Advances in Immunology Volume 73

ADVANCES IN Immunology VOLUME 73 This Page Intentionally Left Blank ADVANCES IN Immunology EDITED BY FRANK J. DIX...

Author: Frank J. Dixon

34 downloads 579 Views 4MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

ADVANCES IN

Immunology VOLUME 73

This Page Intentionally Left Blank

ADVANCES IN

Immunology EDITED BY FRANK J. DIXON The Scripps Research Institute La Jolla, California ASSOCIATE EDITORS

Frederick Alt K. Frank Austen Tadamitsu Kishimoto Fritz Melchers Jonathan W. Uhr

VOLUME 73

ACADEMIC PRESS San Diego London Boston New York Sydney Tokyo Toronto

This Page Intentionally Left Blank

CONTENTS

CONTRIBUTORS

ix

Mechanisms of Exogenous Antigen Presentation by MHC Class I Molecules in Vitro and in Vivo: Implications for Generating CD8⫹ T Cell Responses to Infectious Agents, Tumors, Transplants, and Vaccines

JONATHAN W. YEWDELL, CHRISTOPHER C. NORBURY, AND JACK R. BENNINK I. Introduction II. Supporting Information III. Processing and Presentation of Exogenous Antigens: Reviewing the Recent Literature IV. Conclusion: Basic Questions References

1 4 26 63 64

Signal Transduction Pathways That Regulate the Fate of B Lymphocytes

ANDREW CRAXTON, KEVIN OTIPOBY, AIMIN JIANG, AND EDWARD A. CLARK I. II. III. IV. V.

Introduction B Cell Antigen Receptor Complex Coreceptor Regulation of BCR Signaling Regulation of BCR-Induced Responses by CD40 CD95/Fas-Mediated Signaling and BCR-Mediated Resistance to CD95/Fas-Induced Death VI. General Comments and Concluding Remarks References

79 89 106 122 132 134 135

Oral Tolerance: Mechanisms and Therapeutic Applications

ANA FARIA AND HOWARD L. WEINER I. II. III. IV.

Introduction Mechanisms of Oral Tolerance Immune Functions Affected by Oral Tolerance Factors Affecting Oral Tolerance Induction and Maintenance v

153 158 174 178

vi

CONTENTS

V. VI. VII. VIII. IX. X.

Modulation of Oral Tolerance Nasal Tolerance Other Forms of Antigen-Driven Tolerance Treatment of Autoimmune and Inﬂammatory Diseases in Animals Treatment of Autoimmune Diseases in Humans Future Directions References

201 203 206 208 225 232 232

Caspases and Cytokines: Roles in Inﬂammation and Autoimmunity

JOHN C. REED I. II. III. IV.

Introduction The Caspase Family Caspases and Cytokines Conclusions References

265 265 267 287 287

T Cell Dynamics in HIV-1 Infection

DAWN R. CLARK, ROB J. DE BOER, KATJA C. WOLTHERS, AND FRANK MIEDEMA I. II. III. IV. V. VI. VII.

Introduction Normal T Cell Renewal from Progenitors T Cell Renewal from Progenitors in HIV-1 Infection Getting Quantitative on CD4⫹ T Cell Production Measuring Cell Division with the Ki67 mAb What Is the Cause of CD4⫹ T Cell Depletion in HIV-1 Infection? Appendix: Summarizing in Terms of a Mathematical Model References

301 303 305 309 316 320 323 324

Bacterial CpG DNA Activates Immune Cells to Signal Infectious Danger

HERMANN WAGNER I. Introduction II. Bacterial DNA and CpG Motifs: History of Unraveling Immunobiology III. Binding and Cellular Uptake of ODNs IV. Sequence-Independent Effects of the Backbone V. CpG DNA Sequence-Speciﬁc Effects on B Cells VI. CpG DNA Sequence-Speciﬁc Effects on APCs VII. CpG DNA Effects on T Cells VIII. CpG[S]ODN Effects on NK Cells IX. CpG Motifs Affect Plasmid DNA Biology in Gene Vaccination X. Sequence-Speciﬁc Effects of Poly(G) Motifs XI. Immunosuppressive CpG DNA Motifs XII. CpG-ODN-Mediated Signaling

329 331 333 333 334 336 338 339 340 341 342 342

CONTENTS

XIII. XIV. XV. XVI. XVII. XVIII. XIX. XX. XXI. XXII.

Sensing of Pathogen DNA: Evolutionary Vestige of Foreign DNA CpG DNA Acts as Adjuvant for Th1 Responses CpG DNA Mediates Harmful Effects in Vivo CpG DNA Acts as Adjuvant for Antitumor Responses CpG DNA Reverts Th2-Oriented Pathology CpG DNA Acts as Adjuvant for Mucosal Immunity CpG DNA Causes Extramedullary Hematopoiesis CpG DNA Activates Human Immune Cells CpG DNA Mediates Signaling: Stimulation versus Costimulation CpG DNA Allows MHC Class I-Restricted CTL Responses to Exogeneous Antigens XXIII. Concluding Remarks References

vii 346 348 349 349 350 351 351 352 352 353 355 356

Neutrophil-Derived Proteins: Selling Cytokines by the Pound

MARCO ANTONIO CASSATELLA I. II. III. IV. V. VI. VII. VIII.

Introduction General Features of Cytokine Production by Neutrophils Production of Speciﬁc Cytokines by Neutrophils in Vitro Production of Cytokines by Neutrophils Isolated from Individuals Affected by Different Pathologies Modulation of Cytokine Production in Human Neutrophils Molecular Regulation of Cytokine Production in Neutrophils Cytokine Production by Neutrophils in Vivo Concluding Remarks References

369 369 373 426 440 447 453 476 479

Murine Models of Thymic Lymphomas: Premalignant Scenarios Amenable to Prophylactic Therapy

EITAN YEFENOF I. II. III. IV. V. VI. VII. VIII. IX.

Introduction Immunobiology of the Thymus in Relation to Lymphomagenesis Thymic Lymphomas of AKR Mice Prelymphoma Cells in AKR Mice Carcinogen-Induced Lymphomas Thymic Lymphomas Induced by Fractionated Irradiation RadLV-Induced Lymphomagenesis Preventive Therapy of Prelymphoma Mice Concluding Remarks References

INDEX CONTENTS OF RECENT VOLUMES

511 513 514 515 516 518 520 525 530 531 541 557

This Page Intentionally Left Blank

CONTRIBUTORS

Numbers in parentheses indicate the pages on which the authors’ contributions begin.

Jack R. Bennink (1), Laboratory of Viral Diseases, National Institute for Allergy and Infectious Diseases, Bethesda, Maryland Marco Antonio Cassatella (369), Faculty of Medicine, Section of General Pathology, Department of Pathology, University of Verona, Verona, Italy Dawn R. Clark (301), CLB, Sanquin Blood Supply Foundation, Laboratory for Experimental and Clinical Immunology, Department of Clinical Viro-Immunology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Edward A. Clark (79), Departments of Microbiology and Immunology, University of Washington, Seattle, Washington Andrew Craxton (79), Department of Microbiology, University of Washington, Seattle, Washington Rob J. de Boer (301), Department of Theoretical Biology, University of Utrecht, Utrecht, The Netherlands Ana Faria (153), Center for Neurologic Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts Aimin Jiang (79), Department of Microbiology, University of Washington, Seattle, Washington Frank Miedema (301), CLB, Sanquin Blood Supply Foundation, Laboratory for Experimental and Clinical Immunology, Department of Clinical Viro-Immunology and Department of Human Retrovirology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Christopher C. Norbury (1), Laboratory of Viral Diseases, National Institute for Allergy and Infectious Diseases, Bethesda, Maryland Kevin L. Otipoby (79), Department of Immunology, University of Washington, Seattle, Washington John C. Reed (265), The Burnham Institute, La Jolla, California Hermann Wagner (329), Institute of Medical Microbiology, Immunology, and Hygiene, Technical University of Munich, Munich, Germany ix

x

CONTRIBUTORS

Howard L. Weiner (153), Center for Neurologic Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts Katja C. Wolthers (301), CLB, Sanquin Blood Supply Foundation, Laboratory for Experimental and Clinical Immunology, Department of Clinical Viro-Immunology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands Eitan Yefenof (511), The Lautenberg Center for General and Tumor Immunology, Hadassah Medical Center, The Hebrew University, Jerusalem, Israel Jonathan W. Yewdell (1), Laboratory of Viral Diseases, National Institute for Allergy and Infectious Diseases, Bethesda, Maryland

ADVANCES IN IMMUNOLOGY, VOL. 73

Mechanisms of Exogenous Antigen Presentation by MHC Class I Molecules in Vitro and in Vivo: Implications for Generating CD8ⴙ T Cell Responses to Infectious Agents, Tumors, Transplants, and Vaccines JONATHAN W. YEWDELL, CHRISTOPHER C. NORBURY, AND JACK R. BENNINK Laboratory of Viral Diseases, National Institute for Allergy and Infectious Diseases, Bethesda, Maryland 20892

I. Introduction

A. PARADIGM AND PARADOX CD8-expressing T cells (TCD8⫹) recognize major histocompatability complex (MHC) class I molecules bearing peptides, usually of 8 to 10 residues in length. Genes encoding class I molecules are (along with those encoding MHC class II molecules) the most polymorphic known in mammals: each allomorph1 binds a distinct spectrum of peptides based predominantly on the interaction of two or three of the residues with the class I binding groove. Most cell types in the body constitutively express class I molecules, which present thousands of distinct peptides from a diverse array of housekeeping and differentiation-speciﬁc cellular proteins. Under normal circumstances (i.e., in the absence of autoimmune dyscrasias) these peptides fail to activate TCD8⫹ due either to the absence of TCD8⫹ bearing a complementary T cell receptor (TCR) or the silencing of such autoreactive TCD8⫹ in the thymus or periphery. The MHC class I system evolved to identify cells bearing foreign peptides, which, in the natural world, derive almost exclusively from infectious agents, primarily viruses, but also a number of medically important prokaryotic and eukaryotic organisms. The triggering of armed TCD8⫹2 by an antigenpresenting cell (APC) results in the release of a number of cytokines/ enzymes that either destroy the APC or induce modiﬁcations in cellular metabolism that more subtly interfere with the replication of the agent 1 Mammalian genomes contain two or three loci that encode for class I molecules. Alleles encoded by these loci are collectively referred to as allomorphs. Allomorphs can vary considerably not only in the peptides they present, but also in how and where peptides are loaded. Most of the ﬁndings we review describe the behavior of a limited number of allomorphs that are chosen mostly on the basis of convenience. There is little doubt that at least some of the other allomorphs display differences resulting in important functional consequences. 2 Naive TCD8⫹ are small resting cells that do not express signiﬁcant levels of effector molecules: on activation cells begin to divide madly and synthesize pools of cytokines that reside in secretory vesicles awaiting triggering. Such trigger-happy cells are termed ‘‘armed.’’

1

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

2

JONATHAN W. YEWDELL et al.

seeking shelter and sustenance from the APC. The ubiquitous distribution of class I molecules among the ﬁve vertebrate classes provides compelling evidence for the selective advantages conferred by the MHC class I–TCD8⫹ system. The vast majority of peptides constitutively presented by class I molecules derive from proteins synthesized by the cell’s own ribosomes. Polypeptides that provide these peptides are referred to as ‘‘endogenous’’ antigens, whereas polypeptides derived from all other sources are termed ‘‘exogenous’’ antigens. A reasonable argument can be made that the presentation of endogenous antigens evolved to exploit the absolute dependence of viral replication on ribosomal synthesis of viral proteins, because it is clear that TCD8⫹ lysis of virus-infected target cells (except in unusual circumstances) is based on the generation of determinants from endogenous viral proteins. This, and observations that class I-restricted determinants are not generated in vitro following exposure of cells to soluble proteins (in contrast to MHC class II-restricted determinants), contributed to the paradigm that TCD8⫹ surveillance is limited to endogenous antigens, leaving exogenous antigen surveillance to TCD4⫹ cells. Given a surveillance system limited to the presentation of endogenous antigens, what happens if a virus does not happen to infect cells capable of activating naive TCD8⫹? This is more than a theoretical paradox, because the activation of naive T cells requires costimulatory signals that are provided under most circumstances only by a specialized set of APCs that originate from precursors residing in the bone marrow (BM). Such ‘‘professional’’ APCs (pAPCs) consist of dendritic cells (DCs) and macrophage (M␾)/monocytes. Other cells that normally lack costimulatory capacity (basically all other cell types) are termed nonprofessional APCs (nonpAPCs). Although cases in which a selected virus cannot infect pAPCs remain to be carefully documented, it is highly likely that this occurs, given how ﬁnicky viruses can be in their capacity to express their genes in different cell types.3 B. EARLY CLUES TO THE SOLUTION The initial clue for solving this paradox came hard on the heels of Zinkernagel and Doherty’s revelation of MHC restriction (Zinkernagel and Doherty, 1974). Bevan (1976) discovered ‘‘cross-priming’’: mice immunized with cells expressing foreign minor histocompatability (miH) antigens mounted a self-class I-restricted, miH-speciﬁc response even if the cells 3

Sufﬁciently ﬁnicky, in fact, to preclude studying many medically relevant human viruses in tissue culture (e.g., hepatitis B and C viruses, papilloma virus). The blockade in expression can occur at any of the early steps in viral replication: adsorption/penetration into host cells, or transcription/translation of viral genes if penetration occurs.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

3

lacked the restricting class I molecule. Additional studies with radiation chimeras revealed that cross-priming required BM-derived cells bearing self-MHC class I molecules (Bevan, 1977). These ﬁndings were extended to include Y chromosome-encoded antigens (Gordon et al., 1976), and a deﬁned foreign antigen, simian virus 40 tumor antigen (Gooding and Edwards, 1980). There were other contemporary examples of presentation of exogenous antigens that contradicted the paradigm: TCD8⫹ responses were found to be induced by inactivated viruses (Schrader and Edelman, 1977; Wiktor et al., 1977), or even by puriﬁed proteins ( Jay et al., 1978; Tevethia et al., 1980). When the molecular framework of MHC restriction became apparent through the work of Grey, Unanue, and Townsend and their colleagues (Babbitt et al., 1985; Buus et al., 1986; Townsend et al., 1986), Bevan (1987) considered the difﬁculties in stimulating TCD8⫹ responses to viruses unable to infect pAPCs, and posited that cross-priming represented the actions of ‘‘specialized antigen-presenting cells that phagocytose large cellular debris and shuttle the resulting peptide degradation products to their endogenous class I presenting system.’’ Initial mechanistic insight into the presentation of exogenous antigens came from the ﬁnding that the delivery of exogenous proteins to the cytosol enabled their presentation to TCD8⫹. In collaboration with Hosaka, we found that the ability of cells to present proteins from inactivated viruses depended on the fusion of viral and cellular membranes and the consequent delivery of viral proteins to the cytosol (Yewdell et al., 1988). Shortly after, Bevan and colleagues showed that introduction of chicken ovalbumin (OVA) to the cytosol via osmotic lysis enabled its presentation to TCD8⫹ (Moore et al., 1988) and that splenocytes exposed to OVA in vitro were able to induced OVA-speciﬁc TCD8⫹ (Carbone and Bevan, 1990). With the demonstration by Rock et al. (1990) that a subset of splenocytes were able to process and present exogenous antigens, the study of exogenous antigen processing and presentation was established as an important area in immunobiology. C. AIMS OF THIS REVIEW Seven years ago it was a relatively simple task to review existing knowledge of exogenous antigen presentation (Yewdell and Bennink, 1992).4 In the intervening years there has been steady progress in understanding the immunogenicity and antigenicity of exogenous proteins. Advances have come on many fronts, including the cell biology of antigen processing, DC 4 Particularly since we missed papers. We apologize for sins past and present, and hope that aggrieved parties understand that even with improvements in online literature searches, it is difﬁcult to ﬁnd all of the relevant studies.

4

JONATHAN W. YEWDELL et al.

biology, the introduction of novel carriers (including molecular chaperones) for enhancing TCD8⫹ responses to exogenous antigens, and the mechanisms of in vivo presentation. Concurrently, there has been a greatly increased interest in creating vaccines that elicit TCD8⫹ responses to tumors and infectious diseases that resist intervention by drugs or alternative effector arms of the immune system. Our principal goal in writing this review is to foster research in this important area of cellular immunology. More speciﬁcally, we aim also to accomplish the following goals: 1. Discuss the strengths, and especially the weaknesses, of methods and strategies used to investigate exogenous antigen processing. 2. Review the aspects of cell biology most intimately involved in exogenous antigen processing. 3. Critically review recent basic ﬁndings in the processing and presentation of exogenous antigens in association with classical class I molecules.5 4. Indicate the most important gaps in knowledge, particularly those that can be effectively addressed with the available technology. II. Supporting Information

A. GENERATION OF MHC CLASS I–PEPTIDE COMPLEXES FROM ENDOGENOUS ANTIGENS Exploration of the pathways of exogenous antigen processing and presentation requires a thorough understanding of how endogenous antigens are processed and presented to TCD8⫹. This follows from two considerations. First, for a number of exogenous antigens the two pathways intersect. Second, uncovering processing pathways devoted to exogenous antigens is usually predicated on initial ﬁndings disfavoring the functioning of the classical pathway. Third, class I molecules used for nonclassical presentation are derived or diverted from the endogenous processing pathway. Although many interesting and important details of endogenous antigen processing remain to be explored, the major features have been reasonably well established.6 In this section we provide an overview of the endogenous 5 As opposed to nonclassical class I molecules, which are highly similar in structure to classical class I molecules, but nonpolymorphic and, in some cases, encoded outside the MHC. Nonclassical class I molecules can present peptides, deriviatized peptides, glycolipids, and probably other oligo substances to TCD8⫹. Although many nonclassical class I molecules present antigens derived from exogenous antigens, this is beyond the scope the present review, and we refer readers to several other reviews (Melian et al., 1996; Lindahl et al., 1997; Porcelli et al., 1998; Brossay et al., 1998). 6 The most compelling evidence being that lengthy reviews are now devoted to subtopics in the endogenous antigen-processing pathway.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

5

processing pathway (Fig. 1), referring readers to more comprehensive reviews (York and Rock, 1996; Tanaka et al., 1997; Elliott, 1997; Pamer and Cresswell, 1998; Momburg and Hammerling, 1998), and providing

FIG. 1. The endogenous class I processing pathway: the standard model. This depicts the typical process (described in the text) by which a determinant (grey oval) in a cytosolic antigen comes to be displayed on the plasma membrane in association with a class I molecule. Manipulations that block the pathway are indicated in red, and the locations of various blockades are indicated by short, double lines. Enzymatically active proteasomes are often necessary for the liberation of the determinant [or a precursor that can be further trimmed by additional cytosolic or endoplasmic reticulum (ER) proteases] as determined experimentally by inhibition of presentation using proteasome inhibitors [of which lactacystin and the nitrophenol derivative of carboxybenzyl-leucyl-leucyl-leucine vinyl sulfor (NLVS) are most speciﬁc]. The identities of proteases involved in the production of proteasome inhibitorinsensitive determinants are unknown, as indicated by the black box. Upward of 90% of cytosolic peptides gain access to the ER via an ATP-dependent transporter (TAP), but TAPindependent transport also occurs, often (but not always) via the translocon. The translocon is the pore by which polypeptides are translocated into the ER. Peptides in the ER can bind either to class I molecules or to molecular chaperones. Peptide binding to chaperones has been demonstrated for TAP-transported peptides, and probably also occurs for translocon-transported peptides. Peptide trimming can occur in the ER, particularly residues extending from the NH2 terminus. Transport from the early secretory pathway (ER and ER–Golgi complex intermediate compartment) is blocked by brefeldin A (BFA). The secretory pathway includes the stacks of the Golgi complex and post-Golgi complex secretory vesicles. LLL, Carbobenzoxy-leucinyl-leucinyl-leucinal; LLnL, N-acetyl-L-leucyl-L-leucylL-norleuci.

6

JONATHAN W. YEWDELL et al.

direct references only for the most recent ﬁndings or more controversial issues. Endogenous determinants are (by deﬁnition) produced from products of the ribosomes of the APCs. The question of how endogenous proteins are targeted for peptide generation has important implications for the generation of peptides from exogenous proteins delivered to the cytosol. This is probably the most poorly characterized portion of the class I antigenprocessing pathway. It is uncertain to what extent peptides are derived from native proteins versus defective forms that never achieve a native state (Yewdell et al., 1996). Given the important role that cytosolic molecular chaperones play in the folding of nascent proteins and the disposal of denatured proteins (Hartl, 1996), it is likely that chaperones are heavily involved in targeting defective nascent proteins to the class I processing pathway. Chaperones may also be involved in the targeting of native or once-native proteins for peptide generation. There is much to learn in this area; the evidence for the involvement of chaperones in antigen processing is scanty and indirect in nature (Wells et al., 1998). The major cytosolic protease involved in the disposal of damaged or unwanted proteins is the proteasome, an abundant, complex macromolecular structure with multiple proteolytic activities. Proteasomes are responsible for the degradation of polyubiquitinated proteins, and also degrade nonubiquitinated proteins targeted via poorly deﬁned signals. The role of polyubiquitination in the generation of class I ligands is uncertain (Michalek et al., 1993; Cox et al., 1995), although there are clearly substrates that are processed in a ubiquitin-independent manner (Yellen-Shaw et al., 1997). Proteasomes are involved in the production of many class Iassociated peptides, as most directly demonstrated by the use of proteasome inhibitors (Rock et al., 1994). A fairly high percentage of determinants and ligands (perhaps even greater than 50% for some class I allomorphs) are generated in the presence of proteasome inhibitors, suggesting the presence of other cytosolic proteases active in antigen processing (Cerundolo et al., 1997; Vinitsky et al., 1997; Yellen-Shaw et al., 1997; Luckey et al. 1998; Anto´n et al. 1998). Whether these putative proteases are related to the proteolytic system that replaces the cell-keeping functions of proteasomes in cells adapted to grow in the presence of a proteasome inhibitor (Glas et al., 1998) is uncertain. Cytosolic peptides are separated from nascent class I molecules by the membrane of the endoplasmic reticulum (ER). Most class I ligands are transported from the cytosol through the action of two MHC-encoded gene products (TAP1 and TAP2) that associated to form an ATP-dependent transporter (TAP) localized in the ER. TAP imposes two major constraints on peptides for their efﬁcient transport into the ER: ﬁrst, they must be

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

7

between 8 and 16 residues in length, and second, they must have the proper COOH-terminal residue, which varies among TAPs derived from different species.7 TAP is not the sole ER portal for endogenous peptides: cells lacking TAP are able to load a variable fraction of class I molecules, ranging from 앑5–10 to 50%, depending on the allomorph. A major route for these peptides is the translocon, the abundant proteinaceous pore used to export proteins destined for the secretory pathway, cell surface, or beyond. Some of the most abundant peptides recovered from class I molecules derive from signal sequences of such translocon-exported proteins. These peptides are highly hydrophobic, and allomorphs such as HLA-A*0201, which binds hydrophobic peptides with high afﬁnity, are the least affected by knocking out TAP function. A generally efﬁcient method of loading class I molecules with a given peptide in TAP-expressing or -deﬁcient cells is to place the peptide immediately COOH-terminal to a signal sequence (Anderson et al., 1991). When expressed by vaccinia virus under the control of a standard promoter, virtually all nascent class I molecules can be loaded with such ER-targeted peptides (Porgador et al., 1997; Anto´n et al., 1997). Nonhydrophobic peptides from cytosolic or nuclear proteins can also be presented to TCD8⫹ in a TAP-independent manner, but far less efﬁciently, and it is uncertain whether these peptides access the ER by the translocon, another proteinaceous pore or transporter, or diffusion. Two observations demonstrate the involvement of ER proteases in the trimming of translocon-transported substrates. First, signal sequencederived class I ligands recovered from TAP-deﬁcient cells are obviously processed by some ER protease from the signal sequence (Hughes et al., 1996). Second, peptides can be liberated from other regions of ER-targeted proteins, particularly the COOH terminus (Snyder et al., 1994, 1998; Elliott et al., 1995). Two additional ﬁndings imply that ER proteases are involved in the biogenesis of a portion of TAP-transported peptides. First, TAP transports peptides of up to 16 or so residues with no apparent loss of efﬁciency, yet only a low percentage of peptides recovered from class I molecules are more than 10 residues in length. Second, a number of deﬁned determinants produced as cytosolic minimal minigene products8 are not antigenic, probably due to proteasomal destruction (Fu et al., 1998), 7

Mouse TAP requires a hydrophobic residue, whereas human TAP requires hydrophobic or positively charged residues. These preferences match those exhibited by mouse versus human class I allomorphs; in rats, however, mismatches are known to occur with interesting consequences well worth reading (Powis et al., 1996). 8 Minimal refers to the peptide thought to represent the determinant naturally processed from its normal precursor; usually such ‘‘minimal’’ peptides are synthesized with an NH2terminal Met to initiate translation. The Met is frequently removed (Buchholz et al., 1995), probably by Met-aminopeptidase, which is highly active in the cytosol.

8

JONATHAN W. YEWDELL et al.

implying that the processing of the natural protein entails the generation of an extended precursor translocated by TAP and trimmed in the ER. The creation of peptide–class I complexes9 in the ER is an intricately choreographed affair. The association of heavy chains with ER chaperones begins as nascent heavy chains are threaded through the translocon and continues at least until class I molecules are exported to the Golgi complex. A partial list of chaperones involved includes all-purpose gene products (calnexin, calreticulin, p60) that facilitate folding of numerous glycoproteins, and at least one gene product (tapasin) (Ortmann et al., 1997; Li et al., 1997) that seems to be dedicated to class I assembly.10 Association of heavy chains with 웁2-microglobulin (웁2m) induces conformational alterations that enable high-afﬁnity binding of peptides to the heavy-chain groove. Two pathways of peptide loading can be distinguished: binding to class I molecules tethered to TAP via tapasin, and binding to class I molecules that are not TAP associated. The latter point is inferred from the ﬁnding that tapasin-deﬁcient cells demonstrate only a variable degree of impaired assembly of class I molecules, ranging from nearly complete to undetectable, depending on the allomorph (Grandea et al., 1997; Peh et al., 1998). The potential for peptides to associate with TAP-free class I molecules is also clearly indicated by the association of peptides derived from translocon-transported proteins with class I molecules in TAPdeﬁcient cells. ER chaperones may be directly involved in loading peptides onto class I molecules by either pathway (Srivastava, 1993). TAP-transported peptides can be recovered from numerous ER chaperones (Lammert et al., 1997; Spee and Neefjes, 1997). It is plausible that chaperones function to protect peptides from proteolysis, or act in a concerted manner to deliver antigenic peptides to the binding groove in the proper conformation. The binding of peptide ligands to TAP-associated class I molecules results in their release from TAP and their transport to the cell surface. Ligand-induced release of non-TAP-associated molecules is likely but not well established (Lewis and Elliott, 1998). Once on the cell surface, most complexes are stable for many hours at 37⬚C, where they can be perused by TCD8⫹. Exogenous proteins also have access to the ER. Certain bacterial and plant toxins are transported in a retrograde manner through the Golgi complex to the ER, where they gain entry to the cytosol11 (Lord and 9

From this point on, complex refers to peptide–class I complex unless modiﬁed. General-purpose ER chaperones recognize unfolded proteins by either binding to certain oligosaccharides or to short linear hydrophobic sequences that occupy the interior regions of native proteins (this latter property is shared with cytosolic chaperones). 11 This is accomplished by at least two mechanisms: ﬁrst, possession of the COOH-terminal sequence used for retrieval of ER proteins that have strayed, or second, by hitching a ride with stray cellular ER proteins on their way home. Once in the ER, the toxins are unfolded, and may enter the cytosol via the pathway involved in degradation of ER proteins. 10

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

9

Roberts, 1998). We have shown that class I-binding peptides have access to the ER, probably via a direct vesicular route (Day et al., 1997). Whether large polypeptides have access via this route is uncertain. Once in the ER, substrates can potentially be trimmed in the same manner as peptides entering via TAP or the translocon. Given the large number of peptidereceptive (PR)12 class I molecules on the cell surface, this pathway is not likely to be physiologically signiﬁcant for peptides that bind to class I molecules with high afﬁnity. It may play a role, however, in the generation of peptides that cannot be created through the actions of cell surface or serum proteases. B. EXTRA-ER LOADING OF CLASS I MOLECULES The creation of complexes in extra-ER sites is likely to play a crucial role in the presentation of many exogenous antigens (Fig. 2). Class I molecules exported from the ER do not uniformly retain their ligands: a substantial subset lose their ligands within 30 min of their departure. These molecules may have been loaded with a low-afﬁnity ligand or may have suboptimally bound a peptide that normally binds with higher afﬁnity. Ligand loss results in the generation of class I molecules capable of binding a new peptide. Dissociation may occur mostly within the secretory pathway, because peptide-receptive class I molecules can be easily detected in the Golgi complex immunocytochemically or biochemically (Day et al., 1995). Once on the plasma membrane, class I molecules lacking high-afﬁnity peptides denature with a t1/2 of 앑5 min at 37⬚C in the absence of 웁2m in the surrounding medium (Day et al., 1995). The presence of 웁2m at physiological concentrations increases the t1/2 of heavy chains by at least an order of magnitude. This has an important consequence for the presentation of exogenous antigens, because it is mostly these 웁2m-stabilized cell surface class I molecules that are capable of binding exogenously added peptides (Rock et al., 1991).13 In addition to binding extracellular processed ligands, cell surface class I molecules could in principle associate with endosomally processed peptides. Such endosomally processed antigens could also bind to class I molecules in the endosomes (as described below, both of these processes are thought to occur). A number of class I molecules have been reported to bind peptides at pH values characteristic of endosomes (Stryhn et al., 12

Class I molecules capable of binding exogenous peptides were originally referred to as ‘‘empty,’’ a state that is impossible to establish experimentally. PR emphasizes the operational nature of the deﬁnition. 13 The caveat regarding potential differences in allomorphs made in footnote 1 is particularly relevant here, since these conclusions are based largely on ﬁndings with a very limited number of mouse allomorphs.

10

JONATHAN W. YEWDELL et al.

FIG. 2. Class I trafﬁcking: deﬁned and possible routes. In addition to the well-deﬁned route of class I transport from ER to plasma membrane as depicted in Fig. 1, here the dashed lines depict less well-characterized routes of trafﬁcking to the endosomal/lysosomal compartment. Peptide-receptive (PR) class I molecules are depicted with a R in the binding groove in place of a tightly bound ligand (oval). In addition, the invariant chain (Ii), which may function in some cases to deliver class I molecules to endosomes, is depicted as sitting in the binding groove, although the basis for the interaction of Ii with class I molecules remains to be ascertained. Signiﬁcant amounts of PR class I molecules have been detected in the Golgi complex and cell surface, but their presence in endosomes is less well established. Transport of class I molecules from the Golgi complex can be blocked by monensin (at least in some cells). Brefeldin A (BFA) has complicated effects on endosomes that can result in the inhibition of recycling molecules.

1996),14 but for this to occur, of course, class I molecules must be delivered to endosomes. The delivery of class I molecules to endosomal compartments is poorly characterized, particularly in pAPCs, of which virtually nothing is known. There are only a few studies of the internalization of class I molecules in the absence of antibody cross-linking (which induces ligand internalization). Class I molecules were found to recycle in mouse T cells, but not B cells or a number of permanent cell lines (Hochman et al., 1991). Using a class I-binding peptide conjugated to an oligosaccharide that enables detection with a monoclonal antibody speciﬁc for the oligosaccharide, Motal et al., (1993) reported that a T cell line spontaneously internalized the complex and could reexpress the complex on the cell surface following pronase removal of noninternalized complexes. It was 14

Indeed, the ﬁndings suggested that conditions were near optimal for peptide exchange.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

11

not rigorously excluded, however, that reexpression of complexes was due to the release of the free glycopeptide and its rebinding to cell surface class I molecules. The endosomal trafﬁcking of class I molecules in human cells is also poorly characterized. Class I molecules were found to be spontaneously internalized in human B lymphoblastoid cells and rapidly returned to the cell surface, resulting in low steady-state levels in endosomes (Reid and Watts, 1990). Although it was originally reported that MHC class IIcontaining endosomal compartments in B lymphoblastoid cells were devoid of class I molecules as detected by immunoelectron microscopy (EM) (Neefjes et al., 1990), it was found that this is a peculiarity of the cell line used (M. J. Kleijmeer, personal communication). The functional signiﬁcance of the endosomal class I molecules detected is uncertain, however, particularly because the antibody used for detection binds to unfolded heavy chains, raising the possibility that all of the class I molecules detected were denatured and simply awaiting destruction. On the other hand, class I molecules were also detected in exosomes in this study. Exosomes are small vesicles generated in the lumen of large endosomes by a pinching-off process, with the topological consequence of orienting lumenal domains of integral membrane proteins (like class I molecules) on the outer surface of the exosome. Following the fusion of multivesicular endosomes with the plasma membrane, exosomes are released by cells, providing another route for complexes assembled in endosomes to access TCD8⫹. Originally described in reticulocytes, exosomes are also known to be produced by B cells and DCs. It was initially demonstrated that exosomes are rich in MHC class II and costimulatory molecules, and are able to able to stimulate TCD4⫹ (Raposo et al., 1996). Zitvogel et al. (1998) reported that exosomes can elicit TCD8⫹-mediated antitumor immunity when derived from DCs pulsed with a deﬁned tumor rejection peptide. It remains to be established how class I molecules arrive in the exosome and the extent to which these class I molecules can be loaded with synthetic or naturally processed peptides in the ER, on the cell surface, or within the endosomal system. Several class I allomorphs have been reported to associate with invariant chain15, which possesses intrinsic endosomal localization signals and could potentially direct class I molecules to compartments associated with class II antigen processing (Cerundolo et al., 1992). The expression of invariant 15 The major function of the invariant chain is to promote the biogenesis of MHC class II molecules. By binding to the class II binding groove, it prevents the association of class II molecules with polypeptides in the ER and it also helps route class II molecules to the endosomal compartments, the major site of processing exogenous antigens for association with class II molecules.

12

JONATHAN W. YEWDELL et al.

chain in HeLa cells directs HLA-B27 to endosomal compartments (Sugita and Brenner, 1995). The functional consequences of this routing have not been investigated, and there is mixed evidence regarding the role of invariant chain in peptide loading. Vigna et al. (1996) reported that invariant chain association with HLA allomorphs correlates positively with cell surface expression of complexes. This effect may be strictly exerted in the ER, however. Moreover, splenocytes from ‘‘knockout’’ mice with targeted disruption (⫺/⫺) of genes encoding invariant chain and TAP express assembled cell surface class I molecules and free heavy chains in amounts similar to those in mice with the TAP disruption alone (Tourne et al., 1996), arguing against a major effect of invariant chains in peptide loading of class I molecules, at least among splenocytes.16 C. MODES OF INTERNALIZATION OF EXOGENOUS ANTIGENS Cells may process extracellular antigens in two general locations: the cell surface or intracellularly. The cell surface is capable of only limited processing, probably restricted to proteolysis by proteases produced by the cell or acquired from the surrounding plasma. For proteins or more complex antigens to be disassembled and unfolded, internalization is required. A few proteins and viruses are able to penetrate the plasma membrane, reaching the cytosol (and the endogenous processing pathway). All other antigens, however, must be taken into the endosomal system. This can proceed by a number of distinct mechanisms (Fig. 3). 1. Clathrin-mediated endocytosis/nocytosis (reviewed in Watts and Marsh, 1992; Schmid, 1997). This entails the formation of small vesicles that are coated by cytosolic proteins, predominantly clathrin. Clathrin interacts with the cytosolic domains of cell surface proteins via adaptor proteins (sundry adaptins), which generally recognize simple motifs (including YXRF or LL). This results in clustering of receptors in small invaginations termed ‘‘coated pits.’’ After a size threshold is reached, pinching results in the formation of ⬍150-nm-diameter vesicles that rapidly release clathrin for recycling. In addition to receptors within the coated pits, the ﬂuid in the immediate vicinity is internalized, which is detected as ﬂuid-phase pinocytosis.17 Vesicles fuse with early endosomes, where the fate of the receptor is determined; some immediately recycle to the cell surface, others proceed further into the endocytic pathway, where they may still be recycled, even after trafﬁcking as far down the pathway as the lysosome. In general, conditions progressively favor degradation (decreased 16

It will be of interest, of course, to study various types of pAPCs from invariant chain ⫺/⫺ mice, and also to compare with normal mice the ability to respond to various exogenous antigens. 17 Although endocytosis and pinocytosis can be used interchangeably, we preserve the original deﬁnition of pinocytosis (cell ‘‘drinking’’) and reserve endocytosis for the delivery of plasma membrane-associated substances.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

13

FIG. 3. Exogenous antigen uptake. Exogenous antigens can be processed either extracellularly or following internalization. As in the Figs. 1 and 2, determinants are represented as an oval. Shown are the various modes by which antigens may be internalized. Endocytosis and pinocytosis represent the same processes. In endocytosis, antigen is bound to a membrane receptor (greatly increasing the efﬁciency by concentrating the antigenic ligand); in pinocytosis, antigen is swept along with the ﬂuid present in the vicinity of the budding endosome. As described in the text, endocytosis–pinocytosis can occur via coated pits or via noncoated pits or caveolae. To reﬂect these two possibilities, nascent endosomes are depicted as half-coated, half-uncoated. Phagocytosis refers to the internalization of particulate matter, be it aggregated proteins, beads, or, as depicted here, bacteria (symbolized by a carefree bug). Macropinocytosis refers to the internalization of large amounts of surrounding solvent, a process that immature DCs are particularly adept at. Drugs that block the various processes of internalization are indicated with double, short lines. Internalized antigens may be processed in the endosome/lysosome compartment, where ligands bind PR class I molecules, whose delivery to the compartment is summarized in Fig. 2. Processing can be interfered with by lysosomotropic agents such as NH4Cl (which raises the pH of acid compartments in cells), proton pump inhibitors such as baﬁlomycin (which has the same effect), or by broadly acting protease inhibitors such as leupeptin (which blocks cysteine proteases). Alternatively, internalized antigens can gain access to the cytosol (indicated by three short arrows), where they are processed via the classical pathway.

pH, increased redox potential, more proteolytic) as substances trafﬁc toward lysosomes. 2. Non-clathrin-mediated endocytosis. This consists of at least two distinct processes that together account for a large share of overall ﬂuidphase endocytosis.18 One is similar to clathrin-mediated endocytosis except 18 This conclusion is based on the rate of pinocytosis in cells in which coated-pit internalization is disrupted. It is possible that this perturbation enhances the activity of the alternative pathways, resulting in an overestimation of their function in nonperturbed cells.

