration of axonal branching in cultured rat hippocampal neurons. Brain Res. 661: 117-126. Arakawa, T., Hovan, T.P., Narhi, L.O., Rees, D.C., Schiffer, S.G., Hoist, P.L., Prestrelski, S.J., Tsai, L.B and Fox, G.M. (1993) Production and characterization of an analog of acidic fibroblast growth factor with enhanced stability and biological activity. Protein Eng. 6: 541-546. Araujo, D.M., Lapchak, P.A. and Hefti, F. (1993) Effects of chronic basic fibroblast growth factor administration to rats with partial fimbrial transections on presynaptic cholinergic parameters and muscarinic receptors in the hippocampus Comparison with nerve growth factor. J. Neurochem. 61, 899-910. Araujo, D.M. and Cotman, C.W. (1992a) Basic FGF in as-

Functions of fibroblast growth factors (FGFs) in the nervous system troglial, microglial and neuronal cultures: characterization of binding sites and modulation of release by lymphokines and trophic factors. J. Neurosci. 12: 1668-1678. Araujo, D.M. and Cotman, C.W. (1992b) Beta-amyloid stimulates glial cells in vitro to produce growth factors that accumulate in senile plaques in Alzheimer's disease. Brain Res. 569: 141-145. Asai, T., Wanaka, A., Kato, H., Masana, Y., Seo, M. and Tohyama, M. (1993) Differential expression of two members of FGF receptor gene family, FGFR-1 and FGFR-2 mRNA, in the adult rat central nervous system. Mol. Brain Res. 17: 174-178. Bahr, M., Vanselow, J. and Thanos, S. (1989) Ability of adult rat ganglion cells to regrow axons in vitro can be influenced by fibroblast growth factor and gangliosides. Neurosci. Lett. 96, 197-201. Baird, A., Esch, F., Gospodarowicz, D. and Guillemin, R. (1985) Retina- and eye-derived endothelial cell growth factors: partial molecular characterization and identity with acidic and basic fibroblast growth factors. Biochemistry 24: 7855-7860. Baird, A. (1994a) Fibroblast growth factors: activities and significance of non-neurotrophin neurotrophic activities. Curr. Opin. Neurobiol. 4: 78-86. Baird, A. (1994b) Potential mechanisms regulating the extracellular activities of basic fibroblast growth factor (FGF-2). Mol. Reprod. Dev. 39: 4 3 ^ 8 . Baird, A. and Bohlen, P. (1990) Fibroblast growth factors. In: M.B. Sporn and A.B. Roberts (Eds.), Peptide Growth Factors and Their Receptors, Handbook Experimental Pharmacology 95 (I)y Springer Verlag, Heidelberg, pp. 369-418. Baird, A., Schubert, D., Ling, N. and Guillemin, R. (1988) Receptor- and heparin-binding domains of basic fibroblast growth factor. Proc. Natl. Acad. Sci. USA 85: 2324-2328. Balaci, L., Presta, M., Ennas, M.G., DelliEra, P., Sogos, V., Lauro, G. and Gremo, F. (1994) Differential expression of fibroblast growth factor receptors by human neurones, astrocytes and microglia. NeuroReport 6, 197-200. Baldin, V., Roman, A.-M., Bosc-Bierne, I., Amalric, F. and Bouche, G. (1990) Translocation of bFGF to the nucleus is Gl phase cell cycle specific in bovine aortic endothelial cells. EM^O 7.9:1511-1517. Bean, A.J., Elde, R., Cao, Y., Oellig, C , Tamminga, C , Goldstein, M., Petterson, R.F. and Hokfelt, T. (1991) Expression of acidic and basic fibroblast growth factors in the substantia nigra of rat, monkey and human. Proc. Natl. Acad Sci. USA 88: 10237-10241. Belford, D.A., Hendry, LA. and Parish, C.R. (1993) Investigation of the ability of several naturally occurring and synthetic polyanions to bind to and potentiate the biological activity of acidic fibroblast growth factor. J. Cell Physiol. 157:184-189. Besnard, F., Perraud, F., Sensenbrenner, M. and Labourdette, G. (1989) Effects of acidic and basic fibroblast growth factors on proliferation and maturation of cultured rat oligodendrocytes. Int. J. Dev. Neurosci. 7: 401-409.

363

S. Bieger and K. Unsicker Bieger, S.C., Henkel, A.W. and Unsicker, K. (1995) Localization of basic fibroblast growth factor in bovine adrenal chromaffin cells. J. Neurochem. 64, 1521-1527. Birren, S.J. and Anderson, D. (1990) A v-myc-immortalized sympathoadrenal progenitor cell line in which neuronal differentiation is initiated by FGF but not NGF. Neuron 4: 189-201. Blottner, D. and Unsicker, K. (1990) Maintenance of intermediolateral spinal cord neurons by fibroblast growth factor administered to the medullectomized rat adrenal gland: dependence on intact adrenal innervation and cellular organization of implants. Eur. J. Neurosci. 2: 378-382. Blottner, D., Westermann, R., Grothe, C , Bohlen, P. and Unsicker, K. (1989) Basic fibroblast growth factor in the adrenal gland. Eur. J. Neurosci. 1: 471-478. Boniece, I.R. and Wagner, J.A. (1993) Growth factors protect PC 12 cells against ischemia by a mechanism that is independent of PKA, PKC, and protein synthesis. /. Neurosci. 13:4220-4228. Borja, A.Z., Meijers, C. and Zeller, R. (1993) Expression of alternatively spliced bPGF first coding exons and antisense mRNAs during chicken embryogenesis. Dev. Biol. 157: 110-118. Bost, L.M., Aotaki-Keen, A.E. and Hjelmeland, L.M. (1992) Coexpression of FGF-5 and bFGF by the retinal pigment epithelium in vitro. Exp. Eye Res. 55:121-134. Bouche, G., Gas, N., Prats, H., Baldin, V., Tauber, J.-P., Teissie, J. and Amalric, F. (1987) Basic fibroblast growth factor enters the nucleolus and stimulates the transcription of ribosomal genes in ABAE cells undergoing GO —> Gl transition. Proc. Natl. Acad. Sci. USA 84: 6770-6774. Braddock, P.S., Hu, D.-E., Fan, T.-P.D., Stratford, I.J., Harris, A.L and Bicknell, R. (1994) A structure-activity analysis of antagonism of the growth factor and angiogenic activity of basic fibroblast growth factor by suramin and related polyanions. Br. J. Cancer 69: 890-898. Briistle, O., Aguzzi, A., Talarico, D., Basilico, C , Kleihues, P. and Wiestler, O.D. (1992) Angiogenic activity of the Kfgf/hst oncogene in neural transplants. Oncogene 1: 11771183. Bugler, B., Amalric, F. and Prats, H. (1991) Alternative initiation of translation determines cytoplasmic or nuclear localization of bFGF. Mol. Cell. Biol. 11: 573-577. Bugra, K., Oliver, L., Jacquemin, E., Laurent, M., Courtois, Y. and Hicks, D. (1993) Acidic fibroblast growth factor is expressed abundantly by photoreceptors within the developing and mature rat retina. Eur. J. Neurosci. 5: 1586-1595. Bugra, K., Pollard, H., Charton, G., Moreau, J., Benari, Y. and Khrestchatisky, M. (1994) aFGF, bFGF and fig mRNAs show distinct patterns of induction in the hippocampus following kainate-induced seizures. Eur. J. Neurosci. 6: 5 8 66. Cao, Y., Ekstrom, M., and Pettersson, R.F. (1993) Characterization of the nuclear translocation of acidic fibroblast growth factor. J. Cell Sci. 104: 77-87. Carpenter, E.M., Goddard, J.M., Chisalea, O., Manley, N.R.

and Capecchi M.R. (1993) Loss of Hox-al (Hox-1.6) function results in the reorganization of the murine hindbrain. Development 118: 1064-1075. Caruelle, D., Groux-Muscatelli, B., Gaudric, A., Sestier, C , Coscas, G., Caruelle, J.P. and Barritault, D. (1989) Immunological study of acidic fibroblast growth factor (aFGF) distribution in the eye. J. Cell Biochem. 39: 117-128. Chadi, G., Cao, Y., Pettersson, R.F. and Fuxe, K. (1994) Temporal and spatial increase of astroglial basic fibroblast growth factor synthesis after 6-hydroxydopamine-induced degeneration of the nigrostriatal dopamine neurons. Neuro-

science6\'.%9\-9\0. Chadi, G., Moller, A., Rosen, L., Janson, A.M., Agnati, L.A., Goldstein, M., Ogren, S.O., Pettersson, R.F. and Fuxe, K. (1993) Protective actions of human recombinant basic fibroblast growth factor on MPTP-lesioned nigrostriatal dopamine neurons after intraventricular Infusion. Exp. Brain Res. 97: 145-158. Chadi, G., Tinner, B., Agnati, L.F. and Fuxe, K. (1993) Basic fibroblast growth factor (bFGF, FGF-2) immunoreactivity exists in the noradrenaline, adrenaline and 5-HT nerve cells of the rat brain. A^^Mro^d. Lett. 160: 171-176. Chellaiah, A.T., McEwen, D.G. , Werner, S., Xu, J. and Ornitz, D.M. (1994) FGFR3 alternative splicing in immunoglobulin-like domain III creates a receptor highly specific for aFGF/FGF-1. J. Biol. Chem. 269: 11620-11627. Cheng, B. and Mattson, M.P. (1991) NGF and bFGF protect rat hippocampal and human cortical neurons against hypoglycemic damage by stabilizing calcium homeostasis. Neuronl: 1031-1041. Cheng, B. and Mattson, M.P. (1992) Glucose deprivation elicits neurofibrillary tangle-like antigenic changes in hippocampal neurons: prevention by NGF and bFGF. Exp. Neurol. 117: 114-123. Cheon, H.G., LaRochelle, W.J., Bottaro, D.P., Burgess, W.H. and Aaronson, S.A. (1994) High affinity binding sites for related FGF ligands reside within different receptor Ig-like domains. Proc. Natl. Acad. Sci. 91: 989-993. Chu, C.H. and Tolkovsky, A.M. (1994) Alternative adrenal chromaffin cell fates induced by basic fibroblast growth factor or cyclic AMP in vitro depend on a collaboration with the growth substrate. Neuroscience 59: 43-54. Clements, D.A., Wang, J.-K., Dionne, C.A. and Goldfarb, M. (1993) Activation of fibroblast growth factor (FGF) receptors by recombinant human FGF-5. Oncogene 8: 13111316. Collarini, E.J., Kuhn, R., Marshall, C.J., Monuki, E.S., Lemke, G. and Richardson, W.D. (1992) Down-regulation of the POU transcription factor SCIP is an early event in oligodendrocyte differentiat. Development 116: 193-200. Collazo, D., Takahashi, H. and McKay, R.D. (1992) Cellular targets and trophic functions of neurotrophin-3 in the developing rat hippocampus. Neuron 9: 643-656. Cook, P.W., Ashton, N.M., Karkaria, C.E., Siess, D.C. and Shipley, G.D. (1995) Differential effects of a heparin antagonist (hexadimethrine) or chlorate on amphiregulin,

364 bFGF, and heparin-binding EGF-like growth factor activity. J. Cell Physiol. 163:418-429. Cordeiro, P.G., Seckel, B.R., Lipton, S.A., D'Amore, P.A., Wagner, J. and Madison, R. (1989) Acidic fibroblast growth factor enhances peripheral nerve regeneration in vivo. Surgery 83: 1013-1019. Coulier, P., de Lapeyri^re, O. and Birnbaum, D. (1994) ComplexitE de la famille de facteurs de croissance FGF: la preuve par 9. Medicine/Sciences 9: 1113-1115. Coulier, P., Ollendorf, V., Maries, I., Rosnet, O., Batoz, M., Planche, J., Marchetto, S., Pebusque, M.-J., de Lapyri^re, O. and Birnbaum, D. (1991) The FGF6 gene within the FGF multigene family. Ann. N. Y. Acad. Sci. 638: 53-61. Coulier, P., Pizette, S., Batoz, M. and Birnbaum, D. (1992) Les recepteurs des facteurs de croissance des fibroblastes: une famille qui s'aggrandit. Medicine/Sciences 8: 811818. Courty, J., Loret, C, Moenner, M., Chevallier, B., Lagente, O., Courtois, Y. and Barritault, D. (1985) Bovine retina contains three growth factor activities with different affinity to heparin: eye-derived growth factor I, II, III. Biochimie 67: 265-269. Crossley, P.H. and Martin, G.R. (1995) The mouse Fgf8 gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patterning in the developing embryo. Development 121: 439-451. Cuevas, P., Giminez-Gallego, G., Carceller, P., Cuevas, B. and Crespo, A. (1993) Single topical application of human recombinant basic fibroblast growth factor (rbPGF) promotes neovascularization in rat cerebral cortex. Surg. Neurol. 39: 380-384. Cummings, B.J., Su, J.H. and Cotman, C.W. (1993) Neuritic involvement within bPGF immunopositive plaques of Alzheimer's disease. Exp. Neurol. 124: 315-325. Cummings, B.J., Yee, G.J. and Cotman, C.W. (1992) bPGF promotes the survival of entorhinal layer II neurons after perforant path axotomy. Brain Res. 591: 271-276. Dai, Z. and Peng, H.B. (1992) The influence of basic fibroblast growth factor on acetylcholine receptors in cultured muscle cells. Neurosci. Lett. 144: 14-18. Damon, D.H., Lobb, R.R., D'Amore, P.A. and Wagner, J.A. (1989) Heparin potentiates the action of acidic fibroblast growth factor by prolonging its biological half-life. J. Cell Physiol. 138: 221-226. Davis, J.B. and Stroobant, P. (1990) Platelet-derived growth factors and fibroblast growth factors are mitogens for rat Schwann cells, y. Cell Biol. 110: 1353-1360. Davison, J.M. and Broadley, K.N. (1991) Manipulation of the wound-healing process with basic fibroblast growth factor. Ann. N. Y. Acad. Sci. 638: 306-315. de Sainte Hilaire, Z. and Nicolaidis, S. (1992) Enhancement of slow wave sleep parallel to the satiating effect of acidic fibroblast growth factor in rats. Brain Res. Bull. 29: 525528. Delaunoy, J.P., Langui, D., Ghandour, S., Labourdette, G. and Sensenbrenner, M. (1988) Influence of basic fibroblast

Functions offibroblastgrowth factors (FGFs) in the nervous system growth on carbonic anhydrase expression by rat glial cells in primary culture. Int. J. Dev. Neurosci. 6: 129-136. Deloulme, J.C, Baudier, J. and Sensenbrenner, M. (1991) Establishment of pure neuronal cultures from fetal rat spinal cord and proliferation of the neuronal precursor cells in the presence of fibroblast growth factor. J. Neurosci. Res. 29: 499-509. Depres, G., Jalenques, I. and Romand, R. (1991) Basic fibroblast growth factor in the lower stato-acoustic system of the neonatal and adult rat. NeuroReport 2: 639-642. Dionne, C.A., Jaye, M. and Schlessinger, J. (1991) Structural diversity and binding of FGF receptors. Ann. N. Y. Acad. Sci. 638: 161-166. Doble, B.W. and Kardami, E. (1995) Basic fibroblast growth factor stimulates connexin-43 expression and intercellular communication of cardiac fibroblasts. Mol. Cell. Biochem. 143: 81-87. Dreyer, D., Lagrange, A., Grothe, C, and Unsicker, K. (1989) Basic fibroblast growth factor prevents ontogenetic neuron death in vivo. Neurosci. Lett. 99: 35-38. Du, X., Stull, N.D. and lacovitti, L. (1994) Novel expression of the tyrosine hydroxylase gene requires both acidic fibroblast growth factor and an activator. J. Neursci. 14: 7688-7694. Eccleston, P.A. and Silberberg, D.H. (1985) Fibroblast growth factor is a mitogen for oligodendrocytes in vitro. Dev. Brain /?^5. 21:315-318. Eckenstein, P.P., Shipley, G.D. and Nishi, R. (1991) Acidic and basic fibroblast growth factors in the nervous system: distribution and differential alteration of levels after injury of central versus peripheral nerve. J. Neurosci. 11: 412419. Eckenstein, P.P. (1994) Fibroblast growth factors in the nervous system. /. Neurobiol. 25: 1467-1480. Eckenstein, P.P., Kuzis, K., Nishi, R., Woodward, W.R., Meshul, C, Sherman, L. and Ciment, G. (1994) Cellular distribution, subcellular localization and possible functions of basic and acidic fibroblast growth factor. Biochem. Pharmacol. Al\ 103-110. Eddleston, M. and Mucke, L. (1993) Molecular profile of reactive astrocytes-implications for their role in neurologic disease. Neuroscience 54: 15-36. Edelman, E.R., Mathiowitz, E., Langer, R. and Klagsbrun, M. (1991) Controlled and modulated release of basic fibroblast growth factor. Biomaterials 12: 619-626. El-Husseini, A.E., Paterson, J.A. and Shiu, R.P.C. (1994) Basic fibroblast growth factor (bPGF) and two of its receptors, FGFRl and FGFR2: gene expression in the rat brain during postnatal development as determined by quantitative RT-PCR.Mo/. Cell. Endocrinol 104: 191-200. Elde, R., Cao, Y., Cintra, A., Brelje, T.C., Pelto, H.M., Junttila, T., Fuxe, K., Pettersson, R.F. and Hokfelt, T. (1991) Prominent expression of acidic fibroblast growth factor in motor and sensory neurons. Neuron 7: 349-364. Emoto, N., Gonzalez, A.-M., Walicke, P.A., Wada, E., Simmons, D.M., Shimasaki, S. and Baird, A. (1989) Basic fi-

S. Bieger and K. Unsicker broblast growth factor (FGF) in the central nervous system: Identification of specific loci of basic FGF expression in the rat brain. Growth Factors 2: 21-29. Engele, J. and Bohn, M.C. (1991) The neurotrophic effects of fibroblast growth factors on dopaminergic neurons in vitro are mediated by mesencephalic glia. J. Neurosci. 11: 3070-3078. Engele, J. and Bohn, M.C. (1992) Effects of acidic and basic fibroblast growth factors (aFGF, bFGF) on glial precursor cell proliferation: Age dependency and brain region specificity. Dev. Biol 152: 363-372. Eriksson, A.E., Cousens, L.S. and Matthews, B.W. (1993) Refinement of the structure of human basic fibroblast growth factor at 1.6A resolution and analysis of presumed heparin binding sites by selenate substitution. Protein Sci. 2: 1274-1284. Eriksson, A.E., Cousens, L.S., Weaver, L.H. and Matthews, B.W. (1991) Three-dimensional structure of human basic fibroblast growth factor. Proc. Natl Acad. ScL USA 88: 3441-3445. Ernfors, P., Lonnerberg, P., Ayer-Lelievre, C. and Persson, H. (1990) Developmental and regional expression of basic fibroblast growth factor mRNA in the rat central nervous system. J. Neurosci Res. 27: 10-15. Faktorovich, E.G., Steinberg, R.H., Yasumura, D., Matthes, M.T. and Lavail, M.M. (1992) Basic fibroblast growth factor and local injury protect photoreceptors from light damage in the rat. J. Neurosci 12: 3554-3567. Faktorovich, E.G., Steinberg, R.H., Yasumura, D., Matthes, M.T. and Lavail, M.M. (1990) Photoreceptor degeneration in inherited retinal dystrophy delayed by basic fibroblast growth factor. Nature 347: 83-86. Fantl, W.J., Johnson, D.E. and Williams, L.T. (1993) Signalling by receptor tyrosine kinases. Annu. Rev. Biochem. 62: 453^81. Ferguson, I.A., Schweitzer, J.B. and Johnson, Jr., E.M. (1990) Basic fibroblast growth factor: receptor-mediated internalization, metabolism, and anterograde axonal transport in retinal ganglion cells. /. Neurosci 10: 2176-2189. Ferguson, I.A. and Johnson, Jr., E.M. (1991) Fibroblast growth factor receptor-bearing neurons in the CNS: identification by receptor-mediated retrograde transport. /. Comp. Neurol 313: 693-706. Ferhat, L., Khrestchatisky, M., Roisin, M.-P. and Barbin, G. (1993) Basic fibroblast growth factor-induced increase in zif/268 and c-fos mRNA levels is Ca2+ dependent in primary cultures of hippocampal neurons. J. Neurochem. 61: 1105-1112. Ferrara, N., Ousley, F. and Gospodarowicz, D. (1988) Bovine brain astrocytes express basic fibroblast growth factor, a neurotropic and angiogenic mitogen. Brain Res. 462: 223-232 Ferrari, G., Minotti, M.-C, Toffano, G., Leon, A. and Skaper, S.D. (1989) Basic fibroblast growth factor promotes the survival and development of mesencephalic neurons in culture. Dev. Biol 133: 140-147. Ferrari, G., Toffano, G. and Skaper, S.D. (1991) Epidermal growth factor exerts neuronotrophic effects on dopaminer-

365 gic and GABAergic CNS neurons: comparison with basic fibroblast growth factor. J. Neurosci Res. 30: 4 9 3 ^ 9 7 . Fiddes, J.C, Hebda, P.A., Hayward, P., Robson, M.C, Abraham, J.A. and Klingbeil, C.K. (1991) Preclinical woundhealing studies with recombinant human basic fibroblast growth factor. Ann. N. Y. Acad. Scl 638: 316-328. Figueiredo, B.C., Piccardo, P., Maysinger, D., Clarke, P.B.S. and Cuello, A.C. (1993) Effects of acidic fibroblast growth factor on cholinergic neurons of nucleus basalis magnocellularis and in a spatial memory task following cortical devascularization. Neuroscience 56: 955-963. Finch, P.W., Rubin, J.S., Miki, T., Ron, D. and Aaronson, S.A. (1989) Human KGF is FGF-related with properties of a paracrine effector of cell growth. Science 245: 752-755. Finklestein, S.P., Apostolides, P.J., Caday, C.G., Prosser, J., Philips, M.F. and Klagsbrun, M. (1988) Increased basic fibroblast growth factor (bFGF) immunoreactivity at the site of focal brain wounds. Brain Res. 460: 253-259. Flanders, K.C., Ludecke, G., Renzing, J., Hamm, C , Cissel, D.S. and Unsicker, K. (1993) Effects of TGF-betas and bFGF on astroglial cell growth and gene expression in vitro. Mol Cell Neurosci 4: 406-417. Flaumenhaft, R., Abe, M., Mignatti, P. and Rifkin, D.B. (1992) Basic fibroblast-induced activation of transforming growth factor b in endothelial cells: regulation of plasminogen activator activity. J. Cell Biol 118: 901-909. Florkiewicz, R.Z. and Sommer, A. (1989) Human basic fibroblast growth factor gene encodes four polypeptides: three initiate translation from non-AUG codons. Proc. Natl Acad Scl USA 86: 3978-3981. Florkiewicz., R.Z., Majack, R.A., Buechler, R.D. and Florkiewicz, E. (1995) Quantitative export of FGF-2 occurs through an alternative, energy-dependent, non-ER/Golgi pathway. J. Cell Physiol 162: 388-399. Follesa, P. and Mocchetti, I. (1993) Regulation of basic fibroblast growth factor and nerve growth factor messenger RNA by beta-adrenergic receptor activation and adrenal steroids in rat central nervous system. Mol Pharmacol 43: 132-138. Follesa, P., Wrathall, J.R. and Mocchetti, I. (1994) Increased basic fibroblast growth factor mRNA following contusive spinal cord injury. Mol Brain Res. 22: 1-8. Ford, M.D., Bartlett, P.F. and Nurcombe, V. (1994) Colocalization of FGF-2 and a novel heparan sulphate proteoglycan in embryonic mouse brain. NeuroReport 5: 5 6 5 568. Frautschy, S.A., Walicke, P.A. and Baird, A. (1991) Localization of basic fibroblast growth factor and its mRNA after CNS injury. Brain Res. 553: 291-299. Freese, A., Finklestein, S.P. and DiFiglia, M. (1992) Basic fibroblast growth factor protects striatal neurons in vitro from NMDA receptor mediated neurotoxicity. Brain Res. 575: 351-355. Fressinaud, C , Laeng, P., Labourdette, G., Durand, J. and Vallat, J.-M. (1993) The proliferation of mature oligodendrocytes in vitro is stimulated by basic fibroblast growth

366 factor and inhibited by oligodendrocyte-type 2 astrocyte precursors. Dev. Biol 158: 317-329. Fressinaud, C, Vallat, J.M. and Labourdette, G. (1995) Basic fibroblast growth factor down-regulates myelin basic protein expression and alters myelin compaction of mature oligodendrocytes in vitro. J. Neurosci. Res. 40: 285293. Fuller-Pace, F., Peters, G. and Dickson, C. (1991) Cell transformation by kFGF requires secretion but not glycosylation. J. Cell Biol 115:547-555. Gall, CM., Berschauer, R. and Isackson, P.J. (1994) Seizures increase basic fibroblast growth factor mRNA in adult rat forebrain neurons and glia. Mol Brain Res. 21: 190-205. Gallo, v., Wright, P. and McKinnon, R.D. (1994) Expression and regulation of a glutamate receptor subunit by bFGF in oligodendrocyte precursors, Glia 10: 149-153. Gannoun-Zaki, L., Fieri, I., Badet, J. and Barritault, D. (1994) Interaction of basic fibroblast growth factor (FGF-2) with nonresponsive HeLa cells. Exp. Cell Res. 213: 375-382. Gao, H. and Hollyfield, J.G. (1992) Basic fibroblast growth factor (bFGF) inmiunolocalization in the rodent outer retina demonstrated with an anti-rodent bFGF antibody. Brain Res. 585: 355-360. Gensburger, C, Capo, M., Deloulme, J.C. and Sensenbrenner, M. (1992) Influence of basic fibroblast growth factor on the development of cholinoceptive neurons from fetal rat cerebrum in culture. Dev. Neurosci 14: 278-281. Gensburger, C, Labourdette, G., and Sensenbrenner, M. (1987) Brain basic fibroblast growth factor stimulates the proliferation of rat neuronal precursor cells in vitro. FEBS Lett. 217: 1-5. Gerdes, W., Brysch, W., Schlingensiepen, K.-H. and Seifert, W. (1992) Antisense bFGF ohgodeoxynucleotides inhibit DNA synthesis of astrocytes. NeuroReport 3: 43^6. Goldfarb, M., Bates, B., Drucker, B., Hardin, J. and Haub, O. (1991) Expression and possible functions of the FGF-5 gene. Ann. N. Y. Acad. Scl 638: 38-52. Gomez-Pinilla, F., Cummings, B.J. and Cotman, C.W. (1990) Induction of basic fibroblast growth factor in Alzheimer's disease pathology. NeuroReport 1: 211-214. Gomez-Pinilla, F., Lee, J.W. and Cotman, C.W. (1992) Basic FGF in adult rat brain: cellular distribution and response to entorhinal lesion and fimbria-fornix transection. J. Neurosci 12: 345-355. Gomez-Pinilla, F. and Cotman, C.W. (1993) Distribution of fibroblast growth factor-5 messenger RNA in the rat brain An in situ hybridization study. Brain Res. 606: 79-86. Gomez-Pinilla, F., Vu, L. and Cotman, C.W. (1995) Regulation of astrocyte proliferation by FGF-2 and heparan sulfate in vivo. J. Neurosci 15: 2021-2029. Gonzalez, A.-M., Buscaglia, M. , Ong, M. and Baird, A. (1990) Distribution of basic fibroblast growth factor in the 18-day rat fetus: localization in the basement membranes of diverse tissues. J. Cell Biol 110: 753-765. Gonzalez, A.-M., Buscaglia, M.L., Fuller, J., Dahl, R., Carman, L.S. and Baird, A. (1991a) Local fate and distribution

Functions of fibroblast growth factors (FGFs) in the nervous system of locally infused basic FGF. Ann. N Y. Acad. Scl 416419. Gonzalez, A.-M., Lappi, D.A., Buscaglia, M.L., Carman, L.S., Gage, F.H. and Baird, A. (1991b) Basic FGF-SAP mitotoxin in the hippocampus: specific lethal effect on cells expressing the basic FGF receptor. Ann. N. Y. Acad. Scl 638: 442-444. Gonzalez, A.M., Carman, L.S., Ong, M., Gage, F.H., Shults, C.W. and Baird, A. (1994) Storage, metabolism, and processing of 1251-fibroblast growth factor-2 after intracerebral injection. Brain Res. 665: 285-292. Gray, C.W. and Patel, A.J. (1993) Induction of beta-amyloid precursor protein isoform messenger RNAs by bFGF in astrocytes. NeuroReport 4: 811-814. Grinspan, J.B., Stern, J., Francheschini, B. and Pleasure, D. (1993) Trophic effects of basic fibroblast growth factor (bFGF) on differentiated oligodendroglia: a mechanism for regeneration of the oligodendroglial lineage. J. Neurosci Res. 36: 672-680. Gritti, A., Cova, L., Parati, E.A., Galli, R. and Vescovi, A.L. (1995) Basic fibroblast growth factor supports the proliferation of epidermal growth factor-generated neuronal precursor cells of the adult mouse CNS. Neurosci Lett. 185: 151-154. Grothe, C. and Unsicker, K. (1989) Immunocytochemical localization of basic fibroblast growth factor in bovine adrenal gland, ovary, and pituitary. /. His toe hem. Cytochem. 37:1877-1883. Grothe, C. and Unsicker, K. (1990) Immunocytochemical mapping of basic fibroblast growth factor in the developing and adult rat adrenal gland. Histochemistry 94: 141-147. Grothe, C, Zachmann, K. and Unsicker, K. (1991) Basic FGF-like immunoreactivity in the developing and adult rat brain stem. J. Comp. Neurol 305: 328-336. Grothe, C. and Unsicker, K. (1992) Basic fibroblast growth factor in the hypoglossal system: specific retrograde transport, trophic and lesion-related responses. J. Neurosci Res. 32:317-328. Grothe, C, Otto, D. and Unsicker, K. (1989) Basic fibroblast growth factor promotes in vitro survival and cholinergic development of rat septal neurons: comparison with the effects of nerve growth factor. Neuroscience 31: 649-661 Grothe, C. and Janet, T. (1995) Expression of FGF-2 and FGF receptor type 1 in the adult rat brainstem: effect of colchicine. /. Comp. Neurol 353: 18-24. Guimond, S., Maccarana, M., Olwin, B.B., Lindahl, U. and Rapraeger, A.C. (1993) Activating and inhibitory heparin sequences for FGF-2 (basic FGF): distinct requirements for FGF-1, FGF-2, and FGF-4. J. Biol Chem. 268: 2390623914. Guvakova, M.A., Yakubov, L.A., Vlodavsky, I., Tonkinson, J.L. and Stein, C.A. (1995) Phosphorothioate oligodeoxynucleotides bind to bFGF, inhibit its binding to cell surface receptors, and remove it from low-affinity binding sites on extracellular matrix. J. Biol Chem. 270: 26202627.

S. Bieger and K. Unsicker Hanai, K., Oomura, Y., Kai, Y., Nishikawa, K., Shimizu, N., Morita, H. and Plata-Salaman, C. (1989) Central action of acidic fibroblast growth factor in feeding regulation. Am. J. Physiol. 256: R217-R223. Hanneken, A. and Baird, A. (1992) Immunolocalization of basic fibroblast growth factor: dependence on antibody type and tissue fixation. Exp. Eye Res. 54: 1011-1014. Hara, Y., Tooyama, I., Yasuhara, O., Akiyama, H., McGeer, P.L., Handa, J. and Kimura, H. (1994) Acidic fibroblast growth factor-like immunoreactivity in rat brain following cerebral infarction. Brain Res. 664: 101-107. Hassan, S.M., Kerkhoff, H., Troost, D., Veldman, H. and Jennekens, F.G.I. (1994) Basic fibroblast growth factor immunoreactivity in the peripheral motor system of the rat. Acta Neuropathol. 87: 405-410. Hatten, M.E., Lynch, M., Rydel, R.E., Sanchez, J., JosephSilvestein, J., Moscatelli, D. and Rifkin, D.B. (1988) In vitro neurite extension by granule neurons is dependent upon astroglial-derived fibroblast growth factor. Dev. Biol. 125: 280-289. Haub, O., Drucker, B. and Goldfarb, M. (1990) Expression of the murine fibroblast growth factor 5 gene in the adult rat central nervous system. Proc. Natl. Acad. Sci. USA 87: 8022-8026. Haub, O. and Goldfarb, M. (1991) Expression of the fibroblast growth factor-5 gene in the mouse embryo. Development 112: 391-406. Hawker, J.R. and Granger, H.J. (1994) Nuclear accumulation of exogenous basic fibroblast growth factor in endothelial, fibroblast and myoblast cell lines results in diverse biological responses. In vitro Cell. Dev. Biol. Anim. 30A: 653-663. Haycock, J.W. (1993) Multiple signaling pathways in bovine chromaffin cells regulate tyrosine hydroxylase phosphorylation ar Serl9, Ser31 ,and Ser40. Neurochem. Res. 18: 15-26. Hebert, J.M., Rosenquist, T., Gotz, J. and Martin, G.R. (1994) FGF-5 as a regulator of the hair growth cycle: Evidence form targeted and spontaneous mutations. Cell 78: 10171025. Heikinheimo, M. , Lawshe, A., Shackleford, G.M., Wilson, D.B. and Mac Arthur, C.A. (1994) Fgf-8 expression in the post-gastrulation mouse suggests roles in the development of the face, limbs and central nervous system. Mech. Dev. 48: 129-138. Hendry, I.A. (1992) Retrograde factors in peripheral nerves. Pharmacol. Ther. 56: 265-285. Heuer, J.G., Von Bartheld, C.S., Kinoshita, Y., Evers, P.C. and Bothwell, M. (1990) Alternating phases of FGFreceptor and NGF receptor expression in the developing chicken nervous system. Neuron 5: 283-296. Hicks, D. and Courtois, Y. (1992) Fibroblast growth factor stimulates photoreceptor differentiation in vitro. /. Neurosci. 12: 2022-2033. Hill, C.E., Belford, D.A., Godovac-Zimmermann, J., and Hendry, LA. (1991) Class 1 heparin-binding growth factor promotes the differentiation but not the survival of ciliary neurones in vivo. Dev. Brain Res. 63: 13-19.

367 Hondermarck, H., Courty, J., Boilly, B. and Thomas, D. (1990) Distribution of intravenously administered acidic and basic FGF in the mouse. Experientia 46: 973-974. Honegger, P. and Tenot-Sparti, M. (1992) Developmental effects of basic fibroblast growth factor and platelet-derived growth factor on glial cells in a three-dimensional culture system. /. Neuroimmunol. 40: 295-304. Hou, Y.-L, Yu, A.C.H., Garcia, J.M.R.Z., Aotaki-Keen, A., Lee, Y.-L., Eng, L.F., Hjelmeland, L.J. and Menon, V.K. (1995) Astrogliosis in culture. IV. Effects of basic fibroblast growth factor. /. Neurosci. Res. 40: 359-370. Hughes, R.A., Sendnter, M., Goldfarb, M., Lindholm, D. and Thoenen, H. (1993a) Evidence that fibroblast growth factor 5 is a major muscle-derived survival factor for cultured spinal motoneurons. Neuron 10: 369-377. Hughes, R.A., Sendtner, M. and Thoenen, H. (1993b) Members of several gene families influence survival of rat motoneurons in vitro and in vivo. J. Neurosci. Res. 36: 663671. Humpel, C, Chadi, G., Lippoldt, A., Ganten, D., Fuxe, K. and Olson, L. (1994) Increase of basic fibroblast growth factor (bFGF, FGF-2) messenger RNA and protein following implantation of a microdialysis probe into rat hippocampus. Exp. Brain Res. 98: 229-237. Humpel, C, Lippoldt, A., Chadi, G., Ganten, D., Olson, L. and Fuxe, K. (1993) Fast and widespread increase of basic fibroblast growth factor messenger RNA and protein in the forebrain after kainate-induced seizures. Neuroscience 57: 913-922. Imaizumi, K., Iwata, H., Yoshida, S., Sun, G., Okumura, N. and Shiosaka, S. (1993) Coexistence of amyloid betaprotein precursor and basic fibroblast growth factor in single cells of the rat parietal cortex, hippocampus and basal magnocellular nucleus, y. Chem. Neuroanat. 6: 159-165. Imamura,T., Oka, S., Tanahashi, T. and Okita, Y. (1994) Cell cycle-dependent nuclear localization of exogenously added FGF-1 in BALB/c 3T3 and human vascular endothelial cells. Exp. Cell Res. 215: 363-372. Ishihara, M. (1994) Structural requirements in heparin for binding and activation of FGF-1 and FGF-4 are different from that for FGF-2. Glycobiology 4: 817-824. Ishikawa, R., Nishikori, K. and Furukawa, S. (1991a) Developmental changes in distribution of acidic fibroblast growth factor in rat brain evaluated by a sensitive two-site enzyme immunoassay. J. Neurochem. 56: 836-841. Ishikawa, R., Nishikori, K. and Furukawa, S. (1991b) Appearance of nerve growth factor and acidic fibroblast growth factor with different time courses in the cavity-lesioned cortex of the rat brain. Neurosci. Lett. 127: 70-72. Ishikawa, K., Ohe, U., Okutomi, Y., Takeuchi, T. and Suzuki, M. (1992) Neurotrophic effects of FGFs on peptidecontaining neurons in culture from postnatal rat hypothalamus. Neuroendocrinology 55: 193-198. Itoh, N., Yazaki, N., Tagashira, S., Miyake, A., Ozaki, K., Minami, M., Satoh, M., Ohta, M. and Kawasaki, T. (1994) Rat FGF receptor-4 mRNA in the brain is expressed prefer-

368 entially in the medial habenular nucleus. Mol Brain Res. 21:344-348. Janet, T., Miehe, M., Pettmann, B., Labourdette, G. and Sensenbrenner, M. (1987) Ultrastructural localization of fibroblast growth factor in neurons of rat brain. Neurosci. Lett. 68: 175-180. Jaye, M., Schlessinger, J. and Dionne, C.A. (1992) Fibroblast growth factor receptor tyrosine kinases: molecular analysis and signal transduction. Biochim. Biophys. Acta 1135: 185199. Kan, M., Wang, F., Xu, J., Crabb, J.W., Hou, J. and McKeehan, W.J. (1993) An essential heparin-binding domain in the fibroblast growth factor receptor kinase. Science 259: 1918-1921. Kato, T., Sasaki, H., Katagiri, T., Sasaki, H., Koiwai, K., Youki, H., Totsuka, S. and Ishii, T. (1991) The binding of basic fibroblast growth factor to Alzheimer's neurofibrillary tangles and senile plaques. Neurosci. Lett. 122: 33-36. Kiefer, P., Acland, P., Pappin, D., Peters, G. and Dickson, C. (1994) Competition between nuclear localization and secretory signals determines the subcellular fate of a single CUG-initiated form of FGF-3. EMBO J. 13: 4126-4136. Kimelman, D. and Kirschner, M.W. (1989) An antisense mRNA directs the covalent modification of the transcript encoding fibroblast growth factor in Xenopus oocytes. Cell 59: 687-696. Kinoshita, Y., Kinoshita, C , Heuer, J.G. and Bothwell, M. (1993) Basic fibroblast growth factor promotes adhesive interactions of neuroepithelial cells from chick neural tube with extracellular matrix proteins in culture. Development 119:943-956. Kitchens, D.L., Snyder, E.Y. and Gottlieb, D.I. (1994) FGF and EGF are mitogens for immortalized neural progenitors. J. Neurobiol. 25: 797-807. Kiyota, Y., Takami, K., Iwane, M., Shino, A., Miyamoto, M., Tsukuda, R. and Nagaoka, A. (1991) Increase in basic fibroblast growth factor-like immunoreactivity in rat brain after forebrain ischemia. Brain Res. 545: 322-328. Kniss, D.A. and Burry, R.W. (1988) Serum and fibroblast growth factor stimulate quiescent astrocytes to re-enter the cell cycle. Brain Res. 439: 281-288. Knusel, B., Michel, P.P., Schwaber, J.S. and Hefti, F. (1990) Selective and nonselective stimulation of central cholinergic and dopaminergic development in vitro by nerve growth factor, insulin and the insulin-like growth factors I and II. /. Neurosci. 10: 558-570. Koketsu, N., Berlove, D.J., Moskowitz, M.A., Kowall, N.W., Caday, C.G. and Finklestein, S.P. (1994) Pretreatment with intraventricular basic fibroblast growth factor (bFGF) decreases infarct size following focal cerebral ischemia in rats. Ann. Neurol. 35: 451-457. Kornblum, H.I., Raymon, H.K., Morrison, R.S., Cavanaugh, K.P., Bradshaw, R.A. and Leslie, F.M. (1990) Epidermal growth factor and basic fibroblast growth factor: effects on an overlapping population of neocortical neurons in vitro. Brain Res. 535: 255-263.

Functions of fibroblast growth factors (FGFs) in the nervous system Korsching, S., Auburger, G., Heumann, R., Scott, J. and Thoenen, H. (1985) Levels of nerve growth factor and its mRNA in the central nervous system of the rat correlate with cholinergic innervation. EMBO J. 4: 1389-1393. Koshinaga, M., Sanon, H.R. and Whittemore, S.R. (1993) Altered acidic and basic fibroblast growth factor expression following spinal cord injury. Exp. Neurol. 120: 3 2 48. Kostyk, S.K., DiAmore, P.A., Herman, I.M. and Wagner, J.A. (1994) Optic nerve injury alters basic fibroblast growth factor localization in the retina and optic tract. J. Neurosci. 14: I44i_i449. Kresse, A., Pettersson, R. and Hokfelt, T. (1995) Distribution of acidic fibroblast growth factor mRNA-expressing neurons in the adult mouse central nervous system. /. Comp. Neurol, (in press). Kumon, Y., Sakaki, S., Kadota, O., Matsuda, S., Fujita, H., Yoshimura, H. and Sakanaka, M. (1993) Transient increase in endogenous basic fibroblast growth factor in neurons of ischemic rat brains. Brain Res. 605: 169-174. Kushima, Y., Nishio, C , Nonomura, T. and Hatanaka, H. (1992) Effects of nerve growth factor and basic fibroblast growth factor on survival of cultured septal cholinergic neurons from adult rats. Brain Res. 598: 264-270. Laeng, P., Decimo, D., Pettmann, B., Janet, T. and Labourdette, G. (1994) Retinoic acid regulates the development of oligodendrocyte precursor cells in vitro. J. Neurosci. Res. 39:613-633. Lappi, D.A., Martineau, D. and Baird, A. (1989) Biological and chemical characterization of basic FGF-saporin mitotoxin. Biochem. Biophys. Res. Commun. 160: 917-923. Lavail, M.M., Unoki, K., Yasumura, D., Matthes, M.T., Yancopoulos, G.D. and Steinberg, R.H. (1992) Multiple growth factors, cytokines, and neurotrophins rescue photoreceptors from the damaging effects of constant light. Proc. Natl. Acad. Sci. USA 89: 11249-11253. Lazarini, F., Castelnau, P., Chermann, J.-F., Deslys, J.-P. and Dormont, D. (1994) Modulation of prion protein gene espression by grwoth factors in cultured mouse astrocytes and PC-12 cells. Mol. Brain Res. 22: 268-274. Lennon, A.M., Esfandiari, A., Gavaret, J.-M., Courtin, F. and Pierre, M. (1994) 12-0-tetradecanoylphorbol 13-acetate and fibroblast growth factors in cultured mouse astrocytes and PC-12 cells. Mol. Brain Res. 22: 268-274. Leonard, S., Luthaman, D., Logel, J., Antle, C , Freedman, R. and Hoffer, B. (1993) Acidic and basic fibroblast growth factor mRNAs are increased in striatum following MPTPinduced dopamine neurofiber lesion: assay by quantitative PCR. Mol. Brain Res. 18: 275-284. Lewis, G.P., Erickson, P.A., Guerin, C.J., Anderson, D.H. and Fisher, S.K. (1992) Basic fibroblast growth factor: a potential regulator of proliferation and intermediate filament expression in the retina. J. Neurosci. 12: 3968-3978. Lindholm, D., Castren, E., Berzaghi, M., Blochl, A. and Thoenen, H. (1994) Activity-dependent and hormonal regulation of neurotrophin mRNA levels in the brain -

S. Bieger and K. Unsicker implications for neuronal plasticity. J. Neurobiol 25: 13621372. Lindholm, D., Castren, E., Kiefer, R., Zafra, F. and Thoenen, H. (1992) Transforming growth factor-bl in the rat brain: increase after injury and inhibition of astrocyte proliferation./. Cell Biol 117: 395-400. Lindholm, D., Hartikka, J., da Penha Berzaghi, M., Castren, E., Tzimagiorgis, G., Hughes, R.A. and Thoenen, H. (1994) Fibroblast growth factor-5 promotes differentiation of cultured rat septal cholinergic and raphe serotonergic neurons: comparison with the effects of neurotrophins. Eur. J. Neurosci. 6: 244-252. Lipton, S.A., Wagner, J.A., Madison, R.D. and D'Amore, P.A. (1988) Acidic fibroblast growth factor enhances regeneration of processes by postnatal mammalian retinal ganglion cells in culture. Proc. Natl. Acad. Sci. USA 85: 2388-2392. Liu, H.M. and Chen, H.H. (1994a) c-fos protein expression and ischemic changes in neurons vulnerable to ischemia/hypoxia, correlated with basic fibroblast growth factor immunoreactivity. J. Neuropathol. Exp. Neurol. 53: 598605. Liu, H.M. and Chen, H.H. (1994b) Correlation between fibroblast growth factor expression and cell proliferation in experimental brain infarct - studied with proliferating cell nuclear antigen immunohistochemistry. J. Neuropathol. Exp. Neurol. 53: 118-126. Logan, A. and Berry, M . (1993) Transforming growth factorbeta(l) and basic fibroblast growth factor in the injured CNS. Trends Pharmacol. Sci. 14: 337-343. Logan, A. , Frautschy, S.A., Gonzalez, A.-M. and Baird, A. (1992) A time course for the focal elevation of synthesis of basic fibroblast growth factor and one of its high-affinity receptors (fig) following a localized cortical brain injury. J. Neurosci. 12: 3828-3237. Loret, C, Janet, T., Labourdette, G., Schneid, H. and Binoux, M. (1991) FGFs stimulate IGF binding proteiin synthesis without affecting IGF synthesis in rat astroblasts in primary culture. Glia 4: 378-383. Luo, L., Koutnouyan, H., Baird, A. and Ryan, A.F. (1993) Acidic and basic FGF messenger RNA Expression in the adult and developing rat cochlea. Hear. Res. 69: 182-193. Luo, L., Moore, J.K., Baird, A. and Ryan, A.F. (1995) Expression of acidic FGF mRNA in rat auditory brainstem during postnatal maturation. Dev. Brain Res. 86: 24-34. Mahmood, R., Kiefer, P., Guthrie, S., Dickson, C. and Mason, I. (1995) Multiple roles for FGF-3 during cranial neural development in the chicken. Development 121: 1399-1410. Maiese, K., Boniece, I., Demeo, D. and Wagner, J.A. (1993) Peptide growth factors protect against ischemia in culture by preventing nitric oxide toxicity. J. Neurosci. 13: 3034-3040. Mansour, S.L. (1994) Targeted disruption of int-2 (fgf-3) causes developmental defects in the tail and inner ear. Mol. Reprod. Dev. 39: 62-68. Maries, I., Adelaide, J., Raybaud, F., Mattel, M.-G., Coulier, F., Planche, J., de Lapeyriere, O. and Birnbaum, D. (1989) Characterization of the HST-related FGF.6 gene, a new

369 member of the fibroblast growth factor gene family. Oncogene 4: 335-340. Marin, F., Boya, J., Calvo, J.L., Lopez-Munoz, F. and GarclaMaurino, J.E. (1994) Immunocytochemical localization of basic fibroblast growth factor in the rat pineal gland. J. Pineal Res. 16:44-49. Mason, I.J., Fuller-Pace, F., Smith, R. and Dickson, C. (1994) FGF-7 (keratinocyte growth factor) expression during mouse development suggests roles in myogenesis, forebrain regionalisation and epithelial-mesenchymal interactions. Mech. Dev. 45: 15-30. Matsuda, S., Saito, H. and Nishiyama, N. (1990) Effect of basic fibroblast growth factor on neurons cultured from various regions of postnatal rat brain. Brain Res. 520: 310316. Matsuda, S., Desaki, J., Okumura, N., Shiosaka, S., Imaoka, S. and Sakanaka, M. (1992) Basic fibroblast growth factorlike immunoreactivity in the trigeminal proprioceptive and motor systems. Brain Res. 511: 92-100. Matsuda, S., Okumura, N., Yoshimura, H., Koyama, Y. and Sakanaka, M. (1992b) Basic fibroblast growth factor-like immunoreactivity in Purkinje cells of the rat cerebellum. Neuroscience 50: 99-106. Mattson, M.P. and Rychlik, B. (1990) Glia protect hippocampal neurons against excitatory amino acid-induced degeneration: involvement of fibroblast growth factor. Int. J. Dev. Neurosci. 8: 399^15. Mattson, M.P., Murrain, M., Guthrie, P.B. and Kater, S.B. (1989) Fibroblast growth factor and glutamate: opposing roles in the generation and degeneration of hippocampal neuroarchitecture. J. Neurosci. 9: 3728-3740. Mattson, M.P.and Scheff, S.W. (1994) Endogenous neuroprotection factors and traumatic brain injury: mechanisms of action and implications for therapy. J. Neurotrauma 11:333. Mattson, M.P., Barger, S.W., Cheng, B., Lieberburg, I., Smithswintosky, V.L. and Rydel, R.E. (1993a) betaAmyloid precursor protein metabolites and loss of neuronal Ca2+ homeostasis in Alzheimer's disease. Trends. Neurosci. 16: 409^14. Mattson, M.P., Cheng, B. and Smithswintosky, V.L. (1993b) Mechanisms of neurotrophic factor protection against calcium-and free radical-mediated excitotoxic injury - implications for treating neurodegenerative disorders. Exp. Neurol. 124: 89-95. Mattson, M.P., Kumar, K.N., Wang, H., Cheng, B. and Michaelis, E.K. (1993c) Basic FGF regulates the expression of a functional 71-kDa NMD A receptor protein that mediates calcium influx and neurotoxicity in hippocampal neurons. J. Neurosci. 13: 4575-4588. Mattson, M.P., Tomaselli, K.J. and Rydel, R.E. (1993d).Calcium-destabilizing and neurodegenerative effects of aggregated beta-amyloid peptide are attenuated by basic FGF. Brain Res. 621: 35-49. Matsuo, A., Tooyama, I., Oomura, Y., Akiguchi, I., Hanai, K., Kimura, J. and Kimura, H. (1994) Immunohistochemical

370

Functions of fibroblast growth factors (FGFs) in the nervous system

localization in the rat brain of an epitope corresponding to the fibroblast growth factor receptor-1. Neuroscience 60: 49-66. Mayer, E., Dunnett, S.B., and Fawcett, J.W. (1993a) Mitogenic effect of basic fibroblast growth factor on embryonic ventral mesencephalic dopaminergic neurone precursors. Dev. Brain Res. 12: 253-258. Mayer, M., Bogler, O. and Noble, M. (1993b) The inhibition of oligodendrocytic differentiation of 02-A progenitors caused by basic fibroblast growth factor is overridden by astrocytes. Glia 8: 12-19. Meiners, S., Marone, M., Rittenhouse, J.L. and Geller, H.M. (1993) Regulation of astrocytic tenascin by basic fibroblast growth factor. Dev. Biol. 160: 480-493. McAvoy, J.W., Chamberlain, C.G., delong, R.U., Richardson, N.A. and Lovicu, F.J. (1991) The role of fibroblast growth factor in eye lens development. Ann. N. Y. Acad. Sci. 638: 256-274. McKeehan, W.L. and Kan, M. (1994) Heparan sulfate fibroblast growth factor receptor complex: structure-function relationships. Mol. Reprod. Dev. 39: 69-82. McKinnon, R.D., Matsui, T., Dubois-Dalcq, M., and Aaronson, S.A. (1990) FGF modulates the PDGF-driven pathway of oligodendrocyte development. Neuron 5: 603-614. Middaugh, C.R., Mach, H., Burke, C.J., Volkin, D.B., Dabora, J.M., Tsai, P.K., Bruner, M.W., Ryan, J.A. and Marfia, K.E. (1992) Nature of the interaction of growth factors with suramin. Biochemistry 31: 9016-9024. Mitani, A., Oomura, Y., Yanase, H. and Kataoka, K. (1992) Acidic fibroblast growth factor delays in vitro ischemiainduced intracellular calcium elevation in gerbil hippocampal slices: a sign of neuroprotection. Neurochem. Int. 21: 337-341. Miyamoto, M., Naruo, K.-I., Seko, C, Matsumoto, S., Kondo, T. and Kurokawa, T. (1993) Molecular cloning of a novel cytokine cDNA encoding the ninth member of the fibroblast growth factor family which has a unique secretion property. Mol. Cell. Biol. 13: 4251-4259. Morgan, L., lessen, K.R. and Mirsky, R. (1994) Negative regulation of the Po gene in Schwann cells: suppression of Po mRNA and protein induction in cultured Schwann cells by FGF2 and TGFy31, TGF/S2 and TGF/33. Development 120: 1399-1409. Morrison, R.S., Sharma, A., De Vellis, J. and Bradshaw, R.A. (1986) Basic fibroblast growth factor supports the survival of cerebral cortical neurons in primary culture. Proc. Natl. Acad. Sci. USA 83: 7537-7541. Morrison, R.S. (1991) Suppression of basic fibroblast growth factor expression by antisense oligodeoxynucleotides inhibits the growth of transformed human astrocytes. J. Biol. Chem. 266: 728-734. Morrison, R.S., Yamaguchi, F., Saya, H., Bruner, J.M., Yahanda, A.M., Donehower, L.A. and Berger, M. (1994) Basic fibroblast growth factor and fibroblast growth factor receptor 1 are implicated in the growth of human astrocytomas. /. Neurooncol. 18: 207-216.

Murphy, A.N., Unsworth, E.J. and Stetler-Stevenson, W.G. (1993) Tissue inhibitor of metalloproteinases-2 inhibits bFGF-induced human microvascular endothelial cell proliferation. J. Cell. Physiol. 157: 351-358. Murphy, P.R., Sato, Y. and Knee, R.S. (1992) Phosphorthioate antisense oligonucleotides against basic fibroblast growth factor inhibit anchorage-dependent and anchorageindependent growth of a malignant glioblastoma cell line. Mol. Endocrinol. 6: 877-884. Nakata, N., Kato, H. and Kogure, K. (1993) Protective effects of basic fibroblast growth factor against hippocampal neuronal damage following cerebral ischemia in the gerbil. Brain Res. 605: 354-356. Nakayama, Y., Iwahana, M., Sakamoto, N., Tanaka, N.G. and Osada, Y. (1993) Inhibitory effects of a bacteria-derived sulfated polysaccharide against basic fibroblast growth factor-induced endothelial cell growth and chemotaxis. /. Cell. Physiol. 154: 1-6. Neary, J.T., Whittemore, S.R., Zhu, Q. and Norenberg, M.D. (1994) Synergistic activation of DNA synthesis in astrocytes by fibroblast growth factors and extracellular ATP. J. Neurochem. 63: 490-494. Neufeld, G. and Gospodarowicz, D. (1987) Protamine sulfate inhibits mitogenic activities of the extracellular matrix and fibroblast growth factor, but potentiates that of epidermal growth factor. J. Cell Physiol. 132: 287-294. Nieto-Sampedro, M., Lim, R., Hicklin, D.J. and Cotman, C.W. (1988) Early release of glia maturation factor and acidic fibroblast growth factor after rat brain injury. Neurosci. Lett. 86: 361-365. Niswander, L., Tickle, C, Vogel, A. and Martin, G. (1994) Function of FGF-4 in limb development. Mol. Reprod. Dev. 39: 83-89. Nozaki, K., Finklestein, S.P. and Beal, M.F. (1993) Basic fibroblast growth factor protects against hypoxia-ischemia and NMDA neurotoxicity in neonatal rats. J. Cerebral Blood Flow Metab. 13: 221-228. Nugent, M.A. and Edelman, E.R. (1992) Kinetics of basic fibroblast growth factor binding to its receptor and heparan sulfate proteoglycan: A mechanism for cooperativity. Biochemistry 31: 8876-8883. Oellig, C, Pirvola, U., Taylor, L., Elde, R., Hokfelt, T. and Pettersson, R.F. (1995) Acidic FGF and FGF receptors are specifically expressed in neurons of developing and adult rat dorsal root ganglia. Eur. J. Neurosci. 7: 863-874. Ohuchi, H., Yoshioka, H., Tanaka, A., Kawakami, Y., Nohno, T. and Noji, S. (1994) Involvement of androgen-induced growth factor (FGF-8) gene in mouse embryogenesis and morphogenesis. Biochem. Biophys. Res. Commun. 204: 882-888. Olson, J.J., Polk, D.M., and Reisner, A. (1994) The efficacy and distribution of suramin in the treatment of the 9L glliosarcoma. Neurosurgery 34: 297-308. Olwin, B.B. and Rapraeger, A. (1992) Repression of myogenic differentiation by aFGF, bFGF and k-FGF is dependent on cellular heparan sulfate. J. Cell Biol. 118: 631-639.

S. Bieger and K. Unsicker Olwin, B.B., Arthur, K., Hannon, K., Hein, P., McFall, A., Riley, B., Szebenyl, G., Zhou, Z., Zuber, M.E., Rapraeger, A.C., Fallon, J.E. and Kudla, A.J. (1994) Role of FGFs in skeletal muscle and limb development. Mol. Reprod. Dev. 39:90-101. Oppenheim, R.W., Prevette, D. and Fuller, F. (1992) The lack of effect of basic and acidic FGFs on the naturally occurring death of neurons in the chick embryo. J. Neurosci. 12: 2726-2734. Ortega, S., Schaeffer, M.T., Soderman, D., DiSalvo, J., Linemeyer, D.L., Giminez-Gallego, G. and Thomas, K.A. (1991) Conversion of cysteine to serine residues alters the activity, stability, and heparin dependence of aFGF. /. Biol. Chem. 266: 5842-5846. Oshawa, F., Widmer, H.R., Knusel, B., Denton, T.L. and Hefti, F. (1993) Response of embryonic rat hippocampal neurons in culture to neurotrophin-3, brain-derived neurotrophic factor and badic fibroblast growth factor. Neuroscience 57: 67-77. Otsuka, H., Matsuda, S., Fujita, H., Matsuda, J., Shibata, T., Uryu, K. and Sakanaka, M. (1993) Localization of basic fibroblast growth factor (bFGF)-like immunoreactivity in neural circuits innervating the gastrocnemius muscle, with reference to the direction of bFGF transport. Arch. Histol. Cytol. 56: 207-215. Otto, D., Frotscher, M. and Unsicker, K. (1989) Basic fibroblast growth factor and nerve growth factor administered in gel foam rescue medial septal neurons after fimbria fornix transection. J. Neurosci. Res. 22: 83-91. Otto, D., Unsicker, K. and Grothe, C. (1987) Pharmacological effects of nerve growth factor and fibroblast growth factor applied to the transectioned sciatic nerve on neuron death in adult rat dorsal root ganglia. Neurosci. Lett. 83: 156-160. Otto, D. and Unsicker, K. (1990) Basic FGF reverses chemical and morphological deficits in the nigrostriatal system of MPTP-treated mice. y. N^Mro^d. 10: 1912-1921. Otto, D. and Unsicker, K. (1993a) FGF-2 modulates dopamine and dopamine-related striatal transmitter systems in the intact and MPTP-lesioned mouse. Eur. J. Neurosci. 5: 927932. Otto, D. and Unsicker, K. (1993b) FGF-2-mediated protection of cultured mesencephalic dopaminergic neurons against MPTP and MPP+: specificity and impact of culture conditions, non-dopaminergic neurons, and astroglial cells. J. Neurosci. Res. 34: 1-12. Partanen, J., Vainikka, S., Korhonen, J., Armstrong, E. and Alitalo, K. (1992) Diverse receptors for fibroblast growth factors. Prog. Growth Factor Res. 4: 69-83. Pechan, P.A., Chowdhury, K., Gerdes, W. and Seifert, W. (1993) Glutamate induces the growth factors NGF, bFGF, the receptor FGF-Rl and c-fos messenger RNA expression in rat astrocyte culture. Neurosci. Lett. 153: 111-114. Perraud, F., Labourdette, G., Miehe, M. and Sensenbrenner, M. (1988) Comparison of the morphological effects of acidic and basic fibroblast growth factors on rat astroblasts in culture. J. Neurosci. Res. 20: 1-11.

371 Peters, K.G., Werner, S., Chen, G. and Williams, L.T. (1992) Two FGF receptor genes are differentially expressed in epithelial and mesenchymal tissues during limb formation and organogenesis in the mouse. Development 114: 233-243. Peters, K., Ornitz, D., Werner, S. and Williams, L. (1993) Unique expression pattern of the FGF receptor 3 gene during mouse organogenesis. Dev. Biol. 155: 423-430. Peters, K., Basbaum, A., Williams, L. and Wilcox, J. (1995) Localization of two FGF receptors in the adult mouse central nervous system by in situ hybridization: evidence for different neuronal and glial FGF receptors, (in press). Pettman, B., Labourdette, G., Weibel, M. and Sensenbrenner, M. (1986) The brain fibroblast growth factor (FGF) is localized in neurons. Neurosci. Lett. 68: 175-180. Peulve, P., Laquerriere, A., Paresy, M., Hemet, J. and Tadie, M. (1994) Establishment of adult rat Schwann cell cultures: effect of bFGF, a-MSH, NGF, PDGF, and IGV-p on cell cycle. Exp. Cell Res. 214: 543-550. Piehl, F., Arvidsson, U., Johnson, H., Cullheim, S., Dagerlind, A., Ulfhake, B., Cao, Y., Elde, R., Pettersson, R.F., Terenius, L. et al. (1993) GAP-43, aFGF, CCK and alpha-CGRP and beta-CGRP in rat spinal motoneurons subjected to axotomy and/or dorsal root severance. Eur. J. Neurosci. 5: 1321-1333. Pollock, J.D., Krempin, M. and Rudy, B. (1990) Differential effects of NGF, FGF, EGF, cAMP and dexamethasone on neurite outgrowth and sodium channel expression in PC 12 cells. J. Neurosci. 10: 2626-2637. Pons, S. and Torres-Aleman, I. (1992) Basic fibroblast growth factor modulates insulin-like growth factor-I, its receptor, and its binding proteins in hypothalamic cell cultures. Endocrinology 131: 2271-2278. Powell, P.P., Finklestein, S.P., Dionne, C.A., Jaye, M. and Klagsbrun, M. (1991) Temporal, differential and regional expression of mRNA for basic fibroblast growth factor in the developing and adult rat brain. Mol. Brain Res. 11:11-11. Powell, P.P. and Klagsbrun, M. (1993) Regulation of basic fibroblast growth factor mRNA expression in rat C6 glioma cells. Exp. Cell Res. 209: 224-230. Prats, H., Kaghad, M., Prats, A.C., Klagsbrun, M., Lelias, J.M., Liauzun, P., Chalon, P., Tauber, J.P., Amalric, F., Smith, J.A. and Caput, D. (1989) High molecular mass forms of basic fibroblast growth factor are initiated by alternative CUG codons. Proc. Natl. Acad. Sci. USA 86: 1836-1840. Presta, M., Statuto, M., Isacchi, A., Caccia, P., Pozzi, A., Gualandris, A., Rusnati, M., Bergonzi, L. and Sarmientos, P. (1992) Structure-function relationship of basic fibroblast growth factor: site-directed mutagenesis of a putative heparin-binding and receptor-binding region. Biochem. Biophys. Res. Commun. 185: 1098-1107. Prudovsky, I., Savion, N., Zhan, X., Friesel, R., Xu, J., Hou, J., McKeehan, W.L. and Maciag, T. (1994) Intact and functional fibroblast growth factor (FGF) receptor-1 traffics near the nucleus in reponse to FGF-1. J. Biol. Chem. 269: 31720-31724.

372 Puchacz, E., Stachowiak, E.K., Florkiewicz, R.Z., Lukas, R.J. and Stachowiak, M.K. (1993) Basic fibroblast growth factor (bFGF) regulates tyrosine hydroxylase and proenkephalin mRNA levels in adrenal chromaffin cells. Brain Res. 610: 39-52. Quon, D., Catalano, R. and Cordell, B. (1990) Fibrobalst growth factor induces b-amyloid precursor mRNA in glial but not neuronal cultured cells. Biochem. Biophys. Res. Commun. 167: 96-102. Radhakrishna, M. and Almazan, G. (1994) Protein kinases mediate basic fibroblast growth factor's stimulation of proliferation and c-fos induction in oligodendrocyte progenitors. Mol. Brain Res. 24: 118-128. Rakoczy, P.E., Humphrey, M.F., Cavaney, D.M., Chu, Y. and Constable, I.J. (1993) Expression of basic fibroblast growth factor and its receptor in the retina of Royal College of Surgeons rats - a comparative study. Invest. Ophthalmol. Visual Sci. 34: 1845-1852. Rane, S.G. and Pollock, J.D. (1994) Fibroblast growth factorinduced increases in calcium currents in the PC 12 pheochromocytoma cell line are tyrosine phosphorylation dependent. J. Neurosci. Res. 38: 590-598. Regli, L., Anderson, R.E. and Meyer, F.B. (1994) Basic fibroblast growth factor increases cortical blood flow in vivo. Brain Res. 665: 155-157. Reid, H.H., Wilks, A.F. and Bernard, O. (1990) Two forms of the basic fibroblast growth factor receptor-like mRNA are expressed in the developing mouse brain. Proc. Natl. Acad. Sci. USA 87: 1596-1600. Represa, J., Ledn, Y., Miner, C. and Giraldez, F. (1991) The int-2 proto-oncogene is responsible for induction of the inner ear. Nature 353: 561-563. Riedel, B., Friauf, E., Grothe, C. and Unsicker, K. (1995) Fibroblast growth factor-2-like immunoreactivity in auditory brainstem of the developing and adult rat: correlation with onset and loss of hearing. J. Comp. Neurol. 354: 3 5 3 360. Rifkin, D.B., Moscatelli, D., Roghani, M., Nagano, Y., Quarto, N., Klein, S. and Bikfalvi, A. (1994) Studies on FGF-2: nuclear localization and function of high molecular weight forms and receptor binding in the absence of heparin. Mol. Reprod. Dev. 39: 102-105. Riva, M.A. and Mocchetti, I. (1991) Developmental expression of the basic fibroblast growth factor gene in rat brain. Dev. Brain Res. 62: 45-50. Riva, M.A., Gale, K. and Mocchetti, I. (1992) Basic fibroblast growth factor mRNA increases in specific brain regions following convulsive seizures. Mol. Brain Res. 15: 311-318. Riva, M.A., Donati, E., Tascedda, F., Zolli, M. and Racagni, G. (1994) Short- and long-term induction of basic fibroblast growth factor gene expression in rat central nervous system following kainate injection. Neuroscience 59: 55-65. Rivera, S., Gold, S.J. and Gall, CM. (1994) Interleukin-lb increases basic fibroblast growth factor mRNA expression in adult rat brain and organotypic hippocampal cultures. Mol. Brain Res. 27: 12-26.

Functions of fibroblast growth factors (FGFs) in the nervous system Rizzino, A. and Rosfjord, E. (1994) Transcriptional regulation of the murine k-fgf gene. Mol. Reprod. Dev. 39: 106-111. Roghani, M. and Moscatelli, D. (1992) Basic fibroblast growth factor is internalized through both receptormediated and heparan sulfate-mediated mechanisms. J. Biol. Chem. 267: 22156-22162. Rogister, B., Leprince, P., Pettmann, B., Labourdette, G., Sensenbrenner, M. and Moonen, G. (1988) Brain basic fibroblast growth factor stimulates the release of plasminogen activators by newborn rat cultured astroglial cells. Neurosci. Lett. 91: 321-326. Rosenblatt, S., Irikura, K., Caday, C.G., Finklestein, S.P. and Moskowitz, M.A. (1994) Basic fibroblast growth factor dilates rat pial arterioles. J. Cerebral Blood Flow Metab. 14: 70-74. Ruoslahti, E. and Yamaguchi, Y. (1991) Proteoglycans as modulators of growth factor activities. Cell 64: 867-869. Rusnati, M., Urbinati, C. and Presta, M. (1993) Internalization of basic fibroblast growth factor (bFGF) in cultured endothelial cells: Role of the low affinity heparin-like bFGF receptors./. Cell. Physiol. 154: 152-161. Rydel, R.E. and Greene, L.A. (1987) Acidic and basic fibroblast growth factors promote stable neurite outgrowth and neuronal differentiation in cultures of PC 12 cells. /. Neurosci. 7: 3639-3653. Ryken, T.C., Traynelis, V.C. and Lim, R. (1992) Interaction of aFGF and TGF-b in normal and transformed glia in vitro. J. Neurosurg. 76: 850-855. Sakaguchi, T., Yamada, K., Wanaka, A., Kohmura, E., Yuguchi, T., Taneda, M., Kataoka, K., Tohyama, M. and Hayakawa, T. (1994) Expression of basic fibroblast growth factor receptor messenger RNA in the periinfarcted brain tissue. Res tor. Neurol. Neurosci. 7: 29-36. Sano, M. and Kitajima, S. (1992) Activation of microtubuleassociated protein kinase in PCI 2D cells in response to both fibroblast growth factor and epidermal growth factor and concomitant stimulation of the outgrowth of neurites. J. Neurochem. 58: 837-844. Sasaki, K., Oomura, Y., Suzuki, K., Hanai, K. and Yagi, H. (1992) Acidic fibroblast growth factor prevents death of hippocampal cal pyramidal cells following ischemia. Neurochem. Int. 21:397-402. Sasaki, K., Li, A.J., Oomura, Y., Muto, T., Hanai, K., Tooyama, I., Kimura, H., Yanaihara, N., Yagi, H. and Hori, T. (1994a) Effects of fibroblast growth factors and related peptides on food intake by rats. Physiol. Behav. 56: 2 1 1 218. Sasaki, K., Oomura, Y., Figurov, A., and Yagi, H. (1994b) aFGF facilitates generation of long-term potentiation in rat hippocampal slices. Brain Res. Bull. 33: 505-511. Schmidt, M.F. and Kater, S.B. (1993) Fibroblast growth factors, depolarization, and substratum interact in a combinatorial way to promote neuronal survival. Dev. Biol. 158: 228-237. Schmidt, K.-F., Billek, M., Pietruck, C , Noll, G.N., Goureau, O. and Courtois, Y. (1995) Fibroblast growth factors alter

S. Bieger and K. Unsicker light responses and dark voltage in retinal rods of the frog {Rana temporaria). Neurosci. Lett. 191: 177-180. Schubert, D., Ling, A. and Baird, A. (1987) Multiple influences of a heparin-binding growth factor on neuronal development. J. Cell Biol. 104: 635-643. Seddon, A., Decker, M., Miiller, T., ArmeUino, D., Kovesdi, I., Gluzman, Y. and Bohlen, P. (1991) Structure/activity relationships in basic FGF. Ann. N. Y. Acad. Sci 638: 9 8 108. Seno, M., Sasada, R., Iwane, M., Sudo, K., Kurokawa, T., Ito, K. and Igarishi, K. (1988) Stabilizing basic fibroblast growth factor using protein engineering. Biochem. Biophys. Res. Commun. 151: 701-708. Seno, M., Sasada, R., Kurokawa, T. and Igarishi, K. (1990) Carboxyl-terminal structure of basic fibroblast growth factor significantly contributes to its affinity for heparin. Eur. J. Biochem. 188:239-245. Sensenbrenner, M. (1993) The neurotrophic activity of fibroblast growth factors. Prog. Neurobiol. 41: 683-704. Sensenbrenner, M., Labourdette, G., Pettmann, B., Perraud, F. and Besnard, F. (1987) Neuronal-derived factors regulating glial cell proliferation and maturation. J. Physiol. (Paris) 82: 288-290. Sherman, L., Stocker, K.M., Morrison, R. and Ciment, G. (1993) Basic fibroblast growth factor (bFGI^ acts intracellularly to cause the transdifferentiation of avian neural crest-derived Schwann cell precursors into melanocytes.

Development m-. 1313-1326. Siedlak, S.L., Cras, P., Kawai, M., Richey, P. and Perry, G. (1991) Basic fibroblast growth factor binding is a marker for extracellular neurofibrillary tangles in Alzheimer disease. J. Histochem. Cytochem. 39: 899-904. Sievers, J., Hausmann, B., Unsicker, K. and Berry, M. (1987) Fibroblast growth factors promote the survival of adult rat retinal ganglion cells after transection of the optic nerve. Neurosci. Lett. 76: 157-162. Simpson, C. and Morris, B.J. (1994) Basic fibroblast growth factor induces c-fos expression in primary cultures of rat striatum. Neurosci. Lett. 170: 281-285. Slack, J. (1994) Role of fibroblast growth factors as inducing agents in early embryonic development. Mol. Reprod. Dev. 39: 118-125. Sommer, A. and Rifkin, D.B. (1989) Interaction of heparin with human basic fibroblast growth factor: protection of the angiogenic protein from proteolytic degradation by a glycosaminoglycan. 7. Cell Physiol. 138: 215-220. Soulet, L., Chevet, E., Lemaitre, G., Blanquaert, F., Meddahi, A. and Barritault, D. (1994) FGFs and their receptors, in vitro and in vivo studies: new FGF receptor in the brain, FGF-1 in muscle, and the use of functional analogues of low-affinity heparin-binding growth factors in tissue repair. Mol. Reprod. Dev. 39: 49-55. Spivak-Kroizman, T., Lemmon, M.A., Dikic, I., Ladbury, J.E., Pinchasi, D., Huang, J.,. Jaye, M., Crumley, G., Schlessinger, J. and Lax, I. (1994) Heparin-induced oligomerization of FGF molecules is responsible for FGF receptor

373 dimerization, activation, and cell proliferation. Cell 79: 1015-1024. Stachowiak, M.K., Moffett, J., Joy, A., Puchacz, E., Florkiewicz, R. and Stachowiak, E.K. (1994) Regulation of bFGF gene expression and subcellular distribution of bFGF protein in adrenal medullary cells. J. Cell Biol. 127: 2 0 3 223. Stefanik, D.F., Rizkalla, L.R., Soi, A., Goldblatt, S.A. and Rizkalla, W.M. (1991) Acidic and basic fibroblast growth factors are present in glioblastoma multiforme. Cancer Res. 51:5760-5765. Stemple, D.L., Mahanthappa, N.K. and Anderson, D.J. (1988) Basic FGF induces neuronal differentiation, cell division, and NGF dependence in chromaffin cells: a sequence of events in sympathetic development. Neuron 1: 517-525. Stock, A., Kuzis, K., Woodward, W.R., Nishi, R. and Eckenstein, F. (1992) Localization of acidic fibroblast growth factor in specific subcortical neuronal populations. J. Neurosci. 12:4688-4700. Stocker, K.M., Sherman, L., Rees, S. and Ciment, G. (1991) Basic FGF and TGF-p\ influence commitment to melanogenesis in neural crest-derived cells of avian embryos. Development 111: 635-645. Sullivan, D.E. and Storch, T.G. (1991) Tissue- and development-specific expression of HBGF-1 mRNA. Biochim. Biophys. Acta 1090: 17-21. Sweetnam, P.M., Sanon, H.R., White, L.A., Brass, B.J., Jaye, M. and Whittemore, S.R. (1991) Differential effects of acidic and basic fibroblast growth factors on spinal cord cholinergic, GABAergic, and glutamatergic neurons. J. Neurochem. 57: 237-249. Tagashira, S., Ozaki, K., Ohta, M. and Itoh, N. (1995) Localization of fibroblast growth factor-9 mRNA in the rat brain. Mol. Brain Res. 30: 233-241. Takahashi, J.A., Fukumoto, M., Igarishi, K., Oda, Y., Kikuchi, H. and Hatanaka, M. (1992) Correlation of basic fibroblast growth factor expression levels with the degree of malignancy and vascularity in human gliomas. J. Neurosurg. 76: 792-798. Takami, K., Iwane, M., Kiyota, Y., Miyamoto, M., Tsukuda, R. and Shiosaka, S. (1992) Increase of basic fibroblast growth factor immunoreactivity and its mRNA level in rat brain following transient forebrain ischemia. Exp. Brain Res. 90: 1-10. Takami, K., Kiyota, Y., Iwane, M., Miyamoto, M., Tsukuda, K., Igarishi, K., Shino, A., Wanaka, A., Shiosaka, S. and Tohyama, M. (1993) Upregulation of fibroblast growth factor-receptor messenger RNA expression in rat brain following transient forebrain ischemia. Exp. Brain Res. 97: 185-194. Tan, E.M.L., Hoffren, J., Rouda, S., Greenbaum, S., Fox, IV, J.W., Moore, Jr., J.H. and Dodge, G.R. (1993) Decorin, versican, and biglycan gene expression by keloid and normal dermal fibroblasts: differential regulation by basic fibroblast growth factor. Exp. Cell Res. 209: 200-207. Tanaka, A., Miyamoto, K., Minamino, N., Takeda, M., Sato,

374 B., Matsuo, H. and Matsumoto, K. (1992) Cloning and characterization of an androgen-induced growth factor essential for the androgen-dependent growth of mouse mammary carcinoma cells. Proc. Natl. Acad. Sci. USA 89: 89288932. Tannahill, D., Isaacs, H.V., Close, M.J., Peters, G. and Slack, J.M.W. (1992) Developmental expression of the Xenopus int-2 (FGF-3) gene: activation by mesodermal and neural induction. Development 115: 695-702. Tardieu, M., Gamby, C , Avramoglou, T., Jozefonvicz, J. and Barritault, D. (1992) Derivatized dextrans mimic heparin as stabilizers, potentiators, and protectors of acidic or basic FGF./. Cell Physiol. 150: 194-203. Taylor, S. and Folkman, J. (1982) Protamine is an inhibitor of angiogenesis. Nature 297: 307-312. Terlau, H. and Seifert, W. (1990) Fibroblast growth factor enhances long-term potentiation in the hippocampal shce. Eur. J. Neurosci. 2: 973-977. Thomas, K.A., Ortega, S., Soderman, D., Schaeffer, M.-T., DiSalvo, J., Giminez-Gallego, G., Linemeyer, D., Kelly, L. and Menke, J. (1991) Structural modifications of acidic fibroblast growth factor alter activity, stability, and heparin dependence. Ann. N. Y. Acad. Sci. 638: 9-17. Thompson, S.A. and Fiddes, J.C. (1991) Chemical characterization of the cysteines of basic fibroblast growth factor. Ann. N. Y. Acad. Sci. 638: 78-88. Tooyama, I., Akiyama, H., Mcgeer, P.L., Hara, Y., Yasuhara, O. and Kimura, H. (1991) Acidic fibroblast growth factorlike immunoreactivity in brain of Alzheimer patients. Neurosci. Lett. 121: 155-158. Tooyama, I., Mcgeer, E.G., Kawamata, T., Kimura, H. and Mcgeer, P.L. (1994) Retention of basic fibroblast growth factor immunoreactivity in dopaminergic neurons of the substantia nigra during normal aging in humans contrasts with loss in Parkinson's disease. Brain Res. 656: 165-168. Tooriglia, A. and Blanquet, P.R. (1994) Immunohistochemical analysis of fibroblast growth factor receptor in adult retinal optic fiber and synaptic layers. Neuroscience 60: 969-981. Torriglia, A., Jeanny, J.C. and Blanquet, P.R. (1994) Immunohistochemical analysis of fibroblast growth factor receptor in bovine retina. Neurosci. Lett. Ill: 125-128. Tourbah, A., Baron-Van Evercooren, A., Oliver, L., Raulais, D., Jeanny, J.C. and Gumpel, M. (1992) Endogenous aFGF expression and cellular changes after a demyelinating lesion in the spinal cord of adult normal mice: immunohistochemical study. J. Neurosci. Res. 33: 47-59 Tsai, P.K., Volkin, D.B., Dabora, J.M., Thompson, K.C., Bruner, M.W., Gress, J.O., Matuszewska, B., Keogan, M., Bondi, J.V. and Middaugh, C.R. (1993) Formulation design of aFGF. Pharmacol. Res. 10: 649-659. Unsicker, K., Grothe, C , Westermann, R. and Wewetzer, K. (1992a) Cytokines in neural regeneration. Curr. Opin. Neurobiol. 2: 671-678. Unsicker, K., Prehn, J.H.M., Backhauss, C , Krieglstein, K., Otto, D. and Krieglstein, J. (1992b) In: J. Krieglstein and H. Oberpichler-Schwenk (Eds.), Pharmacology of Cerebral

Functions of fibroblast growth factors (FGFs) in the nervous system Ischemia 1992, Wissenschaftliche Verlagsgesellschaft, Stuttgart, pp. 229-238. Unsicker, K., Reichert-Preibsch, H., Schmidt, R., Pettmann, B., Labourdette, G. and Sensenbrenner, M. (1987) Astroglial and fibroblast growth factors have neurotrophic functions for cultured peripheral and central nervous system neurons. Proc. Natl. Acad. Sci. USA 84: 5459-5463 Unsicker, K. and Westermann, R. (1992) Basic fibroblast growth factor promotes transmitter storage and synthesis in cultured chromaffin cells. Dev. Brain Res. 65: 211-216. Unsicker, K., Grothe, C , Ludecke, G., Otto, D., and Westermann, R. (1993) Fibroblast growth factors: their roles in the central and peripheral nervous system. In: S.E. Loughlin and J. Fallon (Eds.), Neurotrophic Factors, Academic Press, New York, pp. 313-338. Vaca, K., Stewart, S.S. and Appel, S.H. (1989) Identification of basic fibroblast growth factor as a cholinergic growth factor from human muscle. J. Neurosci. Res. 23: 55-63. Vainikka, S., Partanen, J., Bellosta, P., Coulier, F., Birnbaum, D., Basilico, C , Jaye, M. and Alitalo, K. (1992) Fibroblast growth factor receptor-4 shows novel features in genomic structure, ligand binding and signal transduction. EMBO J. 11:4273-4280. Van der Wal, E.A., G6mez-Pinilla, F. and Cotman, C.W. (1994) Seizure-associated induction of basic fibroblast growth factor and its receptor in the rat brain. Neuroscience 60:311-323. Vergara, J., Medina, L., Maulen, J., Inestrosa, N.C. and Alvarez, J. (1993) Nerve regeneration is improved by insulinlike growth factor I (IGF-I) and basic fibroblast growth factor (bFGF). Restor. Neurol. Neurosci. 5: 181-189. Vescovi, A.L., Reynolds, B.A., Eraser, D.D. and Weiss, S. (1993) bFGF regulates the proliferative fate of unipotent (neuronal) and bipotent (neuronal/astroglial) EGFgenerated CNS progenitor cells. Neuron 11: 951-966. Vlodavsky, I., Bar-Shavit, R., Ishai-Michaeli, R., Bashkin, P. and Fuks, Z. (1991) Extracellular sequestration and release of fibroblast growth factor: a regulatory mechanism? Trends Biochem. Sci. 16: 268-271. Volkin, D.B., Tsai, P.K., Dabora, J.M., Gress, J.O., Burke, C.J., Linhardt, R.J. and Middaugh, C.R. (1993a) Physical stabilization of acidic fibroblast growth factor by polyanions. Arc/i. Biochem. Biophys. 300: 30-41. Volkin, D.B., Verticelli, A.M., Marfia, K.E., Burke, C.J., Mach, H. and Middaugh, C.R. (1993b) Sucralfate and soluble sucrose octasulfate bind and stabilize acidic fibroblast growth factor. Biochim. Biophys. Acta 1203: 18-26. Voulgaropoulu, F., Myers, R.L. and Chiu, I.M. (1994) Alternative splicing of fibroblast growth factor 1 (FGF-1) transcripts: a cellular dilemma in determining exon selection and inclusion. DNA Cell Biol. 13: 1001-1009. Walicke, P., Cowan, W.M., Ueno, N., Baird, A. and Guillemin, R. (1986) Fibroblast growth factor promotes survival of dissociated hippocampal neurons and enhances neurite extension. Proc. Natl. Acad. Sci. USA 83: 3012-3016. Walicke, P.A. (1988) Basic and acidic fibroblast growth fac-

S. Bieger and K. Unsicker tors have trophic effects on neurons from multiple CNS regions. J. Neurosci. 8: 2618-2627. Walter, M.A., Kurouglu, R., Caulfield, J.B., Vasconez, L.O. and Thompson, J.A. (1993) Enhanced peripheral nerve regeneration by aFGF. Lymphokine Cytokine. Res. 12: 135-141. Wanaka, A., Johnson, Jr., E.M. and Milbrandt, J. (1990) Localization of FGF receptor mRNA in the adult rat central nervous system by in situ hybridization. Neuron 5: 267-281. Wanaka, A., Milbrandt, J. and Johnson, E.J. (1991) Expression of FGF receptor gene in rat development. Development 111:455-468. Weibel, M., Fages, C, Belakebi, M., Tardy, M. and Nunez, J. (1987) Astroglial growth factor-2 (AGF2) increases atubulin in astroglial cells cultured in a defined medium. Neurochem. Int. 11: 223-228. Weibel, M., Pettman, B., Labourdette, G., Miehe, M., Bock, E. and Sensenbrenner, M. (1985) Morphological and biochemical maturation of rat astroglial cells grown in a chemically defined medium: Influence of an astroglial growth factor. Int. J. Dev. Neurosci. 3: 617-630. Weise, B., Janet, T. and Grothe, C. (1993) Localization of bFGF and FGF-receptor in the developing nervous system of the embryonic and newborn rat. /. Neurosci. Res. 34: 442-453. Weise, B., Unsicker, K. and Grothe, C. (1992) Localization of basic fibroblast growth factor in a subpopulation of rat sensory neurons. Cell Tissue Res. 267: 125-130 Wellstein, A., Zugmaier, III, G., Califano, J.A., Kern, F., Paik, S. and Lippman, M.E. (1991) Tumor growth dependent on Kaposi's sarcoma-derived fibroblast growth factor is inhibited by pentosam poly sulfate. J. Natl. Cancer Inst. 84: 716720. Werner, S., Duan, D.-S.R., DeVries, C., Peters, K.G., Johnson, D.E. and Williams, L.T. (1992) Differential splicing in the extracellular region of fibroblast growth factor receptor 1 generates receptor variants with different ligand-binding specificities. Mol. Cell. Biol. 12: 82-88. Westermann, R., Johannsen, M., Unsicker, K. and Grothe, C. (1990) Basic fibroblast growth factor (bFGF) immunoreactivity is present in chromaffin granules. J. Neurochem. 55: 285-292. Wilcox, B.J. and Unnerstall, J.R. (1991) Expression of acidic fibroblast growth factor mRNA in the developing and adult rat brain. Neuron 6: 397-409. Wilkinson, D.G., Peters, G., Dickson, C. and McMahon, A.P. (1988) Expression of the FGF-related proto-oncogene int-2 during gastrulation and neurulation in the mouse. EMBO J. 7: 691-695. Wirth, S.B., Rufer, M. and Unsicker, K. (1995) Early effects of FGF-2 on glial cells in the MPTP-lesioned striatum. Exp. Neurol. 137: 191-200. Wolswijk, G. and Noble, M. (1992) Cooperation between PDGF and FGF converts slowly dividing 02Aadult progenitor cells to rapidly dividing cells with characteristics of 02Aperinatal progenitor cells. J. Cell Biol. 118: 889-900. Woodward, W.R., Nishi, R., Meshul, C.K., Williams, T.E.,

375

Coulombe, M. and Eckenstein, F. (1992) Nuclear and cytoplasmic localization of basic fibroblast growth factor in astrocytes and CA2 hippocampal neurons. J. Neurosci. 12: 142-152. Yamada, K., Kinoshita, A., Kohmura, E., Sakaguchi, T., Taguchi, J., Kataoka, K. and Hayakawa, T. (1991) Basic fibroblast growth factor prevents thalamic degeneration after cortical infarction. J. Cerebral Blood Flow Metab. 11: 472478. Yamamoto, T., Kardami, E. and Nagy, J.I. (1991) Basic fibroblast growth factor in rat brain: localization to glial gap junctions correlates with connexin43 distribution. Brain Res. 554: 336-343. Yasuda, T., Grinspan, J., Stern, J., Franceschini, B., Bannerman, P. and Pleasure, D. (1995) Apoptosis occurs in the oligodendroglial lineage, and is prevented by basic fibroblast growth factor. J. Neurosci. Res. 40: 306-317. Yayon, A., Klagsbrun, M., Esko, J.D., Leder, P. and Ornitz, D.M. (1991) Cell surface heparin-like molecules are required for binding of bFGF to its high-affinity receptor. Ce//64: 841-848. Yazaki, N., Hosoi, Y., Kawabata, K., Miyake, A., Minami, M., Satoh, M., Ohta, M., Kawasaki, T. and Itoh, N. (1994) Differential expression patterns of mRNAs for members of the fibroblast growth factor receptor family, FGFR-1FGFR-4 in rat brain. J. Neurosci. Res. 37: 445-452. Yokoyama, M., Morrison, R.S., Black, LB. and Dreyfus, C.F. (1994) Septal neuron cholinergic and GABAergic functions: differential regulation by basic fibroblast growth factor and epidermal growth factor. Dev. Brain Res. 78: 201-209. Yoshida, K. and Gage, F.H. (1991) Fibroblast growth factor stimulates nerve growth factor synthesis and secretion by astocytes. Brain Res. 538: 118-126. Yoshida, T., Ishimaru, K., Sakamoto, H., Yokota, J., Hirohashi, S., Igarishi, K., Sudo, K. and Terada, M. (1994) Angiogenic activity of the recombinant hst-1 protein. Cancer Lett. 83: 261-268. Yoshida, T., Sakamoto, H., Miyagawa, K., Sugimura, T. and Terada, M. (1991) Characterization of the hst-1 gene and its product. Ann. N Y. Acad. Sci. 638: 27-37. Zhan, X., Hu, X., Friedman, S. and Maciag, T. (1992) Analysis of endogenous and exogenous nuclear translocation of FGF-1 in NIH 3T3 cells. Biochem. Biophys. Res. Commun. 188:982-991. Zhang, Y., Tatsuno, T., Carney, J.M. and Mattson, M.P. (1993) Basic FGF, NGF, and IGF-II protect hippocampal and cortical neurons against iron-induced degeneration. /. Cerebral Blood Flow Metab. 13: 378-388. Zhou, D. and DiFiglia, M. (1993) Basic fibroblast growth factor enhances the growth of postnatal neostriatal GABAergic neurons in vitro. Exp. Neurol. 122: 171.188. Zurn, A. (1992) Fibroblast growth factor differentially modulates the neurotransmitter phenotype of cultured sympathetic neurons. J. Neurosci. 12: 4195^201.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved. CHAPTER 15

Astroglial neurotrophic and neurite-promoting factors Hans W. Miiller, Ulrich Junghans and Joachim Kappler Molecular

Neurobiology

Laboratory,

Department

of Neurology,

1. Introduction The astrogha comprises a 'family' of related macroglial cells that express different morphological phenotypes with diverse and region-specific functions. This family of astroglial cells includes, for example, fibrous and protoplasmic astrocytes, radial glia, Golgi-Bergmann cells, pineal astrocytes, pituicytes, Gomori-positive astrocytes or perinodal astrocytes (reviewed in Fedoroff, 1990). Astroglial cells commonly express glial fibrillary acidic protein in their intermediate filaments, at least at some stage during their development. Astrocytes are present throughout the entire central nervous system (CNS), but their density varies considerably in various regions (Pope, 1978). Because of their close physical association with neurons, astrocytes and neurons are believed to be functionally interdependent cell types, assuming that astrocytes are neuron-supporting cells. Wellestablished functions of astrocytes include guidance of young neurons migrating through the developing CNS, exchange of nutritional and metabolic material with neurons and clearance or maintenance of neurotransmitter and ion concentrations in the vicinity of neurons (reviewed in Manthorpe and Varon, 1985). Astrocytes interact with neurons (and other cells) within the CNS through a multitude of molecular signals, including diffusible polypeptide factors, membraneassociated macromolecules and extracellular adhesion factors. Thus, astroglial-neuronal cell communication is rather complex. This review will focus on the astroglial cell output of neurotrophic and neurite-promoting factors.

University of Dusseldorf, D-40225 Dusseldorf,

Germany

molecules that support neuronal survival and/or neurite outgrowth. Mostly, these factors have been defined by their in vitro activity. In some instances, however, in vivo confirmation of their regulatory function on neuronal growth in development or regeneration has been demonstrated. Among known astroglia-derived neuroactive molecules are peptide growth factors, glia-derived nexin, amino acid transmitters, neuropeptides, membrane-bound cell adhesion molecules and extracellular matrix (ECM) constituents. Data on biological function(s) of these factors, as well as their regional distribution in the CNS, their target neurons, developmental regulation, responses to injury and protective/regenerative influences on the lesioned brain, will be described. 2. Astrocytes and the action of diffusible neurotrophic factors This section describes astroglia-derived diffusible neurotrophic factors and further reports on some agents that indirectly exert neurotrophic actions through stimulation of astroglial cells to release neurotrophic molecules. More extensive coverage of the specific neurotrophic factors mentioned may be found in other dedicated chapters of this volume. 2.7. Peptide growth factors 2.1.1. Nerve growth factor and other neurotrophins ^-Nerve growth factor (NGF) is the best known member of the neurotrophin family of growth fac-

378

tors (Thoenen, 1991). While it is not certain whether other neurotrophins, such as brain-derived neurotrophic factor, neurotrophin (NT)3 or NT4/5, are produced by astrocytes, expression of NGF mRNA and protein has been demonstrated in cultured astrocytes as well as in vivo (Furukawa et al, 1986; Yamakuni et al., 1987; Gonzales et al., 1990; Lu et al., 1991). Recently the expression of low levels of BDNF, NT-3 and NT-4/5 mRNA in cultured astrocytes has been reported (Condorelli et al., 1994; Leingartner and Lindholm, 1994; Moretto et al., 1994). In addition, astrocytes appear to be sites of neurotrophin action since these cells express truncated forms of the trk B and trk C receptors but no trk A receptor (Altar et al., 1994; Condorelli et al., 1994; Rudge et al., 1994). After CNS or dorsal root fiber lesions, trk A receptor expression was observed in reactive astrocytes (Foschini et al, 1994; Junier et al., 1994). The synthesis and release of NGF by astrocytes appears to be strongly regulated, possibly in relation to cell growth (Furukawa et al., 1987). Selective stimulation of choline acetyltransferase activity has been reported for fetal septal cholinergic neurons in culture and in the developing basal forebrain, as well as adult hippocampus in vivo (Gnahn et al., 1983; Hefti et al., 1984). Furthermore, the intraventricular application of NGF in adult animals can ameliorate cholinergic neuron atrophy (Fischer et al., 1987) and rescue axotomized cholinergic neurons, which normally degenerate when lacking such treatment (Hefti et al., 1989). The distribution of neurotrophin protein and mRNA in brain suggests that these growth factors are synthesized in target regions of neurotrophin projection fibers, and may support maturation and maintenance for neurotrophin sensitive neurons by receptor-mediated uptake and retrograde axonal transport. The expression of NGF and other neurotrophins in astrocytes has been found to be regulated by other growth factors and cytokines, which are known to occur in brain. These regulatory factors include acidic and basic fibroblast growth factor (aFGF and bFGF), epidermal growth factor.

Astroglial neurotrophic and neurite-promoting factors

tumor necrosis factors (TNFs), interferon gamma or interleukin (IL)-l. 2.1.2. Fibroblast growth factors bFGF and aFGF are the most prominent and best characterized members of the heparin-binding growth factor family (Gospodarowicz, 1987; Thomas, 1993). Besides their known mitogenic activity for a variety of mesodermal and neuroectoderm-derived (glial) cells, both bFGF and aFGF have been shown to enhance survival and neurite growth of central and peripheral neurons in vitro (Morrison et al., 1986; Walicke et al., 1986; Unsickeretal., 1987). Evidence supporting the relevance of FGF in vivo comes from experiments in which exogenous bFGF was administered into the rodent brain after transection of the fimbria fornix (Anderson et al., 1988; Otto et al., 1989) or after iV-methyl-4phenyl-l,2,3,6-tetrahydropyridine treatment (Otto and Unsicker, 1990). bFGF treatment rescued septal cholinergic neurons or reversed the chemical and morphological deficits in the dopaminergic nigrostriatal system, respectively, that otherwise would have occurred. While bFGF was first localized immunohistochemically in brain neurons (Pettmann et al., 1986; Janet et al., 1988), bFGF immunoreactivity was demonstrated later in astrocytes of the rat brain (Gomez-Pinilla et al., 1994), as well as in cultured astroglial cells (Araujo and Cotman, 1992). Recently, the regional distribution and developmental expression of bFGF mRNA in the mammalian CNS has been mapped by RNase protection assay (Riva and Mocchetti, 1991). The highest levels of bFGF mRNA were observed in cerebral cortex, hippocampus and spinal cord of adult rat, whereas the levels of bFGF mRNA were low in newbom rats. The time-course of bFGF mRNA expression suggests that this growth factor could be involved in the maturation and/or the maintenance of neural cells. Expression of bFGF in reactive astrocytes is upregulated by L-deprenyl, suggesting a new possibility for therapeutic interventions that increase trophic support for lesioned neurons (Biagini et al., 1994).

H. W. Mailer et al.

Although the precursor molecule does not have a classical signal peptide sequence, bFGF is released from cultured astrocytes (Hatten et al., 1988; Araujo and Cotman, 1992) by an unknown secretory pathway. Release of bFGF from intact neurons thus far has not been demonstrated. Astroglial release of bFGF is modulated by various lymphokines. bFGF can be recovered from ECM both in vivo and in vitro (Baird and Ling, 1987; Bashkin et al., 1989) after treatment with heparin or heparitinase (Saksela and Rifkin, 1990). Binding to ECM constituents may serve as a stable reservoir for bFGF, which supports long-term activities of this growth factor. bFGF interacts with both neurons and astroglial cells in culture, presumably via its highaffinity receptor (Araujo and Cotman, 1992), which was first characterized in baby hamster kidney cells by Neufeld and Gospodarowicz (1985). Acidic FGF and/or bFGF both stimulate proliferation of rabbit retinal astrocytes (Scherer and Schnitzer, 1994) and cultured fetal rat diencephalic astrocytes (Hu and Levine, 1994), respectively. Shifting the expression of the FGF receptor gene type 2 (FGFR2) to type 1 (FGFRla and FGFRl^S) is associated with malignant progression in human astrocytomas (Yamaguchi et al., 1994). Until recently, however, it was an open question whether bFGF acts directly on neurons (Walicke and Baird, 1988) or whether this interaction is mediated through glial cells (Engele and Bohn, 1991). Apparently, the interaction of this growth factor with hippocampal neurons (and astrocytes) is direct, since iodinated bFGF is internalized, intracellularly translocated and processed (Walicke and Baird, 1991). 2.1.3. Ciliary neurotrophic factor Ciliary neurotrophic factor (CNTF) is a very potent survival-enhancing molecule for a variety of neuronal cell populations. The trophic activity initially described for CNTF was confined to the peripheral nervous system (PNS), where CNTF acts on parasympathetic, sympathetic and some sensory neurons (Manthorpe and Varon, 1985). More recently, effects of CNTF on neurons and glial cells in the CNS have been described. CNTF

379

supports survival of developing motor neurons and hippocampal neurons in culture (Arakawa et al., 1990; Ip et al., 1991) and prevents lesion-induced degeneration of motor neurons after axotomy (Sendtner et al., 1990). However, the spectrum of biological activities of CNTF is much broader. CNTF reportedly blocks the division of chick sympathetic precursor cells (Ernsberger et al., 1989), reduces tyrosine hydroxylase activity in cultured sympathetic neurons and, conversely, increases choline acetyltransferase activity in these cells (Saadat et al., 1989). CNTF is a pleiotropic modulator of development in the 0-2A glial cell lineage. This factor promotes oligodendrocyte generation, maturation and survival. However, in the presence of extracellular matrix derived e.g. from endothelial cells, CNTF promotes the differentiation of 0-2A progenitors into type-2 astrocytes (Hughes et al, 1988; Mayer et al., 1994). A sophisticated comparison of sequence data and secondary structure predictions placed CNTF into the *four-helix bundle' family of growth factors (Bazan, 1991). This family comprises cholinergic differentiation factor (identical to leukemiainhibiting factor), IL-6 and a number of other regulators of hematopoiesis. Consistently, neurotrophic effects of lymphokines, such as IL-6, have been reported (Hama et al., 1989; Kushima et al., 1992). Cell contact and exogenous CNTF regulate CNTF mRNA expression in astroglial and glioma cells in vitro (Meyer and Unsicker, 1994). CNTF gene expression can also be stimulated pharmacologically by the administration of R(-)-deprenyl in process-bearing astrocytes after mechanical wounding in vitro suggesting that the rescue of e.g. axotomized neurons is mediated through reactive astrocytes (Seniuk et al., 1994). Regional distribution and rather late developmental expression of CNTF in the rodent CNS suggests that CNTF may not act as a target-derived neurotrophic factor. CNTF immunoreactivity and mRNA were primarily detected in optic nerve, olfactory bulb and spinal cord (Stockli et al., 1991; Dobrea et al., 1992). In other brain regions, both CNTF protein and mRNA levels appear to be much lower. In optic nerve and olfactory bulb,

380

expression of CNTF is confined to a subpopulation of (type-1) astrocytes (Stockli et al., 1991), whereas in spinal cord, branched oligodendrocytelike cells seem to express CNTF (Dobrea et al, 1992). Interestingly, other members of the fourhelix bundle family of growth factors may be produced by astrocytes (i.e. IL-6) in response to IL-1 or TNFa stimulation (Benveniste et al., 1990; Aloisi et al., 1992; Lee et al., 1993). The receptor for CNTF has recently been cloned (Davis et al., 1991) and appears to be exclusively expressed in the nervous system and in skeletal muscle. In contrast to other known receptors, the receptor for CNTF is anchored to the plasma membrane by a glycosyl-phosphatidylinositol linkage (Davis et al., 1991). Its primary structure is most similar to the IL-6 receptor. Furthermore, both CNTF and the structurally related leukemia-inhibitory factor (see above) use the IL-6 signal transducer gpl30 (Ip et al., 1992). These observations raise the possibility that the receptors for CNTF, leukemia-inhibitory factor and hematopoietic cytokines are able to interact with each other and to activate related signaling pathways in diverse cell types (Bazan, 1991; Davis and Yancopoulos, 1993). It has been shown by Lillien et al. (1990) that some biological functions of CNTF, e.g. induction of type-2 astrocyte development, require cooperation with as yet unknown ECM-associated molecule(s). 2,1,4. Insulin and insulin-like growth factors The insulin-like growth factors (IGF I and II) were identified by the molecular analysis of such different biological processes as cartilage growth (somatomedins), multiplication-stimulating activity in cell culture and glucose-uptake in adipose tissue (non-suppressible insulin-like activity) as reviewed by Hepler and Lund (1990) and Wozniak et al. (1993). cDNAs for both IGFs have been cloned, and, at the genomic level, the use of alternative promoters was reported. The biological effects of these insulin homologs are transduced via a tyrosine kinase receptor (IGF I receptor), which is similar to the insulin receptor, and by the structurally unrelated G-protein-coupled IGF II receptor, respectively. This family of growth factors

Astroglial neurotrophic and neurite-promoting

factors

exerts important effects on neurons in vitro: insulin in high concentrations is required for the survival of neurons in serum-free media and may be replaced by IGF I in lower concentrations (RecioPinto and Ishii, 1984; Aizenmann and de Vellis, 1987). Moreover, insulin and IGF II possess a neurite-promoting activity (Collins and Dawson, 1983; Recio-Pinto et al., 1986). Recently, insulin transcripts were detected in the rabbit brain (Devaskar et al., 1994). Cerebral synthesis of IGF I and II, has been demonstrated using in situ hybridization. IGF I mRNA is expressed abundantly in the developing brain of rat, largely confined to large projection neurons of maturing sensory and cerebellar relay systems and non-pyramidal cortical cells (Rotwein et al., 1988; Bondy, 1991), pointing to an involvement of IGF I in neuronal differentiation. In primary cultures derived from embryonic brain, IGF I mRNA was detected in neurons and glial cells (Rotwein et al., 1988; Adamo et al., 1988). In astrocyteconditioned medium, an insulin-like factor with neurotrophic activity could be detected using indirect immunological and receptor-binding criteria, but the molecular identity of the factor has not been elucidated (Kadle et al., 1988). Apparently, astroglial IGF I is involved in the regulation of myelination. IGF I promotes proliferation and oligodendroglial commitment in glial progenitor cells (McMorris and DuboisDalcq, 1988). In animal models for central deand remyelination (cuprizone-treated mice or experimental autoimmune encephalomyelitis in Lewis rats) upregulation of astroglial IGF I mRNA and protein was observed during early recovery (Komoly et al, 1992; Liu et al., 1994). Concomitantly, the IGF I receptor mRNA was transiently increased in oligodendrocytes (Komoly et al., 1992). Similarly, the astroglial IGF I expression is increased after experimental hypoxicischemic injury in rats. Ventricular injection of IGF I could markedly reduce neuronal loss after injury (Gluckman et al, 1992). Since IGF I stimulates the proliferation of astrocytes in vitro (Tranque et al., 1992), a positive feedback regulation of these astroglial rescue mechanisms may exist.

H.W.

Maileretal.

IGF II synthesis in the CNS is largely restricted to the choroid plexus and the leptomeninges, with higher levels in the fetus (Hynes et al., 1988). IGF II immunoreactivity is found in the cerebrospinal fluid, the anterior pituitary and hypothalamic regions (Haselbacher et al., 1985). IGF II mRNA is present in glial, but not in neuronal cell cultures (Rotwein et al., 1988). 2.1.5. Transforming growth factor jS Transforming growth factor p (TGF^) was initially characterized as a promoter of transformed phenotype in fibroblasts. Other biological effects of TGFyS are inhibition of cell proliferation, induction or inhibition of differentiation (depending on the cell type) and immunosuppression. This growth factor is a member of a larger family of homologous growth, morphogenesis and differentiation factors, which comprises TGF^2, TGF)83, inhibins, activins, Mlillerian-inhibiting substance, decapentaplegic product and others (for a review, see Massague, 1990). Interestingly, X-ray analysis of TGF^2 crystals revealed a structure of three intramolecular disulfide bonds (Schlunegger and Griitter, 1992), which is highly reminiscent of the 'cystine knot' in NGF (McDonald and Hendrickson. 1993). Several TGF^S-binding proteins have been found, and for two of them (receptor II and receptor III, ^-glycan), cDNAs have been cloned. Receptor II codes for a serine-threonine kinase, whereas y3-glycan is a proteoglycan (PG) (Massague, 1992). Neurotrophic effects have been described for several members of the TGF^S family in differing systems: TGF^Sl is a survival factor for embryonic motoneurons (Martinou et al., 1990). TGF^Sl and TGF/?2 increase neuronal survival and synthesis of substance P in sensory neurons (Chalazonitis et al., 1992). TGFySl and TGFy32 enhance the gliamediated stimulation of neurite outgrowth of dorsal root ganglion cells due to expression of the neural adhesion molecule LI (Saad et al., 1991), an effect that is mediated via stimulated NGF production in immature astrocytes. On the other hand, Flanders et al. (1991) reported a decrease in survival of embryonic chick ciliary ganglion cells, when TGFy82 and 3 were added to the cultured

381

cells. Neuron-like differentiation has been observed with activin in pi9 teratoma cells (Schubert et al., 1990) and with bone morphogenetic protein 2B in pheochromocytoma PCI2 cells (Paralkar et al., 1992). TGF)8 and related bone morphogenetic proteins (BMPs) induce differentiation and prevent cell death in astrocyte lineage cells (D'Alessandro et al., 1994; D'Alessandro and Wang, 1994). TGF)81, 2 and 3 were localized by immunohistochemistry in the developing nervous system of the mouse. While TGF^l is mainly found in neural crest-derived cells (e.g. meninges; Heine et al., 1987), TGF)82 and 3 co-localize in neuronal pericarya and axons, as well as in glial cells (Flanders et al., 1991). Astrocytes express TGF^l, 2 and 3 mRNA in culture (Constam et al., 1992; MorgantiKossmann et al., 1992) and release TGF^2 into the medium. TGF^l and 3, however, are present in the inactive latent form (Constam et al., 1992). After cortical lesions in adult rat there is an upregulation of TGF^l mRNA in astrocytes and microglia (Lindholm et al., 1992). TGF^l has significant effects not only on neurons, but also on astrocytes. Similarly, as in other cell types, the production of ECM molecules (laminin, fibronectin and collagen) is up-regulated (Toru-Delbauffe et al. 1992), whereas cell proliferation is inhibited. Interestingly, TGF^l stimulates the synthesis of TGFjSl mRNA in astrocytes (Morganti-Kossmann et al., 1992), indicating a positive feedback regulation of astrocytic TGF^. Furthermore, TGF/? inhibits expression of tumor necrosis factor alpha (TNFa) and stimulates interleukin-6 expression in cultured rat primary astrocytes, suggesting that TGF^ is a regulator of astroglial cytokine production under inflammatory conditions in brain (Benveniste et al., 1994). 2.2. Neurotrophic action of neurotransmitters via astrocytes 2.2.1. Norepinephrine and nerve growth factor Activation of the )8-adrenergic receptor through norepinephrine or isoproterenol stimulates C6 glioma cells and astrocyte cultures to express NGF and BDNF (Schwartz and Costa, 1977; Schwartz 1988; Schwartz and Nishiyama, 1994). The same

382

mechanism of regulation could be demonstrated in cultured type-1 astrocytes from cortex, striatum and cerebellum (Schwartz and Mishler, 1990). Furthermore, astrocytic expression of ^-adrenergic receptors has been demonstrated in vivo (Aoki et al., 1987). 2.2.2. Serotonin and SI 00/3 Serotonin (5-hydroxytryptamine, 5-HT) has important effects on the proliferation of different cell populations. 5-HT is detectable early in ontogeny (Shuey et al., 1990). Pharmacologically induced depletion of serotonin in embryogenesis causes suppression, delayed onset or prolongation of neurogenesis in brain areas that receive serotonergic projections (Lauder, 1990). Apparently, the effects of 5-HT are critically concentrationdependent. Low concentrations activate neuronal Sib receptors to inhibit neuronal differentiation and neurite outgrowth (Whitaker-Azmitia and Azmitia, 1986). At higher concentrations, a glial Sja receptor is activated that stimulates astrocytes to produce S 100)8. The disulfide-bonded dimer of this member of the SI00 family of calcium-binding proteins has neurite-extension activity (Kligman and Marshak, 1985). S 100)8 is found in large amounts in astrocytes. Its synthesis is regulated developmentally, and secretion from proliferating astrocytes has been observed during the time of neurite outgrowth of cortical neurons (for a review see Marshak, 1990). These findings suggest regulation of serotonergic neuron growth via a neuronastrocyte-neuron feedback loop (Leslie, 1993). This concept is supported by the observation that SI00)8 deficiency in the astrocytes of retarded Pdn mutant mice causes a developmental defect of serotonergic fibers (Ueda et al., 1994a,b). On the other hand, SI00)8, in the nanomolar range, has been shown to stimulate proliferation of primary astrocytes (Selinfreund et al., 1991). The gene for human SI00)8 is localized on chromosome 21. In Alzheimer's disease and Down's syndrome the regulation of SI00)8 synthesis seems to be disturbed. SI00)3 overexpression of 10- to 20-fold was found in the temporal lobe of brains from Alzheimer's disease patients (Marshak et al., 1992). The cells containing SI00)8 immunoreac-

Astroglial neurotrophic and neurite-promoting

factors

tivity were reactive astrocytes. This finding raises the possibility that SI00)8 contributes to the formation of neuritic plaques. Dunn et al. (1987) reported that the expression of SI00)8 mRNA is inhibited by colchicine and vinblastine in C6 glioma cells. Based on the latter observation, the design of drugs that inhibit the synthesis of SI00)8 may be feasible. 2.2.3. Vasoactive intestinal peptide Vasoactive intestinal peptide (VIP) is a neuroactive 29-amino acid polypeptide, structurally related to glucagon (for a review see Said, 1991). This neuropeptide is released by preganglionic sympathetic neurons, along with acetylcholine, and regulates mitosis, differentiation and survival of cultured sympathetic neuroblasts (Pincus et al., 1990). These direct effects are mediated by a lowaffinity receptor involving adenylate cyclase. Additionally, VIP promotes survival of sensory neurons that are cultivated in the presence of tetrodotoxin, a blocker of fast sodium channels. The latter VIP effect appears to be mediated by astrocytes (Brenneman et al., 1987). VIP stimulates the proliferation of astrocytes and releases a yet unknown neurotrophic factor. While the effects of VIP have been investigated intensively, the proposed astroglial factor still remains elusive. 2.2.4. Perspective: 'gliotransmitters' as possible neurotrophic agents For a growing number of neurotransmitters, direct neurotrophic actions have been reported (for a review see Schwartz, 1992; Schwartz and Taniwaki, 1994). These transmitters are serotonin, acetylcholine, norepinephrine, glutamate and endogenous opioid peptides. Some of these neurotrophic transmitters may also be produced by astrocytes. The family of neurotransmitters synthesized by astrocytes comprises y-aminobutyric acid, glutamate, proenkephalin, neuropeptide Y, somatostatin and others. Martin (1992) has coined the term *gliotransmitter' for such substances. The role of *gliotransmission' in development and function of the mature nervous system has not been firmly established yet. It can be anticipated, however, that neurotrophic activity of astroglia-derived transmit-

H.W. Mailer

etal

ters will play an important role in neuron-glia interactions. Monard and colleagues have characterized a 43kDa glia-derived neurite-promoting factor that inhibits serine proteases (Monard et al., 1973; Guenther et al., 1985). cDNA cloning revealed homology to members of the serpin superfamily of protease inhibitors, e.g. antithrombin III (Sommer et al., 1987). It was postulated that the regulation of proteolytic activity is crucial for the development of growth cones (Monard, 1988). Hirudin, a thrombin inhibitor, mimics the neurite-promoting action of glia-derived nexin (GDN) in cultures of hippocampal neurons, indicating that inhibition of the protease is required for the neurotrophic effect (Farmer et al., 1990). In the adult nervous system, GDN is mainly present in the olfactory bulb, a region where continuous replacement of neurons takes place. Transient forebrain ischemia with selective degeneration of hippocampal CAl pyramidal cells entails long-lasting astroglial production of GDN, predominantly in the vicinity of blood vessels (Hoffmann et al., 1992), suggesting that GDN could be involved in post-lesional repair. In the PNS, GDN mRNA is up-regulated after nerve lesions (Meier et al., 1989). The therapeutic potential of GDN remains to be evaluated. 3. Astroglial membrane-bound molecules supporting neurite groM^th and cell adhesion For many types of neurons, astrocytes provide a suitable substrate for adhesion and neurite outgrowth in vitro (Fallon, 1985; Miller and Smith, 1990; Stichel and Muller, 1992), and for neuronal cell migration in vivo (Hatten, 1990). Using welldefined primary cultures from rat brain, a set of molecular and cellular influences have been defined which seem to be required for long-term survival of brain neurons in vitro (Schmalenbach and Muller, 1993). In the latter study, it was shown that astrocytes provide the full complement of the essential neuron-supporting influences, including diffusible neurotrophic factors, ECM-glycoproteins and cell-bound signals (Schmalen-bach and Muller, 1993).

383

Neural recognition molecules, also called cellcell adhesion molecules (CAMs), are involved in neuron-neuron and neuron-glia interactions, e.g. neuronal migration and neurite outgrowth (Doherty and Walsh, 1989). In general, CAMs can be divided into: (1) molecules of the immunoglobulin superfamily (Ig-family), which operate in a calcium-independent fashion; (2) cadherins (CADs), which bind in a calcium-dependent manner; and (3) other molecules (see below; reviewed in Edelman and Crossin, 1991). Well-studied CAMs of the Ig-family in the nervous system are the neural cell adhesion molecule (N-CAM) and the neuron-glia cell adhesion molecule (LI or NILE). Both molecules are found on neurons and have been suggested to mediate neuron-glia interactions by heterophilic binding mechanisms. Currently, it is unknown whether astroglia express N-CAM and/or LI in vivo (Persohn and Schachner, 1987; Bartsch et al., 1989), but Saad et al. (1991) proved that both CAMs are expressed by astrocytes in vitro. Interestingly, astrocyte-derived TGF)S2 up-regulates LI (via NGF) and down-regulates N-CAM expression by immature, but not mature, astrocytes (see also Section 2.1.5). The neuronal protein Thy-1, the smallest member of the Ig-family, represents a major fraction of the surface membrane protein on long axons. Thy1 inhibits neurite outgrowth on mature astroglia and, thus, may stabilize neuronal connections and suppress axonal sprouting after injury (Tiveron et al., 1992). The astroghal-binding partner of Thy-1, however, is still unknown. Neural CAD (N-CAD) is located both on glial and neuronal cells and binds in a homophilic, calcium-dependent manner (Geiger and Ayalon, 1992). Antibodies against N-CAD inhibit the growth of retinal ganglion cell neurites on astrocytes in culture. Moreover, transfection of fibroblasts with a cDNA for N-CAD converts them from neurite-inhibitory into neurite growthstimulating cells (Takeichi, 1990). Other molecules involved in glia-neuron interactions belong to none of the above-mentioned protein families, e.g. adhesion molecule on glia (AMOG) and Astrotactin. The calcium-

384

independent AMOG is expressed by astrocytes, but not by neurons (Antonicek et al., 1987). AMOG mediates selective neuron-astrocyte interaction. Molecular cloning revealed that AMOG is homologous to the ^ subunit of the Na, K-ATPase. It was speculated that AMOG links cell adhesion with ion transport (Gloor et al., 1990; Schmalzing et al., 1992). Astrotactin, a recognition molecule expressed exclusively by neurons, is involved in CNS neuronal migration on astrocyte surfaces. Astrotactin binds to an as yet unidentified astroglial molecule (Fishell and Hatten, 1991). Recently, a new 135 kDa astroglial surface adhesion protein has been discovered that might play a role in astroglial-guided neuronal migration (Mittal and David, 1994a,b). 4. Astrogiia-derived extracellular matrix molecules that influence adhesion, migration and process formation of neurons 4.1. Laminin Laminin (see also chapter by Nurcombe in this volume), the most abundant non-collagenous protein in basement membranes, was purified initially from mouse Engelbreth-Holm-Swarm (EHS) tumor (Timpl et al., 1979) and a mouse carcinoma cell line (Chung et al., 1979). This multidomain glycoprotein (molecular mass -800 kDa) is composed of the three genetically distinct polypeptide chains A (400 kDa), Bl (215 kDa) and B2 (205 kDa), which are connected by disulfide bonds and associate in a cross-shaped structure. All subunits of mouse laminin have been cloned and entirely sequenced (for reviews see Beck et al, 1990; Paulsson, 1992). Using multiple approaches, active sites for cell adhesion and for neurite outgrowth have been identified. One site for cell adhesion (but not neurite outgrowth) is located on the Bl chain (sequence YIGSR) near the point of chain intersection. A cell attachment site, together with neurite-promoting and heparin-binding activity, has been mapped to the distal part of the long arm. In addition to the ^classical' EHS-tumor laminin, isoforms have been identified at the protein and cDNA levels (Hunter et al., 1989; Ehrig et

Astroglial neurotrophic and neurite-promoting factors

al., 1990). One of them, 5-laminin, is homologous to the Bl chain (designated BI5), whereas another, merosin, turned out to be an A chain isoform (designated Am). B\s and Bl, as well as Am and A chains, are mutually exclusive in formation of heterotrimers. Merosin synthesized, for example, by Schwann cells (glial cells of the PNS) is a strong promoter of neurite outgrowth (Calof and Lander, 1991; Engvall et al., 1992). In brain, laminin is transiently expressed during development. There is clear evidence that glial cells are the source of secreted laminin (Sanes, 1989). The punctuate extracellular deposits of laminin in vivo appear in regions with neuronal migration and nerve pathway formation (Liesi and Silver, 1988; McLoon et al., 1988; Liesi and Risteli, 1989). The expression pattern of laminin in the developing CNS is consistent with functions of this glycoprotein established in vitro, e.g. promotion of neurite extension, cell adhesion and migration (Sanes, 1989; Sephel et al, 1989; Bates and Meyer, 1994). Apparently, glial laminin is a variant of the originally described EHS-laminin: studies by several laboratories have confirmed the observation that glial laminin lacks the A chain (Liesi and Risteli, 1989; Matthiessen et al, 1989). Furthermore, astroglial laminin is associated in a high molecular weight complex with heparan sulfate PG (Matthiessen et al, 1989; Muller et al., 1990, 1991). While the formation and long-lasting persistence of a dense plexus of laminin-immunopositive blood vessels has been observed at the lesion site of a transected fiber tract in the adult rat brain, laminin immunoreactivity is not re-expressed in reactive astrocytes following CNS lesions (Stichel and Muller, 1994). It has been suggested that even the sole expression of laminin B2 chain by cultured astrocytes may be sufficient to stimulate neurite outgrowth of PC12-cells (Wujek et al, 1990). Furthermore, it was demonstrated that ^-laminin is present in developing rat brain and that cultured glia is capable of synthesizing and assembling 'brain-like' 5-laminin into ECM (Chiu et al., 1991; Hunter et al., 1992). The functional implications of the established glial laminin isoforms are still unclear.

385

H.W. Mulleretal

4.2. Fibronectin Fibronectin is a dimeric ECM and plasma glycoprotein, which plays a central role in cell adhesion (reviewed in Ruoslahti, 1988; Muller 1993; see also chapter by Newgreen and Tan in this volume). The primary sequence of fibronectin determined from protein and cDNA, shows a high degree of homology between different species. Nevertheless, comparison of fibronectins from various tissues reveals remarkable differences in their primary structure, due to alternative splicing of a common mRNA precursor transcribed from a single gene. In general, the fibronectin polypeptide is composed of three different kinds of repeated sequences (type I, II and III) assembled into a series of structural domains with distinct binding activities towards collagen, sulfated glycosaminoglycans, fibrin and cell surface receptors of the integrin family. Two polypeptide chains are disulfide bonded near their carboxy-terminus to form the mature fibronectin dimer. As detected by immunohistochemical means, fibronectin is expressed transiently during embryogenesis of the CNS (Rogers et al., 1989; Bixby and Harris, 1991). It is distributed along radial glia in the developing cortex and is thought to be involved in cell adhesion, migration and differentiation (Hatten et al., 1982; Stewart and Pearlman, 1987; Sanes, 1989; Sheppard et al., 1991). So far, the cell type responsible for fibronectin synthesis in the brain is not known, but it was demonstrated in vitro that fibronectin is produced by astrocytes (Price and Hynes, 1985; Liesi et al., 1986). Furthermore, Matthiessen et al. (1989) showed that fibronectin secreted by astrocytes induced neurite growth in cultured hippocampal neurons. Pagani et al. (1991) examined the expression of the known fibronectin mRNA splicing variants during development and aging. As a result, the expression of the V25 alternative splicing variant, which contains a segment responsible for cell migration (Dufour et al., 1988) and binds to a^i integrin (Reichardt and Tomaselli, 1991), was higher in all fetal tissues, as compared with adult tissues, including brain. On the other hand, during aging, a decrease of the EIIIA variant

was observed in brain. The functional significance of the differential expression of the alternative splice forms of fibronectin is not yet understood. 4.3. Tenascin Tenascin, a large extracellular hexameric glycoprotein, is prominently expressed in embryonic tissues during development and growth, as well as in tumors, whereas in normal adult tissues, it is present in very restricted locations (for a review see Erickson and Bourdon, 1989; Faissner et al., 1994). Due to its discovery in different species and tissues by several independent laboratories, tenascin (Chiquet-Ehrismann et al., 1986) is also known as glioma mesenchymal ECM protein (Bourdon et al., 1983), myotendinous antigen (Chiquet and Fambrough, 1984a,b), hexabrachion protein (Erickson and Inglesias, 1984), cytotactin (Grumet et al., 1985) and Jl-220/200 (Kruse et al., 1985; Faissner et al., 1988). Molecular analysis revealed that a tenascin monomer consists of a cysteine-rich aminoterminal region followed by approximately 14 epidermal growth factor-like repeats, at least eight fibronectin type III homologous repeats and a fibrinogen homologous domain at the carboxyterminal. In all species thus far examined (chicken, mouse, human), a tenascin pre-mRNA is transcribed from a single gene and subjected to alternative splicing. This process leads to insertions of additional fibronectin type III homologous repeats into the respective domain, with the result of multiple mRNA species and monomer-polypeptides of different lengths (Jones et al., 1989; Spring et al, 1989; Nies et al., 1991; Siri et al., 1991; Weller et al., 1991). Tenascin monomers are assembled into hexamers by disulfide bridges at their aminoterminal ends and form hexabrachia, as demonstrated by rotary shadowing electron microscopy (Erickson and Inglesias, 1984). Tenascin appears early in neural development and is expressed by glial cells in vitro (Kruse et al., 1985; Faissner et al., 1988) and in vivo (Prieto et al, 1990; Tsukamoto et al., 1991; Tucker and McKay, 1991; Bartsch et al., 1992). It is involved in neuronal migration, e.g. cerebellar granule cell migration (Chuong et al., 1987). During the period

386

of afferent fiber ingrowth, tenascin is expressed transiently by immature-astrocytes in boundaries of vibrissae-related barrel fields in the somatosensory cortex (Steindler et al., 1989a,b; Crossin et al., 1989; Jhaveri et al., 1991; Mitrovic et al., 1994). It has been demonstrated that tenascin mediates neurite elongation of a variety of different nerve cell types when offered as a substrate (Chiquet, 1989; Wehrle and Chiquet, 1990; Lochter et al., 1991). In short-term binding assays, tenascin mediates attachment of neurons to astrocytes (Kruse et al., 1985; Grumet et al., 1985). However, tenascin was also found to be a repulsive substrate for CNS neurons (Faissner and Kruse, 1990) and neurites (Crossin et al., 1990). Exposed to cells in a soluble form, tenascin inhibited neurite outgrowth (Crossin et al., 1990; Lochter et al., 1991). It is thought that the observed contradictory biological activities of tenascin reside in different functional domains located in the fibronectin type III homologous regions of the molecule. Distinct domains involved in neurite outgrowth promotion, cell binding, cell repulsion and granule cell migration have been defined (Chiquet-Ehrismann et al., 1986; Spring et al, 1989; Lochter et al., 1991; Husmann et al., 1992), and it will be interesting to see which signal transducing mechanisms are activated in the different cellular responses. Following trauma to the adult human cerebral cortex, discrete populations of reactive astrocytes upregulate tenascin expression (Brodkey et al., 1995). After postcommissural fornix transection (Lips et al., 1995) and after optic nerve lesion (Ajemian et al., 1994) in adult rat, induced astroglial expression of tenascin is restricted to the border zone of the lesion. It is interesting to note that the local distribution of tenascin immunoreactivity at the lesion site of the transected fornix and the sprouting pattern of the injured axons as shown by Lips et al. (1995) is not compatible with a previously suggested putative axon growth inhibitory function of tenascin in vivo (McKeon et al., 1991). Furthermore, in the tenascin knockout mouse glial scar formation is not altered as compared to wildtype animals (Steindler et al., 1995).

Astroglial neurotrophic and neurite-promoting

factors

4,4. Proteoglycans PCs are proteins that carry at least one covalently bound glycosaminoglycan side chain. The four known types of glycosaminoglycans-heparan sulfate/heparin (HS), chondroitin sulfate/dermatan sulfate (CS/DS), keratan sulfate, hyaluronic acid (HA) are composed of repeating disaccharide units. HS, DS/CS and keratan sulfate are proteinbound glycosaminoglycans, and they all contain sulfate, whereas HA exists as a free glycosaminoglycan and lacks sulfate (reviewed in Gallagher, 1989; Ruoslahti 1989; Ruoslahti and Yamaguchi, 1991; Kjellen and Lindahl, 1991). PGs have different types of core proteins reflecting the various locations and functions of these molecules. Many PGs are constituents of the ECM, some have membrane-embedded core proteins or can be found in intracellular granules. The functions of PGs are diverse: they can bind ECM components, mediate the binding of cells to the matrix and capture soluble molecules, such as growth factors, into the matrix and at cell surfaces. PGs are capable of interacting with other molecules, either by the glycosaminoglycan component or by the core protein. CS-PGs are the most prominent PGs in the brain (Margolis and Margolis, 1989, 1993). During development from immature to adult brain, their localization in neurons and astrocytes changes from predominantly extracellular to exclusively intracellular (Aquino et al., 1984a,b). Gowda et al. (1989) reported the expression of at least six distinct core proteins for CS-PGs. Glial precursor cells in the optic nerve and cerebellum express a membrane-bound CS-PG called NG2. It is maintained in the astrocyte lineage in cell culture, but disappears from oligodendrocytes (Stallcup and Beasley, 1987). Streit et al. (1990) isolated a soluble CS-PG from mouse brain and cultured astrocytes, called astrochondrin (Streit et al., 1993). Astrochondrin carries the L2/HNK-1 and L5 carbohydrate epitopes shared by several adhesion molecules of the Ig-family (reviewed in Schachner, 1989), is transiently expressed and interacts with ECM components, astrocytes and cerebellar granule cells during development.

H.W. Maileretal.

Soluble CS-PG and DS-PG, HS-PG isolated from astrocytes (or the EHS-tumor), as well as individual glycosaminoglycans, strongly enhance embryonic rat neurite outgrowth in culture (LaFont et al., 1992). Interestingly, DS preferentially stimulates dendritic growth, while CS and HS selectively stimulate axonal outgrowth. HS-PG produced by astrocytes and subsequently bound to laminin, but not CS- or DS-PG, enhanced neurite outgrowth by chick sensory neurons to above the outgrowth level that occurs on laminin alone (Johnson-Green et al., 1992). This result is in agreement with earlier observations claiming that neurite outgrowth activities found in conditioned media of various cell types, including astrocytes, involve an HS-PG non-covalently bound to laminin (Lander et al., 1985; Matthiessen et al., 1989). Ard and Bunge (1988) demonstrated that HS-PG and laminin are co-localized on the astrocyte surface in vitro and noted that growing neurites contact areas where these complexes are deposited. Matthew et al. (1985) showed that an antibody to an HS-PG-laminin complex blocks neurite outgrowth in vitro. Moreover, in vivo, this antigen functions in axonal regeneration (Sandrock and Matthew, 1987), as well as in neural crest cell migration (Bronner-Fraser and Lallier, 1988). Faissner et al. (1994) have isolated a CS/DS-PG named DSD-1 from postnatal mouse brain that is expressed by astrocytes and oligodendrocytes and promotes neurite outgrowth of embryonic mesencephalic and hippocampal neurons. PGs may also inhibit neurite outgrowth (for a review see Letoumeau et al., 1992). Following injury to adult rat cerebral cortex, gliotic tissue of the lesion area did not support neurite outgrowth in vitro, but expressed CS-PG and tenascin. However, glial tissue from the lesion of neonatal rats, lacking CS-PG and tenascin, supported neurite outgrowth. These data indicate that CS-PG and/or tenascin may limit regeneration of CNS axons after injury (McKeon et al., 1991). In response to small cerebellar puncture lesions, the mRNA and immunoreactivity of the NG2 proteoglycan was upregulated, suggesting that this CS-PG may contribute to the failure of damaged CNS axons to regenerate. On the other hand, following post-

387

commissural fornix transection in the adult rat, regenerating axons traversed the zone of enhanced CS-PG immunoreactivity at the lesion site (Lips et al., 1995) indicating that CS-PG, at least in the latter lesion paradigm, does not inhibit the advance of sprouting fornix axons. Recently biglycan, a small chondroitin sulfate proteoglycan, was isolated from conditioned medium of cerebral astrocyte cultures (Koops et al, 1995). Besides stimulating neurite growth, this protoglycan markedly enhances the survival but not the adhesion of cultured brain neurons at nanomolar concentrations (Junghans et al., 1995) and facilitates learning following injection into the vicinity of the nucleus basalis magnocellularis of adult rat (Hasenohrl et al, 1995). The expression of the large hyaluronate-binding PG of fibroblasts, versican (Zimmermann and Ruoslahti, 1989), was recently found in rat brain, presumably in astrocytes (Bignami et al., 1993). Versican is thought to bridge between hyaluronate and cells (LeBaron et al., 1992). The molecule is distributed in white matter in a pattern very similar to the glial hyaluronic acid-binding protein (GHAP). GHAP was isolated from brain white matter and is devoid of glycosaminoglycan (Perides et al., 1989). A partial amino acid sequence from GHAP showed almost complete identity with the sequence predicted from a cDNA for the versican Nterminal domain, and it wa^^ suggested that GHAP is a proteolytically processed form of versican (Zimmermann and Ruoslahti, 1989). Another proteoglycan of the aggrecan/versican family, named brevican was found to be expressed in primary cerebellar astrocyte cultures and was subsequently cloned (Yamada et al., 1994). Like other members of the family, it contains a hyaluronic acid-binding domain in its N-terminal region, an EGF-like repeat, a lectin-like and a complement regulatory protein-like domain in its C-terminal region. Since a significant amount of brevican is devoid of glycosaminoglycan, it is considered a 'part-time' proteoglycan. Furthermore, a brain-unique CS-PG, neurocan, was cloned and characterized (Ranch et al., 1992) and the developmentally regulated expression of two variants of this PG were described (Oohira et

388

al., 1994). The spatio-temporal expression pattern suggests that neurocan plays some role in forming early cortical afferent pathways and functional barrel structures in the somatosensory cortex. The recently identified soluble rat brain CS-PG phosphacan (3F8-PG) exhibits striking sequence similarity with the extracellular domain of the receptor type protein tyrosine phosphatase beta (RPTP)8; Grumet et al., 1993; Barnea et al, 1994). Phosphacan is synthesized by glia, inhibits NgCAM-mediated neuronal and glioma cell adhesion and neurite outgrowth (Milev et al., 1994) suggesting that it may play a role in modulating neuronal and glial adhesion, neurite growth and possibly signal transduction in the CNS. Keratan sulfatecontaining glycoforms of phosphacan (phosphacan-KS) are also present in brain (Grumet et al., 1994). The carbonic anhydrase domain of RPTPjS is a functional ligand for the axonal cell recognition molecule contactin (Peles et al., 1995). In addition, several unidentified new proteoglycans of the HS- or Keratan sulfate (KS-) type have been detected in the mammalian CNS. A potentially fibronectin-binding 150-165 kDa HS-PG was localized in a subset of astrocytes of spinal cord (Guiseppetti et al., 1994). A KS-PG with inhibiting activity for neuronal attachment and neurite growth in vitro, associated with astrocytes, marks the boundaries of major functional subdivisions in the lateral geniculate nucleus of the developing ferret (Robson and Geisert, 1994). 4.5. Interaction of proteoglycans with cell adhesion molecules The brain CS-PGs neurocan and phosphacan were recently shown to interact with cell adhesion molecules. The localization of neurocan- and phosphacan-immunoreactivities overlap extensively with that of tenascin, Ng-CAM and N-CAM in embryonic rat spinal cord and early postnatal rat cerebellum, respectively (Grumet et al., 1994; Friedlander et al., 1994; Milev et al., 1994). Neurocan and phosphacan interact with N-CAM, NgCAM and tenascin as shown by aggregation assays

Astroglial neurotrophic and neurite-promoting factors

using microbeads coated with these cell adhesion molecules. The high affinity binding to tenascin is mediated largely through the core proteins (Grumet et al., 1994). Both proteoglycans modulate neuronal and glial cell adhesion, migration and neurite outgrowth, presumably through binding to neural cell adhesion molecules. 5. Summary Astroglial cells express a still increasing plethora of neurotrophic and neurite growth-promoting molecules in vivo and/or in vitro, including: (1) soluble peptide growth factors and proteins, such as neurotrophins, FGFs, CNTF, IGFs, TGF)8, SI00)8 and GDN, (2) amino acid and neuropeptide transmitters, e.g. glutamate, proenkephalin, neuropeptide Y and somatostatin, (3) membranebound cell adhesion molecules, such as N-CAM, LI, N-CAD or AMOG, and (4) the extracellular matrix glycoproteins laminin, fibronectin and tenascin as well as various PGs. While some of the neuroactive factors are localized in discrete regions of the intact CNS, related to specific pathways or the target area of responsive neurons (NGF, CNTF, IGF I, GDN), others are more widely distributed (bFGF, 8100/3, TGF)8, laminin, fibronectin, PGs). Expression of these molecules is regulated during development and/or after lesions of the CNS. However, the regulatory mechanisms underlying synthesis or release of the majority of these factors remain obscure. Some of the astroglial peptide growth factors, including neurotrophins, bFGF, CNTF and IGF I, have been shown to exert protective or regenerative influences on neurons following traumatic, chemical or ischemic lesions in the brain. These observations illustrate the enormous therapeutic potential of astroglial neurotrophic factors, with respect to neurodegenerative diseases. References Adamo, M., Werner, H., Farnsworth, W., Roberts, C.T., Raizada, M. and Le Roith, D. (1988) Dexamethasone reduces steady state insulin-like growth factor I messenger ribonucleic acid levels in rat neuronal and glial cells in primary culture. Endocrinology 123: 2565-2570.

H.W. Maileretal. Aizenmann, Y. and de Vellis, J. (1987) Brain neurons develop in a serum and glial free environment: effects of transferrin, insulin, insulin-like growth factor I and thyroid hormone on neuronal survival, growth and differentiation. Brain Res. 406: 32-42. Ajemian, A., Ness, R. and David S. (1994) Tenascin in the injured rat optic nerve and in non-neuronal cells in vitro: potential role in neural repair. J. Comp. Neurol. 340: 233-242. Aloisi, F., Care, A., BorseUino, G., Gallo, P., Rosa, S., Bassani. A., Cabibbo, A., Testa, U., Levi, G. and Peschle, C. (1992) Production of hemolymphopoetic cytokines (IL-6, IL-8, colony-stimulating factors) by normal human astrocytes in response to IL-1 beta and tumor necrosis factor alpha. J. Immunol. 149: 2358-2366. Altar, C.A., Siuciak, J.A., Wrigth, P., Ip, N.Y., Lindsay, R.M. and Wiegand, S.J. (1994) In situ hybridization of trkB and trkC receptor mRNA in rat forebrain and association with high-affinity binding of [125J]BDNF, [125J]NT-4/5 and [125J]NT-3. Eur. J. Neurosci. 6: 1389-1405. Anderson, K.J., Dam, D., Lee, S. and Cotman, C.W. (1988) Basic fibroblast growth factor prevents death of lesioned cholinergic neurons in vivo. Nature 332: 360-361. Antonicek, H., Persohn, E. and Schachner, M. (1987) Biochemical and functional characterization of a novel neuronglia adhesion molecule that is involved in neuronal migration. 7. Cell Biol. 104: 1587-1595. Aoki, C, Joh, T.H. and Pickel, V.M. (1987) Ultrastructural localization of ^-adrenergic receptor-like immunoreactivity in the cortex and neostriatum of rat brain. Brain Res. 437: 264-268. Aquino, D.A., Margolis, R.U. and Margolis, R.K. (1984a) Immunocytochemical localization of a chondroitin sulfate proteoglycan in nervous tissue. I. Adult brain, retina and peripheral nerve. J. Cell Biol. 99: 1117-1129. Aquino, D.A., Margolis, R.U. and Margolis, R.K. (1984b) Immunocytochemical localization of a chondroitin sulfate proteoglycan in nervous tissue. II. Studies in developing brain. J. Cell Biol. 99: 1130-1139. Arakawa, Y., Sendtner, M. and Thoenen, H. (1990) Survival effect of ciliary neurotrophic factor (CNTF) on chick embryonic motoneurons in culture: comparison with other neurotrophic factors and cytokines. J. Neurosci. 10: 35073515. Araujo, D.M. and Cotman, C.W. (1992) Basic FGF in astroglial, microglial, and neuronal cultures: characterization of binding sites and modulation of release by lymphokines and trophic factors. J. Neurosci. 12: 1668-1678. Ard, M.D. and Bunge, R.P. (1988) Heparan sulfate proteoglycan and laminin immunoreactivity on cultured astrocytes: relationship to differentiation and neurite outgrowth. J. Neurosci. 8: 2844-2858. Baird, A. and Ling, N. (1987) Fibroblast growth factors are present in the extracellular matrix produced by endothelial cells in vitro: implications for a role of heparitinase-like enzymes in the neovascular response. Biochem. Biophys. Res. Commun. 142: 428^35.

389 Barnea, G., Grumet, M., Milev, P., Silvennoinen, O., Levy, J.B., Sap, J. and Schlessinger, J. (1994) Receptor tyrosine phosphatase beta is expressed in the form of proteoglycan and binds to the extracellular matrix protein tenascin. J. Biol. Chem. 269: 14349-14352. Bartsch, S., Bartsch, U., Dorries, U., Faissner, A., Weller, A., Ekblom, P. and Schachner, M. (1992) Expression of tenascin in the developing and adult cerebellar cortex. J. Neurosci. 12: 736-749. Bartsch, U., Kirchhoff, F. and Schachner, M. (1989) Immunohistological localization of the adhesion molecules LI, NCAM, and MAG in the developing and adult optic nerve of mice. /. Comp. Neurol. 284: 451-462. Bashkin, P., Doctrow, S., Klagsbrun, M., Svahn, CM., Folkman, J. and Vlodavsky, I. (1989) Basic fibroblast growth factor binds to subendothelial extracellular matrix and is released by heparitinase and heparin-like molecules. Biochemistry!^: 1737-1743. Bates, C.A. and Meyer, R.L. (1994) Differential effect of serum on laminin-dependent outgrowth of embryonic and adult mouse optic axons in vitro. Exp. Neurol. 125: 99-105. Bazan, J.F. (1991) Neuropoietic cytokines in the hematopoietic fold. Neuron 7: 197-208. Beck, K., Hunter, I. and Engel, J. (1990) Structure and function of laminin: anatomy of a multidomain glycoprotein. FASEBJ.4: 148-160. Benveniste, E.N., Sparacio, S.M., Norris, J.G., Grenett, H.E. and Fuller, G.M. (1990) Induction and regulation of interleukin-6 gene expression in rat astrocytes. J. Neuroimmunol. 30: 201-212. Benveniste, E.N., Kwon, J., Chung, W.J., Sampson, J., Pandya, K., Tang, L.P. (1994) Differential modulation of astrocyte cytokine gene expression by TGF-beta. J. Immunol. 153:5210-5221. Biagini, G., Frasoldati, A., Fuxe, K. and Agnati, L.F. (1994) The concept of astrocyte-kinetic drug in the treatment of neurodegenerative diseases: evidence for L-deprenylinduced activation of reactive astrocytes. Neurochem. Int. 25: 17-22. Bignami, A., Perides, G. and Rahemtulla, F. (1993) Versican, a hyaluronate-binding proteoglycan of embryonal precartilaginous mesenchyma. is mainly expressed postnatally in brain. /. Neurosci. Res. 34: 97-106. Bixby, J.L. and Harris, W.A. (1991) Molecular mechanisms of axon growth and guidance. Annu. Rev. Cell Biol. 7: 117-159. Bondy, C.A. (1991) Transient IGF-I gene expression during the maturation of functionally related central projection neurons. J. Neurosci. 11: 3442-3455. Bourdon, M.A., Wikstrand, D.J., Furthmayr, H., Matthews, T.J. and Bigner, D.D. (1983) Human glioma-mesenchymal extracellular matrix antigen defined by monoclonal antibody. Cancer Res. 43: 2796-2805. Brenneman, D.E., Neale, E.A., Foster, G.A., d'Autremont, S.W. and Westbrook, G.L. (1987) Nonneuronal cells mediate neurotrophic action of vasoactive intestinal peptide. J. Cell Biol. 104: 1603-1610.

390 Brodkey, J.A., Laywell, E.D., O'Brien, T.F., Faissner, A., Stefansson, K., Domes, H.U., Schachner, M. and Steindler, D.A. (1995) Focal brain injury and upregulation of a developmentally regulated extracellular matrix protein. J. Neurosurg.Sl: 106-112. Bronner-Fraser, M. and Lallier, T. (1988) A monoclonal antibody against a laminin-heparan sulfate proteoglycan complex perturbs cranial neural crest cell migration in vivo. J. Cell Biol. 106: 1321-1329. Calof, A.L. and Lander, A.D. (1991) Relationship between neuronal migration and cell-substratum adhesion: laminin and merosin promote olfactory neuronal migration but are anti-adhesive, y. Cell Biol. 115: 779-794. Chalazonitis, A., Kalberg, J., Twardzik, D.R., Morrison, R.S. and Kessler J.A. (1992) Transforming growth factory? has neurotrophic actions on sensory neurons in vitro and is synergistic with nerve growth factor. Dev. Biol. 152: 121-132. Chiquet, M. (1989) Tenascin/Jl/cytotactin: the potential function of hexabrachion proteins in neural development. Dev. Neurosci. 11: 266-275. Chiquet, M. and Fambrough, D.M. (1984a) Chick myotendinous antigen. I. A monoclonal antibody as a marker for tendon and muscle morphogenesis. J. Cell Biol. 98: 1926-1936. Chiquet, M. and Fambrough, D.M. (1984b) Chick myotendinous antigen. II. A novel extracellular glycoprotein complex consisting of large disulfide-linked subunits. J. Cell Biol 98: 1937-1946. Chiquet-Ehrismann, R., Mackie, E.J., Pearson, C.A. and Sakakura, T. (1986) Tenascin: an extracellular matrix protein involved in tissue interactions during fetal development and oncogenesis. Cell 47: 131-139. Chiu, A.Y., Espinosa de los Monteros, A., Cole, R. A., Loera, S. and de Vellis, J. (1991) Laminin and s-laminin are produced and released by astrocytes, Schwann cells, and Schwannomas in culture. Glia 4: 11-24. Chung, A.E., Jaffe, R., Freeman, I.L., Vergnes, J.P., Braginski, J.E. and Carlin, B. (1979) Properties of a basement membrane-related glycoprotein synthesized in culture by a mouse embryonal carcinoma-derived cell line. Cell 16: 277-287. Chuong, C.C, Crossin, K.L. and Edelman G.M. (1987) Sequential expression and differential function of multiple adhesion molecules during the formation of cerebellar cortical layers. J. Cell Biol. 104: 331-342. Collins, F. and Dawson, A. (1983) An effect of nerve growth factor on parasympathetic neurite outgrowth. Proc. Natl. Acad. Sci. USA 80: 2091-2094. Condorellei, D.F., Dell'Albani, P., Mudo, G., Timmusk, T. and Belluardo, N. (1994) Expression of neurotrophins and their receptors in primary astroglial cultures: induction by cyclic AMP-elevating agents. J. Neurochem. 63: 509-516. Constam, D.B., Philipp, J., Malipiero, U.V., ten Dijke, P., Schachner, M. and Fontana, A. (1992) Differential expression of transforming growth factor-beta 1, -beta 2, and beta3 by glioblastoma cells, astrocytes and microglia. J. Immunol. 148: 1404-1410.

Astroglial neurotrophic and neurite-promoting factors Crossin, K.L., Hoffman, S., Tan, S.S. and Edelman, G.M. (1989) Cytotactin and its proteoglycan ligand mark structural and functional boundaries in somatosensory cortex of the early postnatal mouse. Dev. Biol. 136: 381-392. Crossin, K.L., Prieto, A.L., Hoffman, S., Jones, F.S. and Friedlander, D. (1990) Expression of adhesion molecules and the establishment of boundaries during embryonic and neural development. Exp. Neurol. 109: 6-18. D'Alessandro, J.S. and Wang, E.A. (1994) Bone morphogenetic proteins inhibit proliferation, induce reversible differentiation and prevent cell death in astrocyte lineage cells. Growth Factors 11: 45-52. D'Alessandro, J.S., Yetz-Aldape, J. and Wang, E.A. (1994) Bone morphogenetic proteins induce differentiation in astrocyte lineage cells. Growth Factors 11: 53-69. Davis, S. and Yancopoulos, G.D. (1993) The molecular biology of the CNTF receptor. Curr. Opin. Neurobiol. 3: 20-24. Davis, S., Aldrich, T.H., Valenzuela, D.M., Wong, V., Furth, M.E., Squinto, S.P. and Yancopoulos, G.D. (1991) The receptor for ciliary neurotrophic factor. Science 253: 59-63. Devaskar, S.U., Giddings, S.J., Rajaknmar, P.A., Carnaghi, L.R., Menon, R.K. and Zahm, D.S. (1994) Insulin gene expression and insulin synthesis in mammalian neural cells. J. Biol. Chem. 269: 8445-8454. Dobrea, G.M., Unnerstall, J.R. and Rao, M.S. (1992) The expression of CNTF message and immunoreactivity in the central and peripheral nervous system of the rat. Dev. Brain Res. 66: 209-219. Doherty, P. and Walsh, F.S. (1989) Neurite guidance molecules. Curr. Opin. Cell Biol. 1: 1102-1106. Dufour, S., Duband, J.L., Humphries. M.J., Obara, M., Yamada, K.M. and Thiery, J.P. (1988) Attachment, spreading and locomotion of avian neural crest cells are mediated by multiple adhesion sites on fibronectin molecules. EMBO J. 7:2661-2671. Dunn, R. Landry, H., O'Hanlon, D., Dunn, J., AUore, R., Brown, I. and Marks, A. (1987) Reduction in SI00 protein P subunit mRNA in C6 rat glioma cells following treatment with anti-microtubular drugs. J. Biol. Chem. 262: 35623566. Edelman, G.M. and Crossin, K.L. (1991) Cell adhesion molecules: implications for a molecular histology. Annu. Rev. Biochem. 60: 155-190. Ehrig, K., Leivo, I., Argraves, W.S., Ruoslahti, E. and Engvall, E. (1990) Merosin, a tissue-specific basement membrane protein, is a laminin-like protein. Proc. Natl. Acad. Sci. USA 87: 3764-3268. Engele, J. and Bohn, M.C. (1991) The neurotrophic effects of fibroblast growth factors on dopaminergic neurons in vitro are mediated by mesencephalic glia. J. Neurosci. 11: 30703078. Engvall, E., Earwicker, D., Day, A., Muir, D., Manthorpe, M. and Paulsson, M. (1992) Merosin promotes cell attachment and neurite outgrowth and is a component of the neuritepromoting factor of RN22 Schwannoma cells. Exp. Cell Res. 198: 115-123.

H.W.

Maileretal.

Erickson, H.P. and Bourdon, M.A. (1989) Tenascin: an extracellular matrix protein prominent in specialized embryonic tissues and tumors. Anna. Rev. Cell Biol. 5: 71-92. Erickson, H.P. and Inglesias, J.L. (1984) A six-armed oligomer isolated from cell surface fibronectin preparations. Nature (London) 311: 267-269. Ernsberger, V., Sendtner, M. and Rohrer, H. (1989) Proliferation and differentiation of embryonic chick sympathetic neurons: effects of ciliary neurotrophic factor. Neuron 2: 1275-1284. Faissner, A., Clement, A., Lochter, A., Streit, A., Mandl, C. and Schachner, M. (1994) Isolation of a neural chondroitin sulfate . proteoglycan with neurite outgrowth promoting properties. J. Cell Biol. 126: 783-799. Faissner, A. and Kruse, J. (1990) Jl/tenascin is a repulsive substrate for central nervous system neurons. Neuron 5: 627-637. Faissner, A., Kruse, J., Chiquet-Ehrismann, R. and Mackie, E.J. (1988) The high-molecular-weight Jl glycoproteins are immunochemically related to tenascin. Differentiation 37: 104-114. Fallon, J.R. (1985) Preferential outgrowth of central nervous system neurites on astrocytes and Schwann cells as compared with non-glial cells in vitro. J. Cell Biol. 100: 198207. Farmer, L. Sommer, J. and Monard, D. (1990) Glia-derived nexin potentiates neurite extension in hippocampal pyramidal cells in vitro. Dev. Neurosci. 12: 73-80. Fedoroff, S. (1990) Astrocyte reaction in injury and regeneration. In: F.J. Seil (Ed.), Advances in Neural Regeneration Research, Neurology and Neurobiology Series, Vol. 60, Wiley-Liss, New York, pp. 161-170. Fischer, W., Wictorin, K., Bjorklund, A., Williams, L.R., Varon, S. and Gage, F.H. (1987) Amelioration of cholinergic neuron atrophy and spatial memory impairment in aged rats by nerve growth factor. Nature 329: 65-68. Fishell, G. and Hatten, M.E. (1991) Astrotactin provides a receptor system for CNS neuronal migration. Development 113:755-765. Flanders, K.C., Ludecke, G., Engels, S., Cissel, D.S., Roberts, A.B., Kondaiah, P., Lafyatis, R., Sporn, M.B. and Unsicker, K. (1991) Localization and actions of transforming growth factor-)3s in the embryonic nervous system. Development 113: 183-191. Foschini, D.R., Kestler, A.M., Egger, M.D. and Crockett, D.P. (1994) The up-regulation of trkA and trkB in dorsal column astrocytes following dorsal rhizotomy. Neurosci. Lett. 169: 21-24. Friedlander, D.R, Milev, P., Karthikeyan, L., Margolis, R.K., Margolis, R.U. and Grumet, M. (1994) The neuronal chondroitin sulfate proteoglycan neurocan binds to the neural cell adhesion molecules Ng-CAM/Ll/NILE and N-CAM, and inhibits neuronal adhesion and neurite outgrowth. J. Cell Biol. 125:669-680. Furukawa, S., Furukawa, Y., Satayoshi, E. and Hayashi, K. (1986) Synthesis and secretion of nerve growth factor by

391 mouse astroglial cells in culture. Biochem. Biophys. Res Commun. 136: 57-63. Furukawa, S., Furukawa, Y., Satayoshi, E. and Hayashi, K. (1987) Synthesis/secretion of nerve growth factor is associated with cell growth in cultured mouse astroglial cells. Biochem. Biophys. Res. Commun. 142: 395-402. Gallagher, J.T. (1989) The extended family of proteoglycans: social residents of the pericellular zone. Curr. Opin. Cell Biol. 1: 1201-1218. Geiger, B. and Ayalon, O. (1992) Cadherins. Annu. Rev. Cell. Biol. 8: 307-332. Gloor, S., Antonicek, H., Sweader, K.J., Pagliusi, S., Frank, R., Moos, M. and Schachner, M. (1990) The adhesion molecule on glia (AMOG) is a homologue of the subunit of Na, K-ATPase. J. Cell Biol. 110: 165-174. Gluckman, P., Klempt, N., Guan, J., Mallard, C , Sirimanne, E., Dragunow, M., Klempt, M., Singh, K., Williams, C. and Nikolics, K. (1992) A role for IGF-I in the rescue of CNS neurons following hypoxic-ischemic injury. Biochem. Biophys. Res. Commun. 182: 593-595. Gnahn, H., Hefti, F., Heumann, R., Schwab. M. and Thoenen, H. (1983) NGF-mediated increase of choline acetyltransferase (ChAT) in the neonatal forebrain: evidence for a physiological role of NGF in the brain? Dev. Brain Res. 9: 45-52. Gomez-Pinilla, F., Lee, J.W. and Cotman, C.W. (1992) Basic FGF in adult rat brain: cellular distribution and response to entorhinal lesion and fimbria-fornix transection. J. Neurosci 12: 345-355. Gomez-Pinilla, F., Lee, J.W. and Cotman, C.W. (1994) Distribution of basic fibroblast growth factor in the developing rat brain. Neuroscience 61: 911-923. Gonzales, D., Dees, W.L., Hiney, J.K., Ojeda, S.R. and Saneto, R.P. (1990) Expression of y3-nerve growth factor in cultured cells derived from the hypothalamus and cerebral cortex. Brain Res. 511: 249-258. Gospodarowicz, D. (1987) Isolation and characterization of acidic and basic fibroblast growth factor. Methods Enzymol. 147: 106-119. Gowda, D.C., Margolis, R.U. and Margolis, R.K. (1989) Presence of the HNK-1 epitope on poly(A^-acetyllactosaminyl) oligosaccharides and identification of multiple core proteins in the chondroitinsulfate of brain. Biochemistry 28: 4 4 6 8 4474. Grumet, M., Flaccus, A. and Margolis, R.U. (1993) Functional characterization of chondroitin sulfate proteoglycans of brain: interactions with neurons and neural cell adhesion molecules. /. Cell Biol. 120: 815-824. Grumet, M., Hoffman, S., Crossin, K.L. and Edelman, G.M. (1985) Cytotactin, an extracellular matrix protein of neural and nonneural tissues that mediates glia-neuron interaction. Proc. Natl. Acad. Sci. USA 82: 8075-8079. Grumet, M., Milev, P., Sakurai, T., Karthikeyan, L., Bourdon, M., Margolis, R.K and Margolis, R.U. (1994) Interactions with tenascin and differential effects on cell adhesion of neurocan and phosphacan, two major chondroitin sulfate

392 proteoglycans of nervous tissue. /. Biol. Chem. 269: 1214212146. Guenther, J., Nick, H.P. and Monard, D. (1985) A gliaderived neurite-promoting factor with protease inhibitory activity. EMBO J. 4: 1963-1966. Giuseppetti, J.M., McCarthy, J.B. and Letourneau, P.C. (1994) Isolation and partial characterization of cell-surface heparan sulfate proteoglycan from embryonic rat spinal cord. J. Neurosci. Res. 37: 584-595. Hama, T., Miyamoto, M., Tsukui, H., Nishio, C. and Hatanaka, H. (1989) Interleukin-6 as a neurotrophic factor for promoting the survival of cultured basal forebrain cholinergic neurons from postnatal rats. Neurosci. Lett. 104: 340-344. Haselbacher, G., Schwab, M., Pasi, A. and Humbel, R. (1985) Insulin-like growth factor II (IGF II) in human brain: regional distribution of IGF II of an higher molecular mass forms. Proc. Natl. Acad. Sci. USA 82: 2153-2158. Hasenohrl, R.U., Frisch, C., Junghans, U., Muller, H.W. and Huston, J.P. (1995) Facilitation of learning following injection of the chondroitin sulfate proteoglycan biglycan into the vicinity of the nucleus basalis magnocellularis. Behav. Brain Res. 70: 59-67. Hatten, M.E. (1990) Riding the glial monorail: a common mechanism for glial-guided neuronal migration in different regions of the developing mammalian brain. Trends Neurosci. 13: 179-184. Hatten, M.E., Furie, M.B. and Rifkin, D.B. (1982) Binding of developing mouse cerebellar cells to fibronectin: a possible mechanism for the formation of the external granular layer. J. Neurosci. 2: 1195-1206. Hatten, M.E., Lynch, M., Rydel, R.E., Sanchez, J., Silverstein, J., Moscatelli, D. and Rifkin, D.B.(1988) In vitro neurite extension by granule neurons is dependent upon astroglialderived fibroblast growth factor. Dev. Biol. 125: 280289 Hefti, F., David, A. and Hartikka, J. (1984) Chronic intraventricular injections of nerve growth factor elevate hippocampal choline acetyltransferase activity in adult rats with parietal septo-hippocampal lesions. Brain Res. 293: 311-317. Hefti, F., Hartikka, J. and Knusel, B. (1989) Function of neurotrophic factors in the adult and aging brain and their possible use in the treatment of neurodegenerative diseases. Neurobiol. Aging 10: 515-533. Heine, U.I., Munoz, E.F., Flanders, K.C., Ellingsworth, L.R., Lam, H.Y.P., Thompson, N.L., Roberts, A.B. and Sporn, M.B. (1987) Role of transforming growth factor-^ in the development of the mouse embryo. J. Cell Biol. 105: 28612876. Hepler, J.E. and Lund, P.K. (1990) Molecular biology of the insulin-like growth factors. Mol. Neurobiol. 4: 93-127. Hoffmann, M.C., Nitsch, C, Scotti, A.L., Reinhard, E. and Monard, D. (1992) The prolonged presence of glia-derived nexin, an endogenous protease inhibitor, in the hippocampus after ischemia-induced delayed neuronal death. Neuroscience 49: 397-408.

Astroglial neurotrophic and neurite-promoting factors Hu, R.M. and Levin, E.R. (1994) Astrocyte growth is regulated by neuropeptides through Tis 8 and basic fibroblast growth factor. J. Clin. Invest. 93: 1820-1827. Hughes, S.M., Lillien, L.E., Raff, M.C., Rohrer, H. and Sendtner, M. (1988) Ciliary neurotrophic factor induces type-2 astrocyte differentiation in culture. Nature 335: 7073. Hunter, D.D., Shah, V., Merlie, J.P. and Sanes, J.R. (1989) A laminin-like adhesive protein concentrated in the synaptic cleft of the neuromuscular junction. Nature 338: 229-234. Hunter, D.D., Llinas, R., Ard, M., Merlie, J.P. and Sanes, J.R. (1992) Expression of 5-laminin and laminin in the developing rat central nervous system. J. Comp. Neurol. 323: 238251. Husmann, K., Faissner, A. and Schachner, M. (1992) Tenascin promotes cerebellar granule cell migration and neurite outgrowth by different domains in the fibronectin type III repeats. 7. Cell Biol. 116: 1475-1486. Hynes. M.A., Brooks, P.J., VanWyk, J.J. and Lund, P.K. (1988) Insulin-like growth factor II messenger ribonucleic acids are synthesized in the choroid plexus of the rat brain. Mol. Endocrinol. 2: 47-54. Ip, N.Y., Li, Y., van de Stadt, I., Panayotatos, N., Alderson, R.F. and Lindsay, R.M. (1991) Ciliary neurotrophic factor enhances neuronal survival in embryonic rat hippocampal cultures. J. Neurosci. 11: 3124-3134. Ip, N.Y., Nye, S.H., Boulton, T.G., Davis. S., Taga, T., Li, Y., Birren, S.J., Yasukawa, K., Kishimoto, T., Anderson. D.J., Stahl, N. and Yancopoulos, G.D. (1992) CNTF and LIF act on neuronal cells via shared signaling pathways that involve the IL-6 signal transducing receptor component pgl30. CW/69: 1121-1132. Janet, T., Grothe, C, Pettmann, B., Unsicker, K. and Sensenbrenner, M. (1988) Immunocytochemical demonstration of fibroblast growth factor in cultured chick and rat neurons. J. Neurosci. Res. 19: 195-201. Jhaveri, S., Erzurumlu, R.S. and Crossin, K.L. (1991) Barrel construction in rodent neocortex: role of thalamic afferents versus extracellular matrix molecules. Proc. Natl. Acad. Sci. USA 88: 4489^493. Johnson-Green, P.C, Dow, K.E. and Riopelle. R.J. (1992) Neurite growth modulation associated with astrocyte proteoglycans: influence of activators of inflammation. Glia 5: 33-42. Jones, F.S., Hoffman, S., Cunningham, B.A. and Edelman, G.M. (1989) A detailed structural model of cytotactin: protein homologies, alternative splicing, and binding regions. Proc. Natl. Acad Sci. USA 86: 1905-1909. Junghans, U., Koops, A., Westmeyer, A., Kappler, J., Meyer, H.E. and Muller, H.W. (1995) Purification of a meningeal cell-derived chondroitin sulfate proteoglycan with neurotrophic activity for brain neurons and its identification as biglycan. Eur. J. Neurosci. 1: 2341-2350. Junier, M.P., Suzuki, F., Onteniente, B. and Peschanski, M. (1994) Target-deprived CNS neurons express the NGF gene while reactive glia around their axonal terminals contain

H.W. Maileretal. low and high affinity NGF receptors. Mol. Brain Res. 24: 247-260. Kadle, R., Suksang, C, Roberson, E.D. and Fellows, R.E. (1988) Identification of an insulin-like factor in astrocyte conditioned medium. Brain Res. 460: 60-67. Kjellen, L. and Lindahl, U. (1991) Proteoglycans. Structures and interactions. AnfiM. Rev. Biochem. 60: 443-475. Kligman, D. and Marshak, D.R. (1985) Purification and characterization of a neurite extension factor from bovine brain. Proc. Natl. Acad. Sci. USA 82: 7136-7139. Komoly, S., Hudson, L.D., Webster, H.D. and Bondy, C.A. (1992) Insulin-like growth factor I gene expression is induced in astrocytes during experimental demyelination. Proc. Natl. Acad. Sci. USA 89: 1894-1898. Koops, A., Kappler, J., Junghans, U., Kuhn, G., Kresse, H. and Miiller, H.W. (1995) Astrocytes express biglycan, a chondroitin/dermatan sulfate proteoglycan supporting the survival of neocortical neurons. Mol. Brain Res. 41: 65-73. Kruse, J., Keilhauer, G., Faissner, A., Timpl, R. and Schachner, M. (1985) The Jl glycoprotein - a novel nervous system cell adhesion molecule of the L2/HNK-1 family. Nature (London) 316: 146-148. Kushima, Y., Hama, T. and Hatanaka, H. (1992) Interleukin-6 as a neurotrophic factor for promoting the survival of cultured catecholaminergic neurons in a chemically defined medium from fetal and postnatal rat midbrains. Neurosci. Res. 13: 267-280. LaFont, F., Rouget, M., Triller, A., Prochiantz, A. and Rousselet, A. (1992) In vitro control of neuronal polarity by glycosaminoglycans. Development 114: 17-29. Lander, A.D., Fujii, D.K. and Reichardt, L.F. (1985) Laminin is associated with the 'neurite outgrowth-promoting factors' found in conditioned media. Proc. Natl. Acad. Sci. USA 82: 2183-2187. Lauder, J.M. (1990) Ontogeny of the serotonergic system in the rat: serotonin as a developmental signal. Ann. NY. Acad. Sci. S5: 291-314. LeBaron. R.G., Zimmermann, D.R. and Ruoslahti, E. (1992) Hyaluronate binding properties of versican. J. Biol. Chem. 267: 10003-10010. Lee, S.C, Liu, W., Dickson, D.W., Brosnan, C.F. and Berman, J.W. (1993) Cytokine production by human fetal microglia and astrocytes. Differential induction by lipopolysaccharide and IL-1 beta. J. Immunol. 150: 2659-2567. Leingartner, A. and Lindholm, D. (1994) Two promoters direct transcription of the mouse NT-3 gene. Eur. J. Neurosci. 6: 1149-1159. Leslie, F.M. (1993) Neurotransmitters as neurotrophic factors. In: J.H. Fallon and S.E. Loughlin (Eds.), Neurotrophic Factors, Academic Press, San Diego, CA, pp. 565-598. Letourneau, P.C, Condic, M.L. and Snow, D.M. (1992) Extracellular matrix and neurite outgrowth. Curr. Opin. Genet. Dev. 2: 625-634. Liesi, P. and Risteli, L. (1989) Glial cells of mammalian brain produce a variant form of laminin. Exp. Neurol. 105: 8692.

393 Liesi, P. and Silver, J. (1988) Is astrocyte laminin involved in axon guidance in the mammalian CNS? Dev. Biol. 130: 774-785. Liesi, P., Kirkwood, T. and Vaheri, A. (1986) Fibronectin is expressed by astrocytes cultured from embryonic and early postnatal rat brain. Exp. Cell Res. 163: 175-185. Lillien, L.E., Sendtner, M. and Raff, M.C. (1990) Extracellular matrix-associated molecules collaborate with ciliary neurotrophic factor to induce type-2 astrocyte development. J. Cell Biol. 111:635-644. Lindholm, D., Castren, E., Kiefer, R., Zafra, F. and Thoenen, H. (1992) Transforming growth factor-beta I in the rat brain: increase after injury and inhibition of astrocyte proliferation. J. Cell Biol. 117: 395^00. Lips, K., Stichel, C.C. and Muller, H.W. (1995) Restricted appearance of tenascin and chondroitin sulphate proteoglycans after transection and sprouting of adult rat postcommissural fornix. /. Neurocytol. 24: 229-464. Liu, X., Yao, D.L., Bondy, C.A., Brenner, M., Hudson, L.D., Zhou, J. and Webster, H.D. (1994) Astrocytes express insulin-like growth factor-I (IGF-I) and its binding protein, IGFBP-2, during demyelination induced by experimental autoimmune encephalomyelitis. Mol. Cell. Neurosci. 5: 418-430. Lochter, A., Vaughan, L., Kaplony, A., Prochiantz, A., Schachner, M. and Faissner, A. (1991) Jl/tenascin in substrate-bound and soluble form displays contrary effects on neurite outgrowth. 7. Cell Biol. 113: 1159-1171. Loo, D.T., Althoen, M.C. and Cotman, C.W. (1994) Down regulation of nestin by TGF-beta or serum in SFME cells accompanies differentiation into astrocytes. NeuroReport 5: 1585-1588. Lu, B., Yokoyama, M., Dreyfus, C.F. and Black, LB. (1991) NGF gene expression in actively growing brain glia. J. Neurosci. 11:318-326. Manthorpe, M. and Varon, S. (1985) Regulation of neuronal survival and neurite growth in the avian ciliary ganglion by trophic factors. In: G. Guroff (Ed.), Growth and Maturation Factors, Vol. 3, Wiley, New York, pp. 77-117. Margolis, R.K. and Margolis, R.V. (1993) Nervous tissue proteoglycans. Experientia 49: 429-446. Margolis, R.U. and Margolis, R.K. (1989) Nervous tissue proteoglycans. Dev. Neurosci. 11: 276-288. Marshak, D.R. (1990) SlOO beta as a neurotrophic factor. Prog. Brain Res. 86: 169-181. Marshak, D.R., Psce, S.A., Stanley, L.C. and Griffin, W.S. (1992) Increased SlOO beta neurotrophic activity in Alzheimer's disease temporal lobe. Neurobiol. Aging 13: 1-7. Martin, D.L. (1992) Synthesis and release of neuroactive substances by glial cells. Glia 5: 81-94. Martinou, J.C, Le Van Thai, A., Valette, A. and Weber, M.J. (1990) Transforming growth factor )31 is a potent survival factor for rat embryo motoneurons in culture. Dev. Brain Res. 52: 175-181. Massague, J. (1990) The transforming growth factor-)^-family. Annu. Rev. Cell Biol. 6: 597-641.

394 Massague, J. (1992) Receptors for the TGF-y3 family. Cell 69: 1067-1070. Matthew, W.D., Greenspan, R.J., Lander, A.D. and Reichardt, L.F. (1985) Immunopurification and characterization of a neuronal heparan sulfate proteoglycan. J. Neurosci. 5: 1842-1850. Matthiessen, H.P., Schmalenbach, C. and Muller, H.W. (1989) Astroglia-released neurite growth-inducing activity for embryonic hippocampal neurons is associated with laminin bound in a sulfated complex and free fibronectin. Glia 2: 177-188. Mayer, M., Bhakoo, K. and Noble, M. (1994) Ciliary neurotrophic factor and leukemia inhibitory factor promote the generation, maturation and survival of oligodendrocytes in vitro. Development 129: 143-153. McDonald, N.Q. and Hendrickson, W.A. (1993) A structural family of growth factors containing a cystine knot motif. C^//73: 421-424. McKeon, R.J., Schreiber, R.C., Rudge, J.S. and Silver, J. (1991) Reduction of neurite outgrowth in a model of glial scarring following CNS injury is correlated with the expression of inhibitory molecules on reactive astrocytes. J. Neurosci. 11:3398-3411. McLoon, S.C, McLoon, L.K., Palm, S.L. and Furcht, L.T. (1988) Transient expression of laminin in the optic nerve of the developing rat. J. Neurosci. 8: 1981-1990. McMorris, F.A. and Dubois-Dalcq, M. (1988) Insulin-like growth factor I promotes cell proliferation and oligodendroglial commitment in rat glial progenitor cells developing in vitro. J. Neurosci. Res. 21: 199-209. Meier, R., Spreyer, P., Ortmann, R., Harel, A. and Monard, D. (1989) Induction of glia-derived nexin after lesion of a peripheral nerve. Nature 342: 548-550. Meyer, V. and Unsicker, K (1994) Cell density and exogenous CNTF affect CNTF mRNA levels in glial cell cultures. NeuroReport 5: 1946-1948. Milev, P., Friedlander, D.R., Sakurai, T., Karthikeyan, L., Margolis, R.K., Grumet, M. and Margolis, R.U. (1994) Interactions of the chondroitin sulfate proteoglycan phosphacan, the extracellular domain of a receptor-type protein tyrosin phosphatase, with neurons, glia, and neural cell adhesion molecules. /. Cell Biol. 127: 17031715. Miller, R.H. and Smith, G.M. (1990) Cell and molecular interactions that influence astrocyte mediated axon outgrowth. In: F.J. Seil (Ed.), Advances in Neural Regeneration Research, Wiley-Liss, New York, pp. 171-183. Mitrovic, N., Dorries, U. and Schachner, M. (1994) Expression of the extracellular matrix glycoprotein tenascin in the somatosensory cortex of the mouse during postnatal development: an immuncytochemical and in situ hybridization analysis. /. Neurocytol. 23: 364-378. Mittal, B. and David, S. (1994a) A monoclonal antibody that recognizes an adhesion molecule expressed by certain cells of neuroectodermal and mesenchymal origin. Mol. Cell. Neurosci. 5: 63-77.

Astroglial neurotrophic and neurite-promoting factors Mittal, B. and David, S. (1994b) The role of an astrocyte surface molecule in neuronal migration in the developing rat cerebellum. Mol. Cell. Neurosci. 5: 78-86. Monard, D. (1988) Cell-derived proteases and protease inhibitors as regulators of neurite outgrowth. Trends Neurosci. 11:541-544. Monard, D., Solomon, F., Rentsch. M. and Gysin, R. (1973) Glia-induced morphological differentiation in neuroblastoma cells. Proc. Natl. Acad.. Sci. USA 70: 1894-1897. Morganti-Kossmann, M.C., Kossmann, T., Brandes, M.E., Mergenhagen, S.E. and Wahl, S.M. (1992) Autocrine and paracrine regulation of astrocyte function by transforming growth factor beta. 7. Neuroimmunol. 39: 163-173. Moretto, G., Xu, R.Y., Walker, D.G. and Kim, S.U. (1994) Co-expression of mRNA for neurotrophic factors in human neurons and glial cells in culture. J. Neuropathol. Exp. Neurol. 53: 78-85. Morrison, R.S., Sharma, A., DeVellis, J. and Bradshaw, R.A. (1986) Basic fibroblast growth factor supports the survival of cerebral cortical neurons in primary culture. Proc. Natl. Acad. Sci. USA 83: 7537-7541. Muller, H.W. (1993) Adhesion factors. In: J.H. Fallon and S.E. Loughlin (Eds.), Neurotrophic Factors, Academic Press, San Diego, CA, pp. 509-526. Muller, H.W., Matthiessen, H.P. and Schmalenbach, C. (1990) Astroglial factors supporting neurite growth and long-term neuronal survival. In: R.J. Reier, R.P. Bunge and F.J. Seil (Eds.), Advances In Neural Regeneration Research: Proceedings of the Third International Neural Regeneration Research Symposium, Wiley, New York, pp. 147-159. Muller, H.W., Matthiessen, H.P., Schmalenbach, C. and Schroeder, W.O. (1991) Glial support of CNS neuronal survival, neurite growth and regeneration. Rest. Neurol. Neurosci. 2: 229-232. Neufeld, G. and Gospodarowicz, D. (1985) The identification and partial characterization of the fibroblast growth factor receptor of baby hamster kidney cells. J. Biol. Chem. 260: 13860-13868. Nies, D.E., Hemesath, T.J., Kim, J.-H., Gulcher, J.R. and Stefansson, K. (1991) The complete cDNA sequence of human hexabrachion (tenascin). J. Biol. Chem. 266: 28182823. Oohira, A., Matsui, F., Watanabe, E., Kushima, Y. and Maeda, N. (1994) Developmentally regulated expression of a brain specific species of chondroitin sulfate proteoglycan, neurocan, identified with a monoclonal antibody IG2 in the rat cerebrum. Neuroscience 60: 145-157. Otto, D. and Unsicker, K. (1990) Basic FGF reverses chemical and morphological deficits in the nigrostriatal system of MPTP-treated mice. 7. A^eMro5c/. 10: 1912-1921. Otto, D., Frotscher, M. and Unsicker, K. (1989) Basic fibroblast growth factor administered in gel foam rescue medial septal neurons after fimbria-fornix transection. J. Neurosci. Res. 22.^3-91. Pagani, F., Zagato, L., Vergani, C , Casan, G., Sidoli, A. and Baralle, F.E. (1991) Tissue-specific splicing pattern of fi-

H.W. Maileretal. bronectin messenger RNA precursor during development and aging in the rat. J. Cell Biol. 113: 1223-1229. Paralkar, V.M., Weeks, B.S., Yu, Y.M., Kleinmann, H.K. and Reddi, A.H. (1992) Recombinant human bone morphogenetic protein 2B stimulates PC 12 cell differentiation: potentiation and binding to type IV collagen. J. Cell Biol. 119: 1721-1728. Paulsson, M. (1992) Basement membrane proteins: structure, assembly, and cellular interactions. Crit. Rev. Biochem. Mol. Biol. 27: 93-127. Peles, E., Nativ, M., Campbell, P.L., Sakurai, T., Martinez, R., Lev, S., Clary, D.O., Schilling, J., Barnea, G., Plowman, G.D., Grumet, M. and Schlessinger, J. (1995) The carbonic anhydrase domain of receptor tyrosine phosphatase )3 is a functional ligand for the axonal cell recognition molecule contactin. Cell 82: 251-260. Perides, G., Lane, W.S., Andrews, D., Dahl, D. and Bignami, A. (1989) Isolation and partial characterization of a glial hyaluronate-binding protein. J. Biol Chem. 264: 59815987. Persohn, E. and Schachner, M. (1987) Immunoelectron microscopic localization of the neural cell adhesion molecules LI and N-CAM during postnatal development in the mouse cerebellum. J. Cell Biol. 105: 569-576. Pettmann, B., Labourdette, G., Weibel, M. and Sensenbrenner, M. (1986) The brain fibroblast growth factor (FGF) is localized in neurons. Neurosci. Lett. 68: 175-180. Pincus, D.W., DiCicco-Bloom, E.M. and Black, LB. (1990) Vasoactive intestinal peptide regulates mitosis, differentiation and survival of cultured sympathetic neuroblasts. Nature 343: 564-567. Pope, A. (1978) Neuroglia: quantitative aspects. In: E. Schoffeniels, G. Franck, D.B. Towers and L. Hertz (Eds.), Dynamic Properties of Glial Cells, Pergamon Press, Oxford, pp. 13-20. Price, J. and Hynes, R.O. (1985) Astrocytes in culture synthesize and secrete a variant form of fibronectin. J. Neurosci. 5:2205-2211. Prieto, A.L., Jones, F.S., Cunningham, B.A., Crossin, K.L. and Edelman, G.M. (1990) Localization during development of alternatively spliced forms of cytotactin mRNA by in situ hybridization. J. Biol. Chem. I l l : 685-698. Rauch, U., Karthikeyan, L., Maurel, P., Margolis, R.U. and Margolis, R.K. (1992) Cloning and primary structure of neurocan, a developmentally regulated, aggregating chondroitin sulfate proteoglycan of brain. J. Biol. Chem. 261 \ 19536-19547. Recio-Pinto, E. and Ishii, D.N. (1984) Effects of insulin, insulin-like growth factor II and nerve growth factor on neurite outgrowth in cultured human neuroblastoma cells. Brain Res. 2: 323-334. Recio-Pinto, E., Rechler, M.M. and Ishii, D.N. (1986) Effects of insulin, insulin-like growth factor II, and nerve growth factor on neurite formation and survival in cultured sympathetic and sensory neurons. J. Neurosci. 6: 12111219.

395 Reichardt, L.F. and Tomaselli, K.J. (1991) Extracellular matrix molecules and their receptors: functions in neural development. Awnw. Rev. Neurosci. 14: 531-570. Riva, M.A. and Mocchetti, I. (1991) Developmental expression of the basic fibroblast growth factor gene in rat brain. Dev. Brain Res. 19: 45-50. Rogers, S.L., Letourneau, P.C. and Pech, I.V. (1989) The role of fibronectin in neural development. Dev. Neurosci. 11: 248-265. Rotwein, P., Burgess, S.K., Milbrandt, J.D. and Krause. J.E. (1988) Differential expression of insulin-like growth factor genes in rat central nervous system. Proc. Natl. Acad. Sci. USA 85: 265-269. Rudge, J.S., Li, Y., Pasnikowski, E.M., Mattsson, K., Pan, L., Yancopoulos, G.D., Wiegand, S.L.J., Lindsay, R.M. and Ip, N.Y. (1994) Neurotrophic factor receptors and their signal transduction. Eur. J. Neurosci. 6: 693-705. Ruoslahti, E. (1988) Fibronectin and its receptors. Annu. Rev. Biochem. 57: 375^13. Ruoslahti, E. (1989) Proteoglycans in cell regulation. J. Biol. Chem. 264: 13369-13372. Ruoslahti, E. and Yamaguchi, Y. (1991) Proteoglycans as modulators of growth factor activities. Cell 64: 867-869. Saad, B., Constam, D.B., Ortmann, R.. Moos, M., Fontana, A. and Schachner, M. (1991) Astrocyte-derived TGF-2 and NGF differentially regulate neural recognition molecule expression by cultured astrocytes. J. Cell Biol. 115: 473-484. Saadat, S., Sendtner, M. and Rohrer, H. (1989) Ciliary neurotrophic factor induces cholinergic differentiation of rat sympathetic neurons in culture. J. Cell Biol. 108: 18071816. Said, S.I. (1991) Vasoactive intestinal polypeptide. Biological role in health and disease. Trends Endocrinol. Metab. 2: 107-112. Saksela, O. and Rifkin, D.B. (1990) Release of basic fibroblast growth factor-heparan sulfate complexes from endothelial cells by plasminogen activator-mediated proteolytic activity. J. Cell Biol. 110:161-115. Sandrock, A.W. and Matthew, W.D. (1987) An in vitro neurite-promoting antigen functions in axonal regeneration in vivo. Science 237: 1605-1608. Sanes, J.R. (1989) Extracellular matrix molecules that influence neural development. Annu. Rev. Neurosci. 12: 491516. Schachner, M. (1989) Families of neural adhesion molecules. Ciha Found. Symp. 145: 156-169. Scherer, J. and Schnitzer, J. (1994) Growth factor effects on the proliferation of different retinal glial cells in vitro. Dev. Brain Res. SO: 209-221. Schlunegger, M.P. and Griitter, M.G. (1992) An unusual feature revealed by the crystal structure at 2.2 A resolution of human transforming growth factor-^2. Nature 358: 430-434. Schmalenbach, C. and Muller, H.W. (1993) Astroglia-neuron interactions that promote long-term neuronal survival. J. Chem. Neuroanat. 6: 229-237. Schmalzing, G., Kroner, S., Schachner, M. and Gloor, S.

396 (1992) The adhesion molecule on glia (AMOG/beta 2) and alpha 1 subunits assemble to functional sodium pumps in Xenopus oocytes. 7. Biol. Chem. 267: 20212-20216. Schubert, D., Kimura, H., Lacorbiere, M., Vaughan, J., Karr, D. and Fischer, W. H. (1990) Activin is a nerve cell survival molecule. Nature 344: 868-870. Schwartz, J.P. (1988) Stimulation of nerve growth factor mRNA content in C6 glioma cells by a ^-adrenergic receptor and by cyclic AMP. Glia 1: 282-285. Schwartz, J.P. (1992) Neurotransmitters as neurotrophic factors: a new set of functions. Int. Rev. NeurobioL 34: 123. Schwartz, J.P. and Costa, E. (1977) Regulation of nerve growth factor content in C6 glioma cells by y3-adrenergic receptor stimulation. Naunyn-Schmiedberg's Arch. Pharmacol. 300: 123-129. Schwartz, J.P. and Mishler, K. (1990) Beta-adrenergic receptor regulation, through cyclic AMP, of nerve growth factor expression in the rat cortical and cerebellar astrocytes. Cell. Mol. NeurobioL 10: 447-457. Schwartz, J.P. and Nishiyama, N. (1994) Neurotrophic factor gene expression in astrocytes during development and following injury. Brain Res. Bull. 35: 403-407. Schwartz, J.P. and Taniwaki, T. (1994) Heterogeneity of expression of neuropeptide genes by astrocytes: functional implications. Perspect. Dev. NeurobioL 2: 251-257. Selinfreund, R.H., Barger, S.W., Pledger, W.J. and Van Eldik, L.J. (1991) Neurotrophic protein SI00/3 stimulates glial cell proliferation. Proc. NatL Acad. ScL USA 88: 3554-3558. Sendtner, M., Kreutzberg, G. and Thoenen, H. (1990) Ciliary neurotrophic factor prevents the degeneration of motor neurons after axotomy. Nature 345: 440-441. Seniuk, N.A., Henderson, J.T., Tatton, W.G. and Roder, J.C. (1994) Increased CNTF gene expression in process-bearing astrocytes following injury is augmented by R(-)-deprenyl. J. NeuroscL Res. 37: 278-286. Sephel, G.C., Burrons, B.A. and Kleinmann, H.K. (1989) Laminin neural activity and binding proteins. Dev. NeuroscL 11:313-331. Sheppard, A.M., Hamilton, S.K. and Pearlman, A.L. (1991) Changes in the distribution of extracellular matrix components accompany early morphogenetic events of mammalian cortical development. J. NeuroscL 11: 3928-3942. Shuey, D.L., Yavarone, M., Sadler, T.W. and Lauder, J.M. (1990) Serotonin and morphogenesis in the cultured mouse embryo. In: J.M. Lauder (Ed.), Molecular Aspects of Development and Aging of the Nervous System, Plenum Press, New York, pp. 205-215. Siri, A., Carnemolla, B., Saginati, M., Leprini, A., Casari, G., Baralle, F. and Zardi, L. (1991) Human tenascin: primary structure, pre-mRNA splicing patterns and localization of the epitopes recognized by two monoclonal antibodies. Nucleic Acids Res. 19: 525-531. Sommer, J., Gloor, S.M., Rovelli, G.F., Hofsteenge, J., Nick, H.P., Meier, R. and Monard, D. (1987) cDNA sequence coding for a rat glia-derived nexin and its homology to

Astroglial neurotrophic and neurite-promoting factors members of the serpin superfamily. Biochemistry 26: 64076410. Spring, J., Beck, K. and Chiquet-Ehrismann, R. (1989) Two contrary functions of tenascin: dissection of the active sites by recombinant tenascin fragments. Cell 59: 325-334. Stallcup, W.B. and Beasley, L. (1987) Bipotential glial precursor cells of the optic nerve express the NG2 proteoglycan. J. NeuroscL 7: 2737-2744. Steindler, D.A., Cooper, N.G.F., Faissner, A. and Schachner, M. (1989a) Boundaries defined by adhesion molecules during development of the cerebral cortex: the Jl/tenascin glycoprotein in the mouse somatosensory cortical barrel field. Dev. BioL 131: 243-260. Steindler, D.A., Faissner, A. and Schachner, M. (1989b) Brain 'cordones': transient boundaries of glia and adhesion molecules that define developing functional units. Comments Dev. NeurobioL 1: 29-60. Steindler, D.A., Settles, D., Erickson, H.P., Laywell, E.D., Yoshiki, A., Faissner, A., Kusakabe, M. (1995) Tenascin knockout mice: barrels, boundary molecules, and glial scars. J. NeuroscL 15: 1971-1983. Stewart, G.R. and Pearlman, A.L. (1987) Fibronectin-like immunoreactivity in the developing cerebral cortex. J. NeuroscL 1: 3325-3333. Stichel, C.C. and Muller, H.W. (1992) Expression of inherent neuronal shape characteristics after transient sensitivity to epigenetic factors. Dev. Brain Res. 68: 149-162. Stichel, C.C. and Muller, H.W. (1994) Relationship between injury-induced astrogliosis, laminin expression and axonal sprouting in the adult rat brain. J. Neurocytol: 23: 615-630. Stockli, K.A., Lillien, L.E., Naher-Noe, M., Breitfeld, G., Hughes, R.A., Raff, M.C., Thoenen, H. and Sendtner, M. (1991) Regional distribution, developmental changes, and cellular localization of CNTF-mRNA and protein in the rat brain. J. Cell BioL 115: 447-459. Streit, A., Faissner, A., Gehrig, B. and Schachner, M. (1990) Isolation and biochemical characterization of a neural proteoglycan expressing the L5 carbohydrate epitope. J. Neurochem. 55: 1494-1506. Streit, A., Nolte, C, Rasony, T. and Schachnert, M. (1993) Interaction of astrochondrin with extracellular matrix components and its involvement in astrocyte process formation and cerebellar granule cell migration. J. Cell BioL 120: 799-814. Takeichi, M. (1990) Cadherins: a molecular family important in selective cell-cell adhesion. Annu. Rev. Biochem. 59: 237-252. Thoenen, H. (1991) The changing scene of neurotrophic factors. Trends NeuroscL 14: 165-170. Thomas, K.A. (1993) Biochemistry and molecular biology of fibroblast growth factors. In: S.E. Loughlin and J.H. Fallon (Eds,), Neurotrophic Factors, Academic Press, San Diego, CA, pp. 285-312. Timpl, R., Rohde, H., Gehron Robey, P., Rennard, S.I., Foidart, J.-M. and Martin, G.R. (1979) Laminin-a glycoprotein from basement membranes. /. BioL Chem. 254: 9933-9937.

H.W. Maileretal. Tiveron, M.-C, Barboni, E., Rivero, F.B.P., Gormley, A.M., Seeley, P.J., Grosveld, F. and Morris, R. (1992) Selective inhibition of neurite outgrowth on mature astrocytes by Thy-I glycoprotein. Nature (London) 355: 745-748. Toru-Delbauffe, D., Baghdassarian, D., Both, D., Bernard, R., Rouget, P. and Pierre, M. (1992) Effects of TGF beta I on the proliferation and differentiation of an immortalized astrocyte cell line: relationship with extracellular matrix. Exp. Cell Res. 202: 316-325. Tranque, P.A., Calle, R., Naftolin, F. and Robbins, R. (1992) Involvement of protein kinase C in the mitogenic effect of insulin-like growth factor I on rat astrocytes. Endocrinology 131: 1948-1954. Tsukamoto, R., Kusakabe, M. and Saga, Y. (1991) In situ hybridization with non-radioactive digoxigenin-11-UTPlabelled cRNA probes-localization of developmentally regulated mouse tenascin mRNAs. Int. J. Dev. Biol. 35: 25-32. Tucker, R.P. and McKay, S.E. (1991) The expression of tenascin by neural crest cells and glia. Development 112: 1031-1039. Ueda,. S., Gu, X.F., Whitaker-Azmitia, P.M., Naruse, I. and Azmitia, E.G. (1994a) Neuro-glial neurotrophic interaction in the S-100 beta retarded mutant mouse (Polydactyly Nagoya). I. Immunocytochemical and neurochemical studies. Brain Res. 633:275-283. Ueda, S., Hou, X.P., Whitaker-Azmitia, P.M. and Azmitia, E.G. (1994b) neuro-glial neurotrophic interaction in the S100 beta retarded mutant mouse (Polydactyly Nagoya). II. Co-culture study. Brain Res. 633: 284-288. Unsicker, K., Reichert-Preibsch, H., Schmidt, R., Pettmann, B., Labourdette, G. and Sensenbrenner, M. (1987) Astroglial and fibroblast growth factors have neurotrophic functions for cultured peripheral and central nervous system neurons. Proc. Natl. Acad. Sci. USA 84: 5459-5463. Walicke, P.A. and Baird, A. (1988) Neurotrophic effects of basic and acidic fibroblast growth factors are not mediated through glial cells. Dev. Brain Res. 40: 71-79. Walicke, P.A. and Baird, A. (1991) Internalization and processing of basic fibroblast growth factor by neurons and astrocytes. J. Neurosci. 11: 2249-2258.

397 Walicke, P.A., Cowan, W.M., Ueno, N., Baird, A. and Guillemin, R. (1986) Fibroblast growth factor promotes survival of dissociated hippocampal neurons and enhances neurite extension. Proc. Natl. Acad. Sci. USA 83: 3012-3016. Wehrle, B. and Chiquet, M. (1990) Tenascin is accumulated along developing peripheral nerves and allows neurite outgrowth in vitro. Development 110: 401-415. Weller, A., Beck, S. and Ekblom, P. (1991) Amino acid sequence of mouse tenascin and differential expression of two tenascin isoforms during embryogenesis. J. Cell Biol. 112: 355-362. Whitaker-Azmitia, P.M. and Azmitia, E.G. (1986) Autoregulation of fetal serotonergic neural development: role of high affinity serotonin receptors. Neurosci. Lett. 67: 307-312. Wozniak, M., Rydzewski, B., Baker, S.P. and Raizada, M.K. (1993) The cellular and physiological actions of insulin in the central nervous system. Neurochem. Int. 22: 1-10. Wujek, J.R., Haleem-Smith, H., Yamada, Y., Lipsky, R., Lan, Y.T. and Freese, E. (1990) Evidence that the B2 chain of laminin is responsible for neurite outgrowth-promoting activity of astrocyte extracellular matrix. Brain Res. Dev. Brain Res. 55:237-247. Yamada, H., Watanabe, K., Shimonaka, M. and Yamaguchi, Y. (1994) Molecular cloning of brevican, a novel brain proteoglycan of the aggrecan/versican family. J. Biol. Chem. 269: 10119-10126. Yamaguchi, F., Saya, H., Bruner, J.M. and Morrison, R.S. (1994) Differential expression of two fibroblast growth factor-receptor genes is associated with malignant progression in human astrocytomas. Proc. Natl. Acad. Sci. USA 91: 484^88. Yamakuni, T., Ozawa, F., Hishinuma, F., Kuwano, R., Takahashi, Y. and Amano, T. (1987) Expression of ^-nerve growth factor mRNA in rat glioma cells and astrocytes from rat brain. FEBS Lett. 223: 117-121. Zimmermann, D.R. and Ruoslahti, E. (1989) Multiple domains of the large fibroblast proteoglycan, versican. EMBO 7.8:2975-2981.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 16

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival D.N. Ishii, S.F. Pu, G.W. Glazner, H.-X. Zhuang and D.J. Marsh Department of Physiology and Department of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, CO 80523, USA

1. Introduction The freedom to move, to be able to celebrate one's existence through dance, is a grand gift of nature enjoyed by humankind. Yet this gift can be lost all too readily; injury, particularly near the proximal regions of peripheral nerves, can increase the risk of permanent paralysis. The clinical rate of nerve regeneration is approximately 1 mm per day and 1-2 years may be required for axons to traverse from the proximal to the distal reaches of limbs. During such prolonged denervation, motor (Gutmann and Young, 1944; Drachman et al., 1967; Miledi and Slater, 1969) and sensory (Tower, 1932; Kubota et al., 1978; Barker et al., 1986) end-organs may undergo severe atrophy and degenerate to such an extent that successful reinnervation may no longer be possible. For example, the loss of muscle fiber can be essentially complete and only adipose and connective tissues remain (Gutmann and Zelena, 1962). Prolonged disconnection from target end-organs is also detrimental to neurons and motor neurons are eventually lost when axotomy is permanent (Kawamura and Dyck, 1981). A great hope is that one day it might become feasible to improve upon the rate of regeneration and thereby reduce the risk of paralysis. Therefore, it is of considerable interest to the basic and clinical neurosciences to identify those factors that are produced in nerve and nerve target organs, which determine the rate of nerve regeneration. One might also consider that by uncovering the princi-

ples behind successful regeneration in the peripheral nervous system, new approaches may be found to improve upon the severely limited regeneration in the central nervous system. Several excellent reviews summarize the status of research in this area (Salpeter, 1987; Carbonetto and Muller, 1982; Sanes and Covault, 1985; Windebank, 1993). The present commentary is confined, for the most part, to recounting what is known about the role of insulin-like growth factors (IGFs) in peripheral nerve regeneration in mammals. Other neurotrophic factors found in peripheral nerves and nerve target organs are briefly discussed. This discussion has been expanded to include the role of IGFs in motor neuron survival following nerve injury, and the data showing IGFs differ from classic neurotrophic factors in two important ways: (1) they are circulating factors, and (2) their action appears to be ubiquitous for the nervous system rather than confirmed to a few types of neurons. 2. IGF genes, transcripts and proteins Insulin-like activity in serum fractions was originally termed non-suppressible insulin-like activity, somatomedins (Daughaday et al., 1972) and multiplication-stimulating activity (Moses et al., 1980). The amino acid (Rinderknecht and Humbel, 1978a,b; Blundell and Humbel, 1980) and nucleotide sequences (Dull et al., 1984; Bell et al., 1985; Jansen et al., 1985; Soares et al., 1985) revealed conclusively that somatomedins are members of the insulin gene family. This resulted in their re-

400

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

naming as insulin-like growth factors (IGFs). In order to discuss the role of the IGFs in the nervous system, a few comments on their genes, transcripts and proteins are first in order. 2.1. IGF-IIgene The human IGF-II gene is located downstream of the insulin gene, on chromosome 11. There are nine exons in the 30 kb human gene, but only six exons in the mouse and rat genes. Multiple promoters and polyadenylation sites give rise to transcripts of various sizes, regulated in a tissueselective manner. Mature IGF-II is a neutral protein of 67 amino acids. 2.2. IGF'Igene The human and rat IGF-I genes each contain six exons and are in excess of 70 kb in size. The human gene maps to chromosome 12 (Brissenden et al., 1984). At least two promoters, alternative RNA splicing and differential polyadenylation, give rise to a large number of potential transcripts, regulated in a tissue-selective manner. All of the heterogeneous IGF-I mRNAs encode mature IGFI, but differ in coding peptides that flank the amino- and carboxyl-terminal regions of IGF-I. The physiological significance of multiple IGF transcripts is presently under study (Lund et al., 1991). Mature IGF-I is a basic polypeptide of 70 amino acids. Apparently, there is a regulatory mechanism to modulate levels of insulin and IGFs. Infusion of supraphysiological concentrations of rhIGF-I into test subjects produces a substantial decline in IGFII and insulin content in serum (Guler et al., 1989). 2.3. IGF-II mRNAs and proteins in the central nervous system In adult rats, the brain and spinal cord contain the highest concentration of IGF-II mRNAs (Soares et al., 1985, 1986; Murphy et al., 1987). These transcripts and encoded protein are particularly abundant in the choroid plexus (Ichimiya et al., 1988; Stylianopoulou et al., 1988). Other brain

regions are also reported to express IGF-II mRNA, such as hippocampus, cerebral cortex, and thalamus (Hynes et al., 1988; StyHanopoulou et al., 1988; Lee et al., 1992; Logan et al., 1994). The ability to detect low levels of IGF-II transcripts in various brain regions will depend on the sensitivity of in situ hybridization assays. Following ischemic injury, these transcripts are up-regulated in infarcted cortex, particularly in activated macrophages and astrocyte-like cells (Lee and Bondy, 1993; Beilharz et al., 1995). Glial, but not neuronal cells express IGF-II mRNAs (Rotwein et al, 1988). Cerebrospinal fluid contains higher levels of IGF-II activity than does serum (Hossenlopp et al., 1986), suggesting IGF-II accumulates as a result of active transport from the circulation or synthesis from the choroid plexus. A higher molecular weight form, termed 'big IGF-IF, may be a precursor to IGF-II and is detected in brain (Haselbacher et al., 1985). These studies suggest that IGF-II may play a special role in the nervous system. 2.4. IGF-I mRNAs and proteins in the central nervous system The brain also contains IGF-I mRNA (Lund et al.., 1986; Mathews et al., 1986; Shimatsu and Rotwein, 1987) and in rats, these transcripts are particularly abundant at embryonic day 14 during development (Rotwein et al., 1988; Bach et al., 1991). IGF-I mRNA are found in various regions of the brain, including hippocampus, cerebral cortex, olfactory bulb and brainstem (Bondy, 1991; Bach et al., 1991). Both neurons and glia express the IGF-I gene (Rotwein et al., 1988). Growth hormone can regulate both IGF-I and IGF-II mRNAs in brain (Mathews et al., 1986; Hynes et al., 1987). Immunoreactive IGF-I is present in brain (Sara et al., 1982; Noguchi et al., 1987) and IGF-I is found in cerebrospinal fluid (Buckstrom et al, 1984). It is possible that most brain IGF-I is present as a variant form of IGF-I, Des(l-3)IGF-I. This truncated IGF-I, missing the first three amino acids in the B domain, has decreased affinity for IGF-binding proteins (see Section 4) and may be

401

D.N. Ishii et al.

more biologically potent than IGF-I itself (Carlsson-Skwirut et al., 1989; Sara and Hall, 1990). 3. IGF receptors In addition to IGF mRNAs and proteins, IGF receptors are found in brain (Sara et al., 1982; Goodyer et al., 1984; Rosenfeld et al., 1984) and a peak is observed during fetal brain development (Sara et al., 1983; Valentino et al., 1990). The regional distribution of IGF-I (Bohannon et al., 1988) and IGF-II (Smith et al., 1988) binding has been studied by quantitative autoradiography. IGF receptors are found on the shafts and terminals of axons (Boyd et al., 1985; Caroni and Grandes, 1990). Certain properties of the type-I and type-II IGF receptors (Massague et al., 1980; Kasuga et al., 1982; Rechler et al., 1983) and their relationship to nerve regeneration are discussed in Section 9. The type-I (Heidenreich and Brandenburg, 1986) and type-II (McElduff et al., 1987) receptors are smaller in brain than in peripheral tissues, due to decreased glycosylation. The distribution of IGFs in brain, and presence of variant IGF receptors, together suggest that IGFs are likely to play important roles in the nervous system. 4. IGF-binding proteins Analysis of the action of IGFs requires consideration of the role of IGF-binding proteins (IGFBPs) (Martin and Baxter, 1986; Hardouin et al., 1987; Ocrant et al., 1990). Six IGFBPs have been cloned to date and they are related members of an IGFBP gene family. IGFBP-2 favors binding to IGF-II over IGF-I, demonstrating that these IGFBPs have differential affinity for the IGFs. IGFBP-2 is also the major IGFBP found in central nervous system (CNS) tissues and cerebrospinal fluid (Tseng et al., 1989). On the other hand, IGFBP-3 is present in the circulation and binds both IGF-I and IGF-II. Cerebrospinal fluid is enriched in IGFBP-5 (Roghani et al., 1989), which preferentially binds IGF-II (Hossenlopp et al., 1986). These examples show that the distribution of IGFBP is heterogeneous in various tissues. The binding proteins are

produced in many tissues, including neurons and glia (Lamson et al., 1989; Ocrant et al., 1990). The 'big' IGF-II binds to both IGFBP-2 and IGFBP-3. One purpose for circulating IGFBPs may be to sequester and decrease the concentration of free IGFs available to cross-occupy insulin receptors; this might protect against hypoglycemia. Infusion of rhIGF-I, for example, results in an increase in serum IGFBP-2. The half-life of IGFs is also prolonged as a result of sequestration to IGFBPs. Most studies show that IGFBPs bind to and decrease the activity of IGFs, but a few reports indicate that activity can be enhanced. Further study is needed before a clear understanding of the role of these interesting proteins is likely to emerge. 5. Axon regeneration in peripheral nerves The possibility that a soluble neurotrophic factor in nerves may support regeneration was first suggested by the studies of Forssman (1898) and Ramon y Cajal (1928). It was observed that regenerating axons grew towards the distal stump of a transected nerve. This led to the inference that a soluble substance was released from the distal stump. Subsequent investigators prepared nerve extracts that demonstrated the presence of neurotrophic activity (Lundborg et al., 1982; Politis et al., 1982; Longo et al., 1983). Schwann cells were found to be a source of chemoattractants that are capable of orienting growing axons into the distal nerve stump (Kuffler, 1986). Following the momentous discovery of nerve growth factor (fisubunit) (NGF) in the 1940s, various attempts were made to identify and study the actions of nerve-derived neurotrophic factors. The IGFs have emerged recently as the first example of an identified soluble neurotrophic substance produced in nerves that supports motor and sensory axon regeneration in vivo. The literatures describing synaptogenesis, regeneration and sprouting in the peripheral nervous system have generally been separate. The recent advances made in the understanding of the neurobiology of IGFs may now make it easier to appreciate a closer relationship between these phenomena. For example, experimental data suggest that

402

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

the biochemical events underlying axon regeneration and nerve terminal sprouting are quite similar in many ways, and certain aspects of the expression of the IGF genes in muscle observed during development are recapitulated during regeneration. 5.7. Effects of IGFs on neurite outgrowth in vitro The in vitro neurite outgrowth response to IGF may have been observed first by Both well (1982), who used a commercial preparation of IGF. This promising line of research, unfortunately, was discontinued when it was discovered that his commercial preparation contained several unidentified components. Initially unaware of Bothwell's studies, a second laboratory showed unambiguously that low concentrations of highly purified IGF-I and IGF-II could increase the number of cells with neurites and the average length of neurites in cultured human neuroblastoma cells (Recio-Pinto and Ishii, 1984; Ishii and Recio-Pinto, 1987) and embryonic chick sensory and sympathetic neurons (Recio-Pinto et al, 1986). Cultured spinal cord motor neurons are responsive as well (Caroni and Grandes, 1990). Adult neurons might lose responsiveness or may respond to different growth factors, but adult sensory neurons continue to extend neurites in response to IGFs (Fernyhough et al., 1993). 5.2, Effects of IGFs on neuronal survival and mitosis in vitro Two important attributes of neurotrophic factors include the capacity to enhance neurite outgrowth and support survival of neurons. In addition to increasing neurite outgrowth, low concentrations of IGFs are found to enhance the survival of cultured embryonic sensory and sympathetic cells (RecioPinto et al., 1986), as well as fetal cortical neurons (Aizenman and de Vellis, 1987). The effects on neurite outgrowth and survival were observed in cultures essentially devoid of non-neuronal cells (Recio-Pinto et al., 1986). This shows that IGFs can act directly on neurons to bring about these responses.