14

JONATHAN W. YEWDELL et al.

that vesicles have no easily detected coat and the types of proteins recruited into vesicles are uncertain (reviewed in Sandvig and van Deurs, 1994; Bishop, 1998). The other entails the formation of small vesicles (50– 80 nm) from pits at regions of the membrane rich in a protein known as caveolin (the pits are termed ‘‘caveolae’’) (reviewed in Anderson, 1993). Caveolae possess a distinct array of lipids that recruit proteins anchored by glycophosphoinositol (or vice versa!). Caveolin is believed to be transported from the plasma membrane to the ER, but it is uncertain whether this reﬂects fusion of caveolae with the ER or the redistribution of the dissociated caveolin. 3. Macropinocytosis (reviewed in Swanson and Watts, 1995). Very large pinosomes (200–500 nm in diameter) form at the sites of membrane rufﬂing19 —trapping large amounts of the surrounding media. The fate of macropinosomes varies according to cell type: in some cells they intersect the endosomal pathway either in early endosomes or in tubular lysosomes; in other cells they appear to fuse only with each other. 4. Phagocytosis. This is the engulfment of large particulate matter by cells, including unicellular organisms, tissue debris, and man-made particles.20 It is thought to be receptor mediated but clathrin independent. Receptors believed to induce phagocytosis include the mannose receptor (which recognizes oligosaccharides on the surface of bacteria), Fc21 receptors [which enable the internalization of immunoglobulin (Ig)-coated particles], and receptors able to recognize the plasma membrane of apoptotic cells (reviewed in Savil et al., 1993). The conventional form of phagocytosis involves tight ‘‘zippering’’ of receptors around the phagocytosed particle, resulting in little uptake of solute during the internalization. In the past few years another form of phagocytosis has been described, whereby bacteria release protein messengers into the cytosol of targeted cells, causing cell surface rufﬂing and formation of macropinosome-like membrane invaginations. This forms so-called spacious phagosomes, wherein large amounts of external solute are internalized along with the bacterium (Swanson and Bauer, 1995). D. APCS First, some terminology. The crucial feature of pAPCs is that they serve as afferent APCs (afAPCs), i.e., they activate naive or memory TCD8⫹. Activated TCD8⫹ recognize efferent APCs (efAPCs) (most often, but not 19 Rufﬂing refers to the ﬂickering displacement of large sheets of plasma membrane at the leading edge of cells moving across a solid surface. 20 Phagocytosis is the original phenomenon noted by Metchnikoff on inserting foreign bodies into starﬁsh larvae—the ﬁrst recorded observation of cellular immunity. 21 Fc refers to the constant domain of immunoglobulin.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

15

always non-pAPCs), in the case of microbial antigens with the goal of killing or modifying the APC to impede replication of the intracellular parasite. The efﬁcacy of noninfectious vaccine preparations is predicated on the capacity of the antigen to generate effective pAPCs (or enable nonpAPCs to act as afAPCs). The pAPCs are composed of three cell types, B cells, M␾s, and DCs. B cells play a central role in presenting antigens to TCD4⫹, but their role as afAPCs in TCD8⫹ responses is probably limited, and they are discussed in this review only as efAPCs. Armed with these deﬁnitions, we plunge into thumbnail descriptions of DCs and M␾s (detailed information can be found in the reference at the start of each description). 1. M␾s M␾s (reviewed in Gordon, 1998) are ubiquitous sentries, present in all tissues in the absence of overt immune responses. They derive from blood monocytes, which circulate in the blood for approximately 1 day after departing the BM. Large numbers of monocytes are recruited to the sites of local inﬂammation. In the absence of inﬂammation, M␾s are actively phagocytic and endocytic but express low levels of costimulatory accessory molecules. Exposure of M␾s to any of a number of inﬂammatory cytokines triggers modiﬁcations in the cells, including an increase in endocytic activity, expression of costimulatory molecules, secretion of proinﬂammatory cytokines, and generation of reactive compounds capable of destroying phagocytosed and extracellular microorganisms. The act of phagocytosis appears to up-regulate costimulatory molecules (De Bruijn et al., 1996), possibly enabling presentation prior to the establishment of an inﬂammation. 2. DCs Unlike M␾s, which play important direct roles in establishing an inﬂammation and destroying invaders, it is believed that DCs (reviewed in Steinmen, 1998) function primarily as pAPCs. DCs are so-named due to their highly branched structure: DCs in skin (originally described as Langerhans cells) and the trachea (Schon-Hegrad et al., 1991) are sufﬁciently abundant (앑5% of cells in the epidermis) to form a contiguous network, and probably do something similar in a number of other mucosal locations. DCs are present in other tissues, and are probably widely distributed outside of the central nervous system. Precursors are present in BM, blood, and spleen, which is of practical importance, because large numbers of DCs can be obtained by culturing cells from human blood or mouse spleen or BM in the presence of the proper cytokines. DCs derive from BM precursors of both lymphoid and myeloid lineages. The two subsets

16

JONATHAN W. YEWDELL et al.

of cells display distinct surface phenotypes and have been postulated to play different roles in induction and regulation of an immune response (Galy et al., 1995; Winkel et al., 1997; Fazekas de St. Groth, 1998). Peripheral DCs (i.e., DCs present outside of lymphoid organs) express low amounts of class II molecules and costimulatory molecules (CD40, CD54, and CD80) and, at least in vitro, actively endocytose and phagocytose via receptors for Fc, complement,22 and oligosaccharides (such as the mannose receptor). The phagocytic and endocytic capacities of DCs in vivo are difﬁcult to study, although Matsuno et al. (1996) reported uptake of particulates from blood by DCs lining the hepatic sinusoids. Peripheral DCs are considered pre-pAPCs, graduating on activation. In vivo this probably occurs on exposure to proinﬂammatory cytokines such as interleukin 1웁 (IL-1웁) or tumor necrosis factor 움 (TNF-움), but in vitro, simply exposing cells to adherent plastic [or to bacterial lipopolysaccharide (LPS)] triggers activation. Activated cells are much less endocytic and express high levels of MHC class I and class II molecules as well as costimulatory molecules in vitro. The phenotype of activation in vitro is correlated with migration of the cells from the periphery to lymphoid organs, where they express high levels of MHC and costimulatory molecules. Based on these observations, the simple model (much in vogue) is that DCs in the periphery function efﬁciently to acquire and process antigen and once this is done they cease processing antigen and go about presenting it in the optimal anatomic location and manner. 3. The in Vitro–in Vivo Gap There are a number of technical hurdles in understanding the in vivo roles of DCs and M␾s in generating TCD8⫹ responses. 1. It is uncertain to what extent the behavior of cells in vitro accurately reﬂects their properties in situ, or for that matter, how their properties vary dependent on their method of isolation or propagation. Particularly daunting is the in vitro characterization of peripheral pAPCs, because their recovery requires manipulations that alter their properties. 2. In view of the potential differences in function of DCs obtained via different methods, we will refer to DCs by their source and method of preparation: DCX–Y, where X ⫽ B (BM), S (spleen), or P (peripheral blood), and Y ⫽ F (fresh) or C (cultured) with cytokines in vitro (most often granulocyte/monocyte colony-stimulating factor (GM-CSF), used alone, 22 The complement system is one of the major components of the innate immune system and interacts in an intricate manner with the acquired immune system (Bercovich et al., 1997). Activated complement components, which bind to antibodies and the surfaces of microorganisms, play an important role in the endocytosis and phagocytosis of foreign material.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

17

but sometimes in combination with IL-12 and other cytokines). M␾s no doubt also differ depending on their source (at least subtly), and a similar scheme will be used: M␾X–Y, where X ⫽ B, S, or P (peritoneal), and Y ⫽ L (fresh, elicited by Listeria monocytogenes infection), T (fresh, elicited by thioglycolate injection), O (fresh, elicited by mineral oil injection), or C (cultured). In situations in which the method of preparation of cells is vague or not speciﬁed, this is duly noted with a question mark (?). 3. Knockout mice selectively lacking DCs or M␾s (or DCs and M␾s) are not available, nor are there antibodies known to eliminate selectively the cells in vivo. Ablation experiments are limited to the use of compounds whose selectivity is uncertain (discussed below). 4. It is likely that activation of resting TCD8⫹ following immunization is achieved by extremely small numbers of pAPCs, making the direct isolation and characterization of pAPCs bearing antigens derived from an infectious organism or vaccine preparation daunting if not outright impossible, and limiting characterization to in vitro stimulation of TCD8⫹ or adoptive transfer experiments in which the nature of the APC is inferred by antibodybased depletion. E. METHODS FOR STUDYING EXOGENOUS ANTIGEN PRESENTATION 1. Quantitating Complexes Generated from Exogenous Antigens The generation of TCD8⫹ to a given determinant requires the expression of a threshold number of complexes on the APC surface (reviewed in Yewdell and Bennink, 1999). This number varies depending on the properties of the TCD8⫹ expressing complementary TCRs. In theory, there might be an optimal number of complexes for inducing a maximal response of TCD8⫹ with high-afﬁnity TCRs,23 because expression of superoptimal quantities of complexes could activate ‘‘low-afﬁnity’’ TCD8⫹ that interfere with the generation of high-afﬁnity TCD8⫹. Although this has been observed in vitro (Alexander-Miller et al., 1996), the evidence in vivo supports ‘‘the more the merrier’’ school of thought. It is clear that in many situations, antigen delivery is suboptimal, and increases in the number of complexes generated result in enhanced TCD8⫹ responses. For example, expression of determinants as ER-targeted peptides greatly enhances responses to some determinants without compromising the quality of the responses (Restifo et al., 1995). In general, then, vaccines should aim to produce the maximal number of complexes on afAPCs. In evaluating the in vitro antigenicity of various 23

All things being equal, TCD8⫹ with high-afﬁnity TCRs will be more sensitive than TCD8⫹ with lower afﬁnity TCRs. Activation is a complicated process, however, and no doubt there are other factors that come into play. The goal of vaccination is to produce TCD8⫹ that are triggered by the minimal number of relevant complexes.

18

JONATHAN W. YEWDELL et al.

exogenous antigen preparations it is crucial to consider the quantities of complexes formed. This has been sorely lacking in the ﬁeld of exogenous antigen presentation. Most commonly, functional assays have been used to measure antigen presentation. These assays are convenient, economical, and highly sensitive, but are poor, at best, at quantitation. The most commonly used assay is the lysis of target cells (usually assessed by release of 51 Cr) which is basically a quantal assay: cells expressing a threshold number of complexes (or 1000-fold more) are lysed, and those expressing subthreshold amounts are spared. Relative efﬁciencies of antigen processing /presentation can be obtained by performing the assay under limiting conditions, but this provides only the crudest estimate of quantities of complexes (Anto´n et al., 1997). Another assay often used to study presentation of exogenous antigens is TCD8⫹ activation (usually assessed by cytokine release or activation of a reporter gene). The strength of this assay, the detection of presentation by a low fraction of cells, is a great weakness for quantitating the number of complexes generated by cells, and also obviously a liability in determining the fraction of cells capable of presenting the threshold number of complexes required for activation. Complexes can be quantitated by two methods. One is the biochemical method pioneered by Rammensee and colleagues, in which peptides in high-pressure liquid chromatography (HPLC) fractions from acid extracts are titrated in functional assays of TCD8⫹ recognition in comparison to a known amount of peptide (Rammensee et al., 1997).24 This method requires large numbers of cells, making it arduous and expensive (if not outright impossible) for most studies of exogenous antigen presentation, which utilize cells available in limited quantities ex vivo or cells that must be propagated in vitro. Another liability is that intracellular complexes (which may not make it to the cell surface) are included with cell surface complexes in the pool of peptides recovered from cells. Better is the use of monoclonal antibodies (mAbs) speciﬁc for complexes, termed ‘‘T-AGs’’ (for T cell antigen mAbs) by Germain and colleagues (1997). These have been produced from phage libraries (Andersen et al., 1996), or by standard hybridoma methodology (Porgador et al., 1997). Alternatively, multivalent TCRs25 obtained from speciﬁc TCD8⫹ have been 24 This reference is an amazing compendium of MHC-associated ligands presented in all sorts of permutations of source of antigen, allomorph, etc. It is an invaluable resource for MHC-ologists, as is the companion web site maintained at the University of Tuebingen (www.unituebingen.de/uni/kxi/immunol.html). Other web sites chock ﬁlled with MHC ligands and predictive algorithms are www.bimas.dcrt.nih.gov/molbio/hla bind and www. wehih.wehi.edu.au/mhcpep. 25 The afﬁnity of the TCR is too low for cytoﬂuorographic detection, so TCRs must be used in multimeric form to boost the functional avidity of the interaction. T-AGs are selected on the basis of having a sufﬁcient avidity for detection.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

19

used to similar effect (O’Herrin et al., 1997). These reagents enable the rapid and precise cytoﬂuorographic quantitation of complex numbers on individual APCs. Their potential value to studies of exogenous antigen presentation is difﬁcult to overstate, and every effort should be made to produce similar reagents speciﬁc for additional complexes. There are limitations to their use, however, the foremost being difﬁculties in producing mAbs of suitable speciﬁcity and problems inherent to expressing functional TCRs. Another drawback is their sensitivity, being far less than the most sensitive TCD8⫹ (hundreds to thousands of complexes per cell for TAGs versus 1–10 complexes per cell for ‘‘primo’’ TCD8⫹). 2. Tools and Strategies for Dissecting Mechanisms of Presentation a. Inhibitor-User Responsibility Code. Inhibitors of cellular functions are indispensable, powerful tools for mechanistic studies of antigen processing in vitro. With this power, however, comes certain responsibilities. Users should: 1. Not blithely assume that the inhibitors are active in the cells studied, and should perform controls to establish their efﬁcacy. Where possible, these controls should be functional in nature, i.e., demonstrate the predicted effect on the presentation of control antigens to TCD8⫹, because residual metabolic activity undetectable by other means may be sufﬁcient to generate the few complexes needed for TCD8⫹ recognition.26 2. Perform controls on the cells in question, and not another cell that provides a more convenient means of demonstrating the efﬁcacy of the inhibitor. 3. Demonstrate that, when drugs block presentation, the effect is not due to cell death or morbidity. Using reversible inhibitors, the effects should be shown to be reversed; using irreversible inhibitors, at the very least, the viability of the cells should be conﬁrmed. 4. Always consider that in addition to their expected effects, inhibitors can block or enhance other processes, either via direct interaction of the drug,27 or as a secondary effect of blocking the target process.28 b. Establishing How Exogenous Antigens Are Processed. Distinguishing intracellular from extracellular processing. The ﬁrst question that should be addressed in characterizing the mechanism of 26

This is less of a problem using T-AGs because they provide a linear readout of complex expression. 27 An infallible rule: drugs can only become less speciﬁc following their initial description, and invariably do so. 28 For example, proteasome inhibitors rapidly induce molecular chaperones in cells and inhibit ubiquitination, either of which could affect antigen processing.

20

JONATHAN W. YEWDELL et al.

presentation of exogenous antigens is whether the antigen binds to peptidereceptive molecules on the cell surface ( ‘‘extracellular presentation’’) or must be processed intracellularly. Long peptides are capable of binding to class I molecules with afﬁnities that result in dwell times on the order of minutes, and proteases present in the plasma or on the cell surface can trim such polypeptides to higher afﬁnity forms (Ojcius et al., 1994). The presence of class I-binding peptides in protein preparations is the rule rather than the exception, and may persist despite the best efforts for their removal. This question is often addressed in the negative sense, i.e., it is ruled out by evidence favoring the involvement for cytosolic or endosomal processing (as described in the next two sections). There are two other useful methods for assessing extracellular presentation in vitro. 1. Cells are treated with low concentrations of paraformaldehyde. This prevents most cellular metabolic processes, including endo- and exocytosis,29 and actually increases the abilities of cells to bind and present peptides—indeed, so much so that it could potentially result in creating extracellular presentation when this does not occur with live cells. 2. The dependence of exogenous 웁2m for presentation is determined. As mentioned above, exogenous 웁2m greatly enhances the number of PR molecules on the cell surface. If presentation is decreased by the absence of 웁2m, this is reasonable evidence that exogenous presentation is at least a contributing factor, possibly even the sole mechanism of presentation.30 In addition there may be special features of an antigen that can be used to advantage. For example, if a virus that enters the cell by cell fusion is rendered nonantigenic by inactivating its fusion activity, this is good evidence against extracellular presentation. Testing cytosolic involvement. The next question addressed in mechanistic studies of exogenous antigen presentation is whether peptides are generated by the standard cytosolic route or via an alternative pathway. There are ﬁve ways of assessing whether cytosolic processing is necessary for presentation. 1. Proteasome inhibitors. A number of available drugs inhibit proteasomes. The most speciﬁc (and expensive) is lactacystin, which has minimal 29

Again, inhibition of endocytosis should be established experimentally and not taken on faith. 30 The complete loss of presentation provides strong evidence that this is the sole route of presentation. We have to hedge a bit here, however, because the extent that exogenous 웁2m contributes to the generation of PR molecules in endosomes is uncertain. (There is some evidence for this in the processing of hepatitis B virus S antigen, described below).

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

21

effect on other proteases.31 The sensitivity of cells to lactacystin varies considerably among established cell lines. Other less speciﬁc (and more broadly effective) inhibitors are available; these can be used in comparison to chemically similar compounds with minimal proteasome inhibitory activity to control partially for effects on nonproteasomal proteases. If presentation of exogenous antigen is sensitive to proteasome inhibitors and insensitive to nonproteasomal inhibitors, it is probable that cytosolic processing is required for antigen presentation. A lack of inhibition cannot be taken as evidence for nonstandard processing, because, as discussed above, there are now numerous examples of antigens (both endogenous and exogenous) whose cytosolic processing is not inhibited by proteasome inhibitors, and may even be enhanced. 2. Brefeldin A (BFA). BFA inhibits presentation of endogenous antigens by preventing membrane and secreted proteins, including class I molecules loaded with peptides, from leaving the early secretory compartment. It is effective on nearly all cell types. By blocking the export of nascent class I molecules from the ER, however, BFA reduces the number of PR molecules on the cell surface and in endosomes. Moreover, BFA also has direct (complicated) effects on endosomes, including blocking the transport of peptide-loaded class II molecules from endosomes to the plasma membrane (Pond and Watts, 1997). Thus, BFA can potentially inhibit presentation of exogenous antigens that associate with class I molecules in endosomes or on the cell surface, and BFA sensitivity should not be equated with peptide loading in the ER. Because BFA blocks exocytosis rather quickly (easily within minutes), it should be applied to cells as late as possible following exposure of the cells to antigen. Indeed, if prolonged exposure to BFA is necessary to block antigen presentation, this is reasonable evidence for peptide loading in a post-ER compartment. A novel BFA-based method of demonstrating cytosolic processing of exogenous antigens is to demonstrate TAP-dependent, T-AG staining of the ER of ﬁxed and permeabilized cells, the BFA being necessary to accumulate a sufﬁcient number of complexes for detection (Porgador et al., 1997).32 3. Protein synthesis inhibitors. Cycloheximide and other protein synthesis inhibitors have also been used in a manner analogous to BFA use, the 31

The nitrophenol derivative of carboxybenzyl-leucyl-leucyl-leucine vinyl sulfone (NLVS) is a novel proteasome inhibitor that may be as speciﬁc as lactacystin (for sure it is just as expensive): it clearly has different properties compared to lactacystin, however (Bogyo et al., 1997), and should be a useful complement for antigen-processing studies. 32 Actually, this use has yet to be reported, but it should be possible for antigens that produce abundant complexes because we have used T-AGs to detect complexes in the ER of BFA-treated cells producing an ER-targeted peptide (Porgador et al., 1997), and also to demonstrate the association of exogenous peptides with class I molecules in the ER (Day et al., 1997). Note that it would still be necessary to examine whether processing occurs in the cytosol by investigating the TAP/proteasome dependence of staining.

22

JONATHAN W. YEWDELL et al.

idea being that PR molecules in the ER are rapidly depleted. This is a dangerous assumption (!),33 and the time required to deplete the pool of PR molecules in the ER must be determined for each cell line/allomorph combination. 4. TAP-deﬁcient cells. A number of available established cells lines lack expression of one or both TAP subunits, and pAPCs and non-pAPCs can be obtained from TAP1⫺/⫺ mice. Additionally, it is possible to block TAP function efﬁciently in human cells (but not mouse cells) by expressing the herpes simplex virus protein ICP47. Nearly all peptides generated in the cytosol require TAP for their efﬁcient transfer into the ER. If, using cells matched as closely as possible except for the expression of functional TAP, TAP deﬁciency is associated with greatly decreased efﬁciency of presentation, this is good evidence that the determinant must traverse the cytosol to be presented. It is a common misconception that TAP-expressing cells express far less cell surface PR molecules than TAP-deﬁcient cells; indeed, if anything, loss of TAP function is associated with decreased numbers of PR molecules (Day et al., 1995). Thus, a partial loss of presentation associated with TAP-deﬁcient cells could reﬂect either cytosoldependent or cytosol-independent processing. 5. Gelonin and ricin B chain. These are potent enzymes that inactivate ribosomes—so potent that in at least some cell lines only one or several molecules in the cytosol are sufﬁcient to block protein synthesis completely. Both enzymes are membrane impermeant, and therefore can be used as a highly sensitive means of detecting whether an agent or conditions induce the delivery of exogenous proteins to the cytosol. Indeed, they may be too sensitive, because it is easy to envision conditions in which relevant processing occurs endosomally but delivery of endosomal contents to cytosol occurs in quantities insufﬁcient to generate signiﬁcant amounts of peptides, but enough for gelonin or ricin to act. Testing endosomal involvement 1. Raising endosomal pH. The normal function of the endocytic pathway results in the increasing acidiﬁcation of vesicles as they progress from the plasma membrane to lysosomes. Raising endosomal pH reduces the activity of most resident proteases and can also interfere with membrane trafﬁcking. Because all the agents available also raise the pH of other acidic compartments in cells (including the trans-Golgi complex), this must be taken into 33 We have found, for example, that functional PR Dk molecules remain in the ER of L929 cells for more than 24 hr in the presence of a mixture of potent protein synthesis inhibitors ( J. Yewdell and J. Bennink, unpublished results). Somewhat ironically, as discussed below, protein synthesis inhibitors can be used as supporting evidence for bona ﬁde exogenous nature of viral antigens.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

23

account in interpreting their effects. These agents fall into two categories, lysosomotropic weak bases (e.g., NH4Cl),34 which act by becoming trapped in protonated (and therefore buffering) forms in acidic compartments, and proton pump inhibitors (e.g., baﬁlomycin). 2. Inhibitors of endosomal proteases. There are a number of relatively impermeant broadly active protease inhibitors (e.g., leupeptin, blocking some serine and most cysteine proteases; E64, blocking cysteine proteases) whose activity is limited to extracellular and cell surface proteases and cellular compartments reached by pinocytosed material. 3. Inhibitors of endocytosis. Phagocytosis can be blocked by cytochalasin B or D through inhibitory effects on the contraction of actin-containing microﬁlaments (cytochalasin D is the preferred agent because it has less effects on other cellular processes). Coated pit-mediated endocytosis can be blocked by reducing cellular ATP levels (achieved by azide and 2deoxyglucose), interfering with G-proteins (aluminum ﬂuoride), by acidifying the cytoplasm (incubation with amiloride in sodium-free media), or by wortmannin, a phosphatidylinositol 3-kinase inhibitor. The usefulness of each of these four treatments is limited by their rather broad effects on cellular metabolism. A more speciﬁc (but technically demanding) means of blocking endocytosis is the method of horseradish peroxidasemediated ablation (Pond and Watts, 1997): cells are allowed to endocytose transferrin-conjugated enzyme and then are reacted with a substrate in the presence of H2O2 generating a reaction product that forms insoluble precipitates that block endosome function, and as a consequence also block the formation of new endosomes. c. Inactivated Viruses. Inactivated viruses have been used as a source of exogenous antigens in previous studies, and as we argue below, should be studied more intensively in future studies. Many past studies have suffered from the possible residual ability of an ‘‘inactivated’’ preparation to direct protein synthesis from an incompletely inactivated gene. It is insufﬁcient to demonstrate that infectious virus is no longer detectable following inactivation. Inactivation of a single viral gene is often sufﬁcient to block replication without preventing mRNA production from other genes. Inactivated viruses are generally used at high concentration relative to infectious viruses, a condition that favors multiplicity reactivation.35 The sensitivity of TCD8⫹ is usually much greater than biochemical means to detect viral gene expression, so the best control for virus inactivation is the demonstration 34

Chloroquine, often used for this purpose, can have idiosyncratic effects on other cell functions, and should be avoided or used with trepidation. 35 This is the mutual complementation of defective genes following infection of cells with multiple virions.

24

JONATHAN W. YEWDELL et al.

that a determinant from a nonstructural viral gene is not presented to TCD8⫹.36 An additional control that should be performed for in vitro studies is the demonstration that presentation continues in the presence of inhibitors of protein or RNA synthesis.37 d. Ablating pAPCs in Vivo. There have yet to be reports of knockout or mutant mice with a selective loss of a particular subset of pAPCs. It has been observed that Rel-B⫺/⫺ mice lack classical interdigitating DCs in lymphoid organs (Burkly et al., 1995), and M-CSF⫺/⫺ mice lack M␾s in most tissues (Yoshida et al., 1990). However, the status of pAPC populations in these mice remains to be carefully characterized, and in any event, loss of these genes certainly affects other cell lineages, possibly to the further detriment to T cell responses. Currently, the best method for studying the roles of DCs and M␾s in presenting exogenous antigens in vivo is the use of substances that are cytotoxic when internalized, particularly when it can be shown that TCD8⫹ responses are restored on adoptive transfer of a given cell type (Debrick et al., 1991). Three such agents have been used: silica, carrageenan (a long-chain, high-molecular-weight sulfated polygalactose derived from red algae), and liposome-mediated delivery of clondronate. The effects of both silica and carrageenan on phagocytic cells were described prior to the development of reagents capable of distinguishing between types of pAPCs. These agents continue to be used in the absence of detailed characterization of their effects on pAPC subsets. Liposome-mediated delivery of clondronate is the most speciﬁc method of ablating M␾s in vivo (van Rooijen and Sanders, 1994). Although it has been reported that that splenic DCs are unaffected by clondronate treatment (Nair et al., 1995), its effects on DCs in nonlymphoid organs have not been characterized. Because it is these DCs that are most phagocytic (at least allegedly), there is a good chance that they are affected by clondronate, which has no known mechanism of speciﬁcally targeting M␾s, and whose toxicity should be proportional to its rate of internalization. F. ASSESSING THE IMMUNOGENICITY OF EXOGENOUS ANTIGEN PREPARATIONS The major application of studies of exogenous antigen presentation, at least in the short term, is the development of protein-base vaccines that 36

Not even this is a completely foolproof control, however, because the control determinant from the nonstructural protein may be presented or recognized at much lower efﬁciency on infected cells than the determinant in question. 37 As discussed above, inhibition by protein synthesis inhibitor does not necessarily disprove processing of exogenous viral antigens, because the pool of PR class I molecules in the ER will be decreased.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

25

induce memory TCD8⫹ responses. Due to the convenience and power of experimental systems, as well as ﬁnancial and ethical considerations, the initial vaccinees are usually mice. Although the mouse and human immune systems are not identical, they are sufﬁciently similar that immune responses in mice are useful predictors of human responses. A more serious limitation is that the infectivity and pathogenesis of many microorganisms are species speciﬁc, often to the extent that protection experiments are not even possible in mice or other animals, for that matter. Often, lacking evidence of efﬁcacy in animals, vaccine trials must be performed in humans. When tested in mice, vaccines designed to elicit TCD8⫹ responses to speciﬁc determinants have a unique problem related to MHC restriction: most HLA-restricted determinants do not associate with mouse class I molecules, at least not with the seven mouse allomorphs expressed by the most commonly used strains. This problem can be surmounted using transgenic mice expressing HLA allomorphs. Somewhat surprisingly, determinants immunogenic in humans are also immunogenic in transgenic mice, with no known exceptions to date (reviewed in Yewdell and Bennink, 1999). Vaccines based on whole-protein antigens do not face this problem, at least not directly, because most proteins will have determinants restricted by multiple mouse and human allomorphs. Experiments in mice can establish that the protein is processed for TCD8⫹ recognition, but to determine whether a given HLA-associated determinant is immunogenic again requires the use of HLA-transgenic mice. In the subsequent sections we describe a number of protein preparations that are able to induce TCD8⫹ responses in mice. Whatever the nature of the vaccine, it should always be considered that the demonstrated immunogenicity in mice is dependent partially or entirely on contaminants in the preparations with intrinsic adjuvant activity. The most common and notorious, of course, are bacterial endotoxins, which are ubiquitously present in protein preparations, and may be minimized but not eliminated. Although simple assays are available to assess the degree of endotoxin contamination (mandatory, of course, before administering substances to humans), they are infrequently used in mouse studies. A number of inbred mouse strains that are much less sensitive to the effect of endotoxin can be used to assess the contribution of endotoxins to immunogenicity, but again these are used only infrequently. In the end, of course, compounds with similar adjuvant activities may be added to the vaccine formulations, but it is difﬁcult (if not impossible) to evaluate the relative intrinsic immunogenicity of exogenous antigen preparations without controlling for the presence of endotoxins.

26

JONATHAN W. YEWDELL et al.

III. Processing and Presentation of Exogenous Antigens: Reviewing the Recent Literature

A. PROLOGUE In the subsequent sections we detail the numerous mechanisms that are used in generating responses to exogenous antigens. The use of one mechanism over another can be critically inﬂuenced by the precise conditions in which the antigen is introduced into the organism. In evaluating the signiﬁcance of a given mechanism, experimental manipulations and observations can be divided into ‘‘what does occur’’ and ‘‘what can occur’’ categories. The two are intimately entwined and often not easily dissected, but it will still be useful for the reader to keep the distinction in mind. To cite some concrete examples, we would like to understand the role of cross-priming in the TCD8⫹ response to human immunodeﬁciency virus (HIV), but even if there is no role for cross-priming, we would still like to know if cross-priming is involved in the generation of TCD8⫹ responses to HIV proteins encoded by DNA vaccines, and if it is, under what natural circumstances this occurs. Even for the most pragmatic vaccinologist, understanding what does happen in the course of a mouse immune response to vesicular stomatitis virus can provide important clues for improving human vaccines, and even for academics in the loftiest ivory towers, invaluable insight into natural processes can come from the most mundane vaccine trial. B. PARTICULATE ANTIGENS 1. Bacteria a. The Three Paths to Presentation. The processing of bacteriaassociated proteins into class I-associated peptides is relevant for understanding the TCD8⫹ response both to pathogenic bacteria (particularly intracellular bacteria, wherein TCD8⫹ effector function can play an important antibacterial role) and to bacteria genetically modiﬁed to deliver vaccine antigens. There are three basic ways in which bacterial antigens can be processed into class I-binding peptides. First, bacteria can secrete proteins with the capacity to breach cellular membranes, enabling their delivery to the cytosol—possibly along with other bacterial proteins in the vicinity of the translocation site. Second, bacterial proteins can be delivered to the cytosol through the devices of the APC. In both cases, antigen peptides or their immediate precursors are generated by cytosolic proteases. Third, bacterial proteins can be processed in endosomes and bind class I molecules in endosomes or on the cell surface. There is good evidence for the use of each of these pathways for the processing of bacterial antigen in vitro.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

27

b. Purpose-Made Pores. The ﬁrst pathway is classically used by L. monocytogenes, which produces a protein, listerolysin O (LLO), that allows bacteria to escape phagosomes into the cytosol, an important part of the life cycle of the bacterium.38 This enables cytosolic processing and presentation of a number of bacterial proteins, as shown by a number of criteria: requirement for live bacteria that express LLO, sensitivity to BFA and proteasome inhibitors, and TAP dependence (Pamer et al., 1997). Predictably, L. monocytogenes has been shown to facilitate the entry of OVA into the cytosolic processing pathway; this again requires live bacteria and LLO expression, and is TAP dependent (Mazzaccaro et al., 1996; Darji et al., 1997a).39 LLO is not absolutely required for processing of L. monocytogenes proteins, but the processing pathway utilized in this circumstance is not well deﬁned (Szalay et al., 1994; Zwickey and Potter, 1996). Along similar lines, Salmonella typhimurium secretes a number of proteins that assemble to form a pore in host membranes, enabling other secreted bacterial proteins to access the cytosol (Russmann et al., 1998). Insertion of deﬁned TCD8⫹ determinants into one such protein resulted in its presentation by APCs in vitro to TCD8⫹ of corresponding speciﬁcity; such presentation required the expression of the pore proteins and was TAP dependent, solid evidence for cytosolic processing. Inoculation of mice with bacteria expressing this protein primed for memory TCD8⫹ responses, but the immunogenicity of bacteria not expressing a functional pore was not examined, so it is not clear whether this mechanism is also required for the generation of determinants on afAPCs in vivo.40 c. Alternative Routes. The second and third pathways of bacterial antigen presentation have been delineated largely through the efforts of Harding, Pﬁefer, Wick, and their colleagues. Initially, the immunodominant Kb-restricted determinant from OVA corresponding to residues 257– 264 (Ova257–264) was expressed by Escherichia coli in three contexts: a cytoplasmic protein (Crl-OVA), or the peri- or exoplasmic domains of an outer membrane protein (Pfeifer et al., 1993). All three gene products were functional (i.e., could rescue mutant E. coli that do not produce the functional wild-type protein), providing the strongest possible evidence that the fusion proteins behaved in a manner similar to that of natural 38

For a compendium of reviews of the immune response to L. monocytogenes, see Parham (1997). 39 The major point of Mazzaccaro et al. was actually that Mycobacterium tuberculosis does the same thing, the inference being that its proteins are also processed by cytosolic machinery for TCD8⫹ recognition. 40 The ‘‘missing’’ experiment was performed with bacteria expressing a different nominal protein, but the readout was protection of mice from challenge infection with the source virus expressing the nominal protein, and TCD8⫹ priming was not directly examined.

28

JONATHAN W. YEWDELL et al.

proteins. Regardless of the protein context, Ova257–264 was presented to OVA-speciﬁc TCD8⫹ hybridomas by M␾sP–L41 exposed to the manipulated bacteria. Phagocytosis was required for presentation because stimulation was blocked by cytochalasin D and not detected in a nonphagocytic tissue culture cell line nor by exposing M␾P–Ls to the soluble bacterial protein in the absence of bacteria. Most importantly, presentation occurred within 20 min of exposing cells to bacteria42 and was not inhibited by BFA. Determinants produced by M␾P–Ls lacking Kb molecules could be transferred to paraformaldehyde-ﬁxed Kb-expressing M␾P–Ls, a phenomenon aptly (if unappetizingly) termed peptide regurgitation. This indicates that peptides generated in phagosomes could bind to cell surface PR molecules in addition to whatever PR molecules exist in the phagosomes. Peptides could even be transferred to cells across a ﬁlter, indicating that loading can occur simply by diffusion, but it was noted that the efﬁciency of loading in this manner may be far lower than more intimate delivery. In a subsequent report, M␾P–L presentation of Ova257–264 from E. coli producing Crl-Ova protein was found to be enhanced by exogenous 웁2m and insensitive to proteasome inhibitors, supporting its extracytosolic processing43 (Song and Harding, 1996). Paradoxically, TAP-deﬁcient M␾P–Ls required 앑30-fold as much antigen as normal M␾s to achieve similar stimulation of the TCD8⫹ hybridoma. It was shown, however, that the efﬁciency of presentation paralleled the number of PR cell surface molecules, arguing strongly that TAP is required not for ER delivery of cytosolic peptides but for the creation of PR molecules on the cell surface or in endosomes.44 Moreover, the authors observed that BFA could affect presentation, but in a manner consistent with its effects on cell surface PR molecules and not by blocking export of class I molecules loaded with Ova257–264 in the ER.45 Harding and colleagues also demonstrated that M␾P–Ls could generate Ova257–264 when infected with S. typhimurium expressing the determinant in the secreted protein studied in E. coli (Pfeifer et al., 1993). The efﬁciency of presentation of the determinant was similar for the two bacteria, at 41 This terminology was introduced on page 16–17. We recognize that it is an eyeful, and most readers should not expend any effort in continually deciphering the code; our major purpose is not torture but to emphasize that the properties of pAPCs can vary considerably depending on how they are obtained. 42 This is very fast for class I-restricted presentation: the record for presentation of OVA introduced into the cytosol is 45 min (Hosken et al., 1989). 43 It was not established, however, that liberation of Ova257–264 from Crl-OVA delivered to the cytosol or from Crl-OVA synthesized endogenously is sensitive to proteasome inhibitors, which has been shown for OVA. 44 We warned you about this on page 22. 45 Ditto, page 21.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

29

least in terms of the number of bacteria required for stimulation of the hybridoma. In contrast to M␾P–Ls, an epithelial cell line was unable to generate detectable amounts of Ova257–264 from S. typhimurium (Harding and Pfeifer, 1994). This could not be attributed to inefﬁcient internalization of bacteria. The epithelial cells and M␾s were similarly efﬁcient at presenting a synthetic peptide corresponding to Ova257–264. (To discriminate between determinants and their synthetic peptide counterparts, we will preﬁx the determinant with ‘‘sp,’’ e.g., spOva257–264.) This argues that the cells were much less efﬁcient than M␾P–Ls at generating peptides from the bacterial protein (and not that greater numbers of complexes were needed for stimulation46), and therefore that the generation of complexes in M␾P–Ls endosomes reﬂects specialized properties of the endosomes, and not simply the ability of M␾P–Ls to phagocytose. The effects of protein context on M␾P–?s-mediated presentation of bacterial antigens was further characterized by Wick and Pfeiffer (1996). Using an OVA-speciﬁc TCD8⫹ hybridoma, they compared presentation of Ova257–264 derived from E. coli synthesizing OVA, Crl-OVA, or another cytosolic bacterial fusion protein (MBP-OVA). Presentation of each was cytochalasin D sensitive, consistent with endosomal processing of the proteins. Presentation of OVA was 앑30-fold more efﬁcient than presentation of Crl-OVA or MBP-OVA. By contrast, using M␾P–Ls from TAP1⫺/⫺ mice, the three proteins were presented at similar (low) efﬁciencies: for Crl-OVA and MBP-OVA this represented a 3-fold decrease versus presentation by normal M␾s, but for OVA a 100-fold loss in efﬁciency. These ﬁndings indicate that the presentation of a determinant is inﬂuenced by its context—either related to primary or secondary polypeptide structure, intracellular localization, or metabolism—or a combination of these factors. The simplest explanation for the quantitative differences in presentation of OVA versus Crl-OVA or MBP-OVA in wild-type and TAP⫺/⫺ M␾s is that the three are presented at similar efﬁciency by endosomal processing, but that only OVA has ready access to cytosolic processing, which would then be 앑100 times more efﬁcient at generating peptides than the endosomal route.47 As argued below, this is more likely to be 46

Some caution is in order here, though. Two cells may demonstrate identical dose– response curves in presenting synthetic peptides, yet the number of complexes presented may differ widely at a given dose if the cells have greatly differing numbers of PR molecules. Using the present case as an example, if the epithelial cells expressed 100-fold greater levels of PR molecules than M␾s, then they would be 앑100-fold less efﬁcient at presenting Ova257–264 complexes, which would mean that they could generate the same number of complexes from bacteria and not be recognized. 47 This is testable, because cytosolic processing of exogenous (and endogenous) OVA is blocked by proteasome inhibitors, and cells treated with proteasome inhibitors should present OVA with similar efﬁciency as Crl-OVA or MBP.

30

JONATHAN W. YEWDELL et al.

related to the inability of cytosolic proteases to liberate the determinant from Crl-OVA or MBP-OVA than to the unique delivery of OVA to the cytosol. Wick and colleagues also compared M␾P–?s and DCB–C presentation of E. coli producing Crl-OVA to an Ova257–264-speciﬁc hybridoma. Initially they showed that DCB–C generation of Ova257–264 was blocked by cytochalasin and NH4Cl (Svensson and Wick, 1998). Next (M.J. Wick, personal communication), presentation was found to be reduced at least 10-fold in TAP⫺/⫺ cells relative to wild-type cells, and was sensitive to BFA and a proteasome inhibitor. In contrast to M␾P–?s DCB–Cs did not produce peptides that were presented by bystander M␾P–?s. Altogether, these ﬁndings suggest that both M␾s and DCs process phagocytosed bacteria via the cytosolic pathway but that endosomal processing is limited to M␾s48. d. In Vivo Complications. The potential complexities of the in vivo processing of bacterial antigens are illustrated by two studies. Shen et al. (1998) produced L. monocytogenes that synthesize cytosolic or secreted versions of a protein containing a deﬁned viral determinant recognized by TCD8⫹. The targeting of the protein had no appreciable effect on the abilities of the bacteria to induce TCD8⫹ responses to the viral determinant. If, however, mice immunized with the corresponding virus (to induce TCD8⫹ for the viral determinant) were challenged with the bacteria, protection was afforded only against bacteria secreting the protein. Similarly, in vitro, virus-speciﬁc TCD8⫹ recognition of bacterially infected target cells required the secreted form of the protein. The authors’ interpretation follows: 1. The secreted antigen, but not the whole bacterium, has access to the cytosolic processing pathway in all infected cells due to LLO-created pores in endosomal membranes. 2. A subset of cells (yet to be identiﬁed) is able to process endosomally bacteria-associated proteins into class I-binding peptides, which bind either to endosomal or cell surface PR molecules and activate TCD8⫹ responses. 3. An important subset of efAPCs cannot process nonsecreted antigen in this manner, preventing clearance of the infection. This is a compelling argument, but it remains to be demonstrated that secreted and nonsecreted proteins are processed, respectively, in the cytosol and the endosomes. Further, it is possible that the results reﬂect 48 To compare accurately the processing efﬁciencies of M␾s and DCs (or any pair of cells, for that matter), it is necessary to quantitate the number of cell surface complexes (using T-AGs when possible); there are just too many uncertainties in TCD8⫹-based assays.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

31

quantitative differences in the amount of complexes required for TCD8⫹ activation rather than qualitative differences in antigen processing by diverse cell types.49 Darji et al. (1997b) reported that infection of mice with S. typhimurium results in the transfer of bacterial DNA into host cells. The expression of bacteria-derived DNA was detected using genes with mammalian promoters that had been introduced into bacteria. Although it is clear that levels of expression of genes under the control of bacterial promoters will be lower, these important ﬁndings raise the eerie possibility of endogenous presentation of bacterial genes.50 e. Analysis and Prospects. Taken together, these ﬁndings indicate that bacterial antigens can be processed by multiple routes. Proteins from bacteria with speciﬁc means of penetrating phagosomes are probably preferentially processed in the cytosol, whereas proteins from bacteria that remain in phagosomes are largely processed in endosomes, at least in M␾s.51 Such endosomal processing may be of such low efﬁciency in nonM␾s that the cytosolic route is signiﬁcantly more efﬁcient, even for nonpenetrating organisms. Curiously, the targeting of the antigen (cytosol vs. membrane vs. secreted) seems to have little effect on its immunogenicity, although it can have large effects at the level of efAPCs in vivo. This conclusion is based a very small sample size, however, and is further compromised by confounding effects of altered ﬂanking sequences. Assuming there is some truth to it, this could mean either that APCs gain access to intracellular locations in bacteria, or that the nominally internally located proteins are present on the surface of the bacteria, or present in the inocula due, perhaps, to disruption of bacteria during their in vitro growth. In the latter case, there is the possibility that such determinants would not be produced during the course of a real infection. This points to an important limitation of studies in which deﬁned determinants are placed within bacterial proteins. Aside from the possibility of perturbing the protein structure [even if the gene product is functional, as so elegantly demonstrated by Pfeifer et al. (1993), a higher percentage may be misfolded, mistargeted, etc.], the use of nominal determinants 49

Of course, ‘‘qualitative’’ differences in the end reﬂect quantitative differences—another argument for aiming to quantitate complex numbers. 50 Here is a Byzantine (but plausible) scenario: cross-presentation of a bacterial protein expressed endogenously by a non-pAPC. 51 It should be noted, however, that in most of the experiments discussed antigen processing was allowed to proceed for a relatively short period (generally 2 hr or less), which would bias against cytosolic processing if this proceeds at a more leisurely pace than endosomal processing.

32

JONATHAN W. YEWDELL et al.

normally produced by cytosolic processing greatly biases the system away from alternative processing pathways.52 It is nearly inconceivable that the types of ligands resulting from the action of endosomal and cytosolic proteases are identical, and it is possible that the spectra of class I ligands produced overlap only slightly. This is not to detract from the value of these pioneering studies on endosomal processing of bacteria, which, at the very least, deﬁne the possible. It is important, however, to expand these studies to include naturally processed determinants of nonpenetrating bacteria that elicit TCD8⫹ with antibacterial activity. The complexity of bacteria makes it a daunting task to identify determinants, but this has been achieved using expression libraries (Lenz et al., 1996), and ongoing improvements in mass spectroscopy in conjunction with complete sequencing of bacterial genomes may make the biochemical identiﬁcation of antigenically active peptides a feasible approach in the near future. 2. Artiﬁcial Particles Shortly after the publication of the initial report characterizing the processing of bacterial antigens, Rock and colleagues published the ﬁrst study to examine the processing of an antigen coupled to beads (KovacsovicsBankowski et al., 1993). Once again, OVA was used as a model antigen, and presentation was measured by the ability of APCs to stimulate an Ova257–264-speciﬁc hybridoma. Coupling OVA to either iron or latex beads increased the efﬁciency of presentation by three or four orders of magnitude, in terms of the total amount of OVA required to achieve 50% activation of the hybridoma. The optimal size of beads for activation was 2–3 애m; beads that were too large (10 애m) or too small53 were inefﬁcient sources of OVA. Presentation was observed with M␾P–Ts or cloned, retrovirus-transformed M␾ cell lines but not T or B cell lines. Bystander presentation was not detected, suggesting that loading occurred intracellularly. Inhibitor studies suggested that phagocytosis was required for presentation. In a follow up study (Kovacsovics-Bankowski and Rock, 1995), it was reported that M␾ cell line presentation of OVA from beads had all of the hallmarks of cytosolic processing: BFA sensitive, proteasome inhibitor sensitive, and TAP dependent. Direct evidence for the phagosome-tocytosol route was provided by the demonstration that protein synthesis in M␾ lines but not B or T cell lines was blocked by exposure of cells to gelonin-coupled beads. This study was repeated with basically similar results using DCB–Cs transformed with oncogenes to produced cloned, permanent cell lines (Shen et al., 1997). 52 53

The old saw about searching for lost keys under street lamps comes to mind. Not speciﬁed exactly how small.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

33

The presentation of latex-conjugated OVA has also been studied by Harding and Song (1994; 1996) in experiments parallel to those they performed with bacterial antigens. In contrast to results from the Rock lab, presentation was found to occur in a cytosol-independent manner, and peptide regurgitation was detected. Wick and Pfeiffer also had a go at comparing presentation of bead-conjugated antigens to bacterially produced antigens: based on TAP dependence (proteasome inhibitors were not used), their ﬁndings were most consistent with endosomal processing of Crl-OVA and MBP-OVA and cytosolic processing of OVA (Wick and Pfeifer, 1996; Svensson and Wick, 1998). On the face of it, three extremely competent groups of investigators performing basically the same experiment come up with 2 different answers.54 There are numerous methodological differences in precisely how the experiments are done, however, that can easily account for the discrepancies.55 Importantly, neither the beads nor their coupled protein are natural antigens (although beads could possibly be used for vaccination): so the major point of the experiments is to reveal the potential mechanisms of processing phagocytosed antigens. In this, the experiments are completely successful: they reveal that both endosomal and cytosolic processing are possible, and, in the former case, that peptide association may occur with cell surface PR molecules. Additional insight into the mechanism of presentation of phagocytosed antigen comes from Reis e Sousa and Germain (1995), who found that soluble OVA was presented to TCD8⫹ by M␾P–Ts (and by every cell line tested, including L929 cells, capable of phagocytosing latex beads), but only if it was added to cells with latex beads. Based on sensitivity to BFA and inhibition by gelonin, processing was judged to be cytosolic. These ﬁndings imply that phagocytosis of the beads resulted in the transfer of the endosomal contents into the cytosol, and not just the proteins coating the beads. Notably, however, this means of transfer is not particularly efﬁcient, because threshold triggering required 0.3 mg/ml of OVA. This suggests that for this route to be relevant for bacteria, bacteria must nearly completely ﬁll their phagosomes,56 resulting in a high local concentration of bacterial proteins that are released into the phagosome. 54

This is the sort of thing that decreases (unfairly, we hasten to add) the number of invitations proffered to immunologists to speak at cell biology/biochemistry meetings. 55 These include how the M␾s were elicited, the size and composition of beads, the method of coupling the protein to the beads, the source of antigen, how long the cells were incubated with antigen, whether the cells were ﬁxed prior to the TCD8⫹ readout, and the sensitivity of the TCD8⫹. The ﬁrst factor has in fact been reported to inﬂuence the processing of OVA-beads: M␾sP–Ls were found to present OVA 100- to 1000-fold more efﬁciently than M␾sB–Cs the former presenting OVA in a BFA-insensitive manner, the latter in a BFAsensitive manner (Oh et al., 1997). 56 Indeed, as mentioned on page 14, this does happen: it is known as ‘‘tight zippering.’’

34

JONATHAN W. YEWDELL et al.

Reis e Sousa and Germain termed the bead-mediated presentation of OVA ‘‘indigestion,’’ proposing that phagocytic overindulgence in beads resulted in unnatural physical forces in phagosomes leading to phagosome leakage or lysis (Reis e Sousa and Germain, 1995). Although the presence of beads had no apparent effect on the uptake of OVA (as measured cytoﬂuorographically using ﬂuorescein isothiocyanate-conjugated OVA (FITC-OVA), differences in the rate of OVA uptake may have been missed, because internalization was measured only at a single, relatively late time point (2 hr).57 Further, the beads may have diverted OVA to ‘‘leakier’’ compartments than normal pinosomes. In both these cases, the increased cytosolic delivery would result from factors other than ‘‘indigestion.’’ 3. Cell Debris One of the key issues in the presentation of exogenous antigens is the mechanism of cross-priming. It has been known for some time that M␾s recognize and phagocytose apoptotic cells (reviewed in Savill et al., 1993). Because cell death (from apoptosis and necrosis) is common at the site of inﬂammation, and also a direct consequence of many viral infections, dead cells are prime suspects as a/the major source of exogenous antigen. The ﬁrst rumblings of what should be an avalanche of studies on the antigen processing of dead cells were felt just in the past year. Bellone et al. (1997) studied M␾Ps presentation of an undeﬁned mouse leukemia virus determinant expressed by a virus-transformed cell line and its TAP-deﬁcient mutant (respectively, RMA and RMA/S cells). M␾Ps phagocytosed apoptotic RMA cells, and in 앑10–20% of cells, material from biotinylated RMA cells could be detected in the cytosol by microscopy using ﬂuorescent streptavidin. A substantial amount of radiolabeled material could also be recovered from cytosolic factions of M␾Ps incubated with RMA cells labeled with [35S] methionine/cysteine. M␾Ps incubated with apoptotic (but not necrotic58) RMA cells were recognized by the RMA-speciﬁc TCD8⫹ clone. Despite the inability of the TCD8⫹ clone to recognize RMA/S cells, it recognized M␾Ps incubated with the cells, indicating that the antigen was derived either from the source protein or molecular chaperones bearing the peptide. Presentation required contact between the M␾Ps and RMA cells, was blocked by cytochalasin D, and was not detected using TAP⫺/⫺ M␾Ps, all consistent with the cytosolic delivery of phagocytosed material, but, given the possibility of TAP-related differences in the number of PR molecules available for presentation of endosome-derived peptides, falling short of more conclusive evidence (BFA and proteasome inhibitor sensitivity). 57

Also, quantitation may have been affected by differential destruction of FITC-OVA or differences in endosomal pH, because FITC ﬂuorescence is acid sensitive. 58 Whether this was due to absence of M␾s phagocytosis of necrotic cells (anticipated from previous ﬁndings) or to other factors was not reported.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

35

Albert et al. (1998) studied the presentation of inﬂuenza virus (IV) M1 protein from apoptotic cells by human DCP–Cs as determined by the lytic activity of primed TCD8⫹ following 7 days of in vitro culture with DCP–Cs exposed to various antigens. They found that DCP–Cs incubated with apoptotic but not necrotic or live IV-infected cells were able to stimulate M1speciﬁc TCD8⫹. Presentation required the addition of DCP–Cs and occurred with allogeneic IV-infected cells, demonstrating that DCP–Cs presented antigen derived from the apoptotic cells. Crude fractionation of the apoptotic cells revealed that activity was contained with cellular debris and not with material released from the cells in the course of their interaction with IV. They failed to detect presentation of apoptotic cells by peripheral blood monocytes, but whether this was due to their inability to generate complexes from apoptotic cells or from the general inability of these cells to activate TCD8⫹ was not distinguished. The route of DCP–C processing of apoptotic cells was not delineated, and although it was shown that DCP–Cs phagocytosed material from apoptotic cells, it was not reported whether blocking this process with cytochalasin D interfered with presentation.59 Taken together these studies indicate that phagocytosis of apoptotic cells results in the generation of class I peptide complexes from the cells, raising a number of important questions: 1. Does the site of peptide generation and association with class I molecules (endosome vs. cytosol/ER) vary with the properties of the protein antigen (as observed with bacteria and beads)? 2. To what extent is the failure of M␾s and DCP–Cs to process necrotic cells similarly due to decreased or altered phagocytosis,60 proteolytic destruction of the potential antigen incurred as a result of necrosis, or special handling of the antigen during the process of apoptosis (e.g., enhanced creation of peptide–chaperone complexes)? 3. What is the nature of the material from apoptotic cells that is processed by the APC—intact native antigen, intact denatured antigen, or some further intermediate in the determinant generation process? 59 Because cytochalasin D will interfere with the generation of TCD8⫹ to perform this experiment it would have been necessary to use the DCs as target cells or as aldehydeﬁxed stimulators after exposure to the antigen and drug. Indeed, although the authors report the effects of BFA, NH4Cl, and lactacystin on stimulation, it is not clear how this experiment was performed, because inclusion during the culture period would effect TCD8⫹ responses, if not kill the cells outright. 60 Rubartelli et al. (1997) reported that DCB–Cs are able to phagocytose apoptotic bodies but not necrotic cells, whereas M␾s can phagocytose either. Phagocytosis of apoptotic bodies was dissimilar in receptor usage (M␾s could use a more diverse array of receptors) and in the requirement for calcium mobilization (required by DCs, dispensable for M␾s), possibly related to receptor usage.

36

JONATHAN W. YEWDELL et al.

Bellone et al. (1997) make the point that apoptosis is a routine event in organisms, rarely associated with inﬂammation, and question the capacity of cells presenting apoptosed cells for activating naive TCD8⫹, suggesting instead that this may be a preferred route for tolerance generation. This is an excellent idea, testable by immunizing mice with APCs exposed in vitro to apoptotic antigens. C. VIRAL OR VIRUSLIKE ANTIGENS 1. Viruses a. Rekindling the Fire. Inactivated viruses were the ﬁrst deﬁned antigens used in initial studies of exogenous antigen presentation. Subsequent studies with inﬂuenza virus (Yewdell et al., 1988; Vinitsky et al., 1997) and human cytomegalovirus (Riddell et al., 1991) demonstrated that exogenous viral antigen can be processed by the classical pathway following their introduction into the cytosol. By no means, however, do these ﬁndings mean that all, or even most, exogenous viral antigens are processed via this route. As described in the next section, results with Sendai virus suggest that one or more nonclassical processing pathways are involved in presentation. There are at least two reasons why the presentation of other inactivated viruses should be revisited using the powerful tools now available for in vivo and in vitro analysis. First, given the likelihood of noncytosolic processing of some bacterial and viral/virallike antigens (described below), it is probable that some of the viruses will be processed in interesting ways. The variety of receptors bound by viruses makes it likely that some viruses have access to processing compartments that are only inefﬁciently accessed by pinocytosed antigens. Second, inasmuch as numerous vaccines are composed of inactivated viruses and this remains a valid strategy for future vaccines, it behooves us to understand when and how they induce TCD8⫹ responses. b. Return of the Sendai. Sendai virus (SV) occupies a prominent position in the pantheon of exogenous antigens, being the ﬁrst virus shown to elicit TCD8⫹ in vivo and sensitize target cells in vitro when inactivated (Schrader and Edelman, 1977; Koszinowski et al., 1977). SV is a paramyxovirus that attaches to cells via the hemagglutinin/neuraminidase glycoprotein, which binds (and hydrolyses) sialic residues. Viral entry is mediated by the fusion protein, which is active at neutral pH.61 Early ﬁndings indicated that fusion activity was needed for inactivated virus to sensitize target cells (Koszinowski et al., 1977; Sugamura et al., 1978). At the time, it was thought that this reﬂected a requirement for integration of viral glycoproteins into the plasma membrane for their recognition as intact proteins. With the beneﬁt of 20/20 hindsight (and millions of man hours), we know 61

The activities of these two proteins enables the virus to fuse cells; SV was the ﬁrst, and for many years, the best agent for fusing cells.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

37

that this interpretation is incorrect, and it is likely that fusion is required for the delivery of viral proteins to the cytosol. This issue has never been reexamined, however (all of our ﬁndings regarding the requirement for viral fusion in the processing of cytosolic antigens have been made with inﬂuenza virus), and whether or not this occurs, it seems that exogenous SV can be processed via alternative pathways. The ﬁrst inkling of this came from an early study that reported that in vitro stimulation of memory SV-speciﬁc TCD8⫹ with SV preparations rendered incapable of speciﬁcally binding cells required the presence of adherent cells (Koszinowski and Gething, 1980). Further characterization of the properties of SV as an exogenous antigen awaited the efforts of Jondal, Ljunggren, and colleagues. Zhou et al. (1993) reported that when exposed to SV, TAP-deﬁcient T2 cells62 expressing Kb from a transfected gene presented the immunodominant determinant from nucleoprotein (NP324– 51 332) in a BFA-resistant manner, as measured by TCD8⫹-mediated Cr released from target cells. Two other similar enveloped viruses (IV and vesicular stomatitis virus) were not presented by T2 cells (this is expected from the absence of TAP). Although a number of established cell lines were unable to present SV in this manner, the defect in TAP is not related to presentation, because ﬁrst, another TAP-deﬁcient line was unable to presented the virus, and second, T2 cells transfected with TAP genes still presented the virus in a BFA-resistant manner. The failure of some cells, known to be efﬁcient presenters of synthetic peptides, to present SV in the presence of BFA provides good evidence that that presentation is not attributable to peptide contamination of the virus preparation. The nonclassical presentation of SV apparently does not require the function of the two glycoproteins. Virus incubated for 30 min at 56⬚C (shown to compromise glycoprotein function63) or even 100⬚C was presented in vitro, and could prime mice for secondary TCD8⫹ responses to SV64 (Liu et al., 1995, 1997). Extensive studies with IV, which also binds to sialic acids on glycoproteins and glycolipids, indicate that viral glycoproteinmediated attachment to the APC surface enhances class I- and class II62 While on the topic of cell fusion . . . T2 cells were produced by fusing a human T cell line with a 웂-irradiated Epstein–Barr virus transformed cell line (.174) that had been selected for low class I expression. Chromosome 6 in .174 cells lacks 1 MB of the MHC encompassing class II genes, TAP, and LMP genes. Chromosome 6 in T2 cells is exclusively derived from .174 cells. 63 The authors failed to detect activity of the viral glycoproteins, but the assays used were of relatively low sensitivity, and the degree of residual activity was not determined. 64 Priming with 56⬚C treated virus may have been due to residual infectivity. While this seems very unlikely with boiled or autoclaved virus, it is still possible in the absence of controls demonstrating that the response differs qualitatively from that induced by minimal doses of infectious virus.

38

JONATHAN W. YEWDELL et al.

restricted presentation by many orders of magnitude, due to the parallel increase in the rate of internalization (Eisenlohr et al., 1987; Yewdell et al., 1988). If thermally denaturing SV has no deleterious effect on its presentation (which remains to be established, because the efﬁciencies of presenting the various forms of SV were not compared), it would suggest that distinct processing pathways are used by native and denatured SV. As described below, denaturing numerous other soluble antigens enables their BFA/TAP-independent presentation, so this is not a trivial distinction within the SV system. Supporting this possibility, heated SV (unlike native SV) was presented by all cell types tested. These provocative studies suggest that NP324–332 can be generated from SV by at least one nonclassical pathway. Much remains to be learned about this process, including (1) the involvement of viral infectivity, (2) the contributions of endosomal versus cytosolic processing, (3) the role of the viral glycoprotein attachment and fusion activities, and (4) the extent to which NP324–332 represents a special case among NP determinants, determinants from other SV proteins, and determinants from proteins in other membrane viruses. 2. Viruslike Particles Viruslike particles (VLPs) are composed of self-assembling proteins, usually of viral origin. A number of VLPs have been shown to induce TCD8⫹ in vivo. Hepatitis B virus (HBV) surface antigen (S), human immunodeﬁciency virus PR55gag (Wagner et al., 1996), parvovirus VP2 (Sedlik et al., 1997), and yeast transposon proteins (Layton et al., 1996; Bachmann et al., 1996) have all been used alone, or genetically fused with nominal antigens consisting of minimal determinants or longer fragments. This is an extremely promising vaccine approach, well worthy of detailed characterization of the processing pathway utilized.65 To date, however, mechanistic studies on processing and presentation are limited to HBV S particles. HBV S particles can be abundant in the blood of HBV carriers (up to 300 애g/ml); they comprise the viral antigen found by Blumberg in 1963 while investigating human serum ‘‘polymorphism’’ that led to the discovery of HBV.66 They are usually spherical particles 16–25 nm in diameter (rods are also formed) composed predominantly of a single viral gene product (S protein) in a complex with glycolipid. The S particles form in the ER of HBV-infected cells as a normal part of the infectious cycle. For years the HBV vaccine consisted of S particles puriﬁed from the plasma of chronic HBV-infected individuals. The current vaccine is produced in 65 Take note: in a head-to-head comparison of the immunogenicity of peptides and lipopeptides in various adjuvants, recombinant vaccinia virus, plasmid DNA, and VLPs, VLPs were the clear winner in inducing TCD8⫹ responses (Allsopp et al., 1996). 66 And a Nobel prize for Blumberg—a classical example of serendipity in biomedical research.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

39

mammalian cells or yeast transfected with the S gene. The glycolipid particles formed are similar to those naturally produced by HBV-infected cells, but lack the small amounts of other viral gene products present in natural particles. It was originally reported that S particles occasionally induce TCD8⫹ responses in humans, and can be presented by human B cells to S antigen-speciﬁc TCD8⫹ ( Jin et al., 1988; Barnaba et al., 1990). Building on this foundation, Schirmbeck, Reimann, and their colleagues have made a series of intriguing ﬁndings regarding the processing of S particles. Orginally, they reported that immunization of H-2d mice with yeast-derived S particles in the absence of adjuvant primes for a TCD8⫹ response to an S protein determinant restricted by Ld (Schirmbeck et al., 1994b,c). These TCD8⫹ recognize the 12-mer spS28–39. This is atypically long for Ld-associated peptides, which are generally 8 to 10 residues long. Sensitization of target cells with the sp requires a concentration of 5 nM for half-maximal lysis, roughly 1000-fold higher than commonly observed for other Ld-restricted determinants, suggesting that this is not the optimally active peptide, and not the naturally processed peptide, which presumably is smaller. A comparison of the HPLC elution proﬁles of synthetic peptide and naturally processed peptide, which could help to resolve this question, has not been published. It is possible that the properties of the spS28–39 and the natural or optimally presented peptide differ considerably: this consideration hinders evaluation to what extent peptide contamination contributes to the antigenicity of S particle preparations in the experiments described below. Consistent with their immunogenicity, S particles are presented in vitro by cells, as measured by lysis of 51Cr-labeled APCs (Schirmbeck et al., 1995b). Presentation occurs in a TAP-independent, BFA-resistant67 manner, and in contrast to spS28–39, does not occur at 4⬚C. Presentation of the particles at 37⬚C is 앑100-fold-more efﬁcient on a molar basis68 than presentation of the peptide, which points to one or a combination of the following: superefﬁcient generation and loading of the determinant from S particles, destruction of spS28–39 by serum/cell proteases, or the natural peptide being presented much more efﬁciently than spS28–39. Presentation of S particles but not a saturating amount of spS28–39 was blocked by lysosomotropic agents or leupeptin. Leupeptin could be added as late as 30 min after cells were exposed to S particles, suggesting that limiting steps in proteolysis began only at this time. Presentation was resistant to cytochalasin B, and peptide regurgitation was not detected. These latter 67

Although a control establishing that BFA blocked ER export in the target cells was not reported. 68 In terms of the amount of determinants present, which we have taken the liberty to calculate from the values given in the paper.

40

JONATHAN W. YEWDELL et al.

two ﬁndings contrast with the presentation of heat-aggregated S particles,69 whose presentation was similar to that described above for E. coli-produced antigens,70 and which was much less efﬁcient at inducing TCD8⫹ responses in vivo (Schirmbeck et al., 1995a). Thanks to the efforts of the Hansen laboratory, there are well-characterized mAbs that bind to distinct forms of Ld. The mAb 30-5-7 binds preferentially to native Ld, whereas the mAb 64-3-7 binds preferentially to nonnative Ld (Lie et al., 1991). As expected, 30-5-7 blocked presentation of spS28–39 but 64-3-7 had no effect on peptide presentation (Schirmbeck and Reimann, 1996). The story becomes interesting (and more complicated) when the effects of the mAbs on presentation of S particles are examined. Now 64-3-7 blocks presentation, and the mAb need only be present during the ﬁrst 30 min when cells are incubated with S particles—addition after this time having no effect on presentation. By contrast, 30-5-7 is unable to block presentation when present during this window, but is able to block subsequently if added prior to the cytotoxicity assay. mAb 64-3-7 did not affect M␾ presentation of heat-aggregated antigen, consistent with its processing by a distinct pathway, which largely, if not exclusively, produces peptides that bind to cell surface class I molecules, a´ la synthetic peptides. Another interesting feature of S particle presentation is a nearly absolute requirement for exogenous 웁2m (⬎100-fold antigen required for equivalent recognition in its absence) (Schirmbeck et al., 1997). By contrast, the efﬁciency of spS28–39 presentation is not affected by the absence of exogenous 웁2m. Cells lacking endogenous 웁2m expressed very low levels of 643-7⫹ molecules, and were unable to present S particles even in the presence of exogenous 웁2m,71 indicating that presenting molecules derive from molecules that were once associated with endogenous 웁2m. Exogenous 웁2m could be provided by ﬁrst incubating cells at 4⬚C with 웁2m, washing the cells, and then exposing cells at 37⬚C to S particles, indicating that the appropriate PR molecules (or at least their precursors) could be created at the cell surface. Taken together, these extensive experiments are most consistent with the idea that presentation of S particles entails the loading of endosomally generated peptides onto Ld molecules bearing exogenous 웁2m. Because cell surface PR Ld molecules exist without exogenous 웁2m, the implication is that either such molecules are not internalized or are unsuitable for presentation of endosomal peptides due either to unfolding/degradation 69 The S particles were boiled for 1 min; this resulted in the production of aggregates 0.5–1.5 애m in diameter. 70 Another similarity: aggregates were presented by M␾s and not tissue culture cell lines, consistent with a requirement for phagocytosis. 71 Presentation of spS28–39 was also greatly compromised in these cells.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

41

in endosomes or to an inability to return to the cell surface. The blocking experiments indicate that the PR molecule (or its immediate precursor) used in the endosome is 64-3-7⫹, 30-5-7⫺. Such molecules were not detected in association with exogenous 웁2m. 웁2m was shown, however, to be internalized, presumably to endosomes (although this remains to be established), which favors the following scenario: PR molecules capable of presenting endosomally derived antigens are only created in the endosome from free 웁2m binding to 64-3-7⫹ molecules, generating 30-5-7⫹ molecules that are delivered to the plasma membrane. These ﬁndings point to many interesting possibilities and future experiments. 1. It is important to characterize through cell fractionation and microscopy the involvement of endosomes in S particle processing. The acquisition of 30-5-7 reactivity on peptide loading provides the potential handle in these studies to identify directly the site at which loading occurs. Does loading occur in a compartment in which class II loading occurs? If so, are class II accessory proteins involved? 2. Why are S particles so efﬁciently presented? It has been reported that S particles bind to transferrin, which is internalized efﬁciently by cells expressing the transferrin receptor (expressed in high amounts by rapidly dividing cells) (Gagliardi et al., 1994). Does this contribute to the efﬁcient presentation of S particles? 3. What is the determinant naturally processed from S particles? Does it have a special ability to bind Ld molecules in the endosomal environment, or is it unusually well liberated by endosomal proteases? To what extent are other natural or artiﬁcially inserted determinants processed from S particles? 4. Does Ld have an unusual capacity for utilizing this putative endosomal pathway, and if so, is it due to its unusual trafﬁcking to endosomes or stability within endosomes? 5. As with particulate antigens described in the previous section, the same (or similar) determinant generated endosomally (or in phagosomes for heat-aggregated S antigen) is presented by cells expressing endogenous S protein. Presumably, these cells produce S particles, so it is unclear to what extent the determinant is produced endosomally, in the cytosol, or in the secretory pathway. In any event, all the TCD8⫹ used were produced by priming mice with S particles and restimulating with cells expressing endogenous S protein. Would priming and restimulation with S particles result in the generation of TCD8⫹ speciﬁc for determinants uniquely created in endosomes?

42

JONATHAN W. YEWDELL et al.

3. ‘Somes and ‘SCOMs As with VLPs, there are numerous publications describing the capacity of liposome, virosome,72 or immunostimulating complex73 (ISCOM) preparations to prime TCD8⫹ for secondary responses to protein antigens (reviewed in Zhou and Huang, 1998). Fewer papers report the presentation of such preparations in vitro, and precious few have examined the mechanism of presentation, none using the reagents now available (TAP-deﬁcient cells, proteasome inhibitors) that enable reasonably accurate appraisal of the contribution of cytosolic processing to antigen presentation. ISCOM preparations containing full-length proteins appear to be a particularly promising vehicle for immunization, because they have been shown to prime in vivo for TCD8⫹ responses in the absence of adjuvant. The intracellular trafﬁcking of ISCOMs containing biotinylated IV glycoproteins by peritoneal exudate cells (PECs) was studied by EM examination of cell sections labeled with streptavidin–gold (Villacres et al., 1998). Antigen was detected in areas believed to represent cytosol within 5 min of exposing cells to ISCOMs. Cell fractionation studies were consistent with this ﬁnding, but lacked appropriate controls for trafﬁcking of substances known to remain in endosomes, so provide limited support for the EM ﬁndings. On the face of it, these ﬁndings indicate that ISCOMs have the ability to penetrate cells rapidly. This is an important conclusion that should be relatively easy to conﬁrm.74 It is also important to examine whether penetration by ISCOMs requires endocytosis and results in a general loss in membrane integrity such that proteins in the medium surrounding the site of entry gain access to the cytosol. pH-sensitive liposomes (i.e., liposomes that become destabilized and more fusogenic in acidiﬁed endosomes) have been reported to elicit better TCD8⫹ responses to encapsulated antigens compared to pH-insensitive liposomes, and to better sensitize target cells for TCD8⫹-mediated lysis (Nair et al., 1992; Collins et al., 1992). Using immuno-EM, Zhou et al. (1994) demonstrated the presence of OVA in the cytosol of a mouse T cell line incubated with acid-sensitive OVA-loaded liposomes. Lower amounts of OVA were detected in cells incubated with acid-insensitive liposomes: some of this was likely due to decreased penetration (앑5- to 6-fold decrease), some to decreased binding/uptake (앑4- to 5-fold decrease). The major mechanism of delivery is presumed to be fusion of liposome with 72

These are liposomes consisting of viral glycoproteins and lipids. ISCOM are (in the words of Morein, their creator) ‘‘cage-like structures of 30– nm in diameter, composed of Quil A, cholesterol, phosphatidyl choline, and protein’’ (Villacres et al., 1998). Quil A is a partially puriﬁed saponin, a group of plant glycosides. 74 It should be quite easy, for example, to visualize in live cells the delivery of a ﬂuorescent protein targeted to the nucleus. 73

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

43

cellular membranes, which could be visualized; on the other hand, acidsensitive liposomes also release more material into the endosome, which could then egress to the cytosol at (or distal to) sites of liposome fusion. Some of the cytosolic OVA remained associated with liposomal fragments, indicating that penetration results not from a seamless fusion process, but a quite messy one. Presentation of OVA was shown to be BFA sensitive [also reported by Martin et al. (1993)], and NH4Cl sensitive. Liposomes were internalized at 18⬚C, but antigen presentation was not detected at this temperature. After 2 hr of internalization at 18⬚C, presentation was ﬁrst detected as early as 30 min after shifting to 37⬚C, and peaked at some point over the next 90 min. This effect was independent of the dose used over a nonsaturating range, and therefore establishes the maximal rate of presentation as occurring between 30 and 120 min post exposure.75 By acid stripping cell surface complexes, it was determined that the pool of OVA delivered to the cell capable of serving as a source of peptides was depleted within 3 hr of loading.76 This work provides strong evidence against sensitization by peptides, but deﬁnitive evidence for cytosolic (versus endosomal) peptide generation requires demonstration that presentation is TAP dependent or, better yet, proteasome dependent. Despite their enhanced antigenicity, pH-sensitive liposomes demonstrate little improvement in immunogenicity relative to pH-insensitive liposomes. This suggests either that in vivo presentation is mediated by a pAPC insensitive to the differences in the liposomes, or that liposomes are modiﬁed in vivo prior to their arrival at pAPCs, such that differences between the two are nulliﬁed. The nature of the pAPC involved in the presentation of OVA contained by liposomes and virosomes was examined by treating mice with clondronate-liposomes. This inhibited responses to liposomes but not inﬂuenza virosomes (Wijburg et al., 1998). The authors concluded that M␾s were solely responsible for presentation of liposomes but that other APCs were capable of presenting virosomes. This interpretation hinges, of course, on the speciﬁcity of clondronate-liposomes for M␾s, which, as discussed above, is questionable. It seems more likely that this result reﬂects the similar handling of OVA- and clondronate-liposomes by pAPCs, and the ability of the viral hemagglutinin to target virosomes to pAPCs (possibly even M␾s) that inefﬁciently internalize liposomes. The ﬁndings are still interesting, however, because they suggest that targeting liposomes to the cell surface can recruit a different type of pAPC. 75 This experiment is noteworthy for its consideration of antigen dose on the rate of presentation, which is usually performed with a single dose. 76 Also noteworthy as the only experiment to examine the half-life of processible antigen in the cytosol; this is potentially a means for determining whether there exist in the cytosol subsets of exogenous antigens that are selectively processed.

44

JONATHAN W. YEWDELL et al.

D. ‘‘SOLUBLE’’ AND AGGREGATED ANTIGENS 1. Native Proteins Although it was reported 20 years ago that puriﬁed proteins are capable of eliciting TCD8⫹ responses in vivo, the ﬁrst studies to address the mechanism of spontaneous soluble antigen presentation in vitro appeared in 1990. Carbone and Bevan (1990) reported that transfer of splenocytes exposed to OVA in vitro primed OVA-speciﬁc TCD8⫹. Rock and colleagues found the likely basis for this observation: splenocytes (but not a B cell line) could present OVA to a T hybridoma speciﬁc for Ova257–264, as measured by IL-2 release (Rock et al., 1990). The efﬁciency of OVA presentation was reduced more than 16-fold by either aldehyde ﬁxation of cells or treating cells with azide or ricin, whereas peptide presentation was decreased 앑3-fold, indicating that protein synthesis and possibly other cellular processes were required for presentation. Extracellular processing of the protein could not be detected, nor was what would be later be termed peptide regurgitation. Presentation was limited to a small subset of cells expressing class II molecules, small enough, in fact, to preclude detection by 51Cr-release assay, which is rather dicey when the percentage of presenting cells is ⬍10% (Rock et al., 1993). These studies indicated that pAPCs had the capacity to generate determinants from soluble proteins, at least when the proteins were provided to cells under conditions of vast excess (milligram/milliliter levels were required). Insight into this process was provided by Norbury, Watts, and their colleagues, who found that macropinocytosis by M␾B–Cs (Norbury et al., 1995) or DCB–Cs (Norbury et al., 1997) resulted in the delivery of exogenous proteins to the cytosol (visualized microscopically) and the presentation of OVA to an OVA-speciﬁc TCD8⫹ clone. Macropinocytosis by M␾B–Cs occurs at a low rate, and cytosolic delivery and antigen presentation required treatment with phorbol myristic acetate (PMA) to enhance macropinocytosis. By contrast, DCB–Cs (thought to be similar to peripheral DCs) have a high rate of macropinocytosis, and soluble proteins are spontaneously delivered to the cytosol and presented to TCD8⫹, more efﬁciently and rapidly than even PMA-treated M␾B–Cs. Blocking macropinocytosis by a number of manipulations prevented both cytosolic delivery and antigen presentation. Antigen presentation was shown to be TAP dependent and gelonin, BFA, and proteasome inhibitor sensitive, all the hallmarks of cytosolic processing. Chloroquine (which does not affect the rate of macropinocytosis) enhanced, rather than inhibited, antigen presentation, suggesting that endosomal proteolysis is not necessary for peptide generation, and may even be detrimental. Transfer of PMA-treated, OVA-pulsed DCB–Cs (PMA increases the efﬁciency of OVA presentation 앑2-fold) primed TCD8⫹

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

45

for secondary in vitro responses, but whether this resulted from direct stimulation of TCD8⫹ or cross-priming was not examined. The enhanced ability of DCB–Cs relative to M␾B–Cs to present exogenous OVA to TCD8⫹ was conﬁrmed by Mitchell et al. (1998), who extended these ﬁndings by demonstrating that ‘‘maturation’’ of DCB–Cs to a state thought to represent DCs that migrate from the periphery to lymph nodes resulted in a decreased ability to process and present OVA. The cells also presented spOva257–264 somewhat less efﬁciently, however, and it is not clear to what extent the differences observed are due to alterations in antigen processing, numbers of PR cell surface molecules (for spOva257–264 presentation), or general ability to present antigens. Also, potential differences in pinocytosis rates among cell types were not examined. Brossart et al. (1997) found that DCB–Cs, DCS–Cs, and M␾B–Cs incubated for a prolonged period (16 hr) with high concentrations of OVA (2 mg/ ml) were lysed in a TAP-dependent manner by Ova257–264-speciﬁc TCD8⫹. Other than observing that a higher percentage of DCS–Cs was lysed, the efﬁciency of M␾B–C versus DCS–C presentation was not examined further. Exposure of DCs to TNF-움 or LPS, shown by others to decrease rates of pinocytosis in human DCs, resulted in diminished presentation, an effect that could be reversed by interferon 웂 (IFN-웂), which, when given alone, enhanced the rates of presentation. The mechanisms responsible for these phenomena remain to be examined, but these ﬁndings, in conjunction with the studies discussed in the preceding paragraphs, point to the important conclusion that the capacity of DCs to process soluble proteins can vary widely, depending on the maturational state of the cells, which is controlled by cytokine exposure and other factors. The effects of the cytokines are no doubt complicated because they are known to inﬂuence not just pinocytosis rates, but also the levels of class I molecules, processing components, and costimulatory molecules. Weighing the contributions of these alterations on a individual basis will be no mean feat. Probably the most important ﬁnding of these studies is the identiﬁcation of a mechanism for the cytosolic presentation of exogenous proteins. Several key questions remain regarding the cytosolic delivery of exogenous antigens. First, is macropinocytosis a property of DCs present at the antigen-containing site and does it contribute to antigen presentation? Second, is macropinocytosis qualitatively distinct from other forms of uptake regarding the transport of internalized proteins to the cytosol—i.e., in comparison to endocytosis/phagocytosis, is there an increase in the ratio of cytosol:endosome delivery? Third, does macropinocytosis also increase the ability of pAPCs to process antigens endosomally? Question one poses a formidable challenge, but questions two and three are amenable to direct experimental manipulation.

46

JONATHAN W. YEWDELL et al.

2. Special Cases: Receptor-Mediated Uptake and Penetrating Antigens a. Natural Antigens. One of the striking features of the presentation of pinocytosed soluble proteins is its inefﬁciency: milligram/milliliter quantities of antigen are required. Improvements in efﬁciency would seem to be required for this process to be relevant to in vivo responses to pathogens. These could come in two areas: enhanced delivery to a compartment with access to the cytosol, or enhanced delivery to the cytosol through the properties of the antigen. In the former category, Ke and Kapp (1996) reported that a B cell line producing membrane IgM speciﬁc for TNP presented TNP-conjugated OVA or insulin to appropriate TCD8⫹ at least 10to 30-fold more efﬁciently compared to unmodiﬁed antigen. Presentation required speciﬁc binding to the membrane antibody, as indicated by blocking by TNP-BSA or antiidiotypic antibody. TNP-mediated presentation was blocked by BFA or a proteasome inhibitor, consistent with its processing via the cytosol. It was not affected by chloroquine (which blocked presentation of OVA to TCD4⫹, indicating that normal endosomal acidiﬁcation was not required for delivery to the cytosol. These ﬁndings raise several interesting scenarios. First, if normal B cells behave similarly, they could potentially interact with TCD8⫹ in an antigenspeciﬁc manner, exchanging signals that inﬂuence both parties in a positive or negative manner.77 Second, this will not happen if antigen binds to the same or competing antibodies in serum before the antigen can be acquired by speciﬁc B cells. Third, if serum antibody does bind antigen, the complex should be internalized via Fc receptors on pAPCs and processed into class I-binding peptides. This will be a somewhat less efﬁcient way of producing complexes compared to Ig-mediated internalization, because numerous antigens will be internalized by a given APC via Fc receptors. In the cases of viruses or bacteria, however, which possess antigens in high copy number, this process could still deliver signiﬁcant amounts of antigen to endosomes.78 These ﬁndings also demonstrate that endocytois, like macropinocytosis and phagocytosis, results in the delivery of endocytosed antigen to the cytosol. The crucial parameter in weighing the in vivo importance of these processes is the efﬁciency of cytosol delivery relative to the amount of material internalized. Possibly, synergistic effects may occur at the intersec77

TCD8⫹ elimination of speciﬁc B cells was originally suggested by Barnaba et al. (1990) to explain the absence of anti-S particle antibodies in some patients with chronic HBV, based on their ﬁndings that HBV-speciﬁc B cells demonstrated greatly enhanced presentation of S Ag to TCD8⫹ relative to B cells speciﬁc for other antigens. 78 This could explain prior ﬁndings that IV incubated with a completely neutralizing dose of antibody still elicited TCD8⫹ responses (Greenspan et al., 1982).