One unexpected finding was that insuHn stimulates thymidine incorporation (Recio-Pinto and Ishii, 1984) and both insulin and IGFs increase the population density of cultured human neuroblastoma cells (Ishii and Recio-Pinto, 1987). Neuroblastoma cells are most likely of sympathetic origin and it is believed that they may be arrested in the neuroblast stage of development, in part because IGFs are mitogens for sympathetic neuroblasts (DiCicco-Bloom and Black, 1988). However, the effects of IGFs on neurite outgrowth are independent of their effects on cellular proliferation, because neurites are stimulated whether neuroblastoma cells are in the quiescent or actively dividing stage of the cell cycle (Recio-Pinto and Ishii, 1984). The embryonic sensory, sympathetic and motor neurons used in the previously cited studies on neurite outgrowth were postmitotic. 5.3. IGF-dependent peripheral nerve regeneration In an important extension of the in vitro experiments described in Section 5.1, local infusion of IGF-I was shown to increase the distance of regeneration of sensory axons in lesioned rat sciatic nerves (Kanje et al., 1989; Sjoberg and Kanje, 1989). This was followed by the observation that infused IGF-II could increase the distance of motor axon regeneration in crushed sciatic nerves (Near et al, 1992). These effects were not transient. Of chnical importance, IGF-II infused near a lesioned nerve was capable of sustaining an increase in the rate of sensory axon regeneration (Glazner et al., 1993a). Moreover an anti-IGF antiserum, which recognized both IGF-I and IGF-II, caused a sustained decrease in the rate of sensory axon regeneration, showing that endogenous IGF activity in nerves was required constantly to maintain the spontaneous rate of regeneration. IGFs infused locally near the site of nerve crush are unlikely to be active over a distance of more than a few millimeters from the end of catheters. They would form complexes with IGFBPs, be removed by tissue perfusion or otherwise become quickly diluted. This suggests that the axon termi-

D.N. Ishii et al.

nals at or near the growth cone are receptive to locally infused IGFs. Moreover, local infusion of an anti-IGF antiserum can reduce the rate of axon regeneration, showing that endogenous IGFs produced at or near the growth cones are active. Hansson et al. (1986) showed that IGF-I immunoreactivity is increased in lesioned nerves. The immunoreactivity was reported to be associated with reactive Schwann cells, but not other types of cells. It is not known, however, whether IGF-I is produced in, or bound and incorporated by, the Schwann cells. At any rate, these combined data indicate that endogenous IGF-I and IGF-II may both contribute to IGF-dependent sciatic nerve regeneration. 5,4. Up-regulation of IGF gene expression in lesioned sciatic nerves An increase in IGF gene expression in lesioned nerves may support the IGF-dependent nerve regeneration. IGF-I and IGF-II mRNAs per poly(A)+ RNA, as well as per milligram tissue, are increased several-fold in denervated nerve (Glazner et al., 1994). A relatively sharp peak in IGF mRNA content is found at approximately six days after nerve crush. In situ hybridization studies show that IGF-I mRNA is increased early and intensely at the nerve crush site, and later is increased more moderately distal but not proximal to the site of crush (Pu and Ishii, 1993; Pu et al., 1995a). This is associated with the proliferated Schwann cells at and distal to the injury. Interestingly, there is a differential temporal and spatial expression of the IGF-II gene. IGF-II gene expression is not increased at the injury site, but is increased in the distal, intramuscular reaches of the nerves. This increase in IGF-II mRNAs occurs much later than the increase in IGF-I mRNAs, but is also associated with Schwann cells. Thus, there is a timedependent gradient in IGF mRNAs in the nerve, suggesting that IGF-I mRNAs may contribute to a greater extent to the elongation of axons within nerves, whereas IGF-II mRNAs may act upon the tips of axons much later as they return to muscle. IGF mRNAs are detected also in activated macrophages which invade lesioned nerves.

403

Surprisingly few changes in the relative abundance of specific mRNAs are observed in the distal nerves following crush (De Leon et al., 1991). In a cDNA library screened by differential hybridization, only 11 of 2000 cDNA clones were induced, whereas 30 of 5000 clones were repressed. The expression of other neurotrophic factor genes in nerves is discussed in Section 5.8. 5.5. Axon regeneration and down-regulation of IGF mRNAs IGF mRNAs are down-regulated in the distal nerve shortly after axons have regenerated (Glazner et al., 1993b). This down-regulation occurs following nerve crush, but not transection. IGF down-regulation, therefore, is closely associated in some way with the regeneration of axons. However, this down-regulation precedes the reestablishment of functional synapses with muscle. These data indicate that denervation relieves an axon-dependent feedback suppression of IGF gene expression and that regeneration re-establishes this feedback inhibition. A feedback regulation is also observed for IGF gene expression in muscle (see Section 6.2), although it is unclear whether the mechanisms are similar or different. It appears unlikely that a retrograde signal from muscle suppresses IGF gene expression in nerves. The up-regulation of the IGF genes following nerve lesion is not due to interruption of a retrograde signal from muscle, because up-regulation is not observed in the nerve proximal to the lesion. Moreover, the down-regulation of the IGF genes in crushed nerves begins long before axons have reached their muscle targets. 5.6. Schwann cell columns guide the regeneration of axons Events surrounding Wallerian degeneration in the nerve are well described (Lampert, 1967; Salzer and Bunge, 1980; Lubinska, 1982). Schwann cells proliferate in the basement membrane tubes of the formerly myelinated nerve fibers (Asbury and Johnson, 1978). Premitotic incorporation of thymidine in Schwann cells spreads proximo-

404

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

distally within 3-4 days in the distal nerve stumps (Oaklander et al., 1987). When Schwann cell division is prevented, Wallerian degeneration and axon regeneration is inhibited (Hall and Gregson, 1975). The delayed Wallerian degeneration in sciatic nerves of C57BL/01a mice results in impaired sensory, but not motor axon regeneration (Bisby and Chen, 1990). The bands of Bungner contain several axonal sprouts and Schwann cells in a basal lamina tube. The regenerating axons are essentially constrained within the Schwann cell columns and stimulated growth cones are guided by these axonal conduits. The Schwann cell basal lamina undoubtedly provides an important adhesive surface for the advancement of growth cones. Thus, although chemotaxis may be helpful, regenerating axons need not be strongly dependent upon it, because the direction of regeneration may be determined primarily by the basal lamina tubes. Transection destroys the continuity of tubes. The efficiency of regeneration is reduced greatly because it is very difficult to surgically realign the basal laminae columns. Functional specificity may be lost if proximal stump axons enter inappropriate distal stump Schwann cell columns. This, for example, can lead to the growth of motor axons into sensory nerve branches. Nevertheless, preferential reinnervation of the appropriate femoral motor nerve branch is observed (Brushart, 1990). After nerve transection, motor axons form collateral branches that randomly regenerate into sensory, as well as motor, nerve branches, but the specificity of regeneration is later improved by the subsequent retraction of incorrectly projecting collateral axonal branches. 5.7. Neurotrophic role of IGFs in regenerating nerves The dependence of motor and sensory axon regeneration on endogenous IGF in nerves, upregulation of IGF mRNAs in lesioned nerves and the observation that IGF-I immunoreactivity is increased in Schwann cells, together support, rather strongly, the hypothesis that IGFs are solu-

ble nerve-derived neurotrophic factors that help regulate nerve regeneration. In situ hybridization shows that there is a temporal and spatial gradient of IGF mRNAs in the distal nerve (Pu et al., 1995a).Thus, a gradient of IGFs along the longitudinal axis of the distal nerve may help stimulate the regeneration of axons. In addition, it is possible that IGFBPs may produce a microenvironment in which there is a gradient of IGFs near the surface of up-regulated Schwann cells. It remains unclear how sharp the IGF gradient is, or needs to be, along the longitudinal axis of the nerve. We speculate that the high concentrations of IGFs within lesioned nerves may stimulate axonal sprouting and collateral branching during regeneration. The subsequent down-regulation of IGF content may permit retraction of incorrectly projecting collateral branches. This explanation is similar to that for the retraction of incorrect polyneuronal innervation during development. That, too, may involve retraction of inappropriate axonal branches to end-plates after IGF levels decrease. Once IGF levels in nerve have returned towards baseline, following regeneration of axons, it is possible that weaker forces may play a role in the retraction of inappropriate axonal branches. The nature of the presumed weaker forces is not understood. One possibility is that the amount and type of trophic support provided by different branches of nerves and nerve targets may vary in the basal state. Altematively, the correctly projecting motor axons in motor nerve branches may be retained because IGF-II is elevated in the intramuscular reaches of the nerves as well as in the muscle surrounding these nerves. The temporal relationship in which IGF-II is more slowly increased and decreased in muscle, relative to IGF-I in nerve, further supports this possibility (Glazner et al., 1993b; Glazner and Ishii, 1995; Pu et al., 1995a). IGF-II gene expression is not detectably increased in denervated skin (Glazner and Ishii, unpubHshed). Axons in intact nerves normally keep the IGF mRNA content in the non-neuronal cells at a low level and this basal IGF mRNA content may serve a vital function. The Schwann and other cells may

D.N. Ishii et al.

continuously release low levels of IGFs that stimulate IGF receptors found all along the axon shaft. This may be quite important because, in all likelihood, neurons have a trophic factor requirement that is proportional to the length of their axons. The total neurotrophic requirement of neurons may be satisfied, at least in part, by other sources, such as target organs, other cells in proximity to the neural soma, and, potentially, the circulation or cerebrospinal fluid. Interruption of these sources of IGFs may be pathologic for neurons, and recent data implicate a decline in IGF activity in the pathogenesis of diabetic neuropathy (Wuarin et al., 1994; Zhuang et al., 1994; Ishii, 1995; Ishii and Lupien, 1995; Zhuang et al., 1996). 5.8. Role of other nerve-derived soluble neurotrophic factors in nerve regeneration Are other nerve-derived neurotrophic factors capable of supporting regeneration in adult mammalian motor and sensory nerves? Other soluble neurotrophic factors, in fact, are known to be present in nerves and have been found to support neurite outgrowth in cultured embryonic neurons (see Loughlin and Fallon, 1993). These studies provide important indications that other active regenerative factors are present in neural tissue. It is not necessarily the case that factors found to be active in vitro will be found to be active in vivo. Even in vivo, silastic chambers, nerve guidance channels and the like do not quite replicate the specific environment that regenerating axons must traverse in nerves. Tissue culture and nerve guidance channels are informative and valuable research tools, but one would particularly like to find those procedures that can enhance axon growth within its normal environment, the nerve. The current state of the art does not yet permit perfect artificial replication of the myriad of positive and negative influences in nerves that may influence the response to test neutrotrophic substances. The trophic requirement can be very different between embryonic and mature neurons. It is well known, for example, that prenatal sensory neurons are much more dependent on NGF for survival and neurite growth than are postnatal sensory neurons.

405

Thus, in evaluating the potential usefulness of a neurotrophic factor for improving the rate of nerve regeneration, one needs to consider whether such a factor is active on mature neurons in the normal nerve environment in intact mammals. The effects of several well-known nerve-derived neurotrophic factors on vertebrate sensory and motor neurons are briefly discussed next. As will be seen, there is unfortunately, a paucity of information concerning the capacity of these factors to increase regeneration in nerves of mature vertebrates. The reader is referred to other articles in this volume as well as Loughlin and Fallon (1993) for up-to-date discussion on the many other fascinating and important effects of these neurotrophic factors in various neural populations. 5.8.L Nerve growth factor NGF supports neurite outgrowth and survival of cultured embryonic sensory neurons (LeviMontalcini and Angeletti, 1968). It supports neurite growth, but neither survival of ventral spinal cord neurons in vitro (Longo et al., 1982; Wayne and Heaton, 1990) nor in vivo (Oppenheim et al., 1988). Although NGF gene expression and protein are increased in lesioned sciatic nerves of adult rodents (Heumann et al., 1987a,b), infusion of NGF, nevertheless, has no influence on the distance or numbers of regenerating sciatic nerve axons or extent of reinnervation (Diamond et al., 1987; Kanje et al., 1989; Hollowell et al., 1990). These data are consistent with the finding that highaffinity NGF receptors are down-regulated on neurons and retrograde NGF transport is reduced in lesioned nerves (Raivich et al., 1991). Following nerve transection in neonatal rats, NGF administration causes a greater loss of motor, but attenuated loss of sensory neurons (Miyata et al., 1986). NGF administration can also spare the death of adult sensory neurons (Rich et al., 1987). There is controversy concerning the role of NGF in adult neurons, because it is reported that NGF is not required for survival of adult sensory neurons in culture (Lindsay, 1988) and anti-NGF antiserum has no effect on sensory neuron survival in vivo (Gorin and Johnson, 1980; Rich et al., 1984). One

406

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

possibility is that NGF is acting indirectly to support adult sensory neuron survival. 5.8.2. Other neurotrophins NGF, brain-derived neurotrophic factor and neurotrophin-3 are encoded members of the neurotrophin gene family. Brain-derived neurotrophic factor and neurotrophin-3 are capable of supporting the survival of embryonic sensory neurons in vitro (Lindsay et al., 1985; Lindsay, 1988) and in vivo (Hofer and Barde, 1988). Brain-derived neurotrophic factor mRNA is increased in transected sciatic nerve (Meyer et al., 1992), but with a time course much slower than for IGF or NGF mRNAs. Schwann cells and skeletal muscle are tissues in which brain-derived neurotrophic factor immunoreactivity is found. Brain-derived neurotrophic factor and neurotrophin-3 undergo retrograde transport to the dorsal root ganglia and ventral spinal cord, suggesting that they might be active in vivo, but 'evidence that adult neurons respond to brain-derived neurotrophic factor and neurotrophin-3 in vivo is lacking' (DiStefano et al., 1992). 5.8.3, Ciliary neurotrophic factor The actions of ciliary neurotrophic factor have been reviewed recently (Manthorpe et al., 1993). It can support the survival of embryonic sensory neurons in vitro, but not in vivo (Oppenheim et al., 1991). It also prevents the developmentally programmed death of spinal cord motor neurons in embryonic chicks (Wewetzer et al., 1990; Oppenheim et al., 1991) and of brainstem motor neurons, which follows transection of the facial nerve (Sendtneretal., 1990). There are some reports that ciliary neurotrophic factor might enhance neurite outgrowth in cultured neurons, but significant reservation has emerged as to whether this is a specific role (Manthorpe et al., 1993). Ciliary neurotrophic factor mRNA is found in Schwann cells, but not in skeletal muscle (Stockli et al, 1989). These mRNAs are unexpectedly down-regulated in regenerating nerves, which is considered to be inconsistent with a role in neuronal injury, because ciliary neurotrophic factor levels are decreased while the requirement for neu-

rotrophic support is increased (Rabinovsky et al., 1992). In addition, because its cDNA nucleotide sequence does not encode a recognized aminoterminal signal sequence for secretion, it remains unclear how ciliary neurotrophic factor might be released from Schwann cells. 5.8.4. Fibroblast growth factor Fibroblast growth factor does not support the survival of cultured sympathetic or dorsal root ganglion sensory neurons (Unsicker et al., 1987). Also, survival of spinal cord motor, sympathetic and dorsal root ganglion neurons in developing chicks is not increased (McManaman et al., 1990; Oppenheim et al., 1990). Curiously, fibroblast growth factor administration can increase survival of sensory neurons after sciatic nerve transection in adult rats (Otto et al., 1987), although retrograde transport is not observed (Ferguson et al., 1990). This suggests that indirect actions might be involved. Neurite growth and axonal myelination is enhanced when fibroblast growth factor is applied in nerve guidance channels (Danielsen et al., 1988), but a positive effect of fibroblast growth factor on axon regeneration in nerves has not been reported. Additional valuable discussion is available (Unsicker et al., 1993). The pace of research is advancing rapidly for these and other neurotrophic factors. Although they have not been found to enhance peripheral nerve regeneration when administered alone, it remains possible that combinations of other neurotrophic factors, together with IGFs, may further augment the regeneration rate following nerve lesion. 6. Regulation of IGF gene expression in muscle 6.1. Transition in IGF gene expression at the neuromuscular junction Regenerating axons enter a transition zone at the neuromuscular junction; the pattern of IGF gene expression differs in muscle and nerve (Glazner et al, 1994; Glazner and Ishii, 1995). IGF-I mRNA content per milligram tissue is sig-

D.N. Ishii et al.

nificantly higher in nerve than in muscle, whether these tissues are denervated or not. Upon denervation, IGF-I mRNA levels increase substantially more in nerve than in muscle. In contrast the IGFII mRNA content per mg tissue is significantly higher in muscle than nerve, whether these tissues are denervated or not. IGF-II mRNA content increases many-fold in denervated muscle, whereas IGF-I mRNA content changes little. The increase in IGF-II mRNA in muscle clearly cannot be involved in signaling to regenerating axons at a distant nerve lesion. This is evident because axon regeneration begins following a lag time of about 1-2 days, whereas IGF-II mRNA in muscle only begins to accumulate after approximately six days (Glazner and Ishii, 1995). Moreover, IGF-II released from muscle is likely to form complexes and quickly become inactivated by IGFBPs and thereby become unavailable to distant regenerating axons. Neurite outgrowth occurs equally well in response to IGF-I and IGF-II in cultured neurons (Recio-Pinto and Ishii, 1984; Recio-Pinto et al., 1986; Wang et al, 1992). Also, IGF-I and IGF-II both can increase axon regeneration in vivo (Kanje et al., 1989; Near et al., 1992; Glazner et al., 1993a). Thus, axon elongation per se may not be modified greatly by a transition from a predominantly IGF-I environment in nerve to a predominantly IGF-II environment in muscle. This transition might have other consequences for the growth cone, however. As advancing axonal growth cones approach their muscle targets, IGF-II released from muscle may play an important paracrine role. The formation of complexes with IGFBPs is likely to enforce a gradient of free IGF-II close to the surface of muscle and such a gradient may help attract approaching growth cones. As discussed in Section 6.1, the growth cone may be exposed primarily to IGF-I in nerve and primarily to IGF-II in muscle. The transition from nerve to muscle includes additional components of the IGF signaling system. The IGF-I produced in nerves would bind tightly primarily to the type-I IGF receptors on axons and with much lower-affinity for type-II

407

IGF receptors. But the IGF-II released from muscle would have a high affinity for type-II, as well as type-I, IGF receptors. The type-II receptor is also the mannose-6-phosphate receptor (Morgan et al., 1987; MacDonald et al., 1988) and muscle IGF-II may modulate the binding of lysosomal enzymes containing mannose-6-phosphate moieties that bind to the type-II receptors. One potential function of IGF-II in denervated muscle, then, may be to modify the physiology of the growth cone and to signal changes in the neuronal soma. To form a synapse, the growth cone needs to make a transition from a mobile to an immobile state and undergo other biochemical and morphological adjustments. Cytological changes are seen in situ when growth cones reach their targets (Mason and Gregory, 1984; Harris et al., 1985). 6.2, The IGF-II gene is under feedback inhibition in muscle Nerve transection or crush is observed to relieve a feedback inhibition on IGF-II gene expression in muscle (Ishii, 1989; Glazner and Ishii, 1995). Chemical denervation produced by botulinum toxin has the same effect. IGF-II gene expression down-regulates following the re-establishment of synapses after crush. On the other hand, gene expression remains up-regulated following nerve transection, a condition in which synapses do not regenerate. Thus, IGF-II gene expression in adult muscle is under feedback inhibition by neuromuscular activity. The nature of the feedback inhibitory signal has yet to be elucidated. IGF-II mRNA content is very low in mature innervated muscle. Both quantal and non-quantal leakage release of acetylcholine (Katz and Miledi, 1977) may help suppress IGF-II gene expression. On the other hand, other events, such as those associated with muscle contraction, might produce an inhibitory signal. In contrast to the response in neonatal muscle (see Section 8.1), nerve transection does not upregulate IGF-I gene expression in adult muscles. The reason for this developmental loss of regulation by innervation has not been investigated.

408

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

7. IGF role in nerve sprouting Intramuscular sprouts can emerge from motor nerve terminals and nodes of Ranvier under a variety of conditions (Brown et al., 1981). Sprouting is observed during synapse regeneration, muscle paralysis resulting from pre- and postsynaptic blockade and partial denervation. Partially denervated muscle can attract sprouts from intact adjacent motor nerve terminals located within a distance of 200/^m, indicating that a soluble factor is released from partially denervated muscle (Pockett and Slack, 1982). IGFs can increase the sprouting of motor nerve terminals. Administration of IGF-I or IGF-II to undamaged muscles can induce sprouting in rats and GAP43 immunoreactivity is also increased (Caroni and Grandes, 1990). The sprouting induced by botulinum toxin A has a similar basis. This toxin inhibits the release of acetylcholine and induces terminal sprouts that can form synapses on nearby end-plates in rodents (Fex et al., 1966; Duchen and Strich, 1968). By relieving feedback inhibition on the IGF-II gene, botulinum toxin significantly increases IGF-II mRNA content in muscle (Ishii et al., 1989). This, in itself, does not reveal whether IGF-II is the sprouting factor, because other factors might be increased as well. However, Caroni et al. (1994) showed that the sprouting induced by botulinum toxin could be inhibited by infusing an IGFBP that would complex with and neutralize muscle-derived IGF-II. IGF-II mRNAs in muscle are transiently increased during regeneration (Glazner et al., 1993b). Thus, it is not surprising that collateral and terminal sprouts are transiently observed when regenerating motor axons reach muscle (Brown and Ironton, 1978; Taxt, 1983). We propose that conditions that inhibit neuromuscular activity also relieve feedback inhibition on IGF-II gene expression in muscle, and rat or human IGF-II mRNA content can be expected to increase. A prediction of this hypothesis is that sprouting will occur in response to nerve block with tetrodotoxin, or pre- and postsynaptic blocks with botulinum toxin or a-bungarotoxin, respec-

tively. Partial denervation should cause an increase in IGF-II mRNA production in denervated muscle fibers and promote sprouting from nerve terminals on adjacent muscle fibers. Sprouting from motor nerve terminals and intramuscular nodes of Ranvier would be consistent with the expected high intramuscular content of IGF-II in chemically or physically denervated muscles. This hypothesis might also explain the sprouting of motor nerves observed clinically in patients treated for strabismus with botulinum toxin A (Holds et al., 1990) or afflicted with neuromuscular or motor neuron disorders, such as amyotrophic lateral sclerosis. It is not understood clearly, however, why sprouting occurs predominantly from nerve terminals following botulinum toxin treatment, whereas sprouting occurs from both nodes of Ranvier and terminals in nerve regeneration. An IGF dose effect on axons might be responsible, because IGF-I and IGF-II mRNAs are increased in nerve and IGF-II mRNA in muscle, during nerve regeneration. By contrast, one would not expect botulinum toxin to increase IGF mRNA content in nerves. 8. Putative role for IGF in synapse regeneration It is clear that denervated and innervated muscles are in very different receptive states for the establishment of synapses. A foreign nerve routed near the surface of a denervated adult muscle will form many synapses, whereas synapses generally do not form on an innervated adult muscle (Elsberg, 1917; Hoffman, 1951). In this respect, denervated adult muscle is similar to immature muscle. It is worth noting that this receptive state is not limited uniquely to the end-plate region. Functional synapses can form even at ectopic sites anywhere on the surface of denervated mature muscle fibers (Fex et al., 1966). The question is raised as to whether regenerating axons can induce synapses, because denervated muscle becomes generally receptive to innervation. The non-receptive state is equally important, because it discourages further innervation and potential loss of functional specificity. There is a close correlation between the receptive state of

D.N. Ishii et al.

immature and denervated mature muscles and IGF-II gene expression. 8.1. IGF gene expression and synaptogenesis during development Polyneuronal innervation is a condition in which multiple nerve terminals form synapses at a single end-plate (Redfern, 1970). A correlation is found between IGF-II gene expression in developing muscle and polyneuronal innervation (Ishii, 1989). The IGF-II gene is expressed on embryonic day 14 prior to the formation of the first synapse (Kelly and Zacks, 1969). We speculate that the fusion of myoblasts and the formation of myotubes might activate IGF gene expression and thereby signal the onset of synaptogenesis. IGF gene expression increases, together with the accumulation of multiple nerve terminals at single endplates, until the time of birth. It is postulated that development predisposes to polyneuronal innervation, because advancing axons cannot distinguish between innervated and naive myotubes in the muscle environment of extremely high IGF content (Ishii, 1989). In order for IGFs produced in muscle to influence synaptogenesis, it would be necessary for IGF receptors to be present on growth cones, and they are enriched in growth cone preparations from brain (Quiroga et al, 1995). Polyneuronal innervation is lost during the first 2 weeks of life (Brown et al., 1976; O'Brien et al., 1978). There is a 100-fold increase in miniature end-plate potentials between the first and second postnatal week of life (Diamond and Miledi, 1962; Nakajima et al., 1980). This coincides very closely with the postnatal decline in IGF-II gene expression. It is postulated that the increased neuromuscular activity (miniature end-plate potentials), through heightened feedback inhibition, causes down-regulation of IGF-II gene expression. The decline in IGF-II activity results in an environment that can no longer support multiple synapses and polyneuronal innervation is lost. IGF-I gene expression is also transiently induced near the time of birth and IGF-I might contribute further to polyneuronal innervation

409

(Glazner et al., 1993b). IGF-I mRNA (Beck et al., 1987) and immunoreactivity (Jennische and Olivecrona, 1987) are low in myoblasts, high in myotubes and low in mature innervated muscle fibers. The loss of multiple synapses may follow the postnatal decline in the combined activity of IGF-I and IGF-II and possibly other factors. IGF-I gene expression, in contrast to IGF-II gene expression, is down-regulated prior to the increase in miniature end-plate potentials (Glazner et al., 1993b). Nevertheless, both IGF genes are regulated by neuromuscular activity because nerve transection causes IGF-I and -II mRNAs to accumulate in postnatal muscle. Other studies support the hypothesis that IGFs may contribute to polyneuronal innervation. Botulinum toxin inhibits the presynaptic release of acetylcholine. By relieving feedback inhibition caused by neuromuscular activity, it causes upregulation of IGF-II mRNA content in muscle (Ishii et al., 1989). Additionally, this toxin retards the loss of multiple synapses (Brown et al., 1981). Similarly, inhibition of the developmental loss of polyneuronal innervation caused by administration of tetrodotoxin (Thompson et al., 1979) and abungarotoxin (Duxson, 1982) is potentially due to up-regulation of IGF mRNAs. The pattern of IGF gene expression in various organs during development is described elsewhere (Bondy et al., 1990; Leeetal., 1990). 8.2. IGF gene expression and synapse formation in adult muscle As in nerve, soluble neurotrophic activity is found in muscle (Bennett et al., 1980; Dribin and Barrett, 1980; Obata and Tanaka, 1980). Musclederived neurotrophic activity is observed to increase following denervation (Bennett et al., 1980; Henderson et al., 1983). It seems likely that IGFs contribute to soluble, muscle-derived neurotrophic activity. There is a close association between sprouting, formation of multiple synapses and up-regulation of IGF-II gene expression in adult muscle, just as there is in developing muscle. This might be an important regulatory mechanism in the nervous

410

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

system, for example, to increase motor unit size should innervation of an adjacent muscle fiber fail. Those conditions that are predicted to upregulate IGF-II gene expression in mature muscle are known to induce sprouting and synapse formation. By locally infusing tetrodotoxin in a cuff around nerves, Taxt (1983) was able to block Na+channel-dependent action potentials and found that motor nerve terminals sprouted and formed polyneuronal synapses. Likewise, the sprouts induced by botulinum toxin form polyneuronal synapses (Duchen and Strich, 1968; Holland and Brown, 1981). The studies on IGFs, however, have not revealed a causal link with selective elimination of inappropriate polyneuronal synapses (Betz et al., 1990; Van Essen et al., 1990). Selective elimination may be governed by weaker forces that come into play after IGF activity wanes. Following nerve crush, axons regenerate back to end-plates, sprout and transiently form polyneuronal synapses (McArdle, 1975; Benoit and Changeux, 1978; Brown and Ironton, 1978). This observation correlates with the transiently upregulated IGF-II mRNA content in denervated gastrocnemius muscle (Glazner and Ishii, 1995) and soleus muscles (Pu and Ishii, 1995). Taken together, these data support the hypothesis that IGF-II may contribute to synapse regeneration. After axons regenerate back to muscle, the nerve terminals sprout, advance randomly over the surface of muscle fibers, and form polyneuronal synapses predominately in the region of muscle containing end-plates. Recent in situ hybridization studies show that IGF-II mRNA is increased much more at the middle than at the ends of denervated muscles (Pu, Zhuang, Marsh and Ishii, unpublished). IGF-I mRNA is only slightly increased. Immunohistochemistry reveals further that IGF-II is increased mostly in the middle of denervated muscles (Marsh et al., 1994). The middle region of muscle fibers is enriched in end-plates. However, the increases in IGF-II mRNA and protein are not localized exclusively at end-plates, but occur regionally in the middle of muscle. This pattern of expression is precisely that which would support a random sprouting of nerve terminals throughout the region containing end-plates, thereby maximiz-

ing the probability that all end-plates will be reinnervated. The slot blot and in situ hybridization data together suggest that the sprouting factor in muscle during regeneration is more likely to be IGF-II rather than IGF-I. The sprouting from intramuscular nodes of Ranvier may be provoked by the high IGF-II gene expression in muscle as well as in intramuscular nerve. For various fetal and adult rat tissues, including muscle, IGF mRNA and IGF protein content are closely correlated (Romanus et al, 1988). Co-localization of IGF-II mRNAs and translated pre-pro-IGF-II is observed in various tissues during embryogenesis (Lee et al., 1990). It would be enormously important if IGFs were found to contribute to synapse regeneration. The longer the duration of denervation, the more difficult it is to re-establish functional synapses (Gutmann and Young, 1944). This is due, in part, to atrophy, degeneration and disappearance of muscle cells (Gutmann and Young, 1944; Drachman et al., 1967; Miledi and Slater, 1969). Additionally, it involves loss of spinal motor neurons, which is observed clinically in long-term axotomy (Kawamura and Dyck, 1981). A better understanding of the role of IGFs in synapse regeneration may lead to new therapies, which may diminish the specter of permanent loss of function following injury. These results in peripheral nerve may be seen as a general paradigm, and IGFs may play a similar role in the central nervous system. Indeed, deafferentation induces axonal sprouting as well as an increase in IGF mRNA in hippocampus (Guthrie et al., 1995). 8.3. Muscle basal lamina and synapse regeneration Regenerating axons were shown to grow back to the residual muscle basal lamina in vivo, even after destruction and phagocytosis of the muscle fibers (Sanes et al., 1978). Co-culture studies showing that some axons synapse on basal lamina near denervated end-plates are more difficult to interpret because only a small fraction of total axons are observed to behave in this manner (Covault et al., 1987). One interpretation of these

411

D.N. Ishii et al.

data is that the basal lamina may guide regenerating axons back to the previous synaptic site. Several lines of evidence, however, argue against this intriguing and still widely held interpretation. The muscle basal lamina is unlikely to be able to 'guide', or 'attract', growing axons. In particular, the insoluble extracellular matrix has no readily conceivable means to communicate with distant regenerating axons. As discussed in Section 5.6, the Schwann cell columns generally constrain the growth of regenerating axons. Axons ordinarily have little freedom but to return to their original end-plates, irrespective of the muscle basal lamina. This would be true of motor axons, which use the empty Schwann cell columns for guidance back to their former endplates (Hoffman, 1950). On the other hand, terminal sprouts also appear to prefer original end-plate sites, despite the absence of Schwann cell columns for guidance. This phenomenon may have other explanations (see below). The basal lamina hypothesis infers that there is a fixed synaptic site that attracts regenerating axons. However, synaptic sites are not fixed, but can be induced anywhere along the length of a denervated muscle fiber by contact with axons (Elsberg, 1917; Fex et al., 1966; Frank et al., 1974). Agrin released from axons may aid in this process by inducing the clustering of acetylcholine receptors. Finally, the formation of muscle basal lamina in embryos occurs after synaptogenesis (Chiu and Sanes, 1984); the basal lamina would be unavailable to attract axons to a particular site on developing muscle. The basal lamina at the synaptic site contains certain antigens that change in distribution during synaptogenesis. Although one can not exclude a priori the possibility that the mechanism for synapse formation is different during development and regeneration, it would be more efficient if the same mechanism were shared. Are there other roles for the basal lamina during regeneration? Once axons are guided to the muscle basal lamina by Schwann cell columns, adhesive molecules in the muscle basal lamina may play a 'docking' function to help stabilize synapses. Existing end-plates, thereby, may be better able than other surfaces of muscle to retain contacts with

growth cones. This, together with the observation that IGF-II is increased in the middle of muscle fibers, might explain why terminal sprouts seem to prefer former end-plate sites. One hypothesis currently being tested is that IGF content is particularly high at end-plates due to selective IGF gene expression in the nuclei known to be clustered there. The faster formation of synapses with existing end-plates may discourage the further induction of extrasynaptic end-plates by axons. Adhesion to muscle basal lamina can also cause nerve terminals to acquire active zones and accumulate synaptic vesicles (Sanes et al., 1978). 8.4. Relationship of IGF-11 gene expression during nerve regeneration to muscle growth IGF-II gene expression and muscle growth are not positively correlated during regeneration. For example, up-regulation of IGF-II gene expression in denervated muscle occurs while muscle undergoes atrophy (Ishii, 1989; Glazner and Ishii, 1995). The recovery of muscle weight after reestablishment of synapses coincides with the down-regulation of IGF-II mRNAs. Thus, a negative correlation of IGF-II gene expression with muscle weight is observed during synapse regeneration, as well as during postnatal development (Ishii, 1989). The observations cited above do not imply, however, that IGF-II gene expression is not correlated with muscle growth under other conditions. For example, implantation of cells over-expressing growth hormone in rats causes IGF-II mRNAs to accumulate in muscle and muscle to hypertrophy (Turner et al., 1988). Heterozygous mice with the IGF-II gene disrupted by homologous recombination are smaller in body size by 40% and many pups die at or soon after birth (DeChiara et al., 1990). Thus, the IGF genes are under complex regulation to serve several muscle functions. 9. Mechanism of IGF-dependent axon regeneration In cultured neuroblastoma cells, the enhancement of neurite outgrowth by IGF-I and IGF-II is

412

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

closely correlated with occupancy of the type-I IGF receptor (Recio-Pinto and Ishii, 1988a). This receptor has two a- and two ^-subunits, a tyrosine kinase domain on the intracellular portion of the Psubunit and is similar in structure to the insulin receptor (Kasuga et al, 1982; Jacobs et al., 1983). The type-II IGF receptor (Morgan et al., 1987; MacDonald et al., 1988) is a single polypeptide chain with a large extracellular and a short intracellular domain. It is not a tyrosine kinase and may activate G proteins. There is a close correlation between the concentrations of IGFs that increase neurite outgrowth in cultured neuroblastoma cells (Recio-Pinto and Ishii, 1984; Ishii and Recio-Pinto, 1987) and those that increase the expression of genes encoding structural proteins of neurites, such as a- and Ptubulins (Mill et al., 1985; Femyhough et al., 1989; Wang et al., 1992). Tubulin mRNAs accumulate due to stabilization by IGF-I (Femyhough et al., 1989) and IGF-II (Wang et al., 1992). The tubulin heterodimers assemble to form microtubules, essential structural components of axons also involved in axonal transport. Tubulin mRNAs in neurons are increased during the time of rapid neurite growth in developing rodent brain (Bond and Farmer, 1983), as well as during regeneration of the goldfish optic nerve (Neumann et al., 1983). Tubulin mRNAs are also increased in regenerating dorsal root sensory (Hoffman and Cleveland, 1988) and spinal cord motor neurons (Miller et al., 1989) in rats. Thus, the increase in IGF gene expression in regenerating nerve and muscle may contribute to increased tubulin gene expression in neurons. A similar, but more complex, correlation emerges for the expression of the neurofilament light (NFL) and neurofilament medium (NFM) genes. These intermediate filaments are considered to be a major determinant of axonal caliber (Hoffman et al, 1984; Lasek et al., 1984). IGF concentrations that induce neurite growth also increase the expression of the NFL and NFM genes in cultured neuroblastoma cells (Wang et al., 1992). The NFL and NFM mRNAs (Julien et al., 1986) and proteins (Shaw and Weber, 1982; Garden et al., 1987) are increased during neural

development. These are conditions in which the axonal caliber is increasing. On the other hand, the diameter of regenerating axons is reduced during the initial phase of nerve regeneration and recovers thereafter (Kreutzberg and Schubert, 1971). Accordingly, there is a transient down-regulation of NFL and NFM gene expression in neurons following nerve lesion (Hoffman et al., 1987; Goldstein et al., 1988; Wong and Oblinger, 1990). It remains to be determined whether up-regulated IGF gene expression in regenerating nerve and muscle contributes to these changes in tubulin and NF gene expression in neurons. Further details and other mechanistic aspects, such as the role of protein kinase C and second messengers, are discussed elsewhere (Ishiietal., 1991; Ishii, 1993). 10. IGFs prevent motor neuron death following nerve injury The long-term disconnection of motor neurons from muscle is known to result in atrophy and degeneration of motor neurons, both clinically and experimentally (Gutmann and Young, 1944; Kawamura and Dyck, 1981; Greensmith and Vrbova, 1992; Snider et al., 1992). It is critical to identify those factors that may support motor neurons survival in order to avoid permanent paralysis. Curiously, the length of the remaining nerve stump following nerve transection influences motor neuron survival (Schmalbruch, 1984; Kashihara et al., 1987). Survival is greater with longer stumps. IGF-II is required for the survival of motor neurons following axotomy in neonatal rats (Pu et al., 1995b). For example, administration of IGF-II in gel foam to the proximal nerve stump increases, whereas an anti-IGF-II antiserum decreases motor neuron survival. Because IGF is produced in Schwann cells throughout the length of nerve, these results may explain the observation that the degree of motor neuron survival is proportional to the length of the remaining nerve stump following transection. The administration of IGF-I to embryonic chicks can prevent the developmentally programmed death of motor neurons (Neff et al..

413

D.N. Ishii et al.

1993). IGFs are important survival factors in postnatal as well as prenatal life. Polyneuronal innervation is suggested to help secure more neurotrophic factors to support motor neuron survival (Landmesser, 1992). This seems logical because a greater amount of neurotrophic factor might be secured by a larger number of synapses. Motor neuron survival might be dependent on IGFs obtained through multiple sources including (1) polyneuronal innervation during nerve regeneration and development, (2) increased production of IGF-II in muscle during nerve regeneration and development, (3) increased IGF production in nerve Schwann cells and macrophages during regeneration, (4) length of nerve stump producing IGFs, (5) IGFs in the circulation, and (6) IGFs in the cerebrospinal fluid. 11. IGFs are circulating neurotrophic factors The classic neurotrophic factors are produced locally at end-organs and act on a limited number of the various specific types of neurons. For example, the action of NGF in brain is confined to the catecholaminergic neurons. The classic neurotrophic factors (NGF, BDNF, NT3, NT4, etc.) are not present in significant quantities in the circulation. It is well known that both IGF-I and IGF-II are important growth factors in early development. However, circulating IGF levels remain persistently elevated decades after the growth spurt (Hall and Sara, 1984). This suggests roles other than growth for circulating IGFs. We believe that IGFs are maintenance factors for the nervous system. Circulating IGF-II levels are several-fold higher than IGF-I. This predominance in IGF-II might be expected, because the close regulation of IGF-I by growth hormone may be undesirable for a circulating neurotrophic factor. The blood-nerve barrier is not an impediment to the actions of circulating IGFs on peripheral neurons. Sensory and sympathetic ganglia are served by special fenestrated capillaries which permit molecules as large as Mr 44 000 to leave the circulation and enter the ganglia (Jacobs et al., 1976;

Jacobs, 1977), and IGFs are only M^ 7500-7700. Circulating IGFs, like insulin, would be expected to distribute into the extracellular space, and are likely to act on motor nerve terminals. The subcutaneous local injection of IGFs over muscle has been shown to cause sprouting of motor nerve terminals in rats (Caroni and Grandes, 1990). These data support the hypothesis that IGFs are circulating neurotrophic factors, and differ thereby from the classic neurotrophic factors. The IGFs appear to be circulating general support factors for neurons (virtually every type of neuron examined to date is responsive to IGFs), whereas the classic neurotrophic factors may augment this support by acting locally on specific types of neurons. Support would be augmented further by the paracrine/autocrine activity of IGFs. Thus, peripheral neurons would be served by multiple sources of neurotrophic support from the circulation, Schwann cells, and end-organs. A prediction of this hypothesis is that a decline in circulating IGFs may predispose to symmetrical polyneuropathy in disease states. This underlies the specific interrelated theory that decline in circulating IGFs is pathogenic for diabetic neuropathy (Ishii, 1995), the most prevalent form of which involves a symmetrical polyneuropathy afflicting the sensory, sympathetic, and motor systems. The age-dependent decline in circulating IGFs (Hall and Sara, 1984; Tan and Baxter, 1986) is proposed to be responsible for the age-dependent emergence of diabetic neuropathy. Replacement by subcutaneous administration of IGFs can prevent diabetic neuropathy in rats (Zhuang et al, 1996; Ishii and Lupien, 1995). 12. Dual growth and neurotropliie factor activities The IGFs are found to have dual action as neurotrophic factors on nerves and as mitogens and growth factors on nerve target tissues. They differ in this respect from the classic neurotrophic factor, NGF, which is not a mitogen (Mobley et al., 1977). One reason some growth factors might serve simultaneously as neurotrophic factors may be traced to the observations of Hamburger (1934) and Harrison (1935), who found that the sizes of

414

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

the lateral motor columns and the dorsal root ganglia were influenced by the size of the target field that was innervated. Thus, IGFs may act during development to help achieve coordinated growth between the size of the nervous system and the size of a target field that requires innervation (Recio-Pinto and Ishii, 1988b). The complexity of the IGF genes may reflect, in part, the need to regulate gene expression differentially in nerve and nerve target tissues during the various phases of development and in response to growth hormone, innervation and denervation. Acknowledgements We thank Dr. Laura Wuarin for kindly reviewing the manuscript and Diane M. Guertin for assistance in compiling the references. This work was supported in part by Award PA 1-9401 from the American Paralysis Association and Grant R49/CCR811509 from the Centers for Disease Control. References Aizenman, Y. and De Vellis, J. (1987) Brain neurons develop in a serum and glial free environment: effects of transferrin, insulin, insulin-like growth factor-I and thyroid hormone on neuronal survival, growth and differentiation. Brain Res. 406: 32-42. Asbury, A.K. and Johnson, P. (1978) Pathology of Peripheral Nerve. W.B. Saunders, Philadelphia. Bach, M.A., Shen-Orr, Z., Lowe, W.L., Jr., Roberts, C.T., Jr. and LeRoith, D. (1991) Insulin-like growth factor I mRNA levels are developmentally regulated in specific regions of the rat brain. Mol. Brain Res. 10: 43-48. Barker, D., Scott, J.J.A. and Stacey, M.J. (1986) Reinnervation and recovery of cat muscle receptors after long-term denervation. Exp. Neurol. 94: 184-202. Beck, F., Samani, N.J., Penschow, J.D., Thorley, B., Tregear, G.W. and Coghlan, J.P. (1987) Histochemical localization of IGF-I and -II mRNA in the developing rat embryo. Development \{)l'. 175-184. Beilharz, E.J., Bassett, N.S., Sirimanne, E.S., Williams, C.E. and Gluckman, P.D. (1995) Insulin-like growth factor II is induced during wound repair following hypoxic-ischemic injury in the developing rat brain. Mol. Brain Res. 29: 8191. Bell, G.I., Gerhard, D.S., Pong, N.M., Sanchez-Pescador, R. and Rail, L.B. (1985) Isolation of the human insulin-like growth factor genes: insulin-like growth factor II and insu-

lin genes are contiguous. Proc. Natl Acad. Sci. USA 82: 6450-6454. Bennett, M.R., Lai, K. and Nurcombe, V. (1980) Identification of embryonic motoneurons in vitro: their survival is dependent on skeletal muscles. Brain Res. 190: 537542. Benoit, P. and Changeux, J.P. (1978) Consequences of blocking the nerve with a local anesthetic on the evolution of multiinnervation at the regenerating neuromuscular junction of the rat. Brain Res. 149: 89-96. Betz, W.J., Ribchester, R.R. and Ridge, R.M.A.P. (1990) Competitive mechanisms underlying synapse elimination in the lumbrical muscle of the rat. J. Neurobiol. 21: 1-17. Bisby, M.A. and Chen, S. (1990) Delayed Wallerian degeneration in sciatic nerves of C57BL/01a mice is associated with impaired regeneration of sensory axons. Brain Res. 530:117-120. Blundell, T.L. and Humbel, R.E. (1980) Hormone families: pancreatic hormones and homologous growth factors. Nature 2^1:1 ^\-l%l. Bohannon, N.J., Corp, E.S., Wilcox, B.J., Figlewicz, D.P., Dorsa, D.M. and Baskin, D.G. (1988) Localization of binding sites for insulin-like growth factor-I (IGF-I) in the rat brain by quantitative autoradiography. Brain Res. 444: 205-213. Bond, J.F. and Farmer, S.R. (1983) Regulation of tubulin and actin mRNA production in rat brain: expression of a new Ptubulin mRNA with development. Mol. Cell. Biol. 3: 13331342. Bondy, C.A. (1991) Transient IGF-I gene expression during the maturation of functionally related central projection neurons. J. Neurosci. 11: 3442-3455. Bondy, C.A., Werner, H., Roberts, C.T., Jr. and LeRoith, D. (1990) Cellular pattern of insulin-like growth factor-I (IGFI) and type I IGF receptor gene expression in early organogenesis: comparison with IGF-II gene expression. Mol. Endocrinol. 4:1386-1398. Bothwell, M. (1982) Insulin and somatomedin MSA promote nerve growth factor-independent neurite formation by cultured chick dorsal root ganglionic sensory neurons. /. Neurosci. Res. 8:225-231. Boyd, F.T., Jr., Clarke, D.W., Muther, T.F. and Raizada, M.K. (1985) Insulin receptors and insulin modulation of norepinephrine uptake in neuronal cultures from rat brain. J. Biol. Chem. 260: 15880-15884. Brissenden, J.E., Ullrich, A. and Francke, U. (1984) Human chromosomal mapping of genes for insulin-like growth factors I and II and epidermal growth factor. Nature 310: 781-784. Brown, M.C. and Ironton, R. (1970) Sprouting and regression of neuromuscular synapses in partially denervated mammalian muscles. /. Physiol. (London) 278: 325-348. Brown, M.C, Jansen, J.K.S. and Van Essen, D. (1976) Polyneuronal innervation of skeletal muscle in new-born rats and its elimination during maturation. J. Physiol. (London) 261:387-422.

D.N. Ishii et al. Brown, M.C., Holland, R.L. and Hopkins, W.G. (1981) Restoration of focal multiple innervation in rat muscles by transmission block during a critical stage of development. J. Physiol. (London) 318: 355-364. Brushart, T.M.E. (1990) Preferential motor reinnervation: a sequential double-labeling study. Restor. Neurol. Neurosci. 1:281-287. Buckstrom, M., Hall, K. and Sara, V. (1984) Somatomedin levels in cerebrospinal fluid from adults with pituitary disorders. Acta Endocrinol. (Copenhagen) 107: 171-178. Carbonetto, S. and Muller, K.J. (1982) Nerve fiber growth and the cellular response to axotomy. Curr. Top. Dev. Biol. 17: 33-76. Garden, M.J., Trojanowski, J.Q., Schlaepfer, W.W. and Lee, V.M.-Y. (1987) Two-stage expression of neurofilament polypeptides during rat neurogenesis with early establishment of adult phosphorylation patterns. J. Neurosci. 7: 3489-3504. Carlsson-Skwirut, C , Lake, M., Hartmanis, M., Hall, K. and Sara, V.R. (1989) A comparison of the biological activity of the recombinant intact and truncated insulin-like growth factor I. Biochim. Biophys. Acta 1011: 192-197. Caroni, P. (1992) Signalling by muscle-derived IGFs in neuromuscular junction development and regeneration. N.Y. Acad. Sci. Abstr. 17: 27. Garoni, P. and Grandes, P. (1990) Nerve sprouting in innervated adult skeletal muscle induced by exposure to elevated levels of insulin-like growth factors. J. Cell Biol. 110: 1307-1317. Garoni, P. and Schneider, G. (1992) Activity-sensitive signalling by insulin-like growth factors in the developing and regenerating neuromuscular system. Soc. Neurosci. Abstr. 18: 1095. Garoni, P., Schneider, G., Kiefer, M.G. and Zapf, J. (1994) Role of muscle insulin-like growth factors in nerve sprouting: suppression of terminal sprouting in paralyzed muscle by IGF binding protein 4. J. Cell. Biol. 125: 893-902. Ghiu, A.Y. and Sanes, J.R. (1984) Development of basal lamina in synaptic and extrasynaptic portions of embryonic rat muscle. Dev. Biol. 103: 4 5 6 ^ 6 7 . Govault, J., Gunningham, J.M. and Sanes J.R. (1987) Neurite outgrowth on cryostat sections of innervated and denervated skeletal muscle. J. Cell Biol. 105: 2479-2488. Danielsen, N., Pettmann, B., Vahlsing, H.L., Manthorpe, M. and Varon, S. (1988) Fibroblast growth factor effects on peripheral nerve regeneration in a silicone chamber model. J. Neurosci. Res. 20: 320-330. Daughaday, W.H., Hall, K., Raben, M.S., Salmon, W.D., Jr., VanDenBrande, J.L. and VanWyk, J.J. (1972) Somatomedin: proposed designation for sulfation factor. Nature 235: 107. DeGhiara, T.M., Efstratiadis, A. and Robertson, E.J. (1990) A growth-deficiency phenotype in heterozygous mice carrying an insulin-like growth factor II gene disrupted by targeting. Nature 345: 78-80. DeLeon, M., Welcher, A.A., Suter, U. and Shooter, E.M.

415 (1991) Identification of transcriptionally regulated genes after sciatic nerve injury. J. Neurosci. Res. 29: 437-448. Diamond, J. and Miledi, R. (1962) A study of foetal and new-born rat muscle fibers. J. Physiol. (London) 162: 3 9 3 408. Diamond, J., Goughlin, M., Maclntyre, L., Holmes, M. and Visheau, B. (1987) Evidence that endogenous )3-nerve growth factor is responsible for the collateral sprouting, but not the regeneration, of nociceptive axons in adult rats. Proc. Natl. Acad. Sci. USA 84: 6596-6600. DiGicco-Bloom, E. and Black, LB. (1988) Insulin growth factors regulate the mitotic cycle in cultured rat sympathetic neuroblasts. Proc. Natl. Acad. Sci. USA 85: 4 0 6 6 ^ 0 7 0 . DiStefano, P.S., Friedman, B., Radziejewski, G., Alexander, G., Boland, P., Schick, G.M., Lindsay, R.M. and Wiegand, S.J. (1992) The neurotrophins BDNF, NT-3 and NGF display distinct patterns of retrograde axonal transport in peripheral and central neurons. Neuron 8: 983-993. Drachman, D.B., Murphy, S.A., Nigam, M.P. and Hills, J.R. (1967) 'Myopathic' changes in chronically denervated muscles. Arc/z. Neurol. 16: 14-24. Dribin, L.B. and Barrett, J.N. (1980) Gonditioned medium enhances neuritic outgrowth from rat spinal cord explants. Dev. Biol. 74: 184-195. Duchen, L.W. and Strich, S.J. (1968) The effects of botulinum toxin on the pattern of innervation of skeletal muscle in the mouse. Q. J. Exp. Physiol. 53: 84-89. Dull, T.J., Gray, A., Hayflick, J.S. and Ullrich, A. (1984) Insulin-like growth factor II precursor gene organization in relation to insulin gene family. Nature 3 1 0 : 1 1 1 - l S l . Duxson, M.J. (1982) The effect of postsynaptic block on development of the neuromuscular junction in postnatal rats. J. Neurocytol. 11:395-408. Elsberg, G.A. (1917) Experiments on motor nerve regeneration and the direct neurotization of paralyzed muscles by their own and by foreign nerves. Science 45: 318-320. Ferguson, LA., Schweitzer, J.B. and Johnson, E.M., Jr. (1990) Basic fibroblast growth factor: receptor-mediated internalization, metabolism and anterograde axonal transport in retinal ganglion cells. J. Neurosci. 10: 2176-2189. Fernyhough, P., Mill, J.F., Roberts, J.L. and Ishii, D.N. (1989) Stabilization of tubulin mRNAs by insulin and insulin-like growth factor I during neurite formation. Mol. Brain Res. 6: 109-120. Fernyhough, P., Willars, G.B., Lindsay, R.M. and Tomlinson, D.R. (1993) Insulin and insulin-like growth factor I enhance regeneration in cultured adult rat sensory neurones Brain Res. 601: 117-124. Fex, S., Sonesson, B., Twesleff, S. and Zelena, J. (1966) Nerve implants in botulinum poisoned mammalian muscle. J. Physiol. (London) 184: 872-882. Forssman, J. (1898) Uber die Uraschen welche die Wachstumrichtung der peripheren Nervernfasern bei Regeneration bestimmen. Beitr. Pathol. Anat. 24: 56. Frank, E., Jansen, J.K.S., Lomo, T. and Westgaard, R.H. (1974) Maintained function of foreign synapses on hyper-

416

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

innervated skeletal muscle fibers of the rat. Nature 247: 375-376. Glazner, G.W. and Ishii, D.N. (1995) Insulin-like growth factor gene expression in rat muscle during reinnervation. Muscle Nerve 18: 1433-1442. Glazner, G.W., Lupien, S., Miller, J.A. and Ishii, D.N. (1993a) Insulin-like growth factor II increases the rate of sciatic nerve regeneration in rats. Neuroscience 54: 791-797. Glazner, G.W., Wright, W.G. and Ishii, D.N. (1993b) Kinetics of insulin-like growth factor (IGF) mRNA changes in rat nerve and muscle during IGF-dependent nerve regeneration. Soc. Neurosci. Abstr. 19: 253. Glazner, G.W., Morrison, A.E. and Ishii, D.N. (1994) Elevated insulin-like growth factor (IGF) gene expression in sciatic nerves during IGF-supported nerve regeneration. Mol Brain Res. 25: 265-272. Goldstein, M.E., Weiss, S.R., Lazzarini, R.A., Shneidman, P.S., Lees, J.F. and Schlaepfer, W.W. (1988) mRNA levels of all three neurofilament proteins decline following nerve transection. Mol Brain Res. 3: 287-292. Goodyer, C.G., De Stephano, L., Lai, W.H., Guyda, H.J. and Posner, B.I. (1984) Characterization of insulin-like growth factor receptors in rat anterior pituitary, hypothalamus and brain. Endocrinology 114: 1187-1195. Gorin, P.D. and Johnson, E.M., Jr. ( 1980) Effects of longterm nerve growth factor deprivation on the nervous system of the adult rat: an experimental autoimmune approach. Brain Res. 198:27-42. Greensmith, L. and Vrbova, G. (1992) Alterations in nervemuscle interaction during postnatal development influence motorneurone survival in rats. Dev. Brain Res. 69: 125131. Guler, H.-P., Schmid, C, Zapf, J. and Froesch, E.R. (1989) Effects of recombinant insulin-like growth factor I on insulin secretion and renal function in normal human subjects. Proc Natl. Acad. Sci. USA 86: 2868-2872. Guthrie, K.M., Nguyen, T. and Gall, CM. (1995) Insulin-like growth factor I mRNA is increased in deafferented hippocampus: spatiotemporal correspondence of a trophic event with axon sprouting /. Comp. Neurol. 352: 147-160. Gutmann, E. and Young, J.Z. (1944) Re-innervation of muscle after various periods of atrophy. J. Anat. 78: 15-43. Gutmann, E. and Zelena, J. (1962) Morphological changes in the denervated muscle. In: E. Gutmann (Ed.), The Denervated Muscle, Publishing House of the Czechoslovak Academy of Sciences, Prague, pp. 57-102. Hall, K. and Sara, V.R. (1984) Somatomedin levels in childhood, adolescence and adult life. Clin. Endocrinol. Metab. 13:91-112. Hall, S.M. and Gregson, N.A. (1975) The effects of mitomycin C on the process of remyelination in the mammalian peripheral nervous system. Neuropath. Appl. Neurobiol. 1: 149-170. Hamburger, V. (1934) The effects of wing bud extirpation on the development of the central nervous system in chick embryos. J. Exp. Zool. 68: 449^94.

Hansson, H.-A., Dahlin, L.B., Danielsen, N., Fryklund, L., Nachemson, A.K., Polleryd, P., Rozell, B., Skottner, A., Stemme, S. and Lundborg, G. (1986) Evidence indicating trophic importance of IGF-I in regenerating peripheral nerves. Acta Physiol. Scand. 126: 609-614. Hardouin, S., Hossenlopp, P., Segovia, B., Seurin, D., PortoIan, G., Lassarre, C. and Binoux, M. (1987) Heterogeneity of insulin-like growth factor binding proteins and relationships between structure and affinity. I. Circulating forms in man. Eur. J. Biochem. 170: 121-132. Harris, W.A., Holt, C.E., Smith, T.A. and Gallenson, N. (1985) Growth cones of developing retinal cells in vivo, on culture surfaces and in collagen matrices. J. Neurosci. Res. 13: 101-122. Harrison, R.G. (1935) On the origin and development of the nervous system studied by the methods of experimental embryology. Proc. R.. Soc. Lond.. 118: 155-196. Haselbacher, G.K., Schwab, M.E., Pasi, A. and Humbel, R.E. (1985) Insulin-like growth factor II (IGF II) in human brain: regional distribution of IGF II and of higher molecular mass forms. Proc. Natl. Acad. Sci. USA 82: 2153-2157. Heidenreich, K.A. and Brandenberg, D. (1986) Oligosaccharide heterogeneity of insulin receptors. Comparison of Nlinked glycosylation of insulin receptors in adipocytes and brain. Endocrinology 118: 1835-1842. Henderson, C.E., Huchet, M . and Changeux, J.-P. (1983) Denervation increases a neurite-promoting activity in extracts of skeletal muscle. Nature 302: 609-611. Heumann, R., Korsching, S., Bandtlow, C. and Thoenen, H. (1987a) Changes of nerve growth factor synthesis in nonneuronal cells in response to sciatic nerve transection. J. Cell Biol. 104: 1623-1631. Heumann, R., Lindholm, D., Bandtlow, C, Meyer, M., Radeke, M.J., Misko, T.P., Shooter, E. and Thoenen, H. (1987b) Differential regulation of mRNA encoding nerve growth factor and its receptor in rat sciatic nerve during development, degeneration and regeneration: role of macrophages. Proc. Natl. Acad. Sci. USA 84: 8735-8739. Hofer, M.M. and Barde, Y.-A. (1988) Brain-derived neurotrophic factor prevents neuronal death in vivo. Nature 331:261-262. Hoffman, H. (1950) Local re-innervation in partially denervated muscle: a histo-physiological study. Aust. J. Exp. Biol. Med Sci. 28: 383-397. Hoffman, H. (1951) A study of the factors influencing innervation of muscles by implanted nerves. Aust. J. Exp. Biol. Med. Sci. 29: 289-308. Hoffman, P.N. and Cleveland, D.W. (1988) Neurofilament and tubulin expression recapitulates the developmental pattern during axonal regeneration: induction of a specific ^-tubulin isotype. Proc. Natl. Acad Sci. USA 85: 4530-4533. Hoffman, P.N.. Griffin, J.W. and Price, D.L. (1984) Control of axonal caliber by neurofilament transport. /. Cell Biol. 99: 705-714. Hoffman, P.N., Cleveland, D.W., Griffin, J.W., Landes, P.W., Cowan, N.J. and Price, D.L. (1987) Neurofilament gene ex-

D.N. Ishii et al. pression: a major determinant of axonal caliber. Proc. Natl. Acad. Sci. USA 84: 3472-3476. Holds, J.B., Alderson, K., Fogg, S.G. and Anderson, R.L. (1990) Motor nerve sprouting in human orbicularis muscle after botulinum A injection. Invest. Ophthalmol. Vis. Sci. 31:964-967. Holland, R.L. and Brown, M.C. (1981) Nerve growth in botulinum toxin poisoned muscles. Neuroscience 6: 1167-1179. Hollowell, J.P., Villadiego, A. and Rich, K.M. (1990) Sciatic nerve regeneration across gaps within silicone chambers: long-term effects of NGF and consideration of axonal branching. Exp. Neurol. 110: 45-51. Hossenlopp, P., Seurin, D., Segovia-Quinson, B. and Binoux, M. (1986) Identification of an insulin-like growth factorbinding protein in human cerebrospinal fluid with a selective affinity for IGF-II. FEES Lett. 208: 4 3 9 ^ 4 4 . Hynes, M.A., Brooks, P.J. Van Wyk, J.J. and Lund, P.K. (1988) Insulin-like growth factor II messenger ribonucleic acids are synthesized in the choroid plexus of the rat brain. Mol. Endocrinol. 2: 47-54. Hynes, M.A., Van Wyk, J.J., Broors, P.J., D'Ercole, A.J., Jansen, M. and Lund, P.K. (1987) Growth hormone dependence of somatomedin-C/insulin-like growth factor-I and insulin-like growth factor-II messenger ribonucleic acids. Mol. Endocrinol. 1: 233-242. Ichimiya, A., Emson, P.C, Northrop, A.J. and Gilmour, R.S. (1988) Insulin-like growth factor II in the rat choroid plexus. Mol. Brain Res. 4: 167-170. Ishii, D.N. (1989) Relationship of insulin-like growth factor II gene expression in muscle to synaptogenesis. Proc. Natl. Acad. Sci. USA 86: 2898-2902. Ishii, D.N. (1995) Implication of insulin-like growth factors in the pathogenesis of diabetic neuropathy Brain Res. Rev. 20: 47-67. Ishii, D.N. (1993) Neurobiology of insulin and insuUn-like growth factors. In: S.E. Loughlin and J.H. Fallon (Eds.), Neurotrophic Factors, Academic Press, New York, pp. 415-442. Ishii, D.N. and Lupien, S.B. (1995) Insulin-like growth factors protect against diabetic neuropathy: effects on sensory nerve regeneration in rats. /. Neurosci. Res. 40: 138-144. Ishii, D.N. and Recio-Pinto, E. (1987) Role of insulin, insulinlike growth factors and nerve growth factor in neurite formation. In: M.K. Raizada, M.I. Phillips and D. LeRoith (Eds.), Insulin, Insulin-like Growth Factors and Their Receptors in the Central Nervous System, Plenum Press, New York, pp. 315-348. Ishii, D.N., Glazner, G.W., Wang, C. and Fernyhough, P. (1989) Neurotrophic effects and mechanism of insulin, insulin-like growth factors and nerve growth factor in spinal cord and peripheral neurons. In: D. LeRoith and M.K. Raizada (Eds.), Molecular and Cellular Biology of Insulin-like Growth Factors and Their Receptors, Plenum Press, New York, pp. 4 0 3 ^ 2 5 . Ishii, D.N., Wang, C. and Li, Y, (1991) Second messengers mediating gene expression essential to neurite formation di-

All rected by insulin and insulin-like growth factors. In: M.K. Raizada and D. LeRoith (Eds.), Molecular Biology and Physiology of Insulin and Insulin-like Growth Factors, Plenum Press, New York, pp. 361-378. Jacobs, J.M. (1977) Penetration of systemically injected horseradish peroxidase into ganglia and nerves of the autonomic nervous system. /. Neurocytol. 6: 607-612. Jacobs, J.M., MacFarlane, R.M. and Cavanagh, J.B. (1976) Vascular leakage in the dorsal root ganglia of the rat studied with horseradish peroxidase. J. Neurol. Sci. 29: 95-107. Jacobs, S., Kull, F.C., Earp, H.S., Svoboda, M.E., Van Wyk, J.J. and Cuatrecasas, P. (1983) Somatomedin-C stimulates the phosphorylation of the beta subunit of its own receptor. J. Biol. Chem. 258: 9581-9584. Jansen, M., Van Schaik, F.M.A., Van Tol, H., Van Den Brande, J.L. and Sussenbach, J.S. (1985) Nucleotide sequences of cDNAs encoding precursors of human insulinlike growth factor II (IGF-II) and an IGF-II variant. FEBS Lett. 179: 243-246. Jennische, E. and Olivecrona, H. (1987) Transient expression of insulin-like growth factor I immunoreactivity in skeletal muscle cells during postnatal development in the rat. Acta Physiol. Scand. 131: 619-622. Julien, J.-P., Meyer, D., Flavell, D., Hurst, J. and Grosveld, F. (1986) Cloning and developmental expression of the murine neurofilament gene family. Mol. Brain Res. 1: 2 4 3 250. Kanje, M., Skottner, A., Sjoberg, J. and Lundborg, G. (1989) Insulin-like growth factor I (IGF-I) stimulates regeneration of the rat sciatic nerve. Brain Res. 486: 396-398. Kashihara, Y., Kuno, M. and Miyata, Y. (1987) Cell death of axotomized motorneurons in neonatal rats and its prevention by peripheral reinnervation. J. Physiol. 386: 135-148. Kasuga, M., Karlsson, F.A. and Kahn, C.R. (1982) Insulin stimulates the phosphorylation of the 95 000-dalton subunit of its own receptor. Science 215: 185-187. Katz, B. and Miledi, R. (1977) Transmitter leakage from motor nerve endings. Proc. R. Soc. Lond. 196: 159-172. Kawamura, Y. and Dyck, P.J. (1981) Permanent axotomy by amputation results in loss of motor neurons in man. J. Neuropath. Exp. Neurol. 40: 658-666. Kelly, A.M. and Zacks, S.I. (1969) The fine structure of motor endplate morphogenesis. /. Cell Biol. 42: 154-169. Kreutzerg, G.W. and Schubert, P. (1971) Volume changes in the axon during regeneration. Acta Neuropathol. 17: 220226. Kubota, K., Sato, Y., Masegi, T., Kobayashi, M. and Shishido, Y. (1978) Electron microscopic study of the snout muscle spindles of the mole following denervation. Anat. Anz. 144: 81-96. Kuffler, D.P. (1986) Isolated satellite cells of a peripheral nerve direct the growth of regenerating frog axons. J. Comp. Neurol. 249: 57-64. Lampert, P.W. (1967) A comparative electron microscopic study of reactive, degenerating, regenerating and dystrophic axons. J. Neuropathol. Exp. Neurol. 26: 345-368.

418

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

Lamson, G., Pham, H., Oh, Y., Ocrant, I. and Schwander, J. (1989) Expression of the BRL-3A insulin-like growth factor binding protein (rBP-30) in the rat central nervous system. Endocrinology 123: 1100-1102. Landmesser, L. (1992) The relationship of intramuscular nerve branching and synaptogenesis to motoneuron survival. J. Neurobiol. 23: 1131-1139. Lasek, R.J., Oblinger, M.M. and Drake, P.F. (1984) Molecular biology of neuronal geometry: expression of neurofilament genes influences axonal diameter. Cold Spring Harb. Symp. Quant. Biol. 4S:13\--744. Lee, J.E., Pintar, J. and Efstratiadis, A. (1990) Pattern of the insuhn-like growth factor II gene expression during early mouse embryogenesis. Development 110: 151-159. Lee, W.-H. and Bondy, C.A. (1993) Insulin-like growth factors and cerebral ischemia. Ann.. N. Y. Acad. Sci. 679: 418422. Lee, W.-H., Clemens, J.A. and Bondy, C.A. (1992) Insulinlike growth factors in the response to cerebral ischemia. Mol. Cell. Neurosci. 3: 36-43. Levi-Montalcini, R. and Angeletti, P.U. (1968) Nerve growth factor. Physiol. Rev. 48: 534-569. Li, L., Oppenheim, R.W., Lei, M. and Houenou, L.J. (1994) Neurotrophic agents prevent motoneuron death following sciatic nerve section in the neonatal mouse. J. Neurobiol. 25: 759-766. Lindsay, R.M. (1988) Nerve growth factors (NGF, BDNF) enhance axonal regeneration but are not required for survival of adult sensory neurons. J. Neurosci. 8: 23942405. Lindsay, R.M., Thoenen, H. and Barde, Y.-A. (1985) Placode and neural crest-derived sensory neurons are responsive at early developmental stages to brain-derived neurotrophic factor (BDNF). Dev. Biol. 112: 319-328. Logan, A., Gonzalez, A.-M., Hill, D.J., Berry, M., Gregson, N.A. and Baird, A. (1994) Coordinated pattern of expression and localization of insulin-like growth factor-II (IGFII) and IGF-binding protein-2 in the adult rat brain. Endocrinology 135: 2255-2264. Longo, F.M., Manthorpe, M. and Varon, S. (1982) Spinal cord neurotrophic factors (SCNTFs): I. Bioassay of Schwannoma and other conditioned media. Dev. Brain Res. 3: 277-294. Longo, F.M., Manthorpe, M., Skaper, S.D., Lundborg, G. and Varon, S. (1983) Neuronotrophic activities accumulate in vivo within silicon regeneration chambers. Brain Res. 261: 109-117. Loughlin, S.E. and Fallon, J.H. (1993) Neurotrophic Factors. Academic Press, New York. Lubinska, L. (1982) Patterns of Wallerian degeneration of myelinated fibers in short and long peripheral stumps and in isolated segments of rat phrenic nerve. Interpretation of the role of axoplasmic flow of the trophic factor. Brain Res. 233: 227-240. Lund, P.K., Moates-Staats, B.M., Hynes, M.A., Simmons, J.G., Jansen, M., D'Ercole, A.J. and Van Wyk, J.J. (1986) Somatomedin-C/insulin-like growth factor-I and insulin-

like growth factor-II mRNAs in rat fetal and adult tissues. J. Biol. Chem. 261: 14539-14544. Lund, P.K., Hepler, J.E., Hoyt, E.C. and Simmons, J.G. (1991) Physiological relevance of IGF-I mRNA heterogeneity. In: E.M. Spencer (Ed.), Modern Concepts of Insulin-like Growth Factors, Elsevier, New York, pp. 111-120. Lundborg, G., Dahlin, L.B., Danielsen, N., Gelberman, R.H., Longo, F.M., Powell, H.C. and Varon, S. (1982) Nerve regeneration in silicone chambers: influence of gap length and of distal stump components. Exp. Neurol. 76: 361-375. MacDonald, R.G., Pfeffer, S.R., Coussens, L., Tepper, M.A., Brocklebank, CM., Mole, J.E., Anderson, J.K., Chen, E., Czech, M.P. and Ullrich, A. (1988) A single receptor binds both insulin-like growth factor II and mannose-6phosiphsiie. Science 239: 1134-1137. Manthorpe, M., Louis, J.-C, Hagg, T. and Varon, S. (1993) Ciliary neuronotrophic factor In: S.E. Loughlin and J.H. Fallon (Eds.), Neurotrophic Factors, Academic Press, New York, pp. 443-473. Marsh, D.J., Pu, S.-F., Bernstein, B.W. and Ishii, D.N. (1994) Localization of insulin-like growth factor II (IGF-II) protein and mRNA during nerve regeneration. Soc. Neurosci. Abstr. 20: 1498. Martin, J.L. and Baxter, R.C. (1986) Insulin-like growth factor-binding protein from human plasma. Purification and characterization. /. Biol. Chem. 261: 8754-8760. Mason, C.A. and Gregory, E. (1984) Postnatal maturation of cerebellar mossy and climbing fibers: transient expression of dual feature on single axons. J. Neurosci. 4: 1715-1735. Massague, J., Pilch, P.F. and Czech, M.P. (1980) Electrophoretic resolution of three major insulin receptor structures with unique subunit stoichiometrics. Proc. Natl. Acad.. Sci. raA 77: 7137-7141. Mathews, L.S., Norstedt, G. and Palmiter, R.D. (1986) Regulation of insulin-like growth factor I gene expression by growth hormone. Proc. Natl. Acad. Sci. USA 83: 93439347. McArdle, J.J. (1975) Complex end-plate potentials at the regenerating neuromuscular junction of the rat. Exp. Neurol. 49.- 629-638. McElduff, A., Poronnik, P. and Baxter, R.C. (1987) The insulin-like growth factor-II (IGF II) receptor from rat brain is of lower apparent molecular weight than the IGF II receptor from rat liver. Endocrinology 121: 1306-1311. McManaman, J.L., Oppenheim, R.W., Prevette, D. and Marchetti, D. (1990) Rescue of motoneurons from cell death by a purified skeletal muscle polypeptide: effects of the ChAT development factor, CDF. Neuron 4: 891-898. Meyer, M., Matsuoka, I., Wetmore, C, Olson, L. and Thoenen, H. (1992) Enhanced synthesis of brain-derived neurotrophic factor in the lesioned peripheral nerve: different mechanisms are responsible for the regulation of BDNF and NGF mRNA. J. Cell Biol. 119: 45-54. Miledi, R. and Slater, C.R. (1969) Electron microscopic structure of denervated skeletal muscle. Proc. R. Soc. London 114: 253-269.

D.N. Ishii et al. Mill, J.F., Chao, M.V. and Ishii, D.N. (1985) Insulin, insulinlike growth factor II and nerve growth factor effects on tubulin mRNA levels and neurite formation. Proc. Natl Acad. Sci. C/5A 82: 7126-7130. Miller, F.D., Tetzlaff, W., Bisby, M.A., Fawcett, J.W. and Milner, R.J. (1989) Rapid induction of the major embryonic a-tubulin mRNA, Tal, during nerve regeneration in adult rats. J. Neurosci. 9: 1452-1463. Miyata, Y., Kashihara, Y., Homma, S. and Kuno, M. (1986) Effects of nerve growth factor on the survival and synaptic function of la sensory neurons in neonatal rats. J. Neurosci. 6: 2012-2018. Mobley, W.C, Server, A.C., Ishii, D.N., Riopelle, R.J. and Shooter, E.M. (1977) Nerve growth factor. N. Engl J. Med.. 291: 1096-1104 (Part I); 1149-1158 (Part II); 1211-1218 (Part III). Morgan, D.O., Edman, J.C, Standring, D.N., Fried, V.A., Smith, M.C., Roth, R.A. and Rutter, W.J. (1987) Insulinlike growth factor II receptor as a multifunctional binding protein. Nature 329.* 301-307. Moses, A.C., Nissley, S.P., Short, P.A., Rechler, M.M., White, R.M., Knight, A.B. and Higa, O.Z. (1980) Increased levels of multiplication-stimulating activity and insulin-like growth factor, in fetal rat serum. Proc. Natl Acad. Scl USA 11: 3649-3653. Murphy, L.J., Bell, G.I. and Friesen, H.G. (1987) Tissue distribution of insulin-like growth factor I and II messenger ribonucleic acid in the adult rat. Endocrinology 120: 12791282. Nakajima, Y., Kodokoro, Y. and Klier, G. (1980) The development of functional neuromuscular junctions in vitro: an ultrastructural and physiological study. Dev. Biol 11: Sill. Near, S.L., Whalen, L.R., Miller, J.A. and Ishii, D.N. (1992) Insulin-like growth factor II stimulates motor nerve regeneration. Proc. Natl Acad. Scl USA S9: 11716-11720. Neff, N.T. Prevette, D., Houenou, L.J. Lewis, M.E., Glicksman, M.A., Yin, Q.W. and Oppenheim, R.W. (1993) Insulin-like growth factors: putative muscle-derived trophic agents that promote motoneuron survival. J. Neurobiol 24: 1578-1588. Neumann, D., Scherson, T., Ginzburg, I., Littauer, U.Z. and Schwartz, M. (1983) Regulation of mRNA levels for microtubule proteins during nerve regeneration. EEBS Lett. 162: 270-276. Noguchi, T., Kurata, L.M. and Sugisaki, T. (1987) Presence of a somatomedin-C-immunoreactive substance in the central nervous system: immunohistochemical mapping studies. Neuroendocrinology 46.-111-1%1. Oaklander, A.L., Miller, M.S. and Spencer, P.S. (1987) Rapid anterograde spread of premitotic activity along degenerating cat sciatic nerve. J. Neurochem. 48.- 111-114. Obata, K. and Tanaka, H. (1980) Conditioned medium promotes neurite growth from both central and peripheral neurons. Neurosci Lett. 16: 27-33. O'Brien, R.A.D., Ostberg, A.J.C. and Vrbova, G. (1978) Ob-

419 servations on the elimination of polyneuronal innervation in developing mammalian skeletal muscle. J. Physiol (London). 282: 571-582. Ocrant, I., Fay, C.T. and Parmelee, J.T. (1990) Characterization of insulin-like growth factor binding proteins produced in the rat central nervous system. Endocrinology 111: 1260-1267. Oppenheim, R.W., Haverkamp, L.J., Prevette, D., McManaman, J.L. and Appel, S.H. (1988) Reduction of naturally occurring motoneuron death in vivo by a target-derived neurotrophic factor. Science 240: 919-922. Oppenheim, R.W., Prevette, D. and Fuller, F.H. (1990) In vivo treatment with basic fibroblast growth factor during development does not alter naturally occurring neuronal death. Soc. Neurosci. Abstr. 16: 1135. Oppenheim, R.W., Prevette, D., Qin-Wei, Y., Collins, F. and MacDonald, J. (1991) Control of embryonic motoneuron survival in vivo by ciliary neurotrophic factor. Science 251: 1616-1618. Otto, D., Unsicker, K. and Grothe, C. (1987) Pharmacological effects of nerve growth factor and fibroblast growth factor applied to the transectioned sciatic nerve on neuron death in adult rat dorsal root ganglia. Neurosci Lett. 83: 156160. Pockett, S. and Slack, J.R. (1982) Source of the stimulus for nerve terminal sprouting in partially denervated muscle. Neuroscience 1: 3173-3176. Politis, M.J., Ederle, K. and Spencer, P.S. (1982) Tropism in nerve regeneration in vivo. Attraction of regenerating axons by diffusible factors derived from cells in distal nerve stumps of transected peripheral nerves. Brain Res. 253: 112. Pu, S.-F. and Ishii, D.N. (1993) Distribution of insulin-like growth factor mRNAs in rat nerve and muscle during nerve regeneration by in situ hybridization. Soc. Neurosci Abstr. 19.- 253. Pu, S.-F., Zhuang, H.-X. and Ishii, D.N. (1995a) Differential spatio-temporal expression of the insulin-like growth factor genes in regenerating sciatic nerve. Mol Brain Res. 34: 1828. Pu, S.-F., Zhuang, H.-X., Marsh, D.J. and Ishii, D.N. (1995b) Insulin-like growth factor is required for spinal motoneuron survival following sciatic nerve axotomy. Soc. Neurosci Abstr. 11: 1789. Quiroga, S., Garofalo, R.S. and Pfenninger, K.H. (1995) Insulin-like growth factor I receptors of fetal brain are enriched in nerve growth cones and contain a y^-subunit variant. Proc. Natl Acad. Scl USA 92: 4309^312. Rabinovsky, E.D., Smith, G.M., Browder, D.P., Shine, H.D. and McManaman, J.L. (1992) Peripheral nerve injury down-regulates CNTF expression in adult rat sciatic nerves. J. Neurosci Res. 31: 188-192. Raivich, G., Hellweg, R. and Kreutzberg, G.W. (1991) NGF receptor-mediated reduction in axonal NGF uptake and retrograde transport following sciatic nerve injury and during regeneration. A^ewron 7: 151-164.

420

Roles of insulin-like growth factors in peripheral nerve regeneration and motor neuron survival

Ramon y Cajal, S. (1928) Degeneration and Regeneration of the Nervous System, Oxford University Press, London, reprinted 1991. Rechler, M.M., Kasuga, M., Sasaki, N., DeVroede, M.A., Romanus, J.A. and Nissley, S.P. (1983) Properties of insulin-like growth factor receptor subtypes. In: E.M. Spencer (Ed.), Insulin-like Growth Factors/Somatomedins: Basic Chemistry, Biology and Clinical Importance, W. de Gruyter, New York, pp. 459-490. Recio-Pinto, E. and Ishii, D.N. (1984) Effects of insulin, insulin-like growth factor-II and nerve growth factor on neurite outgrowth in cultured human neuroblastoma cells. Brain Res. 302: 323-334. Recio-Pinto, E. and Ishii, D.N. (1988a) Insulin and insulinlike growth factor receptors regulating neurite formation in cultured human neuroblastoma cells. J. Neurosci. Res. \9: 312-320. Recio-Pinto, E. and Ishii, D.N. (1988b) Insulin and related growth factors- effects on the nervous system and mechanism for neurite growth and regeneration. Neurochem. Int. 12: 397-414. Recio-Pinto, E., Rechler, M.M., and Ishii, D.N. (1986) Effects of insulin, insulin-like growth factor-II and nerve growth factor on neurite formation and survival in cultured sympathetic and sensory neurons. J. Neurosci. 6: 1211-1219. Redfern, P.A. (1970) Neuromuscular transmission in newborn rats. J. Physiol. (London) 209: 701-709. Rich, K.M., Yip, H.K., Osborne, P.A., Schmidt, R.E. and Johnson, E.M., Jr. (1984) Role of nerve growth factor in the adult dorsal root ganglia neuron and its response to injury. /. Comp. Neurobiol. 230: 110-118. Rich, K.M., Luszczynski, J.R., Osborne, P.A. and Johnson, E.M., Jr. (1987) Nerve growth factor protects adult sensory neurons from cell death and atrophy caused by nerve injury. J. Neurocytol. 16: 261-268. Rinderknecht, E. and Humbel, R.E. (1978a) The amino acid sequence of human IGF-I and its structural homology with proinsulin. J. Biol. Chem. 253: 2769-2776. Rinderknecht, E. and Humbel, R.E. (1978b) Primary structure of human insulin-like growth factor-II. FEBS Lett. 89: 283286. Roghani, M., Hossenlopp, P., Lepage, P., Balland, A. and Binoux, M. (1989) Isolation from human cerebrospinal fluid of a new insulin-like growth factor-binding protein with a selective affinity for IGF-II. FEBS Lett. 255: 253258. Romanus, J.A., Yang, Y.W., Adams, S.O., Sofair, A.N., Tseng, L.Y., Nissley, S.P. and Rechler, M.M. (1988) Synthesis of insulin-like growth factor II (IGF-II) in fetal rat tissues: translation of IGF-II ribonucleic acid and processing of pre-pro-IGF-II. Endocrinology 122: 709-716. Rosenfeld, R.G., Ceda, G., Wilson, D.M., Dollar, L.A. and Hoffman, A.R. (1984) Characterization of high affinity receptors for insulin-like growth factors I and II on rat anterior pituitary cells. Endocrinology 114: 1571-1575. Rotwein, P., Burgess, S.K., Milbrandt, J.D. and Krause, J.E.

(1988) Differential expression of insulin-like growth factor genes in rat central nervous system. Proc. Natl. Acad. Sci. USA 85: 265-269. Salpeter, M.M. (1987) Development and neural control of the neuromuscular junction and of the junctional acetylcholine receptor. Neurol. Neurobiol. 23: 55-115. Salzer, J.L. and Bunge, R.P. (1980) Studies on Schwann cell proliferation. I. An analysis in tissue culture of proliferation during development, Wallerian degeneration and direct injury. J. Cell Biol. 84: 739-752. Sanes, J.R. and Covault, J. (1985) Axon guidance during reinnervation of skeletal muscle. Trends Neurosci. 8: 523528. Sanes, J.R., Marshall, L.M. and McMahan, U.J. (1978) Reinnervation of muscle fiber basal lamina after removal of myofibers. Differentiation of regenerating axons at original synaptic sites. J. Cell Biol. 78: 176-198. Sara, V.R. and Hall, K. (1990) Insulin-like growth factors and their binding proteins. Physiol. Rev. 70: 591-614. Sara, V.R., Hall, K., Von Holtz, H., Humbel, R., Sjogren, B. and Wetterberg, L. (1982) Evidence for the presence of specific receptors for insulin-like growth factors I (IGF-I) and 2 (IGF-II) and insulin throughout the adult human brain. Neurosci. Lett. 34: 39^4. Sara, V.R., Hall, K., Misaki, M., Fryklund, L., Christensen, N. and Wetterberg, L. (1983) Ontogenesis of somatomedin and insuhn receptors in the human fetus. J. Clin. Invest. 71: 1084-1094. Schmalbruch, H. (1984) Motoneuron death after sciatic nerve section in newborn rats. J. Comp. Neurol. IIA: 252-258. Sendtner, M., Kreutzberg, G.W. and Thoenen, H. (1990) Ciliary neurotrophic factor prevents the degeneration of motor neurons after axotomy. Nature 345: 440-441. Shaw, G. and Weber, K. (1982) Differential expression of neurofilament triplet proteins in brain development. Nature 298: 277-279. Shimatsu, A. and Rotwein, P. (1987) Mosaic evolution of the insulin-like growth factors: organization, sequence and expression of the rat insulin-like growth factor I gene. J. Biol Chem. 262: 7894-7900. Sjoberg, J. and Kanje, M. (1989) Insulin-like growth factor (IGF-I) as a stimulator of regeneration in the freeze-injured rat sciatic nerve. Brain Res. 485: 102-108. Smith, M., Clemens, J., Kerchner, G.A. and Mendelsohn, L.G. (1988) The insulin-like growth factor-II (IGF-II) receptor of rat brain: regional distribution visualized by autoradiography. Brain Res. 445: 241-246. Snider, W.D., Elliott, J.L. and Yan, Q. (1992) Axotomyinduced neuronal death during development. J. Neurobiol. 23: 1231-1246. Soares, M.B., Ishii, D.N. and Efstratiadis, A. (1985) Developmental and tissue-specific expression of a family of transcripts related to rat insulin-like growth factor II mRNA. Nucleic Acids Res. 13: 1119-1134. Soares, M.B., Turken, A., Ishii, D.N., Mills, L., Episkopou, v.. Cotter, S., Zeitlin, S. and Efstratiadis, A. (1986) Rat in-

D.N. Ishii et al. sulin-like growth factor II gene. A single gene with two promoters expressing a multitranscript family. J. Mol. Biol. 192: 737-752. Stockli, K.A., Lottspeich, F., Sendtner, M., Masiakowski, P., Carroll, P., Gotz, R., Lindholm, D. and Thoenen, H. (1989) Molecular cloning, expression and regional distribution of rat ciliary neurotrophic factor. Nature 342: 920-923. Stylianopoulou, F., Herbert, J., Soares, M.B. and Efstratiadis, A. (1988) Expression of the insulin-like growth factor II gene in the choroid plexus and the leptomeninges of the adult rat central nervous system. Proc. Natl. Acad. Sci. USA 85: 141-145. Tan, K. and Baxter, R.C. (1986) Serum insulin-like growth factor I levels in adult diabetic patients: the effect of age. J. Clin. Endocrinol. Metab. 63: 651-655. Taxt, T. (1983) Local and systemic effects of tetrodotoxin on the formation and elimination of synapses in reinnervated adult rat muscle. J. Physiol. (London) 340: 175-194. Thompson, W., Kuffler, D.P. and Jansen, J.K.S. (1979) The effect of prolonged, reversible block of nerve impulses on the elimination of polyneuronal innervation of new-born rat skeletal muscle fibers. Neuroscience 4: 271-281. Tower, S.S. (1932) Atrophy and degeneration in the muscle spindle. Brain 55: 77-90. Tseng, L.Y.-H., Brown, A.L., Yang, Y.W.-H., Romanus, J.A., Orlowski, C.C, Taylor, T. and Rechler, M.M. (1989) The fetal rat binding protein for insulin-like growth factors is expressed in the choroid plexus and cerebrospinal fluid of adult rats. Mol. Endocrinol. 3: 1559-1568. Turner, J.D., Rotwein, P., Novakofski, J. and Bechtel, P.J. (1988) Induction of mRNA for IGF-I and -II during growth hormone-stimulated muscle hypertrophy. Am. J. Physiol. 255:E513-E517. Unsicker, K., Grothe, C, Liidecke, G., Otto, D. and Westermann, R. (1993) Fibroblast growth factors: their roles in the central and peripheral nervous system. In: S.E. Loughlin and J.H. Fallon (Eds.), Neurotrophic Factors, Academic Press, New York, pp. 313-338. Unsicker, K., Grothe, C, Ludecke, G., Otto, D. and Westermann, R. (1993) Fibroblast growth factors: their roles in the central and peripheral nervous system. In: S.E. Loughlin

421 and J.H. Fallon (Eds.), Neurotrophic Factors, Academic Press, New York. Valentino, K.L., Ocrant, I. and Rosenfeld, R.G. (1990) Developmental expression of insulin-like growth factor-II receptor immunoreactivity in the rat central nervous system. Endocrinology 126: 914-920. Van Essen, D.C., Gordon, H., Soha, J.M. and Eraser, S.E. (1990) Synaptic dynamics at the neuromuscular junction: mechanisms and models. J. Neurobiol. 21: 223-249. Wang, C, Li, Y., Wible, B., Angelides, K.J. and Ishii, D.N. (1992) Effects of insulin and insulin-like growth factors on neurofilament mRNA and tubulin mRNA content in human neuroblastoma SH-SY5Y cells. Mol. Brain Res. 13: 289300. Wayne, D.B. and Heaton, M.B. (1990) The response of cultured trigeminal and spinal cord motorneurons to nerve growth factor. Dev. Biol. 138: 473^83. Wewetzer, K., MacDonald, J.R., Collins, F. and Unsicker, K. (1990) CNTF rescues motoneurons from ontogenetic cell death in-vivo, but not in-vitro. NeuroReport 1: 203-206. Windebank, A.J. (1993) Neuronal growth factors in the peripheral nervous system. In: P.J. Dyck, P.K. Thomas, J.W. Griffin, P.A. Low, and J.F. Poduslo (Eds.), Peripheral Neuropathy, 3rd edn., W.B. Saunders, Philadelphia, pp. 377388. Wong, J. and Oblinger, M.M. (1990) Differential regulation of peripherin and neurofilament gene expression in regenerating rat DRG neurons. J. Neurosci. Res. 27: 332-341. Wuarin, L., Guertin, D.M. and Ishii, D.N. (1994) Reduction in insulin-like growth factor (IGF) gene expression in nerves precedes the onset of diabetic neuropathy. Exp. Neurol. 130:106-114. Zhuang, H.-X., Pu, S.-F. and Ishii, D.N. (1994) Insulin-like growth factors (IGFs) prevent diabetic neuropathy (impaired nerve regeneration and hyperalgesia) in rats despite hyperglycemia. Neurosci. Abstr. 20: 415. Zhuang, H.-X., Snyder, C.K., Pu, S.-F. and Ishii, D.N. (1996) Insulin-like growth factors reverse or arrest diabetic neuropathy: effects on hyperalgesia and impaired nerve regeneration in rats. Exp. Neurol. 140: 198-205.

This Page Intentionally Left Blank

Section IV Factors Implicated in Neuronal Damage

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 17

Oxidative stress: free radical production in neural degeneration Mario E. Gotz^ Gabriella Kunig^ Peter Riederer^ and Moussa B.H. Youdim^ ^Clinical Neurochemistry, Department of Psychiatry, University of Wurzburg, D-W-97080, WUrzburg, Germany "Department of Pharmacology, Faculty of Medicine, Technion, Haifa, Israel

1. Definition of neurodegeneration The normal functioning of the central nervous system (CNS) presupposes a well-balanced interaction between different biochemically and structurally linked neuronal systems. When one member of a neuronal circuit is altered in its structural or biochemical entity, an imbalance in the functional system results and a compensatory mechanism must be activated in order to maintain physiological equilibrium. Neurodegenerative processes can involve diverse areas of the CNS. The initial etiology of such disturbances is mostly unknown. Nevertheless, we can distinguish between an idiopathic and a symptomatic form of neurodegenerative disease. A close examination of the latter enables one to make valuable conclusions about the nature of the former. Thus, we can amplify our knowledge by examining function, structure and biochemistry of the CNS in symptomatic cases and by looking for comparable alterations in idiopathic forms. Another way is to mimic the neurodegenerative diseases by means of animal models. Acute symptomatic neurodegeneration can develop during a very short period, a fact from which we deduce the delaying effect of compensatory and regenerative mechanism in slowly developing chronic neurodegeneration. Neurodegeneration appears clinically as a breakdown of functionally connected neuronal circuits, with corresponding alterations in the neurotransmitter system and morphological organization of the affected cell system. In order to understand the pathophysiological consequences of neurodegeneration, it is

necessary to know the function of neuronal loops in their normal state (Fig. 1). Parkinson's disease (PD), Alzheimer's disease (AD), multi-infarct dementia and motoneuron disease (e.g. amyotrophic lateral sclerosis, ALS) represent typical neurodegenerative diseases for which no known etiology has been put forward. Parkinson's disease is characterized by reduced size and velocity of movements. In AD, cognitive impairment is the cardinal clinical symptom. In motoneuron disease, a degeneration of the central pyramidal, the peripheral motor system or both is the reason for the clinical picture. A significant overlap exists between these three disorders. 2. Characterization of neurodegenerative disorders 2.7. Movement disorders 2.1.1. Physiology of the motor system Several events must occur when a motor action is performed. The acting subject develops the intention to reach a goal by motor action. A motor plan is made. The orientation of the body is adapted to the spatial position of the goal and finally, the motor system must coordinate the activities of the different descending pathways. Before efferent motor actions can be started, sensoric afferent information about body and goal position in space must be established so that the motor system is able to estimate the consequences of actions. These afferent messages and efferent commands are linked to a functional neuronal network by the following neocortical areas: the motor cor-

426

Oxidative stress: free radical production in neural degeneration

isocortex

entorhinal region

hippocampui| formation

cortex

I

ventral intra! atriatum

retroaplenial] anterior c i n guiate and other areaa

frontal asaociation and other areaa

promoter and other areaa

motor and other areaa

doraal atriatum

striatum

I!

ventral pallidum ntri

ext. pall.

int. pall.

pallidum

no. ventralia lateralia poaterior

thalamus

|pont. gray

other basal nuclei

cerebell. cortex

cerebell. nuclei

Z3~

input complex loop'

motor loop

• output

A

Fig. 1. Integrative aspects of cognitive and motor loops. Disturbances in cognitive loops predominate in AD, and motor loops are predominantly affected in PD. Abbreviations: nc, nucleus; ext pall, external pallidum; int pall, internal pallidum; nc paranigr, nucleus paranigralis; nc subthal, nucleus subthalamicus; pont gray, pontine gray. Modified from Alexander et al., 1986; Gerlach et al., 1991b; Nieuwenhuys et al., 1991; Braak and Braak, 1993.

tex, the premotor cortex of the frontal lobe, the supplementary motor area and the posterior parietal cortex. The motor cortex covers the precentral gyrus and is organized somatotopically. The premotor cortex, located on the lateral surface of the hemisphere, plays a specific role in the preparation of responsive motor reactions and controls motor responses to tactile stimuli (Fig. 1). The supplementary motor area plays an indirect preparatory role in movement generation. Both prefrontal cortex and supplementary motor areas receive inputs from posterior parietal areas, but differ in their subcortical projections. The prefrontal cortex receives afferents from the cerebellum, the supplementary motor area from thalamic nuclei.

The posterior parietal cortex processes sensory stimuli before initiating a purposeful movement and projects this information to the premotor area and supplementary motor area. The premotor and supplementary areas represent the highest levels in the hierarchy of motor systems because of their integrating function in programming movements. These areas project to the motor cortex, from where the descending upper motoneurons project to the brain stem nuclei via the corticobulbar system and to the spinal cord through the corticospinal system. In the brain stem, the corticobulbar neurons form synaptic contacts with cranial nerve cells. Besides this, the brain stem is an integrative organ for descending motor commands from higher levels and ascending sensory information

M.E. Gotz et al.

from proprioceptive systems and vestibular nuclei. Thus, it provides the postural adjustment of the body. In the spinal cord, the tractus corticospinalis ends on a-motoneurons directly by synaptic contact or indirectly through intemeurons that enhance or inhibit the stimulation of motoneurons (Ghez, 1991). 2.7.2. Clinical manifestation of Parkinson's disease Idiopathic PD is a movement disorder whose symptomatology is defined by three cardinal symptoms: tremor at rest, rigidity and akinesia (Fahn, 1989). Since the first description of the syndrome by James Parkinson in 1817, knowledge about this disease has progressed in several directions, but an exact insight into its etiological mechanism has not yet been possible. The mean age of onset of PD is about 60 years. In the age group over 60, the incidence is increased, but an early onset form also exists. The clinical manifestation of PD first appears in most cases as a unilateral subjective feeling of stiffness and clumsiness or of internal tremulousness, later signs of which are the classical symptoms of tremor, rigor and akinesia. In most cases, clinical diagnosis of PD is correct. Nevertheless, misdiagnoses do occur, especially if additional clinical signs (e.g. gaze paresis and pyramidal signs, which characterize multisystem atrophy (MSA)) are overlooked (Fahn, 1989; Hughles et al., 1992). From the different expression of the symptoms, the subdivision of the disease into three subtypes can be made. In the first, one can see an equal appearance of rigor, tremor and akinesia. Therefore, it is called equivalence type. In the second form, the clinical picture is dominated by rigor and akinesia. This is the rigor/akinesia type. In the tremordominance type, the tremor at rest is the predominant clinical symptom. This subdivision can give us important indications about the prognosis of PD. Firstly, there are indications that, in the tremor-dominant form, there is a slower progress of disease (benign form), whereas a positive correlation between rigidity and malign progress is observed. Secondly, a connec-

All

tion between rigidity/akinesia and parkinsonian bradyphrenia and dementia has been described. The risk of developing a cognitive impairment is much lower in the tremor form. A positive correlation between motor symptoms and intellectual impairment in PD patients leads to the conclusion that damage in the same structures, substantia nigra zona compacta (SNC), may be responsible for both parkinsonian symptoms and the associated dementia (Mortimer et al., 1982). This assumption is supported neuropathologically by semiquantitative estimation of neuronal loss in locus coeruleus (LC) and substantia nigra (SN) (Fig. 1) and cognitive impairment (Caspar and Gray, 1984). Nevertheless, the course of the disease is a progressive one. The beneficial effect of L-3,4-dihydroxyphenylalanine (L-DOPA) therapy has increased patients' life expectancy to a significant degree. The advanced stage of the disease is dominated by the complications of L-DOPA therapy and lack of L-DOPA responsiveness. Indeed, little can be said about the natural history of the disease. 2.1.3. Pathophysiology of Parkinson's disease Lesions in different anatomical sites of the structures that make up the motor loop can be used to create models of movement disorders (Fig. 1). When the subthalamic nucleus is lesioned, a disruption of the indirect pathway results: the decline of glutamatergic stimulation of internal globus pallidus/substantia nigra zona reticulata (SNR) neurons via subthalamico-pallidonigral pathways reveals diminished inhibition of the excitatory thalamocortical pathways. The clinical consequence of this lesion is a proximal ballistic movement of the limbs. Degeneration of striatal neurons, as seen in Huntington's disease (HD), increases the inhibitory activity of the indirect pathway and causes an increase in activity of the direct pathway. This leads to an enhanced activation of the thalamocortical input, evidenced by clinical correlates distal to irregular choreiformic movement of limbs. In hypokinetic disorders, the discharge of efferent striatal neurons is diminished in the direct and increased in the indirect way, so that an inhibition of thalamocortical connections originates from a

428

dual alteration (Alexander and Crutcher, 1990). The puq)ose of internal globus pallidus/SNR, as part of the motor loop, is to make a copy of the corticospinal instruction, sending this copy back to the cortex via thalamocortical connections (Fig. 1). This copying system is markedly important in the automatic execution of learned motor plans. A motor plan consists of a consecutive series of single motor programs that are integrated by neuronal interaction between sensoric perception and motoric action. Whereas single programs are unimpaired in PD, motor planning, an action performed by the basal ganglia, is disrupted. Although the patient is able to perform several single motor acts slowly, he cannot perform the same actions successively as part of a smoothly executed motor plan. In single motor actions, we distinguish fast ballistic and slow movements. The ballistic movements depend mostly on an intact fast interaction of perceptive judgement and motor command. This kind of movement cannot be performed by parkinsonian patients because of the lack of sensorimotor interaction as irreplaceable presupposition in motor planning. Slowly performed movements in PD indicate that motor programming and sensory perception, per se, are still intact (Marsden, 1982). 2.7.4. The pathology of Parkinson's disease The morphologic changes in PD are characterized by neuronal cell death in the pigmented cell nuclei of the brain stem, which frequently appear symmetrically (Jellinger, 1986). One can also find changes in the nucleus dorsalis nervi vagi and nucleus basalis of Meynert (NBM). The most important degenerative process is in the SNC. These neurons project to the striatum, which plays a role in programming of motor actions (Marsden, 1982). As an important functional member of the motor loop (Riederer et al., 1989b), the dopaminergic nigrostriatal connections are primarily involved in the frictionless performance of planned movements, and their disruption results in a heavy imbalance of the motor system. The basis of the neurodegeneration in the SNC is an active pathological process (McGeer et al., 1988). There is a natural age-related cell loss in the SNC (Feamley

Oxidative stress: free radical production in neural

degeneration

and Lees, 1991). Comparing the number of pigmented cells in young (20 years) and aged (90 years) non-parkinsonian subjects, a decline of 47% can be observed. This corresponds to a cell loss of 5% per decade. In parkinsonian patients there is an age-dependent correlation between duration of disease and decline of cell number. The cell loss occurs in an exponential manner, being 10 times greater in the first decade (45%) than in agematched control persons. Whereas the first clinical symptoms appear at a time when dopamine (DA) content is reduced by 80%, at the same time point, the cell number in the SNC is reduced by only 50%. This indicates a partial biochemical dysfunction of the remaining neurons. Comparing pathologically two different subtypes of PD, the rigor/akinesia-dominance type and the tremor-dominance type, more pronounced neuronal cell loss in the medial and lateral part of the SNC, more severe gliosis, dystrophic axons and nigral extraneuronal melanin deposits have been found in the rigor/akinesia-dominance type (Paulus and Jellinger, 1991). The loss of neurons in 80-100% of PD is associated with morphological changes of cytoplasmic structure: these consist of the appearance of Lewy bodies (LB) (Fomo, 1986; Jellinger, 1986), hyalinic cytoplasmic inclusion bodies in several brain regions, such as SNC, LC, NBM, spinal cord, sympathetic ganglia and cerebral cortex (Jellinger, 1986, 1989). The ultrastructure of LB is described as follows: inclusion bodies with an electrondense, amorphous or circulate dense core from which filaments radiate to the neuromelanin particles. LB can appear singly or multiply in the pigmented cell cytoplasm (Forno, 1986). LB are not pathognomonic for PD. 2.7.5. Biochemical alterations in Parkinson's disease Biochemical changes in PD are strongly correlated with the morphological degeneration of neuronal systems. The pathomorphological hallmark of this disease is the decline of dopaminergic nigrostriatal neurons, which histologically appears as loss of melanin-containing neurons in the SNC. Our knowledge of the physiological function of

M.E. Gotzetal.

these neurons in the motor loop explains the clinical signs of reduced size and velocity of volitional movements. Indeed, the chemoarchitectural profile of the motor loop would enable several diverse neurochemical alterations to provoke a picture of hypokinesia (Gerlach et al., 1991a; Braak and Braak, 1993): (1) reduction of dopaminergic nigrostriatal transmission enhances the excitotoxic effect of the corticostriatal afferents; (2) a shift in glutamatergic activity can explain the phenomenon of hypokinesia by inhibiting the direct pathway or exciting the indirect pathway via increased excitatory action of the subthalamic nucleus; and (3) an increased action of striatal cholinergic interneurons could also result in an imbalance of motor regulation and, therefore, in parkinsonlike hypokinesia. The most prominent biochemical finding in PD is the DA decline in the striatum, the main DA containing region (80%) in human brain (Ehringer and Hornykiewicz, 1960; Homykiewicz and Kish, 1986; Agid et al., 1987; Montastruc, 1991; Calne, 1992). Nigrostriatal degeneration in patients with postencephalitic parkinsonism and idiopathic PD differ with respect to the progression of nigrostriatal cell loss (Agid, 1991). Postencephalitic parkinsonism is characterized by an initial rapid cell decline at the onset of the disease, which is followed by a slow, age-related degeneration of nigrostriatal neurons (0.5% cell loss per year). By contrast, in PD, there is a continuous fast progression of nigrostriatal cell death, exceeding that observed in normal aging (Scherman et al., 1989; Agid, 1991; Kish et al.. 1992). However, aging can play an additional role in degeneration of dopaminergic nigrostriatal neurons because of the early continuing natural cell loss of 0.5% per year in healthy persons. Tyrosine hydroxylase (TH) is the key enzyme in DA synthesis, by means of which transmitter production is regulated in relation to the physiological need. The discharge frequency of dopaminergic neurons is correlated positively to the DA synthesis from L-DOPA. A high DA concentration in the synaptic cleft suppresses the firing rate of

429

the neuron, the activity of TH and DA release via presynaptic autoreceptors (Riederer et al., 1989b; Melamed, 1992). By means of TH-staining, using immunohistochemical methods, the dopaminergic neurons and their synaptic boutons can be labelled in a quantitative manner (Pasik et al., 1986), and a decrease of TH-positive cells in the ventral tegmental area and perforant path has also been observed in PD (Torack and Morris, 1990, 1992). The loss of dopaminergic cells in the ventral tegmental area, together with the decrease of DA content in neocortical and limbic structures, demonstrates that mesolimbicocortical tracts degenerate in PD (Dubois et al, 1992), although the DA decline in the mesolimbic system does not reach the degree of cell loss in the nigrostriatal system (Agid, 1991). It is not quite clear how much these alterations contribute to the cognitive slowing and affective alterations in PD patients because, in contrast to the undoubted improvement of motor actions after L-DOPA treatment, the results of this therapy on cognitive and affective functions are very contradictory (Riklan et al., 1976; Dehs et al., 1982; Mortimer et al., 1982). In the nigrostriatal system of PD patients, the activity of remaining TH increases following cell loss (Nagatsu, 1990). The elevated homovanillic acid (3-methoxy-4-hydroxyphenylacetic acid)/DA ratio in the striatum, which begins at a DA reduction of 50-60%, is proof of increased metabolism in the remaining neurons. Another way to compensate for the cell decline in PD is the D2 receptor up-regulation seen in untreated patients, whereas the L-DOPA substitution, depending on the dose, provokes either normalization or down-regulation (Rinne et al., 1981; Guttman and Seeman, 1986; Riederer et al., 1989b; Montastruc, 1991; Playford and Brooks, 1992). The denervation-receptor hypersensitivity develops after a nigrostriatal cell reduction of 90% and more (Riederer et al., 1989b; Agid, 1991; Montastruc, 1991). A positive correlation between decline of D2 receptor density and duration of disease, age and duration of L-DOPA therapy cannot be found (Guttman and Seeman, 1986).