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

47

tion of these pathways that could be put to use by speciﬁcally targeting antigens to receptors on DCs that are internalized into macropinosomes.79 b. Engineered Antigens. In 1990 we suggested that certain toxins and viral proteins capable of penetrating cellular membranes to the cytosol could serve as vehicles for antigen delivery to the class I processing system (Yewdell and Bennink, 1990). This approach has been attempted several times, but the evidence supporting cytosolic processing of the antigens is limited, and the vaccine potential of this strategy remains to be demonstrated.80 Ulmer and co-workers originally reported that Pseudomonas exotoxin was able to deliver IV peptides to the cytosolic processing system (Donnelly et al., 1993). Subsequently, however, these authors recanted: on more detailed examination, presentation was found to occur in a TAP-independent and BFA-resistant manner (Ulmer et al., 1994), i.e., results most consistent with noncytosolic processing or peptide contamination of toxinantigen preparations. Lee et al. (1998) inserted a determinant into the Bfragment of Shiga toxin and found it was presented to appropriate TCD8⫹ by either human EBV-transformed B lymphoblastoid cells or DCP–Cs. Presentation of the protein was blocked by paraformaldehyde ﬁxation of cells or BFA, which failed to affect peptide presentation. (As in other studies, only a single supersaturating peptide concentration was tested.) These ﬁndings suggest that intracellular trafﬁcking of the protein was required for presentation, but provide little further insight into the processing mechanism. Probably the best evidence for cytosolic processing of toxin-based antigen comes from Goletz et al. (1997), who reported that anthrax toxin is capable of efﬁciently sensitizing cells for lysis by gp120-speciﬁc TCD8⫹ —indeed, that it was 100- to 1000-fold more efﬁcient than the optimally binding synthetic peptide.81 Two ﬁndings suggested that this was due to cytosolic delivery of the protein: ﬁrst, a mutation that blocks toxin translocation (but not cell binding) prevented antigen presentation; second, presentation was partially blocked by lactacystin. 79

It is also possible that this would interfere with cytosolic delivery if release of antigen from the macropinosome membrane was inefﬁcient. 80 In addition to what has been published, there are probably numerous failures whose descriptions will be limited for eternity to the tear-stained pages of laboratory notebooks. This melancholy statement is based on personal experience. We have tried this approach in collaborative studies with ﬁve different outstanding laboratories using different carrier proteins with reported cytosol-penetrating activities, crashing and burning each time. So much for prophecies (at least self-fulﬁlling ones). 81 They also mention ‘‘preliminary results’’ that the toxin is able to induce ‘‘vigorous’’ TCD8⫹ following a single subcutaneous injection.

48

JONATHAN W. YEWDELL et al.

Saron et al. (1997) found that a Bordetella pertussis toxin inactivated by the insertion of a determinant from lymphocytic choriomeningitis virus (LCMV) into its catalytic domain elicited virus-speciﬁc TCD8⫹ cells and protected mice against lethal LCMV challenge in a TCD8⫹ dependent manner. Unlike the toxins discussed above, which are thought to make their way to the ER, and then egress to the cytosol, the B. pertussis toxin is believed to translocate directly across the plasma membrane. In vitro studies on the mechanism of toxin processing have not been reported, but it was shown that mutations that interfere with its capacity to penetrate cells reduce its immunogenicity. Making use of a different evolutionary solution for polypeptide translocation, Schutze-Redelmeier et al. (1996) described fusion proteins consisting of the 60-residue DNA binding region of a Drosophila DNA-binding protein genetically fused at its COOH terminus to deﬁned antigenic peptides.82 This polypeptide has the remarkable ability to penetrate the plasma membrane efﬁciently, even at 4⬚C. Penetration is nonsaturable, consistent with its occurrence in a receptor-independent manner. Presentation of deﬁned determinants to TCD8⫹ in vitro occurred in a BFA-sensitive manner, whereas presentation of a limiting amount of the corresponding synthetic peptide was unaffected. The fusion protein was only immunogenic in vivo, however, when injected in sodium dodecyl sulfate (SDS). The authors attribute this SDS effect to enhanced binding of the protein to cells or protection from serum proteases. Equally plausible is that SDS enhances immunogenicity by denaturing the protein, which, as described below, is known to ehnance the immunogenicity of protein antigens. c. Facilitated Delivery. Another strategy for cytosolic delivery of protein antigens is the use of LLO, the L. monocytogenes gene product that forms pores in eukaryotic cell membranes. Immunization of mice with 1 애g of either IV NP, OVA, or 웁-galactosidase with 5 애g LLO in incomplete Freund’s adjuvant was sufﬁcient to induce TCD8⫹ responses easily detected ex vivo by lysis of appropriate target cells83 (Darji et al., 1997a). LLO was shown to enable presentation of each of the proteins in vitro; presentation 82

Only 16 residues are required for the translocation activity; see Derossi et al. (1998) for a review of this intriguing protein. 83 It is much more difﬁcult to achieve activation of TCD8⫹ sufﬁcient to measure ex vivo responses than to prime for secondary responses, so this is quite an impressive response. Although the goal of vaccination is the generation of memory TCD8⫹ such strong primary responses are almost always (we throw in ‘‘almost’’ just to be careful; based on our experience, noi qualiﬁcation is necessary) associated with generation of vigorous memory responses. The authors noted that LLO is biologically active, incducing cytokines and modifying signal transduction, and this could be a signiﬁcant contributor to the immunogenicity of the LLO–antigen mixtue.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

49

was TAP dependent, BFA sensitive, and was blocked by LLO-speciﬁc antibodies that prevent pore formation but do not block LLO binding to cells. Presentation could be achieved when cells were incubated with LLO on ice, washed, and then exposed to NP within 30 min during incubation at 37⬚C. This indicates, ﬁrst, that penetration through the plasma membrane occurred, and second, that the pores were repaired between 30 and 60 min after exposure to LLO.84 The strength of this approach for vaccine use is that it requires only mixing the vehicle with the target antigen, which can be any protein that is sufﬁciently small to enter through LLO-produced pores. It may, however, be limited to a single immunization, if antibodies are induced with LLOneutralizing activity. 3. Denatured/Aggregated Proteins—The Day the Dogma Died85 The coup de graˆce to the dogma that proteins are inefﬁcient TCD8⫹ immunogens comes from studies with denatured antigens. Two groups found that addition of SDS to proteins increased their immunogenicity (Weidt et al., 1994; Schirmbeck et al., 1994a). The likely explanation for this SDS effect was provided by Martinez-Kinader et al. (1995), who, in trying to understand why some batches of ‘‘native’’ OVA primed mice for memory TCD8⫹ responses following footpad injection in the absence of adjuvant, found that nonimmunogenic batches could be rendered immunogenic by a 3-min treatment at 95⬚C, or by injection with SDS.86 The common thread, of course, is that both treatments denature proteins. Permanent T cell lines were able to present such heat-denatured OVA (OVAHD ) in a TAP-independent, BFA-resistant manner. Presentation of OVAHD was blocked by NH4Cl or vinblastine and did not occur at 4⬚C, all suggestive of requirement for endocytosis. Although these treatments did not affect presentation of spOva257–264, the peptide was used in vast excess, which weakens its value as a control to rule out the presence of contaminating antigenic peptides in OVAHD preparations. Similar ﬁndings were reported by Liu et al. (1997), who characterized presentation of heat-denatured SV (SVHD ) by splenocytes, as determined by their capacity to activate memory virus-speciﬁc TCD8⫹ in vitro. Presentation of SVHD was TAP independent, BFA resistant, and chloroquine and cytochalasin D sensitive, i.e., all the hallmarks of endosomal presentation. Injection of SVHD-sensitized wild-type or TAP⫺/⫺ splenocytes primed naive mice for secondary SV-speciﬁc TCD8⫹ responses. This was not attributable 84 Pores remained open for at least 2 hr if cells were incubated on ice. Translocation of NP occurred on ice as well, indicating that diffusion is sufﬁcient for cytosolic loading. 85 Apologies to Don McClean and Buddy Holly. 86 The upshot of this experiment is that claimants of the immunogenicity of native proteins bear the burden of proof that activity is not due to denatured protein in the preparation.

50

JONATHAN W. YEWDELL et al.

to cross-priming because priming was not obtained with either allogeneic splenocytes or 웁2m⫺/⫺ splenocytes (which do not present antigen in vitro). Importantly, this experiment provides the clearest demonstration to date that endosomally processed antigen can activate naive TCD8⫹. The failure to observe cross-priming in this last experiment is notable. There are two explanations, one trivial, the other extremely important. First, it may simply be dose related, i.e., the amount of antigen in the cells is below the threshold needed for cross-priming. Second, it may indicate that cross-priming requires antigen to be delivered to the cytosol, which would mean that cross-priming requires the metabolic modiﬁcation of antigen-perhaps the generation of antigen-derived peptides complexed with chaperones (discussed below), perhaps other things (just what, we leave to the reader’s imagination). The immunogenicity of heat-denatured proteins appears to be the rule rather than the exception. Speidel et al. (1997) reported that boiling enhanced the capacity of OVA, 웁-galactosidase, and human papilloma virus (HPV) E7 protein to prime mice for secondary in vitro TCD8⫹ responses. Boiled or even autoclaved IV was capable of priming for secondary responses, and cross-priming to miH antigens was achieved using boiled spleen cells. Following centrifugation of OVAHD for 1 hr at 160,000 g, immunogenic material was present in the aggregated material in the pellet. The ability of OVAHD to elicit TCD8⫹ was reduced by mixing with alum, which greatly enhanced antibody responses.87 4. Proteins of Unknown Nature in Cellular Sonicates The immunogenicity of cellular sonicates was ﬁrst explored by Staerz and colleagues (Debrick et al., 1991), who, in a groundbreaking study, found that immunization with sonicated splenocytes exposed to OVA induced OVA-speciﬁc memory TCD8⫹. In vivo priming was abrogated by carrageenan treatment, and was recovered by injection of M␾P–Cs, demonstrating the involvement of phagocytic cells in presentation of antigen, presumably acting as afAPCs. The immunogenicity of cellular sonicates was further examined by Bachmann et al. (1994), who were able to induce memory antiviral TCD8⫹ responses to by immunizing mice with crude sonicates from insect cells infected with baculoviruses expressing LCMV glycoprotein (G), or NP, or vesicular stomatitis virus (VSV) nucleocapsid protein (N). Boiled sonicates were still immunogenic. Following centrifugation of the sonicates for 2 hr 87

It is not clear to what extent the antibody responses measured responses to native versus denatured OVA—an important consideration for understanding the mechanism of alum inhibition. This may also impact the ability to boost TCD8⫹ response by booster immunization with heat-denatured antigens.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

51

at 40,000 g, immunization with an equivalent amount of nominal antigen in the supernatant and pellet resulted in more efﬁcient TCD8⫹ responses for pelleted material. The immunogenicity of the supernatant could be enhanced by addition of sonicates of either insect cells or splenocytes, indicating that cell debris had an adjuvant effect. In vivo antibody-mediated depletion of TCD4⫹ did not impair TCD8⫹ responses to sonicates, demonstrating that their ability to enhance TCD8⫹ responses occurred independently of their induction of TCD4⫹ responses. Experiments to control for peptide contamination of N preparations were well designed, but were based on the failure of spN49–62 to induce TCD8⫹ responses. Given that the naturally processed peptide (spN52–59) is expected to bind with much higher afﬁnity to class I molecules, and that this or another peptide that is signiﬁcantly more antigenic than spN49–62 may present in the preparation, the contribution of contaminating peptides to immunogenicity of the insect cell preparations must be considered uncertain. Subsequently, Bachmann et al. (1995) reported that M␾p–?s, but not EL4 cells (a permanent T cell line), pulsed with insect cell sonicates and then washed were able to restimulate primed TCD8⫹ in vitro over the course of a 5-day stimulation. Presentation occurred in a TAP-independent manner, and did not involve detectable peptide regurgitation. Evidence against antigenic peptide contamination of preparations was the loss of activity by ﬁltration through a 30-kDa membrane ﬁlter, and the stimulation of TCD8⫹ by EL4 cells exposed to synthetic peptide.88 The mechanism of presentation was not further characterized. A DC line believed to be similar to immature DCs was also able to present antigens in insect debris in this system (Bachmann et al., 1996). These ﬁndings are difﬁcult to interpret mechanistically because the native/denatured status of the nominal antigen is uncertain, and the adjuvant effect of cell debris could result from many factors.89 The latter is, of course, well worth pursuing, as it could point to mechanisms of recruitment of and encouragement of APCs specialized in antigen uptake and processing. E. QUANTUM THEORY OF EXOGENOUS ANTIGEN PRESENTATION Summarizing the results from pages 36 to 51, four different categories of exogenous antigens, viruses/VLPs, liposomes, ISCOMs, and denatured proteins, are capable of routinely inducing TCD8⫹ responses to diverse determinants in numerous contexts. The conformation of the carrier pro88 This control is valid only if EL4 cells required equal or less peptide than M␾s to attain an equivalent amount of stimulation. 89 Some possibilities: the debris (1) denatures the nominal antigen, (2) enables its delivery to the cytosol via effects of lipids or other amphipathic material, (3) emits Matzingeric danger signals that induce the right sort of inﬂammation. . . .

52

JONATHAN W. YEWDELL et al.

teins varies over the entire spectrum from native to denatured and from monomeric (or reasonably so) to aggregated/polymeric, and their capacity for binding cells and penetrating membranes varies considerably, suggesting that immunogenicity is not due to a shared ability to access special processing pathways. The one consistent feature of these preparations is that endocytosis or phagocytosis of the smallest quanta of the vaccine (e.g., an individual virosome or liposome) results in the delivery of hundreds to thousands of copies of the determinant to an individual pAPC. We propose that the immunogenicity of these preparations results from this high quantum number, i.e., endocytosis by pAPCs is limiting in vivo so that following administration of soluble antigens, no individual pAPC internalizes enough antigen to generate the threshold number of complexes needed to activate naive TCD8⫹. This is consistent with the ﬁnding that soluble antigens are immunogenic if incubated at high concentrations with pAPCs—conditions that we would argue result in much greater uptake by any individual cells than can be achieved in vivo. This theory makes several testable predictions; the latter two, if true, can be applied to enhance vaccine immunogenicity. 1. Multimeric preparations should have a much steeper dose–response curve compared to soluble antigens when pAPCs incubated in vitro with antigen are tested for their immunogenicity. 2. Alternative methods of increasing the quantum number of antigen preparations90 should result in enhanced immunogenicity. 3. Immunogenic potency should increase as the quantum number increases—at least until the sheer size of the complex interferes with its access to pAPCs. F. MOLECULAR CHAPERONES: THE SRIVASTAVA SAGA In 1986 Ullrich et al. (1986) and Srivastava et al. (1986) reported that immunization with molecular chaperones (later revealed to be HSP90 and gp96, respectively) puriﬁed from tumor cells induced by chemical mutagenesis protected mice from challenge with the autologous tumor. Against prevailing wisdom, Srivastava and colleagues persisted in characterizing this phenomenon, ultimately with spectacular success (reviewed in Srivastava et al., 1998). gp96 is an abundant constituent of the ER, expressed probably in all cells in the body. It is highly related in sequence to HSP90, its counterpart in the cytosol/nucleus, and both have been shown to function as molecular chaperones (Nicchitta, 1998). gp96 was reported to be tightly bound in 90

For example, introducing chemical cross-links; avidin cross-linking of multibiotinylated antigen.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

53

approximately 1 : 1 ratio with peptides ranging from 400 to 2000 Da, and to possess ATPase activity (Li and Srivastava, 1993). Another report, however, strongly suggests that this latter property is due to contamination of gp96 preparations with a potent cellular ATPase (Wearsch and Nicchitta, 1997). HSP70 puriﬁed from tumor cells was also found to provide protection against tumor challenge. In this case, taking advantage of the ability of ATP to release peptides from HSP70, it was shown that this resulted in the loss of protective activity (Udono and Srivastava, 1993). gp96-induced tumor rejection was reported to be due to induction of TCD8⫹ speciﬁc for undeﬁned determinants (Udono et al., 1994). Using complexes of gp96 with one of seven different deﬁned determinants, or HSP70 with one of two deﬁned determinants (generated by in vitro loading of puriﬁed protein with peptides), speciﬁc priming for secondary in vitro responses to the corresponding determinants was attained (Blachere et al., 1997). Mice were immunized twice with 20–50 애g of protein (containing an estimated 앑2 ng peptide) with no adjuvant. Immunization following the same protocol with 10 애g of peptide in the absence of chaperones gave no detectable response, demonstrating the potent adjuvant effect of the chaperones. As shown for one of the peptides, this effect required tight association of the chaperones with the peptide, because simply mixing peptide with chaperone and then immunizing was not sufﬁcient for priming.91 Similarly, Ciupitu et al. (1998) found that HSP70 loaded with a peptide from LCMV was able to accelerate TCD8⫹ responses following LCMV infection and accelerate viral clearance. Notably, distinct from the Srivastava laboratory, in which HSP70 puriﬁed from mouse tissues or tumors was used, human HSP70 produced in bacteria was used, demonstrating that the adjuvant effect of HSP70 is probably independent of any speciﬁc contaminant that copuriﬁes with the molecules. The ability of gp96 to prime for TCD8⫹ responses to antigens expressed by the source cell was conﬁrmed by Arnold et al. (1995), who showed that gp96 puriﬁed from cells expressing 웁-galactosidase and miH antigens was able to prime for TCD8⫹ responses to these antigens. Biochemical evidence for the binding of deﬁned antigenic peptides of gp96 was provided by the demonstration by Nieland et al. (1996) that gp96 derived from VSVinfected cells copuriﬁed with the immunodominant viral determinant. In both of these studies, the presence of gp96-associated peptides occurred independently of the expression of the restricting class I molecule, conﬁrming earlier predictions (Srivastava et al., 1994). 91 Loading does not occur spontaneously; for gp96, loading requires either high salt concentration or heating to 50–60⬚C; for HSP70, ADP and Mg2⫹ are required. Even under these ‘‘optimal’’ conditions, loading seems to be limited to 앑1% of molecules.

54

JONATHAN W. YEWDELL et al.

There is but a single report regarding the mechanism of chaperonemediated presentation. Suto and Srivastava (1995) used gp96 puriﬁed from transfected cells expressing VSV N protein or gp96 loaded in vitro with N-derived peptides.92 Half-maximal presentation by M␾P–Os was achieved by incubating cells with 앑2 애g endogenously loaded gp96 for as little as 2 hr at 37⬚C. Only about a third of the cells expressed sufﬁcient numbers of complexes to by lysed by N-speciﬁc TCD8⫹. A number of permanent cell lines were unable to present the same material. M␾P–O presentation was blocked by BFA, or by depleting cellular ATP levels. Although presentation of spN52–59 was not affected by these treatments, the peptide was used in 200-fold excess of the concentration needed to achieve the same presentation as the gp96 preparation, so it is not clear whether the mechanisms used are distinct. Using gp96 loaded in vitro with the naturally processed peptide, Suto and Srivastava (1995) found that 앑300-fold less gp96complexed peptide was required to achieve the same level of target cell sensitization as compared to the synthetic peptide alone (after accounting for the poor efﬁciency of the in vitro loading process of gp96). Remarkably, extending the peptide by 12 amino acids from the natural sequence (to what extent residues are added to each end is not disclosed) did not decrease the efﬁciency of target cell sensitization.93 The extended peptide alone could not sensitize cells for lysis, which argues that gp96 has a qualitative effect on peptide loading, and does not, for example, simply pony up to class I molecules and discharge its cargo. Although these ﬁndings suggest that there is a speciﬁc mechanism for handling the gp96chaperoned peptides,94 further characterization is clearly required. These ﬁndings have raised considerable excitement for academicians and clinicians alike. On the academic side, the MHC independence of chaperone immunogenicity is exactly what is expected for the ‘‘crosspriming factor,’’ as is the incredible potency of the chaperones at eliciting TCD8⫹ responses. Practically speaking, there is great interest in using chaperones as vaccine vehicles for deﬁned determinants or to elicit TCD8⫹ speciﬁc for malignancies when sufﬁcient numbers of cells are available to obtain gp96 for autoimmunization, particularly following the demonstration of 92

The cells used also expressed the restricting class I molecule, heightening concerns regarding the presence of class I-derived peptides in the gp96 preparation: on the other hand, if the peptide is bound to gp96 for these experiments, whether it was originally derived from class I molecules does not matter much. 93 Its surprising that extending the peptide had no discernible effect (negative or positive). The fact that the natural octomeric peptide was isolated from gp96 by Nieland et al. (1996) suggests that the 20-mer binds to gp96 by the core natural peptide and the ﬂanking sequences are removed by serum or cellular proteases. 94 That is, chaperone receptors on the surface of pAPCs—a quarry that is being avidly pursued by a number of laboratories.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

55

some efﬁcacy against preexisting animal tumors, the gold standard in evaluating tumor vaccines (Tamura et al., 1997).95 The ﬁndings also raise a number of questions.96 1. To what extent are the adjuvant effects related to contaminants (such as LPS) introduced during the puriﬁcation and preparation process? 2. What pAPCs present chaperone-protected peptides following immunization with chaperone–peptide preparations? 3. How do the cells transfer peptides from the chaperones to class I molecules, and how many complexes are generated on the APC?97 Are there cell surface receptors for chaperones? Where is the peptide released—the plasma membrane, endosomes, the cytosol? 4. Which, if any, of the molecular chaperones with adjuvant activity participate in immune responses to infectious or tumor antigens, and under what circumstances? 5. Does every molecular chaperone that binds an antigenic peptide (or precursor) exhibit adjuvant activity,98 or, in addition to peptide protection, are special properties required? Possibly relating to this last point, Suzue et al. (1997) reported that a bacterially synthesized fusion protein consisting of mycoabacterial HSP70 with OVA residues 161–276 attached to the amino terminus was able to prime in the absence of adjuvant for OVA-speciﬁc TCD8⫹ responses, and for protection against an OVA-expressing tumor. The adjuvant effect of HSP70 requires covalent attachment: a mixture of HSP70 with OVA161–276 was not immunogenic. This ﬁnding is potentially revealing regarding the mechanism of the chaperone adjuvant effect, because it suggests that the mechanism of peptide binding to the chaperone is irrelevant, and rather that it is the 95

Just to throw a little rain on the parade: the amount of protein required to obtain responses is impressively small, but the responses themselves generally seem much weaker than are induced by infectious vectors. It will be important to quantitate the generation of memory TCD8⫹ using tetramers or intracellular cytokine staining. Also, it remains to be seen whether chaperones (high-tech, expensive, limited to a few quaternary-care cancer centers) are more efﬁcient than boiled tumor cells (cheapest, lowest-possible-tech) at eliciting TCD8⫹ against tumor-speciﬁc antigens. On the other hand, boiled cells have yet to be shown to inﬂuence the rejection of preexisting tumors. Time will tell. 96 A key nagging question: if chaperones bind to antigenic peptides in a class I-independent manner, why are peptides not isolated from cells lacking the appropriate class I molecule when those cells are homogenized in acid conditions, which should elute peptides from chaperones. 97 Do not be surprised if the number turns out to be very small—even a single copy of the right complex on the right APC may be sufﬁcient to trigger a response. 98 As mentioned above, studies of binding of TAP-transported radiolabeled peptides reveal that in addition to gp96, at least four other ER proteins bind to peptides.

56

JONATHAN W. YEWDELL et al.

cellular handling of the chaperone that matters. Although the mycobacterial HSP70 is closely related to mouse HSP70, possibly handled in a different manner. Moreover, it is worth noting that the amount of fusion protein used in these experiments (120 pmol) is similar to the amount of OVAHD (also used in the absence of adjuvant) required to induce a TCD8⫹ response, so the immunogenicity of the fusion protein is not so potent as to require a dedicated, specialized mechanism for its explanation. G. EXOGENOUS PRESENTATION OF ENDOGENOUS ANTIGENS in Vivo 1. Pros and Cons The studies we have reviewed to this juncture deal with the presentation of exogenous protein antigens. Although leaning toward the ‘‘what can happen,’’ the ﬁndings are important in their own right, particularly with regard to vaccine development. Moreover, they provide a clear indication that presentation of exogenous antigens is not only possible, but occurs at reasonable efﬁciency by several routes, providing a mechanistic framework for studying cross-priming. In the ﬁnal section we review studies that begin to come grips with the ultimate ‘‘what does happen’’ issue in the ﬁeld of exogenous antigen presentation: the contribution of cross-priming in the induction of TCD8⫹ responding to endogenous antigens. At present, these studies have been largely limited technically to examining under what conditions crosspriming occurs, a useful if not absolutely necessary prelude to mechanistic analysis. 2. Evidence for Exogenous Antigen Presentation a. Transgenic Mice. In the opening chapter of a wonderful series of experiments performed by a team of Melburnians, the constitutive presentation of self-antigens was studied using transgenic mice expressing a membrane-bound transferrin receptor–OVA chimera under the control of the rat insulin promoter (mOVA-mice) (Kurts et al., 1996). mOVA is expressed at high levels in pancreatic islet 웁 cells, and (for whatever reasons) also in the proximal tubules of the kidneys, and at low levels in the thymus. TCR-transgenic mice producing Ova257–264-speciﬁc TCD8⫹ (OTI) were used as a source of antigen-speciﬁc TCD8⫹. Following adoptive transfer into mOVA-mice, activated OT-I cells selectively accumulated in the draining lymph nodes of the pancreas and kidney. Homing to renal nodes was prevented in a unilateral manner if the respective kidney was removed 7 days prior to transfer. Nephrectomy 4 hr prior to transfer was insufﬁcient to prevent homing of the OT-I cells. Reconstitution of irradiated animals with BM cells from a mouse expressing mutant forms of Kb unable to present Ova257–264 prevented homing to draining nodes, while

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

57

reconstitution with BM cells expressing wild-type Kb supported homing. Together, these experiments convincingly demonstrate that BM-derived pAPCs in local nodes constitutively present antigens from their respective tissues to the immune system.99 These ﬁnding were extended using transgenic mice expressing higher or lower amounts of unmodiﬁed OVA.100 Only in the high expressors was homing of naive OT-I cells to pancreatic nodes noted as observed with the mOVA-mice (Kurts et al., 1998). Nevertheless, the OVA expression in OVAlow mice was sufﬁcient to result in diabetes following transfer of activated OT-I cells. In a clever twist, OVAlow mice were given activated OTI cells to induce islet damage prior to transfer of naive cells labeled with a ﬂuorescent dye that allowed determination of their number of subsequent cell divisions. Naive cells transferred 4 days after activated cells were activated as determined on day 7, and it was shown using radiation chimeras that presentation was performed by BM-derived cells. By contrast, naive cells transferred 1 day after activated cells (and assayed on day 4) were not activated. Based on an estimated 1-day lag for OT-I cells to start proliferating, this indicates that pAPCs capable of activating naive OT-I cells were present in nodes no earlier than 3 days, and no later than 6 days following transfer of activated OT-I cells. These ﬁndings led the authors to conclude that ‘‘cellular destruction can enhance access of exogenous antigens to the cross-presentation pathway.’’ This is a good possibility but nearly equally plausible is that the level of exogenous antigen remains constant and the inﬂammation-induced cytokines either increase the processing efﬁciency of exogenous antigens or decrease the number of complexes required for activation. In either event, the important practical point is that inﬂammation increases the sensitivity of cross-presentation, and implies that foreign antigens in such a context will be presented with relatively high efﬁciency. b. Explants. A number of studies have reproduced the original crosspriming phenomenon of Bevan, with wrinkles that cells were expressing proteins from transgenes with deﬁned class I-restricted peptides, and that differentiated cells were injected [myoblasts (Ulmer et al., 1996) or keratinocytes (Oukka et al., 1997)] in place of splenocytes. Insight into the mechanism of cross-priming was provided by Nevala et al. (1998), who explanted splenocytes in Theracyte immunoisolation devices, which limit communication of the grafted cells with the recipient APCs to 0.4-애m pores. This 99 It is uncertain, however, whether OT-I cells are representative of normal naive T cells. Thus, although pAPCs may constitutively present antigens, it may not be in a manner capable of activating typical naive cells. 100 A reminder that this is a secreted protein, although the generation of Ova257–264 as an endogenous antigen is cytosolic.

58

JONATHAN W. YEWDELL et al.

resulted in cross-priming to both strong and weak miH antigens, demonstrating that intimate cell contact is not required under these conditions, with the implication being that pAPCs can acquire exogenous antigens in packages of less than 0.4 애m. Huang et al. (1996) studied generation of TCD8⫹ responses in A 씮 AXB BM radiation chimeras101 to IV NP expressed by explanted tumors. Marrow cells from normal mice, but not TAP1⫺/⫺ mice, were able to support the generation of NP-speciﬁc TCD8⫹ responses. Such mice were, however, capable of responding to NP if they were infected with a recombinant vaccinia virus (rVV) expressing an ER-targeted version of the peptide. The authors concluded that cross-priming is TAP dependent, probably due to a requirement for cytosolic processing. It is entirely plausible, however, that the absence of TAP exerts its inﬂuence beyond the ER, as discussed above with regard to the presentation of bacterial antigens. Moreover, the rVVs used to demonstrate the capacity of the TCD8⫹ repertoire of the chimeric mice to respond almost certainly express much greater quantities of peptide–class I complexes on pAPCs than would ever be generated from exogenous antigen, and is not clear whether the reconstituted TAP⫺/⫺ mice could respond to pAPCs bearing the lesser amounts of complexes generated during cross-priming. To examine the potential requirement of ER chaperones in crosspriming, Schoenberger et al. (1998) produced cell lines from wild-type and TAP⫺/⫺ mice that express adenovirus (AV) proteins. Both cell lines induced AV-speciﬁc memory TCD8⫹ in syngeneic mice. Because the TAP⫺/⫺ cells were not recognized by AV-speciﬁc TCD8⫹ in vitro, the authors concluded that cross-priming occurs in a TAP-independent manner, and therefore must not require the presentation of peptide by ER chaperones. Only a single dose of cells was used, however, so the possible contribution of any TAP-dependent process to priming may have been missed. Further, the priming observed is technically not cross-priming, because the syngeneic mice were used, and although it is unlikely that the TAP⫺/⫺ cells were able to activate naive TCD8⫹ directly, this is possible, particularly because the character of the cells may change in vivo. Finally, it is still at this point an assumption (though a reasonable one) that the generation of immunogenic gp96 by cells expressing cytosolic antigens requires TAP, because this has not been formally demonstrated. c. Plasmid DNA. A number of studies have examined the mechanisms of presentation of antigens encoded by plasmid DNA under the control 101

Such mice have a T cell repertoire that can recognize antigens in association with either A or B haplotypes.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

59

of the CMV promoter (which is active in most tissues).102 Corr et al. (1996) used BM chimeras to determine the nature of the APC. They found that following intramuscular (i.m.) injection of a plasmid encoding NP into A 씮 AXB mice, only A-restricted, NP-speciﬁc TCD8⫹ were primed, demonstrating that that functional presentation was limited to marrow-derived APCs. Similar ﬁndings were reported by Iwasaki et al. (1997), who further found that including genes encoding the costimulatory molecule B7-2 or GM-CSF together with IL-12 did not enable the presentation of IV NP by non-BM-derived cells following i.m. or gene gun epidermal immunizations.103 These experiments do not distinguish whether the bone marrow-derived pAPCs present exogenous antigens obtained from tissues in the muscle cells or are transfected and are simply presenting the (now) endogenous transgene product. As described in the next section, the latter clearly can occur under some circumstances. There are two studies, however, that suggest that exogenous presentation can also occur. Using an unusual chimera system in which AXB cells are infused into B mice with severe combined immunodeﬁciency,104 Doe et al. (1996) reported that BM-derived cells were required for the generation of TCD8⫹ speciﬁc for HIV gp120 or HSV glycoprotein B, when mice were injected i.m. with a plasmid expressing the transgene. Responses could be obtained if donor cells were given 21 days after DNA injection, which argues against direct transfection of the transferred BM cells, favoring cross-priming.105 Boyle et al. (1997) examined the immunogenicity of plasmids expressing OVA or a cytosolic version (cOVA). Following i.m. injection, better OVAspeciﬁc TCD8⫹ responses were elicited by the OVA-producing plasmid. By contrast, the two plasmids induced similar responses following intradermal (i.d.) injection which has been shown by others (Raz et al., 1994; and see below) to result in transfection of Langerhans cells (i.e., skin DCs). Arguing, ﬁrst, that cytosolic OVA should be presented more efﬁciently than secreted 102

In fact, this promoter is used in all of the transfection studies we discuss, making it ubiquitous in use as well as tissue expression. 103 In this method (protected by the Second Amendment), also known as biolistic delivery, gold particles coated with DNA are introduced into the skin at high velocity using a helium powered gun. 104 These mice have no T cells and consequently do not mount an anti-A alloreactive response. To enable responses, T cells from AXB mice are adoptively transferred. 105 The time lag does not eliminate the possibility of transfection by plasmid, however, and the pAPCs may even be transfected by DNA derived from host cells. Taken to its extreme, this argument can also be applied to all of cross-priming: it is possible that what is transferred is not protein but genetic material (DNA or RNA). Although this is more likely to occur with plasmid DNA, there are no direct experiments that eliminate this possibility (i.e., antibody to the transferred antigen blocks priming), so it should be kept in mind, if only in the subconscious realms.

60

JONATHAN W. YEWDELL et al.

OVA as an endogenous antigen (this is likely based on results from other systems, but was not examined in this study) and, second, that transfected cells should release more sOVA than cOVA (which is probably degraded in the cytosol, but again was not examined), they concluded that presentation following i.m. injection was probably based largely on presentation of exogenous OVA, whereas i.d. injection resulted in presentation of endogenous antigens. This is an important experimental approach because it implicates the exogenous antigen, and not a molecular chaperone bearing peptides, as being important in cross-priming (at least following i.m. DNA injection), and it is well worth verifying the assumptions made in this study and extending the results to other proteins. d. Viruses. In the course of investigating the intracellular processing of a secreted version of IV NP (sNP) produced by recombinant vaccinia viruses, we found that sNP is processed only slightly less efﬁciently in vitro than NP, and was similarly immunogenic when the doses of VV required for priming TCD8⫹ responses were compared (Bacik et al., 1997). Converting an immunodominant determinant to contain an N-linked glycosylation site had a disparate effect on the antigenicity and immunogenicity of sNP. The efﬁciency of peptide production was reduced about 50% (due to glycosylation of the peptide) in vitro, and immunogenicity was reduced more than 100-fold (i.e., 100-fold more virus was required for priming). The same alteration had no deleterious effect on the antigenicity or immunogenicity of NP (expected, because as a cytosolic protein it is not glycosylated). Although enzymatic removal of the N-linked oligosaccharide is possible, it results in the conversion of the determinant to a nonimmunogenic form. There are two major explanations for these ﬁndings. First, the processing of sNP in afAPCs may differ from that in the cells used in the in vitro study such that glycosylation has a more deleterious effect on presentation. Second, the presentation of this determinant from secreted NP may proceed mainly through exogenous presentation of the secreted protein. There is evidence that presentation of VV is mediated by infected efAPCs (discussed below), which would argue against cross-priming, but perhaps the efﬁciency of presentation of sNP as an endogenous antigen is so low that the (presumably less efﬁcient) route of exogenous presentation is pressed into service. 3. Evidence against Exogenous Presentation a. Plasmid DNA. Chattergoon et al. (1998) searched for expression of green ﬂuorescent protein (GFP) in lymph nodes following i.m. injection of a GFP-expressing plasmid. Two weeks after inoculation, a few GFPexpressing cells thought to represent M␾s and DCs were detected in

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

61

draining nodes. Transfected M␾s were also detected in the blood. In all cases, GFP was located in the cytoplasm of cells, and not in endosomes, consistent with its expression as an endogenous antigen (as opposed to its endocytosis from transfected muscle cells). Transfected M␾s were shown to express B7-2, and could be shown to activate naive T cells in vitro, but it was not shown to what extent this reﬂected activation of TCD8⫹. These ﬁndings echoed a prior report that GFP-expressing cells of DC morphology were present in draining nodes 1 day following gene gun immunization with a GFP-expressing plasmid (Condon et al., 1996). Such cells also contained the gold particles used for plasmid delivery. To determine whether these cells were present in the skin at the time of immunization, the skin was painted with rhodamine immediately before immunization. All GFP-expressing cells in the node were rhodamine labeled, but the relevance of this experiment is muddied by the possibility that rhodamine painting enhanced the migration of normally sessile LCs.106 Considerable insight into the mechanism of presentation of DNA gundelivered antigens was provided by Porgador et al. (1998), who used plasmid DNA expressing 웁-gal and located transfected cells histochemically in serial sections. This revealed that 앑200 transfected cells are present in nodes, with most, but not all, transfected cells present in local draining nodes. Dissociated node cells were able to activate a 웁-gal-speciﬁc TCD8⫹ clone, and presentation was abrogated by depletion of class II-expressing cells. Depletion of M␾s or B cells had little effect on presentation, whereas depletion with a DC-speciﬁc mAb eliminated 앑70% of presentation activity. Because this mAb is known to bind poorly to a subpopulation of DCs, these ﬁndings are consistent with DCs representing the major, and possibly the sole, APC in this system. Next, mice were injected with particles cocoated with the 웁-gal-expressing plasmid and a plasmid expressing human CD4, and 24 hr later the dissociated node cells were depleted of cells expressing human CD4. Depletion had a marked impact on the ability of cells to activate the 웁-gal-speciﬁc TCD8⫹, indicating that most APCs in the node present endogenous antigens. In parallel experiments using tissue culture cell lines, gun transfection with the same beads led to coexpression of 웁-gal and CD4 in 앑70% of cells, suggesting that all of the APCs may have presented endogenous antigens. These ﬁndings indicate that DNA immunization frequently results in the transfection of pAPCs. Such transfection is probably favored by biolistic immunization, because pAPCs are well suited for endocytosing the gold particles used for delivery. The ﬁndings of Porgador et al. are consistent with transfection accounting for all priming, but there are important limita106

Very possible, because this happens with ﬂuorescein painting.

62

JONATHAN W. YEWDELL et al.

tions to this conclusion. First, the APCs were tested for their abilities to activate a TCD8⫹ clone, and their ability to activate naive TCD8⫹ remains to be establish by adoptive transfer experiments. Second, nodes were assayed only 24 hr after DNA delivery, and cross-priming make take longer to occur (as suggested by the Melbourne ﬁndings). Finally, the nature of the inﬂammation induced by DNA immunization no doubt differs considerably from that associated with other antigenic stimuli, and capacity of pAPCs for cross-priming may vary according to the milieu. b. Viruses. Adeno-associated virus (AAV) vectors, unlike plasmid DNA or AV vectors, are able to transduce muscle cells stably with 웁-gal following i.m. injection, without detectably inducing TCD8⫹ responses to 웁-gal (Xiao et al., 1996; Kessler et al., 1996). Studying this phenomenon, Jooss et al. (1998) found that induction of 웁-gal-speciﬁc TCD8⫹ by infection with 웁gal-expressing recombinant adenovirus (rAd) results in the destruction of AAV-transduced myocytes, demonstrating that the myocytes can serve as efAPCs. Co-i.m. injection with AAV 웁-gal and rAd not expressing 웁-gal resulted in local inﬂammation of the muscle (including TCD8⫹ inﬁltration) but no induction of 웁-gal-speciﬁc TCD8⫹ or detectable destruction of AAV-transduced, 웁-gal-expressing myocytes. A subpopulation of DCS–Cs (‘‘⬎10%’’) were clearly infected by Ad-웁-gal and were able to induce 웁gal-speciﬁc TCD8⫹ responses on adoptive transfer. By contrast, AAV-웁-gal was unable to induce detectable 웁-gal expression in DCs or M␾s or to enable the cells to elicit in vivo TCD8⫹ responses. The authors concluded that the difference between Ad and AAV in their abilities to induce TCD8⫹ lie in the inability of pAPCs to express AAV-encoded proteins. This conclusion hinges, of course, on the assumption that the in vivo properties of APCs are maintained in vitro. Moreover, it remains possible that to some extent differences between Ad and AAV relate to the amount of 웁-gal produced by myocytes, which was not examined. Although these issues remain to be sorted out, this ﬁne study is in seeming conﬂict with the equally compelling results of the Melbourne group described above, regarding the cross-presentation of OVA in transgenic animals. There are, however, crucial differences in the studies that may inﬂuence the outcome. The most likely culprits are listed: 1. The expression of OVA as secreted or membrane protein bound versus cytosolic expression of 웁-gal. 2. The efﬁciency of processing exogenous OVA versus 웁-gal. 3. Muscle versus kidney or pancreas as a source of exogenous antigen. 4. The activation of transgenic versus normal TCD8⫹.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

63

5. The amounts of peptide class I complexes required for TCD8⫹ activation. There is also evidence regarding the involvement of cross-priming in TCD8⫹ responses to rVVs. It was reported a number of years ago that IV hemagglutinin (HA) expressed under the control of a late viral promoter could be less immunogenic (depending on the mouse strain tested) than when expressed under the control of an early promoter, despite the fact that greater amounts were synthesized under late promoter control107 (Coupar et al., 1986). These ﬁndings were extended by Bronte et al. (1997), who examined the ability of a panel of rVVs expressing 웁-gal under the control of a number of early and late promoters to prolong survival of mice bearing a lethal 웁-gal-expressing tumor108: Given promoters of equal strength, rVVs producing 웁-gal under early promoter control were more effective at prolonging survival, although viruses expressing large quantities of 웁-gal under late promoter control provided a similar extent of protection as the best early promoter rVV. These ﬁndings were correlated with the expression of 웁-gal in mouse DCs–cs: 웁-gal was expressed far better from early than from late promoters, and DCs infected with the early promoter construct were much better at restimulating primed TCD8⫹ in vitro. These latter ﬁndings are consistent with the idea that pAPCs presenting endogenous VV antigens play an important role in inducing TCD8⫹ responses.109 On the other hand, the fact that late constructs are immunogenic argues that cross-priming can also play a role in presentation of VV gene products in vivo, which would agree with the SNP ﬁndings discussed above, as well as the immunogenicity of VV late HA in some mouse strains. IV. Conclusion: Basic Questions

In this review we have labored to ﬂag the areas in exogenous antigen presentation where the ice is thinnest or unformed, and (keeping within this metaphor) have given our opinions on which areas will provide the most promising skating in the future. In this spirit, we conclude by posing questions whose answers we deem most important for achieving a basic understanding of exogenous antigen presentation in vitro and in vivo, conﬁdent that this knowledge will have important practical applications, particularly in the area of vaccine development. 107

Early promoters allow the expression of protein from the onset of viral infection, whereas viral DNA replication is required for expression of late proteins. 108 TCD8⫹ responses were not measured, but in prior work, protection was shown to be mediated by 웁-gal-speciﬁc TCD8⫹. 109 They do not clearly implicate DCs, however, because other cells (notably M␾s) are known to limit the expression of late VV genes.

64

JONATHAN W. YEWDELL et al.

1. How are PR molecules delivered to endosomes or created there? How does this vary with cell type or history of exposure to cytokines? What roles do invariant chain or other molecular chaperones have in the delivery of class I molecules to endosomes? Is there a mechanism that distinguishes complexes created in the endosome and routes them to the cell surface, while complexes derived from the cell surface are sent to their death in lysosomes or prelysosomes? 2. How are class I ligands generated in endosomes? Which proteases are involved? Does this vary with cell type or history of cytokine exposure? Do class I molecules serve as templates for the action of proteases? Is peptide loading facilitated in a manner similar to DM-mediated loading of class II molecules—if so, is there a gene product equivalent to DM? To what extent does the spectrum of ligands generated in endosomes overlap with the ligand spectrum generated from the same protein by the cytosolic processing pathway? 3. How are antigens transported from endosomes to the cytosol? Does this vary depending on how the antigen was internalized (e.g., macropinocytosis versus endocytosis)? 4. Under what situations does cross-priming occur in vivo? How is this inﬂuenced by the activation status of the responding TCD8⫹ (e.g., naive versus memory TCD8⫹, different clones of naive TCD8⫹)? How does the constitutive cross-presentation of antigens compare with cross-presentation under conditions of foreign invasion, and to the extent that differences exist, what are the relative contributions of alterations in antigen-expressing cells versus alterations in pAPCs? 5. How are peptides transferred to pAPCs in cross-priming? How do apoptotic, necrotic, and viable cells compare as sources of antigens? What is the contribution of molecular chaperones to cross-priming and how does this relate to their normal role(s) in antigen processing? To what extent are cross-primed antigens processed endosomally or cytosolically? REFERENCES Albert, M. L., Sauter, B., and Bhardwaj, N. (1998). Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs. Nature (London) 392, 86–89. Alexander-Miller, M. A., Leggatt, G. R., Sarin, A., and Berzofsky, J. A. (1996). Role of antigen, CD8, and cytotoxic T lymphocyte (CTL) avidity in high dose antigen induction of apoptosis of effector CTL. J. Exp. Med. 184, 485–492.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

65

Allsopp, C. E., Plebanski, M., Gilbert, S., Sinden, R. E., Harris, S., Frankel, G., Dougan, G., Hioe, C., Nixon, D., Paoletti, E., Layton, G., and Hill, A. V. (1996). Comparison of numerous delivery systems for the induction of cytotoxic T lymphocytes by immunization. Eur. J. Immunol. 26, 1951–1959. Andersen, P. S., Stryhn, A., Hansen, B. E., Fugger, L., Engberg, L., and Buus, S. (1996). A recombinant antibody with the antigen-speciﬁc major histocompatibility complexrestricted speciﬁcity of T cells. Proc. Natl. Acad. Sci. U.S.A. 93, 1820–1824. Anderson, R. G. (1993). Caveolae: Where incoming and outgoing messengers meet. Proc. Natl. Acad. Sci. U.S.A. 90, 10909–10913. Anderson, K., Cresswell, P., Gammon, M., Hermes, J., Williamson, A., and Zweerink, H. (1991). Endogenously synthesized peptide with an endoplasmic reticulum signal sequence sensitizes antigen processing mutant cells to class I-restricted cell-mediated lysis. J. Exp. Med. 174, 489–492. Anto´n, L. C., Yewdell, J. W., and Bennink, J. R. (1997). MHC class I-associated peptides produced from endogenous gene products with vastly different efﬁciencies. J. Immunol. 158, 2535–2542. Anto´n, L. C., Snyder, H. L., Bennink, J. R., Vinitsky, A., Orlowski, M., Porgador, A., and Yewdell, J. W. (1998). Dissociation of proteasomal degradation of biosynthesized viral proteins from generation of MHC class I-associated antigenic peptides. J. Immunol. 160, 4859–4868. Arnold, D., Faath, S., Rammensee, H. G., and Schild, H. (1995). Cross-priming of minor histocompatibility antigen-speciﬁc cytotoxic T cells upon immunization with the heat shock protein gp96. J. Exp. Med. 182, 885–889. Babbitt, B. P., Allen, P. M., Matsueda, G. R., Haber, E., and Unanue, E. R. (1985). Binding of immunogenic peptides to Ia histocompatibility molecules. Nature (London) 317, 359–361. Bachmann, M. F., Kundig, T. M., Freer, G., Li, Y., Kang, C. Y., Bishop, D. H., Hengartner, H., and Zinkernagel, R. M. (1994). Induction of protective cytotoxic T cells with viral proteins. Eur. J. Immunol. 24, 2228–2236. Bachmann, M. F., Oxenius, A., Pircher, H., Hengartner, H., Ashton-Richardt, P. A., Tonegawa, S., and Zinkernagel, R. M. (1995). TAP1-independent loading of class I molecules by exogenous viral proteins. Eur. J. Immunol. 25, 1739–1743. Bachmann, M. F., Lutz, M. B., Layton, G. T., Harris, S. J., Fehr, T., Rescigno, M., and Ricciardi-Castagnoli, P. (1996). Dendritic cells process exogenous viral proteins and viruslike particles for class I presentation to CD8⫹ cytotoxic T lymphocytes. Eur. J. Immunol. 26, 2595–2600. Bacik, I., Link-Snyder, H., Anto´n, L. C., Russ, G., Chen, W., Bennink, J. R., Urge, L., Otvos, L., Dudkowska, B., Eisenlohr, L., and Yewdell, J. W. (1997). Introduction of a glycosylation site into a secreted protein provides evidence for a novel antigen processing pathway: Transport of precursors of MHC class I resistricted-peptides from the endoplasmic reticulum to the cytosol. J. Exp. Med. 186, 479–487. Barnaba, V., Franco, A., Alberti, A., Benvenuto, R., and Balsano, F. (1990). Selective killing of hepatitis B envelope antigen-speciﬁc B cells by class I-restricted, exogenous antigenspeciﬁc T lymphocytes. Nature (London) 345, 258–260. Bellone, M., Iezzi, G., Rovere, P., Galati, G., Ronchetti, A., Protti, M. P., Davoust, J., Rugarli, C., and Manfredi, A. A. (1997). Processing of engulfed apoptotic bodies yields T cell epitopes. J. Immunol. 159, 5391–5399. Bercovich, B., Stancovski, I., Mayer, A., Blumenfeld, N., Laszlo, A., Schwartz, A. L., and Ciechanover, A. (1997). Ubiquitin-dependent degradation of certain protein substrates in vitro requires the molecular chaperone Hsc70. J. Biol. Chem. 272, 9002–9010.

66

JONATHAN W. YEWDELL et al.

Bevan, M. J. (1976). Cross-priming for a secondary cytotoxic response to minor H antigens with H-2 congenic cells which do not cross-react in the cytotoxic assay. J. Exp. Med. 143, 1283 Bevan, M. J. (1977). In a radiation chimaera, host H-2 antigens determine immune responsiveness of donor cytotoxic cells. Nature (London) 269, 417–418. Bevan, M. J. (1987). Antigen recognition: Class discrimination in the world of immunology. Nature (London) 325, 192–194. Bishop, N. E. (1998). An update on non-clathrin-coated endocytosis. Rev. Med. Virol. 7, 199–209. Blachere, N. E., Li, Z., Chandawarkar, R. Y., Suto, R., Jaikaria, N. S., Basu, S., Udono, H., and Srivastava, P. K. (1997). Heat shock protein–peptide complexes, reconstituted in vitro, elicit peptide-speciﬁc cytotoxic T lymphocyte response and tumor immunity. J. Exp. Med. 186, 1315–1322. Bogyo, M., McMaster, J. S., Gaczynska, M., Tortorella, D., Goldberg, A. L., and Ploegh, H. (1997). Covalent modiﬁcation of the active site threonine of proteasomal 웁 subunits and the Escherichia coli homolog Hsl V by a new class of inhibitors. Proc. Natl. Acad. Sci. U.S.A. 94, 6629–6634. Boyle, J. S., Koniaras, C., and Lew, A. M. (1997). Inﬂuence of cellular location of expressed antigen on the efﬁcacy of DNA vaccination: Cytotoxic T lymphocyte and antibody responses are suboptimal when antigen is cytoplasmic after intramuscular DNA immunization. Int. Immunol. 9, 1897–1906. Bronte, V., Carroll, M. W., Goletz, T. J., Wang, M., Overwijk, W. W., Marincola, F., Rosenberg, S. A., Moss, B., and Restifo, N. P. (1997). Antigen expression by dendritic cells correlates with the therapeutic effectiveness of a model recombinant poxvirus tumor vaccine. Proc. Natl. Acad. Sci. U.S.A. 94, 3183–3188. Brossart, P., Goldrath, A. W., Butz, E. A., Martin, S., and Bevan, M. J. (1997). Virusmediated delivery of antigenic epitopes into dendritic cells as a means to induce CTL. J. Immunol. 158, 3270–3276. Brossay, L., Burdin, N., Tangri, S., and Kronenberg, M. (1998). Antigen-presenting function of mouse CD1: One molecule with two different kinds of antigenic ligands. Immunol. Rev. 163, 139–150. Buchholz, D., Scott, P., and Shastri, N. (1995). Presentation without proteolytic cleavage of endogenous precursors in the MHC class I antigen processing pathway. J. Biol. Chem. 270, 6515–6522. Burkly, L., Hesslon, C., Ogata, L., Reilly, C., Marconi, L. A., Olsen, D., Tizard, R., Cate, R., and Lo, D. (1995). Expression of relB is required for the development of thymic medulla and dendritic cells. Nature (London) 373, 531–536. Buus, S., Colon, S. M., Smith, C., Freed, J. H., Miles, C., and Grey, H. M. (1986). Interaction between a‘‘processed’’ ovalbumin peptide and Ia molecules. Proc. Natl. Acad. Sci. U.S.A. 83, 3968–3971. Carbone, F. R., and Bevan, M. J. (1990). Class I-restricted processing and presentation of exogenous cell-associated antigen in vivo. J. Exp. Med. 171, 377–387. Cerundolo, V., Elliot, T., Elvin, J., Bastin, J., and Townsend, A. (1992). Association of the human invariant chain with H-2 Db class I molecules. Eur. J. Immunol. 22, 2243–2248. Cerundolo, V., Benham, A., Braud, V., Mukherjee, S., Gould, K., Macino, B., Neefjes, J., and Townsend, A. (1997). The proteasome-speciﬁc inhibitor lactacystin blocks presentation of cytotoxic T lymphocyte epitopes in human and murine cells. Eur. J. Immunol. 27, 336–341. Chattergoon, M. A., Robinson, T. M., Boyer, J. D., and Weiner, D. B. (1998). Speciﬁc immune induction following DNA-based immunization through in vivo transfection and activation of macrophages/antigen-presenting cells. J. Immunol. 160, 5707–5718.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

67

Ciupitu, A.-M. T., Petersson, M., O’Donnell, C. L., Williams, K., Jindal, S., Kiessling, R., and Welsh, R. M. (1998). Immunization with a lymphocytic chroiomeningitis virus peptide mixed with heat shock protein 70 results in protective antviral immunity and specﬁc cytotoxic T lymphocytes. J. Exp. Med. 187, 685–691. Collins, D. S., Findlay, K., and Harding, C. V. (1992). Processing of exogenous liposomeencapsulated antigens in vivo generates class I MHC-restricted T cell responses. J. Immunol. 148, 3336–3341. Condon, C., Watkins, S. C., Celluzzi, C. M., Thompson, K., and Falo, L. D. J. (1996). DNAbased immunization by in vivo transfection of dendritic cells. Nature Med. 2, 1122–1128. Corr, M., Lee, D. J., Carson, D. A., and Tighe, H. (1996). Gene vaccination with naked plasmid DNA: Mechanism of CTL priming. J. Exp. Med. 184, 1555–1560. Coupar, B. E. H., Andrew, M. E., Both, G. W., and Boyle, D. B. (1986). Temporal regulation of inﬂuenza hemagglutinin expression in vaccinia virus recombinants and effects on the immune response. Eur. J. Immunol. 16, 1479–1487. Cox, J. H., Galardy, P., Bennink, J. R., and Yewdell, J. W. (1995). Presentation of endogenous and exogenous antigens is not affected by inactivation of E1 ubiquitin-activating enzyme in temperature-sensitive cell lines. J. Immunol. 154, 511–519. Darji, A., Chakraborty, T., Wehland, J., and Weiss, S. (1997a). TAP-dependent major histocompatibility complex class I presentation of soluble proteins using listeriolysin. Eur. J. Immunol. 27, 1353–1359. Darji, A., Guzman, C. A., Gerstel, B., Wachholz, P., Timmis, K. N., Wehland, J., Chakraborty, T., and Weiss, S. (1997b). Oral somatic transgene vaccination using attenuated S. typhimurium. Cell 91, 765–775. Day, P. M., Esquivel, F., Lukszo, J., Bennink, J. R., and Yewdell, J. W. (1995). Effect of TAP on the generation and intracellular trafﬁcking of peptide-receptive major histocompatibility complex class I molecules. Immunity 2, 137–147. Day, P. M., Yewdell, J. W., Porgador, A., Germain, R. N., and Bennink, J. R. (1997). Direct delivery of exogenous MHC class I molecule-binding oligopeptides to the endooplasmic reticulum of viable cells. Proc. Natl. Acad. Sci. U.S.A. 94, 8064–8069. De Bruijn, M. L., Peterson, P. A., and Jackson, M. R. (1996). Induction of heat-stable antigen expression by phagocytosis is involved in in vitro activation of unprimed CTL by macrophages. J. Immunol. 156, 2686–2692. Debrick, J. E., Campbell, P. A., and Staerz, U. D. (1991). Macrophages as accessory cells for class I MHC-restricted immune responses. J. Immunol. 147, 2846–2851. Derossi, D., Chassaing, G., and Prochiantz, A. (1998). Trojan peptides: The penetratin system for intracellular delivery. Trends Cell Biol. 8, 84–87. Doe, B., Selby, M., Barnett, S., Baenziger, J., and Walker, C. M. (1996). Induction of cytotoxic T lymphocytes by intramuscular immunization with plasmid DNA is facilitated by bone marrow-derived cells. Proc. Natl. Acad. Sci. U.S.A. 93, 8578–8583. Donnelly, J. J., Ulmer, J. B., Hawe, L. A., Friedman, A., Shi, X.-P., Leander, K. R., Shiver, J. W., Oliff, A. I., Martinez, D., Montgomery, D., and Liu, M. A. (1993). Targeted delivery of peptide epitopes to class I major histocompatibility molecules by a modiﬁed Pseudomonas exotoxin. Proc. Natl. Acad. Sci. U.S.A. 90, 3530–3534. Eisenlohr, L. C., Gerhard, W., and Hackett, C. J. (1987). Role of receptor-binding activity of the viral hemagglutinin molecule in the presentation of inﬂuenza virus antigens to helper T cells. J. Virol. 61, 1375–1383. Elliott, T. (1997). Transporter associated with antigen processing. Adv.Immunol. 65, 47–109. Elliott, T., Willis, A., Cerundolo, V., and Townsend, A. (1995). Processing of major histocompatibility class I-restricted antigens in the endoplasmic reticulum. J. Exp. Med. 181, 1481– 1491.

68

JONATHAN W. YEWDELL et al.

Fazekas de St.Groth, B. (1998). The evolution of self-tolerance: A new cell arises to meet the challenge of self-reactivity. Immunol. Today 19, 448–454. Fu, T. M., Mylin, L. M., Schell, T. D., Bacik, I., Russ, G., Yewdell, J. W., Bennink, J. R., and Tevethia, S. S. (1998). An endoplasmic reticulum-targeting signal sequence enhances the immunogenicity of an immunorecessive simian virus 40 large T antigen cytotoxic Tlymphocyte epitope. J. Virol. 72, 1469–1481. Gagliardi, M. C., Nisini, R., Benvenuto, R., De Petrillo, G., Michel, M. L., and Barnaba, V. (1994). Soluble transferrin mediates targeting of hepatitis B envelope antigen to transferrin receptor and its presentation by activated T cells. Eur. J. Immunol. 24, 1372– 1376. Galy, A., Travis, M., Cen, D., and Chen, B. (1995). Human T, B, natural killer, and dendritic cells arise from a common bone marrow progenitor cell subset. Immunity 3, 459–473. Glas, R., Bogyo, M., McMaster, J. S., Gaczynska, M., and Ploegh, H. L. (1998). A proteolytic sysytem that compensates for loss of proteasome function. Nature (London) 392, 618–622. Goletz, T. J., Klimpel, K. R., Arora, N., Leppla, S. H., Keith, J. M., and Berzofsky, J. A. (1997). Targeting HIV proteins to the major histocompatibility complex class I processing pathway with a novel gp120–anthrax toxin fusion protein. Proc. Natl. Acad. Sci. U.S.A. 94, 12059–12064. Gooding, L. R., and Edwards, C. B. (1980). H-2 antigen requirements in the in vitro induction of SV40-speciﬁc cytotoxic T lymphocytes. J. Immunol. 124, 1258 Gordon, S. (1998). Macrophages and the immune response. In ‘‘Fundamental Immunology’’ (W. E. Paul, ed.), pp. 533–546. Lippincott-Raven, Philadephia and New York. Gordon, R. D., Mathieson, B. J., Samelson, L. E., Boyse, E. A., and Simpson, E. (1976). The effect of allogeneic presensitization on H-Y graft survival and in vitro cell-mediated responses to H-Y antigen. J. Exp. Med. 144, 810–820. Grandea, A. G., Lehner, P. J., Cresswell, P., and Spies, T. (1997). Regulation of MHC class I heterodimer stability and interaction with TAP by tapasin. Immunogenetics 46, 477–483. Greenspan, N., Hurwitz, J. L., and Doherty, P. C. (1982). Capacity of inﬂuenza virus– monoclonal antibody mixtures to stimulate memory and cytotoxic T lymphocyte populations. Infect. Immun. 38, 645–650. Harding, C. V., and Pfeifer, J. D. (1994). Antigen expressed by Salmonella typhimurium is processed for class I major histocompatibility complex presentation by macrophages but not infected epithelial cells. Immunology 83, 670–674. Harding, C. V., and Song, R. (1994). Phagocytic processing of exogenous particulate antigens by macrophages for presentation by class I MHC molecules. J. Immunol. 153, 4925–4933. Hartl, F. U. (1996). Molecular chaperones in cellular protein folding. Nature (London) 381, 571–580. Hochman, J. H., Jiang, J., Matyus, L., Edidin, M., and Pernis, B. (1991). Endocytosis and dissociation of class I MHC molecules labeled with ﬂuorescent 웁2-microglobulin. J. Immunol. 146, 1862–1867. Hosken, N. A., Bevan, M. J., and Carbone, F. R. (1989). Class I-restricted presentation occurs without internalization or processing of exogenous antigenic peptides. J. Immunol. 142, 1079–1083. Huang, A. Y., Bruce, A. T., Pardoll, D. M., and Levitsky, H. I. (1996). In vivo cross-priming of MHC class I-restricted antigens requires the TAP transporter. Immunity 4, 349–355. Hughes, E. A., Ortmann, B., Surman, M., and Cresswell, P. (1996). The protease inhibitor, N-acetyl-L-leucyl-L-leucyl-L-norleucinal decreases the pool of major histocompatibility complex class I-binding peptides and inhibits peptide trimming in the endoplasmic reticulum. J. Exp. Med. 183, 1569–1578.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

69

Iwasaki, A., Torres, C. A. T., Ohashi, P. S., Robinson, H. L., and Barber, B. H. (1997). The dominant role of bone marrow-derived cells in CTL induction following plasmid DNA immunization at different sites. J. Immunol. 159, 11–14. Jay, G., Jay, F. T., Chang, C., Friedman, R. M., and Levine, A. S. (1978). Tumor-speciﬁc transplantation antigen: Use of Ad2⫹ND1 hybrid virus to identify the protein responsible for simian virus 40 tumor rejection and its genetic orgin. Proc. Natl. Acad. Sci. U.S.A. 75, 3055–3059. Jin, Y., Shih, W. K., and Berkower, I. (1988). Human T cell response to the surface antigen of hepatitis B virus (HBsAg). Endosomal and nonendosomal processing pathways are accessible to both endogenous and exogenous antigen. J. Exp. Med. 168, 293–306. Jooss, K., Yang, Y., Fisher, K. J., and Wilson, J. M. (1998). Transduction of dendritic cells by DNA viral vectors directs the immune response to transgene products in muscle ﬁbers. J. Virol. 72, 4212–4223. Ke, Y., and Kapp, J. A. (1996). Exogenous antigens gain access to the major histocompatibility complex class I processing pathway in B cells by receptor-mediated uptake. J. Exp. Med. 184, 1179–1184. Kessler, P. D., Podsakoff, G. M., Chen, X., McQuiston, S. A., Colosi, P. C., Matelis, L. A., Kurtzman, G. J., and Byrne, B. J. (1996). Gene delivery to skeletal muscle results in sustained expression and systemic delivery of a therapeutic protein. Proc. Natl. Acad. Sci. U.S.A. 93, 14082–14087. Koszinowski, U. H., and Gething, M. J. (1980). Generation of virus-speciﬁc cytotoxic T cells in vitro. II. Induction requirements with functionally inactivated virus preparations. Eur. J. Immunol. 10, 30–35. Koszinowski, U., Gething, M. J., and Waterﬁeld, M. (1977). T-cell cytotoxicity in the absence of viral protein synthesis in target cells. Nature (London) 267, 160–163. Kovacsovics-Bankowski, M., and Rock, K. L. (1995). A phagosome-to-cytosol pathway for exogenous antigens presented on MHC class I molecules. Science 267, 243–245. Kovacsovics-Bankowski, M., Clark, K., Benacerraf, B., and Rock, K. L. (1993). Efﬁcient major histocompatibility complex class I presentation of exogenous antigen upon phagocytosis by macrophages. Proc. Natl. Acad. Sci. U.S.A. 90, 4942–4946. Kurts, C., Heath, W. R., Carbone, F. R., Allison, J., Miller, J. F., and Kosaka, H. (1996). Constitutive class I-restricted exogenous presentation of self antigens in vivo. J. Exp. Med. 184, 923–930. Kurts, C., Miller, J. F., Subramaniam, R. M., Carbone, F. R., and Heath, W. R. (1998). Major histocompatibility complex class I-restricted cross-presentation is biased towards high dose antigens and those released during cellular destruction. J. Exp. Med. 188, 409–414. Lammert, E., Stevanovic, S., Brunner, J., Rammensee, H. G., and Schild, H. (1997). Protein disulﬁde isomerase is the dominant acceptor for peptides translocated into the endoplasmic reticulum. Eur. J. Biochem. 27, 1685–1690. Layton, G. T., Harris, S. J., Myhan, J., West, D., Gotch, F., Hill-Perkins, M., Cole, J. S., Meyers, N., Woodrow, S., French, T. J., Adams, S. E., and Kingsman, A. J. (1996). Induction of single and dual cytotoxic T-lymphocyte responses to viral proteins in mice using recombinant hybrid Ty-virus-like particles. Immunology 87, 171–178. Lee, R. S., Tartour, E., van der Bruggen, P., Vantomme, V., Joyeux, I., Goud, B., Fridman, W. H., and Johannes, L. (1998). Major histocompatibility complex class I presentation of exogenous soluble tumor antigen fused to the B-fragment of Shiga toxin. Eur. J. Immunol. 28, 2726–2737. Lenz, L. L., Dere, B., and Bevan, M. J. (1996). Identiﬁcation of an H2-M3-restricted Listeria epitope: Implications for antigen presentation by M3. Immunitiy 5, 63–72.

70

JONATHAN W. YEWDELL et al.

Lewis, J. W., and Elliott, T. (1998). Evidence for successive peptide binding and quality control stages during MHC class I assembly. Curr. Biol. 8, 717–720. Li, Z., and Srivastava, P. K. (1993). Tumor rejection antigen gp96/grp94 is an ATPase: Implications for protein folding and antigen presentation. EMBO J. 12, 3143–3151. Li, S., Sjogren, H. O., Hellman, U., Pettersson, R. F., and Wang, P. (1997). Cloning and functional characterization of a subunit of the transporter associated with antigen processing. Proc. Natl. Acad. Sci. U.S.A. 94, 8708–8713. Lie, W. R., Myers, N. B., Connolly, J. M., Gorka, J., Lee, D. R., and Hansen, T.H. (1991). The speciﬁc binding of peptide ligand to Ld class I major histocompatibility complex molecules determines their antigenic structure. J. Exp. Med. 173, 449–459. Lindahl, K. F., Byers, D. E., Dabhi, V. M., Hovik, R., Jones, E. P., Smith, G. P., Wang, C. R., Xiao, H., and Yoshino, M. (1997). H2-M3, a full-service class Ib histocompatibility antigen. Annu. Rev. Immunol 15, 851–879. ¨ rvell, C., Lederer, E., Ljunggren, H.-G., and Jondal, M. (1995). HeatLiu, T., Zhou, X., O inactivated Sendai virus can enter multiple MHC class I processing pathways and generate cytotoxic T lymphocyte responses in vivo. J. Immunol. 154, 3147–3155. Liu, T., Chambers, B., Diehl, A. D., Van Kaer, L., Jondal, M., and Ljunggren, H.G. (1997). TAP peptide transporter-independent presentation of heat-killed Sendai virus antigen on MHC class I molecules by splenic antigen-presenting cells. J. Immunol. 159, 5364–5371. Lord, J. M. and Roberts, L. M. (1998). Toxin entry: Retrograde transport through the secretory pathway. J. Cell Biol. 140, 733–736. Luckey, C. J., King, G. M., Marto, J. A., Venketeswaran, S., Maier, B. F., Crotzer, V. L., Colella, T. A., Shabanowitz, J., Hunt, D. F., and Engelhard, V. H. (1998). Proteasomes can either generate or destroy MHC class I epitopes: Evidence for nonproteasomal epitope generation in the cytosol. J. Immunol. 161, 112–121. Martin, S., Niedermann, G., Leipner, C., Eichmann, K., and Weltzien, H.U. (1993). Intracellular processing of hapten-modiﬁed protein for MHC class I presentation: Cytoplasmic delivery by pH-sensitive liposomes. Immunol. Lett. 37, 97–102. Martinez-Kinader, B., Lipford, G. B., Wanger, H., and Heeg, K. (1995). Sensitization of MHC class I-restricted T cells to exogenous proteins: Evidence for an alternative class I restricted antigen presentation pathway. Immunology 86, 287–295. Matsuno, K., Ezaki, T., Kudo, S., and Uehara, Y. (1996). A life stage of particle-laden rat dendritic cells in vivo: Their terminal division, active phagocytosis, and translocation from the liver to the draining lymph. J. Exp. Med. 183, 1865–1878. Mazzaccaro, R. J., Gedde, M., Jensen, E. R., van Santen, H. M., Ploegh, H. L., Rock, K. L., and Bloom, B. R. (1996). Major histocompatibility class I presentation of soluble antigen facilitated by Mycobacterium tuberculosis infection. Proc. Natl. Acad. Sci. U.S.A. 93, 11786–11791. Melian, A., Beckman, E. M., Porcelli, S., and Brenner, M. B. (1996). Antigen presentation by CD1 and MHC-encoded class I-like molecules. Curr. Opin. Immunol. 8, 82–88. Michalek, M. T., Grant, E. P., Gramm, C., Goldberg, A. L., and Rock, K. L. (1993). A role for the ubiquitin-dependent proteolytic pathway in MHC class I-restricted antigen presentation. Nature (London) 363, 552–554. Mitchell, D. A., Nair, S. K., and Gilboa, E. (1998). Dendritic cell/macrophage precursors capture exogenous antigen for MHC class I presentation by dendritic cells. Eur. J. Immunol. 28, 1923–1933. Momburg, F. J., and Hammerling, G. J. (1998). Generation and TAP-mediated transport of peptides for major histocompatibility complex class I molecules. Adv. Immunol. 68, 191–256.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

71

Moore, M. W., Carbone, F. R., and Bevan, M. J. (1988). Introduction of soluble protein into the class I pathway of antigen processing and presentation. Cell 54, 777–785. Motal, U. M. A., Zhou, X., Joki, A., Siddiqi, A. R., Srinivasa, B. R., Stenvall, K., Dahmen, J., and Jondal, M. (1993). Major histocompatibility complex class I-binding peptides are recycled to the cell surface after internalization. Eur. J. Immunol. 23, 3224–3229. Nair, S., Zhou, F., Reddy, R., Huang, L., and Rouse, B. T. (1992). Soluble proteins delivered to dendritic cells via pH-sensitive liposomes induce primary cytotoxic T lymphocyte responses in vitro. J. Exp. Med. 175, 609–612. Nair, S., Buiting, A. M., Rouse, R. J., van Rooijen, N., Huang, L., and Rouse, B. T. (1995). Role of macrophages and dendritic cells in primary cytotoxic T lymphocyte responses. Int. Immunol. 7, 679–688. Neefjes, J. J., Stollorz, V., Peters, P. J., Geuze, H. J., and Ploegh, H. L. (1990). The biosynthetic pathway of MHC class II but not class I molecules intersects the endocytic route. Cell 61, 171–183. Nevala, W. K., Paul, C., and Wettstein, P. J. (1998). Immunodominant minor histocompatibility antigen peptides recognized by cytolytic T lymphocytes primed by indirect presentation. Transplantation 65, 559–569. Nicchitta, C. V. (1998). Biochemical, cell biological and immunological issues surrounding the endoplasmic reticulum chaperone GRP94/gp96. Curr. Opin. Immunol. 10, 103–109. Nieland, T. J. F., Tan, M. C. A. A., Monnee-van Muijen, M., Koning, F., Kruisbeek, A. M., and Van Bleek, G. M. (1996). Isolation of an immunodominant viral peptide that is endogenously bound to the stress protein GP96/GRP94. Proc. Natl. Acad. Sci. U.S.A. 93, 6135–6139. Norbury, C. C., Hewlett, L. J., Prescott, A. R., Shastri, N., and Watts, C. (1995). Class I MHC presentation of exogenous soluble antigen via macropinocytosis in bone marrow macrophages. Immunity 3, 783–791. Norbury, C. C., Chambers, B. J., Prescott, A. R., Ljunggren, H. G., and Watts, C. (1997). Constitutive macropinocytosis allows TAP-dependent major histocompatibility complex class I presentation of exogenous soluble antigen by bone marrow-derived dendritic cells. Eur. J. Immunol. 27, 280–288. Oh, Y. K., Harding, C. V., and Swanson, J. A. (1997). The efﬁciency of antigen delivery from macrophage phagosomes into cytoplasm for MHC class I-restricted antigen presentation. Vaccine 15, 511–518. O’Herrin, S. M., Lebowitz, M. S., Bieler, J. G., al-Ramadi, B. K., Utz, U., Bothwell, A. L., and Schneck, J. P. (1997). Analysis of the expression of peptide–major histocompatibility complexes using high afﬁnity soluble divalent T cell receptors. J. Exp. Med. 186, 1333– 1345. Ojcius, D. M., Langlade-Demoyen, P., Gachelin, G., and Kourilsky, P. (1994). Role for MHC class I molecules in selecting and protecting high afﬁnity peptides in the presence of proteases. J. Immunol. 151, 2798–2810. Ortmann, B., Copeman, J., Lehner, P. J., Sadasivan, B., Herberg, J. A., Grandea, A. G., Riddell, S. R., Tampe´, R., Spies, T., Trowsdale, J., and Cresswell, P. (1997). A critical role for tapasin in the assembly and function of multimeric MHC class I-TAP complexes. Science 277, 1306–1309. Oukka, M., Galour, M., Belkaid, Y., Tricotet, V., Riche, N., Reynes, M., and Kosmatopoulos, K. (1997). Major histocompatibility complex class I presentation of exogenously acquired minor alloantigens initiates skin allograft rejection. Eur. J. Immunol. 27, 3499–4506. Pamer, E., and Cresswell, P. (1998). Mechanisms of MHC class I-restricted antigen processing. Annu. Rev. Immunol. 16, 323–358.

72

JONATHAN W. YEWDELL et al.