430

Besides the cardinal symptoms in PD, many autonomic disturbances, such as hypotension, hypersalivation and constipation, which are a part of a parkinson-plus syndrome, are present and also can be attenuated by DOPA-application (Guttman and Seeman, 1986). Therefore, it can be suspected that these symptoms also result from a central DA deficiency, perhaps related to the observed biochemical effects on DA-containing centers in the hypothalamus and area postrema. In PD brain, there is a significant decrease of serotonin and its metabolite 5-hydroxyindoleacetic acid in fiber tracts from the nuclei raphe to the forebrain, in forebrain areas, basal ganglia, hippocampus and cerebral cortex (Mayeux, 1990) (Fig. 1). A strict attribution of defective transmitter systems to symptoms is not really possible. Nevertheless, it has been proposed that depression in PD is related to a decrease in the serotonergic system (Agid et al., 1987), while the known decline in noradrenergic coerulocortical connections is thought to contribute to depression in PD (Mayeux 1982, 1990) and, furthermore, to provoke a cognitive decline. A very interesting field with regard to the connection between several neurodegenerative diseases is the appearance of dementia in PD patients. As reviewed by Boiler (1985), clinical dementia is a frequent finding, of which the etiologic explanation is far from clear. Pathological investigations have not been able to show a higher incidence of Alzheimer-specific findings in PD than in controls. Nevertheless, a significant percentage (17% younger than 70 years, 35% aged over 70 years) of PD brains shows typical alterations as observed in AD (Jellinger, 1986). Because of the possibility of PD-associated dementia without corresponding AD pathology, the conclusion has been drawn that, analogous to the DA decline of 80-90% necessary for clinical manifestation of motor disturbance, a comparable reduction of acetylcholine (ACh) in several areas must have occurred before the appearance of clinical signs of dementia (Whitehouse et al., 1983a; Jellinger, 1989; Ruberg et al., 1989; Agid, 1991). In the striatum and SNC of PD patients, there is, however, no alteration of the cholinergic system. In contrast, in the striatum, a

Oxidative stress: free radical production in neural degeneration

functional hyperactivity occurs. Some degeneration of cholinergic systems, as measured by choline acetyl-transferase (ChAT) immunohistochemistry, occurs in the NBM-cortical and the septohippocampal pathway. Concerning the ACh receptors, a reduction only of the hippocampal nicotinic receptors has been detected in either AD or PD (Perry, E.K. et al., 1987), whereas the muscarinic receptor density in PD frontal cortex can even be elevated (Ruberg et al., 1989). Signs of ChAT, ACh and nicotine receptor alterations are absent from thalamic nuclei and subthalamic nucleus in both AD and PD. The contradictory role of DA in cognitive disturbances has been discussed recently (Mohr et al., 1989; Dubois et al., 1992), while the reduction of norepinephrine in coerulocortical neurons, which has also been found in demented patients (Cash et al., 1984; Hardy et al., 1985), could also contribute to the dementia in PD. Further parallel biochemical findings between PD and AD are: (1) diminution of substance P in cortical regions of demented parkinsonian patients (Clevens and Beal, 1989) with Alzheimer-specific pathological findings in SNC, SNR, NBM and internal globus pallidus; (2) reduction of cortical corticotropin releasing factor (Whitehouse et al., 1987); and (3) low somatostatin content in cerebrospinal fluid, correlating more with the degree of motor disadvantage than with that of dementia (Strittmatter and Cramer, 1992). Other investigations have also elaborated a decline of somatostatin in frontal cortex and hippocampus of demented parkinsonian patients, which corresponds to the decline of ChAT (Eppelbaum et al., 1983). 2.1.6. Pathophysiology of motoneuron disease Lesions of the corticospinal tract produce both positive and negative signs. The positive signs originate from normally suppressed, inadequate motor responses to sensoric stimuli. Negative signs correspond to the affected ability to perform volitional movements forcefully and rapidly. Lesion of the upper motoneuron results in weakness of groups of muscles and a spastic mus-

M.E. Gotz et ai

cle tone, which is produced by a brain stemmediated increased excitation of a-motoneurons. Lesions of the lower motoneuron affect single muscles. Weakness, atrophy and a slack muscle tone are the clinical symptoms of this damage. A further symptom occurs after degeneration of a motoneuron: the muscle fibers innervated by it discharge sporadically and not in coordination with other muscle fibers that are still controlled by intact motoneurons. These visible single discharges are called fasciculations (Ghez, 1991). 2.7.7. The clinical manifestation of amyotrophic lateral sclerosis In the typical form of ALS, the upper and the lower motoneurons of the spinal cord and the bulbar system are affected. The oculomotor nuclei and the sphincter muscles of the anus and bladder are spared. When the upper motoneuron is predominantly affected, the clinical term 'primary lateral sclerosis' or 'progressive pseudobulbar palsy' is used, while isolated affection of the lower motoneuron is called 'progressive muscle atrophy' or 'progressive bulbar palsy' (Tandan and Bradley, 1985a). In the Scottish motoneuron disease register, ALS, progressive bulbar palsy and progressive muscle atrophy are characterized as clinical subtypes of a single disease, whereas primary lateral sclerosis is considered to be a separate disease (The Scottish Motor Neuron Research Group, 1992). The overall term for all these disturbances is motoneuron disease. ALS is a chronic progressive degenerative disorder, which, in its classical form, appears sporadically. It is noteworthy that there exists a familial form of ALS in 5-10% of the cases (Roe, 1964; Finlayson et al., 1973; Rosen, 1978). This form has been described as being of early onset, with a longer course, as compared with the predominant late onset form, with a shorter course. A study exists that was able to detect a familial ALS form caused by a gene on chromosome 21 (Siddique et al, 1991). The combination of ALS with PD or dementia is not so rare and appears in a familial and sporadic form (Hudson, 1981). Encephalitis lethargica, usually seen in connection

431

with PD, has also sometimes caused a combination of ALS and PD (Hufschmidt et al., 1960). In the Western Pacific form, ALS appears as one member of a complex disease in association with PD and dementia. ALS may occur as a consequence of a remote poliomyelitis infection, as reported in cases of ALS in patients who suffered from a polio infection years ago. Proof of a connection between these two diseases is still lacking. There also sometimes seems to be metal intoxication in ALS patients. The best investigated toxic metal is lead, which has been described as a trigger of ALS symptoms in many cases, although there is no knowledge of a potential pathophysiological mechanism (Rowland, 1984; Tandan and Bradley, 1985b). The average annual incidence rate of ALS is about 1:100 000: a male predominance of 2:1 is described, the mean age of onset is 56 years, the duration of disease 3 years (Hudson, 1981; Tandan and Bradley, 1985a; Li et al., 1990). In the majority of cases (43-37%), the disease starts in the upper or lower limbs; in a minor percentage (20%), in the bulbar muscles (Li et al., 1990). The first symptoms are mostly difficulty in walking or diminution of manual dexterity due to increasing muscle weakness. The brain stem symptoms are difficulty in speaking or swallowing. Finally, the patient becomes unable to speak or swallow food or saliva. The esophageal dysmotility results in dysphagia and predisposes to aspiration. An additional respiratory failure makes respiratory support necessary to keep the patient alive. In the classical form of ALS, the patient's cognition and consciousness is unimpaired until death (Rowland, 1984; Tandan and Bradley, 1985a). 2.1.8. Neuropathology of amyotrophic lateral sclerosis A number of pathological changes are seen in the CNS of ALS patients, with an expected regional distribution. The most prominent of these is a loss of large motoneurons in the motor cortex, brain stem and spinal cord. It has been debated whether this degeneration represents a dying back mechanism, spreading from the axon back to the

432

perikaryon (Appel, 1981; Rowland, 1984). The remaining neurons show atrophy, lipofuscin accumulation or LB-like or eosinophilic (Bunianbodies) intracytoplasmic inclusions (Appel, 1981; Tandan and Bradley, 1985a). A non-obligatory, but frequently seen, change is a cerebral atrophy and abnormal gliosis in cerebral cortex, striatum, pallidum, subthalamic nucleus and SN (Hudson, 1981). The Guamanian ALS is characterized pathologically by many neurofibrillary tangles (NFT) throughout the brain (Kurland, 1988). In familial ALS combined with dementia, there is a symmetrical involvement of amygdala, insula and frontotemporal cortex, whereas the cell loss in spinal cord anterior horn is identical to the sporadic changes seen in uncomplicated ALS. 2.7.9. Biochemistry in amyotrophic lateral sclerosis The biochemical changes in ALS-affected patients have been widely and inconsistently construed. L-Glutamate is the neurotransmitter of the primate tractus corticospinalis; from the degeneration of this system a derangement of glutamatergic systems results (Eisen and Calne, 1992). Since the ALS of Guam is highly suspected to be caused by an excitatory amino acid (EAA) affecting glutamate receptors, a damaging role of glutamate has been proposed in sporadic ALS also (Meldrum and Garthwaite, 1990, 1991; Farooqui and Horrocks, 1991). However, other authors have found a glutamate deficiency in sporadic ALS brain which is absent in ALS of Guam (Plaitakis et al., 1988; Plaitakis, 1990; Perry et al., 1991). A reduction of muscarinic receptor sites in the ventral horn has also been found to be related to this neurodegenerative disorder (Whitehouse et al, 1983b), while the observation of decreased citrate synthase activity of isolated anterior horn cells has led to the proposal that neuronal ACh production may be depressed (Hayashi and Tsubaki, 1982). Other biochemical abnormalities reported for ALS patients have included increased anterior horn levels of ornithine and ammonia (Patten et al., 1982), reduced glycine receptor binding in anterior horn (Hayashi and Tsubaki, 1982) and reduced levels of y-amino-butyric acid (GABA) and

Oxidative stress: free radical production in neural degeneration

norepinephrine in the cerebrospinal fluid. A lack of some specific motoneuron growth factor has also been postulated (Appel, 1981). 2.2. Alzheimer's disease, a cognitive disorder 2.2.1. Physiology of cognition and emotion The hippocampal formation as a relay station in the origin of emotions was proposed by Papez in 1937 in the circuit constructed by himself (Gray, 1987). The gyrus cinguli receives afferents from a mamillothalamic tract and sends inputs to the hippocampal formation, from where neurons project back to the gyrus cinguli. Neocortical areas send their information to the gyrus cinguli via a hippocampo-mamillo-thalamic pathway. In a new version of the classic Papez circuit, the nucleus accumbens has been added as a basal ganglia structure with emotional function that receives afferents from thalamic nuclei and from the gyrus cinguli directly or via a cingulo-hippocampal loop (Fig. 1). While involvement in emotion is one function of the hippocampal formation, contribution to cognition and memory is another. For instance, kindling experiments involving hippocampal structures provoke disturbances of memory (Gaffan, 1987). The pathological and biochemical alterations in the hippocampal formation of Alzheimer patients discussed in Sections 2.2.3 and 2.2.4 show the importance of this anatomical substrate in learning and cognition. As reviewed by Rolls (1990), the hippocampus proper consists of a system of high plasticity, which is modified in its specificity by long-term potentiation. In this process, randomly arriving inputs are filtered and, thus, specified by partly increasing and partly decreasing strength of synapses between the pre- and postsynaptic neuron. This neuronal plasticity means that neurons discharging at high frequency develop closer synaptic linkages with the postsynaptic neuron, whereas subthreshold firing is not able to alter synaptic properties. Because of this process, partly mediated by A^methyl-D-aspartate (NMDA)-responsive receptor activation (Monaghan et al, 1989), the hippocam-

M.E. Gotz et al.

pus becomes a memory and recognition organ for individual experiences. Neocortical inputs entering the hippocampal formation induce synaptic plasticity, and back projections to the neocortex prepare cortical areas for further perceptional stimuli by storing the hippocampus-mediated information. As a consequence of this learning system, the neocortex is enabled to react faster and more specifically upon a known stimulus (Rolls, 1990). 2.2.2. The clinical manifestation of Alzheimer's disease The first case of AD was reported in 1907 by Alois Alzheimer, who investigated a woman who displayed symptoms that today are characteristic of a progressed stage of the disease. After death, he found histologically cortical NFT and plaques, a finding that represents a milestone in the history of this disease (Brun et al., 1990). According to the definition of dementia given by the World Health Organization, this term means 'an acquired global impairment of higher cognitive functions including: memory, the ability to resolve problems of daily living, the performance of sensorimotoric and social functions, language communication and control of emotional reactions without marked reduction of consciousness. This process is mostly progressive but not absolutely irreversible.' The diagnostic criteria for dementia (American Psychiatric Association, 1987) represent a more detailed description of this definition, confirming that the term 'dementia' is adequate only when the above-mentioned disturbances of personality, memory, language, abstract thinking, judgement and social behavior that interfere with work and social activities occur in the absence of delirium, functional or organic impairment, such as major depression (Thompson, 1987). The clinical diagnosis of AD is made in the absence of histopathological findings. However, comparing the percentage of diagnostic pathological and clinical agreement, we see that there has been an increase of correct clinical diagnoses over the last 10 years. Having diagnosed AD in an aged person, other dementias, such as multi-infarct dementia. Pick

433

disease and HD, must be excluded. It is also important to distinguish between dementia and pseudodementia in depressive patients. It is a helpful hint that true dementia worsens at night and becomes more prominent when antidepressant, anxiolytic or hypnotic drugs are prescribed (Thompson, 1987). The age of onset of the disease can be about 40 years, but more usually, it begins at ages over 60 years (Khachaturian, 1985). Two to five percent of individuals over 65 years suffer from dementia, and half of these have AD (Thompson, 1987). It is probable that the rate doubles every 4 5 years from 60 to 90 years of age. Further, a higher incidence in women is apparent. Investigations have led to the conclusion that some AD cases have a positive familial history of the disease. In some cases, there is an upward shift in maternal ages and in some, an increased frequency of Down's syndrome is found in relatives (Henderson, 1990). The course of dementia is progressive and can be subdivided into three phases: in the early phase, there is no cognitive impairment, but the patient complains of other diffuse symptoms. He suffers from anxiety, mild depression, feeling frustrated, insomnia or multiple somatic symptoms. The cognitive deficits can still be compensated. A marked alteration of personality develops, and patients show conspicuous accentuation of certain character qualities. In the middle phase the cognitive decline can no longer be compensated. There is a memory impairment firstly for recent events, later for longdistance events. Arithmetical properties, judgement and social living are disturbed. In the late phase, patients often develop paranoid ideas, are no longer able to recognize their family members, lose completely the capacity to participate in conversation and cannot perform the simplest activities, such as feeding themselves and caring for their bodies. The consequences are reduction of body weight and incontinence (Thompson, 1987). Some patients die in bed, a consequence of decubitus or pneumonia; others die from accidents that occur due to the spatial disorientation.

434

The oldest group of demented patients differs clearly from the younger. In this group, head trauma and cerebrovascular diseases often complicate a mild subclinical Alzheimer dementia. The patients mostly die of non-dementia-related organic diseases (Brun et al., 1990). 2.2.3. Pathology of Alzheimer's disease Cognitive decline is the essential criterion of AD during life, but only after death can the diagnosis be confirmed pathologically. The pathology of normal, aged persons and mildly demented patients is broadly overlapping, so that only after crossing the defined neuropathological threshold can the diagnosis be made (Roth, 1986). The autopsy criteria for AD described by Alois Alzheimer in 1907 were NFT and neuritic plaques. The presence of NFT is the presupposition for the diagnosis of AD. Whereas in brains of normal, aged persons, NFT in greater extent are never found, numerous amyloid and neuritic plaques can frequently be observed (Perry, E.K., 1986; Perry, R.H., 1986). In order to diagnose AD, at least three neocortical regions, amygdala, hippocampal formation, basal ganglia, SN, cerebellar cortex and spinal cord should be examined (Khachaturian, 1985). The distribution of neuritic plaques is irregular, whereas the NFT show a characteristic pattern with minimal interindividual variations. The pathologic process develops in a clear temporal order. The region first affected is the transentorhinal region, followed by the regio entorhinalis, hippocampal formation, isocortical regions and extrapyramidal system (Braak and Braak, 1991) (Fig. 1). The senile plaques represent extracellular round or ovoid structures, their diameters ranging between 1.5 and 20 nm. Typically, these plaques consist of three components: abnormal nerve processes, glial processes and a central or amyloid core. There are different stages in plaque development: primitive plaques consisting of neuritic components only, the classic plaque with central amyloid core surrounded by dystrophic neurites and, in the final stage, the burned-out plaque with

Oxidative stress: free radical production in neural degeneration

a great amyloid core and no or only a small surrounding neuritic border. Biochemically, catecholaminergic and cholinergic components have been detected (Perry, E.K., 1986; Perry, R.H., 1986). Besides amyloid-containing neuritic plaques, pure amyloid deposits, without pathologic neuritic components, also exist. Such deposits can appear without accompanying neurofibrillary changes, whereas severe neurofibrillary changes characteristically accompany amyloid plaques (Braak and Braak, 1991). The NFT are an abnormal intracytoplasmatic accumulation of neurofilaments. They develop within the nerve cell soma, from where they can extend into the dendrites. The parent cell may disappear and the NFT persist as ghost tangles (Braak and Braak, 1991). Ultrastructural investigations of the NFT show that they are composed of altered neurofilament peptides that form paired helical filaments (PHF). The PHF are composed of two identical filaments twisted around each other with a periodicity of 80 nm (Perl and Pendlebury, 1987). The PHF are insoluble, covalent-bound cytoskeletal elements of high stability that, therefore, could disrupt intraneuronal axonal transport and cytoskeletal metabolism. By biochemical and immunochemical techniques it is possible to stain three PHF components: microtubule-associated protein (MAP), tau and ubiquitin. The question is whether the PHF represent intracellular amyloid deposits (Masters and Beyreuther, 1990). Both the PHF and extracellular amyloid plaques are formed from globular subunits that consist of aggregates of A4 protein. Experimentally, the amyloid protein precursor fragments are transformed to their insoluble aggregating form by meta-catalysin oxidation systems. This transformation can be prevented by radical scavengers (Dyrks et al., 1992). The gene of amyloid precursor protein (APP) is located on chromosome 21. It is speculated that the precursor A4 protein (APP) forms intracellular NFT and amyloid plaques in the extracellular space (Masters and Beyreuther, 1990; Beyreuther et al., 1991). On the question of whether NFT consist of A4 protein or not, opinions are still divided, but intracellular NFT are stained by both APP and

M.E. Gotz et al.

NFT antibodies (Muq)hy et al., 1992). As well, in hippocampal and cells of NBM in AD patients, an overexpression of APP was detected by in situ hybridization (Goldgaber and Schmechtel, 1990). The gene location of APP on chromosome 21 is interesting because patients suffering from trisomia 21 develop NFT by age 35-40 years. In cases of sporadic AD, a duplication of chromosome 21 around the A4 protein locus has been reported. This allows the logical conclusion that the pathological changes of both AD and Down's syndrome are due to an increased, A4 protein production (Wisniewski et al., 1988). 2.2,4. Biochemical alterations in Alzheimer's disease In conventional terminology, AD, as cortical dementia, is distinguished from the subcortical forms that accompany diseases of subcortical structures, such as PD, HD, morbus Wilson and progressive supranuclear palsy. This classification neglects the fact that, in both forms, both cortical and subcortical systems are affected: NBM and LC in AD and cortical areas in the subcortical dementias. Therefore, a clear distinction in biochemical findings of the two forms cannot be expected. Rather, a reasonable course of study might be to examine common biochemical characteristics of both forms (Rossor, 1985). A further difficulty is that the typical pathological changes in neocortex and hippocampus represent valuable diagnostic criteria in young AD patients, but can be numerous in brains of nondemented aged persons (Khachaturian, 1985). Thus, where to draw the limits between pathology and physiology of aging becomes very problematic. Clearly, biochemical and pathological data can only be interpreted usefully in combination with clinical observations. The neurotransmitter profiles in normal aging and AD must be interpreted with caution, keeping in mind the influence of terminal conditions (medical treatment, hypoxic states, cachexia) on post mortem measured transmitter content (Hardy et al., 1985). In normal aged brain, a decline of TH in putamen and nucleus caudatus, of ChAT in nucleus caudatus and of glutamic acid decarboxylase

435

and DA in putamen and nucleus caudatus has been described, although the ChAT decline was not consistent. Norepinephrine decline was found in hippocampus, 5-hydroxytryptamine decline was measured in gyrus cinguli. A prominent finding in all investigated regions is an increase of monoamine oxidase (MAO)-B in comparison with that in young control brains; this could be interpreted as one marker for neurodegeneration in the aging brain. Thinking is a complex process that includes several levels and activities. A certain attentional state is necessary to absorb new information. Information must then be transferred from a shortterm memory depot to long-term memory storage. In order to apply stored information usefully, new contents must be compared with that of semantic storages in an associative manner. In order to resolve problems further, we make use of logical convergent and creative divergent thinking. The former depends on high attention and tends to try known solution concepts, the latter appears at a low arousal state and enables one to find new simple solutions. These aspects of thinking can be connected to neurotransmitter systems. Numerous experiments have been able to show that anticholinergic drugs decrease focused attention and inhibit transformation of information from short- to long-term memory storage, whereas cholinergic agonists have the opposite effect. Concerning norepinephrine systems, we know that they facilitate associative thinking (Holttum and Gershon, 1992). Starting from this point, the goal of investigating these neurotransmitter systems in demented persons is a logical consequence. The synthesis of ACh from mitochondrial coenzyme A and extracellular choline is mediated by ChAT (Tucek et al., 1990). By electrical stimulation of cholinergic neurons, a large production of ACh has been seen, despite a complete lack of extracellular choline. Therefore, the suggestion has been made that neurodegeneration develops by an autocannibal mechanism taking the choline of membrane phosphatidylcholine for ACh synthesis, which is followed by membrane destruction (Wurtman et al., 1990).

436

The impairment of tlie cholinergic system in AD is an undoubted fact. ACh histochemistry and ChAT immunohistochemistry in NBM of AD brains show marked reductions, corresponding to cell loss of about 75%. In cortical target areas, there is also a cholinergic pathology, which may provoke the cholinergic deficit in NBM by retrograde degeneration (Perry, E.K., 1986; Perry, R.H., 1986). The hippocampal formation, the anatomic structure most closely related to the cognitive decline in AD, is innervated intrinsically by glutamatergic and GABAergic neurons, and extrinsically by cholinergic projections from frontal forebrain. These topographically organized cholinergic projections originate from the medial septal nucleus and the vertical limb of the diagonal band of Broca. The projections end in CA2/3, hilus and gyrus dentatus of the hippocampus (Emson and Lindvall, 1986). Measuring the ChAT activity in neocortical areas, 70% is estimated to belong to extrinsic projections from basal forebrain and 30% to intrinsic cholinergic neurons (Emson and Lindvall, 1986). Reduced ChAT activity in temporal, frontal and parietal neocortex, hippocampus and NBM from both autopsy and biopsy specimens of AD patients further demonstrates the degeneration of cholinergic cells. Since ChAT is a marker for presynaptic ACh synthesis, the diminution is interpreted as presynaptic axon degeneration. Therefore, the degenerating axons should belong to NBM neurons that project strongly to the frontal neocortex (Adolfsson et al., 1979; Gottfries, 1985a,b; Hardy et al., 1985; Ichimiya et al., 1986; Perry, 1987; Sparks et al., 1988). On the other hand, cases have also been described that show no cell degeneration in NBM, despite cortical ChAT diminution (Etienne et al., 1986; Quirion et al., 1986). Nevertheless, overall there is a relationship between quantity of specific neuropathological findings (NET, plaques), and ChAT-deficiency and choline uptake is also reduced. A neurotransmitter reduction is also seen in the monoaminergic system of AD brains. Whereas norepinephrine and serotonin are shown to be reduced by most investigators (Mann et al., 1980;

Oxidative stress: free radical production in neural degeneration

Ichimiya et al., 1986; Palmer et al., 1987; D'Amato et al., 1987; Blennow et al, 1991), the statements about DA are quite different (Adolfsson et al., 1979). Sometimes a diminution of DA in the striatum and nucleus caudatus exists reproducibly (Rossor, 1985; Bowen and Davison, 1986). Norepinephrine reduction can be due to a severe cell loss in LC of AD brains. Serotonin, its metabolite 5-hydroxyindole-3-acetic acid and norepinephrine are measured to be decreased in hippocampal formation. The substantia innominata may be unaffected (Baker and Reynolds, 1989). The loss of serotonin correlates to many neuropathological findings in serotonergic cells of the nuclei raphe (Quirion et al., 1986). Somatostatin is the only consistently reduced neuropeptide in AD. It is directly involved in histopathological changes, as evidenced by the fact that it has been found in NFT-containing cells and plaques. In frontal and temporal cortex, a decline in content of somatostatin and substance P is possible (Hardy et al., 1985; Quirion et al, 1986). There is no alteration of GABA content in AD post mortem brain, despite occasional detecting of glutamate decarboxylase diminution. Whereas GABA is not affected by perimortal circumstances, reduction of glutamic acid decarboxylase activity is often due to premortal hypoxia (Bowen etal., 1990). The role of EAA in cognition and learning is well established (Monaghan et al, 1989). The high level of EAA receptors in neocortical and hippocampal regions is also undoubted (Young and Egg, 1991). Therefore, in AD, there seems to be a logical connection between cognitive impairment, neuropathologic changes in neocortical and hippocampal areas and alterations of EAA. The neurodegeneration caused by EAA has been shown experimentally, just as has the vulnerability of limbic and cortical neurons to this neurotoxic effect. Instead, corresponding studies of AD brains showed a reduction of EAA-binding sites in hippocampal areas (Jansen et al., 1990; Geddes et al, 1992). As further proof of this connection, we can take investigations that explain that EAA alter the polymerization of cytoskeletal elements in cultured cells.

M.E. Gotz et al.

a fact that may elucidate the origin of NFT in AD (Katoetal., 1992). 2.3. Pathological overlapping of neurodegenerative diseases In approximately 10% of persons over 60 years and also in 15% of Alzheimer patients, LB can be found (Forno, 1986). Another term is the 'diffuse LB disease' which is clearly distinguished from PD: here the LB infiltrate the limbic system and cortical areas; the clinical correlate is dementia with or without PD (Gibb, 1989). Tests exist to distinguish between parkinsonian LB and diffuse LB disease: using antineurofilament antibodies, the LB of the latter disease were stained by rprotein antibodies only (Galloway et al., 1989). The immunoreactivity against other cytoskeletal elements, such as ubiquitin, was the same in PD and diffuse LB disease (Galloway et al., 1989; Dale et al., 1992). Many cases of unclear parkinsonian-like movement disturbances are misdiagnosed as PD (Feamley and Lees, 1991). The striatonigral degeneration, Steel-RichardsonOlszewsky syndrome and corticobasal degeneration can be distinguished from PD by the lack of LB in SNC and the presence of structural pathology outside the SNC (Gibb, 1989). The LB can be taken as a good example for a pathological, pathophysiological correlation between diverse degenerative disorders. Other examples are the NFT and the neuritic plaques. The NFT consist of intracytoplasmatic filaments with antigen determinants of abnormal neurofilament and non-neurofilament protein and of microtubules. The senile plaques are spherical dense structures composed of pre- and postsynaptic degenerating and regenerating neuritic terminals, abnormal synapses, glia, filamentous protein and extracellular amyloid (Jellinger, 1989). NFT and senile plaques are typical structural pathological findings in AD, where they are found in the limbic system, cortical areas and NBM. The number of NFT corresponds to the stage of the disease and the several brain regions are invaded in a specific order. At the beginning of the disease.

437

NFT appear in the regio entorhinalis, then progress to the hippocampus and later involve neocortical areas (Braak and Braak, 1991) (Fig. 1). The NFT are an irreplaceable index proof of AD, whereas senile plaques are not necessarily linked with pathological diagnosis (Braak and Braak, 1991). The finding of single NFT and senile plaques in the brain of non-demented elderly persons is not unusual (Katzman and Saitoh, 1991). Comparing pathological findings in early and late onset PD, a similar Parkinson-specific pathology can be detected, whereas changes typically found in Alzheimer's brains are more frequent in the later onset of PD, a fact that shows a probable decompensation of subtle damages by an additional influence of aging processes (Jellinger, 1986). Agedependent number of NFT and senile plaques correspond to the clinical signs of dementia in PD. Clinically, senile parkinsonism presents dementia, mild parkinsonian signs and a lack of L-DOPA responsiveness. Pathologically, there is a mild cell loss in SNC, LB in 50% and NFT in 60% of the cases (JelHnger, 1989). The prevalence of dementia and pathologically diagnosed Alzheimer's changes in PD patients is 6-fold higher than in the age-matched, healthy population (Boiler, 1985). Also described are cases of L-DOPA responsive, clinically non-demented PD, which on pathological examination, show a severe decline of neurons in SNC, without LB or other pathological findings outside this region, which could support the diagnosis of a PD-like syndrome. At the same time, many NFT and senile plaques were detected in various brain regions. The severity of dementia in PD patients correlates to the severity of Alzheimer's pathology in parkinsonian brains whereby a mild degree of dementia can be ascribed to degeneration of SNC, and a marked degree of cognitive impairment results from additional cortical Alzheimer lesions (Paulus and Jellinger, 1991). The neuropathological distinction between PD/dementia and PD plus dementia is very difficult. One test is to compare the degree of brain atrophy in several regions by photographic analysis of coronare slices. The result shows a global cerebral

438

atrophy in PD plus AD with moderate atrophy of white matter and a more partial cerebral atrophy without impairment of white matter in PD dementia (De la Monte et al., 1989). As we can see, pathological findings in different neurodegenerative disorders overlap to a great extent. As a probably useful model for a common etiopathologic cause in clinically diverse neurodegenerative diseases, the Guam disease can be considered to be a combined syndrome of PD, AD and ALS. This disease shows neuropathological findings of PD and AD: loss of neurons and appearance of LB and NFT in cortical and subcortical structures. In the spinal cord of ALS and PD patients of Guam, NFT can be observed. Distribution and number are similar in both diseases. In immunohistochemistry, a reactivity against anti-r-protein antibodies has been seen, a result that could not be confirmed by examination of the cytoplasm of sporadic ALS and controls (Kato et al., 1992). A hypothesis of abiotrophic interaction between aging and environment in PD, ALS and AD exists (Calne et al., 1986). Another way to reach a unifying hypothesis for several neurodegenerative disorders is to analyse inclusion bodies linked with it. The result is that the inclusion bodies in all these cases are composed of cytoskeletal components, so that neurodegeneration of several functional systems can be subordinated to a common term of cytoskeletal disorders (Calne and Eisen, 1989). 3. Possible biochemical causes of cell degeneration 3.1. Oxidative stress 3.1.1. Reactive oxygen species Metabolically, the brain is one of the most active organs in the body. This is reflected by cerebral O2 consumption in normal, conscious, young men which amounts to 3.5 ml O2 per 100 g brain per min (Sokoloff, 1960). Thus, 2% of total body weight accounts for 20% of the resting total body O2 consumption. Nearly all O2 is utilized for the oxidation of carbohydrates (Sokoloff, 1960) and results in an estimated steady-state turnover of

Oxidative stress: free radical production in neural degeneration

approximately 4 x 10^^ molecules of ATP per min in the entire human brain. Oxygen maintains brain function and is crucial for life. However, O2 supplied at concentrations greater than those in normal air is highly toxic. High pressure O2 can lead to convulsions, which are attributed to an inhibition of the enzyme glutamate decarboxylase by O2 or reactive oxygen species (ROS) (Halliwell and Gutteridge, 1989). Even normal O2 consumption could lead to toxic cellular reactions mediated by oxidative stress. 'Oxidative stress' is an expression used for a process that implicates reactions with biomolecules of O2 or derived substances, such as hydrogen peroxide (H2O2), superoxide [(O2)*'], hydroxyl radicals (OH)' or singlet 0*2. If a reaction is thermodynamically feasible, its reaction rate depends primarily on the concentrations of the reacting partners. Thus, to evaluate effects of ROS on biomolecules, their concentrations and sites of production have to be considered. In the following Section, some general comments on chemistry will be made before discussing the biochemistry of radicals in relation to neurodegeneration. Groundstate O2, is in the triplet or diradical electronic configuration, having two unpaired electrons, each located in a different 7C* antibonding orbital. These two electrons have the same spin quantum number ('parallel spins') in contrast to singlet O2, which has antiparallel spins. A description of molecular orbital chemistry in the biomedical context has been made by Halliwell and Gutteridge (1984). A prerequisite for exergonic reactions is the rule that reacting electrons in an energetic groundstate have to have antiparallel spins. Thus, in order to achieve spin conversion, groundstate O2 must react in a two-step, energydependent process. This is the reason for the slow reactivity of groundstate O2 when this energy is not provided by enzymes or light (McMurry and Groves, 1986). Among all oxygen species, dioxide(l-) [(O2)*] (which is a new term for the old, but still allowed, name 'superoxide'), (HO") and H2O2 are supposed to be the most abundant ROS in biological systems. Although O2 is a diradical and may be represented as (O2):, we prefer to omit the radical dots in the cases of groundstate O2 and of

439

M.E. Gotz etal. NADP NADPH + H

Amines + 0 ^

J

• Products + H^O^

H^O

GSSG'Rd GSSG

^

GSH'Px NADP"*"^

O

1 e 2

Fe

._ ^O 2 + 2

-^ 2 GSH

1e

+ 2 H

• H O 2 2

^^r^

SOD

• 2 H O ^2 H O + O 2 2 ^^^ 2 2

r- Fe

Cu

1 e

H^O^

+ 1H

+

^ ^

I Cu "*"+ O - ^ J _

1 e

CAT

2+

3+

OH

2

^ ^

OH

1H Fig. 2. Possible redox reactions leading to ROS (H2O2; (O2)'; (OH)') and pathways degradating H2O2 directly via CAT or via NADPH-dependent mechanisms utilizing GSH. Adapted from Benzi et al., 1988.

transition metal ions. For a discussion of new nomenclature for oxygen species, see Koppenol (1990). Dioxide(l-), superoxide, (02^ (02)^ is mainly produced in biological systems through oneelectron reduction of triplet O2 mediated by enzymes (Fig. 2). In brain, xanthine oxidase (EC 1.2.3.2) and aldehyde oxidase (EC 1.2.3.1), located in nuclear membranes and cytoplasm, seem to be the only enzymes responsible for (O2)* production (Eqs. (1) and (2)). RCHO + H2O + 2O2 -^ RCOO- + 2(02)^ + 3H+ (1) Xanthine + H2O + 02-^ Urate + (O2)" + H+

(2)

As well, these two enzymes generate H2O2 and are dependent on p02 and pH. Xanthine oxidase can oxidize a variety of substrates, including aldehydes, pteridines, purines and hypoxanthine

(Halliwell and Gutteridge, 1989). Xanthine oxidase can be converted from a dehydrogenase (nonsuperoxide-producing) to the oxidase form during tissue hypoxia (Granger et al., 1981). The relative rates of (62)^ production by these enzymes may vary with the concentrations of the enzymes in various cell types and the availability of substrates and cofactors (for (62)^ production by mitochondria and microsomes see Section 3.2.2). (O2)* can behave as a free radical, a weak nucleophile, a one-electron oxidant or a one-electron reductant (for a review, see Fridovich, 1986a). Free radicals usually are very likely to abstract hydrogen or to add to double bonds; however, as for superoxide, these reactions are slow (Bors et al., 1979). Only if protons (3) are present will the reactive free radical hydrogen dioxide (HO'2) be formed. (02)^ H+<^ HO'2 p/^a = 4.88

(3)

440

Oxidative stress: free radical production in neural degeneration

Thermodynamically {O^)'' tends to dismutate to H2O2 and O2 (Eqs. (4) and (5)). (02)^ + H+ <-> HO'2

(4)

HO'2 +(02)^ + H+ <^ H2O2 + O2

(5)

In aprotic media, (02)"" behaves like a strong nucleophile but in protic media, it is only weakly nucleophilic, due to the presence of protons. Thus, reactions of (O2)* in protic media are determined by kinetic rather than by thermodynamic parameters (concentrations, pH, ionic strength). The most important reaction in terms of biological effects is the dismutation of (02)"' (Eqs. (4) and (5)), which proceeds rather slowly in physiological conditions (ATdis-S X lO^M-^s-i at pH 7.4; Halliwell and Gutteridge, 1989), but can be considerably favoured by superoxide dismutases (SOD) (EC 1.15.1.1) (Chance et al., 1979; K^,, ~ 1.6 x 10^ M-^ s-i). (O2)* can act as a reductant (Eq. (6)) of peroxides only if transition metals (Mn+) are present (Sawyer and Nanni, 1981; Wood, 1988) (Eq. (7)) or of quinones (Eq. (8)) (Sawada et al., 1975). O2 + e- <-^ (O2)' E"" = -0.33 V in protic media (02)^ + H2O2
Mn*

^ HO- + (HO)* + O2

Quinone + (02)^ <-^ Semiquinone radical + O2

(6) (7) (8)

On the other hand, (02)^ can act as a oneelectron oxidant, oxidizing e.g. hydroquinones to semiquinone radicals, or oxidizing ascorbate or epinephrine with concomitant production of H2O2. Thus, if dismutation reactions and scavenging of (02)^ are impaired, degradation of alkylperoxides (ROOK) to alkoxylradicals (RO)^ via Eq. (9) could lead to potent cytotoxic substances. HO'2 + ROOH ^ 0 2 + RO* + HO- + H+

(9)

(02)"" biochemistry is strongly influenced by transition metals. Tyler (1975) reasoned that lipid peroxidation (LPO) in membranes occurs only in the presence of iron (for iron and oxidative stress see Section 4.1.4).

Hydrogen peroxide, iron, copper and hydroxyl radical The metabolism of H2O2 in mammalian organs was reviewed by Chance et al. (1979). Enzymes known in the liver to generate H2O2 are assumed to be also present in the human brain. These oxidases bear flavins (flavin mononucleotide (FMN), flavin adenine dinucleotide (FAD)) or pyridoxale phosphate (Pj) and metal ions as prosthetic groups. Pyridoxamine phosphate oxidase (EC 1.4.3.5) triggers H2O2 production (Eq. (10)). Pyridoxamine-Pi + H2O + 0 2 ^ Pyridoxale-Pi + NH3 + H2O2

(10)

D-Amino acid oxidase (EC 1.4.3.3) (Eq. (11)), converting glycine to glyoxylate, is present in the CNS (Gaunt and De Duve, 1976), but glycine is a very poor substrate for the enzyme (DeMarchi and Johnston, 1969). ^-RCHNH2C00H + H2O + 0 2 - ^ RCOCOOH + H2O2 + NH3

(11)

For R = H the reaction product of Eq. (11), glycolate, is converted by glycolate oxidase (EC 1.1.3.1) to glyoxylate and H2O2. The most prominent oxidase in brain tissue, however, is the flavin-containing MAO (EC 1.4.3.4), which preferentially deaminates primary, secondary and tertiary monoamines (Fig. 2) according to Eq. (12) and is located at the outer mitochondrial membrane (see Section 4.1.1) (Tipton, 1967). RCH2NH2 + H2O + 0 2 - ^ RCHO + H2O2 + NH3 (12) Moreover, a great source of H2O2 production in the intact cell appears to be generated by the autooxidation of chemically reactive compounds during reductive processes associated with the mitochondrial and microsomal electron transport systems and during action of SOD (Chance et a l , 1979; Forman and Boveris, 1982). In rat liver, microsomal compartments make the greatest contribution (about 5% of liver uptake of

441

M.E. Gotz et al.

O2) to the intracellular oxidant load, followed by peroxisomes, mitochondria and cytosol (Boveris et al., 1972). If not enhanced by transition metals or enzymes, such as catalase (CAT) (Eq. (13)) in peroxisomes (EC 1.11.1.6), or by various peroxidases, such as glutathione peroxidase (GSH-Px) (EC 1.11.1.9) in cytoplasm (Eq. (14)), reactivity of H2O2 is rather low. H2O2 + H2O2 ^ 0 2 + 2H2O

(13)

2GSH+ H2O2 -^ GSSG + 2H2O

(14)

Thus, H2O2 or (62)^ may diffuse some distance from their sites of production. Consequently, radical generation by subcellular compartments may be a threat for the whole cell. H2O2 diffusion to reach nuclear DNA may be even enhanced by histidine (his) (Schubert and Wilmer, 1991). Oya and Yamamoto (1988) found that L-his enhanced the induction by H2O2 of chromosomal aberrations eight-fold in human embryonic fibroblasts and suggested involvement of a his-H202 adduct for transport of H2O2. It has been estimated that the steady-state concentrations in normal aerobic liver cells are IQ-i^M to lO-^ M (02)^ and lO'^M to 10-7 M Hp^^ respectively (Chance et al., 1979). Due to its high need for O2, the brain appears to produce (O2)" and H2O2 in similar amounts. If H2O2 is not detoxified by CAT or peroxidases, one-electron reduction results in the formation of the (HO*. This species is assumed to be the most toxic reactive oxygen radical, with an approximate intracellular half-life of 10"^ s (Pryor, 1986) so that reactions with biomolecules become diffusion-controlled. In a variety of biological phenomena, for example aging (Sohal and Allen, 1990; Masoro, 1991; Pacifici and Davies, 1991; Le Bel and Bondy, 1992), cancer (Sahu, 1991; Cerutti, 1991), diabetes (Wolff et al., 1991), phagocytosis and cataractogenesis (Halliwell and Gutteridge, 1989), ischemia-reperfusion injury (Braughler and Hall, 1989; Sussman and Bulkley, 1990; Agardh et al., 1991), quinone toxicity (Powis, 1989; Sinha and Mimnaugh, 1990; O'Brien, 1991), 6-hydroxydopamine (6-OHDA) toxicity (Kostrzewa, 1989) and radiation injury

(Girotti, 1990), the hydroxy radical is assumed to contribute to or to cause toxic processes. Irradiation of water leads to formation of (HO)'. By contrast, in the brain, strong water-soluble electron donors (DH) such as nicotinamide adenine dinucleotide phosphate (NADPH), catechin, hydroquinone, ascorbic acid or glutathione (L-yglutamyl-L-cysteinyl-glycine; GSH) can promote formation of (HO)' from H2O2 in the presence of Cu"^ or some iron complexes (e.g. Fe^+-adenosine diphosphate complexes) according to Eqs. (15) and (16) (Florence, 1984; Kadiiska et al., 1992). DH + Fe3-^(chelate) -^ Fe2+(chelate) + D' + H+ DH + Cu2+(chelate) -> Cu+(chelate) + D* + H+

(15)

H2O2 + Fe2+(chelate) -^ HO" + (HO)' + Fe3+(chelate) H2O2 + Cu+(chelate) -^ HO-(HO)' + Cu2+(chelate) (16) Reaction (16) is a Fenton-type reaction (Fenton, 1894; Croft et al., 1992). There is still controversy as to whether the reaction of Cu"*" and H2O2 leads to the formation of (HO)' or even a Cu^+ species (Bielski and Cabelli, 1991). As for iron-catalysed reactions, Fenton chemistry probably involves (oxygen)iron(2+) [FeO]^"^ intermediates, which are strong oxidants as well. The reaction of Fe3+(chelate) + H2O2 could yield (02)", but is reported to be rather slow (Halliwell and Gutteridge, 1989, 1992; Chiueh et al., 1992). In vivo formation of (HO)' is determined by measurement of hydroxylated salicylic acid (Van Steveninck et al., 1985) or nitrophenol (Florence, 1984). Involvement of oxygen radicals will often be detectable by means of electron spin resonance spectroscopy using spin traps, such as 5,5-dimethyl-l-pyrrolineA^-oxide (Buettner, 1987), if care is taken of possible pitfalls (Buettner, 1987; Makino et al., 1990; Davies et al., 1992). This method has been successfully used to determine that mainly low molecular weight complexes of iron catalyse formation of substantial amounts of (HO)' (Ozaki et al..

442

Oxidative stress: free radical production in neural degeneration

1988) in contrast to complexes of desferrioxamine, an iron chelator often used therapeutically (Desferal®) in experimental hepatic iron overload (Bacon and Britton, 1989). Thus, it is believed that iron-oxygen complexes, rather than free (HO)', are involved in initiating cytotoxic mechanisms, at least those involving stimulation of LPO by decomposing preexisting lipid peroxides or oxidation of proteins and nucleic acids (for effects of ROS on biomolecules, see Section 3.1.3). Singlet dioxygen. The singlet O2 can be generated by an input of energy (e.g. irradiation with light) to normal groundstate triplet O2 or can arise from H2O2 reacting with hypochlorite (CIO"), which can be formed by myeloperoxidase (EC 1.11.1.7) during phagocytosis (Khan et al., 1983). Exposure of cells in culture to high-intensity visible light causes damage especially to mitochondria, which are rich in haem proteins known to be potent sensitizers of singlet O2 formation (Srivastava et al., 1986). Since formation of singlet O2 is predominantly dependent on light or presence of ozone, and its relevance to oxygen-mediated tissue injury in the CNS is not yet elucidated, we shall focus on cell damage that possibly involves reactions of (O2)*, H2O2 and (HO)'-iron species with biomolecules. 3.1.2. Factors scavenging reactive oxygen species Enzymes. Two types of enzymes exist to remove H2O2 within cells (Fig. 2): the haem protein CAT, present in most aerobic cells, which catalyses the degradation of H2O2 to triplet O2 and water (Eq. (13)), and the haem- or selenocysteine-bearing peroxidases utilizing electron donors to reduce H202towater(Eq. (17)). DH2 + H2O2 ^ D + 2H2O

(17)

As mentioned in Section 3.1.1, H2O2 detoxification prevents formation of reactive oxidants like (HO)'. However, in contrast to liver and erythrocytes, which contain high levels of CAT (~ 1300 U/mg protein), the brain contains less than 20U/mg protein (Marklund et al., 1982). CAT is predominantly located in small peroxisomes

(microperoxisomes). In brain mitochondria, there is very little CAT activity (Sorgato and Sartorelli, 1974). Histochemical techniques have revealed that, in the CNS, at least two classes of microperoxisomes exist, mainly in astrocytes, and that Damino acid oxidase (which contributes to generation of H2O2) and CAT are not in the same peroxisome. Studies of the regional distribution of CAT (Brannan et al., 1981) revealed highest activity in hypothalamus and SN and lowest activity in striatum and frontal cortex. The distribution was reported to correspond to the localization of CAT to catecholaminergic nerve cell bodies (McKenna et al., 1976). The observation that ethanol can be metabolized to acetic aldehyde by H202-activated CAT ('compound F) points towards a role for peroxidatic activity of CAT in brain (Cohen, 1983b). It was further suggested that CAT may play a role in the metabolism of lipids (Masters and Holmes, 1977) and is very dominant in the liver. However, current evidence supports the view that, in certain cell types, GSH-Px is probably more important than CAT for protecting cells from H2O2 (Cohen and Hochstein, 1963; Nathan et al, 1980). The relative importance of GSH-Px in brain, compared with that of CAT, remains to be elucidated. Meister (1991) has reviewed the enzymology, metabolism and transport of GSH. GSH is an essential tripeptide present in virtually all animal cells. It is synthesized by the consecutive actions of y-glutamyl-cysteine synthetase and GSH synthase. The rate of the former enzyme is regulated through feedback inhibition by GSH. During action of GSH-transdehydrogenases and GSH-Px, glutathione disulfide (GSSG) is formed. GSH is regenerated via glutathione disulfide reductase (GSSG-Rd) utilizing NADPH resulting from nicotinamide adenine dinucleotide (NADH) by transdehydrogenation and, mainly, from glucose6-phosphate dehydrogenase, which is very specific for NADP+ and is regulated by intracellular contents of ATP, NADPH and ribose-5-phosphate. Due to its high nucleophilicity, GSH forms conjugates with endogenous compounds, such as estrogens and leukotriene A, or with xenobiotics and products of LPO (Spitz et al., 1991). These reactions are often catalysed by GSH transferases.

443

M.E. Gotz et al.

GSH-Px (EC 1.11.1.9) catalyses the reductive destruction of H2O2 (Eq. (17)) and organic hydroperoxides (ROOH), using GSH as an electron donor (Eq. (18)) (for a review on structure and function of peroxidases, see Spallholz and Boylan, 1991). ROOH + 2GSH -» GSSG + ROH + H2O

(18)

Oshino and Chance (1977) pointed out that in contrast to liver peroxisomes, H2O2 is destroyed in mitochondria and cytoplasm mainly by GSH-Px and not by CAT. GSH-Px activity of human brain amounts to approximately 70U/mg protein (Marklund et al., 1982). GSH-Px is a seleniumdependent enzyme (Flohe et al., 1973; Rotruck et al., 1973) and accounts for about one-fifth of total brain selenium (Prohaska and Ganther, 1976). In perfused rat brain, activities of GSH-Px and GSSG-Rd are highest in the striatum. GSH-Px, but not GSSG-Rd, activity was high in SN (Brannan et al, 1980a,b). This points towards a high need for peroxide-detoxifying enzymes in dopaminergic neurons. However, GSH-Px activity is more pronounced in glial cells than in neurons. Comparing primary cultures of murine astrocytes and neurons with respect to their content of total glutathione (GSH + GSSG), differentiated astrocytes contained about 16-fold higher levels (~ 16nmol/mg protein) than neurons (Raps et al., 1989). The overall concentration of GSH in rat brain is about 2mM. The ratio GSH/GSSG is roughly 10:1 or higher in favor of the reduced form (Cooper et al., 1980; Rehncrona et al., 1980). Since MAO is also predominantly localized in glial cells (see Section 4.1.1), deamination of catecholamines appears to be linked to GSH-Px content (Maker et al., 1981; Spina and Cohen, 1989). In addition to reaction with H2O2, the seleniumdependent GSH-Px is postulated to act together with phospholipase A, in converting potentially harmful phospholipid hydroperoxides (LOOH) to free fatty acid alcohols (ROH) via production of lysophospholipids and free fatty acid hydroperoxides (ROOH; Eq. (19)) (van Kuijk et al., 1987).

LOOH

PLA

"""^ ) lysophospholipid + ROOH

ROOH + 2GSH

^^""^^ ) ROH + GSSG + H2O (19)

Moreover, a specific phospholipid hydroperoxide GSH-Px has been described (Ursini et al., 1982), which reduces directly the hydroperoxide moiety of the still esterified fatty acid to phosphatidylglycerol without the necessity of phospholipase A2 activity. This prevents successive formation of prostanoids and lysolipids, which otherwise would affect cellular metabolism and destabilize membranes. GSH is located in a key position of cellular defense against free radical-mediated injury (Benzi et al., 1990, 1991). Protective potency against membrane protein oxidation (Section 3.1.3) (Reglinski et al., 1988), lipid oxidation (Section 3.1.3) (Thomas et al., 1990) and chelation of free haem (Shviro and Shaklai, 1987) has been ascribed to GSH. In addition, maintenance of the proper GSH/GSSG ratio (Miller et al., 1990) may be of significance in the metabolic regulation of the cell. Gilbert (1982) speculated that modulation of the thiol/disulfide ratio in vivo may serve as a 'third messenger' in response to cyclic adenosine monophosphate levels, and that the activity of key enzymes of glycolysis/gluconeogenesis may be regulated in response to changing thiol/disulfide ratios. However, since regeneration of GSH is dependent on NADPH, activity of glucose-6-phosphate dehydrogenase could be rate limiting for the activity of GSSG-Rd (Scott et al., 1991). Since GSH synthase is dependent on ATP, the overall pool of GSH is linked to oxidative phosphorylation, implying that impairment of mitochondrial respiration could lead to decreased synthesis of GSH (Section 3.2.1) (Meister, 1991). Thus, a proper balance of antioxidant enzyme activities and reducing equivalents (NADH, NADPH, GSH, ascorbate) is crucial for optimal cell function and resistance to oxidative stress. SOD (EC 1.15.1.1) are metalloenzymes that are widely distributed among oxygen-consuming organisms (yeasts, plants, animals). McCord and

444

Fridovich (1969) discovered (O2)* to be a substrate for a copper- and zinc-containing protein (Fig. 2), formerly known as *haemocuprein', in which copper is associated with enzymatic activity, while zinc serves as a stabilizer of protein structure. Interestingly, a manganese-dependent (MnSOD) (Keele et al., 1970) and an iron-dependent SOD were first characterized in Escherichia coli (FeSOD) (Yost and Fridovich, 1973). Localization of these enzymes is very different, indicating functional changes during the evolutionary history of SOD. MnSOD in eucaryotic cells is strictly a mitochondrial enzyme in the inner membrane and is synthesized by nuclear genes (Autor, 1982; Wisp et al. 1989). It resembles the FeSOD found in procaryotes, while cytosolic and peroxisomal CuZnSOD (Keller et al., 1991) are different from MnSOD with respect to amino acid sequence and secondary structure (Harris et al., 1980), supporting the idea for an endosymbiotic origin for mitochondria (Steinman and Hill, 1973; Beyer et al., 1991). In rat brain, SOD is homogeneously distributed (Thomas et al, 1976) with respect to brain region. In human grey matter, CuZnSOD amounts to 3.1 jug/mg protein almost equalling the amount of CuZnSOD in liver (Hartz et al., 1973). In brains of mice, total SOD activity was reported to be 408 U/mg protein (Grankvist et al., 1981), while in rat brain only 3 U/mg protein were measured (Peeters-Joris et al., 1975). The subcellular localization of brain SOD is highest in cytoplasm. Mitochondria and microsomes showed only 1315% of cytoplasmic levels. Glial cells from rat cortex contain higher specific activity of SOD than neurons. Several studies (Fridovich, 1986a) have demonstrated a direct toxicity of (02)* without invoking (HO)* produced by the metal catalyzed Haber-Weiss reaction (Eq. (20)). The direct reaction of (02)^ with H2O2 is very unlikely to proceed because of a reaction rate constant of 10"^ M~^ s~^ (Rigoetal., 1977). Fe(chelate)3+ + (O2)' <-> Fe(chelate)2+ + O2 Fe(chelate)2+ + H2O2 -^ Fe(chelate)3+ + HO" + HO' (20)

Oxidative stress: free radical production in neural degeneration

Evidence for toxicity of superoxide is exemplified by the following observations: bovine liver CAT is inactivated by (O2)" but protected by SOD (Kono and Fridovich, 1982); exposure of a purified GSH-Px to an enzymatic source of (O2)* and H2O2 causes inactivation of the enzyme, which is preventable by SOD, but not by CAT, suggesting, in each case, that superoxide-mediated cytotoxicity is not dependent on dismutation to H2O2 (Blum and Fridovich, 1985). SOD provide in vivo cellular protection by virtue of their ability to catalytically dismute (02)^ (Fridovich, 1975, 1986b). Failure of SOD could result in increased production of (O2)* during respiratory bursts of phagocytic leukocytes (Babior, 1982) and could aggravate inflammatory processes and reperfusion injury (McCord, 1987). Thus, biosynthesis of CAT, peroxidases and SOD have to be rigorously controlled to ensure protection. Mechanisms of regulation of enzyme synthesis in eucaryotes depend on many diverse factors, including age, organ, developmental stage, prevailing hormone profile and the availability of active site cofactors. However, just as our understanding of genetic regulation has been helped by studies of bacteria, studies of the bacterial response to H2O2 have given general insight into how cells defend themselves against deleterious oxidants. E. coli and Salmonella typhimurium respond to H2O2 with the induction of synthesis of over 30 proteins, including CAT, GSSG-Rd and alkylhydro-peroxidase. Sensing of the H2O2 is achieved by a DNA-binding protein, the OxyR protein. It is encoded by a genetic locus, the OxyR regulon, and assumed to be a redox-sensing protein capable of reversibly altering its conformation in response to the prevailing redox-environment of the cell. For eucaryotes, the regulatory mechanisms of antioxidant enzymes await elucidation. Since spontaneous or catalytic dismutation of (O2)* by SOD provides cells with H2O2 cellular response must not only elevate SOD activity to counteract (02)"' toxicity, but those of CAT and of GSH-Px as well. If SOD activity is increased selectively in brain of transgenic mice by introduction of the CuZnSOD gene, a significant increase of LPO (measured as malondialdehyde (MDA); Section 3.1.3) in the py-

445

M.E. Gotz et al.

ramidal cells of Amnion's horn and the granule cells of gyrus dentate can be detected (CeballosPicot et al., 1991). In contrast to GSH-Px, levels of CuZnSOD mRNA and protein, as well as susceptibility to LPO increase with age in mice (De Haan et al., 1992), suggesting involvement of ROS in aging, trisomy 21 (Down's syndrome) and possibly neurodegenerative diseases. Antioxidants. Excessive concentrations of ROS can have serious effects on membranes, nucleic acid bases and proteins (Section 3.1.3). If uncontrolled, mutations and membrane damage could lead to cell death. To minimize damage, defensive control systems exist. Besides enzymes, there are hydrophilic- and lipophilic-soluble molecules called 'antioxidants', scavenging free radicals to prevent destruction of cellular biomolecules crucial for cell viability. Non-enzymatic biological antioxidants include tocopherols, carotenoids, quinones, bilirubin, steroids, ascorbate, uric acid, GSH, cysteine and metal-binding proteins, such as ferritin (Krinsky, 1992). Iron-binding proteins (transferrin, haemosideroin and ferritin) (Halliwell and Gutteridge, 1990) remove iron from the cytosol so that it is no longer able to catalyse oxidation of biomolecules (Eq. (21)) through formation of alkoxylradicals (RO)' or (HO)*. ROOH + Fe2+ -^ (RO)' + HO" + Fe^^

(21)

The most important of these seems to be ferritin, which will be discussed later (Section 4.1). Due to their long, conjugated double-bond systems, carotenoids are excellent substrates for radical attack, thus scavenging singlet O2 or alkoxylradicals in membranes. Some of the reaction products have been described recently (Kennedy and Liebler, 1991; Handelman et al, 1991; Mordit et al., 1991), and the role of carotenoids in the risk of lung cancer, coronary heart disease and cataract has been discussed (Canfield et al., 1992; Rousseau et al., 1992). However, to date, there is no study on the role of carotenoids in neurodegeneration. The same is true for bilirubin and uric acid, the end-products of haem metabolism and of pu-

rine metabolism, respectively. In addition to the features of GSH discussed in Section 3.1.2, either GSH or cysteine can react directly with (HO)*, generating thiyl radicals (RS'; reaction rate constant >109 M~i s"0, which can also be formed when GSH is oxidized by peroxidases or by O2 in the presence of copper or iron (Rowley and Halliwell, 1982). Although less reactive than (HO)*, thiyl radicals also react with biomolecules, suggesting that thiol compounds are not ideal antioxidants. Coenzyme Q (Q) in its reduced form (ubiquinol) is known to inhibit LPO in subcellular membranes (Mellors and Tappel, 1966; Forsmak et al., 1991), either by reducing the a-tocopheroxyl radical (TO)* back to a-tocopherol (TOH) (Kagan et al., 1990) or by reacting directly with radicals. However, by far the most information exists, to date, for vitamin E and vitamin C (ascorbate). Vitamin E is the term used for eight naturally occurring fat-soluble nutrients (Fritsma, 1983). Four compounds bear a saturated phytyl side chain and differ only with respect to number and position of methyl groups at the chromanol ring (a-, ^-, y-, and ^-tocopherols). Four other compounds contain phytyl side chains with three double bonds (a-, )8-, y-, and 6-tocotrienols). However, TOH predominates in many species. The phytyl side chain in the 2-position facilitates incorporation and retention of TOH in biomembranes, while the active site of radical scavenging is the 6-hydroxyl group of the chromanol ring (Lucy, 1972; Burton and Ingold, 1981). Since eight stereoisomers exist, the name tocopherol should not be used without clarification of stereochemistry (Horwitt, 1991). However, /?,/?,/?-tocopherols are the only stereoisomers to occur in nature (Cohen et al., 1981; Slover and Thompson, 1981; Vecchi et al., 1990). The most widely accepted physiological function of TOH is its role as a scavenger of free radicals. Thus, it prevents oxidant injury to polyunsaturated fatty acids and thiol-rich proteins in cellular membranes and cytoskeleton. It is thought to preserve the structure and functional integrity of subcellular organelles (Chow, 1991). Each TOH molecule can react with two peroxyl radicals (Eqs. (22) and (23)).

446

Oxidative stress: free radical production in neural degeneration

TOH + R* -» (TO)* + RH

(22)

(TO)' + ROO' ^ ROO-TO adduct

(23)

(TO)' + DH -> TOH + D'

(24)

The first product is tlie (TO)', which is a resonance-stabilized, oxygen-centered radical. It can react with other peroxyl radicals to form stable adducts (Eq. (23)), some of which have already been isolated (Matsumoto et al., 1986), or can react with electron donors (e.g. ascorbate) to become re-reduced to TOH (Eq. (24)) (Bendich et al., 1984, 1986). The absorption, transport and metabolism of TOH in animals has been reviewed on several occasions (Bjomeboe et al., 1989; Drevon, 1991). TOH is transferred from circulating lipoproteins to the brain, spinal cord and peripheral nerves and muscle by unknown mechanisms (Sokol, 1989). There is no uniform distribution of TOH in the central and peripheral nervous system (Vatassery etal., 1984a). In contrast to other brain regions, the cerebellum is particularly active in the metabolism or utilization of TOH (Vatassery, 1987). During experimental TOH deficiency, nerve tissue retains a greater percentage of TOH than do serum, liver and adipose tissue (Goss-Sampson et al., 1988). Morphological and functional studies performed on experimental TOH-deficient rats have revealed axonal dystrophy and degeneration of peripheral nerve. This can be aggravated by increasing dietary polyunsaturated fatty acids providing increased quantities of peroxidizable substrate and reduced by feeding a synthetic antioxidant (ethoxyquin) (Southam et al., 1991). These experiments provide evidence in favour of an antioxidant role for TOH in the nervous system (Nelson, 1987). In brain, TOH is predominantly localized in the mitochondrial, microsomal and synaptosomal fractions (Vatassery et al., 1984b), suggesting that protection by TOH from peroxidative damage to subcellular membranes may be important for mitochondrial energy production or microsomal enzyme activity (Chou and Gairola, 1984).

Ascorbic acid is an extremely water-soluble antioxidant essential for humans, primates and guinea pigs, but not for rodents, which can synthesize it from glucose. Ascorbic acid serves as a cofactor in several iron-dependent hydroxylases (Padh, 1991) important for collagen synthesis, (prolyl- and lysyl-hydroxylases), for carnitine biosynthesis (6-A^-trimethyl-L-lysine-hydroxylase) and for catabolism of tyrosine (4-hydroxyphenylpyruvate-hydroxylase). Two major functions of ascorbate are support of the synthesis of norepinephrine and a-amidation of neurohormones, explaining in part its higher concentrations in brain and endocrine tissues (adrenal gland). The coppercontaining DA-^-hydroxylase (EC 1.14.17.1) catalyses the final step in the synthesis of norepinephrine (Eq. (25)), the hydroxylation of DA. DA + ascorbate + 0 2 ^ norepinephrine + semidehydroascorbate*

(25)

Ascorbate is most likely required by hydroxylases to maintain iron or copper at the active enzyme site in the reduced form, since it is necessary for hydroxylation. The semidehydroascorbate radical is not very reactive (Bielski and Richter, 1975; Rose, 1989). It decays by disproportionation to ascorbate and dehydroascorbate (the latter subsequently degrades to oxalic acid and L-threonic acid), rather than acting as a reactive free radical. Reaction of ascorbic acid with (OH)' is rapid and diffusion-dependent (K~1.2 x 10^-1.3 x 10^0 M-i s-i) (Cabelli and Bielski, 1983). (02)^ oxidizes ascorbic acid with a rate constant of 10"^10^ M-i s-i (Bielski et al., 1985). Besides direct scavenging of radicals, ascorbic acid is known to have a number of physiological effects (Padh, 1991), with a role in leukotriene biosynthesis (Schmidt et al., 1988), tetrahydrofolate reduction (Stone and Townsley, 1973), immunity (Anderson, 1984) and cancer (Wittes, 1985). Many membrane proteins are sensitive to tissue redox state (Levine, 1983), such as the NMDA receptor, thought to be involved in neuronal degeneration in seizure and ischemia (Choi, 1988a). It is inhibited by ascorbate, whereas reductants, such as dithiothreitol and

M.E. Gotzetal

447

penicillamine, which break protein disulfide bonds, potentiate receptor function (Majewska et al., 1990). The mechanism of this effect is not fully understood, but it must be important for survival of cells in cerebral ischemia. Furthermore, an important protective action of ascorbic acid is its ability to act synergistically with TOH in the inhibition of various oxidation reactions (McCay, 1985; Craw and Depew, 1985; Bendich et al., 1986; Burton and Ingold, 1986; Niki, 1987a,b). Packer et al. (1979) have shown in pulse radiolysis studies that in solution (TO)' reacts rapidly with ascorbic acid {K- 1.55 x 10^ M~^ s"0 to yield TOH again. This synergism seems to function in liposomal membranes as well (Scarpa et al., 1984; Doba et al., 1985; Niki et al., 1985). Some studies indicate that ascorbic acid helps to maintain tissue levels of TOH in vivo (Hruba et al., 1982; Bendich et al., 1984). However, by contrast, other in vivo studies found no evidence for an interaction between TOH and ascorbic acid (Yen et al., 1985; Burton et al., 1990). Thus, the interaction of TOH and ascorbate remains an open question for discussion. GSH was observed to protect against LPO in vitro (Reddy et al., 1982; Wefers and Sies, 1988; Graham et al., 1989), probably involving a GSHdependent heat labile factor(s) capable of reducing (TO)'. In addition, GSH is needed to reduce dehydroascorbate to ascorbate (Eq. (26)) by a dehydroascorbate reductsase, or even nonenzymatically (Winkler, 1992). Dehydroascorbate + 2GSH -^ GSSG + ascorbate (26) Moreover, an NADH-dependent semidehydroascorbate reductase is thought to be involved in the regeneration or restoration of ascorbate (Diliberto et a l , 1982; Chow, 1988). Under certain circumstances, ascorbic acid functions as a prooxidant rather than an antioxidant. Similarly to superoxide, ascorbate is able to reduce Fe^+ to Fe^"^, and in the presence of H2O2, it can promote (HO)' production. In vitro concentrations of ascorbate up to 0.2 mM can induce LPO in rat liver microsomes (Samuni et al., 1983; Shinar et al., 1983).

By contrast, at concentrations above 0.2 mM, it protects against LPO. Normal cytosolic concentrations would favour GSH over ascorbic acid as a cytosolic antioxidant in most tissues (McCay, 1985). However, if GSH is compromised in vivo by administration of its antimetabolite, Lbuthionine-(5,/?)-sulfoximine, ascorbate is utilized to protect against cell damage due to GSH deficiency (Martensson and Meister, 1991). If enzymatic or non-enzymatic antioxidants are inactivated or depleted, ROS can trigger various deleterious events, including oxidation of lipids, proteins and nucleic acid bases, as described in the next section. 3.1.3. Consequences of excess of reactive oxygen species Lipid peroxidation. One hypothesis to explain mechanisms of cellular aging and chronic progressive cell degeneration suggests the impairment of enzymatic and/or non-enzymatic anti-oxidant defence (Section 3.1.2), resulting in uncontrolled damage of biomolecules by ROS. In addition, presence of endogenous or exogenous toxins could affect cellular antioxidant defence systems. Thus, a common, but not necessarily primary, cause of oxidation of lipids, proteins and DNA could be an overflow and/or decreased detoxification of ROS. However, primary targets of ROS depend on sites of formation. Since compartmentalization is crucial for cell viability, severe damage to membrane structure could be an irreversible step towards cell death. Impairment of membrane function can be triggered either directly, by oxidation of polyunsaturated fatty acids of lipids (called LPO), or indirectly, by mechanisms leading to decreased lipid synthesis, decreased fatty acid desaturation, impaired redox equilibrium or increased activities of lipases. LPO involves the direct or metal-catalyzed reaction of oxygen and unsaturated fatty acids associated with polar lipids, generating free radical intermediates and semistable peroxides (Tappel, 1973). Since subcellular membranes in brain cells contain high amounts of polyunsaturated fatty acids, formation of a single carbon-centered radical within a membrane can lead to peroxidation of

448

many fatty acids. This can occur when O2 is present. The complex process of LPO is commonly described by three stages: (1) Initiation: the generation of a radical with sufficient reactivity to extract hydrogen atoms from methylene groups of fatty acids [(HO)'; (H02)-]; (2) Propagation: reaction of these radicals to yield another radical, which likewise is capable of generating more radicals (radical chain reaction); (3) Termination: recombination of two radicals or reactions yielding stabilized radicals no longer capable of propagating chain reactions. (HO2)' and (HO)', but not (02)% are able to extract hydrogen from allylic or bis-allylic positions of polyunsaturated fatty acids (Girotti, 1985; Kappus, 1985) The carbon radicals tend to be stabilized by molecular rearrangements to form conjugated dienes In the presence of sufficient amounts of O2, peroxyl radicals are formed (^^=10^\Qio yi-i g-i) i^ media of low hydrogen-donating capacity, the peroxyl radical is free to react further by competitive pathways, resulting in cyclic peroxides, double-bond isomerization or formation of dimers and oligomers (Gardner, 1989). Thus, random peroxidation of, for example, arachidonic acid could give a complex mixture of isomers of cyclic peroxides and hydroperoxides. If peroxidation of free fatty acids is driven enzymatically by cyclooxygenases or lipoxygenases, stereospecific hydroperoxides and endoperoxides are produced, which are precursors of eicosanoids (prostaglandins, thromboxanes, leukotrienes). If the peroxyl radical extracts a hydrogen atom from an adjacent fatty acid to yield another lipid radical (L'), which subsequently reacts with O2, a hydroperoxide (LOOH) chain reaction is propagated. Other peroxyl radical reactions are the )8-scission, intermolecular addition and self-combination. These reactions and those of phenols (e.g. TOH), aromatic amines and conjugated polyenes (e.g. 13carotene) with various radicals (carbon- and oxygen-centered) can terminate radical chain reactions. If LOOHs are not removed by GSHdependent peroxidases (see Section 3.1.2) transition metal ions, especially iron and copper, can

Oxidative stress: free radical production in neural degeneration

catalyse the decomposition of peroxides to form either alkoxyl (LO') alkyl (L') or (OH)^ radicals (Eqs. (27) and (28); K= 1.5x10^ M-^ s'O (Gamier-Suillerot et al., 1984) LOOH + Fe2+ -> Fe3+ + HO" + LO'

(27)

LOOH + Fe3+ -> Fe2+ + H+ + LOO'

(28)

These radicals could initiate a secondary propagation of radical chain reactions called LOOHdependent LPO (Bast and Haenen, 1984). Consequently, iron, or complexes of iron, with low molecular iron chelators stimulate LPO by lipid decomposition reactions (Gutteridge et al., 1984). Moreover, ferritin, an iron-storage protein holding 4500 mol of Fe^"^ per mol of protein, is able to stimulate LPO by releasing Fe^+ (Wills, 1966), and ascorbate enhances the rate of ferritin-stimulated LPO (Gutteridge et al., 1983). In contrast to popular belief, alkoxyl radicals of polyunsaturated fatty acids do not significantly abstract hydrogens, but rather, are channeled into epoxide formation through intramolecular rearrangement (Gardner, 1989). Moreover, besides homolytic reactions of polyunsaturated fatty acids, one has to keep in mind the susceptibility of hydroperoxides to heterolytic transformations, such as nucleophilic displacement and acid-catalysed rearrangement (Gardner, 1989). In 1990, Babbs and Steiner published a computational model of kinetics of LPO in a twocompartment model system (membrane and cytosol), assuming an iron-catalyzed, (O2)* driven Fenton reaction as the initiator of LPO (Eq. (20)) Kinetic interactions of up to 109 simultaneous enzymatic and free radical reactions thought to be involved in the initiation, propagation and termination of LPO were calculated using rate constants from the literature. From these model studies it was concluded that: ' 1. Segregation and concentration of lipids within membrane compartments promote chain propagation; 2. In the absence of antioxidants, computed concentrations of LOOH increase linearly at a rate of 40yaM/min during oxidative stress;

M.E. Gotz. et al.

3. LPO is critically dependent on O2 concentration and the modeled dependence is similar to the experimental function; 4. LPO is rapidly quenched by the presence of TOH-like antioxidants, SOD and CAT; 5. Only small (1 to 50JLLM) amounts of 'free' iron are required for initiation of LPO; 6. Substantial LPO occurs only when cellular defense mechanisms have been weakened or overcome by prolonged oxidative stress. Hence understanding of the balance between free radical generation and antioxidant defense systems is critical to the understanding and control of free radical reactions in biology and medicine.' Dependent on the fatty acid hydroperoxide (primary product of oxidation of unsaturated fatty acids with O2) and on catalytic degradation by either iron complexes or by NADPH cytochrome P450 reductase, a huge range of secondary products of LPO is formed. These include conjugated dienes (Corongiu et al., 1989), hydrocarbon gases (e.g. ethane, ethene from linoleic acid; Burk and Ludden, 1989) and carbonyl compounds (e.g. MDA, alkenals, alkadienals and a-)8-unsaturated aldehydes: Kaneko et al., 1987; Yoshino et al., 1991; Esterbauer et al., 1991). Carbonyl compounds are formed by ^-scission of alkoxyl radicals or thermic- or metal-catalysed degradation of cyclic endoperoxides. The latter process produces MDA. In addition, it is suggested that MDA can also be formed in vivo as a byproduct of eicosanoid biosynthesis (Hecker and Ullrich, 1989). Various techniques exist to evaluate products of LPO in tissues (Gutteridge and Halliwell, 1990; Hageman et al., 1992), but all are limited either with respect to sensitivity, specificity or practicability, since the most accurate assays for measuring lipid peroxides are the most chemically sophisticated, requiring sample preparation under inert gas to ensure no further peroxidation during handling of lipid material (e.g. gas-liquid chromatography/ mass spectrometry; Hughes et al., 1986). Measurement of MDA has been employed to detect and quantify LPO in a variety of chemical and biological matrices (Valenzuela, 1991), but there is increasing doubt of the specificity of MDA as a

449

quantitative indicator of in vivo preformed lipid peroxides (Choi and Yu, 1990; Janero, 1990). In order to release MDA from cyclic endoperoxides, elevated temperature is often applied. The MDA released is trapped by thiobarbituric acid to yield a pink pigment (Nair and Turner, 1984). This step is seldom done under inert gas conditions. The levels of pigments resulting from reaction of MDA and various other aldehydes with thiobarbituric acid (called thiobarbituric acid reactive substances, TEARS) are indicative of levels of TEARS originating from both preformed lipid peroxides in vivo and newly formed peroxides in vitro during incubation (Gotz et al., 1993). Since assay of TEARS is influenced by many experimental conditions (e.g. pH, temperature, O2 antioxidants, buffers, transition metals; for a review, see Janero, 1990), measurement of TEARS is not sufficient to give precise evaluations of LPO in pathophysiological states. At best, it can be an empirical indicator of the potential occurrence of peroxidative lipid injury in vivo and of the susceptibility of tissues to oxidative stress in vitro. Thus, whenever possible, a combination of methods measuring primary and secondary, as well as tertiary, products of LPO (amino acid adducts, nucleotide adducts and glutathionyl conjugates) is advisable. Excellent overviews concerning analytical aspects of monitoring oxidative stress in vivo are provided by Saran and Eors (1991), Hageman et al. (1992) and Pry or and Godber(1992) Oxidation of proteins. ROS can directly oxidize free or protein-bound amino acids, leading to deactivation of enzymes (Stadtman, 1990; Stadtman and Oliver, 1991; Stadtman and Eerlett, 1991). Cysteine, methionine, histidine and tryptophan are preferentially oxidized, resulting in sulfenic, sulfinic or sulfonic acids from thio-containing amino acids and in histidine- and tryptophanendoperoxides, which subsequently degrade (Sies, 1986). Oxidation of thiols in proteins is often involved in regulation of enzyme activity, such as glucose-6-phosphate dehydrogenase, pyruvate kinase, brain adenylate cyclase, y-glutamylsynthetase and others (Elstner, 1990) Carbonyl compounds can be attacked by amino groups. In-

450

crease of MDA in vivo could result in both intraand intermolecular cross links of proteins, giving fluorescent products (conjugated imines, R-N= CH-CH=CH-NH-R', fluorescence maximum at 470 nm with excitation maximum at 395 nm; Tappel, 1973). Interestingly, accumulating lipofuscin pigments in the aging brain and heart (Brunk and Ericsson, 1972; Mann et al., 1978; Brizzee and Ordy, 1979; Masoro, 1981) show characteristic fluorescence spectra similar to those of MDA cross-linked proteins. This may possibly result from interactions between ROS and autophagocytosis (Brunk et al., 1992). Histological and biochemical studies of lipofuscin have provided evidence that they contain lipid-protein adducts, which are extractable by mixtures of chloroform plus methanol (Davies, 1988). Besides lipids and proteins, lipofuscin contains a high concentration of metal ions, such as zinc, copper and iron. Lipofuscin-like fluorophores can result from reactions between oxidized ascorbic acid and glutamine (Yin and Brunk, 1991; Yin, 1992). Histological and ultrastructural evidence in hippocampal pyramidal and Purkinje neurons of rat brain indicates that lipofuscin probably originates from lysosomes (Masoro, 1981; Schlote and Boellaard, 1983) or mitochondria (Glees and Hasan, 1976; Brizzee and Ordy, 1979; Heinsen, 1979). Since lipofuscin deposition is promoted by very different factors, including inherited abnormalities of fat metabolism (e.g. in patients suffering from abetalipoproteinemia), administration of inhibitors of lysosomal proteases or feeding diets deficient in TOH or abnormally rich in polyunsaturated fatty acids (Halliwell and Gutteridge, 1989), it seems more likely that increase in pigments with age results from impairment of lysosomal functions (degradation of lipids and proteins) rather than oxidative stress outside the lysosomes (Stadtman, 1992; Youngman et al, 1992). Interestingly, degeneration of striatal tissue in HD is accompanied by a massive accumulation of the fluorescent pigment lipofuscin in the brain. However, the role of lipofuscin pigment in cellular aging is still unknown (Amenta et al., 1988). Proteins that have been oxidatively modified become excellent substrates for degradation by

Oxidative stress: free radical production in neural degeneration

proteases (Davies and Goldberg, 1987), probably because of concomitant denaturation and subsequent increase in their hydrophobicity (Pacifici et al., 1989). High molecular weight proteolytic complexes, called ingensin, macropain, macrosin, proteasome, multicatalytic protease or macroxyproteinase (Rivett, 1985; Pacifici et al., 1989), are assumed to be responsible for the degradation of oxidatively modified proteins, providing amino acids for de novo synthesis. Such modifications mark enzymes for degradation by proteases (Stadtman, 1992). a-)8-Unsaturated hydroxy-alkenals are far more toxic than is MDA (for a review, see Esterbauer et al., 1991). rraAi5-4-hydroxy-2-nonenal is the most prominent of these (Van Kuijk et al., 1986). It probably results from peroxidation of arachidonic acid (Pryor and Porter, 1990). It has been shown to inhibit protein synthesis and to interfere with growth of bacterial and animal cells in culture. Hydroxyalkenals are mainly detoxified by alcohol and aldehyde dehydrogenases, or by forming adducts with cysteine or GSH, the latter process being catalysed by GSH transferases (Witz, 1989; Spitz et al., 1991). In addition, adducts of trans-4hydroxy-2-nonenal with nucleosides have been identified (Hageman et al., 1992). This makes it clear that oxidative damage to lipids can affect proteins and DNA by secondary products of LPO. In addition, it is likely that a-)8-unsaturated aldehydes are potentially able to serve as cellular messengers interfering with signal transduction pathways. This hypothesis is supported by observations that hydroxyalkenals can stimulate oriented migration of neutrophils (chemotaxis; Curzio, 1988; Curzio et al., 1990) and phospholipase C activity (Rossi et al., 1990). Nucleic acid damage caused by reactive oxygen species. There is increasing interest in the potential role of ROS as mediators of metal-catalysed carcinogenesis (Klein et al., 1991; Kasprzak, 1991) and in genetic changes occurring as a consequence of ionizing radiation, chemical carcinogens and various other tumor promoters (e.g. phorbolesters; Frenkel, 1992). Besides ribonucleic acids, DNA is the most important factor damaged by ROS in vivo

451

M.E. Gotzetal

(Kasai et al., 1986; Adelman et al., 1988; Richter et al, 1988; Simic et al., 1989), resulting in the disruption of transcription, translation and DNA replication. The amount of oxidative damage, even under normal physiological conditions, may be quite extensive, with estimates as high as one base modification per 130 000 bases in nuclear DNA (Richter et al., 1988). Damage to mitochondrial DNA is estimated to be as much as one per 8000 bases (Richter, 1988, 1992). DNA-DNA and DNA-protein cross links, sister chromatid exchange, single- or double-strand breaks and base modifications are reported to occur due to reactions of ROS with DNA (Teebor et al., 1988; Simic et al., 1989). In principle, all four DNA bases can be oxidatively modified, thymidine being most susceptible to ROS. As for the reaction mechanisms, it is thought that H2O2 interacts with metal ions (Fe, Cu) on DNA and the sugar backbone, causing site-specific (HO)*-mediated DNA damage. For example, a strong correlation exists between the concentration of H2O2 in culture medium and the degree of strand breaks in human peripheral lymphocytes (Cochrane et al., 1987). In the presence of Fe^"^ (micromolar range), H2O2 concentrations of even lower than 100/^M can induce strand breaks in cultured cells, an effect that can be inhibited by iron chelators (phenanthroline) or CAT, but not by SOD, suggesting involvement of (HO)' or reactive ironoxygen species. The nucleosides thymidine glycol and 8-hydroxy-2'-deoxyguanosine (80HdG) are considered to be biomarkers of DNA damage by ROS (Simic, 1991; Hageman et al., 1992). These are specific since, in contrast to the free bases, they are not absorbed through the digestive system (Cathcart et al., 1984) and can be measured by HPLC in urine using electrochemical detection (Kasai and Nishimura, 1986; Floyd et al., 1986; Shigenaga and Ames, 1991; Halliwell and Dizdaroglu, 1992). In eucaryotes, several glycosylases, which act on DNA oxidation products, have been characterized, including a 3' repair diesterase in yeast (Johnson and Demple, 1988), a mammalian endonuclease specific for oxidatively modified DNA (Doetsch et al., 1986, 1987) and GSH transferases and peroxidases recognizing thymidine

hydroperoxide as a substrate (Johnson and Demple, 1988). Simic (1991) has pointed out that not only do exogenous factors, such as ionizing radiation or chemicals (bleomycin, adriamycin, benzo(a)pyrene), increase urinary levels of hydroxylated nucleosides, but high dietary caloric intake and high metabolic rates correlate with urinary thymidine glycol and 8-hydroxy-2'-deoxyguanosine excretion (Cathcart et al., 1984; Simic and Bergtold, 1991). It has been further documented that DNA repair is less efficient in older organisms (reviewed by Rao and Loeb, 1992). In contrast to the known inherited metabolic disorders, there is little evidence of DNA damage in relation to the pathophysiology of PD or AD (see Section 5.3). Of course, mutations in nuclear or mitochondrial DNA could be the ultimate cause of disturbed cellular metabolism leading to nerve cell death. On the other hand, chronic exposure of cells to ROS as a consequence of normal aging can be aggravated and accelerated by exogenously or endogenously produced toxins. This could be a cause of damage to biomolecules (Holmes et al., 1992; Harman, 1992). Since the steady-state level of oxidized biomolecules ultimately will depend on the efficiency with which they are removed, much effort is being made to quantify markers of oxidative damage to proteins, lipids and nucleic acids. The topics dealt with in this Section are briefly summarized in Fig. 3. 3.2. Impairment of energy metabolism 3.2.1. Defects in energy metabolism Defects in energy metabolism cause profound disturbances in the function of muscle or the brain. Such defects may be represented by myopathy, encephalopathy or encephalomyopathy, the latter concomitantly affecting both tissue types. In the postabsorptive state, the brain utilizes glucose predominantly, with regional variations in the metabolic rate, depending on the mental or motor task being performed (Sokoloff et al., 1977; Kennedy et al., 1978). Brain concentrations of glycogen are low (O.lg/lOOg fresh weight) (Sokoloff, 1989), and the role of fatty acids as oxidizable fuels for brain metabolism is considered to be negligible.

452

Oxidative stress: free radical production in neural degeneration

iunknown gene defect, inborn or acquired error of metabolism

normal aging of the CNS

decreased expression or translation of ROSdetoxicating enzymes or of subunits of mitoctiondrial respiratory chain

enzymatic overproduction of HjO; deficient defense mechanisms disturbance of iron homeostasis

aluminium accumulation 7 enzyme inhibitors (e.g. quinolines, carbolines) redox cycling xenobiotics autooxidation of catecholamines ?

MAO inhibitors ROS

decreased expression or translation of proteins of repair system

repair I detoxication mechanisms-. enzymes peroxidases. SOD. Kat. GSH-Px. GSSG-Rd lipid and water soluble antioxidant-^ a-tocophero! uric acid ubiquinol ascorbate ^-carotine glutathione GSH lazaroids Fe-chelators GSH-regeneration by Fe-binding proteins glucose-6-P-dehydrogenase GSH-synthesis

<;=

H202;(HO) ;(Qj)" if repair is insufficient

e.g. macroxYorotgiPagg (MOP) phosphQiipasgg redoxendonucleases

if

damage to proteins nucleic acids and lipids

fatty acid hydroperoxides

oxidated nucleotides

* oxidated bases 8-hydroxyguanine thymidine glycol

V oxidatively modified proteins

* conjugated dienes alkanes, epoxides aldehydes

detoxication bv enzvmes QSH-transferases mixed functional oxidases

Fig. 3. Putative pathogenetic causes assumed to contribute to ROS toxicity, consequently leading to damage to biomolecules, if repair mechanisms become insufficient. Putative sites of therapeutic intervention are indicated by double lines crossing arrows.

Thus, brain tissue is extremely sensitive to fluctuations in the blood glucose concentration, since no satisfactory endogenous substitute exists. Only in prolonged fasting are ketone bodies formed in liver (D-)8-hydroxybutyrate and acetoacetate), passively taken up by the brain from the bloodstream and utilized to produce acetyl-coenzyme A. Most diseases with impairment of enzymes of glycolysis and mitochondrial metabolism are inherited and result in serious malfunction of the nervous system. They can be classified into five major groups, resulting in defects of: (1) mitochondrial transport (2) substrate utilization (3) the Krebs cycle (4) oxidation-phosphorylation coupling (5) the mitochondrial respiratory chain. The clinical picture of deficiencies in glucose utilization can be subdivided into groups in which myopathy is the predominant manifestation or in which brain dysfunction predominates (cerebellar ataxia, pyramidal signs and dementia). In this review, however, we focus on the role of putative defects in energy metabolism as a possible factor in neurodegenerative diseases. In the majority of patients suffering from AD or PD, there seems to be no major genetic factor in its

etiology (see, however. Section 5.3). Nevertheless, defects in mitochondrial and microsomal function could be involved causatively in the pathogenetic process, possibly by their ability to produce (O2)* in vivo. 3.2.2. Microsomal production of superoxide The body is continually threatened by toxic substances, which are inhaled, absorbed or ingested. Thus, enzymatic detoxication systems have been developed for elimination of toxic compounds (Minn et al., 1991). For these purposes, and for the catalysis of oxidations of fatty acids and steroids, microsomal membranes (including membranes from lysosomes, peroxisomes and endoplasmic reticulum) possess a large number of enzymes (Jakoby and Ziegler, 1990). Microsomes contain two electron transport systems, one being dependent on NADH and consisting of NADH cytochrome h^ reductase and cytochrome h^ (needed for fatty acid acyl coenzyme A desaturase system), and the other involving NADPHdependent cytochrome P450 reductase (also referred to as NADPH cytochrome c reductase) and many isoenzymes (Nebert et al., 1989). Cytochrome P450 is involved in the oxidation of a wide range of substrates at the expense of O2 (known as

M.E. Gotz et al.

mono-oxygenation or mixed-function oxidation) and requiring a reducing agent (normally NADPH). The reaction stoichiometry is given in Eq. (29). 2NADPH + O2 + RH-^2NADP+ + H2O + R-OH (29) However, normally more NADPH is oxidized and more O2 is consumed than needed. The excess of O2 can lead to production of (O2)* (Gorsky et al., 1984; Zhukov and Archakov, 1982). (02)^ can be generated both from dissociation of the oxygenated complex of reduced cytochrome P450 (Sligar et al., 1974) and from the auto-oxidation of cytochrome P450 reductase containing FAD and FMN (Nakamura and Yamazaki, 1969; Aust et al., 1972). Like other flavoproteins, reaction of flavinsemiquinone with O2 generates (O2)". Although liver is the major organ involved in the P450mediated metabolism, this process has also been detected in the brain (Mesnil et al., 1984) of mice (Ravindranath and Anandatheerthavarada, 1989), rats (Warner et al., 1988) and humans (Bahmre et al., 1992). Immunocytochemical study of the rat brain P450 reductase using an antibody to the rat liver enzyme had revealed the presence of the enzyme in catecholaminergic neurons in SN, LC and the ventrolateral medullary region (Haglund et al., 1984). Spectral quantification has revealed that the level of cytochrome P450 in brain microsomes is approximately 40 pmol/g tissue, which is roughly 0.25% of that found in the liver microsomes of control rats (Warner et al., 1988), and 30120pmol/mg protein in human brain regions obtained at autopsy (Bahmre et al., 1992). Regional variation in NADPH cytochrome c reductase activity was immunohistochemically observed in human brain stem neuronal cell bodies (Ravindranath et al., 1990). Roles of P450 in the CNS may include metabolism of xenobiotics (Das et al., 1981), aromatization of androgens and estrogens by 2- and 4-hydroxylations (Reddy et al., 1974) and formation of catecholestrogens (Fishman and Norton, 1975; Paul et al., 1977). Their action in the CNS is linked to inhibition of catechol-0-methyltransferase (Ball et al., 1972) and inhibition of TH (Lloyd and Weisz, 1978). In

453

addition, it has been proposed (Sasame et al., 1977) that NADPH cytochrome P450 reductase could be involved in the formation of the catecholamine neurotoxin 6-OHDA (Section 4.1.3) or in potentiation of l-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced (MPTP) toxicity in mouse brain slices, in vitro (Pai and Ravindranath, 1991; Section 5.1). These results suggest that MAO is not the only enzyme producing neurotoxic metabolites of MPTP in brain. Moreover, a role for the cytochrome P450 enzymes in formation of carbolines (possible ligands for benzodiazepine receptors; Section 5.1) and of hallucinogenic indoleamine derivatives (A^,A^-dimethyltryptamine) has been suggested (Seth et al., 1990). It is important to learn more about the cytochrome P450 enzyme family in the CNS. This is because they have a putative role in the synthesis and metabolism of endogenous compounds and of xenobiotics, resulting in acute or chronic generation of toxic agents. Interestingly, Armstrong and colleagues (1992) reported that the risk of PD is more than twice as great for individuals with a P450 genetic polymorphism associated with deficient debrisoquine metabolism than in those without, implicating a genetic factor in pathogenesis of PD. 3.2.3, Superoxide production in mitochondria In brain, the oxidation of NADH and FADH2 produced in the Krebs cycle from various substrates in mitochondria is mediated by an electron transport chain consisting of flavoproteins [EFMN], non-haem iron-sulfur proteins [Fe^-Sn], iron- and copper-containing cytochromes (b562, t>566» Ci c, a, a3) and coenzyme Q (ubiquinone, oxidized form of Q; ubiquinol, reduced form of Q) located in the inner mitochondrial membrane of mitochondria (Jung and Brierley, 1983). The electron transfer chain can be resolved into four catalytically active complexes by fractionation with detergents and salt (De Pierre and Ernster, 1977; Fleischer and Packer, 1978). These are complex I, the NADH ubiquinone reductase containing [EFMN] and [FCm-Sn]; complex II, succinateubiquinone reductase; complex III, ubiquinone cytochrome c reductase containing cytochromes

454

b562» b566, Ci and iron-sulfur proteins; and complex IV, cytochrome c oxidase containing cytochromes a, a3 and copper (Fig. 4). Of the redox centers that have been implicated in electron transport, only Q and cytochrome c are not firmly associated with one of these complexes, and the four complexes, together with these two so-called mobile components, can be reconstituted to yield electron transport activity corresponding to that in the native membrane. Free radicals are formed during activity of the mitochondrial electron transfer chain (Boveris and Chance, 1973; Paraidathathu et al, 1992) and the rate of (O2)* formation is proportional to mitochondrial O2 utilization. Considerable amounts of (O2)* are produced when the electron flow is inhibited (antimycin or rotenone). There are two separate sites of (02)* production: the flavoprotein NADH dehydrogenase (located in complex I) and the ubiquinone cytochrome b segment (Boveris et al., 1976; Turrens and Boveris, 1980). Whether (O2)* formation is coupled with autooxidation of ubisemiquinone or with autooxidation of cytochrome b566 is still unclear, but the latter hypothesis is favoured (Nohl and Jordan, 1986; Beyer, 1990; Glinn et al., 1991; Nohl and Stolze, 1992). In contrast, ubiquinone has been shown to act as a potent protectant against free radical damage to subcellular membranes in vitro (Ernster et al., 1992). It is assumed that under 'normal' conditions little of the (62)^ formed escapes the mitochondria due to the high levels of MnSOD within the matrix (Section 3.1.2). However, during aging, decreased levels of GSH and cytochrome aa3 were measured in brains from old rats (Benzi et al., 1992), supporting the theory of increased oxidative stress due to (02)^ production of the respiratory chain as one of several causes of cell aging (Sohal and Sohal, 1991). Endogenous or exogenous inhibitors of the mitochondrial electron transfer chain could cause a continuous chronic oxidative stress to mitochondria, finally leading to cell death. Thus, it seems reasonable to assume that a decrease in enzymic activity in the electron transfer chain, due to a decreased formation of enzymes (Sections 5.2 and 5.3) or due to inhibitors, probably results in a chronic decrease in ATP levels and an increase in (O2)* formation.

Oxidative stress: free radical production in neural degeneration

3.3. Excitotoxin-induced cell death In addition to the acute effects of generalized forms of CNS trauma (e.g. hypoglycemia, hypoxia or ischemia), topologically restricted and cellselective damage occurs within hours or days following brain injury, mainly affecting pyramidal neurons of the hippocampus, neocortical and striatal neurons. To explain this regional pattern of neuropathology, the existence of selective vulnerable structures in brain has to be postulated. The affected targets are those neurons expressing postsynaptic receptors sensitive to EAA (NMDA-, quisqualate-, kainic acid-sensitive and metabotropic receptors; Farooqui and Horrocks, 1991) and, within the last decade, EAA (glutamate, aspartate) have been implicated as mediating damage to neurons and glial cells (Rothman, 1984; Collins, 1987; Choi, 1988a,b; Olney, 1990; Bridges et al., 1992). For example, pretreatment with EAAreceptor antagonists (Rothman, 1984; Simon et al., 1984) prevents the regional damage (Section 8). 3.3.1. Excitotoxicity Excessive and prolonged release of glutamate and/or aspartate from nerve terminals, their insufficient glial clearance from the extracellular space and decreased GABAergic postsynaptic input are prerequisites to regarding EAA as excitotoxic substances (Olney, 1990; see also chapter by Zorumski and Olney in this volume). Vulnerability of postsynaptic neurons to excitotoxin-mediated damage depends on the nature of the receptors, which can be stimulated by NMDA, kainic acid, quisqualate, ibotenate and quinolinic acid directly. In vitro studies indicate that excitotoxin-induced neuronal injury may involve acute swelling of cells due to the depolarization-mediated influx of sodium chloride, water and calcium (Ca^+) (Choi, 1988a,b; Rothman and Olney, 1986). Ca^^ is regarded as the triggering agent of many biological reactions and has attained the status of a second messenger (Berridge, 1975; Rasmussen and Goodman, 1977). It alters membrane stability and permeability (Seeman, 1972) and is involved in nerve impulse propagation by coupling the electrical signal to neurotransmitter release (Llinas and

455

M.E. Gotzet al.

proton gradient at the inner mitochondrial membrane NADPH a-ketoglutarate isocitrate pyruvate malate

uncoupling inhibitors (e.g. 2,4-dinitrophenol)

| F M N (Fe-S)|^^Q^^^yt.b-(Fe-S)-Cyt.Ci|^=D> Cyt.c^=^ |Cyt.a-Cu -=^ Cyt.ag-Cul 02 MPP rotenone MIQ* MBC*

|FAD (Fe-S)| A

2H2O

(O2)"

antimycin A amytal

CO CN" H2S

an adequate proton g r a d i e n t is n e e d e d for - ATP s y n t h e s i s - Ca** - t r a n s p o r t - phosphate-transport

complex I

complex III

complex IV

complex V

NADH ubiquinone oxidoreductase

ubiquinol c y t o c h r o m e c oxidoreductase

cytochrome c oxidase

synthase

ATP

complex II succinate ubiquinone oxidoreductase

Fig. 4. The pathway of electron transfer from various substrates of the Krebs cycle to water in the inner mitochondrial membrane, including the most relevant flavin- and cytochrome-containing enzyme complexes (I-V) with known and putative inhibitors. Potential sites of (O2)'' production are indicated. NADH-DH, NADH-dehydrogenase.