Pamer, E. G., Sijts, A. J., Villanueva, M. S., Busch, D. H., and Vijh, S. (1997). MHC class I antigen processing of Listeria monocytogenes proteins: Implications for dominant and subdominant CTL responses. Immunol. Rev. 158, 129–136. Parham, P. (1997). Immunity to Listeria monocytogenes. Immunol. Rev. 158. Peh, C. A., Burrows, S. R., Barnden, M., Khanna, R., Cresswell, P., Moss, D. J., and McCluskey, J. (1998). HLA-B27-restricted antigen presentation in the absence of tapasin reveals polymorphism in mechanisms of HLA class I peptide loading. Immunity 8, 531–542. Pfeifer, J. D., Wick, M. J., Roberts, R. L., Findlay, K., Normark, S. J., and Harding, C. V. (1993). Phagocytic processing of bacterial antigens for class I MHC presentation to T cells. Nature (London) 361, 359–362. Pond, L., and Watts, C. (1997). Characterization of transport of newly assembled, T cellstimulatory MHC class II–peptide complexes from MHC class II compartments to the cell surface. J. Immunol. 159, 543–553. Porcelli, S. A., Segelke, B. W., Sugita, M., Wilson, I. A., and Brenner, M. B. (1998). The CD1 family of lipid antigen-presenting molecules. Immunol. Today 19, 362–368. Porgador, A., Yewdell, J. W., Deng, Y., Bennink, J. R., and Germain, R. N. (1997). Localization, quantitation, and in situ detection of specﬁc peptide–MHC class I complexes using a monoclonal antibody. Immunity 6, 715–726. Porgador, A., Irvine, K. R., Iwasaki, A., Barber, B. H., Restifo, N. P., and Germain, R. N. (1998). Predominant role for directly transfected dendritic cells in antigen presentation to CD8⫹ T cells after gene gun immunization. J. Exp. Med. 188, 1075–1082. Powis, S. J., Young, L. L., Joly, E., Barker, P. J., Richardson, L., Brandt, R. P., Melief, C. J., Howard, J. C., and Butcher, G. W. (1996). The rat cim effect: TAP allele-dependent changes in a class I MHC anchor motif and evidence against C-terminal trimming of peptides in the ER. Immunity 4, 159–165. Rammensee, H.-G., Bachmann, J., and Stevanovic, S. (1997). ‘‘MHC Ligands and Peptide Motifs.’’ Landes Bioscience, Austin, Texas. Raposo, G., Nijman, H. W., Stoorvogel, W., Liejendekker, R., Harding, C. V., Melief, C. J., and Geuze, H. J. (1996). B lymphocytes secrete antigen-presenting vesicles. J. Exp. Med. 183, 1161–1172. Raz, E., Carson, D. A., Parker, S. E., Parr, T. B., Abai, A. M., Aichinger, G., Gromkowski, S. H., Singh, M., Lew, D., and Yankauckas, M. A. (1994). Intradermal gene immunization: The possible role of DNA uptake in the induction of cellular immunity to viruses. Proc. Natl. Acad. Sci. U.S.A. 91, 9519–9523. Reid, P. A., and Watts, C. (1990). Cycling of cell-surface MHC glycoproteins through primaquine-sensitive intracellular compartments. Nature (London) 346, 655–657. Reis e Sousa, C., and Germain, R. N. (1995). Major histocompatibility complex class I presentation of peptides derived from soluble exogenous antigen by a subset of cells engaged in phagocytosis. J. Exp. Med. 182, 841–851. Restifo, N. P., Bacı´k, I., Irvine, K. R., Yewdell, J. W., McCabe, B., Anderson, R. W., Eisenlohr, L. C., Rosenberg, S. A., and Bennink, J. R. (1995). Antigen processing in vivo and the elicitation of primary CTL responses. J. Immunol. 154, 4414–4422. Riddell, S. R., Rabin, M., Geballe, A. P., Britt, W. J., and Greenberg, P. D. (1991). Class I MHC-restricted ctytoxic T lymphocyte recognition of cells infected with human cytomegalovirus does not require endogenous viral gene expression. J. Immunol. 146, 2795–2804. Rock, K. L., Gamble, S., and Rothstein, L. (1990). Presentation of exogenous antigen with class I major histocompatibility complex molecules. Science 249, 918–921. Rock, K. L., Gamble, S., Rothstein, L., Gramm, C., and Benacerraf, B. (1991). Dissociation of 웁2-microgluobulin leads to the accumulation of a substantial pool of inactive class I MHC heavy chains on the cell surface. Cell 65, 611–620.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

73

Rock, K. L., Rothstein, L., Gamble, S., and Fleischacker, C. (1993). Characterization of antigen-presenting cells that present exogenous antigens in association with class I MHC molecules. J. Immunol. 150, 438–446. Rock, K. L., Gramm, C., Rothstein, L., Clark, K., Stein, R., Dick, L., Hwang, D., and Goldberg, A. L. (1994). Inhibitors of the proteasome block the degradation of most cell proteins and the generation of peptides presented on MHC class I molecules. Cell 78, 761–771. Rubartelli, A., Poggi, A., and Zocchi, M. R. (1997). The selective engulfment of apoptotic bodies by dendritic cells is mediated by the alpha(v)beta3 integrin and requires intracellular and extracellular calcium. Eur. J. Immunol. 27, 1893–1900. Russmann, H., Shams, H., Poblete, F., Fu, Y., Galan, J. E., and Donis, R. O. (1998). Delivery of epitopes by the Salmonella type III secretion system for vaccine development. Science 281, 565–568. Sandvig, K., and van Deurs, B. (1994). Endocytosis without clathrin. Trends Cell Biol. 4, 275–277. Saron, M. F., Fayolle, C., Sebo, P., Ladant, D., Ullmann, A., and Leclerc, C. (1997). Anti-viral protection conferred by recombinant adenylate cyclase toxins from Bordetella pertussis carrying a CD8⫹ T cell epitope from lymphocytic choriomeningitis virus. Proc. Natl. Acad. Sci. U.S.A. 94, 3314–3319. Savill, J., Fadok, V., Henson, P., and Haslett, C. (1993). Phagocyte recognition of cells undergoing apoptosis. Immunol. Today 14, 131–136. Schirmbeck, R., and Reimann, J. (1996). ‘Empty’ Ld molecules capture peptides from endocytosed B surface antigen particles for major histocompatibility complex class Irestricted presentation. Eur. J. Immunol. 26, 2812–2822. Schirmbeck, R., Bo¨hm, W., and Reimann, J. (1994a). Injection of detergent-denatured ovalbumin primes murine class I-restricted cytotoxic T cells in vivo. Eur. J. Immunol. 24, 2068–2072. Schirmbeck, R., Melber, K., Kuhrober, A., Janowicz, Z. A., and Reimann, J. (1994b). Immunization with soluble hepatitis B virus surface protein elicits murine H-2 class I-restricted CD8⫹ cytotoxic T lymphocyte responses in vivo. J. Immunol. 152, 1110–1119. Schirmbeck, R., Melber, K., Mertens, T., and Reimann, J. (1994c). Selective stimulation of murine cytotoxic T cell and antibody responses by particulate or monomeric hepatitis B virus surface (S) antigen. Eur. J. Immunol. 24, 1088–1096. Schirmbeck, R., Bohm, W., Melber, K., and Reimann, J. (1995a). Processing of exogenous heat-aggregated (denatured) and particulate (native) hepatitis B surface antigen for class I-restricted epitope presentation. J. Immunol. 155, 4676–4684. Schirmbeck, R., Melber, K., and Reimann, J. (1995b). Hepatitis B virus small surface antigen particles are processed in a novel endosomal pathway for major histocompatibility complex class I-restricted epitope presentation. Eur. J. Immunol. 25, 1063–1070. Schirmbeck, R., Thoma, S., and Reimann, J. (1997). Processing of exogeneous hepatitis B surface antigen particles for Ld-restricted epitope presentation depends on exogenous 웁2microglobulin. Eur. J. Immunol. 27, 3471–3484. Schmid, S. L. (1997). Clathrin-coated vesicle formation and protein sorting: An integrated process. Annu. Rev. Biochem. 66, 511–548. Schoenberger, S. P., van der Voort, E. I. Krietemeijer, G. M., Offringa, R., Melief, C. J., and Toes, R. E. (1998). Cross-priming of CTL responses in vivo does not require antigenic peptides in the endoplasmic reticulum of immunizing cells. J. Immunol. 161, 3808–3812. Schon-Hegrad, M. A., Oliver, J., McMenamin, P. G., and Holt, P. G. (1991). Studies on the density, distribution, and surface phenotype of intraepithelial class II major histocom-

74

JONATHAN W. YEWDELL et al.

patibility complex antigen (Ia)-bearing dendritic cells (DC) in the conducting airways. J. Exp. Med. 173, 1345–1356. Schrader, J. W., and Edelman, G. M. (1977). Joint recognition by cytotoxic T cells of inactivated Sendai virus and products of the major histocompatibility complex. J. Exp. Med. 145, 523–539. Schutze-Redelmeier, M. P., Gournier, H., Garcia-Pons, F., Moussa, M., Joliot, A. H., Volovitch, M., Prochiantz, A., and Lemonnier, F. A. (1996). Introduction of exogenous antigens into the MHC class I processing and presentation pathway by Drosophila antennapedia homeodomain primes cytotoxic T cells in vivo. J. Immunol. 157, 650–655. Sedlik, C., Saron, M., Sarraseca, J., Casal, I., and Leclerc, C. (1997). Recombinant parvoviruslike particles as an antigen carrier: A novel nonreplicative exogenous antigen to elicit protective antiviral cytotoxic T cells. Proc. Natl. Acad. Sci. U.S.A. 94, 7503–7508. Shen, Z., Reznikoff, G., Dranoff, G., and Rock, K. L. (1997). Cloned dendritic cells can present exogenous antigens on both MHC class I and class II molecules. J. Immunol. 158, 2723–2730. Shen, H., Miller, J. F., Fan, X., Kolwyck, D., Ahmed, R., and Harty, J. T. (1998). Compartmentalization of bacterial antigens: Differential effects on priming of CD8 T cells and protective immunity. Cell 92, 535–545. Snyder, H. L., Yewdell, J. W., and Bennink, J. R. (1994). Trimming of antigenic peptides in an early secretory compartment. J. Exp. Med. 180, 2389–2394. Snyder, H. L., Bacik, I., Yewdell, J. W., Behrens, T. W., and Bennink, J. R. (1998). Promiscuous liberation of MHC class I-binding peptides from the C termini of membrane and soluble proteins in the secretory pathway. Eur. J. Immunol. 28, 1339–1346. Song, R., and Harding, C. V. (1996). Roles of proteasomes, transporter for antigen presentation (TAP), and 웁2-microglobulin in the processing of bacterial or particulate antigens via an alternate class I MHC processing pathway. J. Immunol. 156, 4182–4190. Spee, P., and Neefjes, J. (1997). TAP-translocated peptides speciﬁcally bind proteins in the endoplasmic reticulum, including gp96, protein disulﬁde isomerase and calreticulin. Eur. J. Immunol. 27, 2441–2449. Speidel, K., Osen, W., Faath, S., Hilgert, I., Obst, R., Braspenning, J., Momburg, F., Hammerling, G. J., and Rammensee, H. G. (1997). Priming of cytotoxic T lymphocytes by ﬁve heat-aggregated antigens in vivo: Conditions, efﬁciency, and relation to antibody responses. Eur. J. Immunol. 27, 2391–2399. Srivastava, P. K. (1993). Peptide-binding heat shock proteins in the endoplasmic reticulum: Role in immune response to cancer and in antigen presentation. Adv. Cancer Res. 62, 153–177. Srivastava, P. K., DeLeo, A. B., and Old, L. J. (1986). Tumor rejection antigens of chemically induced sarcomas of inbred mice. Proc. Natl. Acad. Sci. U.S.A. 83, 3407–3411. Srivastava, P. K., Udono, H., Blachere, N. E., and Li, Z. (1994). Heat shock proteins transfer peptides during antigen processing and CTL priming. Immunogenetics 39, 93–98. Srivastava, P. K., Menoret, A., Basu, S., Binder, R. J., and McQuade, K. L. (1998). Heat shock proteins come of age: Primitive functions acquire new roles in an adaptive world. Immunity 8, 657–665. Steinman, R. M. (1998). Dendritic cells. In ‘‘Fundamental Immunology’’ (W. E. Paul, ed.), pp. 547–604. Lippincott-Raven, Philadelphia and New York. Stryhn, A., Pedersen, L. O., Romme, T., Olsen, A. C., Nissen, M. H., Thorpe, C. J., and Buus, S. (1996). pH dependence of MHC class I-restricted peptide presentation. J. Immunol. 156, 4191–4197. Sugamura, K., Shimizu, K., and Bach, F. H. (1978). Involvement of fusion activity of ultraviolet light-inactivated Sendai virus in formation of target antigens recognized by cytotoxic T cells. J. Exp. Med. 148, 276–287.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

75

Sugita, M., and Brenner, M. B. (1995). Association of the invariant chain with major histocompatibility complex class I molecules directs trafﬁcking to endocytic compartments. J. Biol. Chem. 270, 1443–1448. Suto, R., and Srivastava, P. K. (1995). A mechanism for the speciﬁc immunogenicity of heat shock protein-chaperoned peptides. Science 269, 1585–1588. Suzue, K., Zhou, X., Eisen, H. N., and Young, R. A. (1997). Heat shock fusion proteins as vehicles for antigen delivery into the major histocompatibility complex class I presentation pathway. Proc. Natl. Acad. Sci. U.S.A. 94, 13146–13151. Svensson, M., and Wick, M. J. (1999). Classical MHC-1 peptide presentation of a bacterial fusion protein by bone marrow derived dendritic cells. Eur. J. Immunol. 29, 180–188. Swanson, J. A., and Bauer, S. C. (1995). Phagocytosis by zippers and triggers. Trends Cell Biol. 5, 89–93. Swanson, J. A., and Watts, C. (1995). Macropinocytosis. Trends Cell Biol. 5, 428. Szalay, G., Hess, J., and Kaufmann, S. H. (1994). Presentation of Listeria monocytogenes antigens by major histocompatibility complex class I molecules to CD8 cytotoxic T lymphocytes independent of listeriolysin secretion and virulence. Eur. J. Immunol. 24, 1471–1477. Tamura, Y., Peng, P., Liu, K., Daou, M., and Srivastava, P. K. (1997). Immunotherapy of tumors with autologous tumor-derived heat shock protein preparations. Science 278, 117–120. Tanaka, K., Tanahashi, N., Tsurumi, C., Yokota, K. Y., and Shimbara, N. (1997). Proteasomes and antigen processing. Adv. Immunol. 64, 1–38. Tevethia, S. S., Flyer, D. C., and Tijan, R. (1980). Biology of simian virus 40 (SV40) transplantation antigen (TrAg). VI. Mechanism of induction of SV40 transplantation immunity in mice by puriﬁed SV40 T antigen (D2 protein). Virology 107, 13–23. Tourne, S., van Santen, H. M., van Roon, M., Berns, A., Benoist, C., Mathis, D., and Ploegh, H. (1996). Biosynthesis of major histocompatibility complex molecules and generation of T cells in Ii TAP1 double-mutant mice. Proc. Natl. Acad. Sci. U.S.A. 93, 1464–1469. Townsend, A. R. M., Rothbard, J., Gotch, F. M., Bahadur, G., Wraith, D., and McMichael, A. J. (1986). The epitopes of inﬂuenza nucleoprotein recognized by cytotoxic T lymphocytes can be deﬁned with short synthetic peptides. Cell 44, 959–968. Udono, H., and Srivastava, P. K. (1993). Heat shock protein 70-associated peptides elicit speciﬁc cancer immunity. J. Exp. Med. 178, 1391–1396. Udono, H., Levey, D. L., and Srivastava, P. K. (1994). Cellular requirements for tumorspeciﬁc immunity elicited by heat shock proteins: Tumor rejection antigen gp96 primes CD8⫹ T cells in vivo. Proc. Natl. Acad. Sci. U.S.A. 91, 3077–3081. Ullrich, S. J., Robinson, E. A., Law, L. W., Willingham, M., and Appella, E. (1986). A mouse tumor-speciﬁc transplantation antigen is a heat shock-related protein. Proc. Natl. Acad. Sci. U.S.A. 83, 3121–3125. Ulmer, J. B., Donnelly, J. J., and Liu, M. A. (1994). Presentation of an exogenous antigen by major histocompatibility complex class I molecules. Eur. J. Immunol. 24, 1590–1596. Ulmer, J. B., Deck, R. R., Dewitt, C. M., Donnhly, J. I., and Liu, M. A. (1996). Generation of MHC class I-restricted cytotoxic T lymphocytes by expression of a viral protein in muscle cells: Antigen presentation by non-muscle cells. Immunology 89, 59–67. van Rooijen, N., and Sanders, A. (1994). Liposome mediated depletion of macrophages: Mechanism of action, preparation of liposomes and applications. J. Immunol. Methods 174, 83–93. Vigna, J. L., Smith, K. D., and Lutz, C. T. (1996). Invariant chain association with MHC class I: Preference for HLA class I/beta 2-microglobulin heterodimers, speciﬁcity, and inﬂuence of the MHC peptide-binding groove. J. Immunol. 157, 4503–4510.

76

JONATHAN W. YEWDELL et al.

Villacres, M. C., Behboudi, S., Nikkila, T., Lovgren-Bengtsson, K., and Morein, B. (1998). Internalization of iscom-borne antigens and presentation under MHC class I or class II restriction. Cell Immunol. 185, 30–38. Vinitsky, A., Anto´n, L. C., Snyder, H. L., Orlowski, M., Bennink, J. R., and Yewdell, J. W. (1997). The generation of MHC class I-associated peptides is only partially inhibited by proteasome inhibitors: Involvement of nonproteasomal cyosolic proteases in antigen processing? J. Immunol. 159, 554–564. Wagner, R., Deml, L., Schirmbeck, R., Niedrig, M., Reimann, J., and Wolf, H. (1996). Construction, expression, and immunogenicity of chimeric HIV-1 virus- like particles. Virology 220, 128–140. Watts, C., and Marsh, M. (1992). Endocytosis: What goes in and how? J. Cell Sci. 103, 1–8. Wearsch, P. A., and Nicchitta, C. V. (1997). Interaction of endoplasmic reticulum chaperone GRP94 with peptide substrates is adenine nucleotide-independent. J. Biol. Chem. 272, 5152–5156. Weidt, G., Utermo¨hlen, O., Reimann, J., Deppert, W., and Lehmann-Grube, F. (1994). CD8⫹ T lymphocyte-mediated antiviral immunity in mice as a result of injection of recombinant viral proteins. J. Immunol. 153, 2554–2560. Wells, A. D., Rai, S. K., Salvato, M. S., Band, H., and Malkovsky, M. (1998). Hsp72mediated augmentation of MHC class I surface expression and endogenous antigen presentation. Int. Immunol. 10, 609–617. Wick, M. J., and Pfeifer, J. D. (1996). Major histocompatibility class I presentation of ovalbumin peptide 257–264 from exogenous sources: Protein context inﬂuences the degree of TAP-independent presentation. Eur. J. Immunol. 26, 2790–2799. Wijburg, O. L., van den Dobbelsteen, G. P., Vadolas, J., Sanders, A., Strugnell, R. A., and van Rooijen, N. (1998). The role of macrophages in the induction and regulation of immunity elicited by exogenous antigens. Eur. J. Immunol. 28, 479–487. Wiktor, T. J., Doherty, P. C., and Koprowski, H. (1977). In vitro evidence of cell-mediated immunity after exposure of mice to both live and inactivated rabies virus. Proc. Natl. Acad. Sci. U.S.A. 74, 334–338. Winkel, K. D., Kronin, V., Krummel, M. F., and Shortman, K. (1997). The nature of the signals regulating CD8 T cell proliferative responses to CD8alpha⫹ or CD8alpha⫺ dendritic cells. Eur. J. Immunol. 27, 3350–3359. Xiao, X., Li, J., and Samulski, R. J. (1996). Efﬁcient long-term gene transfer into muscle tissue of immunocompetent mice by adeno-associated virus vector. J. Virol. 70, 8098–8108. Yellen-Shaw, A. J., Wherry, E. J., Dubois, G. C., and Eisenlohr, L. C. (1997). Point mutation ﬂanking a CTL epitope ablates in vitro and in vivo recognition of a full-length viral protein. J. Immunol. 158, 3227–3234. Yewdell, J. W., and Bennink, J. R. (1990). The binary logic of antigen processing and presentation to T cells. Cell 62, 203–206. Yewdell, J. W., and Bennink, J. R. (1992). Cell biology of antigen processing and presentation to MHC class I molecule-restricted T lymphocytes. Adv. Immunol. 52, 1–123. Yewdell, J. W., and Bennink, J. R. (1999). Immunodominance in MHC Class I restricted responses. Annu. Rev. Immunol. 17, 51–88. Yewdell, J. W., Anto´n, L. C., and Bennink, J. R. (1996). Defective ribosomal products (DRiPs). A major source of antigenic peptides for MHC class I molecules? J. Immunol. 157, 1823–1826. Yewdell, J. W., Bennink, J. R., and Hosaka, Y. (1988). Cells process exogenous proteins for recognition by cytotoxic T lymphocytes. Science 239, 637–640. York, I., and Rock, K. L. (1996). Antigen processing and presentation by the class I major histocompatibility complex. Annu. Rev. Immunol. 14, 369–396.

ANTIGEN PRESENTATION BY MHC CLASS I MOLECULES

77

Yoshida, H., Hayashi, S., Kunisada, T., Ogawa, M., Nishikawa, S., Okamura, H., Sudo, T., and Shultz, L. D. (1990). The murine mutation osteopetrosis is in the coding region of the macrophage colony stimulating factor gene. Nature (London) 345, 442–444. Zhou, F., and Huang, L. (1998). Delivery of protein antigen to the major histocompatibility complex class I-restricted antigen presentation pathway. J. Drug Targeting 3, 91–109. Zhou, X., Glas, R., Liu, T., Ljunggren, H.-G., and Jondal, M. (1993). Antigen processing mutant T2 cells present viral antigen restricted through H-2Kb. Eur. J. Immunol. 23, 1802– 1808. Zhou, F., Watkins, S. C., and Huang, L. (1994). Characterization and kinetics of MHC class I-restricted presentation of a soluble antigen delivered by liposomes. Immunobiology 190, 35–52. Zinkernagel, R. M., and Doherty, P. C. (1974). Restriction of in vitro T cell-mediated cytotoxicity in lymphocytic choriomeningitis within a syngeneic or semiallogeneic system. Nature (London) 248, 701–702. Zitvogel, L., Regnault, A., Lozier, A., Wolfers, J., Flament, C., Tenza, D., Ricciardi-Castagnoli, P., Raposo, G., and Amigorena, S. (1998). Eradication of established murine tumors using a novel cell-free vaccine: Dendritic cell-derived exosomes. Nature Med. 4, 594–600. Zwickey, H. L., and Potter, T. A. (1996). Peptide epitopes from noncytosolic Listeria monocytogenes can be presented by major histocompatibility complex class I molecules. Infect. Immun. 64, 1870–1872.

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 73

Signal Transduction Pathways That Regulate the Fate of B Lymphocytes ANDREW CRAXTON,* KEVIN L. OTIPOBY,† AIMIN JIANG,* AND EDWARD A. CLARK*,† *Department of Microbiology and †Department of Immunology, University of Washington, Seattle, Washington 98195

I. Introduction

A. SIGNALING PATHWAYS REGULATING CELL FATE The size of any population, whether it be a cell or human population, is controlled by three factors: birth rate, death rate, and net immigration/ emigration rate (Raff, 1996). Thus, a full understanding of how populations of lymphocytes are dramatically increased and decreased during development or during immune responses will require insight into the mechanisms regulating both cell birth and cell death. Research over the past 30 years has led to signiﬁcant progress in understanding how cell proliferation works and is regulated. There are many excellent reviews and books on the cell cycle (e.g., Murray and Hunt, 1993; Sherr, 1996; Dynlacht, 1997; Fisher, 1997; S. I. Reed, 1997; Jackman and Pines, 1997), so this topic will not be discussed further here. However, the complexity and importance of the mechanisms regulating cell death have only recently begun to be fully appreciated (Raff, 1996). Thus, as a prelude to discussing signaling pathways regulating B cell fate, we begin with a brief overview on the rapidly advancing area of programmed cell death. A major breakthrough occurred when studies using the Caenorhabditis elegans model revealed key genes that control the onset of programmed death of 131 of the 1090 somatic cells generated during development of the worm (Horvitz et al., 1983; Hedgecock et al., 1983; Ellis and Horvitz, 1986). Two of these genes, ced-3 and ced-4 (cell death abnormal 3 and 4), are required for cell death to occur, and a third gene, ced-9, is necessary to prevent cells that normally survive from undergoing apoptosis (Hengartner et al., 1992). During the past 5 years there has been an explosion of knowledge related to cell death pathways and the genes regulating them (Yang and Korsmeyer, 1996, Chinnaiyan and Dixit, 1996, 1997; J. C. Reed, 1996, 1997; White, 1996; Chao and Korsmeyer, 1998; Newton and Strasser, 1998; Green, 1998). Links between kinases, death proteases, and Bcl-2 family members have been identiﬁed, underscoring that the induction of cell death, just like the induction of cell proliferation, is a highly regulated process. Critical discoveries include the deﬁnition of a set of death proteases that are essential for many cell death pathways, identiﬁcation of the mammalian 79

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

80

ANDREW CRAXTON et al.

equivalents of ced-3, -4 and -9, and the deﬁnition of new families of natural inhibitors of death proteases (Table I). The mammalian counterpart for ced-3 is related to the interleukin-1converting enzyme (ICE), a cysteine protease (Yuan et al., 1993), and to other mammalian cysteine aspartate-speciﬁc proteases, or ‘‘caspases’’ (e.g., Fernandes-Alnemri et al., 1994; Wang et al., 1994; Alnemri et al., 1996; Alnemri, 1997; Salvesen and Dixit, 1997). The size of the caspase family is growing and now includes at least 12 members (Salvesen and Dixit, 1997). Ced-9 is related to Bcl-2 (Vaux et al., 1992; Hengartner and Horvitz, 1994) and other Bcl-2 family members such as Bcl-x ( J. Reed, 1997; Chao and Korsmeyer, 1998). The size of the Bcl-2 family is also growing and includes at least 15 members—mammalian and viral (Rao and White, 1997)—some of which are generally antiapoptotic (Bcl-2, Bcl-x, and A1), whereas others are usually proapoptotic (Bax, Bik, Bak, and Bad). A mammalian ced-4 equivalent, Apaf-1, has been discovered (Zou et al., 1997), and quite possibly represents the discovery of the ﬁrst member of products of another family of genes. Ced-3, ced-4, and ced-9 have been found to interact directly with each other (Chinnaiyan et al., 1997; James et al., 1997; Wu et al., 1997a; Spector et al., 1997), as have caspase-9, Bcl-xL, and Apaf-1 (Pan et al., 1998). The various members of the Bcl-2 family can form heterodimers with each other (Chao and Korsmeyer, 1998), and caspases may interact with each other (Salvesen and Dixit, 1997). Thus, the possibilities for interactions between the various mammalian homologues of ced-3, -4, and -9 are quite staggering (Chao and Korsmeyer, 1998; Newton and Strasser, 1998; Green, 1998). B. TUMOR NECROSIS FACTOR AND RECEPTOR FAMILIES A unique family of growth-regulating receptors was ﬁrst discovered in 1989 when two groups noted that CD40 and nerve growth factor (NGF) receptor were related (Braesch-Andersen et al., 1989; Stamenkovic et al., 1989). When tumor necrosis factor (TNF) receptors 1 and 2 (TNFR1, TNFR2) and rat OX40 were cloned the following year, the term ‘‘NGF/ TNF receptor family’’ was coined and subsequently adopted. Today it is clear that the TNF superfamily and the NGF/TNF receptor superfamily play essential roles in regulating the fate of cells during development and inﬂammatory processes (Gruss and Dower, 1995; Baker and Reddy, 1996). Both these families expanded dramatically with the recent discovery of the TNF members (Table II), TRAIL/Apo-2 (Wiley et al., 1995; Pitti et al., 1996), RANKL/TRANCE (Anderson et al., 1997; Wong et al., 1997a,b), and LIGHT (Mauri et al., 1998), along with their NGF/TNF family receptors, DR4/TRAIL-R1 (Pan et al., 1997a), DR5/TRAIL-R2 (Pan et al.,

TABLE I MAMMALIAN DEATH PATHWAY-ASSOCIATED GENE FAMILIES Death Gene Family

Examples

Approx. No.a

Special Features

Death domain (DD)

FADD, TRADD, DAP kinase, RIP FADD, FLIP Caspase-8, -10 RAIDD, Ced-4 ARC, Caspase-2 Caspase-1, -3 cIAP1, cIAP2, NAIP, XIAP

7

Cytosolic proteins binding each other or DD receptor subgroup

4

Links receptor complexes to caspase Specialized regulator (FLIP/MRIT/Casper) Possible alternative death pathway

Death effector domain (DED) CARD domain Caspases Inhibitors of apoptosis (IAPs) TRAFs Bcl-2 family Serpins a

TRAF1, 2, 3 Bcl-2, Bax Bcl-x, Bad PAI-2, PI-9

4 12 5

Protease initiators/effectors of death Caspase binding, inhibitors of apoptosis, bind TRAFs

6 15

Antiapoptotic via activation of NF-␬B Antiapoptotic or proapoptotic Associated with mitochondria Inhibit caspase-1/other proteases

2

Numbers include some overlap due to some genes being in more than one group.

TABLE II TNF SUPERFAMILY AND NGF/TNF RECEPTOR SUPERFAMILY MEMBERS TNF Superfamily Member Name

TNF Receptor Superfamily Member

Description

Name

FasL

Activated T cells

CD95/Fas

TNF-움

—

LT-움3/LT-움2/웁1 LT-움1/웁2 LT-움3 only LIGHT TRAIL/Apo-2L

— — — Activated T cells —

TNFR1 TNFR2 TNFR1, TNFR2 LT-웁R HVEM HVEM or LT-웁R TRAIL-R1 (DR4) TRAIL-R2 (DR5) TRAIL-R3 (TRID) TRAIL-R4

Apo-3L

—

CD40L (CD154)

Activated T cells, DCs

OPG, FDCR-1, or OCIF DR3/WSL-1/Apo-3, TRAMP/LARD CD40

RANKL/TRANCE

—

RANK

CD70 4-1BBL

— —

OPG/FDCR1 CD27 4-1BB (CDw137)

CD30L(CD153)

—

CD30

OX40L

DCs, B cells, endothelial cells

OX40 (CD134)

Description Cloned 1991, broadly distributed, has DD, binds FADD and FLICE, CD40-regulated Cloned 1990, has DD, binds TRADD and FADD Cloned 1990, no DD, binds TRAF1/2, activates NF-␬B — Cloned 1994, binds TRAF2/3/5, activates NF-␬B Cloned 1996, binds HSV glycoprotein D, on activated T cells Binds TRAF1/2/3/5, activates NF-␬B Cloned 1997, has DD, widespread tissue distribution Cloned 1997, has DD, widespread tissue distribution Cloned 1997, no cytoplasmic domain, blocks cell death. Cloned 1997, widespread tissue distribution, no DD, activates NF␬B, inhibits cell death Cloned 1997, ﬁbroblasts, FDCs, DCs, soluble and membrane forms, CD40-regulated Cloned 1996, activated T cells, non-Hodgkin’s lymphomas, two isoforms, binds TRADD and FADD, activates NF-␬B Cloned 1989, B, DCs, FDCs, epithelial cells, no DD, binds TRAF2/3/5/6, activates NF-␬B Cloned 1997, widespread tissue distribution, DCs, CD40-regulated, activates NF-␬B Cloned 1991, B, T, and NK cells Cloned 1989, costimulatory molecule on T cells, binds TRAF1/2, activates NF-␬B Cloned 1992, on activated T and B cells, binds TRAF1/2/3, activates NF-␬B Cloned 1990, on activated T cells, binds TRAF2/3, activates NF-␬B

B CELL SIGNAL TRANSDUCTION PATHWAYS

83

1997b; Sheridan et al., 1997; Walczak et al., 1997), TRAIL-R3/TRID/ DcR1, TRAIL-R4 (Degli-Esposti et al., 1997a,b; Pan et al., 1997b; Sheridan et al., 1997), RANK, OPG/FDCR-1, and HVEM (Simonet et al., 1997; Degli-Esposti et al., 1997b; Anderson et al., 1997; Mauri et al., 1998; Yun et al., 1998). The extracellular regions of NGF/TNF receptor family members have the common characteristic of homologous repeating cysteine-rich domains. A subset of NGF/TNF receptor family members (Table II) contain within their cytoplasmic tails a region called the ‘‘death domain’’ (DD) also found within certain cytoplasmic proteins such as FADD and RIP (Table I) (Cleveland and Ihle, 1995). The death domain of CD95, for example, can interact with the death domain of FADD. Cytosol-associated DDs apparently function to activate the death pathway by linking up to caspases and other key death pathway proteins that contain specialized domains called death effector domains (DEDs); a related death pathway uses caspase recruitment domains (CARDs) (Table I). Yet another family of proteins, the so-called inhibitors of apoptosis proteins (IAPs), apparently bind to caspases and the TNF receptor-associated factor (TRAF) family of molecules (see Sections IV,B and IV,F) to regulate death pathways. Various studies have emphasized that cell death phenotypes are regulated at a number of different steps in the death pathway. The IAPs may function at the initiation of the death pathway (Uren et al., 1996) or by directly binding to and inhibiting effector caspases such as caspase-3 and caspase-7 (Deveraux et al., 1997, 1998; Roy et al., 1997). Some IAPs can bind both to membrane-associated TRAF family members involved in initiating TNFR family signaling pathways (Ware et al., 1996) and to downstream effector caspases (Roy et al., 1997), suggesting that they may function at more than one point in the death pathway. Another inhibitor, sentrin, binds directly to the DD in TNFR1 and Fas to block death (Okura et al., 1996). There are at least two members in the sentrin family (Kamitani et al., 1998), and sentrins can apparently be conjugated to other proteins in a manner analogous to protein ubiquitination (Kamitani et al., 1997). Precisely how sentrins may regulate cell death is not yet understood. Yet another key regulator of cell death is the caspaselike molecule called Casper, FLIP, or MRIT (Shu et al., 1997; Wallach, 1997; Han et al., 1997); like the death pathway initiators, caspase-8 and caspase10, FLIP has multiple RNA isoforms that may express proteins with different functions, e.g., blocking an early step in cell death by interfering with caspase-8 or interacting with and regulating Bcl-2 family members. Viral serine proteinase inhibitors (serpins) such as CrmA clearly can inhibit caspase-dependent cell death, but relatively little is known about how mammalian serpin counterparts may regulate death pathways. Cells

84

ANDREW CRAXTON et al.

transfected with a cDNA encoding the serpin, plasminogen activator inhibitor 2 (PAI-2), are resistant to TNF-induced cell death (Kumar and Baglioni, 1991; Dickinson et al., 1995), and PAI-2 can block Mycobacterium aviuminduced apoptosis of macrophages (Gan et al., 1995). The serpin PI-9 is a potent inhibitor of granzyme B and is expressed at high levels in lymphocytes (Sun et al., 1996, 1997). Thus, it is an outstanding candidate for a regulator of B and T lymphocyte fate. The partial list of ⬎70 death pathway-associated genes in Tables I and II (which do not include kinase regulators) emphasizes the complexity of death pathways and genes. A number of these gene products have dual functions, e.g., they may be proapoptotic in one context but antiapoptotic in another. Even within the same death receptor DD-containing subgroup, i.e., the set of receptors for TRAIL, there are both proapoptotic receptors and a ‘‘decoy receptor,’’ DcR1, which interferes with cell death pathways (Pan et al., 1997b; Sheridan et al., 1997). There is no question that the mechanisms regulating cell death pathways and the phenotypes associated with cell death are both numerous and complex. This list also does not take into consideration the possible functions of caspases beyond inducing programmed cell death, such as protecting cells from necrotic death (Vercammen et al., 1998), or the possible other functions of so-called death inducers such as FADD, which may also regulate T cell proliferation (Newton et al., 1998; Zhang et al., 1998a). Nor does room allow for a discussion of the downstream targets of caspases such as serine/threonine kinases (Graves et al., 1998b) survival factors, or an inhibitor of cytosolic DNase (Sakahira et al., 1998; Enari et al., 1998). Clearly much needs to be learned about the multiple functions of death-associated signaling pathways. C. B CELL SURVIVAL AND DEATH Of the many surface molecules on B lymphocytes (Clark and Ledbetter, 1994), three in particular have been clearly implicated in regulating B cell fate: the B cell receptor (BCR) complex, CD40, and Fas/CD95. Signaling through the BCR has different consequences depending on the developmental stage of the B cell and factors such as the strength and duration of the signal (Goodnow, 1996). For example, the BCR can (1) induce cell death, (2) stimulate B cell survival, and (3) protect B cells from Fas/CD95induced cell death (Fig. 1); it is essential both for maintaining normal levels of B cells in mice (Lam et al., 1997), and for induction of tolerance to autoantigens (Goodnow, 1996; Melamed et al., 1998). Just how the various BCR-induced signaling pathways lead to certain phenotypes is not fully understood and is one of the major questions currently facing the ﬁeld of B cell biology (DeFranco, 1997; Law et al., 1996a). We will return

B CELL SIGNAL TRANSDUCTION PATHWAYS

85

FIG. 1. Stages when BCR, CD40, and CD95/Fas signaling regulate the fate of B lymphocytes. Immature B cells about to leave the bone marrow can be induced to die by BCR ligation (left), while signaling via CD40 may rescue immature B cells from death. Newly produced B cells migrating into the spleen arrive at the marginal zone (MZ) sinusoids and then enter a cell-rich zone: the periarteriolar lymphoid sheath (PALS) in the spleen (or paracortex in lymph nodes). Most of these new B cells emigrating to the outer splenic PALS die; signaling via the BCR can prevent cell death and trigger B cells to survive (center) and then enter follicles and form germinal centers (right). The PALS is also a key site for T–B cell and dendritic cell (DC) interactions. CD40 signaling can induce B cells both to divide and to express CD95/Fas so they can be killed via CD95-induced apoptosis (center). However, a signal to the BCR from speciﬁc antigen, e.g., associated with follicular dendritic cells (FDC), can rescue B cells from CD95-induced death via a transcriptiondependent ‘‘Fas blockade’’ (right).

to this fundamental question throughout the review. The distinct functions of the BCR noted above apparently occur in characteristic lymphoid niches (Figs. 1 and 2): BCR signals directly induce cell death of immature selfreactive B cells, which are eliminated after BCR recognition of multivalent self-antigens such as those on cell surfaces (e.g., Hartley et al., 1993; Tsubata et al., 1993, 1994). Precisely when and where during B cell lymphopoiesis cell death occurs is not entirely clear (e.g., Melamed and Nemazee, 1997), but it most likely occurs at the developmental stage when antigen (Ag) receptors are ﬁrst expressed (Lu and Osmond, 1997) and at a somewhat later ‘‘transitional B cell’’ stage (Carsetti et al., 1995). The size of the peripheral pool of B cells—approximately 1 ⫻ 108 longlived cells in mice—must also be carefully regulated. Though the bone marrow produces about 0.2 ⫻ 108 B cells/day, only a small number of these cells can be recruited into the long-lived pool (Osmond, 1991; Sprent,

86

ANDREW CRAXTON et al.

FIG. 2. Mature B lymphocyte development in germinal centers during cell-dependent responses. Resting unmutated naive B cells (Bm1) in the follicular mantle are activated in a T cell-dependent manner, then migrate into B cell follicles, where they differentiate into proliferating centroblasts (Bm3); there they form the germinal center (GC) dark zone, where somatic mutation in the variable region of the immunoglobulin genes occurs. The centroblasts (Bm3) differentiate into centrocytes (Bm4) that undergo positive selection depending on the afﬁnity of their mutated antigen receptors in the GC light zone. Low afﬁnity and autoreactive Bm3 centrocytes that are not selected undergo spontaneous apoptosis. The positively selected high-afﬁnity centrocytes (Bm4) may undergo CD40-induced isotype switching, become protected from Fas-induced apoptosis, and differentiate into either plasma cells (pc) or into memory B cells (Bm5). See MacLennan et al. (1997).

1994, 1997). Newly produced B cells migrate into the spleen, arrive at the marginal zone sinusoids, penetrate the marginal zone sinus, and then come into the outer zone of the periarteriolar lymphoid sheath (PALS). The outer PALS is a site of primary B cell proliferation in response to various antigens as well as a site where anergic, auto or self-reactive B cells may die (Cyster et al., 1994; Fulcher et al., 1996; Liu, 1997). Most newly formed B cells emigrating to the outer PALS die there but some survive if they are rescued and recruited into the long-lived pool. The signals that trigger

B CELL SIGNAL TRANSDUCTION PATHWAYS

87

rescue are not fully understood, but one is almost certainly a positive signal by Ag via the BCR (Gu et al., 1991; Liu, 1997). The self-reactive B cells may die and not home to follicles because they cannot compete with these other B cells (Cyster et al., 1994) or because of excessive signaling by Ag through the BCR (Fulcher et al., 1996). Support for this latter hypothesis has come from Cook et al., (1997), who showed that the arrest of B cells in outer PALS is determined by the magnitude of BCR stimulation. According to these results, any coreceptor or cytokine that can affect the ‘‘magnitude of BCR stimulation’’ may inﬂuence peripheral B cell maturation at this stage. Indeed, B cells that lack the protein tyrosine kinase (PTK) Syk and have defective signaling through the BCR can migrate to T cell-rich zones but fail to enter the recirculating pool (Turner et al., 1997). On the other hand, Schmidt et al., (1998) found that B cells missing the protein tyrosine phosphatase (PTPase) SHP-1 accumulate in T cell zones in the absence of their speciﬁc antigen. The relative roles of antigen receptor-mediated signals versus follicular composition in affecting B cell fate at this juncture thus remain unresolved. For T cell-dependent responses, cell–cell interaction signals such as CD40L–CD40 and FasL–Fas may affect whether peripheral B cells live or die (Fig. 1) (Goodnow, 1996; Cyster et al., 1994; Rathmell et al., 1996; Fulcher et al., 1996). Tolerant B cells in particular may be susceptible to CD95/Fas-mediated killing by FasL⫹ T cells if their antigen receptors are not engaged or have defective signaling pathways (Rothstein et al., 1995; Rothstein, 1996; Rathmell et al., 1996). The PALS, and not the germinal center, may be the region where FasL regulates B cells ( Jacobson et al., 1996), so it is likely that the degree and duration of BCR ligation in the PALS may inﬂuence subsequent signaling and local responses to B cells. In summary, ligation of the BCR may induce death of immature B cells, arrest of B cells in the outer PALS, or resistance to Fas-mediated cell death (Fig. 1). We will discuss examples of coreceptors such as CD22, CD19, and CD40, which may affect BCR signaling and B cell fate, perhaps in the PALS. D. GERMINAL CENTER FORMATION Germinal centers (GCs) are specialized microenvironments formed mainly during primary immune responses to T cell-dependent antigens. The major function of GCs is to promote the birth and survival of memory B cells, which involves processes including somatic mutation, afﬁnity maturation, and isotype class switching. The cellular and molecular mechanisms leading to GC formation in peripheral lymphoid follicles are becoming better understood (MacLennan, 1994; Clark and Ledbetter, 1994; Ahmed and Gray, 1996; MacLennan et al., 1997; Liu and Arpin, 1997; Przylepa

88

ANDREW CRAXTON et al.

et al., 1998). As noted above, most of the new B cells emigrating to the outer splenic PALS die there; they survive only if they are rescued and recruited into the long-lived pool by signals that are not fully understood. Apparently, the rare B cells with receptors for the foreign Ag bind, process, and present soluble Ag. Some CD4⫹ T cells in the T cell-rich region recognize processed Ag on antigen-presenting cells (APCs), particularly on interdigitating dendritic cells (DCs). Activated T cells now expressing CD40L, and probably also DCs, then provide signals from cell adhesion receptors or cytokines to induce a few Ag-speciﬁc B cells to migrate to follicles. Once in the follicles, these B cells associated loosely with FDCs and begin to undergo a dramatic exponential proliferation: according to some estimates they double every 6 hr (MacLennan, 1994). The recirculating B cells are displaced outward to form the follicular mantle. B cell blasts move from the follicle center (with its associated dense FDC network) to one end of the follicle near T cell-rich regions to form the dark zone of the GC (Fig. 2). The dividing centroblasts in the dark zone most likely activate a hypermutation program and then give rise to nondividing centrocytes, which move to the light zone (MacLennan et al., 1997). FDCs in the light zone with associated Ag then select those mutating B cells with high-afﬁnity receptors for Ag. These B cells in the light zone also make cognate interactions with CD4⫹ T cells in GCs, and isotype switching and B cell maturation into plasma cells or memory B cells occurs. There is one major unanswered question in particular (Liu and Arpin, 1997): What are the molecular mechanisms underlying the high rate of proliferation coupled with the high rate of cell death and somatic hypermutation in GCs? E. PERIPHERAL B CELL SUBPOPULATIONS Liu and co-workers have carefully deﬁned the phenotypes of human B cell subsets formed during T cell-dependent humoral responses (see Liu and Arpin, 1997). Following the approach used to deﬁne mouse B cell subsets (Hardy et al., 1991), Liu et al. evaluated human tonsillar B cells and ﬁrst divided them into three major subsets based on their expression of membrane immunoglobulin D (mIgD) and CD38: naive B cells (Bm1/ 2, mIgD⫹ CD38⫺), GC B cells (Bm3/4, IgD⫺ CD38⫹ ), and memory B cells (Bm5, IgD⫺ CD38⫺). The naive B cells were further subdivided on the basis of whether they did (Bm2) or did not (Bm1) express CD23, and GC B cells were subdivided on the basis of whether they did (Bm3) or did not (Bm4) express CD77. A rare IgD⫹ CD38⫹ subset was also deﬁned. These subsets were examined for their location, spontaneous or induced apoptosis, expression of cell death- or survival-associated genes, somatic mutations, and Ig isotype class switching or switched genes. More recently, Tangye et al. (1998) found that memory B cells can be identiﬁed by their

B CELL SIGNAL TRANSDUCTION PATHWAYS

89

expression of CD148 and CD27. Just how the maturation of various B cell subsets is regulated during GC formation is still not well understood. II. B Cell Antigen Receptor Complex

A. INTRODUCTION The BCR consists of two basic components (Fig. 3): (1) membrane immunoglobulins such as membrane IgM (mIgM) or IgD (mIgD), which recognize Ag, and (2) Ig-움 and Ig-웁 disulﬁde-linked heterodimers, which are essential for the development of B lineage cells (Gong and Nussezweig, 1996; Torres et al., 1996) and for mediating BCR signaling pathways (Kurosaki, 1997). The Ig-움/Ig-웁 signaling subunits invariably contain copies of a conserved immunoreceptor tyrosine-based activation motif (ITAM) within their cytoplasmic domains (Cambier, 1995). Following BCR stimulation, tyrosines in the ITAM motifs are phosphorylated; they then recruit and activate key Src homology (SH2) domain-containing proteins essential for activating downstream signaling pathways (Fig. 3). Both Ig-움 and Ig웁 are required for BCR-induced growth arrest and apoptosis (Yao et al., 1995; Tseng et al., 1997), and in order to function Ig-움 and Ig-웁 must be tyrosine phosphorylated on their ITAM motifs (Cambier, 1995). The generation of a functional BCR requires complex genetic processes initiated during the pro-B cell stage of development and proceeding through the pre-B cell stage (Bireland and Monroe, 1997; DeFranco, 1997). [For excellent recent reviews of early B cell development see Burrows and Cooper (1997) and LeBien (1998).] The pre-BCR contains heavy chains paired with ␭5 surrogate light chain/VpreB and has associated with it Ig움/Ig-웁 heterodimers. Because B cells do not develop normally in preBCR-deﬁcient mice (see DeFranco, 1997) or Ig-움- or Ig-웁-deﬁcient mice (Gong and Nussenzweig, 1996; Torres et al., 1996), it is generally accepted, but by no means proved, that some sort of signal through the pre-BCR is required for developmental progression to the late pre-B cell stage. Consistent with this model, Nagata and co-workers (1997) found that ligating Ig-웁 on pro-B cells from recombination activating gene 2 (RAG 2)deﬁcient mice leads to tyrosine phosphorylation of Ig-움 and other key signaling components such as Syk and Vav. Furthermore, when they treated RAG2⫺/⫺ mice in vivo with anti-Ig-웁 monoclonal antibodies (mAbs), they induced pro-B cells to differentiate to the pre-B cell stage. Similarly, expression of functional IgH chains in vivo correlates with a downregulation of RAG1/2 (Grawunder et al., 1995), and expression of IgH chains in pro-B lines leads to down-regulation of terminal deoxynucleotidyl transferase (TdT) (Wasserman et al., 1997), implying that a functional preBCR may drive pro-B cell maturation in part by turning off RAG1/2 and

90

ANDREW CRAXTON et al.