Nicholson, 1975). Ca^"^ regulates an enormous number of enzyme activities (Carvalho, 1982), including protein kinases, endonucleases, proteases and lipases The intracellular Ca^"^ concentration ([Ca^+li) has to be maintained at a low level of about 0.1 /^M (in contrast to extracellular levels of ca. 1 mM; Orrenius et al., 1989) by rigorously controlling Ca^+ entry, intracellular sequestration of Cd?-^ (endoplasmic reticulum, mitochondria) and binding to high affinity binding proteins, such as calmodulin, calretinin and calbindin-D28K (Billingsley et al., 1985; Baimbridge et al, 1992; Heizmann and Braun, 1992). When the energy state of a cell is normal, Ca^"^ is transported out of the cell via the NaVCa^+ antiporter and Ca^"^activated ATPase. However, in the case of an energy deficit (impairment of mitochondrial respiration, hypoxia, hypoglycemia), intracellular [Ca^+]i may rise many-fold due to channel-mediated influx or mobilization of Ca^+ from internal stores by 1,4,5-inositol triphosphate (IP3). IP3 and diacyl-

glycerol are formed during receptor- and Gprotein-coupled activation of phospholipase C. Diacylglycerol, in concert with Ca^^, activates protein kinase C, which catalyses the activation of many proteins (Barnard, 1992). In addition, phospholipase A2 is dependent on Ca^-^ and calmodulin and is suggested to participate in the detoxication of LOOH (Orrenius et al, 1989). However, by its releasing fatty acids (predominantly arachidonic acid) and membrane destabilizers, such as lysophospholipids (Zaleska and Wilson, 1989), the prostaglandins, leukotrienes and thromboxanes are formed, all of which are known to be mediators of inflammatory and allergic reactions. In addition, a high level of [Ca^+]i leads to activation of nonlysosomal proteases (e.g. calpains), which induce the conversion of xanthine dehydrogenase to xanthine oxidase and may help in the production of (O2)" (Dykens et al., 1987; Section 3.1). Cellular targets for these enzymes include cytoskeletal elements and integral membrane proteins (Kosower et

456

Oxidative stress: free radical production in neural degeneration

functional changes, collapse ^^^^^;:^^o\ membrane potential overstimulation of excitatory amino acid receptors

poly-ADP-ribosesynthetase activation

Ca sequestration by subcellular organelles or calcium binding proteins lysosomal enzyme release and further degradation of biomolecules digest of metalloproteins

endogenous or exogenous toxins

protective or destructive regulatory function

Fig. 5. Putative consequences of ROS, of excess intracellular Ca^"^ and impairment of mitochondrial oxidative phosphorylation to membrane structure and function. The putative sites of therapeutic intervention are indicated by double lines crossing the arrows.

al., 1983; Mirabelli et al., 1989; Melloni and Pontremoli 1989) leading to dissociation of actin microfilaments from anchoring proteins in the plasma membrane with subsequent membrane blebbing and increasing membrane permeability (Fig. 5). Finally, Ca^"*" may even activate endonucleases, catalysing nuclear DNA fragmentation, a process that could be involved in apoptosis, or programmed cell death (Nicotera et al., 1989; Fawthropetal., 1991). 33,2. Calcium and mitochondrial function ATP is essential for maintaining the normal voltage gradient across the cell membrane. Thus reduced ATP levels, following impairment of energy metabolism, depolarize cell membranes, thereby permitting intracellular accumulation of sodium. This can relieve the voltage-dependent magnesium block of NMDA channels and cause opening of voltage-dependent Ca^+ channels. Impairment of energy metabolism also prevents ATPdependent extrusion of Ca^"^ and the storage of excess [Ca^+Ji in endoplasmic reticulum and mitochondria by ATP-dependent mechanisms (Blaustein, 1988; Choi, 1988a; Siesjo and Bengtsson, 1989). In cultured neurons, inhibitors of oxidative phosphorylation or of the sodium-

potassium pump allow NMDA or glutamate to become neurotoxic (Novelli et al., 1988). Chemically induced hypoglycemia results in excitotoxic lesions, which can be prevented with NMDA antagonists (Zeevalk and Nicklas, 1991), but which are not accompanied by increased glutamate release. This suggests that ambient glutamate is sufficient to induce excitotoxic damage if intracellular energy metabolism is compromised (Sah et al., 1989; LoTurco et al, 1990; Real, 1992). Depletion of ATP from rat hepatocytes by treatment with potassium cyanide and iodoacetate leads to a sustained elevation of cytosolic free Ca^"*" preceded by depletion of GSH and loss of ATP. 5.5.3. Calcium and oxidative stress It is generally accepted that Ca^+ mobilization is crucial for the activation of phospholipases. However, Sevanian and coworkers (Sevanian et al., 1981; Sevanian and Kim, 1985) demonstrated that phospholipase A2 can also be activated in the absence of elevated Ca^"*" by the presence of peroxidized fatty acids in phospholipids. The degree of phospholipase activation was correlated with the extent of TEARS. Thus, both peroxidized fatty acids and Ca^^ can independently trigger degradation of membrane lipids, but may also act syner-

457

M.E. Gotz et al.

gistically. Certainly, whenever ROS are involved in membrane damage, Ca^"^ must be suspected as a participant. In Fig. 5, the consequences of excess of EAA and Ca^"*" as responses of neurons to various types of CNS trauma (Choi, 1988a,b; Baumgarten and Zimmermann, 1992) are briefly summarized. 3,4. Relationships between oxidative stress, impairment of energy metabolism and calcium cytotoxicity Impairment of mitochondrial ATP regeneration, as a consequence of electron transfer chain inhibition or of a decrease in activities of enzymes of energy metabolism, could enhance (O2)* production in mitochondria. This can also impair clearance of ROS due to less effective detoxification systems (loss of GSH), leading to increased influx of Ca^"^ and peroxidation of lipids and proteins (Fig. 5). Cells overstimulated by excitotoxic inputs or suffering from decreased levels of ATP react by taking up sodium and water, resulting in swelling. Subsequently, the cells are exposed to an increase of cytoplasmic free Ca^"^ via channel-mediated influx, mobilization of Ca^"^ from internal stores resulting from activation of second messengers and alterations in Ca^"^ clearance due to depletion of energy reserves or of ATP resynthesis. Increased Ca^"*" levels activate proteases, lipases and endonucleases, with subsequent degradation of phospholipids and production of prostaglandins known to involve production of ROS. In the final phase, cytoskeletal components and membranes are degraded. Prolonged oxidative stress episodes associated with depletion of the energy reserve may contribute, therefore, to neurodegeneration in a wide variety of pathological conditions (Fig. 5). Although many of the biochemical mechanisms of cell damage are well established, there is little information concerning the primary causes that trigger these mechanisms and induce the selective neurodegeneration in distinct brain regions. The clinical and neuropathological features of neurodegenerative diseases and the possible common mechanisms of neuronal cell death have been discussed. Next, we wish to focus on animal models

of neurodegeneration and on post mortem data from brains of patients who suffered from neurodegenerative disorders. 4. Relevance of oxidative stress to neurodegeneration 4.1. Factors favouring damage by reactive oxygen species 4.1.1. Hydrogen peroxide production by monoamine oxidase The enzyme MAO exists in two forms, termed A and B, in the mammalian brain. Both enzymes are flavoproteins localized in the outer mitochondrial membrane (Youdim et al., 1988). They play a key role in the metabolism of monoamine neurotransmitters and xenobiotic amines (Blaschko et al., 1937; Youdim et al., 1988). The A-form is mainly responsible for the deamination of serotonin and norepinephrine and is pharmacologically defined by its sensitivity to inhibition by clorgyline (Johnston, 1968; Fowler et al., 1980a). MAO-B, on the other hand, is known to deaminate predominantly non-polar amines, phenethylamine and methylhistamine (Glover et al., 1977). The B-form is characterized by its high sensitivity to inhibition by L-deprenyl (Knoll and Magyar, 1972; Riederer et al., 1978; Oreland et a l , 1983; Riederer and Jellinger, 1983; for reviews see Denney and Denney, 1985; Dostert et al., 1989; Youdim and Finberg, 1990; Gerlach et al., 1992). Both enzymes metabolize DA and tyramine. Eq. (12) (Section 3.1.1) depicts oxidation of primary amines by MAO, leading to the production of H2O2, aldehydes and ammonia. As for ammonia and aldehydes, detoxification enzymes, such as glutamine synthetase, aldehyde reductase and aldehyde dehydrogenase, seem to be upregulated if MAO activity increases (StrolinBenedetti et al., 1986), as could be observed in the whole brains of aging rats. However, activities of CAT, GSH-Px and GSSG-Rd responsible for detoxification of H2O2 did not change significantly in brains of old rats compared with young controls. Thus a greater sensitivity to oxidative damage arising from amine oxidation might be expected to

458

accompany aging or various pathological stages in which turnover by MAO is increased, in particular in combination with drugs known to elevate amine concentrations, such as L-DOPA or catecholaminergic uptake inhibitors (StroUn-Benedetti and Dostert, 1989). In human brain, MAO-B increases with increasing age (Robinson et al., 1971; Fowler et al., 1980b; Jossan et al., 1991a) due to an increase in MAO-B concentration. Moreover, this increase in MAO-B activity in brain is further accelerated in neurodegenerative disorders, such as senile dementia of Alzheimer type (Adolfsson et al., 1980; Jossan et al., 1991b), HD (Mann, J.J. et al, 1980b, 1986), ALS (Ekblom et al., 1992) and maybe PD (Schneider et al, 1981; Riederer and Jellinger, 1983). However, increase in MAO-B activity in PD could not be confirmed in a further study (Jellinger and Riederer, 1984), possibly because of treatment of patients with L-DOPA, which is now known to alter MAO-B activity (Mclntyre et al., 1985; LeWitt et al., 1985). The results from animal studies utilizing surgical (Oreland et al, 1983) and neurotoxic lesions (Francis et al., 1985; Jossan et al, 1989) and from investigation of ALS brains have supported the idea that the increase in MAO-B activity is a consequence of gliosis (Oreland et al., 1989). Astrocytes have been shown to be rich in MAO-B activity (Levitt et al., 1982) and astrocytosis has been demonstrated in senile brains (Schechter et al., 1981), and, interestingly, MAO-B activity was detected in astrocytes of senile plaques (Nakamura et al., 1990). The less consistent changes in MAO-A activities may reflect neuronal loss, but this is still uncertain. The human isoforms of MAO have been purified and cloned (Bach et al., 1988; Grimsby et al., 1991). MAO-A and -B are derived from different genes closely linked to each other and located on the short arm of the X-chromosome (Shih et al, 1990). Tissue specificity of MAO-A and -B activity and gene expression has been demonstrated (Stenstrom et al., 1987; Shih et al, 1990). In addition, within the same brain, there is a great difference in localization in different regions of the isoenzymes (Saura et al., 1992). Thus, the relative extents of MAO-A and MAO-B breakdown of DA varies markedly in different brain regions (Glover et al., 1980). The different promoter organization

Oxidative stress: free radical production in neural degeneration

of MAO-A and -B genes provides the basis for their different tissue- and cell-specific expression (Zhu et al., 1992). It is tempting to assume that the knowledge about regional differences in MAO activity could be a clue to the understanding of region-specific brain damage in neurodegeneration. However, immunohistochemical investigations (Glenner et al., 1957; Graham and Karnovsky, 1965; Levitt et al., 1982; Westlund et al., 1985, 1988; Aral et al., 1986; Thorpe et al., 1987; Konradi et al., 1988, 1989) and quantitative enzyme autoradiography (Saura et al., 1992) revealed that the cellular localization of the isoenzymes of MAO in both rat and human brain differs markedly and does not reflect the distribution of the presumed natural substrates (e.g. absence of MAO-B in melanin-containing neurons of the SN) (Konradi et al., 1989; Moll et al., 1990; absence of MAO-A in serotonergic neurons, Saura et al., 1992). In contrast, glial cells and melanin-free neurons contain both MAO-A and MAO-B in SN (Konradi et al., 1989; Moll et al., 1990). Thus, degeneration of melanin-containing neurons of SNC in PD cannot be directly attributed to increased levels of MAO-B. Perhaps MAO-A found in dopaminergic neurons metabolizing DA and producing H2O2 contributes to degeneration (Spina and Cohen, 1989) and/or increased MAO-B activity in glial cells leads to elevated production of ROS (secondary to increased metabolism of increased activity of surviving neurons). This could be followed by depletion of GSH stores subsequent to H2O2 production (Spina and Cohen, 1988; Sandri et al., 1990; Werner and Cohen, 1991) and by oxidation of catecholamines, leading to neuromelanin generation. However, there is no certain answer to this question, and it seems likely that, in addition to H2O2 production by MAO in glial cells, other factors have to be considered to be pathogenetic in the degradative process of catecholaminergic neurons. 4.1.2, Catecholaminergic toxicity and neuromelanin An increased DA turnover in PD (relative rise of acidic metabolites) as a consequence of an 8090% loss of nerve cells in the SNC and amplified

M.E. Gotz et al.

DA-liberation and reuptake in remaining axons of the striatum is a process that may contribute to the progressive loss of DA neurons in PD (Cohen, 1983a). An elevation in DA turnover (Mogi et al., 1988) may be a compensating mechanism in PD to overcome the effects of the loss of dopaminergic neurons (Hornykiewicz and Kisch, 1986). Although it has been shown that long-term administration of L-DOPA does not damage dopaminergic neurons in the mouse (Hefti et al, 1981), studies on two groups of patients with PD, matched for age and with one group which was treated with L-DOPA, provide some evidence for an increase of the striatal DA loss in the advanced decompensatory phase of the disease (Riederer and Wuketich, 1976). In C57B16 mice, DA synthesis is increased during aging to compensate for loss of dopaminergic neurons (Tatton et al., 1991). This could indicate that, in aging or pathological states, surviving neurons contain higher concentrations of catecholamines. It is known that DA is unstable in solution at neutral pH and easily undergoes autooxidation (Rodgers and Curzon, 1975; Graham et al., 1978). In the presence of transition metal complexes, catechols enhance the formation of (HO)' from H2O2 (Iwahashi et al., 1989), as examined by spin-trapping techniques. Thus, the presence of catechols in cells could provide a threat to cell viability, especially if low molecular weight iron complexes are present (Section 4.1.3) (BindoH et al., 1992). Interestingly, Mann and Yates (1983) showed that the more heavily pigmented neurons of the SN appear to be preferentially lost in PD and during the course of aging, when both iron and melanin are known to increase. When comparing the SN of control and parkinsonian brains, Hirsch et al. (1988) demonstrated the greater vulnerability of the population of DA neurons containing neuromelanin to the neurodegenerative process of PD. Their studies also showed a direct relationship between the distribution of pigmented neurons normally present and the distribution of cell loss in the SN of individuals dying with the disease. Recently, an inverse relationship was observed between the percentage of surviving neurons in PD compared with controls and the amount of neuromelanin they contain. Moreover, the largest pig-

459

mented neurons in SN are lost preferentially in PD (Moller, 1992), suggesting that the vulnerability of the dopaminergic neurons is related to their neuromelanin content (Kastner et al., 1992). However, this hypothesis was challenged recently by Gibb (1992). Neuromelanin is not found in the cerebral cortex, thalamus, strio-pallidal complex, cerebellum or spinal cord (Marsden, 1983). It accumulates during life in pigmented brain-stem nuclei, appearing first in cells of the LC around the time of birth and then in SN around the age of 18 months (Foley and Baxter, 1958; Mann and Yates, 1974). In normal subjects, the intracellular content of neuromelanin has been shown to increase with aging, up to 60 years. Then it begins to decrease, presumably due to destruction of melanincontaining cells. Many investigators have suggested that DA is a precursor of neuromelanin, a dark brown pigment, mainly located in the cell bodies of SN and LC (Van Woert et al., 1967; Nordgren et al., 1971; Das et al., 1978). In the skin, the formation of melanin is catalysed by tyrosinase, a bifunctional enzyme, oxygenating tyrosine to DOPA and oxidizing DOPA to DOPA-quinone (Lemer et al., 1949). However, this enzyme does not seem to be present in the SN (Barden, 1969). Thus, autooxidative mechanisms may play a primary role in neuromelanin formation. The exact chemical composition of this pigment is still unknown and debated. Since there seem to be structural similarities of lipofuscin and neuromelanin (Lillie and Yamada, 1960a,b; Van Woert et al., 1967), it is likely that neuromelanin results from the deposition on lysosomes of a melanin derived from the catecholamines DA or norepinephrine, or from deposition on lipofuscin of a melanin derived from lysosomes (Marsden, 1983). Incubation of DA with tissue homogenates results in the formation of cysteinylDA (via glutathionyl-DA) (Ito et al., 1986), and cysteinyl-catechols, such as cysteinyl-DA, cysteinyl-3,4-dihydroxyphenylacetic acid and cysteinylDOPA, are detected in brains of several mammalian species; the ratio of cysteinyl-DA/DA amounts to about 1:40 in human SN (Rosengren et al., 1985; Fomstedt et al., 1989). Most melanins of the skin are co-polymers of indole and cysteinyl

460

DOPA-derived eumelanin and pheomelanins, respectively (Prota et al., 1976). However, the presence of cysteinyl-DA in neuromelanin is still debated (Wakamatsu et al., 1991; Zecca et al., 1992). In contrast, it is believed that cysteinyl-DA or glutathionyl-DA is actually formed in the brain as a protective mechanism in neurons (Fornstedt et al., 1989; Fornstedt and Carlsson, 1991a) to absorb and excrete an overflow of catecholamines (Fornstedt and Carlsson, 1991b). The presence of indoles in neuromelanin has led to the assumption that, once thiols are depleted, neuromelanin formation is favoured (Carstam et al., 1991). Jellinger et al. ( 1992) and Good et al. (1992a) provided evidence for an iron-melanin complex in SNC neurons in parkinsonian brains, but not in LB or in non-melaninized cytoplasm of SNC neurons, and concluded that an iron-melanin interaction could significantly contribute to dopaminergic neurodegeneration in PD. Depending on conditions melanin can significantly increase or decrease the yield of reactive products of iron-catalysed decomposition of H2O2 in vitro as determined by spin trapping of the products (Pilas et al., 1988). They found that for low concentrations of ferrous ions, melanin decreased the yield of (OH)' due to binding of ferrous ions by melanin. In their experiments, ferrous ions bound to melanin did not decompose H2O2 efficiently. Melanin increased the rate of (OH)' production if the predominant form of iron was Fe^"^, presumably due to the ability of melanin to reduce Fe^"*" to Fe^"^. Thus, it is possibly not the presence of melanin per se, but the interaction between catechols, iron and H2O2 that determines the vulnerability of melaninized DA neurons to neurodegeneration in PD (Ben-Shachar and Youdim, 1990; Youdim et al., 1990; Ben-Shachar etal., 1991b) (Section 4.1.3). The participation of catecholamines in neurotoxicity seems very likely (Sandyk and Willis, 1992). However, the commonly accepted cellular marker of parkinsonian pathology, the LB (consisting of pathologically phosphorylated proteins, ubiquitin, phospholipids and sphingomyelin), is not confined to neurons containing neuromelanin and is not correlated with a selected neurotransmitter system. There are areas of the brain.

Oxidative stress: free radical production in neural

degeneration

of which the substantia innominata is an example that contain virtually no neuromelanin or catecholaminergic neurons but whose cells in PD patients are packed with LB and die (Marsden, 1983). Because there is no neuroanatomical evidence that regions like substantia innominata receive dense catecholaminergic innervation, it is hard to imagine that their degeneration in PD invokes a transsynaptic degeneration (Marsden, 1983). Unfortunately, to our knowledge, LB are not well characterized with respect to their chemical nature and origin. Perhaps there is a relationship to lipofuscin pigments and lysosomal disorders or breakdown of cytoskeleton, possibly due to excess Ca^"^, to yield inclusions similar to the NFT or amyloid plaques of AD. Indeed, it was shown recently by immunohistochemistry that LB in cortex and SN contain epitopes similar to the APP found in Alzheimer brains (Arai et al., 1992). So, is there a common cause of neurodegeneration despite the presence of morphologically and biochemically different neuropathological hallmarks? To give an answer, it seems crucial to define the exact compositions, possible modes of formation and features in common of LB, tangles, plaques, lipofuscin and neuromelanin. 4.1.3. 6-Hydroxydopamine model of neurodegeneration This review is concerned with endogenous factors putatively involved in the pathogenesis of various neurodegenerative diseases. However, research is inspired to a great extent by animal models. Functional relationships between experimental parameters are clearly more easily evaluated in animal models or cell culture than with post-mortem tissue. Thus, we at least want to mention some interesting topics dealing with ROS and a neurotoxin derived from catecholamines, 6OHDA (for reviews concerning neurotoxins see Kostrzewa, 1989; Calne, 1991; Herken and Hucho, 1992). 6-OHDA accumulates in catecholaminecontaining neurons and exhibits selective toxicity towards them. 6-OHDA is often administered intrastriatally to rats to induce degeneration of nigrostriatal neurons (Berger et al., 1991; Ichitani

M.E. Gotz et al.

et al., 1991). However, to be selective for the dopaminergic system, norepinephrine uptake has to be inhibited by desmethylimipramine (Kostrzewa, 1989). Two mechanisms for the toxicity of 6OHDA have been proposed. First, auto-oxidation could generate ROS and subsequently oxidize unsaturated fatty acids of lipids or thiol groups of proteins. Second, 6-OHDA uncouples mitochondrial oxidative phosphorylation (Wagner and Trendelenburg, 1971). Whether the neurotoxicity of 6-OHDA can be attributed to the production of ROS or dihydroxyindoles (for a review, see Thoenen and Tranzer, 1973) is not yet defined. Degeneration of nigrostriatal neurons after intracerebral injections of 6-OHDA to rats is potentiated by administration of iron (Ben-Shachar and Youdim, 1991) or after depletion of brain GSH by intracerebral administration of L-buthionine sulfoximine (Pileblad et al., 1989). In contrast, long-term oral administration of TOH (Cadet et al., 1989) or intraventricular injection of desferrioxamine (BenShachar et al., 1991b) attenuated the 6-OHDAinduced depletion of striatal dopamine. Although these experimental conditions are very harsh, the participation of ROS in the toxicity of 6-OHDA seems likely. These experiments could be a rational basis for the use of antioxidants or iron chelators in treatment of diseases suspected to involve ROS, if adverse effects can be minimized. 4.1.4. Iron distribution in brain and its role for oxidative stress As described in Section 3, iron can promote peroxidation of biological macromolecules due to its reactions with ROS and, thus, is of high toxic potential for cells, if it is not kept in a toxicologically inactivated form bound to specific proteins. Only when iron is tightly bound to a chelator is its capacity for promoting LPO minimal. Amongst synthetic chelators of iron, fo/^-(2-aminoethyl)amine-A^,A^,A^',A^'-penta-acetic acid, desferrioxamine, o-phenanthroline and bathophenanthroline are able to complex Fe^+ and, thus, slow down reduction of Fe^"^ to Fe^"^ by reductants like ascorbic acid or (O2)* in vitro, but EDTA is ineffective. Desferrioxamine was originally developed for the treatment of iron overload disease because it binds Fe^+

461

rather selectively, but there are current efforts to create more specific iron chelators that pass the blood-brain barrier (Section 8). The role of ferritin in iron-promoted LPO is ambiguous. Ferritin stimulates LPO proportionately to the amount of iron it contains, provided that mechanisms exist that release iron from ferritin. In liposomes, LPO by ferritin or haemosiderin, presumably a product of proteolytic attack on ferritin in lysosomes, is almost completely inhibited by desferrioxamine, suggesting that it is mediated by released iron ions (Halliwell and Gutteridge, 1989). Recently, it was argued that ferritin serves as a source of iron for oxidative damage (Reif, 1992) in the presence of redox-cycling xenobiotics, such as paraquat, adriamycin or alloxan, which cause (O2)* production (Minotti et al., 1991; Winterbourn et al, 1991). On the other hand, as long as iron is correctly bound to ferritin, it seems that it does not initiate peroxidation of biomolecules. Iron is an essential participant in many metabolic processes, including DNA, RNA and protein synthesis, the formation of myelin and the development of the neuronal dendritic tree, and as a cofactor of many haem and non-haem enzymes. A deficiency in iron metabolism, therefore, would be expected to alter some or all of these processes (Youdim, 1985; Youdim et al., 1991), and excessive accumulation of tissue iron may lead to oxidative stress via the formation of ROS (Halliwell, 1989a,b). Cytotoxicity of iron was confirmed by in vitro studies to cultured neurons (Tanaka et al., 1991; Michel et al., 1992) or by intranigral injection of iron into rats (Ben-Shachar and Youdim, 1991; Sengstock et al., 1992). Moreover, iron causes a time- and concentration-dependent opening of dihydropyridine sensitive Ca^"^ channels in rat cortical synaptosomes, resulting in parallel increased uptake of ^^Ca^"^ and stimulation of LPO, as measured by formation of TBARS in vitro from intact rat cortical synaptosomes. Both the Ca^+ uptake and LPO can be inhibited by dihydropyridines (nifedipine) and iron-chelating agents (desferrioxamine). The ability of iron salts to induce opening of Ca^^ channels, resulting in alteration of intracellular Ca^"^, would support the recent

462

hypothesis that iron could be the agent that induces neurotoxic events in iron-rich regions of the brain. Therefore, in order to be toxic to cells, iron has to be present in the brain in a more or less loosely bound form. This could mean that minimizing the amount of non-haem iron without depleting enzyme bound iron in biological systems is an important part of antioxidant defence 4.1.5. Uptake and distribution of iron and copper in normal and pathological brain There are two major problems associated with the biological use of iron, namely the poor solubility of Fe^"^ at physiological pH and the involvement of iron in potentially harmful redox reactions. These problems have led to the evolution of a variety of elegant biochemical processes, high affinity binding and cellular uptake of iron (Fatemi et al., 1991). Recent investigations (Morris et al., 1992a,b; Crowe and Morgan, 1992; Roberts et al, 1992) suggest that iron uptake into the brain does not involve the transcytotic pathway of transferrinbound iron via transferrin receptors into endothelial cells, but deposition of transferrin-bound iron within endothelial cells followed by recycling of apotransferrin to the circulation. The deposited iron is then delivered to brain-derived transferrin for extracellular transport within the brain and, subsequently, taken up as Fe^"*" via transferrin receptors on neurons and glial cells. Iron is used in several enzymes (Wrigglesworth and Baum, 1988), such as mitochondrial complexes of the respiratory chain, MAO, cytochrome P450, CAT, TH and others, or for storage in ferritin. Reaction of Cu"^ ions with H2O2 appears to generate (HO)' and reactive Cu^+ species in vitro. However, there is doubt whether copper is available to promote production of ROS in vivo. Copper transport seems not to involve transferrin (Thorstensen and Romslo, 1990) but Cu^^ complexes with histidine-imidazoles, a-amino groups of amino acids or nitrogen of peptide bonds of proteins, such as albumin or ceruloplasmin. Following binding, the Cu^"^ is reduced to Cu+, possibly by a reductase or perhaps by ascorbate, and then carried across the membrane into the cell (McArdle, 1992). It has been suggested that ceru-

Oxidative stress: free radical production in neural degeneration

loplasmin may be able to donate copper within cells for incorporation into copper proteins, such as CuZnSOD (Dameron and Harris, 1987a,b). Cerulo-plasmin obviously functions as an important scavenger of excess copper since low ceruloplasmin concentrations in the blood, as observed in Wilson's disease, lead to excess copper in various organs, including the brain, and concomitantly to lack of coordination, to tremors and to progressive mental retardation. If iron or copper were causally involved in neurodegenerative diseases, transition metal distribution in brain should ideally reflect neuropathological changes and perhaps explain the region-specific cell loss. Thus, many investigators were and are still concerned with the question of metal distribution in normal and pathological brain. In principle, it would be necessary to quantitate cellular and subcellular iron levels in pathologically affected brain regions. However, until recently, iron determinations were based on analytical approaches utilizing iron chelators for histochemical staining, spectrophotometric analysis or magnetic resonance brain imaging, which provide data only on regional bulk iron concentrations. Now, more sensitive methods are available, such as laser microprobe mass analysis (LAMMA) (Jellinger et al., 1990) and energy-dispersive radiographic microanalysis (Perl and Good, 1992), which allow identification and localization of cellular structures in histological sections and provide sensitive trace-elemental detection and characterization. Using the LAMMA technology. Good and colleagues (1992a,b) measured increased concentrations of aluminum (Al) and iron in the NFT within tangle-bearing and adjacent neurons in patients with AD. However, another microprobe analytical technique (nuclear microscopy) failed to demonstrate the presence of Al in plaque cores of chemically untreated tissue (Landsberg et al., 1992). It is obvious that each method has advantages and drawbacks, thus making it sometimes difficult to compare reported findings. Using histochemical techniques, the presence of iron in the brain was first detected at the end of the last century and, subsequently, was a subject of intense investigation (for a review, see Hill, 198^. Recently, a de-

M.E. Gotzetal

tailed study of the anatomical distribution of bulk iron in non-pathologic human post mortem brain was published; using Perl's and Tumbull's methods with the diaminobenzidine intensification procedure for the demonstration of non-haem Fe^"^ and Fe^"^, respectively (Morris et al., 1992a). This confirmed and extended the findings of earlier studies, showing highest levels of stainable iron in the extrapyramidal system (globus pallidus, SNR, red nucleus and myelinated fibres of the putamen). Moderate staining with Perl's technique was found in the thalamus, cerebellar cortex and SNC. Microscopically, the non-haem iron appears to be predominantly in glial cells as fine cytoplasmic granules. Neurons, in general, show low reactivity for iron, and this is difficult to discern, often because of the higher reactivity of the surrounding neuropile. In the globus pallidus and SNR, however, neurons with highly stainable iron content are found with granular cytoplasmic iron reactivity similar to that seen in the local glial cells. Although there seem to be no apparent correlations of iron staining with known transmitter systems, the extrapyramidal system is favored in iron uptake and storage. This could point towards involvement of iron in the pathogenesis of disorders involving striatonigral degenerations (Uitti et al., 1989), such as Hallervorden-Spatz disease (Dooling et al., 1974), MSA, progressive supranuclear palsy, ALS, HD and PD. Here, we want to focus on PD as the most frequently occurring disease. 4.1.6. Iron and neurode generation in Parkinson's disease Sofic and colleagues' (1988, 1991) biochemical studies of total iron, Fe^"^ and Fe^^, using spectrophotometry in different brain regions of patients with PD with and without AD, showed an increase of total iron in the SN of patients with PD versus patients with AD and control subjects. Fe^^ in the SN of patients with PD was nearly twice as high as in patients with AD and control subjects; in both the SNC and SNR it was increased by approximately one-third, whereas Fe^"^ levels showed no differences. In the cortex, hippocampus, putamen and globus pallidus, there

463

were no differences in the levels of total iron or Fe^+. The findings of increased total iron in the SN of patients with PD were confirmed by Dexter and colleagues (1989b, 1991, 1992a). Subcellular regions in SN of patients with PD and control subjects were investigated for iron by Jellinger et al. (1992) using transmission electron microscopy and energy dispersive radiographic microanalysis. Only the analysis of neuromelanin in SN neurons of patients with PD showed iron levels that were significantly greater than baseline control levels. No significant demonstration of iron accumulation was observed in the central core or the periphery of LB or in the cytoplasm and neuromelanin of SN neurons of control subjects. These results agree with previous histochemical findings that LB are consistently negative for Fe^"*" (Jellinger et al., 1990). However, they are at variance with the radiographic microanalysis data reported by Hirsch and colleagues (1991), who found higher iron concentrations in LB in SN neurons of patients with PD than in control subjects. An increase in total iron content in the SN seems not to be specific to PD, but is detected in other neurodegenerative diseases affecting the striato-nigral system, namely MSA and progressive supranuclear palsy. Total iron levels were also increased in striatal areas affected by the pathology of those diseases and of HD (Dexter et al., 1991, 1992a). Copper levels were reduced in the SN in PD and were elevated in the putamen of HD. The same authors found no consistent alterations in manganese levels in the basal ganglia in any of these diseases, but increased levels of zinc in SN, caudate nucleus and putamen in PD. Other studies have demonstrated increased iron levels in multiple sclerosis brain (Valberg et al., 1989) and accumulation of iron in the striatum of patients with ALS and AD (Olanow et al., 1989). The potential toxicity of the increased iron load in these disorders would be determined by the extent to which iron is deactivated by binding to ferritin and other moieties. Whether ferritin levels increase or decrease in PD is difficult to judge at this point, since both an increase and a decrease of SN ferritin in PD have been reported, using virtually the same methodology but different polyclonal antibodies to quantitate brain

464

ferritin (Riederer et al, 1989c; Dexter et al., 1990, 1992a). Taken all together, these findings suggest that increased iron levels are likely to be involved in neurodegenerative diseases affecting basal ganglia. However, it is still questionable whether iron acts as primary initiator of nerve cell death in PD or represents a secondary response to another yet unknown pathological cause. Nevertheless, iron in the SN may exist in a form capable of contributing to the toxic processes occurring in PD by stimulating formation of ROS. To date, changes in other metals are difficult to explain, Zinc and Mn are essential in the human diet and needed, for example, by SOD. Perhaps increased levels of Zn could be an expression of an answer by cells to increased levels of (O2)*. 4.1.7, Aluminum neurotoxicity It was the studies of Gutteridge et al. (1985) and Quinlan et al. (1988) that shed light on the possibility that Al could be involved in ROS-mediated cell damage. However, AP+ salts do not themselves stimulate LPO. In the presence of Fe^+ salts {100 juM), Quinlan and colleagues (1988) observed a 3-fold increase of TEARS induced by AP+ salts (300 juM) in rat Hver microsomes. Accumulation of Al and iron not efficiently bound to storage proteins, therefore, could provide a risk factor for brain cells. In contrast to a study by Fleming and Joshi (1987), who found a 5-fold increase of Al content in brain ferritin in AD vs controls, Dedman and colleagues (1992) observed no difference in ferritin isolated from the cerebral cortex of AD patients vs controls. In a further study, Fleming and Joshi (1991) demonstrated a concentration-dependent decrease in the initial rate of iron loading into human brain ferritin in the presence of Al, suggesting that both Al and iron can be stored in ferritin. In contrast, Dedman et al. (1992) showed that brain ferritin from chronic renal-dialysis patients had less than nine atoms of Al per ferritin molecule, despite markedly increased concentrations of Al in the cerebral cortex in these patients. These authors suggested that Al does not accumulate in ferritin in vivo. Some authors have reported that Al is associated with various lesions

Oxidative stress: free radical production in neural degeneration

of Alzheimer's brain-ipofuscin granules in the abnormal processes of some senile plaques and in the cytoplasm of neurons (Duckett and Galle, 1976; Duckett et al., 1985), nuclei of tangle-bearing neurons (Perl and Brody, 1980; Crapper et al., 1980) and amyloid cores of senile plaques and NFT (Candy et al., 1986). On the contrary, McDermott et al. (1979), Markesberry et al. (1981), Stern et al. (1986), Chafi et al. (1991) and Landsberg et al. (1992) did not find prominent concentrations of Al in the brain of Alzheimer patients. Likewise, Dedman et al. (1992) found no increase in bulk Al in the parietal cortex in AD, but an increase in ferritin (38%) and the non-haem iron content (45%) predominantly located in microglial cells (Kaneko et al., 1989) associated with senile plaques (Grundke-Iqbal et al., 1990). Moreover, Connor et al. (1992a,b) demonstrated intense ferritin immunoreactivity in senile plaques and blood vessels of brains from AD patients, suggesting a disruption of brain iron homeostasis. Since Al has also been found in the senile plaques and tangles of the functionally normal elderly (Perl and Brody, 1980; Candy et al., 1986), there are doubts whether Al may be involved in neuronal degeneration and dementia. To date, there is only sparse experimental evidence for a role of Al in ironinduced oxidant stress in vivo. Nevertheless, there is evidence from animal experiments and from cell culture systems that Al affects many biochemical and neurochemical metabolic events (for reviews, see Crapper McLachlan et al., 1991; Van der Voet et al., 1991; Mera, 1991). The experimental administration of Al or its salts by intracerebral or subcutaneous injection results in encephalopathies and in the production of Al-containing NFT (Wisniewski et al., 1980; Trancoso et al., 1982), due to accumulation of neurofilaments in the cell body and processes (axon, dendrites) of large neurons. However, the individual fibrils making up NFT in AD appear as PHF and are ultrastructurally different from the normal neurofilaments and those induced by Al (Munoz-Garcia et al., 1986). On the other hand, AD-type tangles share determinants with normal and Al-induced neurofilaments and also appear to contain MAPs, tau and MAP-2 (Langui et al..

M.E.

Gotzetal

1988). Although Shigematsu and McGeer (1992a) reported an accumulation of APP in damaged neuronal processes and microglia following intracerebral administration of Al salts, the authors state that this would not be specific to Al-induced pathology, but is rather a general response to disturbance of axoplasmic flow, regardless of the causative factors. They conclude that the value of models, such as the Al model of neuroskeletal toxicity, would be in revealing that APP accumulation can be secondary to interruption of axoplasmic transport from any cause, for example toxins, such as colchicine (Shigematsu and McGeer, 1992b), or impairment of mitochondrial energy production. Thus, there has to be a primary, maybe longlasting, event that directly or indirectly precedes neuroskeletal degenerative changes. Epidemiologic studies linking AD to Al concentrations in water supply remain a matter of controversy (Reynolds et al., 1992; Whalley et al., 1992). There are far more important sources of Al in the diet (Davenport and Goodall, 1992). Thus, there is much to do in the future to elicit the real risk of Al for neurodegenerative diseases. Although Al may not be causative for the development of neurofibrillary pathology in AD, its contribution cannot be so far discounted (Shea et al., 1992). 4.2. Factors protecting cells from oxidative stress 4.2.1. Detoxifying enzymes In order to prevent oxidative damage to DNA, proteins and lipids, cells are equipped with different antioxidative enzymes (Section 3.1.2). There is some evidence that these enzymes are altered in PD, whereas for AD very little information exists. In 1975, Ambani and colleagues found in PD patients that the non-GSH-dependent peroxidase was decreased in homogenates from SN, caudate and putamen (reduction ~ 50% of controls), but not in other brain regions, Kish et al. (1985) observed a lower, but significant, decrease of GSH-Px in frontal cortex, putamen, globus pallidus externus and SN (reduction -20% of controls). However, this could not be verified by Marttila et al. (1988). Recently, GSH-Px was reported to be reduced in erythrocytes in PD. This reduction was correlated

465

with the duration of the disease, but not with the age of patients (Johannsen et al., 1991). In contrast, serum levels of GSH-Px and SOD seem to be increased in PD (Kalra et al., 1992). Moreover, GSH-Px activity and levels of the vitamins E, C and A are reported to be decreased in erythrocytes of patients with AD (Jeandel et al., 1989). Since GSH-transport seems not to be affected in PD (GSH-transferase activity is not altered in SN in PD) (Perry and Yong, 1986), decreases in peroxidase and GSH-Px activity would imply a possible increase in susceptibility to oxidative stress of some brain regions in PD. The other important antioxidative enzyme, CAT, is in rat brain, predominantly located in microperoxisomes of catecholaminergic neurons and oligodendrocytes (McKenna et al., 1976). Again, Ambani et al. (1975) detected significant reductions in CAT activity in PD in SN and putamen, but Marttila et al. (1988) could not confirm this. It is evident that the extent of damage would be increased if H202-degrading enzyme activities are decreased or if production of H2O2 is enhanced. As outhned in Section 3.1, SOD generates H2O2. Interestingly, within the SNC, CuZnSOD gene is preferentially expressed in the neuromelaninpigmented neurons (Ceballos et al, 1990). Analyses of SOD activity in homogenates of PD patients have provided conflicting results. Whereas Marttila et al. (1988) reported increased CuZnSOD in temporal cortex, nucleus ruber, thalamus, SN and NBM, but not the caudate nucleus or putamen, Saggu et al. (1989) found an increase only in mitochondrial-MnSOD in SN, but not in cerebellum. The altered activity of MnSOD may be an adaptive increase due to excess formation of (02)"" from the mitochondrial respiratory chain or various enzymes, such as xanthine oxidase (Jenner et al., 1992). In AD patients, Ceballos et al. (1991), using immunohistochemical methods, detected a high level of CuZnSOD protein in large pyramidal neurons of the hippocampus, which are known to be susceptible to degenerative processes in AD. They argued that biochemical pathways leading to (O2)* generation were specially active in these neurons, requiring an active transcription of the CuZnSOD

466

gene. Alternatively, a high cellular CuZnSOD activity might also, by promoting H2O2 production, contribute to the vulnerability of these neurons, in particular within compartments low in GSH-Px or CAT activity (Hirsch, 1992). 4.2.2. Antioxidants GSH-Px needs GSH as a substrate, and many attempts have been made to evaluate the levels of GSH and GSSG in PD. Post mortem changes can dramatically affect measurements of GSH (Reed et al., 1980), with loss of GSH without concomitant increase of GSSG, indicating that peptidase activity could play a role in brain GSH levels. However, careful attention to selection of samples with respect to post mortem delay, sex and age should overcome these problems. In SN, there is a 4050% reduction of GSH (Riederer et al., 1989c; Jenner et al., 1992; Sofic et al, 1992). By contrast, no differences in GSH levels were observed in other brain regions in the same studies. There were no changes in the levels of GSSG, in agreement with the normal levels of GSSG-Rd found in PD (Marttila et al., 1988). In addition, GSHtransferase was reported to be unchanged in postmortem tissue (Perry and Yong, 1986). Thus, the decrease of GSH levels can only be explained by impairment of GSH synthesis (depletion of ATP, enzyme inhibition, decreased transcription or translation rates) or by oxidative degradation, maybe via thiyl radicals. To date, no data are available concerning activity of the rate-limiting synthetic enzyme for GSH formation in SN (yglutamyl cysteine synthetase). A decrease in levels of GSH and of GSH-Px activity could provide a source for H2O2 accumulating in cells. Thus, GSH concentration could be a key factor determining the fate of a cell at the threshold between life and death. It has been argued (Uhlig and Wendel, 1992) that, in order to be really a cause of cell death, GSH levels would have to be decreased to about 10% of that value existing in healthy cells (1 jumoyg GSH, 10 nmol/g GSSG) (Slivka et al., 1987b). GSH depletion to a lesser extent, however, renders cells more susceptible to impairments of cellular metabolism. Interestingly, the levels of GSH are decreased to the same extent in incidental

Oxidative stress: free radical production in neural degeneration

LB disease, considered to reflect early presymptomatic stages of PD (Gibb and Lees, 1988), as in advanced PD, despite a far less intensive neuronal loss in the SN. This could be an indication that impairment of GSH/GSSG equilibrium is an early event in neuronal degeneration. The importance of GSH for mental function is underlined by the observation that patients with GSH synthetase deficiency showed a gradual neurological deterioration of motor functions, retardation of movement, tremor and rigidity and psychomotor retardation beginning in childhood (Jellum et al., 1983). In addition to its antioxidant role, GSH recently was proposed to function as a neuroactive peptide in the CNS (Guo and Shaw 1992). Binding sites for GSH, possibly coupled to inositol phosphate production, were identified in cell membranes of astrocytes and oligodendrocytes. Since GSH is predominantly localized in non-neuronal cells (Slivka et al., 1987a; Raps et al., 1989; Philbert et al., 1991), the severe loss of GSH in SN in PD has to be, at least in part, attributed to impairment of glial functions or to extensive neuronal loss. The latter seems unlikely because, in LB disease, cell loss was very moderate, but GSH depletion was nearly 40% (Jenner et al., 1992). Interestingly, patients with AD and AD plus PD exhibited increased levels of GSH in the hippocampus compared with controls (Adams et al., 1991), and TOH content was doubled in the midbrain of both groups of patients. This raises the question whether these increases result from reactive gliosis or nerve terminal proliferation in response to neuronal loss. Whether this is specific for AD is not known but, as in PD, no elevated GSH levels were reported, despite gliosis. In PD, there were no alterations in levels of TOH in serum (Femandez-Calle et al, 1992) nor in various brain regions, including SN, when compared with control subjects (Dexter et al., 1992b). In addition, brain levels of ascorbate are not altered in PD (Riederer et al., 1989c). In conclusion, decreased activities of GSH-Px and CAT, as well as decreased GSH levels, increased activity of SOD and elevated levels of non-ferritin-bound iron concomitant with a high turnover of catecholamines, may participate in

M.E. Gotzetal.

production of ROS. All these factors may render SN cells in PD more susceptible to hereto undefined toxic noxae and may provoke LPO and/or, as a consequence, lead to increased levels of intracellular Ca^+, with activation of proteases, lipases and endonucleases. As for AD, such a story seems less conclusive since, to date, investigations into the free radical theory of AD are scarce. Nevertheless, possible roles of Al or EAA in the pathogenesis of AD are currently under debate, and an impairment of Ca^"^ homeostasis in AD brains, affecting membrane integrity and cytoskeleton, has been postulated. 4,3. Possible consequences of oxidative stress in the central nervous system If there is a real increase in H2O2 and (O2)* production in PD or AD markers of LPO, hydroxylated nucleotides or signs of protein oxidation should be found in post mortem tissue, at least in those brain regions that undergo the pathogenetic process of degeneration. To date, there is very Httle information available on markers of oxidative stress in neurodegeneration. For PD, only two parameters were studied, namely the substrate for LPO, polyunsaturated fatty acids, and the content of TEARS. Levels of polyunsaturated fatty acids of parkinsonian SN were decreased compared with controls (Dexter et al., 1989a). Thus it appeared that perhaps increased degradation of polyunsaturated fatty acids did occur in parkinsonian SN. This view was confirmed by the finding of a selectively increased level of TEARS in SN of parkinsonian patients. To our knowledge, there is currently no information available concerning oxidatively damaged proteins or nucleotides in parkinsonian SN. There is, however, an indication of increased protein oxidation in aging (Stadtman, 1992) and increased vulnerability of frontal cortex of AD brains to age-related protein oxidation (Smith et al., 1991). In addition, Zemlan et al. (1989) suggested that hydroxy-proline residues found in amyloid deposits of Alzheimer brain arose from free radical-induced oxidation of proline. Furthermore, a selective increase in susceptibility to iron/oxygen- or oxygen-induced for-

467

mation of TEARS in frontal cortex of AD brains could be detected by Subbarao et al. (1990) and Gotz et al. (1992), respectively. This could be indicative of defective defense systems against ROS. Fatty acid peroxides become rapidly deacylated by phospholipase A2. The resulting lysophospholipid leads to labilization of the lipid membrane, if it is not removed or reacylated. Levels of glycerophospholipids, plasmalogens and polyphosphoinositides are markedly decreased in patients with AD compared with age-matched control subjects (Suzuki et al., 1965; Stokes and Hawthorne, 1987; Farooqui et al., 1988a; Gottfries, 1990; Nitsch et al., 1992; Soderberg et al., 1992; Jellinger et al., 1993). This decrease in glycerophospholipids is correlated with elevations of phospholipid degradation metabolites, such as glycerophosphocholine, phosphocholine and phosphoethanolamine, in autopsy samples of AD patients (Earany et al., 1985; Miatto et al., 1986; Pettegrew et al., 1988; Elusztajn et al., 1990). These changes may be associated with elevated activities of lipolytic enzymes in AD (Farooqui et al., 1988b, 1990). As a result of deacylation of fatty acids from lipids, increases in levels of prostaglandins may occur in AD (Iwamoto et al., 1989). Fatty acid composition seems to be altered only in selected phospholipids of frontal grey matter and in hippocampus, with a substantial increase in the relative amounts of the saturated components 14:0, 16:0 and 18:0 paralleled by a decrease in polyunsaturated fatty acids 20:4, 22:4 and 22:6 (Soderberg et al., 1991). In contrast, analysis of fatty acid composition of total lipid fraction from grey or white matter of the frontal cortex of AD brains and occipital lobes (Antuono et al., 1991) showed no differences from controls. This could mean that the free fatty acid pool increases with an increase in eicosanoid production. Eesides elevated formation of lipid peroxides, impairment of phospholipid synthesis, of the microsomal desaturase system (Strittmatter et al., 1974, Jeffcoat, 1979) or of deficiency in uptake or reacylation of fatty acids could account for the above-mentioned changes in lipids (Dhopeshwarkar and Mead, 1973). Thus, the changes in fatty acid composition of selected lipids (phosphatidylcholine and phosphatidylethanola-

468

mine, predominantly located in subcellular membranes) (Soderberg et al., 1991) are very unspecific in nature, albeit consistent with oxidative stress. There may be also a link between EAA and abnormal phospholipid metaboUsm in AD (Section 3.3). Since no correlations have been found between levels of phosphatidylcholine and choline acetyltransferase activity in AD frontal cortex, the changes in membrane phospholipids may not be confined to cholinergic terminals. This supports the hypothesis of a defective biosynthesis or stimulated degradation of phospholipids in AD, pointing towards a generalized defect preceding the formation of neuritic plaques. Since amyloid formation in AD brains requires abnormal processing of the APP (Muller-Hill and Beyreuther, 1989; Esch et al., 1990; Sisodia et al., 1990; Katzman and Saitoh, 1991; Selkoe, 1991), defective membrane metabolism could expose the APP transmembrane domain to proteolytic cleavage, enhanced by increased Ca^+ influx; alternatively, amyloidogenic APP fragments may be poorly anchored to defective membranes and thus, released into the neuropile (Nitsch et al., 1992). Therefore, taking into account the existing knowledge concerning the etiology of AD and PD, further effort is needed to define the physiological importance of the accumulation of iron or Al in specific brain regions, of antioxidant systems and of alterations in membrane composition, in order to confirm or to rule out a role for oxidative stress as a pathogenetic factor in neurodegeneration. 5. Impairment of mitochondrial function 5.7. Tetrahydropyridines, tetrahydroisoquinolines and tetrahydro-/S-carbolineSy neurotoxins producing a parkinson-like syndrome in animals and humans The neurotoxic substance MPTP produces a clinical syndrome strikingly similar to idiopathic PD in humans (Langston et al, 1983), and induces nigrostriatal cell loss with concomitant decrease in DA and its metabolites, 3,4-dihydroxyphenylacetic acid and homovanillic acid (3-methoxy-4hydroxyphenylacetic acid), in monkeys (Burns et

Oxidative stress: free radical production in neural degeneration

al., 1983; Jenner et al., 1984; Langston et al., 1984a), mice (Heikkila et al., 1984a; for reviews, see Heikkila et al., 1989a,b; Gerlach et al., 1991b), cats (Schneider et al, 1986), dogs (Johannessen et al., 1985) and even goldfish (Pollard et al., 1992) Although the chronic progressive nature of the symptomatology of parkinsonism could never be elicited in this model (Birkmayer and Riederer, 1985; Gerlach et al., 1991a), administration of MPTP provides, at present, the most thoroughly investigated animal model for PD. The mechanism of toxicity of MPTP clearly involves bioactivation to A^-methyl-4-phenyl-dihydropyridinium ion (MPDP+) by MAO-B in glia (Heikkila et al., 1984b; Langston et al., 1984b) and serotonin-containing neurons (Westlund et al., 1985). The A^-methyl-4-phenyl-pyridinium ion (MPP+), the degradation product of MPDP+, is an effective substrate for the catecholaminergic synaptosomal uptake system (Javitch et al., 1985; Chiba et al., 1985). Active uptake of MPP+ was also observed in synaptosomes prepared from extrastriatal brain regions (e.g. hypothalamus, hippocampus, cerebellum) (Chiba et al., 1985). At least two possible pathways exist by which MPP+ has been postulated to be toxic: (1) oxidative stress could ensue (Johannessen et al., 1986); or (2) mitochondrial respiration could be inhibited in vulnerable neurons (Nicklas et al., 1985; Ramsay etal., 1986a,b). While some preliminary reports indicated that administration of ascorbic acid (Sershen et al., 1985; Wagner et al., 1985) or TOH (Perry et al., 1985) to mice is partially protective against MPTP toxicity, other reports dispute protection (Smith et al., 1987; Perry, T.L. et al., 1987; Baldessarini et al., 1986; Martinovits et al., 1986; Mihatsch et al., 1991; Gong et al. 1991). However, the toxicity of MPTP has been found to be potentiated by TOH deficiency (Odunze et al., 1990) or after depletion of brain GSH by intracerebroventricular injections of diethylmaleate (Adams et al, 1989), Young et al. (1986), as well as Riederer et al. (1987), found that MPTP slightly lowers mouse brain stem GSH levels, an effect that could be explained by reduced NADPH-dependent or by higher need of GSH by

M.E. Gotz et al.

GSH-peroxidase because of detoxification of putatively increased amounts of hydroperoxides. Therefore, it seems crucial to protect cells from loss of antioxidants and potent reducing substances, such as NADPH or NADH. In turn, depletion of cells from ATP by inhibition of oxidative phosphorylation could result in decreased activity of 5-oxoprolinase, glutamyl-cysteinesynthetase and GSH synthetase, all known to be dependent on ATP (Meister, 1991), resulting in impaired detoxification mechanisms and subsequent cell damage because of oxidative stress. On the other hand, ATP depletion caused by MPP+, potassium cyanide or antimycin A (the latter two substances are known as definite inhibitors of mitochondrial respiratory chain) could lead to decreased GSH concentrations independent of oxidative stress, because substrates for glycolytic production of ATP counteract the GSH depletion caused by mitochondrial respiratory chain inhibitors (Mithofer et al., 1992). Thus, we have to consider a second hypothesis of MPTP neurotoxicity. Metabolism, transport and storage of MPTP within cells is functionally closely connected with mitochondria. Once created by the MAO in glial cells, MPP+ is suspected to be actively concentrated in mitochondria of astrocytes and of dopaminergic neurons (Ramsay and Singer, 1986). Reports of MPP+ interference with mitochondrial function and ATP formation provide an alternative possible mechanism to explain toxicity of MPP+ (for a review, see Gerlach et al., 1991b). MPP+ is a potent inhibitor of oxidation of the NAD+-linked substrates pyruvate/malate and glutamate/malate in isolated rat liver and brain mitochondria, while leaving the oxidation of succinate unaffected (Nicklas et al., 1985). The locus of inhibition of the mitochondrial respiration is assumed to be between the highest potential Fe-S cluster in NADH dehydrogenase and the coenzyme Q located probably at the rotenone-binding site (Ramsay et al., 1991). As a consequence of inhibition of respiration, cellular energy supplies in the form of ATP would rapidly be consumed, followed by depolarization of membranes, probable Ca^"^ influx and overstimulation of Ca^"^-dependent lysosomal enzymes.

469

Whichever mechanism of toxicity is exerted by MPP+, the result could be impairment of 'GSH homeostasis' in mitochondria. Adopting this hypothesis, some investigators have sought to identify possible endogenous compounds similar in structure to MPTP and MPP+ in parkinsonian patients and in animal models of parkinsonism. Amongst these substances, tetrahydroisoquinolines (THIQ) and tetrahydro-)8-carbolines (THBC) gained considerable interest recently, because they produce parkinsonism in animals after long-term treatment (Nagatsu and Yoshida, 1988; Yoshida et al., 1990) and are found in foods, such as cheese, milk, chocolate powder and wine and even endogenously in brain (Makino et al., 1988; Adachi et al., 1991a,b; Rommelspacher et al., 1991). The mechanisms of toxicity of THIQ and THBC are supposed to be similar to that of MPTP. Subsequent to uptake by catecholaminergic transporters and MAO-catalyzed oxidation, the resulting isoquinolinium ions or carbolinium ions and/or A^-methylated forms could accumulate within mitochondria and disturb electron flow in the respiratory chain (Suzuki et al., 1988, 1992; Albores et al., 1990; Collins et al., 1992; Fields et al., 1992). Naoi et al. (1989a,b) and Matsubara et al. (1992a,b) have provided evidence that it is predominantly the //-methylated THIQ and THBC, which are oxidized by MAO. The resulting A^-methyl-isoquinolinium ions (MIQ+ ) or A^-methyl-^-carbolinium ions (MBC+) are effective enzyme inhibitors in vitro (Naoi et al., 1989c; Sayre et al., 1991). Selectivity of THlike immunoreactive cells for MIQ"*" at 100//M was demonstrated (Niijima et al., 1991), suggesting its uptake and accumulation selectively by catecholaminergic neurons. Despite less severe toxicity of MIQ"^ compared with that of MPP+ in vivo (to induce a similar decrease of DA in striatum by intranigral administration of MIQ+ or MPP+, a 20-fold higher concentration of MIQ+ than of MPP+ is needed; Sayre et al., 1991), prolonged accumulation of such compounds over a lifetime from foods or because of severe ethanol uptake (Collins et al., 1990) could putatively be a risk factor for catecholaminergic neurons. This is especially true in subjects with a poor metaboliz-

470

ing capacity involving cytochrome P450 reductase, where brain accumulation of THIQ is enhanced (Ohta et al., 1990). Interestingly, it has been found that, among patients with PD, many are poor debrisoquine metabolizers (Poirier et al., 1987). This implies that PD may be associated with low levels of the specific form of cytochrome P450, which oxidizes debrisoquine and potentially MPTP-like environmental toxins. Recently, however, using immunoassays sensitive to a broad range of compounds structurally related to MPTP and MPP+, Ikeda et al. (1992) could not demonstrate altered immunoactivity in striatum of parkinsonian versus control brains, suggesting that compounds chemically related to MPTP are not likely to exist (at a nanogram range) in PD brain at the time of death. Nevertheless, it can be imagined that such compounds could be initiators of toxic events (genetic defects, impairment of mitochondrial function), leading to loss of a certain amount of cells, but not enough to produce parkinson-like symptoms early in life. However, due to the age-related decline of dopaminergic cells, PD would become manifest at a later stage of life. If isoquinolines or carbolines played a major role in PD, consumers of large amounts of cheese, chocolate powder and wine would be expected to show at least a greater prevalence of PD than others, since, for example, THIQ may be formed by ring cyclization of 2-phenylethylamine in foods containing formaldehyde or in combination with ethanol in beverages. However, this seems not to be the case (Andrade, 1991). The discovery of MPTP has rekindled the environmental toxin theory of PD that started with PD in manganese miners (Barbeau, 1984). Nevertheless, despite a number of epidemiological studies that suggest an unisotrope occurrence of PD in some populations, perhaps related to exposure to environmental toxins (Barbeau et al., 1986; Rajput et al., 1987; Tanner, 1989; Ho et al., 1989; Goldsmith et al., 1990) and the considerable efforts made to elucidate the mechanisms of toxicity of various exogenous and even endogenous compounds, to date it still is not possible to define the ultimate *parkinsonism-inducing toxin'.

Oxidative stress: free radical production in neural degeneration

5.2. Activities of enzymes of the respiratory chain in Parkinson's and Alzheimer's diseases Mitochondria are one of the main generators of ROS (Section 3.2.2). Consequently, at the site of cellular free radical generation, the enzymes of the respiratory chain and the mitochondrial DNA (mtDNA) are particularly susceptible to damage by ROS. The rate of mitochondrial (O2)" and H2O2 generation increases with age in houseflies and in the brain, heart and liver of the rat (Sohal et al., 1990; Sohal and Sohal, 1991; Sohal, 1991; Sohal and Brunk, 1992). In addition, mitochondrial respiratory chain functions in human muscle (Trounce et al., 1989) or human hver (Yen et al., 1989) have been reported to decline with increasing age. Therefore, we have to ask whether these accelerated age-related events are important in the pathology of neurodegenerative diseases with adult onset, such as PD and AD. In PD, several groups of investigators have reported mitochondrial respiratory dysfunctions in brain, muscle and platelets. Using immunoblotting techniques with specific antisera against enzyme complexes I, III and IV, Mizuno et al. (1989) found a decrease in four out of five patients with PD in the 30-, 25- and 24-kDa subunits of complex I. In contrast, Schapira et al. (1989, 1990a,c) and Mann et al. ( 1992a) detected a decrease in complex I activity in the SN of PD, but not in cerebral cortex, cerebellum, globus pallidum, caudate nucleus and tegmentum. Due to the high amount of glial cells in the brain, it has to be assumed that, in addition to that of neuronal cells, glial complex I is also defective. The absence of changes in activities of mitochondrial enzyme complexes I-IV in MSA suggests that complex I deficiency in PD is not due to cell death and, perhaps, may be specific to PD (Schapira et al. 1990a). Immunohistochemical studies in PD showed a fair proportion of the nigral neurons with reduced staining against complex I and, in three patients, against complex II antibodies, whereas staining for complexes III and IV appeared normal (Hattori et al., 1991). Two studies using platelets failed to detect abnormalities in the enzyme activities of respiratory chain complexes (Mann et al., 1992a; Bravi et al..

M.E. Gotz et al.

1992), in contrast to Parker et al. (1989), who reported a 55% decrease in the mean platelet mitochondrial complex I activity, and Yoshino et al. (1992), who published small but significant decreases in platelet complex I and II activities. Studies on skeletal muscle from PD patients have produced conflicting results, also. As summarized by Schapira and Cooper (1992), three studies have shown multiple respiratory chain deficiencies in some patients, pure complex I deficiency in others and normal activities in one patient with advanced disease (Bindoff et al., 1991; Shoffner et al., 1991; Nakagawa-Hattori et al., 1992). Because the reasons for these inconsistent findings are still unknown, it seems impossible, to date, to develop a diagnostic test for PD using blood cells or biopsies from peripheral tissues, such as muscle. A similar inconsistency exists concerning oxidative phosphorylation in AD. Although activities of enzymes of the mitochondrial electron transfer chain are reported to be normal in AD brain, partial uncoupling of oxidative phosphorylation (electron transfer and phosphorylation of adenosine diphosphate are normally functionally linked) (Sims et al., 1987) and overexpression of cytochrome oxidase subunit-3 gene in cerebral temporal cortices (Alberts et al., 1992) have been reported. In addition, substantial decreases of complex IV activity were detected in platelets from five patients with AD (Parker et al., 1990). Although not entirely convincing, these results could point towards a more generalized defect of oxidative phosphorylation in neurodegeneration, with a possible genetic determination. 5.3. Alterations in mitochondrial deoxyribonucleic acid The 16.6 kb human mtDNA codes for two ribosomal ribonucleic acids, 22 transfer ribonucleic acids and 13 peptides, which are part of enzyme complexes of the respiratory chain in the inner mitochondrial membrane (Capaldi, 1988). Mitochondria are largely, but not entirely, maternally inherited (Gyllensten et al., 1991). They proliferate

471

independently of the cell cycle. In mammals, mtDNA mutates much faster than nuclear DNA, possibly because mtDNA is not covered by histones, and is at least transiently attached to the inner mitochondrial membrane, where large amounts of ROS are produced. Therefore, mtDNA is particularly susceptible to oxidative damage. The steady-state level of oxidized bases in mtDNA is about 16 times higher than in nuclear DNA (Richter et al., 1988; Hruszkewycz and Bergtold, 1990). Numerous base modifications are detectable when ROS react with DNA (von Sonntag, 1987). The most studied oxidized base is 8hydroxydeoxyguanosine, which can be measured in the femtomolar range (Halliwell and Aruoma, 1991). ROS generate strand breaks in mtDNA (reviewed by Richter, 1988, 1992), and DNA repair in mitochondria is much less efficient than in the nucleus. These mammalian organelles do not have significant recombinational repair, but may excise damaged bases. Mutations of mtDNA are the cause of some oxidative phosphorylation diseases with prominent basal ganglia pathology, such as Leigh disease (Montpetit et al., 1971; Pincus, 1972) and Leber's disease with dystonia (Novotny et al., 1986). Kearns-Sayre syndrome and mitochondrial encephalomyopathy lactic acidosis and stroke-like episodes may show basal ganglia calcifications, neuronal loss and spongiform histopathological changes (Horwitz and Roessmann, 1978; Driscoll et al., 1987; Ichiki et al., 1988). These mitochondrial diseases are predominantly maternally transmitted (Harding, 1991). It is now evident that large deletions of mtDNA, duplication of mtDNA or point mutations of mtDNA account for the metabolic changes observed in these diseases (Harding, 1991; Wallace, 1989, 1992). However, as for PD, pedigree analysis has identified only a few families with familial PD, suggesting autosomal dominant inheritance with variable penetrance (Golbe et al, 1988, 1990), but providing no evidence of maternal inheritance (Maraganore et al., 1991). Twin studies in PD have not provided very conclusive data on inheritance susceptibility, although there are promising activities in progress (Johnson et al., 1990; Du-

472

voisin and Johnson, 1992). Initial reports of an increase in mtDNA deletion in SN of parkinsonian patients (Ozawa et al., 1990; Ikebe et al., 1990) could not be confirmed by others (Schapira et al., 1990b; Lestienne et al., 1990, 1991), but was attributed to an age-related phenomenon (Yen et al., 1991; Mann et al., 1992b). The role of genetic factors in the etiology of sporadic cases of PD remains to be determined. There is no evidence that major deletions of mtDNA occur in SN in PD, indicating that a further biochemical insult (toxins, ROS) could act directly on mitochondrial enzymes or that the nuclear genome is damaged. Clearly, further effort is needed in this field of research. 6. Excitatory amino acids and neurodegeneration The importance of EAA in physiological function of neural transmission is well known. Glutamate was the first of them to be recognized. In the meantime, a further subdivision of the glutamate receptors has taken place. At the moment, we know of four glutamate receptor subtypes (Watkins et al. 1991): (1) the NMDA, (2) the a-amino-3-hydroxy5-methyl-4-isoxazolepropanoic acid (AMPA), (3) the kainate and (4) the metabotropic receptor. The EAA are involved as neurotransmitters in the sensory input of spinal and supraspinal systems. Another important function of EAA is to contribute to the programming and execution of movements in the motor loop (Riederer et al. 1989b). A further EAA-mediated neuronal function is the process of learning and memory, where the NMDA receptor, in particular, is thought to be important because of its electrophysiological action in long-term potentiation, which introduces synaptic plasticity (Monaghan et al., 1989; CoUingridge and Singer, 1991). Besides the role of EAA in physiological actions, an excitotoxic etiology of neurodegenerative diseases has been proposed in the last few years (see also the chapter by Zorumski and Olney in this volume). A lot of investigations show a noxious effect of EAA, which is mediated by Ca^^ influx into the cytoplasm (Monaghan et al. 1989; Mayer and Miller, 1991). Endogenous and envi-

Oxidative stress: free radical production in neural degeneration

ronmental EAA receptor agonists can cause acute and chronic neurodegenerative diseases, resulting in dysfunction of motion and memory. Environmental diseases, such as Guam disease, neurolathyrism and mussel poisoning, can be used for studying the putative role of excitotoxic substances in PD, ALS and dementia. In Guam disease, a still unidentified ingredient of the cycas seed, perhaps the sago palm toxin 2amino-3-(methylamino)-propanoic acid (Kurland, 1988; Meldrum and Garthwaite, 1990, 1991; Duncan et al. 1990), causes a syndrome characterized by dementia and by aspects of ALS, PD or both, in people living on the islands of Guam and Rota (Zang et al., 1990). The noxious agent acts by binding on the NMDA-receptor ion channel (Meldrum and Garthwaite, 1990, 1991). Although there are valid doubts concerning the participation of the sago palm toxin in the etiology of Guam disease (Spencer et al., 1990), the improvement of MPTP-provoked hypokinesia by treatment with EAA antagonists (Loschmann et al., 1991; Klockgether et al., 1991) supports the general hypothesis of a glutamate dysbalance in PD. Investigation of glutamate receptor density in PD SNC has shown a loss of glutamate-receptor subtypes (Difazio et al., 1992). The heterogeneous distribution of glutamate-receptor subtypes in the striatum and the participation of these receptors in movement function has lead to the hypothesis of glutamate imbalance in motor system disturbances that show opposite clinical pictures (Albin et al., 1991, 1992). One implication for the excitotoxic cell damage in PD and HD is the binding of 6-hydroxylated LDOPA on non-NMDA receptors. Thus, the hypothesis arises that an abnormal derivative of the natural DA precursor may be responsible for destroying nigral and striatal cells in these diseases (Hanson et al., 1985; Olney et al., 1990; Cha et al., 1991). Application of kainic acid, and the more specific quinolinic acid, into the striatum is followed by Huntington-like structural and biochemical changes; excitotoxin injection into the striatum produces locomotor hyperactivity and deficits in learning tests (DiFiglia, 1990).

473

M.E. Gotz et al.

Quinolinic acid is an endogenous NMDA receptor agonist formed from L-tryptophan, thought to be elevated in HD. Although there is still disagreement that neurodegeneration in HD is mediated by altered tryptophan/quinolinic acid ratios, the excitotoxic model of HD is a solid approach to the elucidation of the role of glutamate in movement disorders (Schwarcz et al., 1987; Reynolds et al., 1988; Connick et al., 1988; Brogn and Stoof, 1990; Bakker and Foster, 1991). In AD, impairment of memory and cognition could reflect disturbances in NMDA-mediated long-term potentiation, and a decline of NMDA receptors in hippocampal and cortical regions has been found (Jansen et al., 1990). Under hypercalcemic conditions, stimulation of NMDA receptors can produce formation of cytoskeletal NFT-like structures. Therefore, NFT could be seen to be a consequence of the degenerative process. One of the best investigated environmental diseases is neurolathyrism, a spastic paresis, resulting from damage of the corticospinal tract. A reliable connection between the lathyrus toxin ^-Noxalylamino-L-alanine and the degenerative process of the upper motoneuron has been shown (Mertens, 1947; Spencer et al., 1990). In this case, the neurotoxic process is initiated by binding on non-NMDA receptors (Ross et al., 1989). Implication of EAA in cognitive processes is documented in a further exotoxic environmental disease: in domoic acid poisoning, many people have shown a severe memory impairment after eating mussels (Sutherland et al., 1990; Teitelbaum et al., 1990; Strain and Tasker, 1991). Domoic acid is well known to bind on the kainate receptor. Excitotoxicity also plays a role in epileptic seizures and cerebral ischemia. During ischemic injury, a rapid accumulation of glutamate takes place. This induces a noxious Ca^"^ influx into the cytoplasm. The Ca^+ conductance of NMDA receptors explains the efficacy of NMDA antagonists against hypoxic damage. In addition, AMPA antagonists have also a protective function in hypoxia (Siesjo et al., 1991) In kindling experiments systemic administration of kainic acid in rats induces epileptic seizures (De Veraetal., 1991).

The activation of non-NMDA receptors is sufficient for epileptogenesis, but the latency before the onset of convulsions, their duration and the resulting brain damage depends critically on NMDA participation (Hwa and Avoli, 1991). The action of glutamate in epilepsy is assumed to be the consequence of NMDA-receptor binding and of AMPA-receptor activation as well. In vivo kindling studies provoke epileptic activity. In vitro application of glutamate on hippocampal cell cultures burst firing. These phenomena can be reduced by NMDA antagonists. 7. Current therapy of Parkinson's disease PD is the only neurodegenerative disorder we can symptomatically treat in a very specific and effective manner. Since the first description of the disease by James Parkinson in 1817, a number of procedures for handling this disease have been developed. For a long time, anticholinergic drugs were the only effective treatment of parkinsonian symptoms. The therapeutic mechanism of these drugs is suppression of the relative cholinergic overactivity in the striatum of parkinsonian patients. Since the detection of DA deficiency in PD, neurotransmitter replacement has been the therapy of choice (Coleman, 1992). The transmitter has been applied in form of its precursor L-DOPA, combined with a peripheral decarboxylase inhibitor (e.g. benserazide), so that the drug's efficacy develops only in the CNS. L-DOPA is stored in the remaining nigrostriatal dopaminergic neurons, decarboxylated to DA and then released. It acts postsynaptically on Di and D2 receptors, so that the natural function of the transmitter is imitated and the clinical signs improve. The patients reach a non-fluctuating, constant degree of mobility. Over the years the L-DOPA-storing nigrostriatal neurons continue to diminish, so that DA acts on its receptors in an uncontrolled and pulsatile manner. The clinical correlate of this is an immediate switch of the patients' mobility from bad to good, and vice versa depending on L-DOPA application. The receptors become supersensitive and dyskinetic phases develop (Coleman, 1992). These are

474

also thought to be triggered by DA acting on the D| receptors of remaining descending striatonigral neurons and, in this way, increasing the GABA release in SNR (Robertson, 1992). This undesired L-DOPA effect should be prevented by application of low and frequent L-DOPA doses, not really a successful strategy (Rajput et al., 1984; Poewe et al., 1986a), or by controlled L-DOPA release over the day, a useful but difficult technique (Poewe et al., 1986b; Ceballos-Baumann et al., 1990; Gauthier and Amyot, 1992). L-DOPA therapy may have some disadvantages: the glial MAO-B metabolizes dopamine, decreasing the neurotransmitter content and increasing H2O2 (Fig. 2) (Riederer et al., 1989a; Olanow, 1992). With MAO-B inhibitor treatment, the DA content increases and oxidative damage is diminished (Riederer et al., 1989a; Roller, 1992). This is an important implication of combining LDOPA with L-deprenyl (Coleman, 1992). Protection from the complications of L-DOPA therapy is possible by giving direct DA-receptor agonists, such as the ergot derivates bromocriptine, lisurid, pergolid and apomorphine (Lataste, 1984). The most important effect of combining a D1/D2 agonist, such as bromocriptine, with L-DOPA is a decrease in duration of dyskinetic periods (Montastruc, 1991; Rabey et al., 1991). The motor fluctuations in advanced DOPA-treated patients can be well controlled by subcutaneous apomorphine or lisurid (Poewe et al., 1986b), while an early combination of lisurid with low L-DOPA doses can prevent the development of fluctuations and dyskinesias (Madeja, 1992). A Hmiting factor, especially in lisurid therapy, is the psychotic potency of antiparkinsonian drugs. One possible method of suppressing this side effect is to add the atypical neuroleptic drug clozapine, which does not bind on D2 receptors and, therefore, cannot reverse the effect of dopaminergic drugs (Wolters etal., 1989). In view of the interaction of glutamate and DA of the motor loop in physiological and pathological conditions, glutamate-receptor blockade may offer another avenue for PD treatment (Riederer et al., 1992). Amantadine, which is thought to increase DA release (Coleman, 1992), has been reported to

Oxidative stress: free radical production in neural degeneration

act by blocking the NMDA ion channel (Kornhuber et al, 1991). This well-noted antiparkinsonian drug has also a potent psychotic activity (Danielczyk, 1980). Antagonists at AMPA-receptors, in combination with low-dose L-DOPA, may open new perspectives in PD treatment because of their antiakinetic effect (Loschmann et al., 1991). Another antiparkinsonian drug is piribedil, a direct DA agonist acting in the nigrostriatal and mesolimbic systems; its properties are reduction of tremor and decrease of age-related cognitive decline (Jenner, 1992; Ollat, 1992; Randot and Ziegler, 1992). 8. Protective versus symptomatic therapy 8.1. Antioxidative strategy for neuroprotection If ROS are involved in the initiation or progression of neurodegeneration, then scavengers of ROS should ideally stop the pathogenetic process and preserve the function and integrity of vulnerable neurons. Because iron may promote formation of ROS, its abnormal presence may contribute to oxidative stress. Thus, strategies have been designed to reduce entry of iron into the brain, to increase non-toxic brain storage of iron and to remove iron through chelation. Desferrioxamine, the most widely used chelating agent, however, does not enter the brain after systemic administration, thus diminishing only the peripheral iron (Halliwell, 1989a,b). D-Penicillamine was shown to reduce brain iron content when it was given intraperitoneally to rats. It was shown to be an inhibitor of iron-induced LPO in vitro and in vivo (Ciuffi et al., 1992) and to chelate copper efficiently. For this reason, penicillamine aroused reasonable interest for the treatment of Wilson's disease. In this autosomal recessive disorder, agents that deplete brain copper, such as penicillamine, are effective in slowing basal ganglia damage, supporting restoration of neurological function and preventing the onset of illness in presymptomatic homozygotes (Scheinberg and Sternlieb, 1984). Because iron is involved in many biologically important processes, long-term treatment with iron

M.E. Gotz et al.

475

Some drugs of potential benefit in AD and PD Alzheimer's disease

Parkinson's disease

predominant symptomatic mode of action augmenting acetylcholine cholinesterase inhibitors (e.g. physostigmine) increasing precursors ( l e c i t h i n , a c e t y l - l - c a r n i t i n e )

vasodilators (e.g. captopril) nootropics (e.g. piracetam, oxiracetam)

augmenting dooamine dopamimetics L-DOPA dopaminergic agonists e.g. bromocriptin, lisurid, pergolid, piribedil MAO-inhibitors counteracting dooamine imbalance to other neurotransmitters anticholinergics glutamate antagonists

symptomatic and putative neuroprotective mode of action calcium channel blockers (e.g. nimodipine)

monoamine oxidase B inhibitors (e.g. selegiline)

monoamine oxidase B inhibitors (e.g. selegiline)

glutamate receptor antagonists

glutamate receotor antagonists NMDA receptor blocker (e.g. amantadine) AMPA receptor blockers (e.g. NBQX)

NMDA receptor blocker (e.g. amantadine) AMPA receptor blockers (e.g. NBQX) antioxidants

growth factors (e.g. NGF; BDNF)

alpha-tocopherol, carotenoids ascorbic acid thiols ( N - a c e t y l c y s t e i n e , e b s e l e n )

Ginkgo biloba ?

iron chelators (e.g. aminosteroids; penicillamine)

gangliosides

Fig. 6. Drugs with known and putative effects in treatment of neurodegenerative diseases or in treatment of animal models of these diseases. NBQX, a-amino-3-hydroxy-5-methyl-4-isoxazole-propanoic acid.

chelators could provoke many adverse effects and, therefore, seems not to be feasible. Thus experimental neuroprotective strategies may be focused on preventing deleterious biochemical consequences of increased brain iron. A variety of agents are purported to inhibit iron-dependent oxidative stress in brain, including ascorbic acid, aminosteroids, TOH and L-deprenyl (a selective MAO-B inhibitor). The latter two substances are currently being subjected to clinical trials for the treatment of PD (Fig. 6). Novel steroids (21-aminosteroids, lazaroids and 2-methylaminochromans) were developed for acute treatment of traumatic or ischemic CNS injury (for reviews, see Jacobsen et al., 1990; Hall, 1992). They were shown to be potent inhibitors of iron-dependent LPO in liposomes or of peroxida-

tion of linoleic acid in the presence of the methanol-soluble free radical generator (2,2'-azobis(2,4dimethyl valeronitril)) (Braughler et al. 1987; Braughler and Pregenzer, 1989), presumably by scavenging lipid peroxyl radicals and thus, blocking lipid-radical chain reactions in a manner similar to TOH. Indeed, administration of the lazaroid U74006F (tirilazad mesylate) prevented decreased brain levels of ascorbate and TOH during brain ischemia reperfusion injury in gerbils (Sato and Hall, 1992). The importance of TOH for human brain function becomes dramatically evident in patients with intestinal fat malabsorption syndromes, who suffer from a severe and prolonged deprivation of TOH (Muller and Goss-Sampson, 1990). In vitro TOH supports the survival and neurite extension of neu-

476

rons of fetal rat brain (Nakajima et al., 1991) and thus, could be putatively beneficial for viability of neurons in vivo. This hypothesis has prompted a pilot study of high dose TOH and ascorbate in early PD (Fahn, 1992). Twenty-one patients were treated with antimuscarinics and amantadine, but not with L-DOPA, and received TOH and ascorbic acid up to 3200 U/day and 3000 mg/day, respectively. The end-point for analysis was the time at which L-DOPA or DA agonist had been required (i.e. when symptoms were severe enough to become a threat to employment or to social or physical capacity). Compared with an independent control group, the need for L-DOPA was delayed by 2-3 years in patients receiving vitamins. TOH is part of the large North American multicenter, controlled trial known as DATATOP (Deprenyl And Tocopherol Antioxidant Therapy Of Parkinsonism), which is evaluating 2000 U/day TOH and 10 mg/day L-deprenyl (Section 8.2) in 800 patients with early PD. However, there was no beneficial effect of TOH alone or any synergistic interaction between TOH and L-deprenyl (The Parkinson Study Group, 1989a,b, 1993). By contrast, a sHght beneficial effect on tardive dyskinesia ratings following TOH could be observed in patients who had had this disease for 5 years or less (Egan et al. 1992). 8.2, L'Deprenyl® (Selegiline®), a selective monoamine oxidase-B inhibitor, in the treatment of early Parkinson's disease If oxidative products of catecholamine metabolism indeed do provide free radical production and provoke progressive deterioration of the DA nigrostriatal system, then blockage of formation of H2O2 by MAO and blockage of accumulation and high turnover of DA without depriving the postsynaptic DA receptors of ligand (e.g. by using agonists selectively acting on presynaptic D2receptors, such as pergolid) should result in a halt of the neuronal and clinical decline in PD (Felten et al. 1992). The selective inhibitor of MAO-B, Ldeprenyl, has gained wide acceptance as a useful form of adjunct therapeutic drug in the treatment of PD (Knoll et al., 1978; Knoll, 1987; Gerlach et

Oxidative stress: free radical production in neural

degeneration

al., 1992) and has been reported to be effective in improving the life expectancy of patients with PD (Birkmayer et al. 1985). The pharmacological basis of the therapeutic effect of L-deprenyl was fully reviewed recently (Chrisp et al., 1991; Knoll, 1992a,b; Gerlach et al. 1992) and has been an important topic in some recent international symposia (published in: Riederer and Przuntek, 1987; Rinne and Heinonen, 1991; Lieberman, 1992). As mentioned in Section 8.1, L-deprenyl is currently under investigation in the DATATOP study, where no other drugs, except TOH, are allowed. An unplanned interim analysis of this trial indicated that L-deprenyl reduced the risk of disability requiring L-DOPA therapy by approximately 50% and similarly reduced the loss of full-time employment (The Parkinson Study Group, 1989a,b). Although encouraging, these interim results, and that of the final report (The Parkinson Study Group, 1993), do not necessarily support a neuroprotective effect of L-deprenyl in PD because the observed functional benefits were also accompanied by slight, but statistically significant, improvements in the cHnical measures of PD. Thus, the findings may reflect the symptomatic antiparkinsonian effects of L-deprenyl (Shoulson, 1992) that could be expected to occur also during treatment with anticholinergics, amantadine or DA agonists. These drugs would provide temporary therapeutic benefits and, thus, might delay the need for introduction of a more potent antiparkinsonian drug, such as L-DOPA (Olanow and Calne, 1992). Whether L-deprenyl enhances DA and/or phenylethylamine availability is still a matter of discussion (Paterson et al., 1990). Furthermore, L-deprenyl metabolites might contribute to clinical effects. Virtually all administered L-deprenyl is metabolized to (-)methamphetamine and to a lesser extent to amphetamine (Reynolds et al.. 1978), which enhance the release of DA and block its reuptake. However, they are present in low concentrations, rapidly cleared and are far less potent than the (+)enantiomers, which cannot be generated from Ldeprenyl. It, therefore, appears unlikely that an amphetamine-like action is the only cause of symptomatic effects of L-deprenyl in PD (Heinonen and Lammintausta, 1991). However, a

All

M.E. Gotz et al.

comparison between a selective MAO-B inhibitor, which does not metabolize into amphetamine metabolites, and L-deprenyl will probably give new insight to this question. The findings of Spina and Cohen (1988, 1989) and Cohen and Spina (1989) have demonstrated that increased extracellular DA turnover induces increased concentrations of GSSG. This rise in GSSG can be suppressed by coadministration of Ldeprenyl, indicating that it is caused by the oxidative metabolism of DA. Moreover, inhibition of MAO protects rat brain from hyperbaric oxygen toxicity (Zhang and Piantadosi, 1991), suggesting that H2O2 production by MAO is a biologically significant risk factor for brain degeneration if cellular defense mechanisms are impaired, as is obviously the case in SN in parkinsonian patients. Interestingly, depending on the sex and age of experimental animals, L-deprenyl increases the activities of CuZnSOD, MnSOD in striatum and SN, and to a lesser extent, in cerebral cortices, as well as CAT activity in striatum, but not in the hippocampus of rats (Knoll, 1988; Carrillo et al, 1992a,b). Despite the fact that the mechanisms of action of L-deprenyl on activities of SOD and CAT remain obscure, there could be a hypothetical link to the free radical theory of aging. Knoll and coworkers (Knoll, 1988; Knoll et al., 1989) observed an increased life-span of male rats after long-term administration of L-deprenyl. Nevertheless, because MAO-B is unlikely to be present in dopaminergic neurons in SNC, the importance of MAO-induced production of H2O2 for catecholaminergic cell death must be seriously questioned. Tatton and Greenwood (1991) noted that following administration of MPTP, neuronal degeneration in the SNC of mice could be dramatically attenuated by L-deprenyl, even when administered 72 h after the last dose of MPTP. Thus, a trophic effect similar to that reported with brain-derived neurotrophic factor (BDNF) (Hyman et al. 1991) was attributed to the drug. This was further substantiated by the findings of Salo and Tatton (1992), who reported the number of surviving motoneurons 21 days after axotomy to be 2.2 times higher than in mice not given L-deprenyl. Thus, L-deprenyl can rescue neurons other than

those in the SN and can compensate in part for the loss of target-derived trophic support caused by axotomy. These findings are also encouraging for use of L-deprenyl in treatment of neurodegenerative diseases other than PD. In AD patients, Ldeprenyl has been shown to improve verbal memory (Finali et al., 1991) and a wide variety of other cognitive functions, without frequent or severe adverse effects (Mangoni et al., 1991) 8.3. Maintaining neuronal plasticity The major symptoms of AD result from massive destruction of cholinergic synaptic terminals in cerebral cortex and subcortical structures (Bartus et al., 1982), although other noncholinergic synaptic terminals are also affected. Autopsy studies have shown that age-related loss of cholinergic cells in the NBM has a dropout rate very close to that observed in the LC and the zona compacta of the SN, two non-cholinergic nuclei (McGeer et al., 1977, 1984; Mann et al., 1984a,b). In AD the level of ChAT, the synthesizing enzyme for ACh, is reduced in the neocortical projection fields of basal forebrain cholinergic neurons (Etienne et al., 1986). Perry et al. (1978) found a positive correlation between these cholinergic markers and clinical severity on dementia scales. Thus, present therapeutic strategies are directed towards enhancing cholinergic function by: (1) increasing various precursor molecules utilized for ACh synthesis (choline, phosphatidylcholines, acetyl-L-carnitine (ALC), L-a-glycerylphosphorylcholine) (2) improving ACh function (through administration of aminopyridines to increase reuptake of free choline from synapses, or cholinesterase inhibitors, or muscarinic agonists) and (3) maintaining membrane function and neuronal viability (growth factors, gangliosides, phosphatidylserine) (for reviews, see Cooper, 1991; Holttum and Gershon, 1992). By antagonizing progressive destruction of membrane phosphocholines by a putative mechanism of 'autocannibalism' (Wurtman, 1992), choline precursors should slow the degenerative process. However, results of treatment with choline

478

(Ferris et al., 1982) and with phosphatidylcholines (Bartus et al., 1982) have been reported to be negative or at least equivocal. Although not yet subject to clinical trials for treatment of AD, cytidine-5'-diphosphocholine, a major precursor in the synthesis of phosphatidylcholines, phosphatidylserines and phosphatidylethanolamines in cell membranes, affects the synthesis and levels of cell membrane phospholipids in PC-12 cells when simultaneously incubated with choline (Lopez G.Coviella and Wurtman, 1992) and in brain of mice after long-term treatment for 27 months (Lopez G.-Coviella et al, 1992). In contrast, a large multicenter study from Italy reported improvements in logical intelligence, verbal critical abilities long-term verbal memory and selective attention in AD patients receiving oral ALC for 1 year (Spagnoli et al., 1991; critically reviewed by Bowman, 1992). Although progression of the disease was not halted, it was slowed markedly. Carnitine (3-hydroxy-4-A^-trimethylaminobutyric acid) is a naturally occurring important metabolite in higher organisms that plays a key role in the transport of fatty acids from the cytosol into the mitochondrial matrix for ^-oxidation (Bahl and Bressler, 1987; Bieber, 1988). Carnitine is synthesized from protein-bound lysine, mainly in the liver, kidney and brain (Shug et al., 1982) and is additionally present in plasma and muscle in its free form and as acylcamitine esters (Bieber and Lewin, 1981). Acylcarnitines can be exchanged across subcellular membranes and ALC serves as a pool of acetyl groups from which to regenerate acetylcoenzyme A (Dolezal and Tucek, 1981). ALC is structurally similar to ACh and has been shown to increase ACh synthesis, to promote ACh release (Imperato et al, 1989) and to increase ChAT activity. These modes of action are possibly responsible for the symptomatic effects of ALC. In addition, a putative protective effect of ALC should be considered in neurodegeneration. ALC appears to be effective in reversing certain aging processes in the brain (Sershen et al., 1991) such as reducing some of the morphological changes in lipofuscin of aged rat Purkinje neurons (Dowson et

Oxidative stress: free radical production in neural degeneration

al., 1992) and decreasing lipofuscin accumulation in hippocampal and prefrontal brain areas of aged rats (Badiali De Giorgi et al., 1987; Ramacci et al., 1988). Moreover, ALC as well as carnitine increases the metabolic rate of mitochondria, thereby improving mitochondrial oxygen utilization in experimental brain ischemia (Matsuoka and Igisu, 1992; Rosenthal et al., 1992). ALC mechanisms of action in potentiating brain energy metabolism remain to be elucidated (for a review, see Calvani and Carta, 1991). ALC partially protects nonhuman primates against MPTP toxicity (BodisWollner et al., 1991), which could be explained, in part, by an ALC-induced increase in cytochrome c oxidase activity (Villa and Gorini, 1991; Petruzzella et al., 1992). Extended ALC administration causes an increase in nerve growth factor (NGF) receptors in the striatum of developing rats (De Simone et al., 1991) and in cultured PC-12 cells (Taglialatela et al., 1991, 1992). In cultures of aged dorsal root ganglia, ALC did not affect axonal regeneration, as was seen with NGF, but substantially attenuated the rate of neuronal mortality (Manfridi et al., 1992). Regardless of the ultimate cause of neurodegeneration, therefore, ALC appears to attenuate impairment of mitochondrial energy metabolism in vitro and in neuronal cell cultures, and may support regenerative processes in neurons via growth factor biochemistry. Although AD is not usually considered a metabolic disorder, evidence supporting this hypothesis exists. Abnormalities in a number of brain enzymes involved in glucose metabolism (Friedland et al. 1989; Kalaria and Harik, 1989) and in Ca^^ homeostasis in AD have been detected (Martin et al., 1989). Thus, ALC may be useful to attenuate cellular energy deficiency in aging and AD. A causal relationship of NGF depletion and AD seems unlikely because levels of NGF messenger ribonucleic acid in AD are not decreased (Goedert et al., 1986) and because there is a consistent degenerative effect on neural populations in AD, which are not dependent upon or responsive to NGF (Hefti and Weiner, 1986). Nevertheless, intraventricular administration of NGF produces

479

M.E. Gotz et al.

trophic actions on cholinergic neurons and prevents age-related neuronal atrophy (Hefti and Schneider, 1991), justifying the evaluation of NGF as a therapeutic tool for the treatment of AD. The protective function of NGF, even in the presence of ROS, is underlined by experiments in cultured rat astrocytes expressing NGF and basic fibroblast growth factor, following 0.2-1 mM H2O2 (Pechan et al., 1992). As well, NGF restores normal CAT activity and increases CuZnSOD and seleniumdependent GSH-Px activity in several brain areas of aged rats (Nistico et al., 1992) and cultured PC12 cells (Jackson et al., 1990a,b), suggesting that the effects of different toxic events (impairment of mitochondrial respiratory chain induced by MPF*" or accumulation of ROS) can be attenuated by NGF. Similarly, BDNF (Hyman et al., 1991) selectively protects DA neurons against 6-OHDA and MPP"^ toxicity (Spina et al., 1992) by increasing the activity of GSSG-Rd, but not of CAT, and thereby preventing loss of GSH. Although growth factors are obviously not able to prevent onset of neurodegeneration, they could enforce regenerative processes of already damaged but still viable neurons. They, therefore, have been seriously considered for treatment of degenerative diseases. However, growth factors have to be given intracerebrally because, being peptides, they do not cross the blood-brain barrier, rendering their usage in clinical practice more difficult. Thus, will there perhaps be strategies in the future to stimulate growth factor synthesis at specific brain regions in the adult brain? Another therapeutic approach could be the use of membrane-stabilizing agents, such as gangliosides. Gangliosides are sialic acid-containing glycosphingolipids, which are highly concentrated in neuronal membranes (Leeden, 1984; Mahadik et al., 1992). Morphological, developmental, biochemical and behavioral studies have demonstrated that gangliosides participate in the maturation and repair of neural tissue (Thomas and Brewer, 1990; Bull Zeller and Marchase, 1992). They enhance recovery from MPTP-induced parkinsonism in rodents and primates (Hadjiconstantinou et al., 1989, 1990; Fazzini et

al., 1990; Gupta et al. 1990; Schneider et al., 1992) and have been used for the treatment of stroke and AD (Bassi et al., 1984; Porsche-Wiebking, 1989; Ala et al., 1990). During aging, and in AD, some gangliosides (GMI, GDla) were found to be decreased in NBM, frontal cortex and temporal cortex, but the simple gangliosides GM2 and GM3 were elevated (Kracun et al., 1992a,b). These authors correlated their results with accelerated lysosomal degradation of gangliosides and/or astrogliosis occurring during neuronal death. Despite these intriguing data, gangliosides are unlikely to reverse the disease process. Clearly, further effort is needed to find new compounds, or at least efficient combinations of all these drugs mentioned, in order to slow down the destructive processes in brain. Acknowledgements

The authors wish to thank Dr. Wieland Gsell, Dr. K.W. Lange and Dr. A. Dirr for helpful discussions and R. Burger for excellent technical assistance. This work was supported by a grant from the Bundesministerium fur Forschung und Technologic (BMFT), Germany, grant number 01KL9101-0.

References Adachi, J., Mizoi, Y., Naito, T., Ogawa, Y., Uetani, Y. and Ninomiya, I. (1991a) Identification of tetrahydro-betacarboline-3-carboxylic acid in foodstuffs, human urine and human milk. J, Nutr. 121: 646-652. Adachi, J., Mizoi, Y., Naito, T., Yamamoto, K., Fujiwara, S. and Ninomiya, I. (1991b) Determination of beta-carbolines in foodstuffs by high-performance liquid chromatographymass spectrometry. J. Chromatogr. 538: 331-339. Adams, J.D., Klaidman, L.K. and Odunze, I.N. (1989) Oxidative effects of MPTP in the midbrain. Res. Commun. Subst. Abuse 10: 169-180. Adams, J.D., Klaidman, L.K., Odunze, I.N., Shen, H.C. and Miller, C.A. (1991) Alzheimer's and Parkinson's disease. Brain levels of glutathione, glutathione disulfide, and vitamin E. Mol Chem. Neuropathol. 14: 213-226. Aldeman, R., Saul, R.L. and Ames, B.N. (1988) Oxidative damage to DNA: relation to species metabolic rate and life span. Proc. Natl. Acad. Sci. USA 85: 2706-2708.

480 Adolfsson, R., Gottfries, C.G., Roos, B.E. and Winblad, B. (1979) Changes in the brain catecholamines in patients with dementia of Alzheimer's type. Br. J. Psychiatry 135: 216223. Adolfsson, R., Gottfries, C.G., Oreland, L., Wiberg, A. and Winblad, B. (1980) Increased activity of brain and platelet monoamine oxidase activity in dementia of Alzheimer type. LifeSci.ll: 1029-1034. Agardh, C D . , Zhang, H., Smith, M.-L. and Siesjo, B.K. (1991) Free radical production and ischemic brain damage: influence of postischemic oxygen tension. Int. J. Dev. NeuroscL 9: 127-138. Agid, Y. (1991) Parkinson's disease: pathophysiology. Lancet 337:1321-1327. Agid, Y., Javoy-Agid, F. and Ruberg, M. (1987) Biochemistry of neurotransmitters in Parkinson's disease. In: C D . Marsden and S. Fahn (Eds.), Movement Disorders Vol. 2, Butterworths, London, pp. 166-230. Ala, T., Romero, S., Knight, F., Feldt, K. and Frey, W.H. (1990) GMI treatment of Alzheimer's disease. A pilot study of safety and efficacy. Arch. Neurol. 47: 1126-1130. Alberts, M.J., loannou. P., Deucher, R., Gilbert, J., Lee, J., Middleton, L. and Roses, A.D. (1992) Isolation of a cytochrome oxidase gene overexpressed in Alzheimer's disease brain. Mol. Cell. Neurosci. 3: 461-^70. Albin, R.L., Makowiec, R.L., Hollingsworth, Z., Sakurai, S.Y., Dure, L.S., Penney, J.B. and Young, A.B. (1991) Excitatory amino acidergic pathways and receptors in the basal ganglia. Amino Acids 1: 339-350. Albin, R.L., Makowiec, R.L., Hollingsworth, Z.R., Dure, I.S., Penney, J.B. and Young, A.B. (1992) Excitatory amino acid binding sites in the basal ganglia of the rat: a quantitative autoradiographic study. Neuroscience 46: 35-48. Albores, R., Neafsey, E.J., Drucker, G., Fields, J.Z. and Collins, M.A. (1990) Mitochondrial respiratory inhibition by //-methylated beta-carboline derivatives structurally resembling A^-methyl-4-phenylpyridine. Proc. Natl. Acad. Sci. USA 87: 9368-9372. Alexander, G.E. and Crutcher, M.D. (1990) Functional architecture of basal ganglia circuits: neural substrate of parallel processing. Trends Neurosci. 13: 266-271. Alexander, G.E., De Long, M.R. and Strick, P.L. (1986) Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu. Rev. Neurosci. 9: 357381. Alzheimer, A. (1907) Uber eine eigenartige Erkrankung der Hirnrinde. Allg. Z. Psychiatr. Psych. Gerichtl. Med. 64: 146-148. Ambani, L.M., Vans Woert, M.H. and Murphy, S. (1975) Brain peroxidase and catalase in Parkinson's disease. Arch. Neurol. 32: 114-118. Amenta, D., Ferrante, F., Franch, F. and Amenta, F. (1988) Effects of long-term Hydergine® administration on lipofuscin accumulation in senescent rat brain. Gerontology 34: 250-256. American Psychiatric Association (1987) Diagnostic and

Oxidative stress: free radical production in neural degeneration Statistical Manual of Mental Disorders, American Psychiatric Association, Washington, DC. Anderson, R. (1984) The immunostimulatory, antiinflammatory and anti-allergic properties of ascorbate. Adv. NutrRes.6: 19-45. Andrade, L.A.F. (1991) The role of the environmental factors in the genesis of Parkinson s disease. In: T. Caraceni and G. Nappi (Eds.), Focus on Parkinson's Disease, Masson, Milan, pp. 67-78. Antuono, G., Cherayil, G.D. and Ho, K.C (1991) Alteration in phospholipid (PL) fatty acid (FA) in Alzheimer's disease. Neurology 41 (Suppl. 1): 269. Appel, S.H. (1981) A unifying hypothesis for the cause of amyotrophic lateral sclerosis, parkinsonism and Alzheimer's disease. Ann. Neurol. 10: 499-505. Arai, H., Lee, V.M.-Y., Hill, W.D., Greenberg, B.D. and Trojanowski, J.Q. (1992) Lewy bodies contain beta-amyloid precursor proteins of Alzheimer's disease. Brain Res. 585: 368-390. Arai, R., Kimura, H. and Maeda, T. (1986) Topographic atlas of monoamine oxidase-containing neurons in the rat brain studied by an improved histochemical method. Neuroscience 19: 905-925. Armstrong, M., Daly, A.K., Cholerton, S., Bateman, N. and Idle, J.R. (1992) Mutant debrisoquine hydroxylation genes in Parkinson's disease. Lancet 339: 1017-1018. Aust, S.D., Roerig, D.L. and Pederson, T.C (1972) Evidence for superoxide generation by NADPH-cytochrome c reductase of rat liver microsomes. Biochem. Biophys. Res. Commun.41: 1133-1137. Autor, A.P. (1982) Biosynthesis of mitochondrial superoxide dismutase in Saccharomyces cerevisiae. Precursor form of mitochondrial superoxide dismutase made in the cytoplasm. J. Biol. Chem. 257: 2713-2718. Babbs, C F . and Steiner, M.G. (1990) Simulation of free radical reactions in biology and medicine: a new twocompartment kinetic model of intracellular lipid peroxidation. Free Radic. Biol. Med. 8: 471-485. Babior, B.M. (1982) The enzymatic basis for superoxide production by human neutrophils. Can. J. Physiol. Pharmacol. 60: 1353-1358. Bach, A.W.J., Lan, N . C , Johnson, D.L., Abell, C W . , Bembenek, M.E., Kwan, S.W., Seeburg, P. and Shih, J . C (1988) cDNA cloning of human liver monoamine oxidase A and B: molecular basis of differences in enzymatic properties. Proc. Natl. Acad. Sci. USA 85: 4934-4938. Bacon, B.R. and Britton, R.S. (1989) Hepatic injury in chronic iron overload. Role of lipid peroxidation. Chem. Biol. Interact. 70: 183-226. Badiali De Giorgi, L., Bonvicini, F., Bianchi, D., Bossoni, G. and Laschi, R. (1987) Ultrastructural aspects of ageing rat hippocampus and effects of L-acetyl-carnitine treatment. Drugs Exp. Clin. Res. 13: 185-189. Bahl, J.J. and Bressler, R. (1987) The pharmacology of carnitine. Annu. Rev. Pharmacol. Toxicol. 27: 257-277. Bahmre, S., Anandatheerthavarada, H.K., Shankar, S. K. and

M.E. Gotz et al. Ravindranath, V. (1992) Microsomal cytochrome P450 in human brain regions. Biochem. Pharmacol. 44: 1223-1225. Baimbridge, K.G., Celio, M.R. and Rogers, J.H. (1992) Calcium-binding proteins in the nervous system. Trends NeuroscL 15: 303-313. Baker, G.B. and Reynolds, G.P. (1989) Biogenic amines and metabolites in Alzheimer's disease: noradrenaline, 5hydroxytryptamine and 5-hydroxyindole-3-acetic acid depleted in hippocampus but not in substantia innominata. Neurosci. Lett. 100: 335-339. Bakker, M.H.M. and Foster, A.C. (1991) An investigation of the mechanism of delayed neurodegeneration caused by direct injection of quinolinate into the rat striatum in vivo. Neuroscience 42: 387-395. Baldessarini, R.J., Kula, N.S., Francoeur, D. and Finklestein, S.P. (1986) Antioxidants tail to inhibit depletion of striatal dopamine by MPTP. Neurology 36: 735. Ball, P., Knuppen, R., Haupt, M. and Breuer, H. (1972) Interactions between estrogens and catecholamines. III. Studies on the methylation of catecholestrogens, catecholamines and other catechols by the catechol-0-methyltransferase of human liver. / . Clin. Endocrin. Metab. 34: 736-746. Barany, M., Chang, Y.C., Arus, C , Rustan, T. and Frey, W.H. (1985) Increased phosphoryl-3-choline in post-mortem Alzheimer's brain. Lancet i: 517. Barbeau, A. (1984) Manganese and extrapyramidal disorders. Neurotoxicology 5: 13-36. Barbeau, A., Roy, M., Cloutier, T., Plasse, L. and Paris, S. (1986) Environmental and genetic factors in the etiology of Parkinson's disease. Adv. Neurol. 45: 299-306. Barden, H. (1969) The histochemical relationship of neuromelanin and lipofuscin. J. Neuropathol. Exp. Neurol. 28: 4 1 9 441. Barnard, E.A. (1992) Signal transduction crosstalk. Trends Biol. Sci. 17: 367-368. Bartus, R.T., Dean, R.D., Beer, B. and Lippa, A.S. (1982) The cholinergic hypothesis of geriatric memory dysfunction. Science 217: 4 0 8 ^ 1 7 . Bassi, S., Albizzati, M.G. and Sbacchi, M. (1984) Double blind evaluation of monosialoganglioside (GM) therapy in stroke. / . Neurosci. Res. 12: 4 9 3 ^ 9 8 . Bast, A. and Haenen, G.R.M.M. (1984) Cytochrome P-450 and glutathione: what is the significance of their interrelationship in lipid peroxidation? Trends Biol. Sci. 9: 510-513. Baumgarten, H.G. and Zimmermann, B. (1992) Cellular and subcellular targets of neurotoxins: the concept of selective vulnerability. In: H. Herken and F. Hucho (Eds.), Handbook of Experimental Pharmacology, Vol. 102, Springer, Berlin, pp. 1-22. Beal, M.F. (1992) Does impairment of energy metabolism result in excitotoxic neuronal death in neurodegenerative illness? An«. Neurol. 31: 119-130. Bendich, A., DiApolito, P., Gabriel, E. and Machlin, L.J. (1984) Interaction of dietary vitamin C and vitamin E in guinea pigs immune response to mitogens. J. Nutr. 114: 1588-1593.

481 Bendich, A., Machlin, L.J., Scandurra, O., Burton, G.W. and Wayne, D.N. (1986) The antioxidant role of vitamin C. Adv. Free Radical Biol. Med. 2: 4 1 9 ^ 4 4 . Ben-Shachar, D. and Youdim, M.B.H. (1990) Selectivity of melanized nigro-striatal dopamine neurons to degeneration in Parkinson's disease may depend on iron-melanin interaction. J. Neural Transm. 29: 251-258. Ben-Shachar, D. and Youdim, M.B.H. (1991) Intranigral iron injection induces behavioural and biochemical 'Parkinsonism' in rats. J. Neurochem. 57: 2133-2135. Ben-Shachar, D., Eshel, G., Finberg, J.P.M. and Youdim, M.B.H. (1991a) The iron chelator desferrioxamine (desferal) retards 6-hydroxydopamine-induced degeneration of nigrostriatal dopamine neurons. J. Neurochem. 56: 14411444. Ben-Shachar, D., Riederer, P. and Youdim, M.B.H. (1991b) Iron melanin interaction and lipid peroxidation: implications for Parkinson's disease. J. Neurochem. 57: 16091614. Benzi, G., Pastoris, O. and Villa, R.F. (1988) Changes induced by aging and drug treatment on cerebral enzymatic antioxidant system. Neurochem. Res. 13: 467-478. Benzi, G., Marzatico, F., Pastoris, O. and Villa, R.F. (1990) Influence of oxidative stress on the age-linked alterations of the cerebral glutathione system. J. Neurosci. Res. 26: 120128. Benzi, G., Turti, D., Marzatico, F. and Pastoris, O. (1991) Age-related acute depletion of cerebral glutathione by peroxidative stress. /. Neurosci. Res. 29: 527-532. Benzi, G., Pastoris, O., Marzatico, F., Villa, R.F. Dagani, F. and Curti, D. (1992) The mitochondrial electron transfer alteration as a factor involved in the brain aging. Neurobiol. Aging 13: 361-368. Berger, K., Przedborski, S. and Cadet, J.L. (1991) Retrograde degeneration of nigrostriatal neurons by intrastriatal 6hydroxydopamine injection in rats. Brain Res. Bull. 26: 301-307. Berridge, M.J. (1975) The interaction of cyclic nucleotides and calcium in the control of cellular activity. Adv. Cycl. Nucleotide Res. 6: 1-98. Beyer, R.E. (1990) The participation of coenzyme Q in free radical production and autoxidation. Free Radical Biol. Med. 8: 545-565. Beyer, W., Imlay, J. and Fridovich, I. (1991) Superoxide dismutases. In: W.E. Cohn and K. Moldave (Eds.), Progress in Nucleic Acid Research and Molecular Biology, Vol. 40, Academic Press, New York, pp. 221-253. Beyreuther, K., Bush, A.I., Dyrks, T., Konig, G., Monnig, U., Mulzhaup, G., Prior, R. and Schubert, W. (1991) Mechanisms of amyloid deposition in Alzheimer's disease. Ann. N. Y. Acad. Sci. 640: 129-139. Bieber, L.L. (1988) Carnitine. Annu. Rev. Biochem. 57: 2 6 1 283. Bieber, L.L. and Lewin, M. (1981) Measurement of carnitine and 0-acylcarnitines. Methods Enzymol. 72: 276-287. Bielski, B.H.J, and Cabelli, D.E. (1991) Highlights of current

482 research involving superoxide and perhydroxyl radicals in aqueous solutions. Int. J. Radiat. Biol. 59: 291-319. Bielski, B.H.J, and Richter, H.W. (1975) Some properties of the ascorbate free radical. Ann. N. Y. Acad. Sci. 258: 231237. Bielski, B.H.J., Cabelli, D.E., Arudi, R.L. and Ross, A.B. (1985) Reactivity of H02/(02)" radicals in aqueous solution. J. Phys. Chem. Ref. Data 14: 1041-1100. Billingsley, M., Hanbauer, I. and Kuhn, D. (1985) Role of calmodulin in the regulation of neuronal function. In: A. Lajtha (Ed.), Handbook of Neurochemistry, Vol. 8., Plenum Press, New York, pp. 201-215. Bindoff, L.A., Birch-Machin, M.A., Cartlidge, N.E.F., Parker, Jr., W.D. and Turnbull, D.M. (1991) Respiratory chain abnormalities in skeletal muscle from patients with Parkinson's disease. J. Neurol. Sci. 104: 203-208. Bindoli, A., Rigobello, M.P. and Deeble, D.J. (1992) Biochemical and toxicological properties of the oxidation products of catecholamines. Free Radical Biol. Med. 13: 391-405. Birkmayer, W. and Riederer, P. (1985) Die ParkinsonKrankheit: Biochemie, Klinik, Therapie, 2ndedn., Springer, Vienna. Birkmayer, W., Knoll, J., Riederer, P., Youdim, M.B.H. and Marton, J. (1985) Improvement of life expectancy due to Ldeprenyl addition to Madopar treatment in Parkinson's disease: a long-term study. J. Neural Transm. 64: 113-127. Bjorneboe, A., Bjorneboe, G.-E. and Drevon, C.A. (1989) Absorption, transport and distribution of vitamin E. /. Nutr. 120: 233-242. Blaschko, H., Richter, D. and Schlossmann, H. (1937) The inactivation of adrenaline. J. Physiol. (London) 90: 1-17. Blaustein, M.P. (1988) Calcium transport and buffering in neurons. Trends Neurosci. 11: 465-469. Blennow, K, Wallin, A., Gottfries, C.G., Lekman, A., Karlsson, I., Skoog, I. and Svennerholm, L. (1991) Significance of decreased lumbar CSF levels of HVA and 5-HIAA in Alzheimer's disease. Neurobiol. Aging 13: 107-113. Blum, J. and Fridovich, I. (1985) Inactivation of glutathione peroxidase by superoxide radical. Arch. Biochem. Biophys. 240: 500-508. Blusztajn, J.K., Gonzalez-Coviella, I.L., Logue, M., Growdon, J.H. and Wurtman, R.J. (1990) Levels of phospholipid catabolic intermediates, glycerophosphocholine and glycerophosphoethanolamine, are elevated in brains of Alzheimer's disease but not of Down's syndrome patients. Brain Res. 536: 240-244. Bodis-Wollner, I., Chung, E., Ghilardi, M.F., Glover, A., Onofrj, M., Pasik, P. and Samson, Y. (1991) Acetyl-levocarnitine protects against MPTP-induced parkinsonism in primates. J. Neural Transm. (P-D Sect.) 3: 63-72. Boiler, F. (1985) Parkinson's disease and Alzheimer's disease: are they associated? In: J.T. Hutton and A.D. Kenny (Eds.), Senile Dementia of the Alzheimer Type, A.R. Liss, New York, pp. 119-129. Bors, W., Michel, C. and Saran, M. (1979) Superoxide anions

Oxidative stress: free radical production in neural degeneration do not react with hydroperoxides. FEBS Lett. 107: 403406. Boveris, A. and Chance, B. (1973) The mitochondrial generation of hydrogen peroxide. General properties and effect of hyperbaric oxygen. Biochem. J. 134: 707-716. Boveris, A., Oshino, N. and Chance, B. (1972) The cellular production of hydrogen peroxide. Biochem. J. 128: 617631. Boveris, A., Cadenas, E. and Stoppani, A.O.M. (1976) Role of ubiquinone in the mitochondrial generation of hydrogen peroxide. Biochem. J. 156: 435-444. Bowen, D.M. and Davison, A.N. (1986) Biochemical studies of nerve cells and energy metabolism in Alzheimer's disease. Br. Med. Bull. 42: 75-80. Bowen, D.M., Steele, J.E., Lowe, S.L. and Palmer, A.M. (1990) Tacrine in relation to amino acid transmitters in Alzheimer's disease. In: R. J. Wurtman, S. Corkin, J.H. Growdon and E. Ritter-Walker (Eds.), Advances in Neurology, Alzheimer's Disease. Vol. 51, Raven Press, New York, pp. 91-102. Bowman, B.A.B. (1992) Acetyl-carnitine and Alzheimer's disease. Nutr. Rev. 50: 142-144. Bkaak, H. and Braak, E. (1991) Neuropathological staging of Alzheimer-related changes. Acta Neuropathol. 82: 239259. Braak, H. and Braak, E. (1993) Anatomy of the human basal ganglia. In: I. Szelenyi (Ed.), Inhibitors of Monoamine Oxidase B, Pharmacology and Clinical Use in Neurodegenerative Disorders, Birkhauser, Basel, pp. 3-23. Brannan, T.S., Maker, H.S., Raes, I. and Weiss, C. (1980a) Regional distribution of glutathione, reductase in the adult rat brain. Brain Res. 200: 474-477. Brannan, T.S., Maker, H.S., Weiss, C. and Cohen, G. (1980b) Regional distribution of glutathione peroxidase in the adult rat brain. J. Neurochem. 35: 1013-1014 Brannan, T.S., Maker, H.S. and Raes, LP. (1981) Regional distribution of catalase in the adult rat brain. J. Neurochem. 36: 307-309. Braughler, J.M. and Hall, E.D. (1989) Central nervous system trauma and stroke. I. Biochemical considerations for oxygen radical formation and lipid peroxidation. Free Radical Biol. Med. 6: 2S9-301. Braughler, J.M. and Pregenzer, J.F. (1989) The 21aminosteroid inhibitors of lipid peroxidation reactions with lipid peroxy and phenoxyl radicals. Free Radical Biol. Med. 7: 125-130. Braughler, J.M., Pregenzer, J.F., Chase, R.L., Duncan, L.A., Jacobsen, E.J. and McCall, J.M. (1987) Novel 21aminosteroids as potent inhibitors of iron-dependent lipid peroxidation./. Biol. Chem. 262: 10438-10440. Bravi, D., Anderson, J.J., Dagani, F., Davis, T.L., Ferrari, R., Gillespie, M. and Chase, T.N. (1992) Effect of aging and dopaminomimetic therapy on mitochondrial respiratory function in Parkinson's disease. Mov. Disord. 7: 228-231. Bridges, R.J., Hatalski, C.G., Shim, S.N., Cummings, B.J., Vijayan, V., Kundi, A. and Cotman, C.W. (1992) Gliotoxic

M.E. Gotz et al. actions of excitatory amino acids. Neuropharmacology 31: 899-907. Brizzee, K.R. and Ordy, J.M. (1979) Age pigments, cell loss and hippocampal function. Mech. Ageing Dev. 9: 143-162. Brogn, R.P.M. and Stoof, J.C. (1990) The quinolinic acid hypothesis in Huntington's chorea. /. Neurol. Sci. 95: 2 9 38. Brun, A., Gustafson, L. and Englund, E. (1990) Subcortical pathology of Alzheimer's disease. In: R.J. Wurtman, S. Corkin, J.H. Growdon and E. Ritter-Walker (Eds.), Advances in Neurology, Alzheimer's Disease, Vol. 51, Raven Press, New York, pp. 73—77. Brunk, U. and Ericsson, J.L.E. (1972) Electron microscopical studies on rat brain neurons. Localization of acid phosphatase and mode of formation of lipofuscin bodies. J. Ultras true t. Res. 38: 1-15. Brunk, U.T., Jones, C.B. and Sohal, R.S. (1992) A novel hypothesis of lipofuscinogenesis and cellular aging based on interactions between oxidative stress and autophagocytosis. Mutat. Res. 275: 3 9 5 ^ 0 3 . Buettner, G.R. (1987) Spin trapping: ESR parameters of spin adducts. Free Radical Biol. Med. 3: 259-303. Bull Zeller, C. and Marchase, R.B. (1992) Gangliosides as modulator of cell function. Am. J. Physiol. 262 (Cell Physiol. 31): C1341-C1355. Burk, R.F. and Ludden, T.M. (1989) Exhaled alkanes as indices of in vivo lipid peroxidation. Biochem. Pharmacol. 38: 1029-1032. Burns, R.S., Chiueh, C.C, Markey, S.P., Eberth, M.H., Jacobowitz, D.M. and Kopin, I.J. (1983) A primate model of parkinsonism: selective destruction of dopaminergic neurons in the pars compacta of the substantia nigra by A^methyl-4-phenyl-l,2,3,6-tetrahydropyridine. Proc. Natl. Acad. Sci. USA 80: 4546-4550. Burton, J.R. and Ingold, K.U. (1981) Autoxidation of biological molecules. The antioxidant activity of vitamin E and related chain-breaking phenolic antioxidants in vitro. /. Am. Chem. Soc. 103: 6472-6477. Burton, G.W. and Ingold, K.U. (1986) Vitamin E: application of the principles of physical organic chemistry to the exploration of its structure and function. Ace. Chem. Res. 19: 194-201. Burton, G.W., Wronska, U., Stone, L., Foster, D.O. and Ingold, K.U. (1990) Biokinetics of dietary RRR-Silphatocopherol in the male guinea pig at three dietary levels of vitamin C and two levels of vitamin E. Evidence that vitamin C does not 'spare' vitamin E in vivo. Lipids 25: 199210. Cabelli, D.E. and Bielski, B.H.J. (1983) Kinetics and mechanism for the oxidation of ascorbic acidlascorbate by H02/(02)" radicals. A pulse radiolysis and stopped-flow photolysis study. J. Phys. Chem. 87: 1809-1812. Cadet, J.L., Katz, M., Jackson Lewis, V. and Fahn, S. (1989) Vitamin E attenuates the toxic effects of intrastriatal injection of 6-hydroxydopamine (6-OHDA) in rats: behavioral and biochemical evidence. Brain Res. 476: 10-15.

483 Calne, D.B. (1991) Neurotoxins and degeneration in the central nervous system. Neurotoxicology 12: 335-340. Calne, D.B. (1992) The nature of Parkinson's disease. Neurochem. Int. 20 (Suppl.): IS-3S. Calne, D.B. and Eisen, A. (1989) The relationship between Alzheimer's disease, Parkinson's disease and motoneuron disease. Can. J. Neurol. Sci. 16: 547-550. Calne, D.B., Eisen, A., McGeer, E. and Spenser, P. (1986) Alzheimer's disease, Parkinson's disease and motoneuron disease: abiotropic interaction between ageing and environment? Lancerii: 1067-1070. Calvani, M. and Carta, A. (1991) Clues to the mechanism of action of acetyl-L-carnitine in the central nervous system. Dementia 2: 1-6. Candy, J., Oakley, A., Klinowski, J., Carpenter, T., Perry, R., Atack, J., Perry, E., Blessed. G., Fairbrain, A. and Edwardson, J. (1986) Aluminosilicates and senile plaques formation in Alzheimer's disease. Lancet i: 354-357. Canfield, L.M., Forage, J.W. and Valenzuela, J.G. (1992) Carotenoids as cellular antioxidants. Proc. Soc. Exp. Biol. Med. 200: 260-265. Capaldi, R. (1988) Mitochondrial myopathies and respiratory chain proteins. Trends Biochem. Sci. 13: 144-148. Carrillo, M.-C, Kanai, S., Nokubo, M., Ivy, G.O., Sato, Y. and Kitani, K. (1992a) (-)Deprenyl increases activities of superoxide dismutase and catalase in striatum but not in hippocampus: the sex and age-related differences in the optimal dose in the rat. Exp. Neurol. 116: 286-294. Carrillo, M.-C, Kitani, K, Kanai, S., Sato, Y. and Ivy, G.O. (1992b) The ability of (-)deprenyl to increase superoxide dismutase activities in the rat is tissue and brain region selective. Life Sci. 50: 1985-1992. Carstam, R., Brink, C , Hindemith-Augustsson, A., Rorsman, H. and Rosengren, E. (1991) The neuromelanin of the human substantia nigra. Biochim. Biophys. Acta 1097: 152160. Carvalho, A.P. (1982) Calcium in the nerve cell. In: A. Lajtha (Ed.), Handbook of Neurochemistry, Chemical and Cellular Architecture, Vol. 1, 2nd edn., Plenum Press, New York, pp. 69-115. Cash, R., Ruberg, M., Raisman, M. and Agid, Y. (1984) Adrenergic receptors in Parkinson's disease. Brain Res. 322: 269-275. Cathcart, R., Schwiers, E., Saul, R.L. and Ames, B.N. (1984) Thymine glycol and thymidine glycol in human and rat urine: a possible assay for oxidative damage. Proc. Natl. Acad. Sci. USA Si: 5633-5637. Ceballos, I., Lafon, M., Javoy-Agid, F., Hirsch, E.C., Nicole, A., Sinet, P.M. and Agid, Y. (1990) Superoxide dismutase and Parkinson's disease. Lancet 335: 1035-1036. Ceballos, I., Javoy-Agid, F., Delacourte, A., Defossez, A., Lafon, M., Hirsch, E.C., Nicole, A., Sinet, P.M. and Agid, Y. (1991) Neuronal localization of copper-zinc superoxide dismutase protein and mRNA within the human hippocampus from control and Alzheimer's disease brains. Free Radical Res. Commun. 12/13: 571-580.

484 Ceballos-Baumann, A.O., Von Kummer, R., Eckert, W. and Weicker, H. (1990) Controlled-release L-dopa^ense^azide (Madopar HBS): clinical observations and L-dopa and dopamine plasma concentrations in fluctuating parkinsonian patients. J. Neurol. 237: 24-28. Ceballos-Picot. I., Nicole, A., Briand, P., Grimber, G., Delacourte, A., DeFossez, A., Javoy-Agid, F., Lafon, M., Blouin, J.L. and Sinet, P.M. (1991) Neuronal-specific expression of human copper-zinc superoxide dismutase gene in transgenic mice: animal model of gene dosage effects in Down's syndrome. Brain Res. 552: 198-214. Cerutti, P.A. (1991) Oxidant stress and carcinogenesis. Eur. J. Clin. Invest. 21: 1-5. Cha, J.J., Dure, L.S., Sakurai, S.Y., Penney, J.B. and Young, A.B. (1991) 3,4,6-Trihydroxyphenylalanine (6-hydroxyDOPA) displaces [^H]AMPA binding in rat striatum. Neurosci. Lett. 132: 55-58. Chafi, A.H., Hauw, J.-J., Rancurel, G., Berry, J.-P. and Galle, C. (1991) Absence of aluminium in Alzheimer's disease brain tissue: electron microprobe and ion microprobe studies. Neurosci. Lett. 123: 61-64. Chance, B., Sies, H. and Boveris, A. (1979) Hydroperoxide metabolism in mammalian organs. Physiol. Rev. 59: 527605. Chiba, K., Trevor, A.J. and Castagnoli, N. (1985) Active uptake Of MPP"*", a metabolite of MPTP, by brain synaptosomes. Biochem. Biophys. Res. Commun. 128: 12281232. Chiueh, C.C., Krishna, G., Tulsi, P., Obata, T., Lang, K., Huang, S.J. and Murphy, D.L. (1992) Intracranial microdialysis of. salicylic acid to detect hydroxyl radical generation through dopamine autooxidation in the caudate nucleus-effects of MPP+. Free Radical Biol. Med 13: 581-583. Choi, D.W. (1988a) Calcium-mediated neurotoxicity: relationship to specific channel types and role in ischemic damage. Trends Neurosci. 11: 465-469. Choi, D.W. (1988b) Glutamate neurotoxicity and diseases of the nervous system. Neuron 1: 623-634. Choi, J.-H. and Yu, B.P. (1990) Unsuitability of TBA test as a lipid peroxidation marker due to prostaglandin synthesis in the aging kidney. Age 13: 61-64. Chow, C.K. (1988) Interrelationship of cellular antioxidant defense systems. In: C.K. Chow (Ed.), Cellular Antioxidant Defense Mechanisms. Vol. 2, CRC Press, Boca Raton, PL, pp. 217-237. Chow, C.K. (1991) Vitamin E and oxidative stress. Free Radical Biol. Med. 11: 215-232. Chow, C.K. and Gairola, C. (1984) Influence of dietary vitamin E and selenium on metabolic activation of chemicals to mutagens. J. Agric. Food Chem. 32: 443-447. Chrisp, P., Mammen, G.J. and Sorkin, M. (1991) Selegiline: a review of its pharmacology, symptomatic benefits and protective potential in Parkinson's disease. Drugs Aging 1: 228-248. Ciuffi, M., Gentilini, G., Franchi-Micheli, S. and Zilletti, L. (1992) D-Penicillamine affects lipid peroxidation and iron

Oxidative stress: free radical production in neural degeneration content in the rat brain cortex. Neurochem Res. 17: 12411246. elevens, R.A. and Beal, M.F. (1989) Substance P-like immunoreactivity in brains with pathological features of Parkinson's and Alzheimer's disease. Brain Res. 486: 387390. Cochrane, C.G., Schraufstatter, I.U., Hyslop, P. and Jackson, J. (1987) Cellular biochemical events in oxidant injury. In: B. Halliwell (Ed.), Oxygen Radicals and Tissue Injury, Proceedings of a Brook Lodge Symposium, Augusta, MI, Federation of American Societies For Experimental Biology, Bethesda, MD, pp. 49-54. Cohen, G. (1983a) The pathobiology of Parkinson's disease: biochemical aspects of dopamine neuron senescence. J. Neural Transm. 19 (Suppl.): 89-103. Cohen, G. (1983b) Catalase, glutathione peroxidase, superoxide dismutase and cytochrome P-450. In: A. Lajtha (Ed.), Handbook of Neurochemistry, Enzymes in the Nervous System, Vol. 4, Plenum Press, New York, pp. 315-329. Cohen, G. and Hochstein, P. (1963) Glutathione peroxidase: the primary agent for the elimination of hydrogen peroxide in erythrocytes. Biochemistry 2: 1420-1428. Cohen, G. and Spina, M.B. (1989) Deprenyl suppresses the oxidant stress associated with increased dopamine turnover. Ann. Neurol. 26: 689-690. Cohen, N., Scott, C.G., Neukom, R.J., Lopresti, G., Weber, G. and Saucy, G. (1981) Total synthesis of all eight stereoisomers of alpha-tocopheryl acetate. Determination of their diastereoisomeric and enantiomeric purity by gas chromatography. Helv. Chim. Acta 64: 1158-1173. Coleman, R.J. (1992) Current drug therapy for Parkinson's disease. A review. Drugs Aging 2: 112-124. Collingridge, G.L. and Singer, W. (1991) Excitatory amino acid receptors and synaptic plasticity. In: D. Lodge and G.L. Collingridge (Eds.), Trends in Pharmacological Science, The Pharmacology of Excitatory Amino acids. Special Report, Elsevier, Amsterdam, pp. 24-48. CoUins, M.A., Ung-Chun, N., Cheng, B.Y. and Pronger, D. (1990) Brain and plasma tetrahydroisoquinolines in rats: effects of chronic ethanol intake and diet. /. Neurochem. 55: 1507-1514. Collins, M.A., Neafsey, E.J., Matsubara, K., Cobuzzi, Jr., R.J. and Rollema, H. (1992) Indole-N-methylated betacarbolinium ions as potential brain-bioactivated neurotoxins. Brain Res. 570: 154-160. Collins, R.C. (1987) Neurotoxins and the selective vulnerability of brain. In: P. Jenner (Ed.), Neurotoxins and Their Pharmacological Implications, Raven Press, New York, pp. 1-17. Connick, J.H., Stone, T.W., Carla, V. and Moroni, F. (1988) Increased kynurenic acid levels in Huntington's disease. Lancet ii: 1373. Connor, J.R., Menzies, S.L., St. Martin, S.M. and Mufson, E.J. (1992a) A histochemical study of iron, transferrin and ferritin in Alzheimer's diseased brains. J. Neurosci. Res. 31: 75-83.

M.E. Gotz et al. Connor, J.R., Snyder, B.S., Beard, J.L., Fine, R.E. and Mufson, E.J. (1992b) Regional distribution of iron and ironregulatory proteins in the brain in aging and Alzheimer's disease. /. Neurosci. Res. 31: 327-335. Cooper, J. K. (1991) Drug treatment of Alzheimer's disease. Arch. Int. Med. 151: 245-249. Cooper, A.J.L., Pulsinelli, W.A. and Duffy, I.E. (1980) Glutathione and ascorbate during ischemia and postischemic reperfusion in rat brain. J. Neurochem. 35: 1242-1245. Corongiu, P.P., Banni, S. and Dessi, M.A. (1989) Conjugated dienes detected in tissue hpid extracts by second derivative spectrophotometry. Free Radical Biol. Med. 7: 183-186. Crapper, D.R., Quittkat, S., Krishnan, S.S., Dalton, A.J. and De Boni, U. (1980) Intranuclear aluminium content in Alzheimer's disease, dialysis encephalopathy and experimental aluminium encephalopathy. Acta Neuropathol. 50: 1924. Crapper McLachlan, D.R.C., Kruck, T.P., Lukiw, W.J. and Krishnan, S.S. (1991) Would decreased aluminium ingestion reduce the incidence of Alzheimer's disease? Can. Med. Assoc. J. 145: 793-804. Craw, M.T. and Depew, M.C. (1985) Contributions of electron spin resonance spectroscopy to the study of vitamins C, E and K. Rev. Chem. Intermed. 6: 1-31. Croft, S., Gilbert, B.C., Lindsay Smith, J.R. and Whitwood, A.C. (1992) An E.S.R. investigation of the reactive intermediate generated in the reaction between Fe° and H2O2 in aqueous solution. Direct evidence for the formation of the hydroxyl radical. Free Radical Res. Commun. 17: 21-39. Crowe, A. and Morgan, E.H. (1992) Iron and transferrin uptake, by brain and cerebrospinal fluid in the rat. Brain Res. 592: 8-16. Curzio, M. (1988) Interaction between neutrophils and 4hydroxyalkenals and consequences on neutrophil motility. Free Radic. Res. Commun. 5: 55-66. Curzio, M., Esterbauer, H.. Di Mauro, C. and Dianzani, M.U. (1990) Influence of the lipid peroxidation product 4hydroxynonenal on human neutrophil migration. Int. J. Tissue React. 6: 13-18. Dale, G.E., Probst, A., Luthert, P., Martin, J., Anderton, B.H. and Leigh, P.N. (1992) Relationship between Lewy bodies and pale bodies in Parkinson's disease. Acta Neuropathol. 83: 525-529. D'Amato, R.J., Zweig, R.M., Whitehouse, P.J., Wenk, G.L., Singer, H., Mayeux, R., Price, D.L. and Snyder, S.H. (1987) Aminergic systems in Alzheimer's disease and Parkinson's disease. Ann. Neurol. 22: 229-236. Dameron, C.T. and Harris, E.D. (1987a) Regulation of aortic CuZn-superoxide dismutase with copper. Biochem. J. 248: 661-668. Dameron, C.T. and Harris. E.D. (1987b) Regulation of aortic CuZn-superoxide dismutase with copper. Ceruloplasmin and albumin re-activate and transfer copper to the enzyme in culture. Biochem. J. 248: 669-675. Danielczyk, W. (1980) Die Mono- und Kombinationstherapie des Parkinson-Syndroms mit Amantadinen. In: P.A. Fischer

485 (Ed.), Parkinson-Syndrom: Kombinations- und Begleittherapien, Schattauer, Stuttgart, pp. 125-136. Das, K.C., Abramson, M.B. and Katzman, R. (1978) Neuronal pigments: spectroscopic characterization of human brain melanin. /. Neurochem. 30: 601-605. Das, M., Seth, P.K. and Mukhtar, H.J. (1981) Characterization of microsomal aryl hydrocarbon hydroxylase of rat brain. J. Pharmacol. Exp. Ther. 216: 156-161. Davenport, A. and Goodall, R. (1992) Aluminum and dementia. Lancet i: 1236. Davies, K.J.A. (1988) Protein oxidation, protein cross-linking, and proteolysis in the formation of lipofuscin. In: J. Nagy (Ed.), Lipofuscin-1987: State of the Art, Excerpta Medica, Amsterdam, pp. 109-133. Davies, K.J.A. and Goldberg, A.L. (1987) Proteins damaged by oxygen radicals are rapidly degraded in extracts of red blood cells. J. Biol. Chem. 262: 8227-8234. Davies, M.J., Gilbert, B.C., Stell, J.K. and Whitwood, A.C. (1992) Nucleophilic substitution reactions of spin adducts. Implications for the correct identification of reaction intermediates by EPR/spin trapping. /. Chem. Soc. Perkin Trans. 2: 333-335. Dedman, D.J., Treffry, A., Candy, J.M., Taylor, G.A.A., Morris, C M . , Bloxham, C.A., Perry, R.H., Edwardson, J.A. and Harrison, P.M. (1992) Iron and aluminium in relation to brain ferritin in normal individuals and Alzheimer's disease and chronic renal-dialysis patients. Biochem. J. 287: 5 0 9 514. De Haan, J.B., Newman, D. and Kola, I. (1992) Cu/Zn superoxide dismutase mRNA and enzyme activity, and susceptibility to lipid peroxidation, increases with aging in murine brains. Mo/. Brain Res. 13: 179-187. De La Monte, S.M., Wells, S.E., Hedley Whyte, T. and Growdon, J.H. (1989) Neuropathological distinction between Parkinson's dementia and Parkinson's plus Alzheimer's disease. Ann. Neurol. 26: 309-320. Delis, D., Direnfeld, L., Alexander, M.P. and Kaplan, E. (1982) Cognitive fluctuations associated with on-off phenomenon in Parkinson's disease. Neurology 32: 1049-1052. De Marchi, W.J. and Johnston, G.A.R. ( 1969) The oxidation of glycine by D-amino acid oxidase in extracts of mammalian central nervous tissue. J. Neurochem. 16: 355-361 Denney, R.M. and Denney, C.B. (1985) An update on the identity crisis of monoamine oxidase: new and old evidence for the independence of MAO-A and -B. Pharmacol. Ther. 30: 227-259. De Pierre, J.W. and Ernster, L. (1977) Enzyme topology of intracellular membranes. Annu. Rev . Biochem. 46: 2 0 1 262. De Simone, R., Ramacci, M.T. and Aloe, L. (1991) Effect of acetyl-L-carnitine on forebrain cholinergic neurons of developing rats. Int. J. Dev. Neurosci. 9: 3 9 ^ 6 . De Vera, N., Artigas, F., Serratosa, J. and Martinez, E. (1991) Changes in polyamine levels in rat brain after systemic kainic acid administration: relationship to convulsant activity and brain damage. J. Neurochem. 57: 1-8.

486 Dexter, D.T., Carter, C.J., Wells, F.R., Javoy-Agid, F., Agid, Y., Lees, A., Jenner, P. and Marsden, CD. (1989a) Basal lipid peroxidation in substantia nigra is increased in Parkinson's disease. J. Neurochem. 52: 381-389. Dexter, D.T., Wells, F.R., Lees, A.J., Javoy-Agid, F., Agid, Y., Jenner, P. and Marsden, CD. (1989b) Increased nigral iron content and alterations in other metal ions occurring in brain in Parkinson's disease. J. Neurochem. 52: 1830-1836. Dexter, D.T., Carayon, A., Vidailhet, M., Ruberg, M., Agid, F., Agid, Y., Lees, A.J., Wells, F.R., Jenner, P. and Marsden, CD. (1990) Decreased ferritin levels in brain in Parkinson's disease. J. Neurochem. 55: 1 6-20. Dexter, D.T., Carayon, A., Javoy-Agid, F., Agid, Y., Wells, F.R., Daniel, S.E., Lees, A.J., Jenner, P. and Marsden, CD. (1991) Alterations in the levels of iron, ferritin, and other trace metals in Parkinson's disease and other neurodegenerative diseases affecting the basal ganglia. Brain 114: 1953-1975. Dexter, D.T., Jenner, P., Schapira, A.H.V. and Marsden, CD. (1992a) Alterations in levels of iron, ferritin, and other trace metals in neurodegenerative diseases affecting the basal ganglia. Ann. Neurol. 32 (Suppl.): 94-100. Dexter, D.T., Ward, R.J., Wells, F.R., Daniel, S.E., Lees, A.J., Peters, T.J., Jenner, P. and Marsden, CD. (1992b) Alphatocopherol levels in brain are not altered in Parkinson's disease. Ann. Neurol. 32: 591-593. Dhopeshwarkar, G.A. and Mead, J.F. (1973) Uptake and transport of fatty acids into the brain and the role of the blood-brain barrier system. Adv. Lipid Res. 11: 109-142. Difazio, M.C, Hollingsworth, Z., Young, A. and Penney, J.B. (1992) Glutamate receptors in the substantia nigra of Parkinson's disease brains. Neurology 42: 402-406. DiFigila, M. (1990) Excitotoxic injury of the neostriatum: a model for Huntington's disease. Trends Neurosci. 13: 286289. Diliberto, Jr., E., Dean, G., Carter, C and Allen, P.L. (1982) Tissue subcellular and submitochondrial distributions of semidehydroascorbate reductase: possible role of semidehydroascorbate reductase in cofactor regeneration. J. Neurochem. 39: 563-568. Doha, T., Burton, G.W. and Ingold, K.U. (1985) Antioxidant and co-antioxidant activity of vitamin C The effect of vitamin C, either alone or in the presence of vitamin E or a water-soluble vitamin E analogue, upon the peroxidation of aqueous multilamellar phospholipid liposomes. Biochem. Biophys. Acta 835: 298-303. Doetsch, P.W., Helland, D.E. and Haseltine, W.A. (1986) Mechanism of action of a mammalian DNA repair endonuclease. Biochemistry 25: 2212-2220. Doetsch, P.W., Henner, W.D., Cunningham, R.P., Toney, J.H. and Helland, D.E. (1987) A highly conserved endonuclease activity present in Escherichia coli, bovine, and human cells recognizes oxidative DNA damage at sites of pyrimidines. Mol. Cell. Biol. 7: 26-32. Dolezal, V. and Tucek, S. (1981) Utilization of citrate, acetylcarnitine, acetate, pyruvate and glucose for the synthesis of

Oxidative stress: free radical production in neural degeneration acetylcholine in rat brain slices. J. Neurochem. 36: 13231330. Dooling, E.C, Schoene, W.C and Richardson, Jr., E.P. (1974) Hallervorden-Spatz syndrome. Arch. Neurol. 30: 70-83. Dostert, P., Strolin-Benedetti, M. and Tipton, K.F. (1989) Interactions of monoamine oxidase with substrates and inhibitors. Med. Res. Rev. 9: 45-89. Dowson, J.H., Wilton-Cox, H., Cairns, M.R. and Ramacci, M.T. (1992) The morphology of lipopigment in rat Purkinje neurons after chronic acetyl-L-carnitine administration: a reduction in aging-related changes. Biol. Psychiatry 32: 179-187. Drevon, CA. (1991) Absorption, transport and metabohsm of vitamin E. Free Radical Res. Commun. 14: 229-246. Driscoll, P.F., Larsen, P.D. and Gruber, A.B. (1987) MELAS syndrome involving a mother and two children. Arch. Neurol. 44: 971-973. Dubois, B., Pillon, B. and Agid, Y. (1992) Deterioration of dopaminergic pathways and alterations in cognition and motor function. J. Neurol. 239 (Suppl.): S9-S12. Duckett, S. and Galle, P. (1976) Mise en evidence de r aluminium dans les plaques seniles de la maladie d'Alzheimer, etude a la microsonde de Castaing. C.R. Acad. Sci. 282: 393-395. Duckett, S., Galle, P. and Fiori, C (1985) Electron probe microanalysis of normal and pathological neuronal tissue with wave length dispersive X ray spectrometry. In: H.H.O. Gabay (Ed.), Metal Ions in Neurology and Psychiatry, A.R. Liss, New York, pp. 367-396. Duncan, M.W., Steele, J.C, Kopin, I.J. and Markey, S.P. (1990) 2-Amino-3-(methylamino)-propanoic acid (BMAA) in cycad flour: an unlikely cause of amyotrophic lateral sclerosis and parkinsonism-dementia of Guam. Neurology 40: 767-772. Duvoisin, R.C and Johnson, W.G. (1992) Hereditary Lewybody parkinsonism and evidence for a genetic etiology of Parkinson's disease. Brain Pathol. 2: 309-320. Dykens, J.A., Stern, A. and Trenkner, E. (1987) Mechanism of kainate toxicity to cerebellar neurons in vitro is analogous to reperfusion injury. J. Neurochem. 49: 1222-1228. Dyrks, T., Dyrks, E., Hartmann, T., Masters, C and Beyreuther, K. (1992) Amyloidogenicity of beta A4bearing amyloid protein precursor fragment by metalcatalysed oxidation. 7. Biol. Chem. 267: 18210-18217. Egan, M.F., Hyde, T.M., Albers, G.W., Elkashef, A., Alexander, R.C, Reeve, A., Blum, A., Saenz, R.E. and Wyatt, R.J. (1992) Treatment of tardive dyskinesia with vitamin E. Am. J. Psychiatry 149:11?>-111. Ehringer, H. and Hornykiewicz, O. (1960) Verteilung von Noradrenalin und Dopamin (3-Hydroxytyramin) im Gehirn des Menschen und ihr Verhalten bei Erkrankungen des extrapyramidalen Systems. Clin. Invest. 24: 1236-1239. Eisen, A. and Calne, D. (1992) Amyotrophic lateral sclerosis, Parkinson's disease and Alzheimer's disease: phylogenetic disorders of the human neocortex sharing many characteristics. Can. J. Neurol. Sci. 19: 117-120.

M.E. Gotz et al. Ekblom, J., Jossan, S.S., Gillberg, P.G., Oreland, L. and Aquilonius, S.M. (1992) Monoamine oxidase-B in motor cortex: changes in amyotrophic lateral sclerosis. Neuroscience 49: 763-769. Elstner, E.F. (1990) Der Sauerstojf: Biochemie, Biologic, Medizin, Bibliographisches Institut & F.A. Brockhaus AG, Mannheim, pp. 272-273. Emson, P.C. and Lindvall, O. (1986) Neuroanatomical aspects of neurotransmitters affected in Alzheimer's disease. In: M. Roth and L.L. Iversen (Eds.), Alzheimer's Disease and Related Disorders, British Medical Bulletin, London, pp. 5 7 62. Eppelbaum, J., Ruberg, M., Moyse, M., Javoy-Agid, F., Dubois, E. and Agid, Y. (1983) Somatostatin and dementia in Parkinson's disease. Brain Res. 278: 376-379. Ernster, L., Forsmark, P. and Nordenbrand, K. (1992) The mode of action of lipid soluble antioxidants in biological membranes: relationship between the effects of ubiquinol and vitamin E as inhibitors of lipid peroxidation in submitochondrial particles. Bio Factors 3: 241-248. Esch, F.S., Keim, P.S., Beattie, E.G., Blacher, R.W., Culwell, A.R., Oltersdorf, T., McClure, D. and Ward, P.J. (1990) Cleavage of amyloid beta peptide during constitutive processing of its precursor. Science 248: 1122-1124. Esterbauer, H., Schaur, R.J. and Zollner, H. (1991) Chemistry and biochemistry of 4-hydroxynonenal, malondialdehyde and related aldehydes. Free Radical Biol. Med. 11: 8 1 128. Etienne, P., Robitaille, Y., Wood, P., Gauthier, S., Nair, N.P.V. and Quirion, R. (1986) Nucleus basalis neuronal loss, neuritic plaques and choline acetyltransferase activity in advanced Alzheimer's disease. Neuroscience 19: 12791291. Fahn, S. (1989) The history of Parkinsonism. Mov. Disord. 4 (Suppl. 1): 2-10. Fahn, S. (1992) A pilot trial of high-dose alpha-tocopherol and ascorbate in early Parkinson's disease. Ann. Neurol. 32 (Suppl.): 128-132. Farooqui, A.A. and Horrocks, L.A. (1991) Excitatory amino acid receptors, neural membrane phospholipid metabolism and neurological disorders. Brain Res. Rev. 16: 171-191. Farooqui, A.A., Liss, L. and Horrocks, L.A. (1988a) Neurochemical aspects of Alzheimer's disease: involvement of membrane phospholipids. Metab. Brain Res. 3: 19-35. Farooqui, A.A., Liss, L. and Horrocks, L.A. (1988b) Stimulation of lipolytic enzymes in Alzheimer's disease. Ann. Neurol. 23: 306-308. Farooqui, A.A., Liss, L. and Horrocks, L.A. (1990) Elevated activities of lipases and lysophospholipases in Alzheimer's disease. Dementia 1: 208-214. Fatemi, S.J.A., Kadir, F.H.A., Williamson, D.J. and Moore, G.R. (1991) The uptake, storage, and mobilization of iron and aluminium in biology. Adv. Inorg. Chem. 36: 409-448. Fawthrop, D.J., Boobis, A.R. and Davies, D.S. (1991) Mechanisms of cell death. Arch. Toxicol. 65: 437-444. Fazzini, E., Durso, R., Davoudi, H., Szabo, G.K. and Albert,

487 M.L. (1990) GMI gangliosides alter acute MPTP-induced behavioral and neurochemical toxicity in mice. J. Neurol. Sci. 99: 59-68. Fearnley, J.M. and Lees, A.J. (1991) Aging and Parkinson's disease: substantia nigra regional selectivity. Brain 114: 2283-2301. Felten, D.L., Felten, S.Y., Steece-Collier, K., Date, I. and Clemens, J.A. (1992) Age-related decline in the dopaminergic nigrostriatal system: the oxidative hypothesis and protective strategies. Ann. Neurol. 32 (Suppl.): 133-136. Fenton, H.J.H. (1894) Oxidation of tartaric acid in the presence of iron. J. Chem. Soc. Trans. 65: 899-903. Fernandez-Calle, P., Molina, J.A„ Jimenez-Jimenez, F.J., Vazquez, A., Pondal, M., Garcia-Ruiz, P.J., Urra, D.G., Domingo, J. and Codoceo, R. (1992) Serum levels of alphatocopherol (vitamin E) in Parkinson's disease. Neurology 42: 1064-1066. Ferris, S.H., Reisberg, B., Crook, T., Friedman, E., Schneck, M.K., Mir, P., Sherman, K.A., Corwin, J., Gershon, S. and Bartus, R.T. (1982) Pharmacological treatment of senile dementia: choline, L-DOPA, piracetam and choline plus piracetam. In: S. Corkin (Ed.), Alzheimer's Disease: A Report of Progress, Raven Press, New York, pp. 4 7 5 ^ 8 1 . Fields, J.Z., Albores, R.R., Neafsey, E.J. and Collins, M.A. (1992) Inhibition of mitochondrial succinate oxidationsimilarities and differences between N-methylated betacarbolines and MPP"*". Arch. Biochem. Biophys. 294: 539543. Finali, G., Piccirilli, M., Oliani, C. and Piccinin, G.L. (1991) L-Deprenyl therapy improves verbal memory in amnesic Alzheimer patients. Clin. Neuropharmacol. 14: 523-536. Finlayson, M.H., Guberman, A. and Martin, J.B. (1973) Cerebral lesions in familial amyotrophic lateral sclerosis and dementia. Acta Neuropathol. 26: 237-246. Fishman, J. and Norton, B. (1975) Catecholestrogen formation in the central nervous system of the rat. Endocrinology 96: 1054-1058. Fleischer, S. and Packer, L. (1978) Methods Enzymology, Vol. 53, Academic Press, New York. Fleming, J.T. and Joshi, J.G. (1987) Ferritin: isolation of aluminum-ferritin complex from brain. Proc. Natl. Acad. Sci. USA 84: 7850-7866. Reming, J.T. and Joshi, J.G. (1991) Ferritin: the role of aluminum in ferritin function. NeurobioL Aging 12: 4 1 3 ^ 1 8 . Flohe, L., Gunzler, W.A. and Schock, H.H. (1973) Glutathione peroxidase: a selenoenzyme. FEBS Lett. 32: 132134. Florence, T.M. (1984) The production of hydroxyl radical from hydrogen peroxide. J. Inorg. Biochem. 22: 221-230. Floyd, R.A., Watson, J.J., Wong, P.K., Miller, D.H. and Richard, R.C. (1986) Hydroxyl free radical adduct of deoxyguanosine: sensitive detection and mechanism of formation. Free Radical Res. Commun. 1: 163-172. Foley, J.M. and Baxter, D. (1958) On the nature of pigment granules in the cells of the locus coeruleus and substantia nigra. J. Neuropathol. Exp. Neurol. 17: 568-598.

488 Forman, H.J. and Boveris, A. (1982) Superoxide radical and hydrogen peroxide in mitochondria. In: W.A. Pryor (Ed.), Free Radicals in Biology. Vol. 5, Academic Press, New York, pp. 65-90. Forno, L.S. (1986) The Lewy body in Parkinson's disease. In: M.D. Yahr and K.J. Bergmann (Eds.), Advances in Neurology, Parkinson's Disease. Vol. 4, Raven Press, New York, pp. 35-43. Fornstedt, B. and Carlsson, A. (1991a) Effects of inhibition of monoamine oxidase on the levels of 5-5-cysteinyl adducts of catechols in dopaminergic regions of the brain of the guinea pig. Neuropharmacology 30: 463-468. Fornstedt, B. and Carlsson, A (1991b) Vitamin C deficiency facilitates 5-5-cysteinyldopamine formation in guinea pig striatum. J. Neurochem. 56: 407^14. Fornstedt, B., Brun, A., Rosengren, E. and Carlsson, A. (1989) The apparent autoxidation rate of catechols in dopamine-rich regions of human brains increases with the degree of depigmentation of substantia nigra. J. Neural. Transm. (P-D Sect.) 1: 279-295. Forsmak, P., Aberg, F., Norling, B., Nordenbrand, K., Dallner, G. and Ernster, L. (1991) Inhibition of lipid peroxidation by ubiquinol in submitochondrial particles in the absence of vitamin E. FEBS Lett. 285: 39-43. Fowler, C.J., Oreland, L., Marcusson, J. and Winblad, B. (1980a) Titration of human brain monoamine oxidase-A and -B by clorgyline and L-deprenyl. NaunynSchmiedeberg 's Arch. Pharmacol. 311: 263-272. Fowler, C.J., Oreland, L., Marcusson, J. and Winblad, B. (1980b) The effect of age of the activity and molecular properties of human brain monoamine oxidase. /. Neural. Transm. 49: 1-20. Francis, A., Pearce, L.B. and Roth, J.A. (1985) Cellular localization of MAO-A and MAO-B in brain: evidence from kainic acid lesions in striatum. Brain Res. 334: 59-64. Frenkel, K. (1992) Carcinogen-mediated oxidant formation and oxidative DNA damage. Pharmacol. Ther. 53: 127166. Fridovich, I. (1975) Superoxide dismutases. Annu. Rev. Biochem. 44: 147-159. Fridovich, I. (1986a) Biological effects of the superoxide radical. Arc/z. Biochem. Biophys. lAl: 1-11. Fridovich, I. (1986b) Superoxide dismutases. Adv. Enzymol. 58: 62-97. Friedland, R.P., Jagust, W.J., Huesman, R.H., Koss, E., Knittel, B., Mathias, C.A., Ober, B.A., Mazoyer, B.M. and Budinger, T.F. (1989) Regional cerebral glucose transport and utilization in Alzheimer's disease. Neurology 39: 14271434. Fritsma, G.A. (1983) Vitamin E and autoxidation. Am. J. Med. Technol. 49: 453-456. Gaffan, D. (1987) Amnesia, personal memory, and the hippocampus: experimental neuropsychological studies in monkeys. In: S.M. Stahl, S.D. Iversen and E.C. Goodman (Eds.), Cognitive Neurochemistry, Oxford University Press, Oxford, pp. 46-56.

Oxidative stress: free radical production in neural degeneration Galloway, P.G., Berceron, C. and Perry, G. (1989) The presence of tau distinguishes Lewy bodies of diffuse Lewy body disease from those of idiopathic Parkinson's disease. Neurosci. Lett. 100: 6-10. Gardner, H.W. (1989) Oxygen radical chemistry of polyunsaturated fatty acids. Free Radical Biol. Med. 7: 65-86. Garnier-Suillerot, A., Tosi, L. and Paniago, E. (1984) Kinetic and mechanism of vesicle lipoperoxide decomposition by Fe(II). Biochim. Biophys. Acta 794: 307-312. Caspar, P. and Gray, F. (1984) Dementia in idiopathic Parkinson's disease. Acta Neuropathol. 64: 43-52. Gaunt, G.L. and De Duve, C. (1976) Subcellular distribution of D-amino acid oxidase and catalase in rat brain. J. Neurochem. 26: 749-759. Gauthier, S. and Amyot, D. (1992) Sustained release antiparkinson agents: controlled release levodopa. Can. J. Neurol. Sci. 19: 153-155. Geddes, J.W., Ulas, J., Brunner, L.C., Choe, W. and Cotman, C.W. (1992) Hippocampal excitatory amino acid receptors in elderly, normal individuals and those with Alzheimer's disease: non-NMDA-receptors. Neuroscience 50: 23-34. Gerlach, M., Gsell, W. and Riederer, P. (1991a) Anatomische, biochemische und funktionelle Strukturen physiologischer Neurotransmitter-Regelkreise. In: H. Beckmann and M. Osterheider (Eds.), Neurotransmitter und psychische Erkrankungen, Springer, Berlin, pp. 1-18. Gerlach, M., Riederer, P., Przuntek, H. and Youdim, M.B.H. (1991b) MPTP mechanisms of neurotoxicity and their implications for Parkinson's disease. Eur. J. Pharmacol. Mol. (Pharmac. Sect.) 208: 273-286. Gerlach, M., Riederer, P. and Youdim, M.B.H. (1992) The molecular pharmacology of L-deprenyl. Eur. J. Pharmacol. Mol. (Pharmac. Sect.) 226: 97-108. Ghez, C. (1991) Voluntary movement. In: E.R. Kandel, J.H. Schwartz and T.M. Jessel (Eds.), Principles of Neural Science,, Elsevier, Amsterdam, pp. 609-624. Gibb, W.R.G. (1989) Neuropathology in movement disorders. J. Neurol. Neurosurg. Psychiatry (Suppl.): 55-67. Gibb, W.R.G. (1992) Melanin, tyrosine hydroxylase, calbindin and substance P in the human midbrain and substantia nigra in relation to nigrostriatal projections and differential neuronal susceptibility in Parkinson's disease. Brain 7?^^. 581:283-291. Gibb, W.R.G. and Lees, A.J. (1988) The relevance of the Lewy body to the pathogenesis of idiopathic Parkinson's disease. J. Neurol. Neurosurg. Psychiatry 51: 745752. Gilbert, H.F. (1982) Biological disulfides: the third messenger? Modulation of phosphofructokinase activity by thiol/disulfide exchange./. Biol. Chem. 251: 12086-12091. Girotti, A.W. (1985) Mechanisms of lipid peroxidation. Free Radical Biol. Med. 1: 87-95. Girotti, A.W. (1990) Photobiology school; photodynamic lipid peroxidation in biological systems. Photochem. Photobiol. 51:491-509. Glees, P. and Hasan, M. (1976) Lipofuscin in neuronal aging

M.E.

Gotzetal.

and disease. In: Doerr (Ed.), Normal and Pathological Anatomy, Thieme, Stuttgart, pp. 1-68. Glenner, G.G., Burtner, H.J and Brown, G.W. (1957) The histochemical demonstration of monoamine oxidase activity by tetrazolium salts. J. His toe hem. Cy toe hem. 5: 5 9 1 600. Glinn, M., Ernster, L. and Lee, C.P. (1991) Initiation of lipid peroxidation in submitochondrial particles: effect of respiratory inhibitors. Areh. Bioehem. Biophys. 290: 57-65. Glover, V., Sandler, M., Owen, F. and Riley, G.J. (1977) Dopamine is a monoamine oxidase B substrate in man. Nature 265:80-81. Glover, V., Elsworth, J.D. and Sandler, M. (1980) Dopamine oxidation and its inhibition by (-)deprenyl in man. J. Neural Transm. 16: 163-172. Goedert, M., Fine, A., Hunt, S.P. and Ullrich, A. (1986) Nerve growth factor mRNA in peripheral and central rat tissues and in the human central nervous system: lesion effects in the rat brain and levels in Alzheimer's disease. Brain Res. 1:85-92. Golbe, L.I., Farrell, T.M. and Davis, RH. (1988) Case-control study of early life dietary factors in Parkinson's disease. Areh. Neurol. 45: 1350-1353. Golbe, L.I., Di lorio, G., Bonavita, V., Miller, D.C. and Duvoisin, R.C. (1990) A large kindred with autosomal dominant Parkinson's disease. Ann. Neurol. 27: 276-282. Goldgaber, D. and Schmechtel, D.E. (1990) Expression of the amyloid y3-protein precursor gene. In: R.J. Wurtman, S. Corkin, J.H. Growdon and E. Ritter-Walker (Eds.), Advances In Neurology, Vol. 51, Alzheimer's Disease, Raven Press, New York, pp. 163-169. Goldsmith, J.R., Herishanu, Y., Abarbanel, J.M. and Weinbaum, Z. (1990) Clustering of Parkinson s disease points to environmental etiology. Arch. Environ. Health 45: 88-94. Gong, L., Daigneault, E.A., Acuff, R.V. and Kostrzewa, R.M. (1991) Vitamin E supplements fail to protect mice from acute MPTP neurotoxicity. NeuroReport 2: 544-546. Good, P.F., Olanow, C.W. and Perl, D.P. (1992a) Neuromelanin-containing neurons of the substantia nigra accumulate iron and aluminum in Parkinson's disease: a LAMM A study. Brain Res. 593: 343-346. Good, P.F., Perl, D.P., Bierer, L.M. and Schmeidler, J. (1992b) Selective accumulation of aluminum and iron in the neurofibrillary tangles of Alzheimer's disease: a laser microprobe (LAMMA) study. Ann. Neurol. 31: 286-292. Gorsky, L.D., Koop, D.R. and Coon, M.J. (1984) On the stoichiometry of the oxidase and monooxygenase reactions catalyzed by liver microsomal cytochrome P-450. J. Biol. C/i^m. 259:6812-6817. Goss-Sampson, M.A., McEvilly, C.J. and Muller, D.P.R. (1988) Longitudinal studies of the neurobiology of vitamin E and other antioxidant systems, and neurological function in the vitamin E deficient rat. J. Neurol. Sei. 87: 25-35. Gottfries, C.G. (1985a) Alzheimer's disease and senile dementia: biochemical characteristics and aspects of the treatment. Psyehopharmaeology 245-251.

489 Gottfries, C.G. (1985b) Transmitter deficits in Alzheimer's disease. Neurochem. Int. 7: 565-566. Gottfries, C.G. (1990) Neurochemical aspects of dementia disorders. Dementia 1: 56-64. Gotz, M.E., Freyberger, A., Hauer, E., Burger, R., Sofic, E., Gsell, W., Heckers, S., Jellinger, K., Hebenstreit, G., Frolich, L., Beckmann, H. and Riederer, P. (1992) Susceptibility of brains from patients with Alzheimer's disease to oxygen-stimulated lipid peroxidation and differential scanning calorimetry. Dementia 3: 213-222. Gotz, M.E., Dirr, A., Freyberger, A., Burger, R. and Riederer, P. (1993) The thiobarbituric acid assay reflects susceptibility to oxygen-induced lipid peroxidation in vitro rather than levels of lipid hydroperoxides in vivo: a methodological approach. Neurochem. Int. 22: 255-262. Graham, D.G., Tiffany, S.M., Bell, Jr., W.R. and Gutknecht, W.F. (1978) Autoxidation versus covalent binding of quinones as the mechanism of toxicity of dopamine. 6hydroxydopamine and related compounds toward CI300 neuroblastoma cells in vitro. Mol. Pharmacol. 14: 644-653. Graham, K.S., Reddy, C.C. and Scholz, R.W. (1989) Reduced glutathione effects on alpha-tocopherol concentration of rat liver microsomes undergoing NADPH-dependent lipid peroxidation. Lipids 24: 909-914. Graham, R.C. and Karnovsky, M.J. (1965) The histochemical demonstration of monoamine oxidase by coupled peroxidatic oxidation. J. Histochem. Cytochem. 13: 604-605. Granger, D.N., Rutili, G. and McCord, J. M. (1981) Superoxide radicals in feline intestinal ischemia. Gastroenterology 81:22-29. Grankvist, K., Marklund, L. and Taljedal, I.-B. (1981) CuZnsuperoxide dismutase, Mn-superoxide dismutase, catalase and glutathione peroxidase in pancreatic islets and other tissues in the mouse. Bioehem. J. 199: 393-398. Gray, J.A. (1987) The neuropsychology of emotion and personality. In: S.M. Stahl, S.D. Iversen and E.C. Goodman (Eds.), Cognitive Neurochemistry, Oxford University Press, Oxford, pp. 171-190. Grimsby, J., Chen, K., Wang, L.-J., Lan, N.C. and Shih, J.C. (1991) Human monoamine oxidase A and B genes exhibit identical exon-intron organization. Proc. Natl. Acad. Sei. /75A 88: 3637-3641. Grundke-Iqbal, L., Fleming, J., Tung, Y.-C, Lassman, H., Iqbal, K. and Joshi, J.G. (1990) Ferritin is component of the neuritic (senile) plaque in Alzheimer dementia. Acta Neuropathol. 81: 105-110. Guo, N. and Shaw, C. (1992) Characterization and localization of glutathione binding sites on cultured astrocytes. Mol. Brain Res. 15: 207-215. Gupta, M., Schwarz, J., Chen, X.L. and Roisen, F.J. (1990) Gangliosides prevent MPTP toxicity in mice - an immunocytochemical study. Brain Res. 527: 330-334. Gutteridge, J.M.C. and Halliwell, B. (1990) The measurement and mechanism of lipid peroxidation in biological systems. Trends Biol. Sei. 15: 129-135. Gutteridge, J.M.C, Halliwell B., Harrison, P., Treffry, A. and

490 Blake, D.R. (1983) Effect of ferritin containing fractions with different iron loading on lipid peroxidation. Biochem. J. 209; 557-560. Gutteridge, J.M.C., Halliwell, B. and Rowley, D.A. (1984) Catalytic iron complexes in biological material: a potential for oxygen radical damage. Life Chem. Rep. 2 (Suppl.): 1526. Gutteridge, J.M.C., Quinlan, G.J., Clark, I. and Halliwell, B. (1985) Aluminium salts accelerate peroxidation of membrane lipids stimulated by iron salts. Biochim. Biophys. Acta 835: AA\-AA1. Guttman, M. and Seeman, P. (1986) Dopamine D2 receptor density in parkinsonian brain is constant for duration of disease, age and duration of L-DOPA therapy. In: M.D. Yahr and K.J. Bergmann (Eds.), Advances in Neurology, Parkinson's Disease, Vol. 45, Raven Press, New York, pp. 51-57. Gyllensten, U., Wharton, D., Josefsson, A. and Wilson, A.C. (1991) Paternal inheritance of mitochondrial DNA in mice. Nature 352: 255-257. Hadjiconstantinou, M., Mariani, A.P. and Neff, N.H. (1989) GMl ganglioside-induced recovery of nigrostriatal dopaminergic neurons after MPTP: an immunohistochemical study. Brain Res. 484: 297-303. Hadjiconstantinou, M., Weihmuller, F.B., Bruno, J.P., Mariani, A.P. and Neff, N.H. (1990) Recovery of dopaminergic function following MPTP-induced neurodegeneration by exogenous GMl ganglioside. In: L.A. Horrocks (Ed.), Trophic Factors and the Nervous System, Raven Press, New York, pp. 293-305. Hageman, J.J., Bast, A. and Vermeulen, N.P.E. (1992) Monitoring of oxidative free radical damage in vivo: analytical aspects. Chem -Biol. Interact. 82: 243-293. Haglund, L., Kohler, C, Haaparanta, T., Goldstein, M. and Gustafsson, J.A. (1984) Presence of NADPH-cytochrome P-450 reductase in central catecholaminergic neurons. Nature 307: 259-262. Hall, E.D. (1992) Novel inhibitors of iron-dependent lipid peroxidation for neurodegenerative disorders. Ann. Neurol. 32 (Suppl.): 137-142. Halliwell, B. (1989a) Oxidants and the central nervous system: some fundamental questions. Acta Neurol. Scand. 126: 23-33. Halliwell, B. (1989b) Protection against tissue damage in vivo by desferrioxamine. What is its mechanism of action? Free Radical Biol. Med. 7: 645-651. Halliwell, B. and Aruoma, O.I. (1991) DNA damage by oxygen-derived species. Its mechanism and measurement in mammalian systems. FBBS Lett. 281: 9-19. Halliwell, B. and Dizdaroglu, M. (1992) The measurement of oxidative damage to DNA by HPLC and GC/MS techniques. Free Radical Res. Commun. 16: 75-87. Halliwell, B. and Gutteridge, J.M.C. (1984) Oxygen toxicity, oxygen radicals, transition metals and disease. Biochem. J. 219: 1-14. Halliwell, B. and Gutteridge, J.M.C. (1989) Free Radicals in

Oxidative stress: free radical production in neural degeneration Biology and Medicine, 2nd edn,, Clarendon Press, Oxford, UK. Halliwell, B. and Gutteridge, J.M.C. (1990) The antioxidants of human extracellular fluids. Arch. Biochem. Biophys. 280: 1-8. Halliwell, B. and Gutteridge, M.C. (1992) Biologically relevant metal ion-dependent hydroxyl radical generation: an update. FEBSLett. 307: 108-112. Handelman, G.J., Van Kuijik, F.J.G.M., Chatterjee, A. and Krinsky, N.I. (1991) Characterization of products formed during the autoxidation of beta-carotene. Free Radical Biol. Med. 10: 427-437. Hanson, C, Rorsman, H., Rosengren, E. and Wittbjer, A. (1985) Production of 6-hydroxydopa by human tyrosinase. Acta Dermato-Venereol. 65: 154-157. Harding, A.E. (1991) Neurological disease and mitochondrial genes. Trends Neurosci. 14: 132-138. Hardy, J., Adolfsson, R., Alafuzoff, I., Bucht, G., Marcusson, J. and Nyberg, P. (1985) Transmitter deficits in Alzheimer's disease. Neurochem. Int. 7: 545-563. Harman, D. (1992) Free radical theory of aging. Mutat. Res. 275: 257-266. Harris, J.I., Auffret, A.D., Northrop, F.D. and Walker, J.E. (1980) Structural comparisons of superoxide dismutase. Eur. J. Biochem. 106: 297-303. Hartz, J.W., Funakoshi, S. and Deutsch, H.F. (1973) The levels of superoxide dismutase and catalase in human tissues as determined immunochemically. Clin. Chim. Acta 46: 125-132. Hattori, N., Tanaka, M., Ozawa, T. and Mizuno, Y. (1991) Immunohistochemical studies on complexes I, II, III and IV of mitochondria in Parkinson's disease. Ann. Neurol. 30: 563-571. Hayashi, H. and Tsubaki, T. (1982) Enzymatic analysis of individual anterior horn cells in amyotrophic lateral sclerosis and Duchenne muscular dystrophy. J. Neurol. Sci. 57: 133-142. Hecker, M. and Ullrich, V. (1989) On the mechanism of prostacyclin and thromboxane A2 biosynthesis. /. Biol. Chem. 264: 141-150. Hefti, F. and Schneider, L.S. (1991) Nerve growth factor and Alzheimer's disease. Clin. Neuropharmacol. 14 (Suppl. 1): 62-76. Hefti, F. and Weiner, W.J. (1986) Nerve growth factor and Alzheimer's disease. Ann. Neurol. 20: 275-281. Hefti, F., Melamed, E., Bhawan, J. and Wurtman, R.J. (1981) Longterm administration of L-DOPA does not damage dopaminergic neurons in the mouse. Neurology 31: 11941195. Heikkila, R.E., Hess, A. and Duvoisin, R.C. (1984a) Dopaminergic neurotoxicity of l-methyl-4-phenyl-1,2,3,6tetrahydropyridine in mice. Science 224: 1451-1453. Heikkila, R.E., Manzino, L., Cabbat, F.S. and Duvoisin, R.C. (1984b) Protection against dopaminergic neurotoxicity of 1methyl-4-phenyl-l,2,3,6-tetrahydropyridine by monoamine oxidase inhibitors. Nature 311: 467-469.

M.E. Gotzetal. Heikkila, R.E., Sieber, B.-A., Manzino, L. and Sonsalla, P.K. (1989a) Some features of the nigrostriatal dopaminergic neurotoxin l-methyl-4-phenyl-l ,2,3,6-tetrahydropyridine (MPTP) in the mouse. Mol Chem. NeuropathoL 10: 171183. Heikkila, R.E., Sonsalla P.K. and Duvoisin, R.C. (1989b) Biochemical models of Parkinson's disease. In: A.B. Boultone, G.B. Baker and A.V. Juorio (Eds.), Neuromethods, Drugs as Tools in Neurotransmitter Research. Vol. 12, Humana Press, Clifton, NJ, pp. 351-384. Heinonen, E.H. and Lammintausta, R. (1991) A review of the pharmacology of selegiline. Acta Neurol. Scand. 84 (Suppl. 136): 44-59. Heinsen, H. (1979) Lipofuscin in the cerebellar cortex of albino rats: an electron microscopic study. Anat. Embryol. 155: 333-345. Heizmann, C.W. and Braun, K. (1992) Changes in Ca^+ binding proteins in human neurodegenerative disorders. Trends Neurosci. 15: 259-264. Henderson, A.S. (1990) Epidemiology of dementia disorders. In: R.J. Wurtman, S. Corkin, J.H. Growdon and E. RitterWalker (Eds.), Advances in Neurology, Alzheimer's Disease. Vol. 51, Raven Press, New York, pp. 15-25. Herken, H. and Hucho, F. (1992) Selective Neurotoxicity Springer, New York. Hill, J.M. (1988) The distribution of iron in the brain. In: M.B.H. Youdim (Ed.), Brain Iron. Neurochemical and Behavioural Aspects, Taylor & Francis, London, pp. 1-24. Hirsch, E.C. (1992) Why are nigral catecholaminergic neurons more vulnerable than other cells in Parkinson's disease? Ann. Neurol. 32 (Suppl): 88-93. Hirsch, E., Graybiel, A.M. and Agid, Y.A. (1988) Melanized dopaminergic neurons are differentially susceptible to degeneration in Parkinson's disease. Nature 334: 345-348. Hirsch, E.C, Brandel, J.-P., Galle, P., Javoy-Agid, F. and Agid, Y. (1991) Iron and aluminium increase in the substantia nigra of patients with Parkinson's disease, an X-ray microanalysis. J. Neurochem. 56: 446-451. Ho, S.C, Woo, J. and Lee, CM. (1989) Epidemiologic study of Parkinson's disease in Hong Kong. Neurology 39: 13141318. Holmes, G.E., Bernstein, C and Bernstein, H. (1992) Oxidative and other DNA damages as the basis of aging: a review. Mutat. Res. 275: 305-315. Holttum, J.R. and Gershon, S. (1992) The cholinergic model of dementia, Alzheimer type: progression from the unitary transmitter concept. Dementia 3: 174-185. Hornykiewicz, O. and Kish, S.J. (1986) Biochemical pathophysiology of Parkinson's disease. In: M.D. Yahr and K.J. Bergmann (Eds.), Advances in Neurology, Parkinson's Disease, Vol. 45, Raven Press, New York, pp. 19-33. Horwitt, M.K. (1991) The term alpha-tocopherol should not be used without clarification. Am. J. Clin. Nutr. 54: 760770. Horwitz, S.J. and Roessmann, U. (1978) Kearns-Sayre syndrome with hypoparathyroidism. Ann. Neurol 3: 513-518.

491 Hruba, F., Novakova, V. and Ginter, E. (1982) The effect of chronic marginal vitamin C deficiency on the alphatocopherol content of the organs and plasma of guinea pigs. Experientia 38: 1454-1455. Hruszkewycz, A.M. and Bergtold, D.S. (1990) The 8hydroxyguanine content of isolated mitochondria increases with lipid peroxidation. Mutat. Res. 244: 123-128. Hudson, A.J. (1981) Amyotrophic lateral sclerosis and its association with dementia, parkinsonism and other neurological disorders: a review. Brain 114: 217-247. Hufschmidt, H.J., Schaltenbrand, G. and Solcher, H. (1960) Uber Muskelatrophien im Zusammenhang mit postencephalitischem Parkinsonismus. Deutsche Zeitschr. Nervenheilkunde 181: 335-344. Hughes, H., Smith, C.V., Tsokos-Kuhn, J.O. and Mitchell, J.R. (1986) Quantitation of lipid peroxidation products by gas chromatography-mass spectrometry. Anal. Biochem. 152: 107-112. Hughles, A.J., Daniel, S.E., Kliford, L. and Lees, A. (1992) Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico pathological study of 100 cases. J. Neurol. Neurosurg. Psychiatry 55: 181-184. Hwa, G.G.C and Avoli, M. (1991) The involvement of excitatory amino acids in neocortical epileptogenesis: NMD A and non-NMDA receptors. Exp. Brain Res. 86: 248-256. Hyman, C, Hofer, M., Barde, Y.-A., Juhasz, M., Yancopoulos, G.D., Squinto, S.P. and Lindsay, R.M. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350: 230-232. Ichiki, T., Tanaka, M. and Nishikimi, M. (1988) Deficiency of subunits of complex I and mitochondrial encephalomyopaihy. Ann. Neurol. 23: 287-294. Ichimiya. Y., Arai, H., Kosaka, K. and lizuka, R. (1986) Morphological and biochemical changes in the cholinergic and monoaminergic systems in Alzheimer-type dementia. Acta NeuropathoL 70: 112-116. Ichitani, Y., Okamura, H., Matsumoto, Y., Nagatsu, I. and Ibata, Y. (1991) Degeneration of the nigral dopamine neurons after 6-hydroxydopamine injection into the rat striatum. Brain Res. 549: 350-353. Ikebe, S., Tanaka, M., Ohno, K., Sato, W., Hattori, K., Kondo, T., Mizuno, Y. and Ozawa, T. (1990) Increase of deleted mitochondrial DNA in the striatum in Parkinson's disease and senescence. Biochem. Biophys. Res. Commun. 170: 1044-1048. Ikeda, H., Markey, C.J. and Marrey, S.P. (1992) Search for neurotoxins structurally related to l-methyl-4-phenylpyridine (MPP"*") in the pathogenesis of Parkinson's disease. Brain Res. 575: 285-298. Imperato, A., Ramacci, M.T. and Angelucci, L. (1989) Acetyl-L-carnitine enhances acetylcholine release in the striatum and hippocampus of awake freely moving rats. Neurosci. Lett. 107: 251-255. Ito, S., Fujita, K., Yoshioka, M., Sienko, D. and Nagatsu, T. (1986) Identification of 5-5- and 2-5-cysteinyldopamine and 5-S-glutathionyldopamine formed from dopamine by

492 high-performance liquid chromatography with electrochemical detection. 7. Chromatogr. 375: 134-140. Iwahashi, H., Morishita, H., Ishii, T., Sugata, R. and Kido, R. (1989) Enhancement by catechols of hydroxyl-radical formation in the presence of ferric ions and hydrogen peroxide J. Biochem. 105: 429-434. Iwamoto, N., Kobayashi, K. and Kosaka, K. (1989) The formation of prostaglandins in the post-mortem cerebral cortex of Alzheimer-type dementia patients. J. Neurol 236: 8084. Jackson, G.R., Apffel, L., Werrbach-Perez, K. and Perez Polo, J.R. (1990a) Role of nerve growth factor in oxidantantioxidant balance and neuronal injury. I Stimulation of hydrogen peroxide resistance. J. Neurosci. Res. 25: 360368. Jackson, G.R., Werrbach-Perez, K. and Perez Polo, J.R. (1990b) Role of nerve growth factor in oxidant-antioxidant balance and neuronal injury. II. A conditioning lesion paradigm. /. Neurosci. Res. 25: 369-374. Jacobsen, E.J., McCall, J.M., Ayer, D.E., Van Doornik, F.J., Palmer, J.R., Belonga. K.L., Braughler, J.M., Hall, E.D., Houser, D.J., Krook, M.A. and Runge, T.A. (1990) Novel 21-aminosteroids that inhibit iron-dependent lipid peroxidation and protect against central nervous system trauma. J. Med. Chem. 33: 1145-1151. Jakoby, W.B. and Ziegler D.M. (1990) The enzymes of detoxication. J. Biol. Chem. 265: 20715-20718. Janero, D.R. (1990) Malondialdehyde and thiobarbituric acid, reactivity as diagnostic indices of lipid peroxidation and peroxidative tissue injury. Free Radical Biol. Med. 9: 515540. Jansen, K.L.R., Faull, R.L.M., Dragunow, M. and Syneks, B.L. (1990) Alzheimer's disease: changes in hippocampal A^-methyl-D-aspartate, quisqualate, neurotensine, adenosine, benzodiazepine, serotonin and opioid receptors, an autoradiographic study. Neuroscience 39: 613-627. Javitch, J.A., D'Amato, R.J., Strittmatter, S.M. and Snyder, S.H. (1985) Parkinsonism-inducing neurotoxin A^-methyl-4phenyl-l,2,3,6-tetrahydropyridine: uptake of the metabolite A^-methyl-4-phenylpyridine by dopamine neurons explains selective toxicity. Proc. Natl. Acad. Sci. USA 82: 21732177. Jeandel, C, Nicolas, M.B., Dubois, E., Nabet-Belleville. F., Penin, F. and Cuny, G. (1989) Lipid peroxidation and free radical scavengers in Alzheimer's disease. Gerontology 35: 275-282. Jeffcoat, R. (1979) Biosynthesis of unsaturated fatty acids and its control in mammalian liver. Essays Biochem. 15: 1-36. Jellinger, K. (1986) Overview of morphological changes in Parkinson's disease. In: M.D. Yahr and K.J. Bergmann (Eds.), Advances in Neurology, Parkinson's Disease, Vol. 45, Raven Press, New York, pp. 1-17. Jellinger, K. (1989) Pathology of Parkinson's syndrome. In: D.B. Calne (Ed.), Handbook of Experimental Pharmacology, Springer, New York, pp. 47-112. Jellinger, K. and Riederer, P. (1984) Dementia in Parkinson's

Oxidative stress: free radical production in neural degeneration disease and (pre)senile dementia of Alzheimer type: morphological aspects and changes in intracerebral MAO activity. Acfv. Neurol. 40: 199-210. Jellinger, K., Paulus, W., Grundke-Iqbal, I., Riederer, P and Youdim, M.B.H. (1990) Brain iron and ferritin in Parkinson's and Alzheimer's diseases. J. Neural Transm 2: 327340. Jellinger, K., Kienzl, E., Rumplmair, G., Riederer, P., Stachelberger, H., Ben-Shachar, D. and Youdim, M.B.H. (1992) Iron-melanin complex in substantia nigra of parkinsonian brains: an X-ray microanalysis. J. Neurochem. 59: 1168-1171. Jellinger, K., Kienzl, E., Puchinger, L. and Stachelberger, H. (1993) Changes of phospholipids in Alzheimer's disease brain. In: B. Corain, K. Iqbal, M. Nicolini, B. Winblad, H. Wisniewski and P. Zatta (Eds.), Alzheimer's Disease: Advances in Clinical and Basic Research, Wiley, New York, pp.315-323. Jellum, E., Marstein, L., Skullerud, K. and Munthe, E. (1983) Glutathione in pyroglutamic aciduria (5-oxoprolinuria) and rheumatoid arthritis. In: A. Larsson (Ed.), Functions of Glutathione: Biochemical, Physiological, Toxicological, and Clinical Aspects, Raven Press, New York, pp. 347353. Jenner, P. (1992) Parkinson's disease: pathological mechanisms and actions of piribedil. /. Neurol. 239 (Suppl. 1): S2-S8. Jenner, P., Rupinak, N.M.J., Rose, S., Kelly, E., Kilpatrick, G., Lees, A. and Marsden, CD. (1984) l-Methyl-4-phenyl1,2,3,6-tetrahydropyridine-induced parkinsonism in the common marmoset. Neurosci. Lett. 50: 85-90. Jenner, P., Dexter, D.T., Sian, J., Schapira, A.H.V. and Marsden. CD. (1992) Oxidative stress as a cause of nigral cell death in Parkinson's disease and incidental Lewy body disease. Ann. Neurol. 32 (Suppl.): 82-87. Johannessen. J.N., Chiueh, C C and Bacon. J.P. (1985) Neurochemical effect of MPTP in the dog: effects of pargyline pretreatment. Soc. Neurosci. Abstr. 11: 631. Johannessen, J.N., Adams, J.D., Schuller, H., Bacon, J. and Markey, S.P. (1986) 1 -Methyl-4-phenylpyridine (MPP+) induces oxidative stress in the rat. Life Sci. 38: 743-749. Johannsen, P., Velander, G., Mai, J., Thorling, E.B. and Dupont, E. (1991) Glutathione peroxidase in early and advanced Parkinson's disease. J. Neurol. Neurosurg. Psychiatry 54: 679-682. Johnson, A.W. and Demple, B. (1988) Yeast DNA 3-repair diesterase is the major cellular apurinic/aparimidinic endonuclease: substrate specificity and kinetics. J. Biol. Chem. 263: 18017-18022. Johnson, W.G., Hodge, S.E. and Duvoisin, R. (1990) Twin studies and the genetics of Parkinson's disease - a reappraisal. Mov. Disord. 5: 187-194. Johnston, LP. (1968) Some observation upon a new inhibitor of monoamine oxidase. Biochem. Pharmac. 17: 1285-1297. Jossan, S.S., Hiraga, Y. and Oreland, L. (1989) The cholinergic neurotoxin ethylcholine mustard aziridinium (AF64A)

M.E.

Gotzetal.

induces an increase in MAO-B activity in the rat brain. Brain Res. 416:291-291. Jossan, S.S., Gillberg. P.G., D'Argy, R., Aquilonius, S.M., L a n o s t r o m , B., Halldin, C. and Oreland, L. (1991a) Quantitative localization of human brain monoamine oxidase B by large section autoradiography using L[^H]deprenyl. Brain Res. 547: 69-76. Jossan, S.S., Gillberg, P.G., Gottfries, C.G., Karlsson, I. and Oreland. L. (1991b) Monoamine oxidase B in brains from patients with Alzheimer's disease: a biochemical and autoradiographical study. Neuroscience 45: 1-12. Jung, D.W. and Brierlee, G.P. (1983) Oxidative phosphorylation. In: A. Lajtha (Ed.), Handbook of Neurochemistry, Metabolism in the Nervous System, Vol. 3. 2nd edn., Plenum Press, New York, pp. 295-319. Kadiska, M.B., Hanna, P.M., Hernandez, L. and Mason, R.P. ( 1992) In vivo evidence of hydroxyl radical formation after acute copper and ascorbic acid intake: electron spin resonance spin-trapping investigation. Mol. Pharmacol. 42: 723-729. Kagan, V., Serbinova, E. and Packer, L. (1990) Antioxidant effects of ubiquinones in microsomes and mitochondria are mediated by tocopherol recycling. Biochem. Biophys. Res. Commun. 169: 851-857. Kalaria, R.N. and Harik, S.I. (1989) Reduced glucose transporter at the blood brain barrier and in cerebral cortex in Alzheimer's disease. J. Neurochem. 53: 1083-1088. Kalra, J., Rajput, A.H., Mantha, S.V. and Prasad, K. (1992) Serum antioxidant enzyme activity in Parkinson's disease. Mol. Cell. Biochem. 110: 165-168. Kaneko, T., Honda, S., Nakano, S.I. and Matsuo, M. (1987) Lethal effects of a linoleic acid hydroperoxide and its autoxidation products, unsaturated aliphatic aldehydes on human diploid fibroblasts. Chem.-Biol. Interact. 63: 127137. Kaneko, Y., Kitamoto, T., Tateishi, J. and Yamaguchi, K. (1989) Ferritin immunohistochemistry as a marker for m\crogiia. Acta Neuropathol. 79: 129-136. Kappus, H. (1985) Lipid peroxidation: mechanisms, analysis, enzymology and biological relevance. In: H. Sies (Ed.), Oxidative Stress, Academic Press, New York, pp. 2 7 3 310. Kasai, H. and Nishimura, S. (1986) Hydroxylation of guanine in nucleosides and DNA at the C-8 position by heated glucose and oxygen radical forming agents. Environ. Health Perspect.61: 111-116. Kasai, H., Grain, P.P., Kuchino, Y., Nishimura, S., Ootsuyama, A. and Tanooka, H. (1986) Formation of 8hydroxyguanine moiety in cellular DNA by agents producing oxygen radicals and evidence for its repair. Carcinogenesis 1: 1849-1851. Kasprzak, K.S. (1991) The role of oxidative damage in metal carcinogenesis. Chem. Res. Toxicol. 4: 604-615. Kastner, A., Hirsch, E.G., Lejeune, O., Javoy-Agid, F., Rascol, O. and Agid, Y. (1992) Is the vulnerability of neurons in the substantia nigra of patients with Parkinson's disease

493 related to their neuromelanin content? J. Neurochem. 59: 1080-1089. Kato, S., Hirano, A., Liena, J.F. and Yen, S.-H. (1992) Ultrastructural identification of neurofibrillary tangles in the spinal cord in guamanian amyotrophic lateral sclerosis and parkinsonism-dementia complex on Guam. Acta Neuropathol. 83: 277-282. Katzman, R. and Saitoh, T. (1991) Advances in Alzheimer's disease. Fed. Proc. Am. Soc. Exp. Biol. 5: 278-286. Keele, B.B., McCord, J.M. and Fridovich, I. (1970) Superoxide dismutase from Escherichia coli B: a new manganesecontaining enzyme. J. Biol. Chem. 245: 6176-6181. Keller, G.A., Warner, T.G., Steimer, K.S. and Hallewell, R.A. (1991) Cu, Zn superoxide dismutase is a peroxisomal enzyme in human fibroblasts and hepatoma cells. Proc. Natl. Acad. Sci. USA 88: 7381 -7385. Kennedy, C., Sakurada, O., Shinohara, M., Jehle, J. and Sokoloff, L. (1978) Local cerebral glucose utilization in the normal conscious Macaque monkey. Ann. Neurol. 4: 2 9 3 301. Kennedy, T.A. and Liebler, D.C. (1991) Peroxyl radical oxidation of beta-carotene: formation of beta-carotene epoxides. Chem. Res. Toxicol. 4: 290-295. Khachaturian, Z.S. (1985) Diagnosis of Alzheimer's disease. Arch. Neurol. 42: 1097-1105. Khan, A.U., Gebauer, P. and Hager, L.P. (1983) Chloroperoxidase generation of singlet molecular oxygen observed directly by spectroscopy in the 1 to \.6fim. region. Proc. Natl. Acad. Sci. USA. 80: 5195-5197. Kish, S.J., Morito, C. and Hornykiewicz, O. (1985) Glutathione peroxidase activity in Parkinson's disease brain. Neurosci. Lett. 58: 343-346. Kish, S.J., Shannak, K., Raijput, A., Deck, J.H. and Hornykiewicz, O. (1992) Aging produces a specific pattern of striatal dopamine loss: implications for the etiology of Parkinson's disease. J. Neurochem. 58: 643-648. Klein, C.B., Frenkel, K. and Costa, M. (1991) The role of oxidative processes in metal carcinogenesis. Chem. Res. Toxicol. 4: 592-604. Klockgether, T., Turski, L., Honore, T., Zhang, Z., Gash, D.M., Kurlan, R. and Greenamyre, J.T. (1991) The AMPA receptor antagonist NBQX has antiparkinsonian effect in monoamine-depleted rats and MPTP monkeys. Ann. Neurol. 30: 717-723. Knoll, J. (1987) (-)Deprenyl (Selegiline, Movergan) facilitates the activity of the nigrostriatal dopaminergic neuron. /. Neural. Transm. 25 (Suppl.): 45-66. Knoll, J. (1988) The striatal dopamine dependency of life span in male rats. Longevity study with (-)deprenyl. Mech. Ageing Dev. 46: 237-262. Knoll, J. (1992a) Pharmacological basis of the therapeutic effect of (-)deprenyl in age-related neurological diseases. Med. Res. Rev. 12: 505-524. Knoll, J. (1992b) The pharmacological profile of (-)deprenyl (Selegiline) and its relevance for humans: a personal view. Pharmacol. Toxicol. 70: 317-321.

494 Knoll, J. and Magyar, K. (1972) Some puzzling pharmacological effects of monoamine oxidase inhibition. Adv. Biochem. Psychopharmacol. 5: 393-408. Knoll, J., Ecsery, Z., Magyar, K. and Satory, E. (1978) Novel (-)deprenyl-derived selective inhibitors of B-type monoamine oxidase. The relation of structure to their action. Biochem. Pharmacol. 27: 1739-1747. Knoll, J., Dallo, J. and Yen, T.T. (1989) Striatal dopamine, sexual activity and life span. Longevity of rats treated with (-)deprenyl. Life Sci. 45: 525-531. Koller, W.C. (1992) Initiating treatment of Parkinson's disease. Neurology 42 (Supp. 1): 33-38. Kono, Y. and Fridovich, I. (1982) Superoxide radical inhibits catalase. J. Biol. Chem. 257: 5751-5754. Konradi, C , Svoma, E., Jellinger, K., Riederer, P., Denney, R. and Thibault, J. (1988) Topographic immunocytochemical mapping of monoamine oxidase-A, monoamine oxidase-B and tyrosine hydroxylase in human post mortem brain stem. Neuroscience 26: 791-802. Konradi, C , Kornhuber, J., Frolich, L., Fritze, J., Heinsen, H., Beckmann, H., Schulz, E. and Riederer, P. (1989) Demonstration of monoamine oxidase-A and -B in the human brainstem by a histochemical technique. Neuroscience 33: 383^00. Koppenol, W.H. (1990) What is in a name? Rules for radicals. Free Radical Biol Med. 9: 225-227. Kornhuber, J., Bormann, J., Hubers, M., Rusche, K. and Riederer, P. (1991) Effects of the 1-amino-adamantanes at the MK-801-binding site of the NMDA-receptor gated ion channel: a human postmortem brain study. Eur. J. Pharmacol. Mol. (Pharmac. Sect.) 206: 297-300. Kosower, N.S., Glaser, T. and Kosower, E.M. (1983) Membrane-mobility agent-promoted fusion of erythrocytes: fusibility is correlated with attack by calcium-activated cytoplasmic proteases on membrane proteins. Proc. Natl. Acad. Sci. USA 80: 7542-7546. Kostrzewa, R.M. (1989) Neurotoxins that affect central and peripheral catecholamine neurons. In: A.B. Boulton, G.B. Baker and A.V. Juorio (Eds.), Neuromethods, Drugs as Tools in Neurotransmitter Research, Vol. 12, Humana Press, Clifton, NJ, pp. 1 ^ 8 . Kracun, I., Kalanj, S. Talan-Hranilovic, J. and Cosovic, C. (1992a) Cortical distribution of gangliosides in Alzheimer's disease. Neurochem. Int. 20: 433-438. Kracun, I., Rosner, H., Drnovsek, V., Vukelic, Z., Cosovic, C , Trbojevic-Cepe, M. and Kubat, M. (1992b) Gangliosides in the human brain development and aging. Neurochem. Int. 20:421-431. Krinsky, N.I. (1992) Mechanism of action of biological antioxidants. Proc. Soc. Exp. Biol. Med. 200: 248-254. Kurland, L.-T. (1988) Amyotrophic lateral sclerosis and Parkinson's disease complex on Guam linked to an environmental neurotoxin. Trends Neurosci. 11: 51-54. Landsberg, J.P., McDonald, B. and Watt, F. (1992) Absence of aluminium in neuritic plaque cores in Alzheimer's disease. Nature 360: 65-68.

Oxidative stress: free radical production in neural degeneration Langston, J.W., Ballard, P.A., Tetrud, J.W. and Irwin, I. (1983) Chronic parkinsonism in human due to a product of meperidine-analog synthesis. Science 219: 979-980. Langston, J.W., Forno, L.S., Rebert, C.S. and Irwin, I. (1984a) Selective nigral toxicity after systemic administration of 1methyl-4-phenyl-l,2,3,6-tetrahydropyridine (MPTP) in the squirrel monkey. Brain Res. 292: 390-394. Langston, J.W., Irwin, I., Langston, E.B. and Forno, L.S. (1984b) Pargyline prevents MPTP-induced parkinsonism in primates. Science 225: 1480-1482. Langui, D., Anderton, B.H., Brion, J.-P. and Ulrich, J. (1988) Effects of aluminium chloride on cultured cells from rat brain hemispheres. Brain Res. 438: 67-76. Lataste, S. (1984) The history and pharmacology of dopamine agonists. Can. J. Neurol. 11: 118-123. Le Bel, C.P and Bondy, S.C. (1992) Oxidative damage and cerebral aging. Prog. Neurobiol. 38: 601-609. Leeden, R.W. (1984) Biology of gangliosides: neuritogenic and neuronotrophic properties. J. Neurosci. Res. 12: 147159. Lerner, A.B., Fitzpatrick, T.B., Calkins, E. and Summerson, W.H. (1949) Mammalian tyrosinase: preparation and properties./. Biol. Chem. 178: 185-195. Lestienne, P., Nelson, J., Riederer, P., Jellinger, K. and Reichmann, H. (1990) Normal mitochondrial genome in brain from patients with Parkinson's disease and complex I defect./. Neurochem. 55: 1810-1812. Lestienne, P., Nelson, I., Riederer, P., Reichmann, H. and Jellinger, K. (1991) Mitochondrial DNA in postmortem brain from patients with Parkinson's disease. J. Neurochem. 56: 1819. Levine, R.L. (1983) Oxidative modification of glutamine synthetase: characterization of the ascorbate model system. J. Biol. Chem. 258: 11828-11833. Levitt, P., Pintar, J.E. and Breakfield, X.O. (1982) Immunocytochemical demonstration of monoamine oxidase B in brain astrocytes and serotonergic neurons. Proc. Natl. Acad. Sci. USA 79:6385-6389. LeWitt, P.A., Oxenkrug, G.F., Mclntyre, I.M. and McCauley, R.M. (1985) Peripheral carbidopa affects monoamine oxidase activity. Neurology 35: 1258-1259. Li, T.M., Aberman, E. and Swash, M. (1990) Clinical features and associations of 560 cases of motor neuron disease. J. Neurol. Neurosurg. Psychiatry 53: 1043-1045. Lieberman, A. ( 1992) Emerging perspectives in Parkinson's disease. Neurology 42 (Suppl. 4): 5-7. Lillie, R.D. and Yamada, H. (1960a) Histochemical studies on the neuromelanins. Okajimas Folia Anat. Jap. 36: 155-163. Lillie, R.D. and Yamada, H. (1960b) On the yellow brown pigment of the substantia nigra, locus coeruleus and dorsal vagal nucleus of a monkey (Macaca mulatta) Okajimas Folia Anat. Jap. 36: 181-183. Llinas, R. and Nicholson, C. (1975) Calcium role in depolarization-secretion coupling: an aequorin study in squid giant synapse. Proc. Natl. Acad. Sci. USA 72: 187-190. Lloyd, T. and Weisz, J. (1978) Direct inhibition of tyrosine

M.E. Gotz et al. hydroxylase activity by catechol estrogens. /. Biol Chem. 253: 4841-4843. L6pez G.-Coviella, I. and Wurtman, R.J. (1992) Enhancement by cytidine of membrane phospholipid synthesis. J. Neurochem. 59: 338-343. Lopez G.-Coviella, I., Agut, J., Ortiz, J.A. and Wurtman, R.J. (1992) Effects of orally administered cytidine 5'diphosphate choline on brain phospholipid content. /. Nutr. Biochem. 3: 313-315. Loschmann, P.A., Lange, K.W., Kunow, M., Jahning, P., Honore, T., Turski, L., Wachtel, H. and Marsden, C D . (1991) Synergistic effects of the AMPA-antagonist NBQX and the AMPA-antagonist CPP with L-DOPA in experimental models of Parkinson's disease. J. Neural Transm. (P-D 5^cr.j 3: 203-213. LoTurco, J.J., Mody, I. and Krieglstein, A.R. (1990) Differential activation of glutamate receptors by spontaneously released transmitter in slices of cortex. Neurosci. Lett. 114: 265-271. Lucy, J.A. (1972) Functional and structural aspects of biological membranes: a suggested role for vitamin E in the control of membrane permeability and stability. Ann. N. Y. Acad. Sci. 203: 4-n . Madeja, U.D. (1992) Der Dopamin-Agonist Lisurid in der Therapie des Morbus Parkinson. Acta Histochem. 8 (Suppl. XLII): 25-31. Mahadik, S.P., Bharucha, V.A., Stadlin, A., Ortiz, A. and Karpiak, S.E. (1992) Loss and recovery of activities of alpha"^ and alpha isoenzymes of (Na"*" + K'^)-ATPase in cortical focal ischemia: GMl ganglioside protects plasma membrane structure and function. J. Neurosci. Res. 32: 209-220. Majewska, M.D., Bell, J.A. and London, E.D. (1990) Regulation of the NMDA receptor by redox phenomena, inhibitory role of ascorbate. Brain Res. 537: 328-332. Maker, H.S., Weiss, C , Silides, D. and Cohen, G. (1981) Coupling of dopamine oxidation (monoamine oxidase activity) to glutathione oxidation via the generation of hydrogen peroxide in rat brain homogenates. /. Neurochem. 36: 589-593. Makino, K., Hagiwara, T., Hagi, A., Nishi, M. and Murakami, A. (1990) Cautionary note for DMPO spin trapping in the presence of iron ion. Biochem. Biophys. Res. Commun. 172: 1073-1080. Makino, Y., Ohta, S., Tachikawa, O. and Hirobe, M. (1988) Presence of tetrahydroisoquinoline and 1-methyltetrahydroisoquinoline in foods: compounds related to Parkinson's disease. Life Sci. 43: 373-378. Manfridi, A., Forloni, G.L., Arrigoni-Martelli, E. and Mancia, M. (1992) Culture of dorsal root ganglion neurons from aged rats: effects of acetyl-L-carnitine and NGF. Neuro science 10: 321-329. Mangoni, A., Grassi, M.P., Frattola, L., Piolti, R., Bassi, S., Motta, A., Marcone, A. and Smirne, S. (1991) Effects of a MAO-B inhibitor in the treatment of Alzheimer disease. Eur. Neurol. 31: 100-107. Mann, D.M.A. and Yates, P.O. (1974) Lipoprotein pigments -

495 their relationship to aging in the human nervous system. IL The melanin content of pigmented nerve cells. Brain 97: 489^98. Mann, D.M. and Yates, P.O. (1983) Possible role of neuromelanin in the pathogenesis of Parkinson's disease. Mech. Aging Dev. 21: 193-203. Mann, D.M.A., Yates, P.O. and Stamp, J.E. (1978) The relationship between lipofuscin pigment and ageing in the human nervous system. J. Neurol. Sci. 37: 83-93. Mann, D.M.A., Lincoln, J., Yates, P.O., Stamp, J.E. and Toper, S. (1980) Changes in the monoamine containing neurons of the human CNS in senile dementia. Br. J. Psychiatry 236: 533-541. Mann, D.M., Yates, P.O. and Marcyniuk, B. (1984a) A comparison of changes in the nucleus basalis and locus coeruleus in Alzheimer's disease. J. Neurol. Neurosurg. Psychiatry 47:201-203. Mann, D.M., Yates, P.O. and Marcyniuk, B. (1984b) Changes in nerve cells of the nucleus basalis of Meynert in Alzheimer's disease and their relationship to ageing and to the accumulation of lipofuscin pigment. Mech. Ageing Dev. 25: 189-204. Mann, J.J., Stanley, M., Gerson, S. and Rossor, M. (1980) Mental symptoms in Huntington's disease and a possible primary aminergic neuron lesion. Science 210: 1369-1371. Mann, J.J., Kaplan, R.D. and Bird, E.D. (1986) Elevated post mortem monoamine oxidase B activity of the caudate nucleus in Huntington's disease compared to schizophrenics and controls. J. Neural Transm. 65: 277-283. Mann, V.M., Cooper, J.M., Krige, D., Daniel, S.E., Schapira, A.H.V. and Marsden, C D . (1992a) Brain, skeletal muscle and platelet mitochondrial function in Parkinson's disease. Brain 115: 333-342. Mann, V.M., Cooper, J.M. and Schapira, A.H.V. (1992b) Quantitation of a mitochondrial DNA deletion in Parkinson's disease. FEBS Lett. 299: 218-222. Maraganore, D.M., Harding, A.E. and Marsden, C D . (1991) A clinical and genetic study of familial Parkinson's disease. Mov. Disord. 6:205-211. Markesberry, W.R., Ehmann, W.D., Hossain, T.LM., Alaudin, M. and Goodin, D.T. (1981) Instrumental neutron activation analysis of brain aluminium in Alzheimer's disease and aging. Ann. Neurol. 10: 511-516. Marklund, S.L., Westman, G., Lundgren, E. and Roos, G. (1982) Copper- and zinc-containing superoxide dismutase, manganese-containing superoxide dismutase, catalase, and glutathione peroxidase in normal and neoplastic human cell lines and normal human tissues. Cancer Res. 42: 19551961. Marsden, C D . (1982) The mysterious motor function of the basal ganglia: The Robert Wartenburg Lecture. Neurology 32:514-538. Marsden, C D . (1983) Neuromelanin and Parkinson's disease. /. Neural Transm. 19 (Suppl.): 121-141. Martensson, J. and Meister, A. (1991) Glutathione deficiency decreases tissue ascorbate levels in newborn rats: ascorbate

496 spares glutathione and protects. Proc. Natl. Acad. Sci. USA 88: 4656-4660. Martin, C.N., Singh, S. and Wood, P.J. (1989) Calcium metabolism in Alzheimer's disease. Gerontology 35: 153157. Martinovits, G., Melamed, E., Cohen, O., Rosenthal, J. and Uzzan, A. (1986) Systemic administration of antioxidants does not protect mice against the dopaminergic neurotoxicity of l-methyl-4-phenyl-l,2.3,6-tetrahydropyridine. Neurosci. Lett. 69: 192-197. Marttila, R.J., Lorentz, H. and Rinne, U.K. (1988) Oxygen toxicity protecting enzymes in Parkinson's disease. Increase of superoxide dismutase-like activity in the substantia nigra and basal nucleus. J. Neurol. Sci. 86: 321-331 Masoro, E.J. (1981) Handbook of Physiology in Aging, CRC Press, Boca Raton, FL. Masoro, E.J. (1991) Biology of aging: facts, thoughts, and experimental approaches. Lab. Invest. 65: 500-510. Masters, C.J. and Beyreuther, K. (1990) Protein abnormalities in neurofibrillary tangles: their relation to the extracellular amyloid deposits of the A4 protein in Alzheimer's disease. In: R.J. Wurtman, S. Corkin, J.H. Growdan and E. RitterWalker (Eds.), Advances in Neurology, Alzheimer's Disease, Vol. 51, Raven Press, New York, pp. 151-161. Masters, C.J. and Holmes, R.S. (1977) The metabolic roles of peroxisomes in mammalian tissues. Int. J. Biochem. 8: 549553. Matsubara, K., Collins, M.A. and Neafsey, E.J. (1992a) Mono-N-methylation of 1,2,3,4-tetrahydro-beta-carbolines in brain cytosol: absence of indole methylation. J. Neurochem. 59:505-510. Matsubara, K., Collins, M.A. and Neafsey, E.J. (1992b) Novel 5-adenosylmethionine-dependent indole-A^-methylation of beta-carbolines in brain particulate fractions. J. Nerdroc/iem. 59:511-518. Matsumoto, S., Matsuo, M., litaka, Y. and Niki, E. (1986) Oxidation of a vitamin E model compound 2,2,5,7,8pentamethylchroman-6-ol, with the r^rr-butylperoxy radical. J. Chem. Soc. Chem. Commun. 14: 1076-1077. Matsuoka, M. and Igisu, H. (1992) Preservation of energy metabolites by carnitine in the mouse brain under ischemia. Brain Res. 590: 334-336. Mayer, M.L. and Miller, R.J. (1991) Excitatory amino acid receptors, second messengers and regulation of intracellular Ca^"*" in mammalian neurons. In: D. Lodge and G.L. Collingridge (Eds.), Trends in Pharmacological Sciences. The Pharmacology of Excitatory Amino Acids, Special Report, Elsevier, Amsterdam, pp. 36-41. Mayeux, R. (1982) Depression and dementia in Parkinson's disease. In: CD. Marsden and S. Fahn (Eds.), Movement Disorders, Butterworths, London, pp. 75-95. Mayeux, R. (1990) The serotonin hypothesis for depression in Parkinson's disease. In: M.B. Streifler, A.D. Korczyn, E. Melamed M.B.H. and Youdim (Eds.), Advances in Neurology: Parkinson's Disease: Anatomy, Pathology and Therapy, Vol. 53, Raven Press, New York, pp. 163-165.