FIG. 3. BCR-mediated signaling pathways and their regulation by CD22 and the CD19/ CD21 complex. Following BCR ligation by either cell surface antigens or polyvalent T cellindependent antigens, ITAM motifs on Ig-움 and Ig-웁 become tyrosine-phosphorylated, leading to the recruitment of both Lyn and Syk. CD22 and CD19 are also tyrosine-phosphorylated after BCR ligation, probably by Lyn. Activated Syk induces PLC-웂1 activation via a mechanism that may involve either association with the adaptor molecule BLNK (not shown) or, alternatively, via direct binding of the C-terminal SH2 domain of PLC-웂1 to Syk or through the binding of both Syk and PLC-웂1 to tyrosine-phosphorylated CD22. PLC-웂1 activation results in PIP2 hydrolysis with the concomitant formation of DAG, which activates PKC isozymes, and IP3, which mediates the release of intracellular Ca2⫹. Increases in intracellular Ca2⫹ result in the activation of downstream effectors, including JNK/SAPK, calcineurin, and Ca2⫹/calmodulin-dependent protein kinase II. The PTPase SHP-1 binds to ITIM motifs in the cytoplasmic tail of CD22 and inactivates PLC-웂1 by dephosphorylating critical tyrosine residues. BCR-induced CD19 phosphorylation leads to the recruitment of PI3K, which mediates PKB activation and may also play a role in Btk activation via its ability to regulate PtdIns(3,4,5)P3 levels. PtdIns(3,4,5)P3 /Ins(1,3,4,5)P4 5-phosphatase (SHIP) may also play a role in PKB activation. BCR-induced p38 MAPK activation requires either Lyn or Syk, whereas Ras activation is involved in BCR-mediated ERK stimulation.

TdT expression. Furthermore, the pre-BCR may function to maintain allelic exclusion (ten Boekel et al., 1998). Although these results strongly suggest that the pre-BCR plays an essential role in B cell development, it is still not clear what leads to pre-BCR ligation in vivo. An alternative possibility, that Ig-움/Ig-웁 heterodimers are essential because they are needed to interact with and signal through other receptors during B cell development, has not been extensively tested. In addition, the allelic exclu-

B CELL SIGNAL TRANSDUCTION PATHWAYS

91

sion evident in pre-B cells does not require the ␭5 component of the preBCR (ten Boekel et al., 1998), suggesting that some other receptor or mechanism distinct from the pre-BCR is involved. Once both functional H and L chains are produced, the resulting BCR in some way promotes further developmental progression to the immature B cell stage. Using the Cre recombinase conditional gene targeting system (Kuhn et al., 1995), Lam et al. (1997) ablated the expression of transgenic BCRs; mature B cells without BCRs underwent apoptosis in vivo, demonstrating that some kind of signaling through the BCR is essential for mature B cells to survive. B. PROTEIN TYROSINE KINASES Precisely how the BCR may transmit a persistent ‘‘survival’’ signal versus a proliferative or a death signal is not known (Neuberger, 1997). In spite of its apparently simple structure, following cross-linking the BCR can activate a remarkably diverse set of signaling pathways (Law et al., 1996a; DeFranco, 1997; Kurosaki, 1997; Monroe, 1998). The key initial event is the activation of PTKs, which initiate most if not all of the subsequent signaling events (Fig. 3). Because the BCR has no intrinsic PTK activity, it uses three distinct families of nonreceptor SH2 domain-containing cytoplasmic PTKs: (1) the Src family PTKs, especially Lyn but probably also, at discrete stages of differentiation, Blk, Fyn, Lck, and Fgr, (2) Syk of the Syk/ZAP-70 family, and (3) Btk of the Tec family (Kurosaki, 1997; Law et al., 1996a; Reth and Wienands, 1997). Early models proposed that one PTK—depending on the model, either Lyn or Syk, because both associate with the BCR—initiates signaling through the BCR. For example, Syk can be found constitutively associated with the BCR (Law et al., 1994) and after BCR ligation may be activated by homodimerization and transphosphorylation (Law et al., 1996a). Conversely, Lyn is especially effective at phosphorylating ITAMs and may physically associate with and activate Syk (Bolen, 1995; Cambier, 1995; Sidorenko et al., 1995). However, this simple model is contradicted by the fact that the BCR can transmit signals in either Lyn⫺/⫺ or Syk⫺/⫺ B cells (Takata et al., 1994, 1995; Chan et al., 1997, 1998; Wang et al., 1996; Turner et al., 1995, 1997; Muthukkumar et al., 1997; Jiang et al., 1998), but not in Lyn⫺/⫺ Syk⫺/⫺ cells (Takata and Kurosaki, 1996; Jiang et al., 1998). This suggests that the BCR activates both Syk-dependent and Lyndependent signaling pathways and that Syk and Lyn may be activated simultaneously, albeit via distinct mechanisms. The central questions now are ‘‘Which pathways are regulated by which PTKs?’’ and ‘‘Which pathways are required at certain stages of B cell differentiation?’’ Studies with Sykor Lyn-deﬁcient mice can only partially address these questions. Condi-

92

ANDREW CRAXTON et al.

tional gene targeting of PTKs at deﬁned stages of B cell development might provide some answers. C. DOWNSTREAM SIGNAL TRANSDUCTION PATHWAYS A striking feature of BCR signaling is that a variety of distinct signaling pathways are activated upon receptor cross-linking (Fig. 3). Phospholipase C-웂 (PLC-웂) activation leads to phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P2] hydrolysis, resulting in the formation of two second messengers: inositol 1,4,5-trisphosphate [Ins(1,4,5)P3], which triggers the release of intracellular calcium, and diacylglycerol (DAG), which results in the activation of protein kinase C (PKC) isozymes. Stimulation of the phosphatidylinositol 3-kinase (PI3K) pathway results in activation of the serine/threonine kinase PKB/Akt (Craxton et al., 1999). Activation of the Ras pathway mediates ERK stimulation and ultimately links receptor activation to transcriptional activation (Bireland and Monroe, 1997; Hashimoto et al., 1998). In addition, BCR ligation activates at least three mammalian subfamilies of MAP kinases, namely, the extracellular signal-regulated kinases (ERKs), the c-jun amino-terminal kinases ( JNKs) or stress-activated protein kinases (SAPKs), and p38 MAPKs (Figs. 3 and 5). Although it is not yet known exactly how these various ‘‘signaling events are integrated to direct the appropriate biological response of B cells to antigen’’ (DeFranco, 1997), adaptor proteins such as Grb2, BLNK (B cell linker protein), and Grap (Grb2-related adaptor protein) are likely to play critical roles. 1. The PLC-웂 Pathway BCR-dependent PtdIns(4,5)P2 hydrolysis is mediated by PLC-웂1 and PLC-웂2 tyrosine phosphorylation and subsequent activation. Genetic studies using DT40 chicken B cells indicate that both Syk and Btk are required for activation of at least the PLC-웂2 isoform (Takata et al., 1994; Takata and Kurosaki, 1996). Although it is presently unclear why both Syk and Btk are necessary for PLC-웂2 activation, it is possible that the two PTKs may either phosphorylate PLC-웂2 on distinct sites, both of which are required for activation; alternatively, one PTK may be responsible for tyrosine phosphorylation and activation of PLC-웂2 whereas the other PTK may mediate PLC-웂2 translocation to the plasma membrane by an undeﬁned mechanism. The novel adaptor molecule BLNK has been proposed as a putative scaffolding molecule that integrates BCR-induced Syk and PLC-웂 activation, because Syk-mediated tyrosine phosphorylation of BLNK recruits PLC-웂 to BLNK, and thereby facilitates PLC-웂 phosphorylation and activation (Fu et al., 1998). Another potential mechanism for Syk-mediated PLC-웂 translocation to the plasma membrane is through

B CELL SIGNAL TRANSDUCTION PATHWAYS

93

the binding of the C-terminal SH2 domain of PLC-웂 to Syk, which has previously been tyrosine phosphorylated within its linker region, and which subsequently binds to phosphorylated ITAMs (Sillman and Monroe, 1995; Law et al., 1996c). Alternatively, Syk and PLC-웂 may directly associate with tyrosine-phosphorylated CD22 (Law et al., 1996a, b). Thus, as CD22 also associates with the BCR (see below), PLC-웂 may indirectly interact with the BCR complex via CD22. However, this is unlikely to be a major mechanism of PLC-웂 translocation to the plasma membrane because CD22⫺/⫺ mice have enhanced BCR-induced Ca2⫹ responses relative to wild-type mice (O’Keefe et al., 1996; Otipoby et al., 1996; Sato et al., 1996a; Nitschke et al., 1997). A Ca2⫹-independent isoform of PKC, PKC-애, associates with the BCR and coprecipitates with both Syk and PLC-웂1/2 (Sidorenko et al., 1996). PKC-애 was activated on BCR engagement, most likely as a consequence of increased DAG production. Genetic analysis using the DT40 chicken B cell line indicated that Syk, PLC-웂2, and, to some extent, Btk were required for BCR-induced PKC애 activation (Sidorenko et al., 1996). Phosphorylation of Syk by PKC-애 decreases the ability of Syk to phosphorylate PLC-웂1 in vitro, suggesting that PKC-애 may function in a negative feedback loop to regulate BCR-mediated signaling. 2. The PI3K Pathway Another class of second messenger molecules is generated via phosphorylation of phosphoinositides on the D-3 position by PI3K (Toker and Cantley, 1997; Vanhaesebroeck et al., 1997; Duronio et al., 1998). Multiple forms of PI3K exist in mammalian cells and can be divided into three main classes, I, II, and III (Vanhaesebroeck et al., 1997). One PI3K subfamily, PI3K I, phosphorylates PtdIns, PtdIns(4)P, and PtdIns(4,5)P2 to generate PtdIns(3)P, PtdIns(3, 4)P2 and PtdIns(3, 4, 5)P3, respectively (Franke et al., 1997a). Class I PI3Ks can form heterodimeric complexes with adaptor molecules that link them to different upstream signaling pathways and are mediators of signaling by either receptors with intrinsic or associated tyrosine kinase activity or, alternatively, receptors linked to heterotrimeric G proteins (Vanhaesebroeck et al., 1997). BCR engagement activates PI3K (Gold et al., 1992a; Gold and Aebersold, 1994; Kuwahara et al., 1996). The PTKs Fyn and Lyn, and the coreceptor CD19, appear to be involved in PI3K regulation following BCR crosslinking (Yamanashi et al., 1992; Tuveson et al., 1993; Pleiman et al., 1994). BCR engagement enhances the association between PI3K and the Srcfamily PTK Lyn. Binding of the SH3 domain of either Fyn or Lyn to the p85 regulatory subunit of PI3K enhances PI3K activity (Yamanashi et al., 1992; Pleiman et al., 1994). A proportion of total cellular CD19 may be

94

ANDREW CRAXTON et al.

associated with the BCR complex; BCR engagement leads to its tyrosine phosphorylation, which in turn enhances the ability of CD19 to recruit and activate PI3K (Carter et al., 1997). Studies with CD19⫺/⫺ mice and CD19 transgenic mice directly demonstrate that CD19 expression is required for both BCR-mediated PI3K activation and maximal phosphoinositide hydrolysis and calcium mobilization (Buhl et al., 1997). Thus, after BCR ligation, PI3K is activated by simultaneously binding to phosphorylated CD19 via its p85 subunit SH2 domain and also by association with the SH3 domain of Fyn and/or Lyn. This complex may also contain the Cbl adaptor protein (Kim et al., 1995; Beckwith et al., 1996b). In response to other stimuli, PI3K may also be activated by direct interaction of Ras with the p110 catalytic subunit of PI3K (Rodriguez-Viciana et al., 1994, 1996). Whether PI3K is also regulated by its interaction with Ras following BCR cross-linking is unknown. In addition to increasing PtdIns(3, 4, 5)P3 levels transiently, BCR engagement also leads to a prolonged elevation of PtdIns(3, 4)P2 levels (Gold and Aebersold, 1994). This may result from either phosphorylation of PtdIns4P by PI3K or removal of the 5-phosphate group from PtdIns(3, 4, 5)P3 by SHIP, an inositol polyphosphate 5-phosphatase (Scharenberg et al., 1998). Although details of the mechanism by which SHIP may regulate the PI3K pathway remain to be elucidated, a recent study showed that the level of PtdIns(3, 4, 5)P3, which may be counterbalanced by both PI3K and SHIP, plays a central role in regulating Btk localization and calcium inﬂux (Scharenberg et al., 1998; Bolland et al., 1998). The functions of PI3K in BCR signaling are relatively poorly deﬁned. While the PI3K inhibitor wortmannin blocks anti-Ig-induced growth inhibition of the human RL B cell line, suggesting a negative role for the PI3K pathway in BCR-mediated signaling (Beckwith et al., 1996a); wortmanninmediated inhibition of PI3K activity induces apoptosis in normal and neoplastic B lymphocytes that are in cell cycle (Padmore et al., 1996), suggesting a positive role for the PI3K pathway in BCR signaling. Additional studies are clearly required to further deﬁne the role of the PI3K pathway in BCR signaling. Studies with multiple cell types and stimuli have revealed multiple downstream targets of PI3K, including the serine/threonine kinase PKB/Akt (Burgering and Coffer., 1995; Franke et al., 1995; Datta et al., 1996) and possibly some PKC isoforms; these include the novel PKC-␧, PKC-␩ and the atypical PKC-␨ (Nakanishi et al., 1993; Toker et al., 1994) as well as the PKC-related kinase PRK1 (Palmer et al., 1995) and p70S6K (Weng et al., 1995). The regulation of PKB/Akt is particularly interesting because PKB/Akt is activated by a variety of stimuli including PDGF, EGF, insulin, NGF, and stresses and has been implicated in promoting cell survival

B CELL SIGNAL TRANSDUCTION PATHWAYS

95

(Dudek et al., 1997; Kulik et al., 1997; Marte and Downward, 1997; Alessi and Cohen, 1998; Downward, 1998). Recent studies suggest that PKB/ Akt may prevent cell death by mediating serine phosphorylation of the ‘‘proapoptotic’’ Bcl-2 family member Bad. Phosphorylated Bad may subsequently be sequestered in the cytosol by the phosphoserine-binding protein 14-3-3, which thereby facilitates the formation of additional Bcl-2/Bcl-x heterodimers and thereby enhances cell survival (Datta et al., 1997; del Peso et al., 1997). Though such a model is quite appealing, these studies did not address the issue of whether PKB/Akt mediates the phosphorylation of endogenous Bad (Datta et al., 1997; del Peso et al., 1997). Indeed, a recent report shows that some cytokines (e.g., IL-4) that activate PKB/Akt do not appear to phosphorylate Bad. Moreover, other cytokines such as GM-CSF phosphorylate Bad, but apparently do so via a PI3K-independent pathway (Scheid and Duronio, 1998). Hence, the mechanism by which PKB/Akt functions to enhance cell survival requires further investigation. PKB/Akt has also been reported to translocate to the nucleus following stimulation. Thus, it is possible that nuclear targets of PKB/Akt may perhaps contribute to its survival function (Andjelkovic et al., 1997; Meier et al., 1997). Activation of PKB/Akt appears to be regulated by a multistep mechanism that requires phosphorylation on threonine 450, membrane translocation, and binding of 3-phosphorylated phosphoinositides to its pleckstrin homology (PH) domain and subsequent phosphorylation on threonine 308 and serine 473 (Alessi et al., 1996; Stokoe et al., 1997; Bellacosa et al., 1998). The ability of SHIP to regulate PtdIns(3, 4, 5)P3 and PtdIns(3, 4)P2 levels reciprocally suggests that SHIP may play at least some role in BCR-induced PKB/Akt activation, although this remains to be demonstrated (Fig. 3). Although BCR ligation activates PI3K and there is a deﬁned link between PI3K and PKB/Akt, the activation of PKB/Akt in response to BCR ligation has not been demonstrated. We have shown that PKB/Akt is both phosphorylated on serine 473 and activated following BCR ligation in multiple B cell lines (Craxton et al., 1999). Furthermore, genetic studies using the chicken DT40 cell line and a series of PTK-deﬁcient mutants suggest that both Syk and Btk are required for BCR-mediated PKB/Akt phosphorylation and activation. 3. The Ras Pathway BCR engagement activates Ras (Harwood and Cambier, 1993; Lazarus et al., 1993). In B cells, Ras activation may involve both activation of guanine nucleotide exchange factors (GEFs), which promote the transition from an inactive GDP-bound to an active GTP-bound state, and inhibition of Ras GTPase-activating proteins (Ras GAPs), which stimulate GTPase

96

ANDREW CRAXTON et al.

inactivation. The signaling requirements for Ras activation and the role of BCR-mediated Ras activation have been the subject of recent review articles (DeFranco, 1997; Kurosaki, 1997). Thus, this section will be restricted to recent developments related to Ras activation in response to BCR ligation. Previous studies have strongly suggested that the adaptor molecule, Shc, may be required for BCR-mediated Ras activation. BCR cross-linking induces rapid tyrosine and serine phosphorylation of Shc and inducible assembly of a multimolecular complex consisting of Shc, the prototypic adaptor protein Grb2, and mSOS, a GEF for Ras (Saxton et al., 1994; Lankester et al., 1994; Smit et al., 1994). However, Shc does not appear to be required for BCR-induced ERK activation, a downstream target of the Ras pathway, at least in chicken DT40 B cells (Hashimoto et al., 1998). BCR-induced ERK stimulation was, however, markedly reduced in Grb2deﬁcient B cells, demonstrating that Grb2, but not Shc, is an important mediator of the BCR-induced ERK response (Hashimoto et al., 1998). The related Grb2 adaptor molecule, Grap, which shares the ability of Grb2 to bind Sos1, does not, however, appear to be involved in BCR-induced ERK activation and quite possibly Ras activation (Hashimoto et al., 1998). These ﬁndings suggest that Grb2 recruitment to the plasma membrane in response to BCR crosslinking may be mediated via an Shc-independent mechanism. One candidate adaptor molecule that may be involved in regulating both BCR-induced Grb2 membrane translocation and, possibly, Ras activation is BLNK (Fu et al., 1998). BLNK associates with Grb2 and Sos1 to form a BLNK/Grb2/Sos1 complex (Fu et al., 1998). Moreover, BLNK both translocates to the membrane fraction in response to BCR cross-linking and does not appear to associate with Shc, suggesting that BLNK/Grb2 complexes represent a Shc-independent mechanism for the regulation of Grb2/Sos signaling pathways, such as the Ras pathway. However, BLNK overexpression does not appear to augment BCR-induced ERK activation (Fu et al., 1998). Accordingly, there may be at least one additional signaling pathway in B cells that regulates BCR-mediated Ras activation via Grb2/Sos complexes or, alternatively, perhaps either Shc or BLNK is sufﬁcient for BCR-mediated Ras activation. In T cells, overexpression of SLP-76, an adaptor molecule that shares 33% homology with BLNK, potentiates TCR-mediated ERK activation (Musci et al., 1997). Though SLP-76 does not appear to be expressed in B lymphocytes, B cells may express another functional counterpart of SLP-76, which is the primary mediator of BCR-induced Ras and ERK activation. Another T cell linker molecule that may play a role in TCR-induced activation of the Ras pathway, and that may also have a B cell counterpart, is LAT (linker for activation of T cells) (Zhang et al., 1998b). LAT binds to Grb2 following TCR ligation, indicating that LAT may regulate plasma membrane recruit-

B CELL SIGNAL TRANSDUCTION PATHWAYS

97

ment of Grb2/Sos complexes and perhaps the Ras pathway (Zhang et al., 1998b). Preliminary studies, however, suggest that LAT may not play a role in TCR-mediated ERK activation (Zhang et al., 1998b). 4. Other Adaptor Protein-Mediated Pathways a. Grb2-Related Molecules. Along with the prototypic adaptor protein, Grb2, there are many other adaptor proteins that can integrate different cellular responses (Birge et al., 1996; Koretzky, 1997; Pawson and Scott, 1997; Pawson, 1998). Although these molecules do not contain intrinsic enzymatic activity, they generally function to direct intermolecular protein– protein interactions. After lymphocyte activation, Grb2 may bind to a number of different molecules, including the PTPase SHP-2 (Tailor et al., 1996), perhaps the SH2 domain-containing leukocyte protein of 76 kDa, SLP-76 ( Jackman et al., 1995; see below), the 120-kDa proline-rich adaptor, Cb1 (Buday et al., 1996), and the GEF, Vav, either directly or indirectly (Nel et al., 1995; Ye and Baltimore, 1994). Thus, Grb2 is likely to have functions beyond simply facilitating Ras activation. This is not surprising because it may be argued that adaptors such as Grb2 must have evolved so that signals could either diverge to multiple effectors or be integrated via many inputs (Mayer and Gupta, 1998). Earlier studies with Shc in activated B cells found that Shc associates with a 130- to 145-kDa molecule (Saxton et al., 1994; Smit et al., 1994), which was subsequently identiﬁed as the SH2 domain-containing phosphatase SHIP (see Kavanaugh et al., 1996; Tridandapani et al., 1997a,b). Tridandapani et al. (1997a) found that conditions that facilitated contact between the BCR and Fc웂RIIB receptors inhibited activation of the Ras pathway and its downstream effectors, Raf-1 and the ERKs. These authors suggested that the reduction in Ras activation was a consequence of increased association of SHIP with tyrosine-phosphorylated Shc, which reduces Shc/Grb2 interactions and thereby inhibits Shc/Grb2-mediated Ras activation (Tridandapani et al., 1997b). The binding of Shc to Grb2 is greatly reduced in either Lyn⫺/⫺ or Syk⫺/⫺ B cell lines (Nagai et al., 1995). One explanation for these ﬁndings proposes that Lyn is required for tyrosine phosphorylation of ITAM motifs on Ig-움 in the BCR complex to which Shc may bind (D’Ambrosio et al., 1996a), whereas Syk may function to tyrosine phosphorylate Shc (Nagai et al., 1995; Richards et al., 1996). Smit et al. (1996a) found that BCR ligation induced two associations, one between Vav, a GEF for the Rho family of small GTPases (Crespo et al., 1997; Reif and Cantrell, 1998), and the adaptor protein Crk, and the other between C3G, which exerts nucleotide exchange activity on the Rasrelated Rap1 protein (e.g., Gotoh et al., 1995; Boussiotis et al., 1997), and the adaptor Crk-L. It has been suggested that c-Cbl may interact with and

98

ANDREW CRAXTON et al.

regulate these complexes (Smit et al., 1996a,b; Panchamoorthy et al., 1996), but it is not entirely clear how. Because SLP-76 is a likely integrator for T cell signals (Koretzky, 1997), counterparts for SLP-76 in B cells may well serve a similar function. Chan and co-workers (Fu and Chan 1997; Fu et al., 1998) have identiﬁed a potential B cell-associated counterpart of SLP-76, known as BLNK. BLNK interacts with Grb2, Vav, and PLC웂 after BCR ligation (Fu and Chan, 1997; Fu et al., 1998) and has approximately 33% amino acid sequence identity with SLP-76 (Fu et al., 1998). BLNK is tyrosine phosphorylated after BCR ligation, probably by Syk, because Syk, but neither Btk or Lyn, can efﬁciently phosphorylate BLNK in vitro. Moreover, tyrosine phosphorylation of BLNK does not occur in Syk⫺/⫺ cells but can be induced in either Lyn⫺/⫺ or Btk⫺/⫺ cells (Fu et al., 1998). Interestingly, though BLNK can associate with Grb2/Sos complexes, it does not associate with Shc, suggesting that it may be involved in a Shcindependent Ras activation pathway. BLNK can also facilitate the tyrosine phosphorylation of PLC-웂1 and release of [Ca2⫹]i as well as increases in NF-AT transcriptional activity (Fu et al., 1998; Ishiai et al., 1999). Thus, BLNK, like SLP-76 (Koretzky, 1997), may play a key role in the activation of NF-AT (Fig. 4) and ultimately in regulating B cell proliferation (Ranger et al., 1998; Yoshida et al., 1998).

FIG. 4. Grb2 and GrpL adaptor protein-mediated pathways in T and B cells. This model proposes that after ligation of the TCR on T cells or the BCR on B cells the ZAP-70 or Syk tyrosine kinases are activated, which in turn leads to tyrosine phosphorylation of LAT or CD22, respectively. Grb2 binds to LAT in T cells or CD22 in B cells, setting in motion activation of the Ras pathway leading to ERK activation. Grb2 also interacts with BLNK in B cells, which may potentiate the NF-AT pathway by either promoting tyrosine phosphorylation of PLC-웂 and release of intracellular free calcium (Ca2⫹ ) or activation of Vav. GrpL, an adaptor related to Grb2, may also activate Vav through its interaction with SLP-76.

B CELL SIGNAL TRANSDUCTION PATHWAYS

99

Feng et al. (1996) have identiﬁed a molecule, Grap, which has high sequence similarity to Grb2 and, like Grb2, also has an SH3–SH2–SH3 domain structure. However, in contrast to Grb2, which is ubiquitously expressed, Grap has a restricted pattern of expression and is most highly expressed in the thymus and spleen. Grap is also associated with Sos1, primarily through its N-terminal SH3 domain, suggesting that it may function to couple signals from receptor and cytoplasmic tyrosine kinases to the Ras signaling pathway. In Jurkat cells, TCR activation leads to associations between Grap and pp36 (probably LAT) and, to a lesser degree, Shc (Trub et al., 1997). Trub and co-workers also found that Grap is expressed in B cell lines and that the SH2 domain of Grap has binding properties similar to those of Grb2. Thus, Grap is likely to have important signaling functions in B cells. Yet another lymphocyte-speciﬁc 52-kDa adaptor, TSAd, which also contains both SH2 and SH3 domains has been described recently (Spurkland et al., 1998). We have also identiﬁed another novel lymphocyte-associated adaptor, which we term GrpL, for Grb2-related protein of the lymphoid system (Law et al., 1999; Liu et al., 1999). GrpL, like Grb2 and Grap, has an N-terminal SH3 domain followed by a SH2 domain and a C-terminal SH3 domain. In contrast to Grb2 and Grap, GrpL also contains a proline-rich region between the SH2 and C-terminal SH3 domains that could potentially interact with SH3 domain-containing proteins. Human GrpL (hGrpL) and mouse GrpL (mGrpL) proteins have approximately 88% overall sequence identity, suggesting that GrpL is highly conserved during mammalian evolution. Unlike Grb2, which is broadly distributed (Lowenstein et al., 1992), GrpL expression is relatively restricted to hematopoietic tissues and thus is more similar to Grap than to Grb2 (Feng et al., 1996; Trub et al., 1997). GrpL transcripts of 1.5 and 4.0 kb in size are expressed in thymus at high levels and also in bone marrow. Interestingly, GrpL transcripts are expressed in naive B cells but not in activated GC B cells (Solow et al., in preparation). Thus, GrpL may function at deﬁned stages of lymphocyte differentiation. GrpL-speciﬁc antisera immunoprecipitate proteins of approximately 38–40 kDa from both B and T cell lines. In addition, on cell activation, tyrosine-phosphorylated SLP-76 is found associated with GrpL (Law et al., 1999). SLP-76 can also be bound in vitro by a Grb2 SH2 domain fusion protein and may, in addition, bind PLC-웂1 and Vav ( Jackman et al., 1995; Wu et al., 1996). SLP-76 may function to integrate antigen receptor signals required for the activation of NF-AT and cell survival (Wu et al., 1996; see Koretzky, 1997), and GrpL promotes SLP-76-regulated activation of NF-AT (Law et al., 1999; Liu et al., 1999). The tyrosines Y113 and Y128 are essential for the Vav SH2 domain to bind SLP-76 and for augmentation of TCR-induced NF-AT activity (Fang et al., 1996; Raab

100

ANDREW CRAXTON et al.

et al., 1997; Musci et al., 1997). However, whereas GrpL binds to SLP76 in vivo, Grb2 does not; conversely, Grb2 binds to both LAT and Sos1 but GrpL does not (Law et al., 1999). Thus, lymphocytes may utilize at least two adaptor protein pathways (Fig. 4). Because both BLNK (Fu and Chan, 1997) and GrpL (Law et al., 1999) are expressed in B cells, it is quite possible that GrpL–BLNK may play a similar role in regulating signaling. b. The Vav Guanine-Nucleotide Exchange Factor. The Rho family of GTPases plays many important roles in lymphocyte activation. The best characterized Rho family GEF member is Vav, which is expressed speciﬁcally in hematopoietic cells. Vav is a 95-kDa protein whose structure includes both SH2 and SH3 domains and a Dbl homology region, which is characteristic of Rho family GTPases (Reif and Cantrell, 1998). Vav is tyrosine-phosphorylated in response to either TCR or BCR ligation and can associate with either ZAP-70 and SLP-76 in T cells (Wu et al., 1997b) or Syk in B cells (Deckert et al., 1996). The critical role of Vav is underscored by the ﬁnding that antigen receptor-mediated proliferative responses in vitro are severely reduced in Vav-deﬁcient B or T cells (Tarakhovsky et al., 1995; Fischer et al., 1995). Moreover, overexpression of Vav or SLP-76 potentiates TCR-induced NF-AT activation (Wange and Samelson, 1996; Wu et al., 1996; Raab et al., 1997), suggesting that Vav may play a key role in promoting lymphocyte proliferation. Vav may serve to function very proximal to antigen receptor signaling. After TCR ligation of Vav⫺/⫺ T cells, mobilization of intracellular calcium, phosphorylation of SLP-76, recruitment of the actin cytoskeleton to the CD3 ␨ chain and cap formation are impaired (Fischer et al., 1998; Holsinger et al., 1998). Moreover, nuclear translocation of NF-AT in both Vav⫺/⫺ T and B cells does not occur in response to antigen receptor stimulation, whereas antigen receptor-induced activation of either ERK or JNK is normal in Vav⫺/⫺ B and T cells (Fischer et al., 1998; Holsinger et al., 1998). Because Vav⫺/⫺ lymphocytes are dramatically impaired in actin polymerization and proliferation, Vav may regulate an actin-dependent pathway that is required for lymphocyte growth. c. A Few Words about Pleckstrin Homology Domains. In proteins that recognize speciﬁc sequences on their target proteins, a number of modular domains, such as SH2 and SH3, have been identiﬁed and described previously, as well as PTB, PDZ, WW, and 14-3-3 domains (Pawson and Scott, 1997; Pawson, 1998). PH domains, by contrast, are small protein modules of approximately 120 residues that bind both to proteins and to charged head groups of speciﬁc polyphosphoinositides; thus they may be able to regulate targeting of signaling molecules to speciﬁc regions of the plasma

B CELL SIGNAL TRANSDUCTION PATHWAYS

101

membrane. Critical signaling molecules with PH domains, including PLC웂1, dynamin, and Sos, have been shown to bind directly to PtdIns(4,5)P2 or Ins(1,4,5)P3 (Toker and Cantley, 1997), whereas the PH domain of Akt/ PKB binds most strongly to PtdIns(3,4)P2 (Franke et al., 1997a,b) and the PH domain of Btk binds to both PtdIns(3,4, 5)P3 and Ins(1,3,4,5)P4 (Salim et al., 1996; Rameh et al., 1997). The PKC-애 isoform, which interacts with Syk and PLC-웂1, also contains a PH domain (Sidorenko et al., 1996). The binding afﬁnity of PH domains overall is relatively poor. Thus, one attractive model is that other adjacent domains, such as SH2 or PTB domains, may bind to their motifs and facilitate PH domain binding (Pawson and Scott, 1997). PH domains are essential for the normal function of certain lymphocytic kinases such as Akt/PKB (Franke et al., 1997b) and Btk (Li et al., 1995; Khan et al., 1995; Bolland et al., 1998; Lemmon and Ferguson, 1998). Interestingly, the PH domain of Btk is also required for it to interact with a novel B cell-associated protein, BAP-135 (Yang and Desiderio, 1997). We have identiﬁed a 32-kDa adaptor protein that is relatively restricted to B lymphocytes and dendritic cells; we call this protein BAM32 for B cell adaptor molecule, 32 kDa in size (Marshall et al., 1999). BAM-32 contains a single SH2 domain and one PH domain. It is likely to be yet another example of the growing number of adaptor molecules that have specialized functions restricted to certain cell lineages. 5. MAP Family Kinase Pathways A common feature of BCR-activated second messenger pathways is that each involves initial activation of a PTK followed by activation of effector molecules such as phospholipases, lipid kinases, or serine/threonine kinases. The serine/threonine kinases provide one link to transcriptional events, because the activity of many transcription factors is regulated by serine/threonine phosphorylation and dephosphorylation. One of these linkers is the family of MAP kinases. Three structurally related MAPK subfamilies have been identiﬁed in mammalian cells: the p42 and p44 ERKs, the c-jun N-terminal kinases ( JNKs or SAPKs), and the p38 MAPKs. Because each of these three types of MAP kinases phosphorylates distinct but also overlapping sets of transcription factors (Su and Karin, 1996; Treisman, 1996), they are good candidates for transducing signals from different receptors, including BCR ligation, into distinct cellular responses in B cells. Hence, the transcription factors Elk, Ets-1/2, and c-fos are substrates for ERKs; Elk-1, c-jun, and ATF2 are substrates for JNK/SAPKs; and ATF2, c-jun, Elk, CHOP, MEF2C, and Max are substrates for p38 MAPK. MAP kinases have been implicated in both mitogenic and apoptotic responses. In PC-12 cells, activation of JNK and p38 MAPK correlates with apoptosis whereas ERK activation correlates with cell growth (Xia et

102

ANDREW CRAXTON et al.

al., 1995). JNK activation has been shown to correlate with apoptosisinducing signals in several systems (Chen et al., 1996; Verheij et al., 1996). Healy et al. (1997) found that whereas both ERK and JNK1 were activated during positive signaling in naive B cells, only ERK, but not JNK1, was activated in tolerant B cells on BCR stimulation by the same ligand, suggesting that the activation of distinct patterns of MAP kinases may mediate positive versus negative signaling in B lymphocytes. In the human B cell line B104, ERK2 is activated rapidly and transiently on mIgM ligation whereas activation of JNK and p38 MAPK is both delayed and sustained and requires new gene transcription (Graves et al., 1996, 1998a). In these B cells, the activation of JNK and p38 MAPK, but not ERK, correlated with the induction of apoptosis by sIgM stimulation (Graves et al., 1996). Furthermore, blockade of p38 MAPK prevented BCR-induced cell death (Graves et al., 1998a). Nevertheless, the role of speciﬁc MAPKs in BCR-induced apoptosis may vary among different cell lines. Thus, in WEHI-231 B cells, BCR-induced apoptosis did not appear to require p38 MAPK (Salmon et al., 1997), but rather correlated with ERK2 activation (Lee and Koretzky, 1998). The mechanisms by which different stimuli achieve differential activation of MAP kinases and subsequently different biological responses are only just beginning to become apparent. One study has revealed that different classes of MAPKs—JNKs, ERKs, and p38 MAPKs—have differential requirements for targeting to Elk-1, a common substrate for all three MAP kinase families (Yang et al., 1998). The speciﬁcity of these individual MAP kinases may be achieved by formation of different scaffolding complexes: a scaffolding molecule, MP1 (MEK partner 1), can speciﬁcally bind MEK1 and ERK1 and facilitate their activation, whereas another scaffolding molecule, JIP1 ( JNK-interacting protein 1), selectively binds MLK/MKK7/JNK to form a signaling module and facilitate signaling (Schaeffer et al., 1998; Whitmarsh et al., 1998). Using the DT40 chicken B cell line and a variety of mutant lines deﬁcient in upstream effectors such as Syk, Lyn, Btk, and PLC-웂2, we, as well as Kurosaki and colleagues, have begun to deﬁne the role of distinct upstream effectors in the activation of different MAPKs during BCR signaling ( Jiang et al., 1998; Hashimoto et al., 1998). BCR-mediated ERK2 activation was partially inhibited in PLC-웂2-deﬁcient cells or in DT40 cells transfected with a dominant negative form of Ras (Ras N17). Interestingly, a complete block of ERK2 activity was seen only in PLC-웂2-deﬁcient cells transfected with Ras N17 (Hashimoto et al., 1998), suggesting that ERK2 may be activated via two distinct pathways: a Ras-dependent pathway and a PLC웂2-mediated, PKC-dependent pathway (Fig. 5). BCR-mediated activation

B CELL SIGNAL TRANSDUCTION PATHWAYS

103

FIG. 5. Activation of MAP family kinases in response to BCR ligation. BCR cross-linking leads to activation of the protein tyrosine kinases Lyn, Syk, and Btk. Loss of either Syk or Btk leads to abrogation of calcium mobilization, although only Syk is essential for PLC-웂2 phosphorylation. PLC-웂2 phosphorylation leads to PLC-웂2 activation and subsequently to calcium mobilization and PKC activation. Syk but not Lyn is essential for BCR-mediated ERK2 activation and Btk may function as a regulator of the ERK2 pathway. Activation of PKCs rather than a calcium signal is necessary for ERK2 activation. A second pathway for ERK activation requires Ras, and both Ras and PKCs are required for maximal ERK activation. Though Shc is not required for BCR-mediated ERK2 activation, Grb2 is necessary, suggesting another adaptor protein is required for ERK activation. In contrast, Rac1 but not Ras is involved in BCR-mediated JNK/p38 MAPK activation. BCR-mediated JNK1 activation depends on both Syk and Btk but not Lyn, via a calcium- and PKC-dependent pathway. In contrast, either Syk or Lyn is sufﬁcient for BCR-mediated p38 MAPK activation.