Oxidative stress: free radical production in neural degeneration McArdle, H.J. (1992) The transport of iron and copper across the cell membrane: different mechanisms for different metals? Proc. Nutr. Soc. 51: 199-209. McCay, P.B. (1985) Vitamin E: interactions with free radicals and ascorbate. An/tM. Rev. Nutr. 5: 323-340. McCord, J.M. (1987) Oxygen-derived radicals: a link between reperfusion injury and inflammation. Fed. Proc. 46: 24022406. McCord, J.M. and Fridovich, I. (1969) Superoxide dismutase: an enzymatic function for erythrocuprein (hemocuprein). J. Biol. Chem. 244: 6049-6055. McDermott, J.R., Smith, A.I., Iqbal, K. and Wisniewski, H.M. (1979) Brain aluminium in aging and Alzheimer's disease. Neurology 29: S09-S\4. McGeer, P.L., McGeer, E.G. and Suzuki, J. (1977) Aging and extrapyramidal function. Arc/i. Neurol. 34: 33-35. McGeer, P.L., McGeer. E.G., Suzuki, J., Dolman, C.E. and Nagai, T. (1984) Aging, Alzheimer's disease and the cholinergic system of the basal forebrain. Neurology 34: 741745. McGeer, P.L., Itagaki, S., Akiyama, H. and McGeer, E.C. (1988) Rate of cell death in parkinsonism indicates active neuropathological process. Ann. Neurol. 24: 574-576. Mclntyre, I.M., McCauley, R.B., Filipowicz, C. and Oxenkrug, G.F. (1985) Carbidopa effect on rat brain monoamine oxidase and pineal melatonin. Biol. Psychiatry 20: 809811. McKenna, O., Arnold, G. and Holtzman, E. (1976) Microperoxisome distribution in the central nervous system of the X2X. Brain Res. 117: 181-194. McMurry, T.J. and Groves, J.T. (1986) Metalloporphyrin models for cytochrome P-450. In: P.R. Oritz DeMontellano (Ed.), Cytochrome P-450: Structure, Mechanism and Biochemistry, Plenum Press, New York, pp. 1-28. Meister, A. (1991) Glutathione deficiency produced by inhibition of its synthesis, and its reversal; applications in research and therapy. Pharmacol. Ther. 51: 155-194. Melamed, E. (1992) Biochemical and functional differences between dopamine formed from endogenous tyrosine and exogenous L-DOPA in nigrostriatal dopaminergic neurons. Neurochem. Int. 20(Suppl.): 115S-117S. Meldrum, B. and Garthwaite, J. (1990) Excitatory amino acid neurotoxicity and neurodegenerative disease. Trends Pharmacol Sci. 11:378-387. Meldrum, B. and Garthwaite, J. (1991) Excitatory amino acid neurotoxicity and neurodegenerative disease. In: D. Lodge and G.L. Collingridge (Eds.), Trends in Pharmacological Sciences, The Pharmacology of Excitatory Amino Acids. Special Report, Elsevier, Amsterdam, pp. 54-61. Melloni, E. and Pontremoli, S. (1989) The calpains. Trends Neurosci. 12: 438-444. Mellors, A. and Tappel, A.L. (1966) The inhibition of mitochondrial peroxidation by ubiquinone and ubiquinol. J. Biol. Chem. 2A\^.A^5^-4^56. Mera, S.L. (1991) Aluminium, amyloid, and Alzheimer's disease. Med. Lab. Sci. 48: 283-295.

M.E. Gotz et al. Mertens, H.G. (1947) Zur Klinik des Lathyrismus. Nervenarzt 18: 493-499. Mesnil, M., Testa, B. and Jenner, P. (1984) Xenobiotic metabolism by brain mono-oxygenase and other cerebral enzymes. In: B. Testa (ed.), Advances in Drug Research, Academic Press, London, pp. 95-207. Miatto, O., Gonzalez, R.G., Buonanno, F. and Growdon, J.H. (1986) In vitro ^^P NMR spectroscopy detects altered phospholipid metabolism in Alzheimer's disease. Can. J. Neurol Sci. 13: 535-539. Michel, P.P., Vyas, S. and Agid, Y. (1992) Toxic effects of iron for cultured mesencephalic dopaminergic neurons derived from rat embryonic brains. J. Neurochem. 59: 118127. Mihatsch, W., Russ, H., Gerlach, M., Riederer, P. and Przuntek, H. (1991) Treatment with antioxidants does not prevent loss of dopamine in the striatum of MPTP-treated common marmosets: preliminary observations. J. Neural Transm. (PD Sect.) 3: 173-178. Miller, R.M., Sies, H., Park, E.M. and Thomas, J.A. (1990) Phosphorylase and creatine kinase modification of thiol disulfide exchange and by xanthine oxidase-initiated 5thiolation. Arch. Biochem. Biophys. 276: 355-363. Minn, A., Ghersi-Egea, J.-F., Perrin, R., Leininger, B. and Siest, G. (1991) Drug metabolizing enzymes in the brain and cerebral microvessels. Brain Res. Rev. 16: 65-82. Minotti, G., Di Gennaro, M., D'Ugo, D. and Granone, P. (1991) Possible sources of iron for lipid peroxidation. Free Radical Res. Commun. 12/13: 99-106. Mirabelli, F., Salis, A., Vairetti, M., Bellomo, G., Thor, H. and Orrenius, S. (1989) Cytoskeletal alterations in human platelets exposed to oxidative stress are mediated by oxidative and calcium-dependent mechanisms. Arch. Biochem. Biophys. 270: 478-488. Mithofer, K., Sandy, M.S., Smith, M.T. and Di Monte, D. (1992) Mitochondrial poisons cause depletion of reduced glutathione in isolated hepatocytes. Arch. Biochem. Biophys. 295:132-136. Mizuno, Y., Ohta, S., Tanaka, M., Takamiya, S., Suzuki, K., Sato, T., Oya, H., Ozawa, T. and Kagawa, Y. (1989) Deficiencies in complex I subunits of the respiratory chain in Parkinson's disease. Biochem. Biophys. Res. Commun. 163: 1450-1455. Mogi, M., Harada, M., Kiuchi, K., Kojima, K., Kondo, T., Narabayashi, H., Rausch, W.-D., Riederer, P., Jellinger, K. and Nagatsu, T. (1988) Homospecific activity (activity per enzyme protein) of tyrosine hydroxylase increases in Parkinson's brain. J. Neural Transm. 72: 77-81. Mohr, E., Fabbrini, G., Williams, J., Schlegel, J., Cox, C., Fedio, P. and Chase, T.N. (1989) Dopamine and memory function in Parkinson's disease. Mov. Disord. 4: 113-120. Moll, G., Moll, R., Riederer, P., Gsell, W., Heinsen, H. and Denney, R.M. (1990) Immunofluorescence cytochemistry on thin frozen sections of human substantia nigra for staining of monoamine oxidase A and monoamine oxidase B: a pilot study. J. Neural Transm. 32 (Suppl.): 61-11.

497 MoUer, A. (1992) Mean volume of pigmented neurons in the substantia nigra in Parkinson's disease. Acta Neurol. Scand. 85 (Suppl. 137): 37-39. Monaghan, D.T., Bridges, R.J. and Cotman, C.W. (1989) The excitatory amino acid receptors: their classes, pharmacology, and distinct properties in the function of the central nervous system. Annu. Rev. Pharmacol. Toxicol. 29: 365402. Montastruc, J.L. (1991) Recent advances in the clinical pharmacology of Parkinson's disease. Therapie 46: 293-303. Montpetit, V.J.A., Andermann, F., Carpenter, S., Fawcett, J.S., Zborowska-Sluis, D. and Giberson, H.R. (1971) Subacute necrotizing encephalomyelopathy: a review and a study of two families. Brain 94: 1-30. Mordit, R.C., Walton, J.C, Burton, G.W., Hughes, L., Ingold, K.U. and Lindsay, D.A. (1991) Exploratory study of betacarotene autoxidation. Tetrahedron Lett. 32: 4203-4206. Morris, CM., Candy, J.M., Oakley, A.E, Bloxham, C.A. and Edwardson, J.A. (1992a) Histochemical distribution of nonhaem iron in the human brain. Acta Anat. 144: 235-257. Morris, CM., Keith, A.B., Edwardson, J.A. and Pullen, R.G.L. (1992b) Uptake and distribution of iron and transferrin in the adult rat brain. J. Neurochem. 59: 300-306. Mortimer, A.J., Pirozzolo, F.J., Hansch, E.C and Webster, D.D. (1982) Relationship of motor symptoms to intellectual deficits in Parkinson's disease. Neurology 32: 133-137. Muller, D.P.R. and Goss-Sampson, M.A. (1990) Neurochemical, neurophysiological and neuropathological studies in vitamin E deficiency. Crit. Rev. Neurobiol. 5: 239-265. Miiller-Hill, B. and Beyreuther, K. (1989) Molecular biology of Alzheimer's disease. AnnM. Rev. Biochem. 58: 287-307. Munoz-Garcia, D., Pendlebury, W.W., Kessler, J.B. and Perl, D.P. (1986) An immunocytochemical comparison of cytoskeletal proteins in aluminium-induced and Alzheimertype neurofibrillary tangles. Acta Neuropathol. 70: 243248. Murphy, G.M., Greenberg, B.D., Ellis, W.G., Forno, L.S., Salamat, S.M., Gonzalez-DeWitt, P.A., Lowery, D.E. and Tinklenberg, J.R. (1992) Alzheimer's disease. )3-Amyloid precursor protein expression in the nucleus basalis of Meynert. Am. J. Pathol. 141: 357-361. Nagatsu, T. (1990) Changes of tyrosine hydroxylase in Parkinsonian brains and in the brain of MPTP-treated mice. In: M.B. Streifler, A.D. Korczyn, E. Melamed and M.B.H. Youdim (Eds.), Advances in Neurology: Parkinson's Disease Anatomy, Pathology and Therapy, Vol. 53, Raven Press, New York, pp. 207-214. Nagatsu, T. and Yoshida, M. (1988) An endogenous substance of the brain, tetrahydroisoquinoline, produces parkinsonism in primates with decreased dopamine, tyrosine hydroxylase and biopterin in the nigrostriatal regions. Neurosci.Lett. 87: 178-182. Nair, V. and Turner, G.A. (1984) The thiobarbituric acid test for lipid peroxidation: structure of the adduct with malondialdehyde. Lipids 19: 804-805. Nakagawa-Hattori, Y., Yoshino, H., Kondo, T., Mizuno, Y.

498 and Horai, S. (1992) Is Parkinson's disease a mitochondrial disorder?/. Neurol. Sci. 107: 29-33. Nakajima, M., Kashiwagi, K., Hayashi, Y., Saito, M., Kawashima, T., Furukawa, S. and Kobayashi, K. (1991) Alpha-tocopherol supports the survival and neurite extension of neurons cultured from various regions of fetal rat brain. Neurosci. Lett. 133: 49-52. Nakamura, S. and Yamazaki, I. (1969) One-electron transfer reactions in biochemical systems. IV. A mixed mechanism in the reaction of milk xanthine oxidase with electron acceptors. Biochim. Biophys. Acta 189: 29-37. Nakamura, S., Kawamata, T., Akiguchi, I., Kameyama, M., Nakamura, N. and Kimura, H. (1990) Expression of monoamine oxidase B activity in astrocytes of senile plaques. Acta Neuropathol. 80: 419-425. Naoi, M., Matsuura, S., Parvez, H., Takahashi, T., Hirata, Y., Minami, M. and Nagatsu, T. (1989a) Oxidation of A^methyl-l,2,3,4-tetrahydroisoquinoline into the A^-methylisoquinolinium ion by monoamine oxidase. J. Neurochem. 52:653-655. Naoi, M., Matsuura, S., Takahashi, T. and Nagatsu, T. (1989b) An A^-methyltransferase in human brain catalyses //-methylation of 1,2,3,4-tetrahydroisoquinoline into A^methyl-l,2,3,4-tetrahydroisoquinoline, a precursor of a dopaminergic neurotoxin, A^-methyhsoquinolinium ion. Biochem. Biophys. Res. Commun. 161: 1213-1219. Naoi, M., Takahashi, T., Parvez, H., Kabeya, R., Taguchi, E., Yamaguchi, K., Hirata, Y., Minami, M. and Nagatsu, T. (1989c) A^-Methylisoquinolinium ion as an inhibitor of tyrosine hydroxylase, aromatic L-amino acid decarboxylase and monoamine oxidase. Neurochem. Int. 15: 315-320. Nathan, C.F., Arrick, B.A., Murray, H.W., DeSantis, N.M. and Cohn, Z.A. (1980) Tumor cell anti-oxidant defenses. /. Exp. Med. 153: 766-782. Nebert, D.W., Nelson, D.R. and Feyereisen, R. (1989) Evolution of cytochrome P-450 genes. Xenobiotica 19: 11491160. Nelson, J.S. (1987) Effects of free radical scavengers on the neuropathology of mammalian vitamin E deficiency. In: O. Hayaishi and M. Mino (Eds.), Clinical and Nutritional Aspects of Vitamin E, Elsevier, Amsterdam, pp. 157-159. Nicklas, W.J., Vyas, I. and Heikkila, R.E. (1985) Inhibition of NADH-linked oxidation in brain mitochondria by 1-methyl4-phenyl-pyridine, a metabolite of the neurotoxin, 1methyl-4-phenyl-l,2,3,6-tetrahydropyridine. Life Sci. 36: 2503-2508. Nicotera, P., McConkey, D.J., Dypbukt, J.M., Jones, D.P. and Orrenius, S. (1989) Calcium-activated mechanisms in cell killing. Drug Metab. Rev. 20: 193-201. Nieuwenhuys, R., Voogd, J. and Huijzen (1991) Das Zentralnervensystem des Menschen, Springer, Berlin. Niijima, K., Araki, M., Ogawa, M., Suzuki, K., Mizuno, Y., Nagatsu, I., Kimura, H., Yoshida, M. and Nagatsu, T. (1991) A^-Methylisoquinolinium ion (NMIQ"*") destroys cultured mesencephalic dopamine neurons. Biog. Amines 8: 61-57.

Oxidative stress: free radical production in neural degeneration Niki, E. (1987a) Antioxidants in relation to hpid peroxidation. Chem. Phys. Lipids 44: 227-253. Niki, E. (1987b) Interaction of ascorbate and alphatocopherol. Ann. N Y. Acad. Sci. 498: 186-199. Niki, E., Kawakami, A., Yamamoto, Y. and Kamiya, Y. (1985) Oxidation of lipids. VIII. Synergistic inhibition of oxidation of vitamin E and vitamin C. Bull. Chem. Soc. Jpn. 58: 1917-1975. Nistico, G., Ciriolo, M.R., Fiskin, K., lannone, M., De Martino, A. and Rotiho, G. (1992) NGF restores decrease in catalase activity and increases superoxide dismutase and glutathione peroxidase activity in the brain of aged rats. Free Radical Biol. Med. 12: 177-181. Nitsch, R.M., Blusztajn, J.K., Pittas, A.G., Slack, B.E., Growdon, J.H. and Wurtman, R.J. (1992) Evidence for a membrane defect in an Alzheimer's disease brain. Proc. Natl. Acad. Sci. USA 89: 1671-1675. Nohl, H. and Jordan, W. (1986) The mitochondrial site of superoxide formation. Biochem. Biophys. Res. Commun. 138: 533-539. Nohl, H. and Stolze, K. (1992) Ubisemiquinones of the mitochondrial respiratory chain do not interact with molecular oxygen. Free Radical Res. Commun. 16: 409-419. Nordgren, L., Rorsman, H., Rosengren, A.-M. and Rosengren, E. (1971) L-DOPA and dopamine in the pigment of substantia nigra. Experientia 27: 1178-1179. Novelli, A., Reilly, J.A., Lysko, P.G. and Henneberry, R.C. (1988) Glutamate becomes neurotoxic via the A^-methyl-Daspartate receptor when intracellular energy levels are reduced. Brain Res. 451: 205-212. Novotny, E.J., Singh, G., Wallace, D.C., Dorfman, L.J., Louis, A., Sogg, R.L. and Steinman, L. (1986) Leber's disease and dystonia: a mitochondrial disease. Neurology 36: 1053-1060. O'Brien, P.J. (1991) Molecular mechanisms of quinone cytotoxicity. Chem.-Biol. Interact. 80: 1-41. Odunze, I.N., Klaidman, L.K. and Adams, J.D. (1990) MPTP toxicity in the mouse brain and vitamin E. Neurosci. Lett. 108: 346-349. Ohta, S., Tachikawa, O., Makino, Y., Tasaki, Y. and Hirobe, M. (1990) Metabolism and brain accumulation of tetrahydroisoquinoline (TIQ) a possible parkinsonism inducing substance, in an animal model of a poor debrisoquine metabolizer. Life Sci. 46: 599-605. Olanow, C.W. (1992) A rationale for dopamine agonists as primary therapy for Parkinson's disease. Can. J. Neurol. Sci. 19: 108-112. Olanow, C.W. and Calne, D. (1992) Does selegiline monotherapy in Parkinson's disease act by symptomatic or protective mechanisms? Neurology 42 (Suppl. 4): 13-26. Olanow, C.W., Holgate, R.C. and Murtaugh, R. (1989) MR imaging in Parkinson's disease and aging. In: D.B. Calne, G. Comi and D. Crippa (Eds.), Parkinson's Disease and Aging, Vol. 26, Raven Press, New York, pp. 155-164. OUat, H. (1992) Dopaminergic insufficiency reflecting cerebral ageing: value of a dopaminergic agonist, piribedil. J. Neurol. 239 (Suppl. 1): S13-S16.

M.E.

Gotzetal

Olney, J.W. (1990) Excitotoxic amino acids and neuropsychiatric disorders. Annu. Rev. Pharmacol. Toxicol. 30: 4 7 71. Olney, J.W., Zorumski, C.F., Stewart, G.R., Price, M.T., Wang, G. and Labruyere, J. (1990) Excitotoxicity of LDOPA and 6-OH-DOPA: implications for Parkinson's and Huntington's disease. Exp. Neurol 108: 269-272. Oreland, L., Arai, Y. and Stenstrom, A. (1983) The effect of deprenyl (selegiline) on intra and extraneuronal dopamine oxidation. Acta Neurol. Scand. 95: 81-85. Oreland, L., Fowler, C.J., Carlsson, A. and Magnusson, T. (1989) The effect of hemitransection of rats upon the brain monoamine oxidase MAO-A and MAO-B activity. Life Sci. 26: 139-146. Orrenius, S., McConkey, D.J., Bellomo, G. and Nicotera, P. (1989) Role of Ca^"^ in toxic cell killing. Trends Pharmacol. Sci. 10:281-285. Oshino, N. and Chance, B. (1977) Properties of glutathione release observed during reduction of organic hydroperoxide, demethylation of aminopyrine and oxidation of some substances in perfused rat liver, and their implications for the physiological function of catalase. Biochem. J. 162: 509-525. Oya, Y. and Yamamoto, K. (1988) The biological activity of hydrogen peroxide. IV. Enhancement of its clastogenic actions by coadministration of L-histidine. Mutat. Res. 198: 233-240. Ozaki, M., Kawabata, T. and Awai, M. (1988) Iron release from haemosiderin and production of iron-catalysed hydroxyl radicals in vitro. Biochem. J. 250: 589-595. Ozawa, T., Tanaka, M., Ikebe, S., Ohno, K., Kondo, T. and Mizuno, Y. (1990) Quantitative determination of deleted mitochondrial DNA relative to normal DNA in parkinsonian striatum by a kinetic PCR analysis. Biochem. Biophys. Res. Commun. 172: 4 8 3 ^ 8 9 . Pacifici, R.E. and Davies, K.J.A. (1991) Protein, lipid and DNA repair systems in oxidative stress: the free-radical theory of aging revisited. Gerontology 37: 166-180. Pacifici, R.E., Salo, D.C. and Davies, K.J.A. (1989) Macroxyproteinase (MOP): a 670-kDa proteinase complex that degrades oxidatively denatured proteins in red blood cells. Free Radical Biol. Med. 1: 521-536. Packer, J.E., Slater, T.F. and Willson, R.L. (1979) Direct observation of a free radical interaction between vitamin E and vitamin C. Nature 278:131-13S. Padh, H. (1991) Vitamin C: newer insights into its biochemical functions. Nutr. Rev. 49: 65-70. Pai, K.S. and Ravindranath, V. (1991) Protection and potentiation of MPTP-induced toxicity by cytochrome P-450 inhibitors and inducer: in vitro studies with brain slices. Brain Res. 555: 239-244. Palmer, A.M., Wilcock, G.K., Esiri, M.M., Francis, P.T. and Bowen, D.M. (1987) Monoaminergic innervation of the frontal and temporal lobes in Alzheimer's disease. Brain Res. 401: 23\-23S. Paraidathathu, T., De Groot, H. and Kehrer, J.P. (1992) Pro-

499 duction of reactive oxygen by mitochondria from normoxic and hypoxic rat heart tissue. Free Radical Biol. Med. 13: 289-297. Parker, Jr., W.D., Boyson, S.J. and Parks, J.K. (1989) Abnormalities of the electron transport chain in idiopathic Parkinson's disease. Ann. Neurol. 26: 719-723. Parker, Jr., W.D., Filley, C M . and Parks, J.K. (1990) Cytochrome oxidase deficiency in Alzheimer's disease. Neurology A0\ 1302-1303. Parkinson, J. (1817) Essay on the Shaking Palsy. Printed by Whittingham and Rowland, Goswell Street, for Sherwood Neely and Jones, Paternoster Row, London. Pasik, P., Pasik, T. and Holstein, G.R. (1986) Ultrastructural chemoanatomy of the basal ganglia: an overview. In: M.D. Yahr and K.J. Bergmann (Eds.), Advances in Neurology, Parkinson's Disease. Vol. 45, Raven Press, New York, pp. 59-65. Paterson, I.A., Juorio, A.V. and Boulton, A.A. (1990) 2Phenyl-ethylamine: a modulator of catecholamine transmission in the melanin central nervous system? /. Neurochem. 55: 1827-1837. Patten, B.M., Kurlander, H.M. and Evans, B. (1982) Free amino acid concentrations in spinal tissue from patients dying of motor neuron disease. Acta Neurol. Scand. 66: 594-599. Paul, S.M., Axelrod, J. and Diliberto, Jr., E.J. (1977) Catecholestrogen-forming enzyme of brain: demonstration of a cytochrome P-450 monooxygenase. Endocrinology 101:1604-1610. Paulus, W. and Jellinger, K. (1991) The neuropathologic basis of different clinical subgroups of Parkinson's disease. J. Neuropathol. Exp. Neurol. 50: 743-755. Pechan, P.A., Chowdhury, K. and Seifert, W. (1992) Free radicals induce gene expression of NGF and bFGF in rat astrocyte culture. NeuroReport 3: 469-472. Peeters-Joris, C , Vandevoorde, A.-M. and Baudhuin, P. (1975) Subcellular localization of superoxide dismutase in rat liver. Biochem. J. 150: 31-39. Perl, D.P. and Brody, A.R. (1980) Alzheimer's disease: X-ray spectrophotometric evidence of aluminium accumulation in neurofibrillary tangle-bearing neurons. Science 208: 2 9 7 299. Perl, D.P. and Good, P.F. (1992) Comparative techniques for determining cellular iron distribution in brain tissues. Ann. Neurol32 (Supply. 16-Sl. Perl, D.P. and Pendlebury, W.W. (1987) Neuropathology of Alzheimer's disease and related dementias. In: H.Y. Meltzer (Ed.), Psychopharmacology: The Third Generation of Progress, Raven Press, New York, pp. 881-885. Perry, E.K (1986) The cholinergic hypothesis ten years on. Br. Med. Bull. 42: 63-69. Perry, E.K. (1987) Cortical neurotransmitter chemistry in Alzheimer's disease. In: H.Y. Meltzer (Ed.), Psychopharmacology: The Third Generation of Progress, Raven Press, New York, pp. 887-895. Perry, E.K., Tomlinson, B.E, Blessed, G., Bergman, K., Gib-

500 son, P.H. and Perry, R.H. (1978) Correlation of cholinergic abnormalities with senile plaques and mental test scores in senile dementia. Br. Med. J. ii: 1403. Perry, E.K, Perry, R.H., Smith, C.J., Dick, D.J., Candy, J.M., Edwardson, J.A., Fairbairn, A. and Blesseds, G. (1987) Nicotinic receptor abnormalities in Alzheimer's and Parkinson's diseases. J. Neurol. Neurosurg. Psychiatry 50: 806809. Perry, R.H. (1986) Recent advances in neuropathology. Br. Med. Bull. 42: 34-41. Perry, T.L. and Yong, V.W. (1986) Idiopathic Parkinson's disease, progressive supranuclear palsy and glutathione metabolism in the substantia nigra of patients. Neurosci Lett. 67: 269-274. Perry, T.L., Yong, V.W., Clavier, R.M., Jones, K., Wright, J.M., Foulks, J.G. and Wall, R.A. (1985) Partial protection from the dopaminergic neurotoxin //-methyl-4-phenyl1,2,3,6-tetrahydropyridine by four different antioxidants in mouse. Neurosci. Lett. 60: 109-114. Perry, T.L., Yong, V.W., Hansen, S., Jones, K., Bergeron, C , Foulks, J.G. and Wright, J.M. (1987) Alpha-tocopherol and y5-carotene do not protect marmosets against the dopaminergic neurotoxicity of l-methyl-4-phenyl-1,2,3,6tetrahydropyridine. J. Neurol. Sci. 81: 321-331. Perry, T.L., Bergerson, C , Steele, J.C, McLachlan, D.R. and Hansen, S. (1991) Brain amino acid contents are dissimilar in sporadic and Guamanian amyotrophic lateral sclerosis. J. Neurol. Sci. 99: 3-8. Petruzzella, V., Baggetto, L.G., Renin, F., Cafagna, F., Ruggiero, F.M., Cantatore, P and Gadaleta, M.N. (1992) In vivo effect of acetyl-L-carnitine on succinate oxidation, adenine nucleotide pool and lipid composition of synaptic and nonsynaptic mitochondria from cerebral hemispheres of senescent rats. Arc/i. Gerontol. Geriatr. 14: 131-144 Pettegrew, J.W., Panchalingam, K., Moossy, J., Martinez, J., Rao, G. and Boiler, F. (1988) Correlation of phosphorous31 magnetic resonance spectroscopy and morphology finding in Alzheimer's disease. Arch. Neurol. 45: 10931096. Philbert, M.A., Beiswanger, C M . , Waters, D.K., Reuhl, K.R. and Lowndes, H.E. (1991) Cellular and regional distribution of reduced glutathione in the nervous system of the rat: histochemical localization by mercury orange and ophthaldialdehyde-induced histofluorescence. Toxicol. Appl. Pharmacol. 107: 215-227. Pilas, B., Sarna, T., Kalyanaraman, B. and Swartz, H.M. (1988) The effect of melanin on iron associated decomposition of hydrogen peroxide. Free Radical Biol. Med. 4: 285-293. Pileblad, E., Miagnusson, T. and Fornstedt, B. (1991) Reduction of brain glutathione by L-buthionine sulfoximine potentiates the dopamine-depleting action of 6-hydroxydopamine in rat striatum. /. Neurochem. 52: 978-980. Pincus, J.H. (1972) Subacute necrotizing encephalomyelopathy (Leigh's disease): a consideration of clinical features and etiology. Dev. Med. Child Neurol. 14: 87-101.

Oxidative stress: free radical production in neural

degeneration

Plaitakis, A. (1990) Glutamate dysfunction and selective motoneuron degeneration in amyotrophic lateral sclerosis: a hypothesis. An«. Neurol. J. 28: 3-8. Plaitakis, A, Constantakakis, E. and Smith, J. (1988) The neuroexcitotoxic amino acids glutamate and aspartate are altered in spinal cords and brain in amyotrophic lateral sclerosis. Ann. Neurol. 24: 446-449. Playford, E.D. and Brooks, D.J. (1992) In vivo and in vitro studies of the dopaminergic system in movement disorders. Cerebrovasc. Brain Metab. Rev. 4: 144-171. Poewe, W.H., Lees, A.J. and Stern, G.M. (1986a) Low-dose L-dopa therapy in Parkinson's disease: a 6-year follow-up sWidy. Neurology ?>6'. 1528-1530. Poewe, W.H., Lees, A.J. and Stern, G.M. (1986b) Treatment of motor fluctuations in Parkinson's disease with an oral sustained-release preparation of L-Dopa: clinical and pharmacological observations. Clin. Neuropharmacol. 9: 4 3 0 439. Poirier, J., Roy, M., Campanella, G., Cloutier, T. and Paris, S. (1987) Debrisoquine metabolism in parkinsonian patients treated with antihistamine drugs. Lancet ii: 386. Pollard, H.B., Dhariwal, K., Adeyemo, O.M., Markey, C.J., Caohuy, H., Levine, M., Markey, S. and Youdim, M.B.H. (1992) A parkinsonian syndrome induced in the goldfish by the neurotoxin MPTP. Fed. Proc. Soc. Exp. Biol. 6: 31083116. Porsche-Wiebking, E. (1989) New A^-methyl-D-aspartate antagonists for the treatment of stroke. Drug Dev. Res. 17: 367-375. Powis, G. (1989) Free radical formation by antitumor quinones. Free Radical Biol. Med. 6: 63-101. Prohaska, J.R. and Ganther, H.E. (1976) Selenium and glutathione peroxidase in developing rat brain. J. Neurochem. 27: 1379-1387. Prota, G., Rorsman, H., Rosengren, A.-M. and Rosengren, E. (1976) Phaeomelanic pigments from a human melanoma. Experientia 32: 970-971. Pryor, W.A. (1986) Oxy-radicals and related species: their formation, lifetimes, and reactions. Annu. Rev. Physiol. 48: 657-667. Pryor, W.A. and Godber, S.S. (1991) Noninvasive measures of oxidative stress status in humans. Free Radical Biol. Med. 10: 177-184. Pryor, W.A. and Porter, N.A. (1990) Suggested mechanisms for the production of 4-hydroxy-2-nonenal from the autoxidation of polyunsaturated fatty acids. Free Radical Biol. Med. 8: 541-543. Quinlan, G.J., Halliwell, B., Moorhouse, C.P. and Gutteridge, J.M.C. (1988) Action of lead(II) and aluminium(III) ions on iron-stimulated lipid peroxidation in liposomes, erythrocytes and rat liver microsomal fractions. Biochim. Biophys. Acta 962: 196-200. Quirion, R., Martel, J.C, Robitaille, Y., Etienne, P., Wood, P., Nair, N.P.V. and Gauthier, S. (1986) Neurotransmitter and receptor deficits in senile dementia of the Alzheimer type. Can. J. Neurol. Sci. 13: 503-510.

M.E. Gotz. et al. Rabey, J.M., Schwartz, M., Graff, E., Harsat, A. and Vered, Y. (1991) The influence of bromocriptine on the pharmacokinetics of levodopa in Parkinson's disease. Clin. Neuropharmacol. 14:514-522. Rajput, A.H., Stern, W. and Laverty, W.H. (1984) Chronic low-dose levodopa therapy in Parkinson's disease. Neurology 34: 991-996 Rajput, A.H., Uitti, R.J., Stern, W., Laverty, W., O'Donnell, K., O'Donnell, D., Yuen, W.K. and Dua, A. (1987) Geography, drinking water chemistry, pesticides and herbicides and the etiology of Parkinson's disease. Can. J. Neurol. Sci. 14:414-418. Ramacci, M.T., De Rossi, M., Lucreziotti, M.R., Mione, M.C. and Amenta, F. (1988) Effects of long-term treatment with acetyl-L-carnitine on structural changes of ageing rat brain. Drugs Exp. Clin. Res. 14: 112-115. Ramsay, R.R. and Singer, T.P. (1986) Energy-dependent uptake of yV-methyl-4-phenylpyridinium, the neurotoxic bioactivation product of l-methyl-4-phenyl-1,2,3,6tetrahydropyridine, by mitochondria. J. Biol. Chem. 261: 7585-7587. Ramsay, R.R., Salach, J.I., Dadgar, J. and Singer, T.P. (1986a) Inhibition of mitochondrial NADH dehydrogenase by pyridine derivatives and its possible relation to experimental and idiopathic Parkinsonism. Biochem. Biophys. Res. Commun. 135:269-275. Ramsay, R.R., Salach, J.I. and Singer, T.P. (1986b) Uptake of the neurotoxin l-methyl-4-phenylpyridine (MPP"*") by mitochondria and its relation of the inhibition of the mitochondrial oxidation of NAD'*"-linked substrates by MPP"^. Biochem. Biophys. Res. Commun. 134: 743-748. Ramsay, R.R., Krueger, M.J., Youngster, S.K., Gluck, M.R., Casida, J.E. and Singer, T.P. (1991) Interaction of 1methyl-4-phenylpyridinium ion (MPP"*") and its analogs with the rotenone/piericidin binding site of NADH dehydrogenase. 7. Neurochem. 56: 1184-1190. Randot, P. and Ziegler, M. (1992) Activity and acceptability of piribedil in Parkinson's disease: a multicentre study. /. Neurol. 239 (Suppl. 1): S28-S34. Rao, K.S. and Loeb, L.A. (1992) DNA damage and repair in brain: relationship to aging. Mutat. Res. 275: 317-329. Raps, S.P., Lai, J.C.K., Hertz, L. and Cooper, A.J.L. (1989) Glutathione is present in high concentrations in cultured astrocytes but not in cultured neurons. Brain Res. 493: 3 9 8 401. Rasmussen, H. and Goodman, D.B.P. (1977) Relationship between calcium and cyclic nucleotides in cell activation. Physiol. Rev. 57: 421-509. Ravindranath, V. and Anandatheerthavarada, H.K. (1989) High activity of cytochrome P-450 linked aminopyrine A^demethylase in mouse brain microsomes and associated sex-related differences. Biochem. J. 261: 769-773. Ravindranath, V. Anandatheerthavarada, H.K. and Shankar, S.K. (1990) NADPH cytochrome P-450 reductase in rat, mouse and human brain. Biochem. Pharmacol. 39: 10131018.

501 Reddy, C.C, Scholz, R.W., Thomas, C.E. and Massaro, E.J. (1982) Vitamin E dependent reduced glutathione inhibition of rat liver microsomal lipid peroxidation. Life Sci. 31: 571-576. Reddy, V.V.R., Naftolin, F. and Ryan, K.J. (1974) Conversion of androstenedione to estrone by neural tissues from fetal and neonatal rats. Endocrinology 94: 117-121. Reed, D.J., Babson, J.R., Beatty, P.W., Brodie, A.E., Ellis, W.W. and Potter, D.W. (1980) High performance liquid chromatography analysis of nanomole levels of glutathione, glutathione disulfide, and related thiols and disulfides. Anal. Biochem. 106: 55-62. Reglinski, J., Hoey, S., Smith, W.E. and Sturrock, R.D. (1988) Cellular response to oxidative stress at sulfhydryl group receptor sites on the erythrocyte membrane. /. Biol Chem. 263:12360-12366. Rehncrona, S., Folbergrova, J., Smith, D.S. and Siesjo, B.K. (1980) Influence of complete and pronounced incomplete cerebral ischemia and subsequent recirculation on cortical concentrations of oxidized and reduced glutathione in the rat. J. Neurochem. 34: 4 7 7 ^ 8 6 . Reif, D.W. (1992) Ferritin as a source of iron for oxidative damage. Free Radical Biol. Med. 12: 417-427. Reynolds, F., Dewan, V. and Morgan, B. (1992) Is aluminium a dementing ion? Lancet i: 713-714. Reynolds, G.P., Elsworth, J.D., Blau, K., Sandler, M., Lees, A.J. and Stern, G.M. (1978) Deprenyl is metabolized to methamphetamine and amphetamine in man. Br. J. Clin. Pharmacol. 6: 542-544. Reynolds, G.P., Pearson, S.J., Halket, J. and Sandler, M. (1988) Brain quinolinic acid in Huntington's disease. J. Neurochem. 50: 1959-1960. Richter, C. (1988) Do mitochondrial DNA fragments promote cancer and aging? FEBSLett. lAX: 1-5. Richter, C. (1992) Reactive oxygen and DNA damage in mitochondria. Mutat. Res. 275: 249-255. Richter, C , Park, J.W. and Ames, B.N. (1988) Normal oxidative damage to mitochondrial and nuclear DNA is extensive. Proc. Natl. Acad Sci. USA 85: 6465-6467. Riederer, P. and Jellinger, K. (1983) Neurochemical insights into, MAO inhibitors, with special reference to deprenyl. Acta Neurol. Scand. 95 (Suppl.): 43-55. Riederer, P. and Przuntek, H. (1987) MAO-B-inhibitor selegiline (/?-(-)-deprenyl): new therapeutic concept in the treatment of Parkinson's disease. J. Neural Transm. 25 (Suppl.): 1-194. Riederer, P. and Wuketich, S. (1976) Time course of nigrostriatal degeneration in Parkinson's disease. J. Neural Transm. 38:277-301. Riederer, P., Youdim, M.B.H., Rausch, W.-D., Birkmayer, W., Jellinger, K. and Seemann, D. (1978) On the mode of action of L-deprenyl in the human central nervous system. J. Neural Transm. 43: 217-226. Riederer, P., Strolin-Benedetti, M., Dostert, P., Sofic, E., Heuschneider, G. and Guffroy, C. (1987) Do glutathione and ascorbic acid play a role in the neurotoxicity of 1-

502 methyl-4-phenyl-l,2,3,6-tetrahydropyriciine (MPTP)? Pharmacol Toxicol. 60 (SuppL): 39. Riederer, P., Konradi, C , Hebenstreit, G. and Youdim, M.B.H. (1989a) Neurochemical perspectives to the function of monoamine oxidase. Acta Neurol. Scand. 126: 4 1 ^ 5 . Riederer, P., Sofic, E., Konradi, C , Kornhuber, J., Beckmann, H., Dietl, M., Moll, G. and Hebenstreit, G. (1989b) The role of brain dopamine. In: E. Fliickinger, E.E. Miiller and M.O. Thorner (Eds.), Basic and Clinical Aspects ofNeuroscience, Springer, Heidelberg, pp. 1-17. Riederer, P., Sofic, E., Rausch, W.D., Schmidt, B., Reynolds, G.P., Jellinger, K. and Youdim, M.B.H. (1989c) Transition metals, ferritin, glutathione, and ascorbic acid in parkinsonian brains. J. Neurochem. 52: 515-520. Riederer, P., Lange, K.W., Kornhuber, J. and Danielczyk, W. (1992) Glutamatergic-dopaminergic balance in the brain: its importance in motor disorders and schizophrenia. Drug Res. 42: 265-268. Rigo, A., Stevenato, R., Finazzi-Agro, A. and Rotilio, G. (1977) An attempt to evaluate the rate of the Haber-Weiss reaction by using hydroxy 1 radical scavengers. FEES Lett. 80: 130-132. Riklan, M., Whelihan, W. and Cullinan, T. (1976) Levodopa and psychometric tests performance in parkinsonism 5 years later. Neurology 26: 173-179. Rinne, U.K and Heinonen, E.H. (1991) New approaches to the treatment of early Parkinson's disease. Acta Neurol Scand. 84(Suppl. 136): 1-98. Rinne, U.K, L6nnberg, P. and Koskinen, V. (1981) Dopamine receptors in the parkinsonian brain. J. Neural Transm. 51: 97-106. Rivett, A.J. (1985) Purification of a liver alkaline protease which degrades oxidatively modified glutamine synthetase: characterization as a high molecular weight cysteine protease. J. Biol Chem. 260: 12600-12606. Roberts, R., Sandra, A., Siek, G.C., Lucas, J.J. and Fine, R.E. (1992) Studies of the mechanism of iron transport across the blood-brain barrier. Ann. Neurol 32 (Suppl.): 43-50. Robertson, H.A. (1992) Synergistic interactions of Dp and D2-selective dopamine agonists in animal models for Parkinson's disease: sites of action and implications for the pathogenesis of dyskinesias. Can. J. Neurol Scl 19: 147152. Robinson, D.S., Davis, J.M., Nils, A., Ravaris, C.L. and Sylvester, D. (1971) Relation of sex and aging to monoamine oxidase activity of human brain, plasma and platelets. Arch. Gen. Psychiatry 24: 536-539. Rodgers, A.D. and Curzon, D. (1975) Melanin formation by human brain in vitro. J. Neurochem. 24: 1123-1129. Roe, P.F. (1964) Familial motor neuron disease. J. Neurol Neurosurg. Psychiatry 27: 140-143. Rolls, E.T. (1990) Functions of neuronal networks in the hippocampus and of backprojections in the cerebral cortex in memory. In: J.L. McGaugh, N.M. Weinberger and G. Lynch (Eds.), Brain Organization and Memory, Oxford University Press, Oxford, pp. 184-210.

Oxidative stress: free radical production in neural

degeneration

Rommelspacher, H., May, T. and Susilo, R. (1991) )3-Carbolines and tetrahydroisoquinolines: detection and function in mammals. Planta Med. 57 (Suppl. 1): 85-92. Rose, R.C. (1989) The ascorbate redox potential of tissues: a determinant or indicator of disease? News Physiol ScL 4: 190-195. Rosen, A.D. (1978) Amyotrophic lateral sclerosis clinical features and prognosis. Arch. Neurol 35: 639-642. Rosengren, E., Linder-Eliasson, E. and Carlsson, A. (1985) Detection of 5-5-cysteinyldopamine in human brain. J. Neural Transm. 63: 247-253. Rosenthal, R.E., Williams, R., Bogaert, Y.E., Getson, P.R. and Fiskum, G. (1992) Prevention of post-ischemic canine neurological injury through potentiation of brain energy metabolism by acetyl-L-carnitine. Stroke 23: 1312-1318. Ross, S.M., Roy, D.N. and Spencer, P.S. (1989) P-Noxalylamino-L-alanine action on glutamate receptors. J. Neurochem. 53: 710-715. Rossi, M.A., Fidale, F., Garramone, A., Esterbauer, H. and Dianzani, M.U. (1990) Effect of 4-hydroxyalkenals on hepatic phosphatidylinositol-4,5-biphosphate-phospholipase C. Biochem. Pharmacol 39: 1715-1719. Rossor, M.N. (1985) Transmitter deficits in Alzheimer's disease. Neurochem. Int. 7: 567-570. Roth, M. (1986) The association of the clinical and neurobiological findings and its bearing on the classification and etiology of Alzheimer's disease. Br. Med. Bull 42: 42-50. Rothman, S.M. (1984) Synaptic release of excitatory amino acid neurotransmitter mediates anoxic neuronal death. J. NeuroscL4: 1884-1891. Rothman, S.M. and Olney, J.W. (1986) Glutamate and the pathophysiology of hypoxic-ischemic brain damage. Ann. Neurol 19:105-111. Rotruck, J.T., Pope, A.L., Ganther, H.E., Swanson, A.B., Hafeman, D. and Hoekstra, W.G. (1973) Selenium: biochemical role as a component of glutathione peroxidase. Science 179: 588-590. Rousseau, E.J., Davison, A.J. and Dunn, B. (1992) Protection by p-csLToiene and related compounds against oxygenmediated cytotoxicity and genotoxicity: implications for carcinogenesis and anticarcinogenesis. Free Radical Biol Med. 13: 407-433. Rowland, L.P. (1984) Motor neuron disease and amyotrophic lateral sclerosis. Trends NeuroscL 4: 110-112. Rowley, D.A. and Halliwell, B. (1982) Superoxide-dependent formation of hydroxyl radicals in the presence of thiol compounds. FEBS Lett. 138: 33-36. Ruberg, M., Ploska, A., Yavoy-Agid, F. and Agid, Y. (1989) Muscarinic binding and choline acetyltransferase activity in parkinsonian subjects with reference to dementia. Brain Res. 232: 129-139. Saggu, H., Cooksey, J., Dexter, D.T., Wells, F.R., Lees, A., Jenner, P. and Marsden, C D . (1989) A selective increase in particulate superoxide dismutase activity in parkinsonian substantia nigra. J. Neurochem. 53: 692-697. Sah, P., Hestrin, S. and Nicoll, R.A. (1989) Tonic activation

M.E. Gotz et al. of NMDA receptors by ambient glutamate enhances excitability of neurons. Science 246: 815-818. Sahu, S.C. (1991) Role of oxygen free radicals in the molecular mechanisms of carcinogenesis: a review. Environ. Carcinog. Ecotoxicol. Rev. C9: 83-112. Salo, P.T. and Tatton, W.G. (1992) Deprenyl reduces the death of motoneurons caused by axotomy. J. Neurosci. Res. 31:394-400. Samuni, A., Aronovitch, J., Godinger, D., Chevion, M. and Czapski, G. (1983) On the cytotoxicity of vitamin C and metal ions. Eur. J. Biochem. 137: 119-124. Sandri, G., Panfili, E. and Ernster, L. (1990) Hydrogen peroxide production by monoamine oxidase in isolated rat-brain mitochondria: its effect on glutathione levels and Ca^"*" efflux. Biochim. Biophys. Acta 1035: 300-305. Sandyk, R. and Willis, G.L. (1992) Amine accumulation: a possible precursor of Lewy body formation in Parkinson's disease. Int. J. Neurosci. 66: 61-74. Saran, M. and Bors, W. (1991) Direct and indirect measurements of oxygen radicals. Klin. Wochenschr./Clin. Invest. 69: 957-964. Sasame, H.A., Ames, M.M. and Nelson, S.D. (1977) Cytochrome P-450 and NADPH-cytochrome c reductase in rat brain: formation of catechols and reactive catechol metabolites. Biochem. Biophys. Res. Commun. 78: 919-926. Sato, P.H. and Hall, E.D. (1992) Tirilazad mesylate protects vitamins C and E in brain ischemia-reperfusion injury. J. Neurochem. 58: 2263-2268. Saura, J., Kettler, R., Da Prada, M. and Richards, J.G. (1992) Quantitative enzyme radioautography with -^H-Ro 41-1049 and ^H-Ro 19-6327 in vitro: localization and abundance of MAO-A and MAO-B in rat CNS, peripheral organs and human brain. J. Neurosci. 12: 1977-1999. Sawada, Y., lyanagi, T. and Yamazaki, I. (1975) Relation between redox potential and rate constants in reactions coupled with the system oxygen-superoxide. Biochemistry 14: 3761-3764. Sawyer, D.T. and Nanni, Jr., E.J. (1981) Redoxchemistry of dioxygen species and their chemical reactivity. In: M.A.J. Rodgers and E.L. Powers (Eds.), Oxygen and Oxy-Radicals in Chemistry and Biology, Academic Press, New York, pp. 15-44. Sayre, L.M., Wang, P., Arora, P.K., Riachi, N.J., Harik, S.I. and Hoppel, G.L. (1991) Dopaminergic neurotoxicity in vivo and inhibition of mitochondrial respiration in vitro by possible endogenous pyridinium-like substances. J. Neurochem. 51:2106-2115. Scarpa, M., Rigo, A., Mairino, M., Fulvio, F. and Gregolin, C. (1984) Formation of alpha-tocopherol radical and recycling of alpha-tocopherol by ascorbate during peroxidation of phosphatidylcholine liposome. An electron paramagnetic resonance study. Biochim. Biophys. Acta 801: 215-219. Schapira, A.H.V. and Cooper, J.M. (1992) Mitochondrial function in neurodegeneration and ageing. Mutat. Res. 275: 133-143. Schapira, A.H.V., Cooper, J.M., Dexter, D., Jenner, P., Clark,

503 J.B. and Marsden, C D . (1989) Mitochondrial complex I deficiency in Parkinson's disease. Lancet i: 1269. Schapira, A.H.V., Cooper, J.M., Dexter, D., Clark, J.B., Jenner, P. and Marsden, C D . (1990a) Mitochondrial complex I deficiency in Parkinson's disease. J. Neurochem. 54: 8 2 3 827. Schapira, A.H.V., Holt, I.J., Sweeney, M., Harding, A.E., Jenner, P. and Marsden, C D . (1990b) Mitochondrial DNA analysis in Parkinson's disease. Mov. Disord. 5: 294297. Schapira, A.H.V., Mann, V.M., Cooper, J.M., Dexter, D., Daniel, S.E., Jenner, P., Clark, J.B. and Marsden, C D . (1990c) Anatomic and disease specificity of NADH CoQi reductase (complex I) deficiency in Parkinson's disease. J. Neurochem. 55: 2142-2145. Schechter, R., Yen, S.C. and Terry, R.D. (1981) Fibrous astrocytes in senile dementia of the Alzheimer's type. J. Neuropathol. Exp. Neurol. 40: 95-101. Scheinberg, I.H. and Sternlieb, I. (1984) Wilson's Disease, W.B. Saunders, Philadelphia. Scherman, D., Desnos, C , Darchen, F., Pollak, P., JavoyAgid, F. and Agid, Y. (1989) Striatal dopamine deficiency in Parkinson's disease: role of aging. Ann. Neurol. 26: 5 5 1 557. Schlote, W. and Boellaard, J.W. (1983) Role of lipopigment during aging of nerve and glial cells in the human central nervous system. In: J. Cervos-Navarro and H.J. Sarkander (Eds.), Brain Aging: Neuropathology and Neuropharmacology, Raven Press, New York, pp. 21-14. Schmidt, K.H., Steinhilber, D., Moser, U. and Roth, H.J. (1988) L-Ascorbic acid modulates 5-lipoxygenase activity in human polymorphonuclear leukocytes. Int. Arch. Allergy Appl. Immunol. 85: 4 4 1 ^ 4 5 . Schneider, G., Oepen, H. and Von Wedel, H.R. (1981) MAOAktivitat in verschiedenen Hirngebieten und Korperorganen von Patienten mit Mb Huntington und Mb Parkinson. Arch. Psychiatr. Nervenkr. 230: 5-15. Schneider, J.S., Yuwiler, A. and Markham, C H . (1986) Production of a Parkinson-like syndrome in the cat with Nmethyl-4-phenyl-l,2,3,6-tetrahydropyridine (MPTP): behaviour, histology and biochemistry. Exp. Neurol. 91: 2 9 3 307. Schneider, J.S., Pope, A., Simpson, K., Taggart, J., Smith, M.G. and Di Stefano, L. (1992) Recovery from experimental Parkinsonism in primates with GMl ganglioside treatment. Science 256: 843-846. Schubert, J. and Wilmer, J.W. (1991) Does hydrogen peroxide exist 'free' in biological systems? Free Radical Biol. Med. 11:545-555. Schwarcz, R., Okuno, E., Speciale, C , Kohler, C and Whetsell, W.O. (1987J Neuronal degeneration in animals and man: the quinolinic acid connection. In: P. Jenner (Ed.), Neurotoxins and Their Pharmacological Implications, Raven Press, New York, pp. 20-32. Scott, M.D., Zuo, L., Lubin, B.H. and Chiu, D.T.-Y. (1991) NADPH, not glutathione, status modulates oxidant sensi-

504 tivity in normal and glucose-6-phosphate dehydrogenase deficient erythrocytes. Blood 11: 2059-2064. Seeman, P. (1972) The membrane actions of anesthetics and tranquilizers. Pharmacol. Rev. 24: 583-655. Selkoe, D.J. (1991) The molecular pathology of Alzheimer's disease. Neuron 6: 487^98. Sengstock, G.J., Olanow, C.W., Dunn, A.J. and Arendash, G.W. (1992) Iron induces degeneration of nigrostriatal neurons. Brain Res. Bull. 28: 645-649. Sershenn, H. Reithn, M.E.A., Hashim, A. and Lajtha, A. (1985) Protection against l-methyl-4-pheny 1-1,2,3,6tetrahydropyridine neurotoxicity by the antioxidant ascorbic acid. Neuropharmacology 24: 1257-1259. Sershen, H., Harsing, L.G. Banay-Schwartz, M., Hashim, A., Ramacci, M.T. and Lajtha, A. (1991) Effect of acetyl-1carnitine on the dopaminergic system in aging brain. J. Neurosci. Res. 30: 555-559. Seth, P.K., Dhawan, A., Parmar, D. and Das, M. (1990) Cytochrome P450 catalyzed reactions in brain. In: C.C. Reddy, G.A. Hamilton and K.M. Madyastha (Eds.), Biological Oxidation Systems, Vol. 1, Academic Press, London, pp. 133146. Sevanian, A. and Kim, E. (1985) Phospholipase A2 dependent release of fatty acids from peroxidized membranes. Free Radical Biol. Med. 1: 263-271. Sevanian, A., Stein, R.A. and Mead, J.F. (1981) Metabolism of epoxidized phosphatidylcholine by phospholipase A2 and epoxide hydrolase. Lipids 16: 781-789. Shea, T.B., Beermann, M.L. and Nixon, R.A. (1992) Aluminum alters the electrophoretic properties of neurofilament proteins: role of phosphorylation state. J. Neurochem. 58: 542-547. Shigematsu, K. and McGeer, P.L. (1992a) Accumulation of amyloid precursor protein in damaged neuronal processes and microglia following intracerebral administration of aluminum salts. Brain Res. 593: 117-123. Shigematsu, K. and McGeer, P.L. (1992b) Accumulation of amyloid precursor protein in neurons after intraventricular injection of colchicine. Am. J. Pathol. 140: 787-794. Shigenaga, M.K. and Ames, B.N. (1991) Assays for 8hydroxy-2'-deoxyguanosine: a biomarker of in vivo oxidative damage. Free Radical Biol. Med. 10: 211-216. Shih, J.C, Grimsby, J. and Chen, K. (1990) The expression of human MAO-A and B genes. J. Neural Transm. 32 (Suppl.): 41-47. Shinar, E., Navok, T. and Chevion, M. (1983) The analogous mechanisms of enzymatic inactivation induced by ascorbate and superoxide in the presence of copper. J. Biol. Chem. 258: 14778-14783. Shoffner, J.M., Watts, R.L., Juncos, J.L., Torroni, A. and Wallace, D.C. (1991) Mitochondrial oxidative phosphorylation defects in Parkinson's disease. Ann. Neurol. 30: 332339. Shoulson, I. (1992) An interim report of the effect of selegiline (L-deprenyl) on the progression of disability in early Parkinson's disease. Eur. Neurol. 32 (Suppl. 1): 46-53.

Oxidative stress: free radical production in neural degeneration Shug, A.L., Schmidt, M.J., Golden, G.T. and Fariello, R.G. (1982) The distribution and role of carnitine in the mammalian brain. Life Sci. 31: 2869-2874. Shviro, Y. and Shakai, N. (1987) Glutathione as a scavenger of free hemin. A mechanism of preventing red cell membrane damage. Biochem. Pharmacol. 36: 3801-3807. Siddique, T., Figlewicz, D.A., Pericak-Vance, M.A., Haines, J.L., Rouleau, G., Jeffers, A.J., Sapp, P., Hung, W., Bebout, J., McKenna-Yasek, D., Deng, G., Horvitz, H.R., Gusella, J.E., Brown, R.H., Roses, A.D. et al. (1991) Linkage of a gene causing familial amyotrophic lateral sclerosis to chromosome 21 and evidence of genetic-locus heterogeneity. N. Engl. J. Med. 324: 1381-1384. Sies, H. (1986) Biochemie des oxidativen Stress. Angew. Chem. 98: 1061-1075. Siesjo, B.K. and Bengtsson, F. (1989) Calcium fluxes, calcium antagonists and calcium-related pathology in brain ischemia, hypoglycemia and spreading depression: a unifying hypothesis. /. Cerebral Blood Flow Metab. 9: 127140. Siesjo, B.K., Memezawa, H. and Smith, M.L. (1991) Neurotoxicity: pharmacological implications. Fund. Clin. Pharmacol. 5: 755-767. Simic, M. (1991) DNA damage, environmental toxicants, and rate of aging. Environ. Carcinog. Exotoxicol. Rev . C: 113153. Simic, M.G. and Bergtold, D.S. (1991) Dietary modulation of DNA damage in humans. Mutat. Res. 250: 17-24. Simic, M.G., Bergtold, D.S. and Karam, L.R. (1989) Generation of oxy radicals in biosystems. Mutat. Res. 214: 3-12. Simon, R.P., Swan, J.H., Griffiths, T. and Meldrum, B.S. (1984) Blockade of A^-methyl-D-aspartate receptors may protect against ischemic damage in the brain. Science 226: 850-852. Sims, N.R., Finegan, J.M., Blass, J.P., Bowen, D.M. and Neary, D. (1987) Mitochondrial function in brain tissue in primary degenerative dementia. Brain Res. 436: 30-38. Sinha, B.K. and Mimnaugh, E.G. (1990) Free radicals and anticancer drug resistance: oxygen free radicals in the mechanisms of drug cytotoxicity and resistance by certain tumors. Free Radical Biol. Med. 8: 567-581. Sisodia, S.S., Koo, E.H., Beyreuther, K., Unterbeck, A. and Price, D.L. (1990) Evidence that beta-amyloid in Alzheimer's disease is not derived by normal processing. Science 248: 492-495. Sligar, S.G., Lipscomb, J.D., Debrunner, P.G. and Gunsalus, I.e. (1974) Superoxide anion production by the autoxidation of cytochrome P-450cam- Biochem. Biophys. Res. Commun. 61:290-296. Slivka, A., Mytilineou, C. and Cohen, G. (1987a) Histochemical evaluation of glutathione in brain. Brain Res. 409: 275284. Slivka, A., Spina, M.B. and Cohen, G. (1987b) Reduced and oxidized glutathione in human and monkey brain. Neurosci. Lett. 14: 112-118. Slover, H.T. and Thompson, R.H (1981) Chromatographic

M.E. Gotz et al. separation of the stereoisomers of alpha-tocopherol. Lipids 16: 268-275. Smith, C D . , Carney, J.M., Starke-Reed, P.E., Oliver, C.N., Stadtman, E.R., Floyd, R.A. and Markesbery, W.R. (1991) Excess brain protein oxidation and enzyme dysfunction in normal aging in Alzheimer disease. Proc. Natl. Acad. Sci. USA 88: 10540-10543. Smith, M.T., Ekstrom, G., Sandy, M.S. and DiMonte, D. (1987) Studies on the mechanism of l-methyl-4-phenyl1,2,3,6-tetrahydropyridine cytotoxicity in isolated hepatocytes. Life Sci. 40: 741-748. Soderberg, M., Edlund, C , Kristensson, K. and Dallner, G. (1991) Fatty acid composition of brain phospholipids in aging and in Alzheimer's disease. Lipids 26: 4 2 1 ^ 2 5 . Soderberg, M., Edlund, C , Alafuzoff, I., Kristensson, K. and Dallner, G. (1992) Lipid composition in different regions of the brain in Alzheimer's disease/senile dementia of Alzheimer's type. /. Neurochem. 59: 1646-1653. Sofic, E., Riederer, P., Heinsen, H., Beckmann, H., Reynolds, G.P., Hebenstreit, G. and Youdim, M.B.H. (1988) Increased iron(III) and total iron content in post mortem substantia nigra of parkinsonian brain. /. Neural Transm. 74: 199205. Sofic, E., Paulus, W., Jellinger, K., Riederer, P. and Youdim, M.B.H. (1991) Selective increase of iron in substantia nigra zona compacta of parkinsonian brains. J. Neurochem. 56: 978-982. Sofic, E., Lange, K.W., Jeslinger, K. and Riederer, P. (1992) Reduced and oxidized glutathione in the substantia nigra of patients with Parkinson's disease. Neurosci. Lett. 142: 128130. Sohal, R.S. (1991) Hydrogen peroxide production by mitochondria may be a biomarker of aging. Mech. Ageing Dev. 60: 189-198. Sohal, R.S. and Allen, R.G. (1990) Oxidative stress as a causal factor in differentiation and aging: a unifying hypothesis. Exp. Gerontol. 25: 499-522. Sohal, R.S. and Brunk, U.T. (1992) Mitochondrial production of pro-oxidants and cellular senescence. Mutat. Res. 275: 295-304. Sohal, R.S. and Sohal, B.H. (1991) Hydrogen peroxide release by mitochondria increases during aging. Mech. Ageing Dev. 57: 187-202. Sohal, R.S., Arnold, L.A. and Sohal, B.H. (1990) Age-related changes in antioxidant enzymes and prooxidant generation in tissues of the rat with special reference to parameters in two insect species. Free Radical Biol. Med. 10: 495-500. Sokol, R.J. (1989) Vitamin E and neurologic function in man. Free Radical Biol. Med. 6: 189-207. Sokoloff, L. (1960) The metabolism of the central nervous system in vivo. In: J. Field, H.W. Magoun and V.E. Hall (Eds.), Handbook of Physiology - Neurophysiology Vol. 3, American Physiological Society, Washington, DC, pp. 1843-1864. Sokoloff, L. (1989) Circulation and energy metabolism of the brain. In: G. Siegel, B. Agranoff, R.W. Albers and P. Mo-

505 linoff (Eds.), Basic Neurochemistry, Raven Press, New York, pp. 565-590. Sokoloff, L., Reivich, M., Kennedy, C , Des Rosiers, M.H., Patlak, C.S., Pettigrew, K.D., Sakurada, O. and Shinohara, M. (1977) The [^^^]deoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. /. Neurochem. 28: 897-916. Sorgato, M.C. and Sartorelli, L (1974) Oxygen radicals and hydrogen peroxide in rat brain mitochondria. FEBS Lett. 45: 92-95. Southam, E., Thomas, P.K., King, R.H.M., Goss-Sampson, M.A. and Muller, D.P.R. (1991) Experimental vitamin E deficiency in rats, morphological and functional evidence of abnormal axonal transport secondary to free radical damage. Brain 114: 915-936. Spagnoli, A., Lucca, U., Menasce, G., Bandera, L., Cizza, G., Forloni, G., Tettamanti, M., Frattura, L., Tiraboschi, P., Comelli, M., Senin, U., Longo, A., Petrini, A., Brambilla, G., Belloni, A., Negri, C , Cavazzuti, E., Salsi, A., Calogero, P., Parma, E., Stramba-Badiale, M., Vitali, S., Andreoni, G., Inzoli, M.R., Santus, G., Caregnato, R., Peruzza, M., Favaretto, M., Bozeglav, C , Alberoni, M., De Leo, D., Serraiotto, L., Baiocchi, A., Scoccia, S., Culotta, P. and leracitano, D. (1991) Long-term acetyl-L-carnitine treatment in Alzheimer's disease. Neurology 41: 17261732. Spallholz, J.E. and Boylan, L.M. (1991) Glutathione peroxidase: the two selenium enzymes. In: J. Everse, K.E. Everse and M.B. Grisham (Eds.), Peroxidases in Chemistry and Biology, Vol. I CRC Press, Boca Raton, FL, pp. 2 5 9 291. Sparks, D.L., DeKosky, S.T. and Markesbery, W.R. (1988) Aminergic-cholinergic alterations in hypothalamus. Arch. Neurol. 45: 994-999. Spencer, P.S., Allen, R.G., Kisby, G.E. and Ludolph, A.C. (1990) Excitotoxic disorders. Science 248: 144. Spina, M.B. and Cohen, G. (1988) Exposure of striatal synaptosomes to L-DOPA increases levels of oxidized glutathione. J. Pharmacol. Exp. Ther. 241: 502-507. Spina, M.B. and Cohen, G. (1989) Dopamine turnover and glutathione oxidation: implications for Parkinson's disease. Proc. Natl. Acad. Sci. USA 86: 1398-1400. Spina, M.B., Squinto, S.P., Miller, J., Lindsay, R.M. and Hyman, C. (1992) Brain-derived neurotrophic factor protects dopamine neurons against 6-hydroxydopamine and A^methyl-4-phenylpyridinium ion toxicity: involvement of the glutathione system. J. Neurochem. 59: 99-106. Spitz, D.K., Sullivan, S.J., Malcolm, R.R. and Roberts, R.J. (1991) Glutathione dependent metabolism and detoxification of 4-hydroxy-2-nonenal. Free Radical Biol. Med. 11: 415-423. Srivastava, R.B., Misra, R.B. and Joshi, P.C. (1986) Photosensitized generation of singlet oxygen and superoxide radicals by selected dyestuffs, food additives and their metabolites. P/z6>ro^/ocAem. P/ioro^/op/zj.?. 11:129-137.

506 Stadtman, E.R. (1990) Metal ion-catalyzed oxidation of proteins: biochemical mechanism and biological consequences. Free Radical Biol. Med. 9: 315-325. Stadtman, E.R. (1992) Protein oxidation and aging. Science 257: 1220-1224. Stadtman, E.R. and Berlett, B.S. (1991) Fenton chemistry: amino acid oxidation. / . Biol. Chem. 266: 17201-17211. Stadtman, E.R. and Oliver, C.N. (1991) Metal-catalyzed oxidation of proteins. J. Biol. Chem. 266: 2006-2008. Steinman, H.M. and Hill, R.L. (1973) Sequence homologies among bacterial and mitochondrial superoxide dismutases. Proc. Natl. Acad Sci. USA 70: 3725-3729. Stenstrom, A., Hardy, J. and Oreland, L. (1987) Intra- and extra-dopamine-synaptosomal localization of monoamine oxidase in striatal homogenates from four species. Biochem. Pharmacol. 18: 2931-2935. Stern, A., Perl, D., Munoz-Garcia, D., Good, R., Abraham, C. and Selkoe, D. (1986) Investigation of a silicon and aluminium content in isolated senile plaque cores by laser microprobe mass analysis (LAMMA). J. Neuropathol. Exp. Neurol. 45. 361. Stokes, C.E. and Hawthorne, J.N. (1987) Reduced phosphoinositide concentration in anterior temporal cortex of Alzheimer's disease brains. J. Neurochem. 48: 1018-1021. Stone, K.J. and Townsley, B.H. (1973) The effect of Lascorbate on catecholamine biosynthesis. Biochem. J. 131: 611-613. Strain, S.M. and Tasker, R.A.R. (1991) Hippocampal damage produced by systemic injections of domoic acid in mice. Neuroscience 44: 543-552. Strittmatter, M.M. and Cramer, H. (1992) Parkinson's disease and dementia: clinical and neurochemical correlations. NeuroReport 3: 413-416. Strittmatter, P., Spatz, L., Corcoran, D., Rogers, M.J., Setlow, B. and Redline, R. (1974) Purification and properties of rat liver microsomal stearyl coenzyme A desaturase. Proc. Natl. Acad. Sci. USA 11: 4565-4569. Strolin-Benedetti, M. and Dostert, P. (1989) Monoamine oxidase, brain ageing and degenerative diseases. Biochem. Pharmacol. 38: 555-561. Strolin-Benedetti, M., Cao Danh, H. and Dostert, P. (1986) Age-related changes in brain MAO and in enzymes involved in detoxication processes of MAO-generated compounds. In: G. Biggio, P.P. Spano, G. Toffano and G.L. Gessa (Eds.), Modulation of Central and Peripheral Transmitter Function, Liviana Press, Padova, pp. 255-267. Subbarao, K.V., Richardson, J.S. and Ang, L.C. (1990) Autopsy samples of Alzheimer's cortex show increased peroxidation in vitro. J. Neurochem. 55: 342-345. Sussman, M.S. and Bulkley, G.B. (1990) Oxygen-derived free radicals in reperfusion injury. In: L. Packer and A.N. Glazer (Eds.), Methods in Enzymology: Oxygen Radicals in Biological Systems Part B Oxygen Radicals and Antioxidants, Vol. 186, Academic Press, New York, pp. 711-723. Sutherland, R.J., Hoesing, J.M. and Wishaw, I.Q. (1990) Domoic acid, an environmental toxin, produces hippocampal

Oxidative stress: free radical production in neural degeneration damage and severe memory impairment. Neurosci. Lett. 120: 221-223. Suzuki, K., Katzman, R. and Korey, S.R. (1965) Chemical studies on Alzheimer's disease. J. Neuropathol. Exp. Neurol. 24: 211-224. Suzuki, K., Mizuno, Y. and Yoshida, M. (1988) Inhibition of mitochondrial NADH-ubiquinone oxidoreductase activity and ATP synthesis by tetrahydroisoquinoline. Neurosci. Lett. 86: 105-108. Suzuki, K., Mizuno, Y., Yamauchi, Y., Nagatsu, T. and Yoshida, M. (1992) Selective inhibition of complex I by A^methylisoquinolinium ion and A^-methyl-l,2,3,4-tetrahydroisoquinoline in isolated mitochondria prepared from mouse brain. J. Neurol. Sci. 109: 219-223. Taglialatela, G., Angelucci, L., Ramacci, M.T., WerrbachPerez, K., Jackson, G.R. and Perez-Polo, J.R. (1991) Acetyl-L-carnitine enhances the response of PC 12 cells to nerve growth factor. Dev. Brain Res. 59: 221-230. Taglialatela, G., Angelucci, L., Ramacci, M.T., WerrbachPerez, K., Jackson, G.R. and Perez-Polo, J.R. (1992) Stimulation of nerve growth factor receptors in PC 12 by acetyl-L-carnitine. Biochem. Pharmacol. 44: 577-585. Tanaka, M., Sotomatsu, A., Kanai, H. and Hirai, S. (1991) DOPA and dopamine cause cultured neuronal death in the presence of iron. J. Neurol. Sci. 101: 198-203. Tandan, R. and Bradley, W.G. (1985a) Amyotrophic lateral sclerosis: Part 1. Clinical features, pathology, and ethical issues in management. Ann. Neurol. 18: 271-280. Tandan, R. and Bradley, W.G. (1985b) Amyotrophic lateral sclerosis: Part 2. Etiopathogenesis. Ann. Neurol. 18: 4 1 9 431. Tanner, C M . (1989) The role of environmental toxins in the etiology of Parkinson's disease. Trends Neurol. Sci. 12: 4 9 54. Tappel, A.L. (1973) Lipid peroxidation damage to cell components. Fed. Proc. 32: 1874-1879. Tatton, W.G. and Greenwood, C.E. (1991) Rescue of dying neurons: a new action for deprenyl in MPTP Parkinsonism. /. Neurosci. Res. 30: 666-672. Tatton, W.G., Greenwood, C.E., Salo, P.T. and Seniuk, N.A. (1991) Transmitter synthesis increases in substantia nigra neurons of the aged mouse. Neurosci. Lett. 131: 179-182. Teebor, G.W., Boorstein, R.J. and Cadet, J. (1988) The repairability of oxidative free radical mediated damage to DNA: a review. Int. J. Radiat. Biol. 54: 131-150. Teitelbaum, J.S., Zatorre, R.J., Carpenter, S., Gendron, D., Evans, A.C., Gjedde, A. and Cashman, N.R. (1990) Neurologic sequelae of domoic acid intoxication due to the ingestion of contaminated mussels. N. Engl. J. Med. 332: 1781-1787. The Parkinson Study Group (1989a) DATATOP: a multicenter controlled clinical trial in early Parkinson's disease. Arch. Neurol. 46: 1052-1060. The Parkinson Study Group (1989b) Effect of deprenyl on the progression of disability in early Parkinson's disease. N. Eng. J. Med. 321: 1364-1371.

M.E. Gotz. et al. The Parkinson Study Group (1993) Effects of tocopherol and deprenyl on the progression of disability in early Parkinson's disease. N. Engl J. Med. 328: 176-183. The Scottish Motor Neuron Research Group (1992) The Scottish motor neuron disease register: a prospective study of adult onset motor neuron disease in Scotland. Methodology, demography and clinical features of incident cases in 1989. J. Neurol. Neurosurg. Psychiatry 55: 536-541. Thoenen, H. and Tranzer, J.P. (1973) The pharmacology of 6hydroxydopamine. AnnM. Rev. Pharmacol. 13: 169-180. Thomas, J.P., Maiorino, M., Ursini, F. and Girotti, A.W. (1990) Protective action of phospholipid hydroperoxide glutathione peroxidase against membrane-damaging lipid peroxidation. In situ reduction of phospholipid and cholesterol hydroperoxides. J. Biol. Chem. 265: 454-461. Thomas, P.D. and Brewer, G.J. (1990) Gangliosides and synaptic transmission. Biochim. Biophys. Acta 1031: 277-289. Thomas, T.N., Priest, D.G. and Zemp, J.W. (1976) Distribution of superoxide dismutase in rat brain. J. Neurochem. 27: 309-310. Thompson, T.L. (1987) Dementia. In: R.E. Hales and S.C. Yudofsky (Eds.), Textbook of Neuropsychiatry, The American Psychiatry Press, New York, pp. 107-124. Thorpe, L.W., Westlund, K.N., Kochersperger, L.M., Abell, C.W. and Denney, R.M. (1987) Immunocytochemical localization of MAO-A and -B in human peripheral tissues and brain. J. Histochem. Cytochem. 35: 23-32. Thorstensen, K. and Romslo, I. (1990) The role of transferrin in the mechanism of cellular iron uptake. Biochem. J. 211: 1-10. Tipton, K.F. (1967) The sub-mitochondrial localization of monoamine oxidase in rat liver and brain. Biochim. Biophys. Acta 5: 910-920. Torack, R.M. and Morris. J.C. (1990) Tyrosine hydroxylase (TH) immunoreactivity in human mesolimbic system. Neurosci. Lett. 116:75-80. Torack, R.M. and Morris, J.C. (1992) Tyrosine hydroxylaselike (TH) immunoreactivity in Parkinson's disease and Alzheimer's disease. J. Neural Transm. 4: 165-171. Trancoso, J . C , Price, D.L., Griffin, J.W. and Parhad, I.M. (1982) Neurofibrillary axonal pathology in aluminium toxicity. A^A?.. Neurol. 12: 278-283. Trounce, I., Byrne, E. and Marzuki, S. (1989) Decline in skeletal muscle mitochondrial respiratory chain function: possible factor in aging. Lancet i: 637-639. Tucek, S., Ricny, J. and Dolezal, V. (1990) Advances in the biology of cholinergic neurons. In: R.J. Wurtman, S. Corkin, J.H. Growdon and E. Ritter-Walker (Eds.), Advances in Neurology, Alzheimer's Disease, Vol. 51, Raven Press, New York, pp. 109-115. Turrens, J.F. and Boveris, A. (1980) Generation of superoxide anion by the NADH-dehydrogenase of bovine heart mitochondria. Biochem. J. 191: 4 2 1 ^ 2 7 . Tyler, D.D. (1975) Role of superoxide radicals in the lipid peroxidation of intracellular membranes. FEBS Lett. 51: 180-183.

507 Uhlig, S. and Wendel, A. (1992) The physiological consequences of glutathione variations. Life Sci. 51: 10831094. Uitti, R.J., Rajput, A.H., Rozdilsky, B., Bickis, M., Wollin, T. and Yuen, W.K. (1989) Regional metal concentrations in Parkinson's disease, other chronic neurological diseases, and control brains. Can. J. Neurol. Sci. 16: 310-314. Ursini, F., Maiorino, M., Valente, M., Ferri, L. and Gregolin, C. (1982) Purification from pig liver of a protein which protects liposomes and biomembranes from peroxidative degradation and exhibits glutathione peroxidase activity on phosphatidylcholine hydroperoxides. Biochim. Biophys. ActallO: 197-211. Valberg, L.S., Flanagan, P.R., Kertesz, A. and Ebers, G.C. (1989) Abnormalities in iron metabolism in multiple sclerosis. Can. J. Neurol. Sci. 16: 184-186. Valenzuela, A. (1991) The biological significance of malondialdehyde determination in the assessment of tissue oxidative stress. Life Sci. 48: 301-309. Van Der Voet, G.B., Marani, E., Tio, S. and De Wolff, F.A. (1991) Aluminium neurotoxicity. In: W. Graumann and J. Drukker (Eds.), Progress in Histo- and Cytochemistry, Vol. 23, Gustav Fischer, Stuttgart, pp. 235-242. Van Kuijk, F.J.G.M., Thomas, D.W., Stephens, R.J. and Dratz, E.A. (1986) Occurrence of 4-hydroxy-alkenals in rat tissues determined as pentafluorbenzyloxime derivatives by gas chromatography-mass spectrometry. Biochem. Biophys. Res. Commun. 139: 144-149. Van Kuijk, F.J.G.M., Sevanian, A., Handelman, G.J. and Dratz, E.A. (1987) A new role for phospholipase A2; protection of membranes from lipid peroxidation damage. Trends Biochem. Sci. 12: 31-34. Van Steveninck, J., Van Der Zee, J. and Dubbelman, T.M.A.R. (1985) Site of specific and bulk-phase generation of hydroxyl radicals in the presence of cupric ions and thiol compounds. Biochem. J. 232: 309-311. Van Woert, M.H., Prasad, K.N. and Borg, D.C. (1967) Spectroscopic studies of substantia nigra pigment in human subjects. J. Neurochem. 14: 707-716. Vatassery, G.T. (1987) Selected aspects of the neurochemistry of vitamin E. In: O. Hayaishi and M. Mino (Eds.), Clinical and Nutritional Aspects of Vitamin E, Elsevier, Amsterdam, pp. 147-155. Vatassery, G.T., Angerhofer, C.K. and Knox, C.A. (1984a) Effect of age on vitamin E concentrations in various regions of the brain and a few selected peripheral tissues of the rat, and on the uptake of radioactive vitamin E by various regions of the rat brain. J. Neurochem. 43: 4 0 9 ^ 1 2 . Vatassery, G.T., Angerhofer, C.K., Knox, C.A. and Deshmukh, D.S. (1984b) Concentrations of vitamin E in various neuroanatomical regions and subcellular fractions, and the uptake of vitamin E by specific areas, of rat brain. Biochim. Biophys. Acta 792: 118-122. Vecchi, M., Walther, W., Glinz, E., Netscher, T., Schmid, R., Lalonde, M. and Vetter, W. (1990) Chromatographische Trennung und quantitative Bestimmung aller acht Stereoi-

508 someren von alpha-Tocopherol. Helv. Chim. Acta 73: 7 8 2 789. Villa, R.F. and Gorini, A. (1991) Action of L-acetylcarnitine on different cerebral mitochondrial populations from hippocampus and striatum during aging. Neurochem. Res. 16: 1125-1132. Von Sonntag, C. (1987) The Chemical Basis of Radiation Biology. Taylor & Francis, London. Wagner, G.C., Jarvis, M.F. and Carelli, R.M. (1985) Ascorbic acid reduces the dopamine depletion induced by MPTP. Neuropharmacology 2A\ 1261-1262. Wagner, K. and Trendelenburg, U. (1971) Effect of 6hydroxydopamine on oxidative phosphorylation and on monoamine oxidase activity. Naunyn-Schmiedeberg's Arch. Pharmacol. 269: 110-116. Wakamatsu, K., Ito, S. and Nagatsu, T (1991) Cysteinyidopamine is not incorporated into neuromelanin. Neurosci. Lett. 131:57-60. Wallace, D.C. (1989) Mitochondrial DNA mutations and neuromuscular diseases. Trends Genet. 5: 9-14. Wallace, D.C. (1992) Mitochondrial genetics: a paradigm for aging and degenerative diseases? Science 256: 628-632. Warner, M., Kohler, C , Hansson, T. and Gustafsson, J.A. (1988) Regional distribution of cytochrome P-450 in the rat brain: spectral quantitation and contribution of P-450b,e and P-450c,d. / . Neurochem. 50: 1057-1065. Watkins, J . C , Krogsgaard-Larsen, P. and Honore, T. (1991) Structure-activity relationships in the development of excitatory amino acid receptor agonists and competitive antagonists. In: D. Lodge and G.L. Collingridge (Eds.), Trends in Pharmacological Sciences, The Pharmacology of Excitatory Amino Acids. Special Report, Elsevier, Amsterdam, pp. 4-12. Wefers, H. and Sies, H. (1988) The protection by ascorbate and glutathione against microsomal lipid peroxidation is dependent on vitamin E. Eur. J. Biochem. 174: 353-357. Werner, P. and Cohen, G. (1991) Intramitochondrial formation of oxidized glutathione during oxidation of benzylamine by monoamine oxidase. FEBS Lett. 280: 44-46. Westlund, K.N., Denney, R.M., Kochsperger, L.M., Rose, R.M. and Abell, C.W. (1985) Distinct MAO-A and MAO-B populations in the primate brain. Science 230: 181-183. Westlund, K.N., Denney, R.M., Rose, R.M. and Abell, C.W. (1988) Localization of distinct monoamine oxidase A and monoamine oxidase B cell populations in human brainstem. Neuroscience 25: 439-456. Whalley, L.J., McGonigal, G. and Thomas, B. (1992) Aluminum and dementia. Lancet i: 1235-1236. Whitehouse, P.J., Heldreen, J . C , White, C.L. and Price, D.L. (1983a) Basal forebrain neurons in the dementia of Parkinson's disease. Ann. Neurol. 13: 243-248. Whitehouse, P.J., Wamsley, J.K. and Zarbin, M.A. (1983b) Amyotrophic lateral sclerosis: alterations in neurotransmitter receptors. Ann. Neurol. 14: 8-16. Whitehouse, P.J., Vale, W.W., Zweig, R.M., Singer, H.S., Mayeux, R., Kuhar, M.J., Price, D.L. and De Souza, E.B.

Oxidative stress: free radical production in neural degeneration (1987) Reductions in corticotropin releasing factor-like immunoreactivity in cerebral cortex in Alzheimer's disease, Parkinson's disease, and progressive supranuclear palsy. Neurology 37: 905-909. Wills, E.D. (1966) Mechanisms of lipid peroxide formation in animal tissue. Biochem. J. 99: 667-676. Winkler, B.S. (1992) Unequivocal evidence in support of the nonenzymatic redox coupling between glutathione/glutathione disulfide and ascorbic acid/dehydroascorbic acid. Biochim. Biophys. Acta 1117: 287-290. Winterbourn, C.C, Vile, G.F. and Monteiro, H.P. (1991) Ferritin, lipid peroxidation and redox cycling xenobiotics. Free Radical Res. Commun. 12/13: 107-114. Wisniewski, H., Sturman, J.A. and Sherk, J.W. (1980) Aluminium chloride-induced neurofibrillary changes in the developing rabbit: a chronic animal model. Ann. Neurol. 8: 479-490. Wisniewski, H.M., Merz, G.S., Rabe, A., Barcikowska, M., Moretz, R.C. and Devine-Gage, E.A. (1988) Current hypotheses of Alzheimer disease neuropathology and dementia. Drug Dev. Res. 15: 115-121. Wisp, J.R., Clark, J.C, Burhans, M.S., Kropp, K.E., Korfhagen, T.R. and Whitsett, J.A. (1989) Synthesis and processing of the precursor for human mangano-superoxide dismutase. Biochim. Biophys. Acta 994: 30-36. Wittes, R.E. (1985) Vitamin C and cancer. N Engl. J. Med. 312:178-179. Witz, G. (1989) Biological interactions of alpha-y3-unsaturated aldehydes. Free Radical Biol. Med. 7: 333-349. Wolff, S.P., Jiang, Z.Y. and Hunt, J.V. (1991) Protein glycation and oxidative stress in diabetes mellitus and aging. Free Radical Biol. Med. 10: 339-352. Wolters, E.C, Hurwitz, T.A., Peppard, R.F. and Calne, D.B. (1989) Clozapine: an antipsychotic agent in Parkinson's disease. Clin. Neuropharmacol. 12: 83-90. Wood, P.M. (1988) The potential diagram for oxygen at pH 7. Biochem. J. 253: 287-289. Wriggles worth, J.M. and Baum, H. (1988) Iron-dependent enzymes in the brain. In: M.B.H. Youdim (Ed.), Brain Iron, Neurochemical and Behavioural Aspects, Taylor & Francis, London, pp. 25-66. Wurtman, R.J. (1992) Choline metabolism as a basis for the selective vulnerability of cholinergic neurons. Trends Neurosci. 15:117-122. Wurtman, R.J., Blusztajn, J.K., Ulus, I.H., Coviella, I.L.G., Buyukuysal, R.L., Growdon, J.H. and Slack, B.E. (1990) Choline metabolism in cholinergic neurons: implications for the pathogenesis of neurodegenerative disorders. In: R.J. Wurtman, S. Corkin, J.H. Growdon and E. RitterWalker (Eds.), Advances in Neurology, Alzheimer's Disease, Vol. 51, Raven Press, New York, pp. 117-125. Yen, J.T., Ku, P.K., Pond, W.G. and Miller, E.R. (1985) Response to dietary supplementation of vitamin C and vitamin E in weanling pigs fed low vitamin E-selenium diets. Nutr. Rep. Int. 31: S11-SS5. Yen, T.C, Chen, Y.S., King, K.L., Yeh, S.H and Wei, Y.H.

M.E. Gotz et al. (1989) Liver mitochondrial respiratory functions decline with age. Biochem. Biophys. Res. Commun. 165: 994-1003. Yen, T.C., Su, J.H., King, K.L. and Wei, Y.H. (1991) Age associated 5 kb deletion in human liver. Biochem. Biophys. Res. Commun. 178: 124-131. Yin, D. (1992) Lipofuscin-like fluorophores can result from reactions between oxidized ascorbic acid and glutamine. Carbonyl-protein cross-linking may represent a common reaction in oxygen radical and glycosylation-related ageing processes. Mech. Ageing Dev. 62: 35-46. Yin, D. and Brunk, U.T. (1991) Oxidized ascorbic acid and reaction products between ascorbic and amino acids might constitute part of age pigments. Mech. Ageing Dev. 61: 9 9 112. Yoshida, M., Niwa, T. and Nagatsu, T. (1990) Parkinsonism in monkeys produced by chronic administration of an endogenous substance of the brain, tetrahydroisoquinoline: the behavioral and biochemical changes. Neurosci. Lett. 119: 109-113. Yoshino, H., Nakagawa-Hattori, Y., Kondo, T. and Mizuno, Y. (1992) Mitochondrial complex I and 11 activities of lymphocytes and platelets in Parkinson's disease. J. Neural Transm. (P-D Sect.) 4: 27-34. Yoshino, K., Sano, M., Fujita, M. and Tomita, I. (1991) Production of aliphatic aldehydes on peroxidation of various types of lipids. Chem. Pharmacol. Bull. 39: 1788-1791. Yost, F.J. and Fridovich, I. (1973) An iron-containing superoxide dismutase from Escherichia coli B. J. Biol. Chem. 248: 4 9 0 5 ^ 9 0 8 . Youdim, M.B.H. (1985) Brain iron metabolism: biochemical and behavioural aspects in relation to dopaminergic neurotransmission. In: A. Lajtha (Ed.), Handbook of Neurochemistry, Vol. 10, Plenum Press, New York, pp. 731-765. Youdim, M.B.H. and Finberg, J.P.M. (1990) New directions in monoamine oxidase A and B: selective inhibitors and substrates. Biochem. Pharmacol. 41: 155-162. Youdim, M.B.H., Finberg, J.P.M. and Tipton, K.F. (1988) Monoamine oxidase. In: U. Trendelenburg and N. Weiner (Eds.), Handbook of Experimental Pharmacology, Vol. 90/1, Springer, Berlin, pp. 119-192. Youdim, M.B.H., Ben-Shachar, D. and Riederer, P. (1990) The, role of monoamine oxidase, iron-melanin interaction and intracellular calcium in Parkinson's disease. J. Neural Transm. 32 (Suppl.): 239-248.

509 Youdim, M.B.H., Ben-Shachar, D. and Riederer, P. (1991) Iron in brain function and dysfunction with emphasis on Parkinson's disease. Eur. Neurol. 31 (Suppl. 1): 34—40. Young, A.B. and Egg, G.E. (1991) Excitatory amino acid receptors in the brain: membrane binding and receptor autoradiographic approaches. In: D. Lodge and G.L. Collingridge (Eds.), Trends in Pharmacological Sciences. The Pharmacology of Excitatory Amino Acids, Special Report, Elsevier, Amsterdam, pp. 18-24. Young, V.W., Perry, T.L. and Krisman, A.A. (1986) Depletion of glutathione in brainstem of mice caused by A^methyl-4-phenyl-l,2,3,6-tetrahydropyridine is prevented by antioxidant treatment. Neurosci. Lett. 63: 56-60. Youngman, L.D., Park, J.-Y.K. and Ames, B.N. (1992) Protein oxidation associated with aging is reduced by dietary restriction of protein or calories. Proc. Natl. Acad. Sci. USA 89:9112-9116. Zaleska, M.M. and Wilson, D.F. (1989) Lipid hydroperoxides inhibit reacylation of phospholipids in neuronal membranes. J. Neurochem. 52: 255-260. Zang, Z.X., Anderson, D.W. and Mantel, N. (1990) Geographic, patterns of parkinsonism-dementia complex on Guam. Arc/z. Neurol. 47: 1069-1074. Zecca, L., Mecacci, C , Seraglia, R. and Parati, E. (1992) The chemical characterization of melanin contained in substantia nigra of human brain. Biochim. Biophys. Acta 1138: 6 10. Zeevalk, G.D. and Nicklas, W.J. (1991) Mechanisms underlying initiation of excitotoxicity associated with metabolic inhibition. J. Pharmacol. Exp. Ther. 257: 870-878. Zemlan, F.F., Thienhaus, O.J. and Bosmann, H.B. (1989) Superoxide dismutase in Alzheimer's disease. Possible mechanism for paired helical filament formation. Brain Res. 476: 160-162. Zhang, J. and Piantadosi, C.A. (1991) Prevention of H2O2 generation by monoamine oxidase protects CNS O2 toxicity. J. Appl. Physiol. 71: 1057-1061. Zhu, Q.-S., Grimsby, J., Chen, K. and Shih, J.C. (1992) Promoter organization and activity of human monoamine oxidase (MAO) A and B genes. J. Neurosci. 12: 4437—4446. Zhukov, A.A. and Archakov, A.I. (1982) Complete stoichiometry of the free NADPH oxidation in liver microsomes. Biochem. Biophys. Res. Commun. 109: 813-818.

This Page Intentionally Left Blank

C. Bell (Editor) Chemical Factors in Neural Growth, Degeneration and Repair © 1996 Elsevier Science B.V. All rights reserved.

CHAPTER 18

Excitotoxic neuronal damage and neuropsychiatric disorders Charles F. Zorumski and John W . Olney Department of Psychiatry, Washington University School of Medicine, St. Louis, MO 63110, USA

1. Introduction Glutamate (Glu) and related excitatory amino acids (EAA) are thought to serve as the primary fast excitatory neurotransmitters in the vertebrate central nervous system (CNS) and to participate in several forms of synaptic plasticity, including long-term potentiation (LTP) and long-term depression (LTD) of excitatory synaptic responses (Kuba and Kumamoto, 1990). LTP and LTD are candidate mechanisms thought to underlie certain types of learning and memory. During development, EAA also help to shape the morphology of neurons by altering the branching and outgrowth of neurites (MacDonald and Johnston, 1990). Paradoxically, EAA also have neurotoxic properties which produce several forms of neurodegeneration. EAA neurotoxicity is often referred to as 'excitotoxicity', a term that connotes the role of an excitatory mechanism and of excitatory receptors in the toxic process. Excitotoxins have been implicated in a wide variety of neuropsychiatric disorders based on evidence which in some cases remains tentative but in others is quite strong. In this chapter we review information about the classes of EAA receptors in the CNS and the evidence linking these receptors to human neuropsychiatric disorders. 2. Excitatory amino acid receptors EAA interact with at least five classes of receptors in the vertebrate CNS (Zorumski and Thio, 1992). These receptors are named according to exogenous agonists that are relatively selective

for the various classes. These include A^-methyl-Daspartate (NMDA), a-amino-3-hydroxy-5-methyl4-isoxazolepropionic acid (AMPA), kainate, trans-(±)-1 -amino-1,3-cyclopentane-dicarboxylic acid (ACPD), and 2-amino-4-phosphonobutyrate (APB, AP4). The first three classes promote the opening of cation selective channels and are called 'ionotropic' receptors. ACPD and APB receptors are coupled by a guanine nucleotide binding protein (G-protein) to second messenger systems and are referred to as 'metabotropic' receptors. The ionotropic receptors mediate fast excitatory synaptic transmission and have been implicated in excitotoxicity syndromes. ACPD and APB receptors have not been directly linked to human neurodegenerative disorders but observations in animals suggest that ACPD receptors also harbor toxic potential. 2,1. NMDA receptors Glu is the most likely native transmitter at EAA receptors and among the proposed transmitter candidates, has the highest affinity for NMDA receptors (Olverman et al., 1984). Aspartate (Asp) is an endogenous transmitter candidate that acts with relative selectivity at NMDA sites. NMDA receptors are inhibited competitively by structural analogs of Glu that include 2-amino-5phos-phonovalerate (APV, APS) and 3-((±)-2carboxypiperazin-4-yl)propyl-1 -phosphonic acid (CPP). NMDA receptors contain several other binding sites at which responses can be regulated. These include at least two distinct sites where positive modulators, including glycine (Johnson

512

and Ascher, 1987) and the polyamines, spermine and spermidine (Williams et al., 1990), act to enhance NMDA responses. Both the glycine and polyamine sites modulate NMDA receptor desensitization (Mayer et al., 1989; Lerma, 1992) and glycine appears to be a required cofactor for NMDA receptor activation (Kleckner and Dingledine, 1988). Agents that interfere with the action of glycine or polyamines are non-competitive NMDA antagonists. Certain dissociative anesthetics, including phencyclidine (PCP), ketamine and dizocilpine (MK-801) also block NMDA responses non-competitively by binding to a distinct site within the NMDA ion channel complex. The dissociative anesthetics act by a long-lived, voltage-dependent channel block mechanism in which the anesthetic molecule becomes trapped in the ion channel (Huettner and Bean, 1988). This block can be most readily relieved by opening the channel at depolarized membrane potentials. The divalent cations, Mg^"^ and Zn^+, regulate NMDA responses non-competitively by blocking NMDA ion channels at two separate loci within the receptorchannel protein (Westbrook and Mayer, 1987). The actions of Mg^"*" are highly voltage dependent and account for both the marked non-linearity of the NMDA current-voltage {IV) relationship (Mayer and Westbrook, 1985) and the observation that NMDA receptors primarily participate in synaptic transmission when neurons are depolarized from their resting membrane potential (Forsythe and Westbrook, 1988). An important feature of NMDA ion channels is their permeability to Ca^"^. Estimates using constant electrical field assumptions suggest that these channels have the highest Ca^"^ permeability of any ligand-gated ion channel in the mammalian CNS (Mayer and Westbrook, 1987). The Ca^^ influx provided by NMDA channels is important in the role that these channels play in both synaptic plasticity and excitotoxicity. Interestingly, Cd?-^ also appears to regulate current flow through NMDA channels by promoting a slow, voltage-dependent form of desensitization during periods of prolonged receptor activation (Mayer and Westbrook, 1985;Zorumskietal., 1989). Five distinct NMDA receptor subunits have

Excitotoxic neuronal damage and neuropsychiatric disorders

been cloned to date (NMDARl, NMDAR2A-D), and at least eight splice variants for NMDARl exist adding to the complexity of this receptor family. NMDARl subunits can form homomeric receptors whereas NMDAR2 subunits function only in heteromeric combinations (Hollmann and Heinemann, 1994). Regional differences in the expression of these subunits are likely to influence the physiological and pharmacological properties of NMDA receptors and perhaps susceptibility of neurons to excitotoxic insults. 2,2, Kainate/AMPA receptors Kainate and AMPA receptors are collectively referred to as 'non-NMDA' receptors. Although in some preparations kainate and AMPA activate the same receptor-channel complex where AMPA, Glu and quisqualate (Quis) gate rapidly desensitizing responses and kainate gates non-desensitizing responses (Patneau and Mayer, 1991; Thio et al., 1991), there is evidence supporting the existence of separate kainate and AMPA receptors in some regions of the CNS (Monaghan et al., 1989). Nine different subunits for AMPA (GluRl^) and kainate (GluR5-7, KAl-2) receptors have been cloned and provide additional evidence for distinct receptor structures (Barnes and Henley, 1992; Hollmann and Heinemann, 1994). AMPA and kainate receptors directly gate cation selective channels that typically have low permeability to Ca^^ and are not blocked by Mg^+, Zn^"^, or dissociative anesthetics. IV relationships for AMPA/kainate currents are nearly linear over the range of physiological membrane potentials. A population of kainate-activated receptors (termed Type II kainate receptors) with greater Ca^"^ permeability and an inwardly rectifying IV relationship have been described in hippocampal neurons (lino et al., 1990) and AMPA-activated receptors with higher Ca^"^ permeability have been described in retinal neurons (Gilbertson et al., 1991). Receptor cloning studies indicate that the low Ca^"*" permeability and linear IV relationship of most AMPA receptors results from the presence of a specific subunit, termed GluR-2 (Hollman et al., 1991). In the absence of GluR-2, currents gated by

513

C.F. Zorumski and J.W. Olney

recombinant AMPA receptors have a higher Ca^"^ permeability and an inwardly rectifying IV relationship. Interestingly, a single amino acid substitution in a region of the second transmembrane segment of the GluR-2 protein accounts for the differences in these properties (Hume et al., 1991; Verdoorn et al, 1991). Most antagonists for AMPA/kainate receptors are non-selective and also block NMDA receptors. The most useful antagonists have been the quinoxalinediones, 6-,7-dinitroquinoxaline-2,3-dione (DNQX) and 6-cyano-7-nitro-quinoxaline-2,3dione (CNQX) (Honore et al., 1988). These agents competitively inhibit non-NMDA receptors but at higher concentrations block NMDA responses through an action at the glycine site. Certain 2,3 benzodiazepines, including GYKI 52466, appear to be more selective, non-competitive antagonists at non-NMDA receptors (Tarnawa et al., 1989; Donevan and Rogawski, 1993; Zorumski et al., 1993) and may be selective for AMPA receptors in hippocampal neurons (Paternain et al., 1995). Compared to NMDA receptors, less is known about the regulation of responses mediated by nonNMDA receptors. Rapid desensitization is one mechanism that modifies the participation of AMPA receptors in fast excitatory synaptic transmission (Tang et al., 1991; VykUcky et al., 1991; Thio et al., 1992) and acute excitotoxicity (Zorumski et al., 1990). This desensitization occurs on a time scale of 10 ms or less following receptor activation (Trussell and Fischbach, 1989; Tang et al., 1989). Several agents that inhibit this rapid desensitization have been described and have proven useful in studying the physiological importance of the process. These agents include the lectins concanavalin A (con A) and wheat germ agglutinin (WGA), that act by binding to carbohydrate residues on or near AMPA receptors to inhibit the conformational changes underlying desensitization (Thio et al., 1993). Other agents, including the nootropic drug, aniracetam (Tang et al., 1991), and the benzothiadiazide derivatives, diazoxide (Yamada and Rothman, 1992) and cyclothiazide (Yamada and Tang, 1993), inhibit desensitization by less certain mechanisms. Interestingly, cyclothiazide alters desensitization of re-

combinant AMPA receptors but not desensitization of recombinant kainate receptors (Partin et al., 1993). 2.3. ACPD (metabotropic) receptors ACPD receptors are linked by G-proteins to the phosphoinositide-protein kinase C (PI-PKC) and adenylate cyclase-cyclic AMP second messenger systems (Conn and Desai, 1991; Winder and Conn, 1992). Actions on the PI system promote the release of Ca^+ from intracellular stores (Conn and Desai, 1991). The most selective agonist for ACPD receptors is lS,3R-trans-ACPD, although Glu, ibotenate and Quis also act at these receptors. A related class of metabotropic receptors is activated by APB. Seven different metabotropic Glu receptors (mGluRl-7) have been cloned to date (Hollmann and Heinemann, 1994; Saugstad et al., 1994). The mGluRl and mGluR5 receptors activate PI turnover while mGlur2 and mGluR3 are negatively coupled to adenylate cyclase. The mGlur4, mGluR6 and mGluRV receptors are likely to be sites activated by APB. The cloned metabotropic receptors have seven membrane spanning regions like other G-protein coupled receptors, but are larger than other G-protein coupled receptors and show no amino acid sequence identity to other G-protein-linked receptors. The most selective antagonists for metabotropic receptors are phenylglycine derivatives, including a-methyl-4-carboxyphenylglycine (MCPG) (Watkins and Collingridge, 1994). 2-Amino-3-phosphonopropionate (AP3) and APB appear to act as non-competitive antagonists at some metabotropic receptors but the mechanism underlying the inhibition remains uncertain (Schoepp et al., 1990b). Additionally, since APB has agonist actions at some sites, this agent is less useful for probing the function of metabotropic receptors (Schoepp et al., 1990a). Interestingly, NMDA also inhibits the PI production mediated by metabotropic receptors, although the mechanism of this interaction is uncertain (Palmer et al., 1988). In some CNS regions, metabotropic receptors exert greater influences on neuronal activity in immature animals than in adults (Nicoletti et al.,

514

1986). Consistent with this, metabotropic receptors may play important roles in development of the visual cortex by helping to define the critical period for ocular dominance changes (Dudek and Bear, 1989). Additionally, in the hippocampus, ACPD receptors contribute to LTP in immature (Izumi et al., 1991b) but not adult rats, while contributing to LTD in the hippocampus of adult animals (Stanton et al., 1991). ACPD receptors may also be important regulators of fast excitatory synaptic transmission by presynaptically inhibiting the release of Glu (Baskys and Malenka, 1991). Some of these presynaptic effects are likely to occur at receptors that are also activated by APB. It is less clear how or whether metabotropic receptors participate in excitotoxicity, but some experiments suggest that activation or inhibition of these sites in developing animals can have toxic consequences. 2.4. APB receptors APB depresses Glu release presynaptically in certain regions (Forsythe and Clements, 1990) and blocks voltage-gated Ca^"^ currents (Trombley and Westbrook, 1992). These actions appear to occur at metabotropic receptors that are distinct from ACPD sites (possibly mGluR4, 6 or 7) (Saugstad et al., 1994). In the retina, APB receptors play a prominent role in synaptic transmission (Slaughter and Miller, 1981; Miller and Slaughter, 1986). There is presently no evidence that APB receptors participate in excitotoxic effects of EAA, although the presynaptic inhibitory effects could help to diminish Glu release during periods of intense synaptic activation. 3. Mechanisms of excitatory amino acidmediated neuronal damage Lucas and Newhouse (1957) initially observed that subcutaneous injections of Glu damage neurons in the immature mouse retina. Subsequent studies demonstrated that oral or subcutaneous administration of Glu not only damages retinal neurons but also neurons in brain regions lacking bloodbrain barriers (Olney, 1969b). In structure-activity

Excitotoxic neuronal damage and neuropsychiatric

disorders

experiments, it was found that the most effective neurotoxins were kainate > A^-methyl-aspartate (NMA) > homocysteate > Asp = Glu. This rank order closely paralleled the ability of these agents to depolarize neurons (Curtis and Watkins, 1960), and agents lacking a depolarizing action had no toxic effect (Olney et al., 1971). This relationship between excitatory and neurotoxic actions prompted the 'excitotoxic' hypothesis which postulates that EAA-induced damage results from prolonged membrane depolarization and the accompanying changes in membrane permeability and ion homeostasis. Experiments performed over the past twenty five years have supported and extended this hypothesis providing evidence for several forms of EAA-mediated damage. 3.1. Acute excitotoxicity Exposure of neurons to high concentrations of EAA for prolonged periods causes profound swelling of dendrites and somata with relative sparing of axons. In cultured hippocampal neurons (Rothman, 1985) and isolated retinae (Olney et al., 1986), this acute neuronal damage is dependent upon the presence of Na^ and Cl~ but not Ca^"*". This suggests that the damage is mediated by the influx of Na+ through EAA-gated ion channels and the redistribution of CI" and water. The osmotic overload results in disruption of the neuronal membrane. The role of Ca^"^ in acute excitotoxicity is less clear and may depend on the neuronal type. In the experiments described above, Ca^"^ removal had no effect on acute neuronal swelling. However, in cerebellar slices Ca^"*" contributes to the acute damage (Garthwaite and Garthwaite, 1986). 3.2. Delayed excitotoxicity A slower form of neuronal damage occurs in cultured neurons following brief exposures to NMDA agonists (Choi et al, 1987). This damage take hours to develop and is dependent on the presence of extracellular Ca^"^ (Choi, 1987). Treatments that hasten the recovery of resting intracellular Ca^"^ levels diminish the extent of the EAA-induced cell loss (Manev et al., 1989;

515

C.F. Zorumski andJ.W. Olney

DeErasquin et al., 1990). Additionally, in cultured cerebellar and hippocampal neurons, a delayed increase in intracellular Ca^"^ appears to play an important role in promoting neurodegeneration. Following removal of the excitotoxin and the initial EAA-mediated rise in Ca^"^ levels there is a recovery to baseline. In neurons that are destined to die there is a characteristic secondary, delayed Ca^"^ rise. The delayed Ca^"^ rises and the cell death depend on the presence of extracellular Ca^"^ during the five min EAA exposure (Randall and Thayer, 1992). The cascade of events following the initial Ca^"^ increases that lead to cell death are not certain. Several lines of evidence suggest that nitric oxide (NO) may be an important mediator of delayed toxicity. NO inhibitors block delayed NMDAmediated neuronal loss in cultured cortical neurons (Dawson et al., 1991) and hippocampal slices (Izumi et al., 1991a). Previously, Garthwaite and colleagues (1988) found that NMDA receptors promote the Ca^^-dependent release of NO in CNS neurons. The short half life of NO suggests that if this agent is involved in delayed neurotoxicity, it likely activates longer-lived cellular processes or is released in an ongoing fashion for some critical period (Moncada et al., 1991). Possible targets of NO include glycolytic enzymes (particularly glyceraldehyde-3-phosphate dehydrogenase) and the mitochondrial electron transport chain. Another possibility is that NO promotes poly-adenosine diphosphate (poly-ADP) ribosylation of nuclear proteins in response to DNA damage, leading to nicotinamide adenine dinucleotide (NAD) depletion and cellular energy compromise. Inhibitors of poly(ADP-ribose) synthetase block NMDA and NO-mediated neurotoxicity in cultured cortical neurons (Zhang et al., 1994). Furthermore, the secondary elevations of intracellular Ca^"*" observed in delayed toxicity paradigms could provide a signal for ongoing production of NO. NO synthase, the enzyme that catalyzes the production of NO from L-arginine, has reduced nicotinamide adenine dinucleotide phosphate (NADPH) diaphorase activity (Hope et al, 1991) or co-localizes with NADPH diaphorase in neurons (Bredt et al, 1991). NADPH diaphorase-containing neurons are

resistant to NMDA-mediated toxicity (Koh and Choi, 1988), suggesting that neurons that produce NO may be relatively resistant to its toxic effects, though the mechanisms responsible for this lack of sensitivity are unknown. 3.3, Slow excitotoxicity In addition to NMDA-mediated delayed toxicity, non-NMDA agonists also damage neurons in a Ca^-^-dependent fashion during prolonged agonist exposures. This damage can be at least partially attenuated by voltage-gated Ca^"^ channel blockers (Weiss et al., 1990), suggesting that the depolarization produced by non-NMDA agonists provides a stimulus for activation of these channels. Whether Type II kainate receptors, which have a relatively high Ca^"^ permeability (lino et al., 1990), contribute to the toxicity is unknown. Slowly developing non-NMDA toxicity may be important for understanding the role of EAA in slowly progressive neurodegenerative disorders. 3.4. NMDA antagonist toxicity Certain NMDA antagonists, including the dissociative anesthetics, have profound behavioral effects in man. Olney et al. (1989) found that these agents also produce pathological changes in the CNS. Both competitive and non-competitive NMDA antagonists produce cytoplasmic vacuoles in pyramidal neurons of the posterior cingulate and retrosplenial cortices. In rodents, lower doses of MK-801 produce vacuoles that are prominent 2 h or more after treatment with MK-801 but subside by 24 h. An apparent tolerance to vacuole formation develops after repeated administration of low doses of NMDA antagonists. Higher doses of MK801 produce irreversible damage to cingulate neurons. The vacuoles are prevented by pretreatment with antimuscarinic agents and drugs that augment the function of y-aminobutyric acidA ( G A B A A ) receptors, including benzodiazepines and barbiturates (Olney et al, 1991). The ability of anticholinergic drugs to block vacuole formation correlates with potency for binding to M3 muscarinic receptors (Bolden et al., 1992). NMDA antagonist

516

toxicity appears to reflect disruption of a complex circuit through which Glu, acting at a NMDA receptor on a GABAergic intemeuron, maintains tonic inhibition over the release of acetylcholine from axon terminals that innervate pyramidal neurons in layers III and IV of the posterior cingulate cortex. Blockade of the NMDA receptor inactivates the GABAergic interneuron, thereby disinhibiting release of acetylcholine onto the cingulate pyramidal neuron. Thus, the proximate cause of the pathomorphological changes in cingulate neurons appears to be excessive cholinergic stimulation of these neurons. Recent experiments indicate that the circuitry responsible for vacuole formation is more complex than originally believed, with the possible involvement of sigma and kainate receptors. Other transmitter systems may also be involved in a modulatory role. Since certain NMDA antagonists (PCP, ketamine) are associated with the production of acute psychotic symptoms and certain drugs (benzodiazepines and barbiturates) block both the neurotoxic effects of NMDA antagonists in rats and psychotomimetic effects in humans, it is possible that a common mechanism underlies these two types of side effects (Olney et al., 1991). 3.5. ACPD receptor toxicity Metabotropic (ACPD) receptors have been associated with neurotoxicity that results from either activation or inhibition of receptors. Fix et al. (1993) found that infant rats injected subcutaneously with AP3 develop with an almost complete absence of optic nerves. The primary effect appears to be on the retina which shows degeneration over a period of five to seven daily APS injections. In the brain, the toxic effects of AP3 are largely confined to regions lacking blood-brain barriers, but some animals exhibit more distributed lesions. Injections of 1S,3R trans-ACPD into the lateral ventricle of rats induces vacuolar changes in large lateral septal nucleus neurons. These changes are reminiscent of those induced by NMDA antagonists in the cingulate cortex. It has been proposed that the ACPD-induced lesions in the lateral septum and the NMDA antagonist lesions in the cin-

Excitotoxic neuronal damage and neuropsychiatric

disorders

gulate cortex share a common pathological mechanism, excessive activation of the PI second messenger system (Price et al., 1992). 4. Excitotoxins and neurodegenerative illnesses A recurring theme in the acute neurodegenerative syndromes is that oxidative stress and acute energy failure act in concert with EAA to cause neuronal death (Coyle and Puttfarcken, 1993). During periods when energy metabolism is compromised (status epilepticus, trauma, hypoglycemia and ischemia), neurons are particularly sensitive to excitotoxic damage (Novelli et al., 1988), and alterations in energy metabolism may underlie the selective vulnerability of some neurons to degeneration. The enhanced susceptibility to neuronal damage in these conditions may result from the generation of specific free radicals that have neurotoxic properties. Additionally, low energy conditions impair the function of ion pumps allowing changes in transmembrane ion concentrations that disrupt glutamate transport and allow accumulation of toxic levels of EAA in the vicinity of neurons (Lipton and Rosenberg, 1993). These conditions also produce neuronal membrane depolarization which relieves the Mg^+ block of NMDA ion channels and increases the probability that even normal extracellular concentrations of Glu will produce significant current flow through NMDA channels. The combination of membrane depolarization and increasing extracellular Glu levels is likely to have disastrous consequences on neurons. Alterations in energy metabolism may also be important in certain chronic neurodegenerative disorders to be discussed below (Section 4.2.2), where defects in mitochondrial function may predispose to selective neuronal vulnerability (Beal et al., 1993). 4.1. Environmental excitotoxins 4.1.1. Neuroendocrinopathies Converging lines of evidence indicate that Glu and EAA are important regulators of neuroendocrine function (Brann, 1995). Rodents treated with Glu in infancy sustain acute damage to the arcuate

C.F. Zorumski and J.W. Olney

nucleus of the hypothalamus (AH) which is a neuroendocrine regulatory center. As these animals grow to adulthood, they manifest a syndrome characterized by obesity, skeletal shortening and reproductive failure, with reduced plasma concentrations of luteinizing hormone (LH) and growth hormone (GH) (Olney, 1969a,b). Although Glu treatment in infancy does not directly damage the pituitary, Glu-treated animals exhibit a smaller anterior lobe of the gland in adulthood, suggesting the loss of an important trophic influence during development (Olney and Price, 1980). Destruction of AH neurons is not the only way that Glu influences neuroendocrine function. Doses of Glu lower than those required to injure or kill AH neurons activate these neurons and thereby influence the neuroendocrine parameters that they regulate. For example, parenteral administration of Glu in doses that are nontoxic causes an acute elevation of plasma LH in young adult male rats and an elevation of both LH and GH in prepubescent monkeys (Olney et al., 1976a; Medhamurthy et al., 1990). In female monkeys, NMA induces a similar increase in the release of LH, prolactin and follicle stimulating hormone (Gay and Plant, 1987). The LH-releasing effect of NMA is dependent on AH neurons and is inhibited by NMDA antagonists (Price et al, 1979). In primates and lower species, administration of NMA during the prepubescent period accelerates the onset of puberty, whereas NMDA antagonists, when similarly administered, retard the onset of puberty (Plant et al., 1989; Urbansky and Ojeda, 1990). Thus it is clear that even without damaging CNS neurons, Glu and related EAA have profound effects on neuroendocrine regulatory status. This raises the question whether it is wise to continue exposing immature humans, on a repetitive basis throughout their formative years, to foods that contain high concentrations of added Glu. This question deserves very careful consideration in view of evidence that ingestion of a given amount of Glu results in much higher and more sustained blood Glu elevations in humans than in rodents or monkeys (Stegink et al., 1979). Although in vivo studies pertaining to Glu neurotoxicity have typically relied on either subcutaneous or feeding tube administration, Olney et al.

517

(1980) demonstrated that when weanling mice are water deprived overnight, they voluntarily ingest sufficient quantities of Glu or Asp in drinking water to produce hypothalamic damage. This signifies that the developing hypothalamus is sensitive to both the excitatory and neurotoxic effects of Glu following voluntary oral intake and that this sensitivity extends well beyond the neonatal period. An additional determinant of vulnerability is suggested by two lines of evidence, one showing that hypothalamic damage induced by Glu in the immature brain, including AH, is mediated exclusively by NMDA receptors, and the other showing that NMDA receptors in the immature brain are hypersensitive to excitotoxic stimulation (MacDonald et al., 1988; Ikonomidou et al., 1989a; Wang et al., 1990). This provides an explanation for the long recognized fact that infant rodents are much more sensitive to Glu-induced hypothalamic damage than adults. 4.7.2. Amyotrophic lateral sclerosisParkinsonism-dementia complex ofguam Certain individuals on Guam and other South Sea Islands exhibit an increased incidence of amyotrophic lateral sclerosis (ALS), parkinsonism and dementia. The absence of inheritable or transmissible factors in this syndrome led to a search for environmental causes and the identification of the false sago palm (Cycas circinalis) as a likely source of the causative agent. This plant contains a number of potential neurotoxins including ^-A^methylamino-L-alanine (BMAA), an agent with weak excitotoxic properties (Spencer et al., 1990). When BMAA is gavage fed to monkeys for several weeks, the animals develop signs of movement disorder with damage in the motor cortex and spinal cord (Spencer et al., 1987). In mice, BMAA acutely damages cerebellar neurons, producing lesions that have the pathological appearance of Glu-type injury (Seawright et al., 1990). In vitro cell culture studies demonstrate that BMAAmediated neuronal damage results from activation of both NMDA and non-NMDA ionotropic receptors. Additionally, BMAA has potent effects at metabotropic EAA receptors, activating PI turnover at concentrations lower than those required to

518

activate NMDA receptors (Copani et al, 1990). This raises the possibility that the PI second messenger system and the release of intracellular Cd?-^ may be important in the toxic effects of this agent. The BMAA molecule lacks the terminal acidic group that is a critical basis for the neurotoxic actions of other EAA. Weiss and Choi (1988) observed that the excitatory and toxic properties of BMAA depend on the presence of bicarbonate ions and suggested that bicarbonate may change BMAA into a molecule with much greater excitotoxic potency. Consistent with this, Nunn and colleagues (1991) reported that BMAA is transformed by bicarbonate into an a-amino-carbamate that resembles the NMDA molecule. Although BMAA has been an instructive agent, it is unclear whether it is the principal toxin in cycad plants (Kisby et al., 1992). Studies using cycad extracts found no correlation between BMAA content and toxicity in cultured neurons. However, the cycad samples had a high content of Zn^"^, and Zn^"^ produced a dose-dependent damage of neurons that parallelled the Zn^+ content of the cycad extracts (Duncan et al., 1992). It is possible that a complex mechanism involving Zn^^, BMAA and perhaps other cycad toxins contributes to the ALSparkinsonism-dementia complex of Guam. 4,L3, Neurolathyrism Neurolathyrism is a crippling upper motor neuron disorder (Spencer et al., 1986), which, like the Guamanian ALS syndrome, cannot be attributed to genetic or infectious causes. This disorder is endemic to Africa and Asia. Epidemiological studies suggest that the causative agent is contained in the seeds of the chickling pea {Lathyrus sativus) which become a dietary staple during periods of famine in regions where the disorder is endemic (Spencer et al.. 1986). The seeds of the chickling pea contain ^-A^-oxalylamino-L-alanine (BOAA), an agent that has the structural features of an EAA and produces excitotoxic brain lesions when administered to immature rodents (Olney et al., 1976b). BOAA is a fairly selective AMPA receptor agonist, displaying preference for cortical AMPA receptors (Ross et al., 1989). In cortical cell cultures, BOAA kills neurons slowly, taking days of exposure at

Excitotoxic neuronal damage and neuropsychiatric

disorders

moderate concentrations to produce significant neurodegeneration (Weiss et al., 1989). Monkeys maintained on a diet containing chick pea pellets or BOAA develop features of neurolathyrism (Spencer et al., 1986). Signs of the disorder appear months after the start of the chick pea diet and sooner after starting a BOAA-enriched diet. The primate disorder, like the human syndrome, appears to be self-limiting with stabilization of symptoms after cessation of exposure to the toxin (Spencer et al., 1990). 4.1.4. Domoate poisoning In 1987, an outbreak of food poisoning occurred in eastern Canada. Affected individuals displayed acute gastrointestinal and neurological symptoms including confusion, memory impairment and seizures (Perl et al., 1990; Teitelbaum et al., 1990). Three patients died as a result of the poisoning and a number of survivors were left with severe anterograde amnesia. Elderly males were at highest risk for developing memory impairments. Among cases examined pathologically, neuronal loss was found in the hippocampus and amygdala. The syndrome was traced to the ingestion of cultured blue mussels harvested from Prince Edward Island. These mussels contained high concentrations of domoate, a potent neuroexcitant and neurotoxin that is a structural analog of kainate. The source of the domoate was a plankton, Nitzchia pungens, which was found in the water of the island at the time. Nitzchia pungens synthesizes large quantities of domoate which becomes concentrated in the flesh of mussels that filter feed upon the plankton. Domoate exhibits many of the physiological and toxicological effects of kainate. After systemic injection, kainate-treated animals experience prolonged seizures, and brain damage involving the hippocampus and amygdala among other areas (Lothman and Collins, 1981). Systemic injections of domoate produce similar seizures and brain damage in rats suggesting that the seizures play an important role in the neurodegenerative process (Stewart et al., 1990). Receptor binding and physiological studies have shown that domoate is likely to exert its effects through actions on kainate

C.F. Zorumski and J.W. Olney

receptors (Slevin et al., 1983; Stewart et al., 1990). These receptors are heavily concentrated in the CA3 region of the hippocampus (Monaghan and Cotman, 1982) which may explain the hippocampal damage. However, based on analogy to the kainate syndrome in animals, it appears that damage in regions outside the hippocampus results from the prolonged seizures and the release of Glu at NMDA receptors. This conclusion is based on evidence that NMDA receptor antagonists attenuate seizure-mediated brain damage, including that produced by kainate (Clifford et al., 1989, 1990). Other evidence suggests a role for NMDA receptors in the domoate syndrome, in that mussel extracts are more toxic than purified domoate to cultured neurons. The enhanced toxicity may be due to high concentrations of Glu and Asp in the mussel tissue and the ability of domoate to augment the toxicity of these other excitotoxins (Novelli et al., 1992). However, it is unclear whether Glu and Asp in the mussels actually gain access to the CNS unless domoate alters the permeability of bloodbrain barriers. 4.1.5. L- Cysteine toxicity L-cysteine (Cys) is an amino acid present normally in the CNS and environment. Following systemic administration to infant rats, Cys produces acute neurodegeneration (Olney et al., 1972a). Cys also crosses the placenta and produces neuronal damage in the fetal brain when administered to pregnant rodents. The lesions produced by Cys resemble those produced by Glu, but differ in affecting more widespread areas of the CNS. Whereas systemic Glu damages only CNS regions lacking blood-brain barriers, Cys produces damage throughout the forebrain suggesting that this agent readily penetrates blood-brain barriers. Damage typically involves the neocortex, hippocampus, septum, caudate and thalamus. Interestingly, the distribution and pattern of lesions produced by Cys resemble those produced by perinatal hypoxia/ischemia suggesting that this agent could serve as a model for studying mechanisms underlying developmental neuropathology syndromes. Cys exerts its actions primarily at NMDA receptors, although non-NMDA receptors con-

519

tribute at higher concentrations (Olney et al., 1990a). The physiological and toxic effects of Cys are much more prominent in the presence of bicarbonate ions. Like BMAA, another bicarbonatedependent toxin, Cys lacks the dicarboxylic acid structure that is typical of EAA, but is converted in the presence of bicarbonate to an a-aminocarbamate resembling other excitotoxins (Nunn et al., 1991). The observation that patients with ALS, Parkinson's disease and Alzheimer's disease have elevated Cys/sulfate ratios in blood has stimulated clinical interest in Cys (Heafield et al., 1990), although a study using a different means of assaying Cys was unable to reproduce the findings in patients with ALS (Perry et al., 1991). 4.2. Endogenous excitotoxins 4.2.1. Acute neurodegenerative

syndromes

Seizure-related brain damage. Prolonged seizures occurring in pathways using Glu as a neurotransmitter produce dendrosomal neuronal damage similar to that produced by EAA (Collins and Olney, 1982; Olney et al., 1983). Typically this damage requires about one hour of continuous seizure activity to become manifest. Seizure-related brain damage (SRBD) is blocked by the non-competitive NMDA antagonists PCP, ketamine and MK-801 (Clifford et al., 1989, 1990) further supporting the hypothesis that an excitotoxic process underlies this form of damage. Interestingly, NMDA antagonists protect against SRBD without completely arresting electrical seizure activity in the protected brain region. This signifies that the protection afforded by NMDA antagonists cannot be attributed to a general anticonvulsant effect, that it is possible to block the NMDA receptor-mediated component of seizure activity without abolishing other components, and that the component mediated by NMDA receptors is primarily responsible for SRBD. Acute trauma. The neuronal damage resulting from CNS trauma may be mediated, at least in part, by excitotoxins. Concussive brain injury results in significant elevations of extracellular Glu

520

levels (Faden et al., 1989; Katayama et al., 1990). Additionally, extracellular potassium is also elevated setting up a situation in which neurons are depolarized and NMDA receptors are activated. Furthermore, the increase in extracellular K"*" depolarizes glial cells and inhibits glutamate transport, leading to accumulation of glutamate in the extracellular space (Danbolt, 1994). Thus, following trauma, neurons are exposed to conditions in which significant ion flux through NMDA channels can occur. In cortical cell cultures, traumatic neuronal injury is blocked by EAA antagonists (Tecoma et al., 1989), and in vivo the behavioral effects of head or spinal cord trauma can be reduced by treatment with either NMDA or nonNMDA receptor antagonists (Faden et al., 1989; Wrathall et al., 1992). Hypoglycemia, In man and animals, acute hypoglycemia can damage the CNS. Sandberg et al. (1986) found that there is an increased release of EAA during hypoglycemia suggesting that EAA may be involved in the damage. The histological picture of hypoglycemic brain damage is similar to Glu-mediated cytopathology (Auer et al., 1985). In animals, lesions of Glu pathways innervating vulnerable brain regions, or the administration of NMDA antagonists protect against hypoglycemic neurodegeneration (Wieloch 1985; Wieloch et al., 1985). Hypoxia/ischemia. Several lines of evidence suggest that anoxic or ischemic brain damage is mediated by an excitotoxic process. First, in hippocampal cell cultures, anoxic cell death is dependent on excitatory synaptic transmission and EAA antagonists protect against this damage (Rothman, 1984). Second, acute ischemia in vivo produces significant elevations of extracellular Glu and Asp in the rat hippocampus (Benveniste et al., 1984) and intra-hippocampal injections of NMDA antagonists protect neurons from acute ischemic injury (Simon et al., 1984). Third, acute hypoxic/ischemic conditions produce a type of histopathology in the CNS that resembles that produced by excitotoxins (Ikonomidou et al., 1989b). In addition to these observations, there is now con-

Excitotoxic neuronal damage and neuropsychiatric disorders

siderable pharmacological evidence using EAA antagonists to indicate that excitotoxins, acting at both NMDA and non-NMDA receptors, contribute substantially to the neuronal insult in in vivo and in vitro stroke models (see Choi, 1990 for review). 4.2.2. Chronic neurodegenerative syndromes Inborn errors of metabolism. Sulfite oxidase deficiency is a rare disorder resulting from the accumulation of cysteine-5-sulfate (CSS) (Mudd et al., 1967). This leads to neuronal degeneration that is manifest clinically as blindness and spastic quadriplegia. Death typically occurs early in infancy. CSS is a structural analogue of Glu that displays excitotoxic activity when systemically administered to infant rats or when injected into the brains of adult rats (Olney et al., 1975). CSS toxicity is blocked by NMDA antagonists. Adult onset olivopontocerebellar degeneration (OPCD) is a syndrome characterized by ataxia and motor incoordination. Individuals with this disorder have a deficiency in Glu dehydrogenase that impairs the metabolism of Glu. Ingestion of Glu causes abnormally high levels of the amino acid in blood (Plaitakis et al, 1984). A similar accumulation of Glu occurring in the CNS could cause a slowly developing neuronal degeneration. Amyotrophic lateral sclerosis. ALS is a chronic neuromuscular disorder characterized by muscle wasting, fasciculations and spastic paraparesis, resulting from degeneration of motoneurons in the cerebral cortex and spinal cord. In patients with ALS, ingestion of Glu leads to abnormally high blood Glu levels (Plaitakis and Caroscio, 1987) and, in addition, ALS patients have abnormally high levels of Glu, Asp and A^-acetyl-aspartylglutamate (NAAG) in their cerebrospinal fluid (CSF) (Rothstein et al., 1990). However, these individuals, unlike patients with OPCD, do not have a defect in Glu dehydrogenase. Rather, some patients with ALS have a defect in a high affinity Glu transporter (Rothstein et al., 1992) that is a principal means for clearing Glu from the extracellular space. In animals, inhibitors of Glu transport are neurotoxic, presumably because of the in-

C.F. Zorumski and J.W. Olney

creased extracellular accumulation of endogenous excitotoxins (McBean and Roberts, 1985), and prolonged inhibition of glutamate transport produces a slowly developing neurodegeneration that is mediated primarily by non-NMDA receptors (Rothstein et al, 1993). Recent clinical studies have found that riluzole, an agent that diminishes glutamate release, slows the progression of illness in some patients with ALS further supporting the involvement of an excitotoxic process in the disorder (Bensimon et al., 1994). While it is reasonable to propose that the elevated EAA levels in ALS are of pathological importance, it remains to be explained why motoneurons are selectively vulnerable. Recent developments may help to answer this question. Patients with a familial form of ALS have a defect in the gene that encodes Cu/Zn superoxide dismutase (SOD), an important enzyme in free radical scavenging (Deng et al., 1993). Transgenic mice with mutations in the Cu/Zn SOD gene overexpress a form of SOD and develop a progressive agerelated defect in motor function with similarities to ALS (Ripps et al., 1995). Although the mechanisms involved in disease formation are not certain, selective vulnerability of motoneurons to certain free radicals is an important area of investigation. Huntington's and Parkinson's diseases. Several lines of evidence suggest that Huntington's disease (HD) and Parkinson's disease (PD), may result from excitotoxic insults. Injection of kainate into the rat striatum produces pathological and biochemical changes characteristic of HD and kainate injection has been used as an animal model for the disease (Coyle et al., 1978). Intrastriatal injection of high doses of Glu produces similar damage suggesting that a local defect in the Glu system may be responsible for the striatal damage in HD (Olney, 1979). An alternative hypothesis postulates that quinolinate, an excitotoxin found naturally in the brain, may be the causative agent (Real et al., 1986) since this agent is more potent in destroying striatal neurons than other CNS neurons and spares a population of striatal cells that are also spared in HD (Schwarcz and Kohler, 1983).

521

However, arguing against the quinolinate hypothesis is evidence that quinolinate levels are not elevated in the CSF or striatum of HD patients (Reynolds et al, 1988). An interesting observation regarding the aspiny striatal cells that are spared in HD is that these neurons contain NADPH diaphorase (NO synthase) (Ferrante et al., 1985). Although these neurons are resistant to NMDAmediated toxicity (Koh and Choi, 1988), they are likely to release NO and thus could serve as 'killer' cells for other neurons in the local environment. This hypothesis is tentative since there is presently no conclusive evidence linking NO to the neuronal damage in HD. Olney et al. (1990b) found that Ldihydroxyphenylalanine (L-DOPA), the natural precursor of dopamine, and its derivative, 6hydroxy-DOPA, exhibit excitotoxic activity primarily by actions at non-NMDA receptors. Other evidence suggests that the toxicity of 6-hydroxyDOPA results from oxidation to a quinone derivative (Rosenberg et al., 1991). These observations may be relevant to the pathophysiology of both HD and PD since the neurons that selectively degenerate in PD contain DOPA while those that degenerate in HD receive input from DOPAcontaining neurons (Albin et al., 1989). It is possible that DOPA or 6-hydroxy-DOPA could be generated in excessive amounts in the cell bodies of DOPA-containing neurons. If these toxins leaked from dopaminergic neurons in the substantia nigra or from dopaminergic terminals in the striatum into the extracellular space they could promote neuronal degeneration. An alternate possibility is that Cys could play a role in PD. This agent exerts excitotoxic activity at NMDA receptors (Olney et al., 1990a) and there is evidence suggesting a metabolic disturbance in patients with PD leading to elevated Cys levels (Heafield et al, 1990). A role for NMDA receptors in parkinsonism is also supported by studies focusing on the methamphetamine and l-methyl-4phenyl-l,2,3,6-tetrahydropyridine (MPTP) models of PD. The NMDA antagonist, MK-801, prevents the toxic effects of methamphetamine (Sonsalla et al., 1989) and MPTP (Turski et al., 1991) on nigrostriatal neurons. Additionally, ganglioside

522

GMl, an agent that prevents delayed excitotoxic cell damage in culture (DeErasquin et al., 1990), attenuates the motor symptoms in primates treated with MPTP (Schneider et al, 1992). In the MPTP model, interest has also centered on the ability of l-methyl-4-phenylpyridinium (MPP+), a toxic metabolite of MPTP, to alter energy metabolism by inhibiting the mitochondrial complex I respiratory chain. The resulting alteration in cellular energy increases the sensitivity of neurons to excitotoxic damage (Storey et al., 1992). Previously, Novelli et al. (1988) demonstrated that energy depletion enhances the neurotoxic effects of Glu in cultured neurons. Nigrostriatal neurons selectively accumulate MPP+ via dopamine transporters accounting for the cellular specificity in MPTPinduced damage (Irwin and Langston, 1985). Based on these observations, it is possible that an underlying metabolic defect renders specific neurons more susceptible to excitotoxic insult in patients with PD and HD (Beal, 1992). Alzheimer's disease. A consistent feature of Alzheimer's disease (AD) is the loss of basal forebrain cholinergic neurons (Bartus et al, 1982). These neurons are also destroyed by injection of excitotoxins into basal forebrain regions (Coyle et al., 1983) or by focal application of excitotoxins to the cerebral cortex (Sofroniew and Pearson, 1985), suggesting that either directly or indirectly, excitotoxins could mediate the loss of cholinergic cells in AD (Maragos et al., 1987). Interestingly, paired helical filaments of the type that comprise the neurofibrillary tangles seen in AD are found in cultured human spinal cord neurons following Glu exposure (DeBoni and McLachlan, 1985). Additionally, in cultured rat cortical neurons Glu causes an increase in the abnormally phosphorylated tau proteins that are found in paired helical filaments (Sindou et al., 1994). Thus it is possible that the development of paired helical filaments could represent a reaction to excitotoxins. In cultured rat hippocampal neurons, Glu exposure causes increased expression of antigens recognized by the Alz-50 and 5E2 antibodies (Mattson, 1990). These antigens are also expressed at increased levels in AD lesions. Furthermore, y3-amyloid protein.

Excitotoxic neuronal damage and neuropsychiatric disorders

which is of considerable interest in the pathophysiology of AD, appears to enhance the sensitivity of cultured cortical neurons to excitotoxic damage (Koh et al., 1990). This enhanced toxicity may result from destabilization of intracellular Ca^"^ homeostasis, resulting in higher basal levels of Ca^+ and increased responses to EAA (Mattson et al., 1992). The development of a potential animal model for AD in transgenic mice that overexpress a mutant amyloid precursor protein offers the hope of understanding the interactions between amyloid plaque formation and excitotoxicity (Games etal., 1995). In biopsies obtained from AD subjects early in the disease process there is a loss of Glu uptake sites, which could lead to an accumulation of EAA in the CNS (Procter et al., 1988). A loss of uptake sites would be expected to cause elevated levels of Glu in cerebrospinal fluid in AD. However studies examining this issue have produced conflicting data. A major problem in these studies has been the failure to study patients according to disease stage, because it is likely that EAA levels change as the disease progresses. Interestingly, Pomara et al. (1992) found elevations of EAA in cerebrospinal fluid taken from AD patients early in the disease course. It is also possible that other agents acting at EAA receptors which are not routinely measured in amino acid assays could contribute to the neuronal damage in AD. One such agent is the endogenous phosphomonoester, L-phosphoserine, an agent that acts at multiple EAA receptors and is elevated in the brains of AD patients (Klunk et al., 1991). Alternatively, based on the finding of abnormalities in sulfur metabolism in AD patients, it is possible that Cys could be involved in the disorder (Heafield et al., 1990). The interaction between altered intracellular energy metabolism and excitotoxicity also needs to be considered in AD. There is evidence that energy metabolism is impaired with aging (Bowling et al., 1993) and AD (Sims et al, 1985). As discussed previously, this renders neurons particularly vulnerable to NMDA-mediated excitotoxicity. Alternatively, or perhaps in parallel, it is important to consider whether NMDA receptor hypofunction which causes damage to posterior cin-

C.F. Zorumski and J.W. Olney

gulate cortex neurons following administration of NMD A receptor antagonists (Olney et al., 1989, 1991) could contribute to AD. NMDA receptor function diminishes with age (Tamaru et al., 1991), a situation that is likely to be worsened in AD. This may lead to conditions that unleash a toxic process similar to that accompanying NMDA receptor antagonists in which corticolimbic neurons are at risk for damage. Another possible tie between the pathology of AD and excitotoxins may be provided by the dementia pugilistica syndrome. This disorder is associated with the sport of boxing and results from repeated head trauma (Corsellis, 1978). The syndrome shares certain neuropathological features with AD, including the presence of abundant 6amyloid protein and neurofibrillary tangles in the brain (Roberts et al., 1990). It is currently believed that the brain damage resulting from CNS trauma may be, at least in part, EAA mediated. Dementia pugilistica thus may represent a model for understanding how an excitotoxic insult could lead to a chronic neurodegenerative illness. Acquired immunodeficiency syndrome encephalopathy. Neurological symptoms occur commonly in patients with the acquired immunodeficiency syndrome (AIDS). The CNS lesions in these subjects include white matter and glial lesions as well as neuronal loss. Although the CNS involvement in these individuals is likely to be multifactorial, there is interest in whether EAA could play a role in the neuronal loss. This interest stems from several observations. First, the CSF and serum of patients infected with human immunodeficiency virus-1 (HIV-1) exhibit increased levels of quinolinate, a weak excitotoxin acting at NMDA receptors (Heyes et al., 1990). Second, the HIV-1 envelope glycoprotein, gpl20, is toxic to cultured retinal ganglion cells and this toxicity is prevented by Ca^+ channel blockers and NMDA receptor antagonists (Lipton, 1991). Gpl20 does not alter physiological responses to EAA but appears to act synergistically with EAA to promote neurodegeneration (Lipton et al., 1991). This action, in some ways, is reminiscent of the enhancing effects of the ^-amyloid protein in cultured neurons (Koh et al..

523

1990). In cultured rat cortical neurons, NO contributes to Gpl20-induced neurodegeneration since NO inhibitors block the toxicity produced by the combination of Gpl20 and glutamate (Dawson et al., 1993). 5. Conclusions Several features of EAA action in the CNS are clear. First, it is clear that EAA act at several different receptor types, some of which are coupled directly to ion channels and others of which are linked to second messenger systems. It is also clear that our knowledge of EAA receptor types and structure is in its infancy. Second, the evidence for an EAA serving as a fast excitatory neurotransmitter is strong and EAA receptors are involved in several forms of synaptic plasticity that may be important in neurodevelopment and behavior. Thus, it is unequivocal that EAA serve important physiological functions in the CNS. Third, EAA can be toxic to neurons in many CNS regions and several forms of EAA-mediated toxicity exist. These observations lead to the conclusion that EAA are truly 'Jekyll-Hyde' molecules, playing vital physiological roles but also mediating the death of neurons if left unchecked. In this chapter, we have discussed evidence that is consistent with a role for EAA in a variety of human neurodegenerative disorders. In some cases, particularly the acute neurodegenerative disorders and certain syndromes produced by exogenous toxins, the evidence is becoming increasingly compelling. Evidence linking excitotoxins to chronic neurodegenerative illnesses remains both tentative and tantalizing, signifying the need for intensive new research. A major hope for the future is that through the use of molecular biological approaches it will be possible to design more selective pharmacological tools that will allow excitotoxic hypotheses to be tested with greater precision. These tools should also help to determine how ubiquitous agents like the EAA can produce focal damage in the CNS. The observation of potentially pathway specific toxins, such as 6-hydroxy-DOPA, is providing some insight into this issue. However, more infor-

524

mation about regional differences in receptor structure and function is likely to be enlightening. Better structural information about EAA receptors also offers the hope of generating more selective anti-excitotoxic drugs that can be used to combat neurodegeneration while having less influence on the normal synaptic actions of EAA. The development of more selective anti-excitotoxic agents may also arise from a better understanding of the cascade of events that follow activation of EAA receptors and ultimately lead to neuronal death. Thus efforts aimed at defining the Ca^+-dependent processes activated in delayed toxicity and prolonged exposure paradigms are likely to be profitable. Acknowledgements This work was supported in part by NIMH Research Scientist Development Award MH00964 (C.F.Z.), NIMH Research Scientist Award MH38894 (J.W.O.) and grants AG05681, AGl 1355 and MH45493. References Albin, R.L., Young, A.B. and Penney, J.B. (1989). The functional anatomy of basal ganglia disorders. Trends Neurosci. 12: 366-375. Auer, R.N., Kalimo, H., Olsson, Y. and Siesjo, B.K. (1985). The temporal evolution of hypoglycemic brain damage: II. Light and electron microscopic findings in the rat hippocampus. Acta Neuropathol (Berlin) 67: 25-36. Barnes, J.M. and Henley, J.M. (1992). Molecular characteristics of excitatory amino acid receptors. Prog. Neurobiol. 39: 113-133. Bartus, R.T., Dean, R.C., Beer, B. and Lippa, A.S. (1982). Cholinergic hypothesis of geriatric memory dysfunction. Science 217: 408-417. Baskys, A. and Malenka, R.C. (1991). Agonists at metabotropic glutamate receptors presynaptically inhibit EPSCs in neonatal rat hippocampus. J. Physiol. (London) 444: 687-701. Beal, M.F. (1992). Does impairment of energy metabolism result in excitotoxic neuronal death in neurodegenerative illnesses. Ann. Neurol. 31: 119-130. Beal, M.F., Hyman, B.T. and Koroshetz, W. (1993). Do defects in mitochondrial energy metabolism underlie the pathology of neurodegenerative disorders. Trends Neurosci. 16: 125-131. Beal, M.F., Kowall, N.W., Ellison, D.W., Mazurek, M.F. and Swartz, K.J. (1986). Replication of the neurochemical char-

Excitotoxic neuronal damage and neuropsychiatric disorders acteristics of Huntington's disease by quinolinic acid. Nature 321: 168-171. Bensimon, G., Lacomblez, L., Meininger, V., and the ALS/Riluzole Study Group. (1994). A controlled trial of riluzole in amyotrophic lateral sclerosis. N. Engl. J. Med. 330:585-591. Benveniste, H., Drejer, J., Schousboe, A. and Diemer, N.M. (1984). Elevation of the extracellular concentrations of glutamate and aspartate in rat hippocampus during transient cerebral ischemia monitored by intracerebral microdialysis. J. Neurochem. 43: 1369-1374. Bolden, C, Cusack, B. and Richelson, E. (1992). Antagonism by antimuscarinic and neuroleptic compounds at the five cloned human muscarinic cholinergic receptors expressed in Chinese hamster ovary cells. J. Pharmacol. Exp. Ther. 260: 576-580. Bowling, A.C., Mutisya, E.M., Walker, L.C., Price, D.L., Cork, L.C. and Beal, M.F. (1993). Age dependent impairment of mitochondrial function in primate brain. J. Neurochem. 60: 1964-1967. Brann, D.W. (1995). Glutamate: a major excitatory transmitter in neuroendocrine regulation. Neuroendocrinology 61: 213-225. Bredt, D.S., Glatt, C.E., Hwang, P.M., Fotuhi, M., Dawson, T.M. and Snyder, S.H. (1991). Nitric oxide synthase protein and mRNA are discretely localized in neuronal populations of the mammalian CNS together with NADPH diaphorase. Neuron 1\6\5-62A. Choi, D.W. (1987). Ionic dependence of glutamate neurotoxicity. J. Neurosci. 7: 369-379. Choi, D.W. (1990). Cerebral hypoxia: some new approaches and unanswered questions. J. Neurosci. 10: 2493-2501. Choi, D.W., Maulucci-Gedde, M. and Kriegstein A.R. (1987). Glutamate neurotoxicity in cortical cell culture. J. Neurosci. 7: 357-368. Clifford, D.B., Olney, J.W., Benz, A.M., Fuller, T.A. and Zorumski, C.F. (1990). Ketamine, phencyclidine and MK801 protect against kainic acid induced seizure-related brain damage. Epilepsia 31: 382-390. Clifford, D.B., Zorumski, C.F. and Olney, J.W. (1989). Ketamine and MK-801 prevent degeneration of thalamic neurons induced by focal cortical seizures. Exp. Neurol. 105: 272-279. Collins, R.C. and Olney, J.W. (1982). Focal cortical seizures cause distant thalamic lesions. Science 218: 177-179. Conn, J.P. and Desai, M.A. (1991). Pharmacology and physiology of metabotropic glutamate receptors in mammalian central nervous system. Drug Dev. Res. 24: 207-229. Copani, A., Canonico, P.L. and Nicoletti, F. (1990). p-Nmethylamino-L-alanine (L-BMAA) is a potent agonist of metabolotropic glutamate receptors, Eur. J. Pharmacol. 181:327-328. Corsellis, J.A.N. (1978). Posttraumatic dementia. In: R. Katzman, R.D. Terry and K.L. Bick (Eds.), Alzheimers Disease, Senile Dementia and Related Disorders, Raven Press, New York, pp. 125-133.

C.F. Zorwnski andJ.W. Olney Coyle, J.T., McGeer, E.F., McGeer, P.L. and Schwarcz, R. (1978). Neostriatal injections: a model for Huntington's chorea. In: E. McGeer, J.W. Olney and P. McGeer, P. (Eds.), Kainic Acid as a Tool in Neurobiology, Raven Press, New York, pp. 139-159. Coyle, J.T., Price, D.L. and DeLong, M.A. (1983). Alzheimer's disease: a disorder of cortical cholinergic innervation. Science 219: 1184-1190. Coyle, J.T. and Puttfarcken, P. (1993). Oxidative stress, glutamate and neurodegenerative disorders. Science 262: 689-695. Curtis, D.R., and Watkins, J.C. (1960). The excitation and depression of spinal neurons by structurally related amino acids. J. Neurochem. 6: 117-141. Danbolt, N.C. (1994). The high affinity uptake system for excitatory amino acids in the brain. Prog. Neurobiol. 44: 377-396. Dawson, V.L., Dawson, T.M., London, E.D., Bredt, D.S. and Snyder, S.H. (1991). Nitric oxide mediates glutamate neurotoxicity in primary cortical cultures. Proc. Natl Acad. Sci. f/5A 88: 6368-6371. Dawson, V.L., Dawson, T.M., Uhl, G.R. and Snyder, S.H. (1993). Human immunodeficiency virus type 1 coat protein neurotoxicity mediated by nitric oxide in primary cortical cultures. Proc. Natl Acad. Scl USA 90: 3256-3259. DeBoni, U. and McLachlan, D.R.C. (1985). Controlled induction of paired helical filaments of the Alzheimer type in cultured human neurons by glutamate and aspartate, J. Neurol ScL 6S: 105-118. DeErasquin, G.A., Manev, H., Guidotti, A., Costa, E. and Brooker, G. (1990). Gangliosides normalize distorted single-cell intracellular free Ca-^"^ dynamics after toxic doses of glutamate in cerebellar granule cells. Proc. Natl Acad. ScL

USA SI:

soil-soil.

Deng, H.-X., Hentafi, A., Tainer, J.A., et al. (1993). Amyotrophic lateral sclerosis and structural defects in Cu,Zn superoxide dismutase. Science 261: 1047-1051. Donevan, S.D. and Rogawaki, M.A. (1993). GYKI 52466, a 2,3-benzodiazepine, is a highly selective, noncompetitive antagonist of AMPA/kainate receptor responses. Neuron 10:51-59. Dudek, S.M. and Bear, M.F. (1989). A biochemical correlate of the critical period for synaptic modification in the visual cortex. Science 246: 673-675. Duncan, M.W., Marini, A.M., Walters, R., Kopin, I.J. and Markey, S.P. (1992). Zinc, a neurotoxin to cultured neurons, contaminates cycad flour prepared by traditional Guamanian methods. J. Neuro scl 12: 1523-1537. Faden, A.I., Demediuk, P., Panter, S. and Vink, R. (1989). The role of excitatory amino acids and NMDA receptors in traumatic brain injury. Science 244: 798-800. Ferrante, R.J., Kowall, N.W., Beal, M.F., Richardson, E.P., Bird, E.D. and Martin, J.B. (1985). Selective sparing of a class of striatal neurons in Huntington's disease. Science 230: 561-563. Fix, A.S., Schoepp, D.D., Olney, J.W., Vestre, W.A., Griffey, K.I., Johnson, J.A. and Tizzano, J.P. (1993). Neonatal ex-

525 posure to D,L-2-amino-3-phosphonopropionate (D,L-AP3) produces lesions in the eye and optic nerves of adult rats. Dev. Brain Res. 75: 223-233. Forsythe, I.D. and Clements, J.D. (1990). Presynaptic glutamate receptors depress excitatory monosynaptic transmission between mouse hippocampal neurones. J. Physiol (London) 429: 1-16. Forsythe, I.D. and Westbrook, G.L. (1988). Slow excitatory postsynaptic currents mediated by A^-methyl-D-aspartate receptors on cultured mouse central neurones. J. Physiol (London) 396: 515-533. Games, D., Adams, D., Alessandrini, R., et al. (1995). Alzheimer-type neuropathology in transgenic mice overexpressing V717F y3-amyloid precursor protein. Nature 373: 523-527. Garthwaite, J., Charles, S.L. and Chess-Williams, R. (1988). Endothelium-derived relaxing factor release on activation of NMDA receptors suggests role as intercellular messenger in the brain. Nature 336: 385-388. Garthwaite, G. and Garthwaite, J. (1986). Neurotoxicity of excitatory amino acid receptor agonists in rat cerebellar slices: dependence on calcium concentration. Neurosci. Lett. 66: 193-198. Gay, V.L. and Plant, T.M. (1987). N-Methyl-D-aspartate elicits hypothalamic gonadotropin-releasing hormone release in prepubertal male rhesus monkeys (macaca mulatto), Endocrinology 120:2289-2296. Gilbertson, T.A., Scobey, R. and Wilson, M. (1991). Permeation of calcium ions through non-NMDA glutamate channels in retinal bipolar cells. Science 251: 1613-1615. Heafield, M.T., Fearn, S., Steventon, G.B., Waring, R.H., Williams, A.D. and Sturman, S.G. (1990). Plasma cysteine and sulfate levels in patients with motor neurone, Parkinson's and Alzheimer's disease. Neurosci Lett. 110: 216-220. Heyes, M.P., Mefford, I.N., Quearry, B.J., Dedhia, M. and Lackner, A. (1990). Increased ratio of quinolinic acid to kynurenic acid in cerebrospinal fluid of D retrovirusinfected rhesus macaques: relationship to clinical and viral status. Ann. Neurol 27: 666-675. Hollmann, M., Hartley, M. and Heinemann, S. (1991). Ca^"^ permeability of KA-AMPA-gated glutamate receptor channels depends on subunit composition. Science 252: 8 5 1 853. Hollmann, M. and Heinemann, S. (1994). Cloned glutamate receptors. AnnM. Rev. Neurosci 17: 31-108. Honore, T., Davies, S.N., Drejer, J., Fletcher, E.J., Jacobsen, P., Lodge, D. and Nielsen, F.E. (1988). Quinoxalinediones: potent competitive non-NMDA glutamate receptor antagonists. Science 241, 701-703. Hope, B.T., Michael, G.J., Knigge, K.M. and Vincent, S.R. (1991). Neuronal NADPH diaphorase is a nitric oxide synthase. Proc. Natl Acad. Scl USA 88: 2811-2814. Huettner, J.E. and Bean, B.P. (1988). Block of //-methyl-Daspartate-activated current by the anticonvulsant MK-801: selective binding to open channels. Proc. Natl Acad. Scl USA S5: 1307-1311.

526 Hume, R.I., Dingledine, R. and Heinemann, S.F. (1991). Identification of a site in glutamate receptor subunits that controls calcium permeability. Science 253: 1028-1031. lino, M., Ozawa, S. and Tsuzuki, K. (1990). Permeation of calcium through excitatory amino acid receptor channels in cultured rat hippocampal neurones. J. Physiol. (London) 424: 151-165. Ikonomidou, C, Mosinger, J.L., Salles, K., Labruyere, J. and Olney, J.W. (1989a). Sensitivity of the developing rat brain to hypobaric/ischemic damage parallels sensitivity to A^methyl-D-aspartate neurotoxicity. J. Neurosci. 9: 28092818. Ikonomidou, C, Price, M.T., Mosinger, J.L., Friedrich, G., Labruyere, J., Salles, K.S. and Olney, J.W. (1989b). Hypobaric-ischemic conditions produce glutamate-like cytopathology in infant rat brain. J. Neurosci. 9: 1693-1700. Irwin, I. and Langston, W. (1985). Selective accumulation of MPP"*" in the substantia nigra: a key to neurotoxicity? Life Sci. 36: 207-212. Izumi, Y., Benz, A.M., Clifford, D.B. and Zorumski, C.F. (1991a). Nitric oxide inhibitors attenuate A^-methyl-Daspartate excitotoxicity in rat hippocampal slices. Neurosci. Lett. 135: 227-230. Izumi, Y., Clifford, D.B. and Zorumski, C.F. (1991b). 2Amino-3-phosphonopropionate blocks the induction and maintenance of long-term potentiation in rat hippocampal slices. Neurosci. Lett. 122: 187-190. Johnson, J.W. and Ascher, P. (1987). Glycine potentiates the NMD A response in cultured mouse brain neurons. Nature 325:529-531. Katayama, Y., Becker, D.P., Tamura, T. and Hovda, D.A. (1990). Massive increase in extracellular potassium and the indiscriminate release of glutamate following traumatic brain injury. J. Neurosurg. 73: 889-900. Kisby, G.E., Ellison, M. and Spencer, P.S. (1992). Content of the neurotoxins cycasin (methylazoxymethanol-y^-Dglucoside) and BMAA (y5-//-methylamino-L-alanine) in cycad flour prepared by Guam Chamorros. Neurology 42: 1336-1340. Kleckner, N.W. and Dingledine, R. (1988). Requirement for glycine in activation of NMDA-receptors expressed in Xenopus oocytes. Science 241: 835-837. Klunk, W.E., McClure, R.J. and Pettegrew, J.W. (1991). LPhosophoserine, a metabolite elevated in Alzheimer's disease interacts with specific L-glutamate receptor subtypes, J. Neurochem. 56: 1997-2003. Koh, J.Y. and Choi, D.W. (1988). Vulnerabihty of cultured cortical neurons to damage by excitotoxins: differential susceptibility of neurons containing NADPH-diaphorase. J. Neurosci. 8: 2153-2163. Koh, J.Y., Yang, L.L. and Cotman, C.W. (1990). )3-Amyloid protein increases the vulnerability of cultured cortical neurons to excitotoxic damage. Brain Res. 533: 315-320. Kuba, K. and Kumamoto, E. (1990). Long-term potentiation in vertebrate synapses: a variety of cascades with common subprocesses. Prog. Neurobiol. 34: 197-269.

Excitotoxic neuronal damage and neuropsychiatric disorders Lerma, J. (1992). Spermine regulates A^-methyl-D-aspartate receptor desensitization. Neuron 8: 343-352. Lipton, S.A. (1991). Calcium channel antagonists and human immunodeficiency virus coat protein-mediated neuronal injury. Ann. Neurol. 30: 110-114. Lipton, S.A. and Rosenberg, P.A. (1994). Excitatory amino acids as a final common pathway for neurologic disorders. N. Engl. J. Med. 330: 613-622. Lipton, S.A., Sucher, N.J., Kaiser, P.K. and Dreyer, E.B. (1991). Synergistic effects of HIV coat protein and NMDA receptor-mediated neurotoxicity. Neuron 7: 111-118. Lothman, E.W. and Collins, R.C. (1981). Kainic acid induced limbic seizures: metabolic, behavioral, electroencephalographic and neuropathological correlates. Brain Res. 218: 299-318. Lucas, D.R. and Newhouse, J.P. (1957). The toxic effect of sodium L-glutamate on the inner layers of the retina. AMA Arch Ophthalmol. 58: 193-201. MacDonald, J.W., Silverstein, F.S. and Johnston, M.V. (1988). Neurotoxicity of A^-methyl-D-aspartate is markedly enhanced in developing rat central nervous system. Brain Res. 459: 200-203. MacDonald, J.W. and Johnston, M.V. (1990). Physiological and pathophysiological roles of excitatory amino acids during central nervous system development. Brain Res. Rev. 15: 41-70. Manev, H., Favaron, M., Guidotti, A. and Costa, E. (1989). Delayed increase of Ca^"*" influx elicited by glutamate: role in neuronal death. Mol. Pharmacol. 36: 106-112. Maragos, W.F., Greenamyre, J.T., Penney, J.B. and Young, A.B. (1987). Glutamate dysfunction in Alzheimer's disease: an hypothesis. Trends Neurosci. 10: 65-68. Mattson, M.P. (1990). Antigenic changes similar to those seen in neurofibrillary tangles are elicited by glutamate and calcium influx in cultured hippocampal neurons. Neuron 4: 105-117. Mattson, M.P., Cheng, B., Davis, D., Bryant, K., Lieberburg, I. and Rydel, R.E. (1992). )3-Amyloid peptides destabilize calcium homeostasis and render human cortical neurons vulnerable to excitotoxicity. /. Neurosci. 12: 376-389. Mayer, M.L., Vyklicky, Jr., L. and Clements, J. (1989). Regulation of NMDA receptor desensitization in mouse hippocampal neurons by glycine. Nature 338: 425-427. Mayer, M.L. and Westbrook, G.L. (1985). The action of Nmethyl-D-aspartic acid on mouse spinal neurones in culture. J. Physiol. (London) 361: 65-90. Mayer, M.L. and Westbrook, G.L. (1987). Permeation and block of A^-methyl-D-aspartate acid receptor channels by divalent cations in mouse cultured central neurones. J. Physiol. (London) 394: 501-527. McBean, G.J. and Roberts, P.J. (1985). Neurotoxicity of Lglutamate and DL-threo-3- hydroxyaspartate in the rat striatum. J. Neurochem. 44: 247-254. Medhamurthy, R., Dichek, H.L., Plant, T.M., Bernardini, I. and Cutler, G.B. (1990). Stimulation of gonadotropin secretion in prepubertal monkeys after hypothalamic excita-

C.F. Zorumski and J.W. Olney tion with aspartate and glutamate. J. Clin. Endocrinol. Metab.ll: 1390-1392. Miller, R. and Slaughter, M. (1986). Excitatory amino acid receptors of the retina: diversity of subtypes and conductance mechanisms. Trends Neurosci. 9: 211-218. Monaghan, D.T., Bridges, R.J. and Cotman, C.W. (1989). The excitatory amino acid receptors: their classes, pharmacology, and distinct properties in the function of the central nervous system. Annu. Rev. Pharmacol. Toxicol. 29: 3 6 5 402. Monaghan, D.T. and Cotman, C.W. (1982). The distribution of [-^HJ-kainic acid binding sites in rat CNS as determined by autoradiography. Brain Res. 252: 91-100. Moncada, S., Palmer, R.M.J, and Higgs, E.A. (1991). Nitric oxide: physiology, pathophysiology, and pharmacology. Pharm. Rev. 43: 109-142. Mudd, S.H., Irreverre, F. and Laster, L. (1967). Sulfite oxidase deficiency in man: demonstration of the enzyme dtfect Science 156: 1599-1602. Nicoletti, F., ladarola, M.J., Wroblewski, J.T. and Costa, E. (1986). Excitatory amino acid recognition sites coupled with inositol phospholipid metabolism: developmental changes and interaction with a 1-adrenoceptors. Proc. Natl. Acad. Sci. USA 83: 1931-1935. Novelli, A., Kispert, J., Fernandez-Sanchez, M.T., Torreblanca, A. and Zitko, V. (1992). Domoic acid-containing toxic mussels produce neurotoxicity in neuronal cultures through a synergism between excitatory amino acids. Brain Res. 577: 41-48. Novelli, A., Reilly, J.A., Lysko, P.G. and Henneberry, R.C. (1988). Glutamate becomes neurotoxic via the A^-methyl-Daspartate receptor when intracellular energy levels are reduced. Brain Res. 451: 205-212. Nunn, P.B., Davis, A.J. and O'Brien, P. (1991). On carbamate formation and the neurotoxicity of L-a-amino acids. Science 25\: 1619. Olney, J.W. (1969a). Glutamate-induced retinal degeneration in neonatal mice. Electron microscopy of the acutely evolving lesion. J. Neuropathol. Exp. Neurol. 28: 4 5 5 ^ 7 4 . Olney, J.W. (1969b). Brain lesions, obesity and other disturbances in mice treated with monosodium glutamate. Science 164:719-721. Olney, J.W. (1979). Excitotoxic amino acids and Huntington's disease. In: T.N. Chase, A. Wexler and A. Barbeau (Eds.), Advances of Neurology, Huntington's Disease,Wo\. 23, Raven Press, New York, pp. 609-624. Olney, J.W., Cicero, T.J., Meyer, E.F. and DeGubareff, T. (1976a). Acute glutamate induced elevations in serum testosterone and leutenizing hormone. Brain Res. 112: 420424. Olney, J.W., DeGubareff, T. and Sloviter, R.S. (1983). Epileptic brain damage in rats induced by sustained electrical stimulation of the perforant path. II. Ultrastructural analysis of acute hippocampal pathology. Brain Res. Bull. 10: 699712. Olney, J.W., Ho, O.L. and Rhee, V. (1971). Cytotoxic effects

527 of acidic and sulphur-containing amino acids on the infant mouse central nervous system. Exp. Brain Res. 14: 61-76. Olney, J.W., Ho, O.L., Rhee, V. and Schainker, B. (1972a). Cysteine induced brain damage in infant and fetal rodents. Brain Res. 45:309-313. Olney, J.W., Labruyere, J. and DeGubareff, T. (1980). Brain damage in mice from voluntary ingestion of glutamate and ^spSLTiate. Neurobehav. Toxicol. 2: 135-129. Olney, J.W., Labruyere, J. and Price, M.T. (1989). Pathological changes induced in cerebrocortical neurons by phencyclidine and related drugs. Science 244: 1360-1362. Olney, J.W., Labruyere, J., Wang, G., Wozniak, D.F., Price, M.T. and Sesma, M.A. (1991). NMDA antagonist neurotoxicity: mechanism and prevention. Science 254: 15151518. Olney, J.W., Misra, C.H. and DeGubareff, T. (1975). Cysteine-iS-sulfate: a brain damaging metabolite in sulfite oxidase deficiency. J. Neuropathol. Exp. Neurol. 34: 167-176. Olney, J.W., Misra, C.H. and Rhee, V. (1976b). Brain and retinal damage from the lathyrus excitotoxin, ^-N-oxa\y\-La,y3-diaminopropionic acid (ODAP). Nature 264: 659661. Olney, J.W. and Price, M.T. (1980). Neuroendocrine interactions of excitatory and inhibitory amino acids. Brain Res. Bull. 5: 361-368. Olney, J.W., Price, M.T., Samson, L. and Labruyere, J. (1986) The role of specific ions in glutamate neurotoxicity. Neurosci. Lett. 65: 65-71. Olney, J.W., Zorumski, C , Price, M.T. and Labruyere, J. (1990a). L-Cysteine, a bicarbonate sensitive endogenous excitotoxin. Science 248: 596-599. Olney, J.W., Zorumski, C.F., Stewart, G.R., Price, M.T., Wang, G. and Labruyere, J. (1990b). Excitotoxicity of LDOPA and 6-OH-DOPA: implications for Parkinson's and Huntington's diseases. Exp. Neurol. 108: 269-272. Olverman, H.J., Jones, A.W. and Watkins, J.C. (1984). LGlutamate has higher affinity than other amino acids for [^H]-D-AP5 binding sites in rat brain membranes. Nature 307: 460-462. Ormandy, G.C. (1992). Inhibition of excitatory amino acidstimulated phosphoinositide hydrolysis in rat hippocampus by L-aspartate-y3-hydroxamate. Brain Res. 572: 103-107. Palmer, E., Monaghan, D.T. and Cotman, C.W. (1988). Glutamate receptors and phosphoinositide metabolism: stimulation via quisqualate receptors is inhibited by A^-methyl-Daspartate receptor activation. Mol. Brain Res. 4: 161-165. Partin, K.M., Patneau, D.K., Winters, C.A., Mayer, M.L. and Buonanno, A. (1993). Selective modulation of desensitization at AMPA versus kainate receptors by cyclothiazide and concanavalin A. Neuron 11: 1069-1082. Paternain, A.V., Morales, M. and Lerma, J. (1995). Selective antagonism of AMPA receptors unmasks kainate receptormediated responses in hippocampal neurons. Neuron 14: 185-189. Patneau, D.K. and Mayer, M.L. (1991). Kinetic analysis of interactions between kainate and AMPA: evidence for acti-

528 vation of a single receptor in mouse hippocampal neurons. Neuron 6: 785-798. Perl, T.M., Bedard, L., Kosatsky, T., Hockin, J.C., Todd, E.C.D. and Remis, R.S. (1990). An outbreak of toxic encephalopathy caused by eating mussels contaminated with domoic acid. N. Engl. J. Med. 322: 1775-1780. Perry, T.L., Krieger, C , Hansen, S. and Tabatabaei, A. (1991). Amyotrophic lateral sclerosis: fasting plasma levels of cysteine and inorganic sulfate are normal, as are brain contents of cysteine. Neurology 41: 487-^90. Plaitakis, A., Berl, S. and Yahr, M. (1984). Neurological disorders associated with deficiency of glutamate dehydrogenase. Ann. Neurol. 15: 144-153. Plaitakis, A. and Caroscio, J.T. (1987). Abnormal glutamate metabolism in amyotrophic lateral sclerosis. Ann. Neurol. 22: 575-579. Plant, T.M., Gay, V.L., Marshall, G.R. and Arslan, M. (1989) Puberty in monkeys is triggered by chemical stimulation of the hypothalamus. Proc. Natl. Acad. Sci. USA 86: 25062510. Pomara, N., Singh, R., Deptula, D., Chou, J.C.-Y., Schwartz, M.B. and LeWitt, P.A. (1992). Glutamate and other CSF amino acids in Alzheimer's disease. Am. J. Psychiatry 149: 251-254. Price, M.T., Ikonomidou, C., Labruyere, J., Izumi, Y. and Olney, J.W. (1992). Neurotoxicity linked to the glutamate metabotropic receptor. Soc. Neurosci. Ahstr. 18: 83. Price, M.T., Olney, J.W., Anglim, M. and Buchsbaum, S., (1979). Reversible action of A^-methyl aspartate on gonadotropin neuroregulation. Brain Res. 176: 165-168. Procter, A.W., Palmer, A.M., Frances, P.T., Lowe, S.L., Neary, D., Murphy, E., Doshi, R. and Bowen, D.M. (1988). Evidence of glutamatergic denervation and possible abnormal metabolism in Alzheimer's disease. J. Neurochem. 50: 790-802. Randall, R.D. and Thayer, S.A. (1992). Glutamate-induced calcium transient triggers delayed calcium overload and neurotoxicity in rat hippocampal neurons. /. Neurosci. 12: 1882-1895. Reynolds, G.P., Pearson, S.J., Halket, J. and Sandler, M. (1988). Brain quinolinic acid in Huntington's disease. /. Neurochem. 50: 1959-1960. Ripps, M.E., Huntley, G.W., Hof, P.R., Morrison, J.H. and Gordon, J.W. (1995). Transgenic mice expressing an altered murine superoxide dismutase gene provide an animal model of amyotrophic lateral sclerosis. Proc. Natl. Acad. Sci. USA 92: 689-693. Roberts, G.W., Allsop, D. and Bruton, C. (1990). The occult aftermath of boxing. J. Neurol. Neurosurg. Psychiatry 53: 373-378. Rosenberg, P.A., Loring, R., Xie, Y., Zaleskas, V. and Aizenman, E. (1991). 2,4,5-Trihydroxyphenylalanine in solution forms a non-N-methyl-D-aspartate glutamatergic agonist and neurotoxin. Proc. Natl. Acad. Sci. USA 88: 4865^869. Ross, S.M., Roy, D.N. and Spencer, P.S. (1989). ^-N-

Excitotoxic neuronal damage and neuropsychiatric

disorders

oxalylamino-L-alanine action on glutamate receptors. J. Neurochem. 53: 710-715. Rothman, S.M. (1984). Synaptic release of excitatory amino acid neurotransmitter mediates anoxic neuronal death, J. Neurosci. 4: 1884-1891. Rothman, S.M. (1985). The neurotoxicity of excitatory amino acids is by passive chloride influx. J. Neurosci. 5: 14831489. Rothstein, J.D., Jin, L., Dykes-Hoberg, M., and Kuncl, R.W. (1993). Chronic inhibition of glutamate uptake produces a model of slow neurotoxicity. Proc. Natl. Acad. Sci. USA 90: 6591-6595. Rothstein, J.D., Martin, L.J. and Kuncl, R.W. (1992). Decreased glutamate transport by the brain and spinal cord in amyotrophic lateral sclerosis. N. Engl. J. Med. 326: 14641468. Rothstein, J.D., Tsai, G., Kuncl, R.W., Clawson, L., Cornblath, D.R., Drachman, D.B., Pestronk, A., Stauch, B.L. and Coyle, J.T. (1990). Abnormal excitatory amino acid metabolism in amyotrophic lateral sclerosis. Ann. Neurol. 28: 18-25. Sandberg, M.J., Butcher, S.P. and Hagberg, H. (1986). Extracellular overflow of neuroactive amino acids during severe insulin-induced hypoglycemia: in vivo dialysis of the rat hippocampus. J. Neurochem. 47: 178-184. Saugstad, J.A., Kinzie, J.M., Mulvihill, E.R., Segerson, T.P. and Westbrook, G.L. (1994). Cloning and expression of the L-2-amino-4-phosphonobutyric acid-sensitive class of metabotropic glutamate receptors. Mol. Pharmacol. 45: 3 6 7 372. Schneider, J.S., Pope, A., Simpson, K., Taggart, J., Smith, M.G. and DiStefano, L. (1992). Recovery from experimental parkinsonism in primates with Gj^j ganglioside treatment. Science 256: 843-846. Schoepp, D., Bockaert, J. and Sladeczek, F. (1990a). Pharmacological and functional characteristics of metabotropic excitatory amino acid receptors. Trends Pharmacol. Sci. 11: 508-515. Schoepp, D.D., Johnson, B.G., Smith, E.C.R. and McQuaid, L.A. (1990b). Stereoselectivity and mode of inhibition of phosphoinositide-coupled excitatory amino acid receptors by 2-amino-3-phosphonopropionic acid. Mol. Pharmacol. 38: 222-228. Schwarcz, R. and Kohler, C. (1983). Differential vulnerability of central neurons of the rat to quinolinic acid. Neurosci. Lett. 38: 84-90. Seawright, A.D., Brown, A.W., Nolan, C.C. and Cavanagh, J.B. (1990). Selective degeneration of cerebellar cortical neurons caused by cycad neurotoxin y3-methylaminoalanine (L-BMAA) in rats. Neuropathol. Appl. Neurobiol. 16: 153169. Simon, R.P., Swan, J.H., Griffiths, T., Meldrum, B.S. (1984). Blockade of A^-methyl-D-aspartate receptors may protect against ischemic damage in the brain. Science 226: 850852. Sims, N.R., Finegan, J.M. and Blass, J.P. (1985). Altered

C.F. Zorumski and J.W. Olney glucose metabolism in fibroblasts from patients with Alzheimer's disease. N. Engl. J. Med. 313: 638-639. Sindou, Ph., Lesort, M., Couratier, Ph., Yardin, C , Esclaire, F. and Hugon, J. (1994). Glutamate increases tau phosphorylation in primary neuronal cultures from fetal rat cerebral cortex. Brain Res. 646: 124-128. Slaughter, M.M. and Miller, R.F. (1981). 2-Amino-4phosphonobutyric acid: a new pharmacological tool for retina research. Science 211: 182-185. Slevin, J.T., Collins, J.F. and Coyle, J.T. (1983). Analogue interactions with the brain receptor labeled by [^H]kainic acid. Brain Res. 265: 169-172. Sofroniew, M.V. and Pearson, R.C.A. (1985). Degeneration of cholinergic neurons in the basal nucleus following kainic or N-methyl-D-aspartic acid application to the cerebral cortex in the rat. Brain Res. 339: 186-190. Sonsalla, P.K., Niklas, W.J. and Heikkila, R.E. (1989). Role for excitatory amino acids in methamphetamine-induced nigrostriatal dopaminergic toxicity. Science 243: 3 9 8 ^ 0 0 . Spencer, P.S., Allen, R.G., Kisby, G.E. and Ludolph, A.C. (1990). Excitotoxic disorders. Science 248: 144. Spencer, P.S., Ludolph, A., Dwivedi, M.P., Roy, D.N., Hugon, J. and Schaumbure, H.H. (1986). Lathyrism: evidence for a role of the neuroexcitatory amino acid BOAA. Lancet 2: 1066-1067. Spencer, P.S., Nunn, P.B., Hugon, J., Ludolph, A . C , Ross, S.M., Roy, D.N. and Robertson, R.C. (1987). Guam amyotrophic lateral sclerosis-Parkinsonism-dementia linked to a plant excitant neurotoxin. Science 237: 517-522. Stanton, P.K., Chattarji, S. and Sejnowski, T.J. (1991). 2Amino-3-phosphonopropionic acid, an inhibitor of glutamate-stimulated phosphoinositide turnover, blocks induction of homosynaptic long-term depression, but not potentiation, in rat hippocampus. Neurosci. Lett. 127: 61-66. Stegink, L.D., Reynolds, W.A., Filer, L.J., Baker, G.L., Daabees, T.T. and Pitkin, R.M. (1979). Comparative metabolism of glutamate in the mouse, monkey and man. In: L.J. Filer, Jr., M.R. Kare, S. Garattini, W.A. Reynolds (Eds.), Glutamic Acid: Advances in Biochemistry and Physiology, Raven Press, New York, pp. 85-102. Stewart, G.R., Zorumski, C.F., Price, M.T. and Olney, J.W. (1990). Domoic acid: a dementia-inducing excitotoxic food poison with kainic acid receptor specificity. Exp. Neurol. 110: 127-138. Storey, E., Hyman, B.T., Jenkins, B., Brouillet, E., Miller, J.M., Rosen, B. and Beal, M.F. (1992). l-Methyl-4-phenylpyridinium produces excitotoxic lesions in rat striatum as a result of impairment of oxidative metabolism. /. Neurochem. 58: 1975-1978. Tamaru, M., Yoineda, Y., Ogita, K., Shimizu, H. and Nagata, Y. (1991). Age-related decreases of the A^-methyl-Daspartate receptor complex in the rat cerebral cortex and hippocampus. Brain Res. 542: 83-90. Tang, C.-M., Dichter, M. and Morad, M. (1989). Quisqualate activates a rapidly inactivating high conductance ionic channel in hippocampal neurons. Science 243: 1474-1477.

529 Tang, C.-M., Shi, Q.-Y., Katchman, A. and Lynch, G. (1991). Modulation of the time course of fast EPSCs and glutamate channel kinetics by aniracetam. Science 254: 288-290. Tarnawa, L, Farkas, S., Berzsenyi, P., Pataki, A. and Andrasi, F. (1989). Electrophysiological studies with a 2,3 benzodiazepine muscle relaxant: GYKI 52466. Eur. J. Pharmacol. 167: 193-199. Tecoma, E.S., Monyer, H., Goldberg, M.P. and Choi, D.W. (1989). Traumatic neuronal injury in vitro is attenuated by NMDA antagonists. Neuron 2: 1541-1545. Teitelbaum, J.S., Zatorre, R.J., Carpenter, S., Gendron, D., Evans, A.C, Gjedde, A. and Cashman, N.R. (1990). Neurologic sequelae of domoic acid intoxication due to the ingestion of contaminated mussels. A^. Engl. J. Med. 322: 1781-1787. Thio, L.L., Clark, D.B., Clifford, D.B. and Zorumski C F . (1992). Wheat germ agglutinin enhances EPSCs in cultured postnatal rat hippocampal neurons by blocking ionotropic quisqualate receptor desensitization. J. Neurophysiol. 68: 1930-1938. Thio, L.L., Clifford, D.B. and Zorumski, C.F. (1991). Characterization of quisqualate receptor desensitization in cultured postnatal rat hippocampal neurons. /. Neurosci. 11: 34303441. Thio, L.L., Clifford, D.B. and Zorumski, C.F. (1993). Blockade of ionotropic quisqualate receptor desensitization in rat hippocampal neurons by wheat germ agglutinin and other lectins. Neuroscience 52: 3 5 ^ 4 . Trombley, P.Q. and Westbrook, G.L. (1992). L-AP4 inhibits calcium currents and synaptic transmission via a G-proteincoupled glutamate receptor. J. Neurosci. 12: 2 0 4 3 2050. Trussell, L.O. and Fischbach, G.D. (1989). Glutamate receptor desensitization and its role in synaptic transmission. Neuron 3:209-218. Turski, L., Bressler, K., Rettig, K.-J., Loschmann, P.-A. and Wachtel, H. (1991). Protection of substantia nigra from MPP"*" neurotoxicity by A^-methyl-D-aspartate antagonists. Nature 349: 4\4-4lS. Urbanski, H.F. and Ojeda, S.R. (1990). A role for N-methylD-aspartate (NMDA) receptors in the control of LH secretion and initiation of female puberty. Endocrinology 126: 1774-1776. Verdoorn, T.A., Burnashev, N., Monyer, H., Seeburg, P.H. and Sakmann, B. (1991). Structural determinants of ion flow through recombinant glutamate receptor channels. Science 252: 1715-1718. Vyklicky, Jr., L., Patneau, D.K. and Mayer, M.L. (1991). Modulation of excitatory synaptic transmission by drugs that reduce desensitization at AMPA/kainate receptors. Neuron 1: 91 \-9H4. Wang, G.J., Labruyere, J., Price, M.T. and Olney, J.W. (1990). Extreme sensitivity of infant animals to glutamate toxicity: role of NMDA receptors. Soc. Neurosci. Abstr. 16: 198. Watkins, J. and Collingridge, G. (1994). Phenylglycine de-

530

rivatives as antagonists of metabotropic glutamate receptors. Trends Pharmacol. Sci. 15: 333-342. Weiss, J.H. and Choi, D.W. (1988). Beta-A^-methylamino-Lalanine neurotoxicity: requirement for bicarbonate as a cofactor. Science 241: 973-975. Weiss, J.H., Hartley, D.M., Koh, J. and Choi, D.W. (1990). The calcium channel blocker nifedipine attenuates slow excitatory amino acid neurotoxicity. Science 247: 14741477. Weiss, J.H., Koh, J.Y. and Choi, D.W. (1989). Neurotoxicity of )3-N-methylamino-L-alanine (BMAA) and ^-Noxalylamino-L-alanine (BOAA) on cultured cortical neurons. Brain Res. 497: 64-71. Westbrook, G.L. and Mayer, M.L. (1987). Micromolar concentrations of Zn^"^ antagonize NMDA and GABA responses of hippocampal neurons. Nature 328: 640-643. Wieloch, T. (1985). Hypoglycemia induced neuronal damage prevented by an A(-methyl-D-aspartate antagonist. Science 230:681-683. Wieloch, T., Engelsen, B., Westerberg, E. and Auer, R. (1985). Lesions of the glutamatergic cortico-striatal projection in the rat ameliorate hypoglycemic brain damage in the striatum. Neurosci. Lett. 58: 25-30. Williams, K., Dawson, V.L., Romano, C, Dichter, M.A. and Molinoff, P.B. (1990). Characterization of poly amines having agonist, antagonist, and inverse agonist effects at the polyamine recognition site of the NMDA receptor. Neuron 5: 199-208. Winder, D.G. and Conn, P.J. (1992). Activation of metabotropic glutamate receptors in the hippocampus increases cyclic AMP accumulation. J. Neurochem. 59: 375-378.

Excitotoxic neuronal damage and neuropsychiatric disorders Wrathall, J.R., Teng, Y.D., Choiniere, D. and Mundt, D.J. (1992). Evidence that local non-NMDA receptors contribute to functional deficits in contusive spinal cord injury. Brain Res. 586: 140-143. Yamada, K.A. and Rothman, S.M. (1992). Diazoxide reversibly blocks glutamate desensitization and prolongs excitatory postsynaptic currents. J. Physiol. (London) 458: 409423. Yamada, K.A. and Tang, C.-M. (1993). Benzothiadiazides inhibit rapid glutamate receptor desensitization and enhance glutamatergic synaptic currents. J. Neurosci. 13: 39043915. Zhang, J., Dawson, V.L., Dawson, T.M. and Snyder, S.H. (1994). Nitric oxide activation of poly(ADP-ribose) synthetase in neurotoxicity. Science 263: 687-689. Zorumski, C.F. and Thio, L.L. (1992) Properties of vertebrate glutamate receptors: calcium mobilization and desensitization. Prog. Neurobiol. 39: 295-336. Zorumski, C.F., Thio, L.L., Clark, G.D. and Clifford, D.B. (1990). Blockade of desensitization augments quisqualate excitotoxicity in hippocampal neurons. Neuron 5: 61-66. Zorumski, C.F., Yamada, K., Price, M.T. and Olney, J.W. (1993). A benzodiazepine recognition site associated with the non-NMDA glutamate receptor. Neuron 10: 61-67. Zorumski, C.F., Yang, J. and Fischbach, G.D. (1989). Calcium-dependent, slow desensitization distinguishes different types of glutamate receptors. Cell. Mol. Neurobiol. 9: 95-104.

Subject index

A-CAM, 48, 60 A7, 100 ACPD receptors {see Metabotropic receptors) Acquired immunodeficiency syndrome (AIDS), 523 ACTH 1-13 (a-melanocyte-stimulating hormone (a-MSH)), 312 ACTH 4-10, 312 ACTH 4-9 analog Org 2766, 313ff ACTH 4-9, 312ff ACTH peptides, 31 Iff binding sites for, 313 similar to MSH binding site, 314 effect on acetylcholine release, 314-315 effect on serotonin release, 319 effect on neuromuscular development, 311,316 effect on dopaminergic neurons, 328 stimulation of axon growth by, 318, 328 stimulation of regeneration by, 315 structure-activity relationships, 314ff Activins, 299 Acute excitotoxicity, 514 Acute neurodegenerative syndromes, 519 Acute psychotic symptoms, 516 Adenoviruses, 104 Adhesion molecule on glia (AMOG), 384 Adhesion molecules, 45ff, 149, 152, 354, 383 Adhesion-inhibitory molecules, 59 Adrenal chromaffin cells, 346 Adult onset olivopontocerebellar degeneration (OPCD), 520 Age, 90, 173, 385, 413, 445, 450, 459, 467, 479 Age-related neuronal degeneration, 315 ALS {see Amyotrophic lateral sclerosis) Aluminum neurotoxicity, 464 Alzheimer's disease, 173, 209, 358, 360, 425, 432, 458, 460, 462, 464-467, 471, 477, 479, 519, 522 antioxidant therapy, 474 BDNF and, 209 biochemical changes, 435 clinical manifestation, 433 pathology, 434 potential neurotrophin therapies, 241, 300, 359 Amantadine, 474 AMPA receptors, 472, 474, 511, 518 {see also Kainate) antagonists at, 513 Amyloid, 434 Amyotrophic lateral sclerosis (ALS), 211, 300, 425, 458, 517-520 biochemical changes, 432 clinical manifestation, 431 pathology, 432 potential neurotrophin therapies, 211, 235, 239, 300, 359 Amyotrophic lateral sclerosis-Parkinsonism-dementia complex of Guam, 438, 517

Angiogenesis, 354 Angiotensin converting enzyme, 174 Annulin, 20 Anterograde trophic signalling, 329 Antioxidants, 445, 461, 466, 474 {see also Free radicals) APB receptors, 514 Aplysia, 24 Apoptosis {see also Programmed cell death. Developmental cell death), 89ff, 354, 456 active process, 93 and mitosis, 110 Bcl-2 in, 98 DNA fragmentation, 105 in ageing, 90 in cell cycle, 108 in insects, 104 induced by T lymphocytes, 90 inhibited by survival (growth) factors, 93 macrophages and, 107 metabolic inhibition and, 93 models of, 91 morphology. 111 signal transduction in, 95 APP, 465 Ascorbate, 443, 446^47, 468, 475 Aspartate-directed proteases, 103 Astrocytes, 72, 182, 280, 299, 346, 353, 355, 357, 360, 377ff, 400, 466 implication in neurotrophic action of neurotransmitters, 381 interaction with neurons, 377 production of neurotrophic factors, 377ff role in axon growth, 383 Astroglia {see Astrocytes) Astrotactin, 384 Auditory system, 346 Autonomic disturbances, 430 Autonomic nervous system origin, 45 Axon fasciculation, 5 genes involved in, 31 inappropriate, 6 Axon guidance cues {see Guidance cues) Axon guidance in invertebrates, 3ff axon repulsion in, 23 calcium in, 26 filipodia adhesion, 24 role of glia, 34 role of muscle targets, 34 Trk receptors in, 26 Axonal regeneration {see Regeneration) B cell lymphoma/leukaemia 2 gene (Bcl-2), 95, 98, 107,109

532 and reactive oxygen species, 101 involvement in cell death, 98 Bad, 100 Bak, 100 Basal laminae, 19 Basement membrane, 68 Bax, 100, 109 Brain derived neurotrophic factor (BDNF), 91, 99, 138, 171, 203ff, 219, 256, 378, 382, 406, 477, 479 effects on a number of neuron populations, 205 expression of mRNA, 208 extremely low abundance, 203 gene, 205 heterodimer with NT-3, 204 high mRNA in the hippocampus, 208 neurodegenerative diseases and, 209ff receptors, 206 retrogradely transported to motor neurons, 204 supports survival of dopaminergic neurons, 204 supports survival of motor neurons, 205 supports survival of sensory neurons, 204 upregulation by seizures or ischaemia, 208 Botulinum toxin, 408, 410 Brevican, 387 c-fos, 97, 177, 182, 189, 274, 302, 353 c-jun, 97, 177, 189, 273, 302 {see also Jun) c-kit, 139 c-myb, 21A c-myc,98, 110, 125,134,274 Cachexia, 281,300 Cadherins (CADs), 383 Caenorhabditis elegans, 3ff, 81, 91, 93, 98, 102, 110 Calcium, 26, 48, 69, 75, 95, 105-106, 315, 352, 360, 454, 456457, 460-461, 468^69, 472, 512, 515, 522-523 and oxidative stress, 456 in regulating axon growth, 26 Calcium/calmodulin-dependent protein kinase type II, 210 Carnitine (3-hydroxy-4-A^-trimethylaminobutyric acid), 478 Carotenoids, 445 Catalase (CAT), 441- 442, 444, 449, 451, 457, 466, 477 Catecholaminergic toxicity and neuromelanin, 458 ced {see Cell death genes) Cell death {ced) genes, 91, 98, 102, 106 Cell adhesion molecules (CAMs) {see Adhesion molecules) Cell surface glycosyltransferases, 76 Ceramide (A^-acyl sphingosine), 96 Cerebellum, 78, 130, 205, 222, 345 Cerebral cortex, 345, 378, 400 Chemoaffinity hypothesis, 149 Chemotropism vs substrate-bound guidance, 8 Chick embryo extract, 252 Chick-quail marker system, 131, 252 Chickling pea, 518 Cholinergic differentiation factor (CDF/LIF), 135, 155, 252, 265ff as a neurotrophic factor, 278 as a repair factor, 279 as defence factor, 280 effects on glia, 279

in bone metabolism, 277 in haemopoiesis, 275 in implantation, 275 in inflanmiatory reactions, 281 in phenotypic specialization, 155, 265ff Chondroitin sulfate-proteoglycan (CS-PG), 49 Ciliary neurotrophic factor (CNTF), 91, 128, 137, 155, 160, 252, 274, 278, 293ff, 379, 406 actions of, 296 and differentiation, 298 as injury factor, 295 in wound repair, 296 receptors, 300 signal transduction, 301 similarity to LIF and IL-6, 300 sources of, 294 structure, 293 synthesis of, 294 Collagens, 47, 55,108 Collagen type I, 52, 55 Collagen type IV, 54-55, 68, 75 Connectin, 34 Copper, uptake and distribution in brain, 462 Cricket, 10 Cross-species comparisons, ix Crustaceans, 4, 10, 13 Cysteine toxicity, 519 Cytochrome P450, 452-453, 470 Cytokines, 178, 273 Cytotactin, 47, 56 D (differentiation) factor, 265 Delayed excitotoxicity, 514 Dementia in Parkinsonian patients, 430, 437 Dementia pugilistica syndrome, 523 L-Deprenyl® (Selegiline®), 457, 474, 476 Detoxifying enzymes, 465 Developmental cell death, 89, 149, 351 {see also Apoptosis) Diabetic neuropathy, 405, 413 Differentiation factors, 25Iff, 298, 351, 353 dopaminergic, 252 muscle-derived, 253, 256 therapeutic manipulation of, 261 Differentiation-inhibitory activity (DIA), 265 Differentiation-retarding factor (DRF), 265 L-3,4-Dihydroxy-phenylalanine (L-DOPA), 427, 437, 459, 473, 476, 521 1,25-Dihydroxyvitamin D, 181 Disintegrins, 75 Dissociative anesthetics, 512, 515 Domoate poisoning, 518 Dopamine role in cognitive disorders, 430 turnover in Parkinsonism, 458 Dopaminergic differentiation factors, 255 Dopaminergic neurons, 378, 521 Dopaminergic nigrostriatal system, 378 {see also Parkinsonism) Dorsal root neurons {see Sensory neurons) Down's syndrome, 435, 445

533 Drosophila, 3, 128, 141 E-cadherin, 48 ElB, 104 Endogenous excitotoxins, 519 Endonucleases, 105 y3-Endorphin, 312 Entactin, 53 Enteric ganglia, 346 Environmental excitotoxins, 516 Epidermal growth factor, 378 receptor, 343 Essential hypertension, 174 Excitatory amino acids, 51 Iff and neurodegeneration, 472 Excitotoxic hypothesis, 514 Excitotoxins acute vs delayed actions, 514 and neurodegenerative illnesses, 516ff cause of neuronal damage, 454, 472, 511 environmental, 516 mechanisms of action, 514 role of calcium in action of, 514 Extracellular matrix (ECM), 20, 46, 67, 107-108, 149, 267, 343, 353,379,384,411 ECM-nerve growth factor interaction, 67 False sago palm {Cycas circinalis), 517 Fas, 92, 94 Fasciclins, 20, 24, 27, 31, 33, 35 Fate map of the neural crest, 131 Ferritin, 445 Fibroblast growth factors (FGFs), 67, 77, 125, 127, 135-137, 159-161, 255, 294, 339ff, 378, 406 angiogenesis and, 353 axon transport of, 347 beneficial effects of exogenous, 358 effect in denervated muscle, 407 effects on glial cells, 353 functions in nervous system, 339, 350 genetics, 339 interaction with other factors, 344, 357 internalization, 344 mechanisms of action, 352 potential therapeutic uses of, 358ff protection from axotomy-induced death by, 359 receptors, 341 regulation of expression, 340, 347, 357 structure, 340 subcellular localization, 346 tissue localization, 344 Fibronectin, 47, 52, 73, 108, 385 Focal adhesion kinase (FAK), 107 Free radical theory of aging, 477 Free radicals, 174, 359, 425ff, 516, 521 and excitotoxins, 454, 472 as causes of neurotoxicity, 452 effect on mitochondria, 468 effects on cell metabolism, 451

generated by dopamine, 458 in Alzheimer's disease, 470 in Parkinsonism, 470 protection against, 445, 465, 474 scavengers of, 442 sites of theraputic interventions against, 452 G-protein, 161, 182, 380, 530 G-Sema, 24 Gangliosides, 479 GAP-13, 163 GAP-43, 408 Gene manipulation, ix, 232-233, 237, 261, 274, 386-387, 411 Glial cells, 17, 280, 299, 350, 353, 377, 458, 463, 468, 470 Glial fibrillary acidic protein (GFAP), 124 Glial progenitors, 126 Gliogenesis, 124, 353 Gliotransmitters, 382 Glucocorticoid hormones, 179 effects on NGF expression, 179 Glutamate toxicity, 359 Glutamate, 315, 357, 427, 432, 436, 454, 456, 472, 511, 514, 516-517, 519-522 dietary, potential hazards of, 517 receptor classification, 511 Glutathione (L-y-glutamyl-L-cysteinyl-glycine; GSH), 441, 443, 445, 447, 448, 450-451, 4 5 6 ^ 5 8 , 465-466, 468, 479 Glutathione peroxidase (GSH-Px), 441-442, 444 Glycosyl-phosphatidylinositol-anchored alkaline phosphatase, 20 GMEM, 56 Gpl20,523 Grasshopper, 3ff Growth factors {see also listings under individual factors) and signal transduction, 93 definition, vi interaction with transcriptional regulators, 140 role in development, 127 Growth hormone, 400 Growth substrates, 11 Growth-associated protein B-50 (GAP-43), 315 Guidance cues, 3ff, 149 ablation of, 22 molecular nature of, 22, 32, 35 on epithelial cells, 19 on trachea, 21 Guidepost cells, 12, 17 H-Sema III, 24 Haemosideroin, 445 Hematopoietic system, 276 Heparan sulfate, 32, 68-69, 343, 386 Heparin, 32, 69, 255, 343, 361 Hepatocyte-stimulating factor III, 267 Hexabrachion, 56 High-affinity 67 kDa laminin-binding protein, 76 Hippocampus, 77, 125, 130-131, 205, 208, 345, 352, 355, 358, 400, 430, 432, 434-435, 454, 468, 473, 514 HO", 438

534 Hox genes, 140 Human placental serum, 252 Huntington's disease, 360, 427, 458, 521 Hyaluronan, 58 Hydrogen peroxide, 440, 438, 441, 443, 451, 457, 459-460, 465467,477 6-Hydroxy-DOPA, 472, 521 6-Hydroxydopamine, 328, 358, 460 5-Hydroxytryptamine {see Serotonin) Hypoglycemia, 454, 520 Hypothalamus, 345 ICAM, 273 ICE: Interleukin \p converting enzyme, 95, 102, 108 IL-1 {see Interleukin-1) Immediate early response genes, 97, 101, 110, 273, 302 {see also c-fos, c-jun) Insulin-like growth factors, 127, 135, 156, 343, 380, 399ff after axotomy, 403 and nerve regeneration, 408, 411 as survival factors, 402, 412 binding proteins for, 401 circulating neurotrophins, 413 distribution in nervous system, 400 effect on neurite outgrowth, 402 gene expression and developmental synaptogenesis, 409 and neuromuscular synapses, 409 feedback inhibition of, 407 neurotrophic role, 404, 413 polyneuronal innervation, role in, 409 receptors, 401 regulation by growth hormone, 400 regulation by muscle, 406, 409 synthesis, 381 Integration of cognitive and motor loops, 426 Integrins, 47, 52, 74, 80, 107, 385 Interaction between growth factors and transcriptional regulators, 140 Interferon gamma, 378 Interleukin (IL)-l, 179, 294, 378, 380 Interleukins, 126, 251, 268, 272, 357 Intestinal ganglioneuromatosis, 174 Invertebrates advantages for neural analysis, vi axon guidance in, 3ff lonotropic receptors, 511 Iron, 460, 464, 466- 467, 474 and neurodegeneration in Parkinson's disease, 463 binding proteins, 445 chelators, 451 distribution in brain, 461 uptake in brain, 462 Ischemia, 208, 354, 454, 475, 516, 520 Jl-200/220, 56 Janus kinase (Jak), 271, 301 Jun-B, 273-274 Jun-D, 273

Kainate,472,518, 521 Kainate/AMPA receptors, 454, 512 L-CAM, 48 LI, 36 Labelled pathways hypothesis, 9, 27 Laminin, 33, 47, 53, 67ff, 137, 384 and neural cell migration, 71, 78, 80 and neural precursor proliferation, 77 at the neuromuscular junction, 72, 78-79 defects in muscular dystrophy, 70 in axon guidance, 81 in cerebellum, 78 in hippocampus, 77 in nerve regeneration, 79 in olfactory ssytem, 78 in retina, 73, 76, 79 isoforms, 70 receptors for, 54 structure, 69 Leech, 3ff Leukemia inhibitory factor (LIF), 135, 137, 155, 252, 265ff, 299 {see also Cholinergic differentiating factor/LIF) and bone cell function, 266 receptors for, 268 structure, 267 Leukotrienes, 455 Lewy bodies, 428, 437, 463 Lipid peroxidation, 257, 440, 442, 444, 447-449, 461, 464, 467, 474-475 Lipofuscin, 450 y3-Lipotropin(y3-LPH),312 Long-term depression, 511, 514 Long-term potentiation, 352, 473, 511, 514 Macrophages, 107, 179, 400 Manganese miners, 470 MAPS, 352, 464 Markers of oxidative stress in neurodegeneration, 467 Mcl-U 100 Melanocortins, 31 Iff {see also ACTH) administration, 315 influences on neuromuscular growth, 317 treatment with, 325 Melanocyte lineage, 133,139 Melanocyte-stimulating hormone, 140 {see also ACTH) Membrane-associated neurotransmitter stimulating factor, 252 Metabotropic receptors, 454, 472, 511,513 l-Methyl-4-phenyl-l,2,3,6-tetrahydropyridine (MPTP), 260, 358, 378, 453, 468, 472, 477-478, 521 A^-Methyl-4-phenylpyridinium ion (MPP+), 210, 260,468,479, 522 ^-N-Methylamino-L-alanine (BMAA), 517 Monoamine oxidase, 443, 457 Monoamine oxidase (MAO)-B, 435, 458, 474 Motor neurons, 4, 91, 346, 381, 402 axon regeneration, 402 development, 311 diseases, 210 survival, 412

535 Mouse sarcoma S 180 cells, 50 Movement disorders, 425 Multiple sclerosis, 463 Multipotentiality, 124, 129 Muscle-derived differentiation factor (MDF), 253, 257 Myc, 92 MyDU6,213 Myelin basic protein (MBP), 134 Myelin, 295, 380 Myeloid differentiation primary response, 273 Myotendinous antigen, 56 Myristoylated alanine-rich C kinase substrate (MARCKS), 163 N-cadherin, 48, 80 N-CAM, 48, 60, 80, 141, 383, 388 N-myc, 134 NADPH, 4 4 1 ^ 4 3 , 449, 452-453, 468 Nerve growth factor (NGF), 8, 80, 91, 94, 96-97, 99, 104, 109, 127, 134, 150, 154, 156, 203, 209, 219, 227, 251, 345, 352, 358,377,401,405 ageing and, 173 Alzheimer's disease and, 173 angiotensin enhances expression, 174 as survival factor, 136 diabetic neuropathies and, 173 essential hypertension and, 174 ganglioneuromatosis and, 174 genetic ataxia and, 173 inCNS, 183 low tissue levels of, 171 maturity vs development, 186 neurofibromatosis and, 174 physiology of, 172 promoter, 177 protection against free radicals by, 174 regulation of expression, 183 rheumatoid arthritis and, 174 role in adulthood, 405 sex differences, 179 therapeutic possibilities, 175 Nerve growth factor receptors, 108, 298, 478 (see also p75, Trk) Nerve growth factor synthesis agents affecting, 178 axotomy and, 186 depolarization and, 183 negative feedback control of, 189 neurotransmitters and, 187 regulation of, 17Iff serum stimulates, 178 sites, 178 steroids and, 178 Netrins, 9 Neural CAD (N-CAD), 383 Neural crest, 78, 80, 123ff, 219, 252, 381 derivatives, 45 development, 45ff, 123ff fate map, 131 migration, 47 mouse sarcoma SI80 cells as analogue, 50

Neural tube, 123 Neurite growth, 163, 328, 383 cessation, 163 Neuritic plaques, 434 Neuroblastoma cells, 402, 411 (see also PC12) Neurocan, 388 Neuroendocrinopathies, 516 Neurofibrillary tangles (NFT), 432, 434, 437, 462, 464 Neurofilaments, 412 Neuroglian, 31, 35 Neurolathyrism, 518 Neuromelanin, 459, 463 Neuromuscular junction, 72, 76, 79, 316, 320 (see also Motor neuron) Neuron-glia cell adhesion molecule (LI or NILE), 383 Neuropsychiatric disorders, 511 Neurotactin, 31, 35 Neurotibromatosis, 174 Neurotrophic factors (see also under individual neurotrophins) control of availability, 159 definition, vii deprivation causes cell death, 149 transport, 138, 149ff types, 161 Neurotrophic theory of cell death, 151 Neurotrophin-3 (NT-3), 91, 99, 136, 138, 171, 203, 219ff, 378, 406 Neurotrophin-4/5, 91, 171, 204, 219ff, 378 Neurotrophin-6, 204, 220 NF (nuclear factor)-IL6, 273 Ng-CAM, 388 Nicotinamide adenine dinucleotide phosphate (see NADPH) Nidogen/entactin, 68 Nitric oxide (NO), 521,523 NMDA receptors, 352, 432, 446, 454, 511, 517, 519523 activation, 315 agonists, 514 and non-NMDA receptors, 520 antagonists, 456, 515-517, 519-520 toxicity, 515 protect against hypoglycemic neurodegeneration, 520 neurotoxicity, 359 subunits, 512 Olfactory system, 78, 400 Oligodendrocytes, 299, 353, 466 Ontogenetic cell death (see Developmental cell death) y3-A^-Oxalylamino-L-alanine (BOAA), 518 Oxidative stress, 101, 210, 425, 438 (see also Free radicals) markers of, 467 p53, 108 p75 receptor, 91, 94, 163, 207, 220 tissue distribution, 223 Parasympathetic ganglia, 45, 346 Parasympathetic lineage, 138 Parkinson's disease, 205, 210, 259, 360, 425, 430, 437, 453, 458, 465-468, 470-471, 476, 517, 519, 521

536 antioxidants in therapy of, 474 biochemical changes in, 428 clinical manifestations of, 427 dementia in, 430, 437 Deprenyl and, 476 in manganese miners, 470 postencephalitic, 429 potential neurotrophin therapies, 205, 360 therapy, 473 PC12 cells, 75, 80, 91, 96, 105, 351, 381, 384, 478 Peripheral nervous system, origin of, 45 Phagocytosis, 106 Phenotype, 68, 251,351 Phenotypic determination, vii, 153, 155, 25Iff, 296 Phosphacan, 388 Phosphorylated NILE, 352 Phosphotidylinositol-linked proteins, 37 Platelet-derived growth factors, 294, 343 Poly-(ADP-ribose) polymerase (PARP), 103 Polyunsaturated fatty acids, 450 Postencephalitic Parkinsonism, 429 Precursor A4 protein (APP), 434 Primary lateral sclerosis, 431 Programmed cell death, 412, 456 {see also Apoptosis, Developmental cell death) defective may underlie psychosis, vi Progressive bulbar palsy, 431 Progressive muscle atrophy, 431 Progressive pseudobulbar palsy, 431 Proopiomelanocortin (POMC), 312 Prostaglandins, 455 Protein kinases, 48, 95 Proteoglycans, 57, 386 Proto-oncogenes {see Immediate early response genes) Reactive oxygen species, 442, 438 {see also Free radicals, Superoxide) redox reactions leading to, 439 Regeneration, 149, 154, 156, 241-242, 295, 311, 321, 402, 404 Retina, 76, 79, 73, 125, 346 Retrophin, 156 Rheumatoid arthritis, 174 S6 protein kinase (PP90rsk), 273 Schwann cell lineage, 354 Schwann cells, 80-81, 134, 179, 181,190, 295, 401, 403^04 Schwannoma, 70 Seizures, 184, 208, 349, 358 Semaphorins, 24, 31,33 Senile plaques, 437, 464 Sensory lineage, 137 Sensory neurons, 4, 8, 45, 91,104,138, 345-346, 381-382, 402, 412 Serotonin, 8, 23, 241, 382, 430, 468 development of neurons utilising, 319 SGF (sweat gland factor), 278 Signal transduction in growth cone navigation, 25 pathways involved in apoptosis, 95

Silverfish, 13 Snail, 8, 23, 26 Somatomedins, 399 SPARC/BM40, 68 Spontaneously hypertensive rat, 174 Steroids and nerve growth factor synthesis, 178 stimulate nerve terminal branching, 318 Streptozotocin-induced diabetes, 173 Substrate-bound guidance, 8 Superoxide dismutase (SOD), 443^44, 449, 451, 465, 477, 479, 521 Superoxide, 438 {see also Free radicals. Reactive oxygen species) production in mitochondria, 453 Symmetrical polyneuropathy, 413 Sympathetic cholinergic differentiating factor, 155 Sympathetic neurons, 45, 91, 98, 109, 346 {see also Neural crest) Sympathoadrenal lineage, 136, 351 Synaptic connectivity, 150, 153 T lymphocytes, 90, 92, 94 tau, 464 Tenascin, 56, 385, 388 Tetrahydro-y3-carbolines (THBC), 469 Tetrahydroisoquinolines (THIQ), 469 Thalamus, 345, 400 Thromboxanes, 455 Thy-1, 352 Thyroid hormone, 181 /w-11,302 a-Tocopheroxyl (TOH), 445^47, 449^50, 466, 468, 475 Toll, 35 Topoisomerase II, 105 Transferrin, 445 Transforming growth factors (TGFs), 48, 129, 344, 357, 381 Transglutaminase, 106 Trk neurotrophin receptors 26, 220, 378 TrkA, 92-93, 207, 220, 378 TrkB, 207, 209, 220, 378 TrkC, 207, 220, 378 Tumor necrosis factors (TNFs), 94, 378, 381 Tyrosine kinase domain, 343 Tyrosine kinase receptors, 162 Tyrosine phosphatases, 31 Up-regulation of IGF gene expression, 403 of NGF synthesis, 183 \-myc, 299 Vasoactive intestinal peptide, 382 Versican, 387 Vitamin E {see a-Tocopherol) Vitronectin, 55, 107 Wallerian degeneration, 404 Woodlouse, 10, 13,15 Wound repair, 296