of Ras and PKCs may converge at the level of Raf-1 to mediate ERK activation (Gold et al., 1992b; Tordai et al., 1994; Kawauchi et al., 1996). In contrast, BCR-mediated JNK1 activation was completely abolished in Syk-, Btk-, or PLC-웂2-deﬁcient DT40 cells but maintained in Lyndeﬁcient cells ( Jiang et al., 1998). Consistent with these observations, BCR-mediated JNK1 activation required intracellular calcium and PMAsensitive PKCs. Interestingly, dominant negative Rac1 (Rac1 N17) also blocked JNK1 activation, suggesting a model in which the cooperation of PKCs, intracellular calcium, and Rac1 facilitates JNK1 activation (Fig. 5). Surprisingly, BCR-mediated p38 MAPK activation was maintained in each PTK-deﬁcient cell line, suggesting that no single PTK is essential for p38 MAPK activation. Consistent with these observations, BCR-mediated p38 MAPK activation was abolished in Lyn/Syk double-deﬁcient cells, demonstrating that either Lyn or Syk was sufﬁcient to activate p38 MAPK ( Jiang et al., 1998). Thus, distinct MAP kinase family members are differentially regulated during BCR-mediated signaling. DT40 cells undergo apoptosis after BCR engagement, and the activation of one MAPK, JNK1, strongly correlated with BCR-mediated apoptosis. BCR-mediated apoptosis was

104

ANDREW CRAXTON et al.

blocked in Syk-, Btk-, or PLC-웂2-deﬁcient DT40 cells in which JNK1 activation was also abolished. Further experiments with either MAPKdeﬁcient cells or knockout mice would increase our understanding of the molecular mechanisms by which the activation of MAPKs inﬂuences B cell development, B lymphocyte activation, and cell death. D. BCR PATHWAYS LEADING TO CELL DEATH 1. Introduction and the WEHI-231 Model Many of the initial studies deﬁning events associated with BCR-induced death have utilized the mouse B cell line, WEHI-231 (Sonenshein, 1997; Scott et al., 1996). Although this line clearly may not reﬂect all modes of BCR-induced or activation-induced cell death, it has been a useful model. For instance, although some studies have suggested that overexpression of the c-myc oncogene can promote apoptosis in certain cell types, in WEHI-231 cells, BCR-mediated death is associated with a decline in cmyc mRNA and protein expression (Levine et al., 1986; Sonenshein, 1997). Stabilization of c-myc protein prevents BCR-induced death in WEHI-231 (Fischer et al., 1994), and other signals that promote the death of B cells or B cell lines, such as transforming growth factor 웁1, induce a decline in c-myc expression prior to the onset of apoptosis (Fischer et al., 1994; Arsura et al., 1996). The decline in c-myc levels is related to decreases in NF-␬B/Rel binding activity required for maintaining c-myc levels (Sonenshein, 1997). Ligating mIgM on primary mature splenic B cells in vitro can induce cell death (Illera et al., 1993; Parry et al., 1994), but this requires extensive cross-linking. Norvell et al. (1995) compared the ability of anti-IgM to induce apoptosis in immature mIgM⫹D⫺ versus mature mIgM⫹D⫺ B cells and found that immature B cells were more sensitive to BCR-induced death. The increased sensitivity of immature B cells may be related to the fact that these B cells, unlike mature B cells, express neither cyclin E nor Cdk2 after BCR ligation and enter the late G1 phase of the cell cycle (Carman et al., 1996; Monroe, 1998). The relative role of signals inducing cell cycle arrest versus caspase death pathways in promoting B cell death remains to be determined. Factors such as IL-4 (Illera et al., 1993; Parry et al., 1994; Carman et al., 1996) or CpG oligodeoxyribonucleotides (Yi et al., 1998), which can promote entry into the cell cycle, also protect B cells from apoptosis, but it is not clear how. Scheuermann and co-workers have proposed that different elements in the BCR signaling pathway may be responsible for cell cycle arrest versus cell death: the PTK Lyn may be essential for cell cycle arrest, whereas Syk and release of [Ca2⫹]i may be essential for apoptosis (Scheuermann et al., 1994; Scheuermann and Uhr,

B CELL SIGNAL TRANSDUCTION PATHWAYS

105

1995). Yao et al. (1995) have even proposed that the growth arrest and apoptotic pathways induced via the BCR may be mediated by different components within the BCR complex. Grumont et al. (1998) compared the level of spontaneous apoptosis and BCR-induced cell cycle progression in B cells taken from mice deﬁcient in either NF-␬B1 or c-Rel. Spontaneous B cell death was higher in NF-␬B1-deﬁcient B cells compared to c-Reldeﬁcient or wild-type B cells, whereas c-Rel, but not NF-␬B1, was essential for mIgM-induced B cell proliferation. After BCR ligation, c-Rel⫺/⫺ B cells arrested in the late G1 stage of the cell cycle and underwent apoptosis, demonstrating that c-Rel is required for cell cycle entry and the prevention of BCR-induced death. There was no clear dissociation between BCRinduced cell cycle arrest and apoptosis in these deﬁcient mice. Unlike TCR-induced death in T cells (Dhein et al., 1995), BCR-induced death is direct and does not involve secondary interactions between FasL and CD95/Fas (Lens et al., 1996; Racila et al., 1996; Scott et al., 1996). However, the CD95 and BCR death pathways have a number of similarities. Both activate the cleavage of caspases such as caspase-3 (Lens et al., 1998; Graves et al., 1998a; Andjelic and Liou, 1998) and its substrates poly(ADPribose) polymerase (Lens et al., 1998) and the STE 20 homologue Mst-1 (Graves et al., 1998b, and unpublished data), and both are blocked by caspase inhibitors such as z-VAD-fmk. However, BCR-induced death, unlike CD95-induced death, is not blocked by a dominant negative form of FADD (Lens et al., 1998) but is blocked by actinomycin D (Graves et al., 1998a). Similarly, BCR-induced death of immature mouse B cells is blocked by cycloheximide (Norvell et al., 1995). Thus, new gene transcription and protein synthesis may be required, in some instances, in order for the BCR to induce death. In this regard, it is interesting that some key death caspases have restricted tissue distribution in B cells: the effector caspase-3 is not expressed in long-lived mantle zone B cells but is expressed in short-lived germinal center B cells (Krajewska et al., 1997; Krajewski et al., 1997). Which caspases are required and when they act to mediate BCR-induced death are presently unclear. 2. Signaling via mIgM versus mIgD Why some B cells express both mIgM and mIgD BCR complexes remains an enigma, because the function of mIgD is not clear. It is known that mIgD is not essential for B cell development, because IgD null mice produce B cells (Roes and Rajewsky, 1993; Nitschke et al., 1993). One hypothesis proposes that cross-linking of either mIgD or mIgM leads to different B cell responses. Distinct proteins associate with mIgM (Terashima et al., 1994) or mIgD (Kim et al., 1994), which presumably could mediate signaling via different pathways. In addition, whereas mIgM and

106

ANDREW CRAXTON et al.

mIgD both associate with Ig-움/Ig-웁 heterodimers (Venkitaraman et al., 1991) critical for signaling (Cambier, 1995), the Ig-움 associated with mIgD is slightly larger and more acidic than that associated with mIgM (Venkitaraman et al., 1991; Campbell et al., 1991). In addition, the CH3 domain of membrane IgD has a conserved N-linked glycosylation site not found in mIgM (Lin and Putnam, 1981), which may inﬂuence mIgD interactions with other B cell-associated molecules (Campbell et al., 1991). Whether the ligation of mIgM and mIgD transmits distinct signals is somewhat controversial. Some studies found that cross-linking either mIgM or mIgD induces similar B cell responses, e.g., activation of Src family PTKs, release of [Ca2⫹]i, increased surface expression of MHC class II molecules, or cell proliferation (Vitetta et al., 1980; Mond et al., 1981; Cambier et al., 1994; Pezzutto et al., 1988; Harnett et al., 1989; Gold et al., 1991; Brink et al., 1992). In contrast, other studies reported clear differences in mIgM- and mIgD-mediated signaling pathways (Vitetta et al., 1980; Tisch et al., 1988; Ales-Martinez et al., 1988; Mongini et al., 1989; Ishigami et al., 1992; Kim et al., 1992; Haggerty et al., 1993; Graves et al., 1996). For example, cross-linking mIgD on several B cell lines, unlike anti-IgM, did not induce growth inhibition or apoptosis (Tisch et al., 1988; Ales-Martinez et al., 1988; Ishigami et al., 1992; Kim et al., 1992; Haggerty et al., 1993; Graves et al., 1996), even though signaling mediated via either receptor induced new protein tyrosine phosphorylation (Haggerty et al., 1993) or calcium mobilization and ERK2 activation (Graves et al., 1996). In addition, in the human B104 line, mIgM ligation activated p38 MAPK and JNK whereas ligating mIgD did not (Graves et al., 1996). Though Monroe and colleagues observed that mature IgM⫹D⫹ B cells are less susceptible to mIgM-induced death, they showed that immature B cells expressing both mIgM and mIgD did not differ in their apoptotic responses to mIgM versus mIgD ligation (Norvell and Monroe, 1996). Their results do not support the model that mIgD ligation can protect B cells from mIgM-induced death (Carsetti et al., 1995). Similarly, crosslinking mIgD on the human B104 line did not protect these cells from mIgM-induced death (E. A. Clark., unpublished data). However, these results do not exclude the possibility that at some stages of B cell differentiation, ligating mIgD has a different effect on the B cell compared to ligating mIgM. III. Coreceptor Regulation of BCR Signaling

Engagement of the BCR by Ag leads to many possible fates, including proliferation, differentiation, and death or anergy of the B cell (Goodnow, 1996). These fates are regulated to a large extent by the interaction of Ag

B CELL SIGNAL TRANSDUCTION PATHWAYS

107

with the BCR. Within the past few years, the importance of other B cell-associated surface molecules in inﬂuencing BCR-induced signals has become apparent. The involvement—or lack of involvement—of so-called coreceptors such as CD22, CD19, CD21, and Fc웂RIIB helps to determine the fate of B cells by qualitatively and quantitatively altering signals initiated by the BCR. The cytoplasmic tails of these coreceptors can recruit several different types of signaling molecules to the membrane, including PTKs, PTPases, lipid kinases, lipid phosphatases, and phospholipases (O’Rourke et al., 1997). Engagement of these coreceptors and the recruitment of additional molecules involved in signal transduction may alter BCRgenerated signals, and thereby allow the B cell to respond in a manner appropriate to its extracellular milieu (Cyster and Goodnow, 1997). Exactly how coreceptors regulate these responses is now under intense investigation. The generation and characterization of mice and B cell lines with targeted gene disruptions has led to a better understanding of these processes. The possible role of CD19 (Tedder et al., 1997a) and Fc웂RIIB (Coggeshall, 1998) in modulating BCR-mediated pathways has been examined, thus we focus mainly on CD22 here. A. CD22 CD22 is a B lymphocyte-restricted surface glycoprotein that functions as both an adhesion and signal-transducing molecule (Law et al., 1994; Tedder et al., 1997b). Its surface expression remains low during early B cell development and increases as B cells mature in the periphery (Dorken et al., 1986; Andersson et al., 1996; Erickson et al., 1996; Nitschke et al., 1997). CD22 physically associates with the antigen receptor, albeit with low stoichiometry, and becomes rapidly phosphorylated on tyrosine residues following BCR cross-linking (Schulte et al., 1992; Leprince et al., 1993; Peaker and Neuberger, 1993). Several molecules have been shown to associate with CD22 following BCR engagement, including Lyn, Syk, the PTPase SHP-1, and PLC-웂 (Campbell and Klinman, 1995; Doody et al., 1995; Law et al., 1996b; Tuscano et al., 1996a). The cytoplasmic tail of CD22 contains six conserved tyrosine residues that serve as possible substrates for tyrosine kinases (Stamenkovic and Seed, 1990; Wilson et al., 1991; Torres et al., 1992). Three of these tyrosines are within potential immunoreceptor tyrosine-based inhibitory motifs (ITIMs) with the amino acid sequence V/ IxYxxL/V (in single-letter amino acid code, x is any amino acid) (Doody et al., 1995; Thomas, 1995; Unkeless and Jin, 1997). Tyrosyl-phosphorylated ITIMs probably serve as potential binding sites for SH2 domain-containing phosphatases, including SHP-1 (Thomas, 1995; Unkeless and Jin, 1997). Unphosphorylated and tyrosyl-phosphopeptides corresponding to the amino acid sequences surrounding the six tyrosines within the cytoplasmic

108

ANDREW CRAXTON et al.

tail of murine CD22 have been used for competition studies (Doody et al., 1995). At high concentrations, three of the six tyrosyl-phosphopetides were able to compete for SHP-1 binding to CD22, whereas the analogous nonphosphopeptide sequences did not (Doody et al., 1995). These data suggest that after BCR-induced tyrosine phosphorylation of CD22, SHP1 recognizes and binds these sites within CD22. Furthermore, these same three tyrosyl-phosphopeptides, but not the analogous nonphosphopeptides, increased the phosphatase activity of puriﬁed SHP-1 fusion proteins in vitro (Doody et al., 1995). Interestingly, these phosphopeptides—containing pY783, pY843, and pY863 in murine CD22—correspond to the three ITIM sequences (Doody et al., 1995; Thomas, 1995). Finally, Law et al. have shown that SHP-1 SH2 domain fusion proteins can bind directly to phosphorylated CD22 (Law et al., 1996b). We have mapped the binding regions on CD22 to which SHP-1 binds in vivo: mCD22 with single Y to F mutations at either Y843 or Y863 do not bind SHP-1 after BCR ligation, whereas mutants at Y783 or Y828 associate with SHP-1 (Otipoby et al., 1999). Collectively, these data suggest that SHP-1 binds directly to the phosphorylated ITIM sequences within CD22 and that this binding increases SHP-1 phosphatase activity. As indicated previously, several other signaling molecules associate with CD22, although the molecular details of these interactions are not well deﬁned. 1. CD22-Deﬁcient Mice To better understand the in vivo functions of CD22, we and three other groups created CD22-deﬁcient mice (O’Keefe et al., 1996; Otipoby et al., 1996; Sato et al., 1996a; Nitschke et al., 1997). Cd22⫺/⫺ mice are born healthy and have no apparent decrease in survival as they age (O’Keefe et al., 1996; Otipoby et al., 1996; Sato et al., 1996a; Nitschke et al., 1997). Early B cell development in the bone marrow appears normal in cd22⫺/⫺ mice, but the number of mature, recirculating B cells in the bone marrow is dramatically reduced (Otipoby et al., 1996; Sato et al., 1996a; Nitschke et al., 1997). Peripheral B cells are present in normal numbers but are phenotypically altered in that they express decreased levels of mIgM and increased levels of MHC class II, indicative of preactivated B cells (O’Keefe et al., 1996; Otipoby et al., 1996; Sato et al., 1996a; Nitschke et al., 1997). Furthermore, splenic B cells from CD22-deﬁcient mice are 10fold more sensitive to BCR-induced calcium mobilization, despite reduced mIgM expression (O’Keefe et al., 1996; Otipoby et al., 1996; Sato et al., 1996a; Nitschke et al., 1997). Also, sera from unchallenged cd22⫺/⫺ mice contain approximately 2-fold higher levels of IgM but normal levels of the IgG subtypes relative to unchallenged cd22⫹/⫹ mice (O’Keefe et al., 1996;

B CELL SIGNAL TRANSDUCTION PATHWAYS

109

Sato et al., 1996a). These data suggest that CD22 may function as a negative regulator of BCR signaling, presumably through the recruitment of SHP-1. Interestingly, antibody responses to thymus-independent type 2 (TI-2) antigens in cd22⫺/⫺ mice are decreased, whereas antibody responses to thymus-dependent (TD) antigens are normal (Otipoby et al., 1996; Nitschke et al., 1997). Consistent with the in vivo antibody responses, B cells from CD22-deﬁcient mice proliferate at reduced levels in response to in vitro treatment with anti-IgM but normally in response to anti-CD40 mAb (Otipoby et al., 1996; Sato et al., 1996a). These data demonstrate that CD22 may also positively regulate signals from the antigen receptor. This possible BCR signal-enhancing function of CD22 has not been fully addressed. The phenotype of CD22-deﬁcient mice is strikingly similar in many aspects to mice deﬁcient in Lyn expression (Lyn⫺/⫺) (Nishizumi et al., 1995; Hibbs et al., 1995; Chan et al., 1997) and mice reduced in SHP1 phosphatase activity (SHP-1 mev/mev) (Cyster and Goodnow, 1995). Therefore, a summary of the phenotypes of these two mutant lines of mice will be discussed with speciﬁc attention given to B cell defects. 2. Lyn-Deﬁcient Mice Lyn, an abundant Src family kinase of B cells, has been shown to play an important role in BCR signaling (Yamanashi et al., 1991; Takata et al., 1994; Kurosaki, 1997) (see Section II,B). To better understand the function of Lyn, several groups have developed Lyn-deﬁcient mice, which have no apparent defects during early B cell development in the bone marrow, but have a dramatic reduction in the number of conventional (IgD⫹, CD5⫺, Mac1⫺) B cells in the periphery (Nishizumi et al., 1995; Hibbs et al., 1995; Chan et al., 1997). This defect becomes more obvious as the mice age (Wang et al., 1996; Chan et al., 1997; Nishizumi et al., 1998). In wild-type mice, B cells emigrating from the bone marrow express high levels of CD24 and low levels of CD45R (Allman et al., 1993). These B cells presumably receive signals in the periphery that allow them to mature into CD24lo, CD45Rhi, recirculating B cells (Allman et al., 1993). Analyses of peripheral B cell subsets in Lyn-deﬁcient mice have revealed that this mature, CD24lo, CD45Rhi population is dramatically reduced, suggesting a partial block in B cell maturation or survival at this stage (Nishizumi et al., 1995; Hibbs et al., 1995; Chan et al., 1997). Consistent with the decrease in conventional peripheral B cell numbers, the turnover rate of Lyn⫺/⫺ splenic B cells is higher than normal (Chan et al., 1997). Despite this reduction in conventional B cells, aged Lyn⫺/⫺ mice develop splenomegaly and produce autoantibodies, a development that eventually leads to autoimmune disorders and high mortality (Nishizumi et al., 1995; Hibbs et al., 1995; Chan et al., 1997). The expanded population in aged Lyn-deﬁcient

110

ANDREW CRAXTON et al.

mice contains B cells expressing CD45R, IgM, Mac1, and CD5, but not IgD, characteristic of the so-called B1 population, normally found at high levels only in the peritoneum of wild-type mice (Nishizumi et al., 1995; Chan et al., 1997; Kantor and Herzenberg, 1993). Lyn-deﬁcient mice are born with relatively normal levels of serum IgM and IgA, but these levels become greatly elevated beginning from 2 weeks of age (Nishizumi et al., 1998). Surprisingly, on challenge with either TD or TI antigens, the responses of Lyn⫺/⫺ mice are similar to those of wild-type mice (Nishizumi et al., 1995; Hibbs et al., 1995). Two studies have shown augmented proliferation of Lyn-deﬁcient B cells in response to F(ab⬘)2 anti-IgM using B cells puriﬁed from either 4to 8-week-old or 7- to 11-week-old mice, an age prior to the onset of splenomegaly (Wang et al., 1996; Chan et al., 1997). In addition to a hyperproliferative response to F(ab⬘)2 anti-IgM, B cells from Lyn-deﬁcient mice displayed an elevated calcium response and increased activation of JNK and ERK following BCR stimulation (Chan et al., 1997; Cornall et al., 1998; Chan et al., 1998; Nishizumi et al., 1998). Collectively, these data suggest that Lyn is required for negative feedback regulation of signals generated via the BCR. This suggestion is supported by the ﬁnding that Lyn⫺/⫺ B cells expressing an antigen receptor speciﬁc for hen egg lysozyme (HEL) are more sensitive to induction of anergy than are wild-type B cells (Cornall et al., 1998). 3. SHP-1-Deﬁcient mev/mev Mice Motheaten (me) and motheaten-viable (mev) mice are two mutants that arose from spontaneous allelic mutation at the loci encoding the SHP-1 PTPase (see Shultz et al., 1993; Tsui et al., 1993; Bignon and Siminovitch, 1994). The SHP-1 me/me mutation leads to complete loss of SHP-1 protein expression and phosphatase activity, whereas mev/mev homozygotes have greatly reduced phosphatase activity (10–20% of wild-type) due to aberrant splicing of the mRNA encoding SHP-1 (Kozlowski et al., 1993; Shultz et al., 1993; Tsui et al., 1993). Not surprisingly, the phenotype of mev/mev homozygotes is similar to but less severe than the phenotype of me/me homozygotes (Shultz, 1988). These mice have high mortality rates, with mean animal life spans of 3 weeks and 9 weeks for SHP-1 me/me and SHP-1 mev/mev mice, respectively (Shultz et al., 1984). Both strains of mice develop autoimmune disorders and contain dramatically elevated levels of serum IgM and autoantibodies (Shultz and Green, 1976; Sidman et al., 1986). This correlates with an expansion of the CD5⫹ population of B cells in the periphery (Sidman et al., 1986). Conventional B cells are essentially absent in the periphery of SHP-1 me/me and SHP-1 mev/mev mice, whereas the CD5⫹ population is greatly expanded, leading to only

B CELL SIGNAL TRANSDUCTION PATHWAYS

111

marginally reduced numbers of IgM⫹ cells in the spleen of SHP-1 mev/ mev mice (Sidman et al., 1978a; Davidson et al., 1979; Cyster and Goodnow, 1995). Interestingly, mice with either SHP-1 mutation are also severely immunodeﬁcient, with impaired responses to TD and TI antigens (Shultz and Green, 1976; Sidman et al., 1978b; Davidson et al., 1979; Shultz, 1988). Early studies showed that splenocytes from SHP-1 me/me mice, compared with those from control mice, are completely unresponsive to LPS-induced proliferation (Sidman et al., 1978b). Furthermore, SHP-1 me/me splenocytes inhibited the proliferation of control splenocytes when cultured together and stimulated with LPS (Sidman et al., 1978b). This demonstrates a trans effect of the SHP-1 me/me defect on normal B cells. Additional studies using puriﬁed splenic B cells instead of total splenocytes showed that B cells from SHP-1 me/me and SHP-1 mev/mev mice respond normally to LPS (Pani et al., 1995). Interestingly, puriﬁed B cells from SHP-1 me/ me and SHP-1 mev/mev mice relative to control mice are actually hyper responsive to BCR stimulation, proliferating to wild-type levels with approximately one-tenth the amount of anti-IgM (Pani et al., 1995). SHP-1 is normally expressed in multiple hematopoietic cell types (Shen et al., 1991; Matthews et al., 1992; Plutzky et al., 1992; Yi et al., 1992). Analysis of defects intrinsic to SHP-1 me/me and SHP-1 mev/mev B cells is greatly complicated by trans effects mediated by these other cell types. To address this problem, Sidman et al. (1989) reconstituted lethally irradiated wild-type mice with a mixture of 50% wild type, 50% SHP-1 mev/ mev bone marrow in order to ‘dilute’ the trans effects on the lymphoid compartment by other cell types. This study showed that the SHP-1 mev/ mev-derived B cells, but not the wild-type-derived B cells in the same environment, secrete elevated levels of serum IgM, suggesting a B cellautonomous defect in SHP-1 mev/mev mice that leads to elevated secretion of IgM. Cyster and Goodnow (1995) extended these ﬁndings by crossing SHP1 mev/mev mice with transgenic mice that express a BCR speciﬁc for HEL and making similar mixed bone marrow chimeras. They found that irradiated wild-type mice reconstituted with 50% wild type, 50% SHP-1 mev/mev bone marrow still showed some signs of dominant trans effects on bone marrow and splenic B cell subpopulations and also the splenic architecture. In contrast, when irradiated wild-type mice were reconstituted with 80% wild type, 20% SHP-1 mev/mev bone marrow mixtures, these trans affects were not observed. Cyster and Goodnow (1995) also showed elegantly that in the absence of the autoantigen HEL, immature SHP-1 mev/mev B cells in the bone marrow are generated at levels similar to those of wild-type B cells, whereas splenic SHP-1 mev/mev B cell numbers

112

ANDREW CRAXTON et al.

are decreased 10-fold relative to those of wild-type B cells. Furthermore, the SHP-1 mev/mev B cells present in the spleen displayed decreased levels of mIgM, normal levels of mIgD, and increased levels of MHC class II. On stimulation through the BCR, SHP-1 mev/mev B cells responded with an exaggerated elevation of [Ca2⫹]i. These B cell-autonomous defects in SHP-1 mev/mev B cells result in an increased sensitivity to deletion and anergy induction by speciﬁc antigen (Cyster and Goodnow, 1995). These data clearly demonstrate that one of the cis functions of SHP-1 in B cells is to regulate signals from the antigen receptor and to help establish the threshold for deletion and anergy induction. A ﬁnal point from this study is that the absence of normal levels of SHP-1 enables the B1 population to develop in a cell-autonomous fashion. Otherwise, normal anti-HEL transgenic mice contain no B1 cells in the peritoneum; somehow, the transgenic Ig receptor prevents development of this lineage. In contrast, mice reconstituted with mixed transgenic wildtype and transgenic SHP-1 mev/mev bone marrow contain CD5⫹ B cells in their peritoneum. These B1 B cells are derived entirely from SHP-1 mev/mev bone marrow (Cyster and Goodnow, 1995). 4. A Common Biochemical Pathway Involving Lyn, CD22, and SHP-1 The similarity of the defects seen in Lyn⫺/⫺, cd22⫺/⫺, and SHP-1 mev/ mev mice (Table III) suggests that a biochemical pathway involving a PTK, coreceptor, and PTPase may be disrupted in these three strains of mice. Recent biochemical and genetic data strongly support this hypothesis. The BCR-induced tyrosine phosphorylation of CD22 is dramatically reduced, but not completely abolished, in B cells from Lyn-deﬁcient mice (Cornall et al., 1998; Smith et al., 1998; Nishizumi et al., 1998; Chan et al., 1998). Furthermore, SHP-1 recruitment to CD22 after BCR cross-linking is undetectable in Lyn⫺/⫺ B cells (Cornall et al., 1998; Smith et al., Nishizumi et al., 1998; Chan et al., 1998). Cornall et al. also showed that Lyn, CD22, and SHP-1 are limiting components of a biochemical pathway that regulates mIgM and MHC class II expression and sensitivity to autoantigen-induced anergy (Cornall et al., 1998). Lyn⫹/⫺, CD22⫹/⫺, SHP-1⫹/mev mice—which express one-half the wild-type levels of each of these molecules—were bred to generate mice heterozygous for either each of the three components, two of the three components, or all of the three components of this proposed pathway. Defects became more pronounced as the number of heterozygous loci increased until the triple-heterozygous mice displayed defects similar to those displayed by Lyn⫺/⫺ and SHP-1 mev/mev mice (Cornall et al., 1998). Taken together, the polygenic nature of these defects and biochemical data suggest a common signal transduction pathway involving Lyn, CD22, and SHP-1.

113

B CELL SIGNAL TRANSDUCTION PATHWAYS

TABLE III COMPARISON OF THE PHENOTYPES AND in Vivo/in Vitro RESPONSES OF LYN⫺/⫺, CD22⫺/⫺, AND SHP-1 mev/mev MICE RELATIVE TO WILD-TYPE CONTROL MICE Response Surface IgM expression level MHC class I expression level Calcium mobilization in response to F(ab⬘)2 antiIgM Serum Ig levels Number of mature B cells in bone marrow Life-span of peripheral B cells Autoimmune disorders Conventional B cell number CD5⫹ B cell number Proliferation in response to F(ab⬘)2 anti-IgM Proliferation in response to LPS

Lyn⫺/⫺

Cd22⫺/⫺

SHP-1 mev/ mev

Decreased Increased Increased

Decreased Increased Increased

Decreased Increased Increased

Dramatically increased Decreased

Slightly increased Decreased

Dramatically increased —

Decreased

Decreased

—

Yes Decreased Increased Increased

No Normal Normal Decreased

Yes Decreased Increased Increased

Normal

Increased

Normal

Precisely why Lyn is required for maximal phosphorylation of CD22 and recruitment of SHP-1 is unknown, but at least three possible models can be envisioned. The simplest model is that Lyn directly phosphorylates the tyrosines within CD22 required for SHP-1 binding—probably Y843 and/or Y863 (Fig. 6a). Thus, SHP-1 would bind directly to these phosphotyrosines via its SH2 domains, whereby its PTPase activity increases due to this binding and its proximity to the lipid bilayer (Zhao et al., 1993; Doody et al., 1995; Law et al., 1996b). Alternatively, Lyn may phosphorylate tyrosine residues in the cytoplasmic tail of CD22 that are not directly involved in SHP-1 recruitment and activation. This phosphorylation may lead to binding and activation of another PTK that phosphorylates the tyrosine residues required for direct SHP-1 binding (Fig. 6b). Finally, Lyn may not phosphorylate CD22, but may be required for the activation of another PTK that phosphorylates CD22, thereby indirectly promoting the recruitment of SHP-1 (Fig. 6c). We present these three alternatives to underscore the range of possibilities that might account for the phenotypes reported. Exactly how SHP-1 regulates surface IgM and MHC class II levels and the responses leading to release of intracellular free calcium after binding

114

ANDREW CRAXTON et al.

A

B FIG. 6. Possible models for Lyn-dependent tyrosine phosphorylation of CD22 and the recruitment of SHP-1. (A) Lyn may directly phosphorylate tyrosine residues within the cytoplasmic tail of CD22 required for SHP-1 recruitment. Antigen-induced cross-linking of the BCR (1) induces activation of the Src family PTK Lyn (2). Activated Lyn directly phosphorylates tyrosine residues within the cytoplasmic tail of CD22 (3) required for the recruitment and activation of the PTPase SHP-1 (4). Activated SHP-1 subsequently regu-

B CELL SIGNAL TRANSDUCTION PATHWAYS

115

C FIG. 6 (Continued). lates, among other events, an increase in intracellular free calcium, surface IgM expression, and MHC class II expression (5). (B) Lyn may directly phosphorylate CD22 but may indirectly induce the recruitment of SHP-1. Antigen-induced (1) activated Lyn (2) directly phosphorylates tyrosine residues within the cytoplasmic tail of CD22 (3) at residues required for the recruitment and activation of another PTK (4). This PTK subsequently phosphorylates additional sites within the cytoplasmic tail of CD22 required for SHP-1 recruitment and activation (6). Activated SHP-1 regulates B cell responses as described in (A) above (7). (C) Lyn may indirectly induce the phosphorylation of CD22. Engagement of the BCR by antigen (1) activates Lyn (2), which directly activates another PTK (3). This PTK phosphorylates the cytoplasmic tail of CD22 (4) and induces the recruitment and activation of SHP-1 (5), allowing SHP-1 to regulate signals from the BCR (6).

CD22 is not understood. Because inducible phosphorylation of CD22 is still observed at reduced levels in B cells from Lyn-deﬁcient mice (Cornall et al., 1998; Smith et al., 1998; Nishizumi et al., 1998; Chan et al., 1998), one prediction of all three models is that a PTK other than Lyn is able to utilize the cytoplasmic tail of CD22 as a substrate. Syk is the most likely candidate as it may physically associate with CD22 (Law et al., 1996b; Tuscano et al., 1996a). The phosphorylation of CD22 by this PTK may not be sufﬁcient in the absence of Lyn to recruit SHP-1. Finally, the involvement of Lyn in regulating the recruitment of molecules other than SHP-1 to CD22 remains to be addressed. 5. Regulation of MAP Family Kinases by CD22 Cross-linking CD22 by itself does not induce proliferation or an increase in [Ca2⫹]i but does induce tyrosine phosphorylation of CD22 (Pezzutto et

116

ANDREW CRAXTON et al.

al., 1987, 1988; Leprince et al., 1993; Law et al., 1996b). Several studies have demonstrated that preincubation of B cells with monoclonal antibody speciﬁc for CD22 reduces the amount of anti-IgM needed for subsequent BCR-induced proliferation, and [Ca2⫹]i increases (Pezzutto et al., 1987, 1988; Doody et al., 1995). Recent studies using a similar experimental approach, suggest that CD22 may regulate BCR-induced activation of the MAPK family of serine/threonine kinases. Preincubation of murine splenic B cells with a cross-linking CD22 mAb, followed by activation through the BCR, leads to elevated ERK2 and JNK activation relative to preincubation with a control mAb (Tooze et al., 1997; Chan et al., 1998). Tooze et al. used biotinylated monovalent Fab fragments of ␬- and CD22-speciﬁc mAbs, which bind their respective epitopes without cross-linking (Tooze et al., 1997). Cross-linking was subsequently induced by addition of avidin. They found that cross-linking ␬ alone activates ERK2, JNK, and p38 MAPK, whereas cross-linking CD22 by itself leads to only modest increases of ERK2 and JNK activity (Tooze et al., 1997). When avidin and biotinylated Fab fragments of both ␬ and CD22 mAbs were used to co-cross-link CD22 with the BCR, the levels of ERK2 and JNK activity were reduced to the amount seen by cross-linking CD22 alone. The activity of p38 MAPK was unchanged relative to biotinylated anti-␬ Fab fragments alone. Hence, cocross-linking CD22 with the BCR appears to dampen the BCR-induced activation of both ERK2 and JNK but not p38 MAPK. Whether coligation of CD22 with the BCR to this degree in vitro has a counterpart in vivo is unclear. As indicated previously, only a small percentage of CD22 (0.2 to 2%) normally associates with the BCR (Leprince et al., 1993; Peaker and Neuberger, 1993). Though CD22 appears to regulate MAPK activation, exactly how it does so is unclear. Antibody cross-linking and co-cross-linking experiments are difﬁcult to interpret. Two models have been proposed to explain these data. The ﬁrst model proposes that cross-linking CD22 actually generates a ‘‘positive’’ subthreshold or incomplete signal that synergizes with signals from the BCR (Pezzutto et al., 1987, 1988; Tuscano et al., 1996b). An alternative model, which is consistent with the hypothesis that CD22 is simply a ‘‘negative’’ regulator, states that preincubation of B cells with CD22 mAb excludes CD22 from the BCR complex, thus preventing its negative regulatory inﬂuence and allowing for a decreased threshold for BCR signaling (Doody et al., 1995; Tooze et al., 1997). According to this model, cross-linking CD22 in the absence of BCR engagement leads to spontaneous signaling from the BCR due to lack of repression by CD22 resulting in the low level of MAPK activation observed by CD22 crosslinking alone.

B CELL SIGNAL TRANSDUCTION PATHWAYS

117

Neither model completely explains certain observations. We have found that BCR-induced ERK2 and JNK, but not p38 MAPK activation, is slightly elevated in B cells from CD22-deﬁcient mice relative to wild-type mice (Otipoby et al., 1999). This is not inconsistent with the hypothesis that CD22 normally functions as a ‘‘negative regulator’’ of the BCR by recruiting SHP-1 PTPase. As described above, genetic and biochemical data strongly suggest that an inhibitory pathway involving Lyn, CD22, and SHP-1 exists in B cells. If CD22 functions to regulate MAPK activation solely in a Lynand SHP-1-dependent manner, one would predict that preincubation of B cells from Lyn⫺/⫺ mice with CD22 mAb relative to control mAb would not lead to elevated BCR-induced MAPK activation because the inhibitory pathway is already disrupted. However, this is not the case. Chan et al. (1998) found that CD22 is still capable of regulating BCR-induced ERK and JNK activation in the absence of Lyn. Again, whether this Lynindependent regulation occurs in a ‘‘positive’’ or ‘‘negative’’ manner is unclear due the experimental approach of cross-linking CD22 with mAb prior to stimulation through the BCR. Because CD22 is still phosphorylated after BCR engagement in B cells from Lyn⫺/⫺ mice, albeit at much reduced levels compared with those in wild-type mice (Cornall et al., 1998; Smith et al., 1998; Nishizumi et al., 1998; Chan et al., 1998), CD22 may regulate MAPK activation via a separate, Lyn-independent pathway. A comparison of the phenotypes of Lyn⫺/⫺, CD22⫺/⫺, and SHP-1 mev/mev mice (Table III) also suggests that CD22 may mediate functions distinct from a Lyn 씮 CD22 씮 SHP-1 pathway. For example, B cells from CD22⫺/⫺ mice, unlike Lyn⫺/⫺ and SHP1 mev/mev mice, proliferate strongly in response to LPS and have normal numbers of conventional and CD5⫹ B1 B cells. In considering a possible second BCR-activated pathway that is regulated by CD22, we were inﬂuenced by the ﬁnding of Hashimoto et al. (1998), who observed that Shc⫺/⫺ B cells, when triggered via the BCR, have a normal pattern of ERK2 activation, whereas Grb2⫺/⫺ B cells have reduced BCR-induced activation of ERK2. This suggests that B cells must have an Shc-independent pathway for activating ERK2 via Grb2. A 36-kDa membrane-associated T cell-associated molecule, pp36, has been shown to interact with Grb2/Sos and thereby promote activation of the Ras pathway; pp36 binds to the SH2 domains of Grb2, PLC-웂1, and the p85 subunit of PI3K (Buday et al., 1994; Sieh et al., 1994; Fukazawa et al., 1995). pp36 or LAT was recently cloned and characterized by Zhang et al. (1998b). LAT is a T/NK-cell restricted membrane protein with a hydrophobic anchor and a long cytoplasmic tail, which becomes tyrosine-phosphorylated probably by ZAP-70 in response to TCR ligation. As expected, tyrosinephosphorylated LAT binds Grb2, the p85 subunit of PI3K, and PLC-

118

ANDREW CRAXTON et al.

웂1. When LAT is mutated at consensus tyrosines (YVNV) for Grb2 binding, mutant LAT acts as a dominant negative inhibitor to block both TCRinduced AP-1 and NF-AT transcriptional activation, suggesting that LAT functions to link Grb2/Sos to the membrane fraction. Our preliminary data suggest that CD22 may play at least one function analogous to that of LAT in B cells. We were interested in identifying B cell surface molecules that have a sequence analogous to the Grb2-binding region in LAT, and noted that CD22 appeared to have such a region (Fig. 7). As predicted, after BCR ligation, Grb2 binds phosphorylated CD22 to initiate an activation pathway (Otipoby et al., 1999). Thus, it appears that CD22 may have multiple signaling functions, which may explain speciﬁc differences between CD22⫺/⫺ mice and Lyn⫺ or SHP-1-deﬁcient mice (Table III). Thus, the Grb2-mediated pathway would also provide a means by which CD22 could regulate ERK and possibly JNK activity in a Lyn-independent manner (Chan et al., 1997, 1998). B. CD19 CD19 is expressed by early B cell precursors and mature peripheral B cells and can form a complex with CD21, Leu-13, and CD81 (Bradbury et al., 1992; Tedder et al., 1997a). The cytoplasmic tail of CD19 rapidly becomes tyrosine-phosphorylated following BCR cross-linking and may associate with several signaling molecules, including the Src family tyrosines kinases (Lyn, Fyn, and Lck), PI3K, and Vav (Uckun et al., 1993; Tuveson et al., 1993; van Noesel et al., 1993; Chalupny et al., 1993, 1995; Weng et al., 1994). Moreover, coligation of CD19 with the BCR leads to a decreased threshold for BCR-induced proliferation (Carter and Fearon, 1992). These data suggest that CD19 may regulate several signaling pathways and may help establish the threshold of signaling by the BCR (O’Rourke et al., 1997; Tedder et al., 1997a).

FIG. 7. Homology of the Grb2-binding domain of murine CD22 with murine LAT. Two tyrosines within the cytoplasmic tail of murine LAT, Y171 and Y191, are required for association with the SH2 domain of Grb2. These two tyrosine residues reside within a motif with sequence homology, over a limited region, with the sequence surrounding Y828 of murine CD22.

B CELL SIGNAL TRANSDUCTION PATHWAYS

119

Studies of genetically altered mice have lent support for this hypothesis. CD19-deﬁcient mice have been generated by targeted gene disruption. These mice show a dramatic reduction in peritoneal CD5⫹ B cell number and a reduction in conventional B cell number in the periphery (Rickert et al., 1995; Engel et al., 1995; Sato et al., 1997). CD19 is also required for antibody responses to T cell-dependent antigens and the formation of germinal centers (Rickert et al., 1995; Engel et al., 1995; Sato et al., 1997). Transgenic mice that overexpress CD19 (CD19tg) have demonstrated that even small changes in CD19 expression levels can have signiﬁcant effects on B cell function (Sato et al., 1997). Overexpression of CD19 by two- to threefold leads to impaired development and reduced numbers of conventional B cells in the periphery (Zhou et al., 1994). On the other hand, the number of CD5⫹ B cells in the peritoneum and spleen increases as the levels of CD19 increase (Sato et al., 1996b), demonstrating the importance of CD19 in CD5⫹ B cell development. The antibody responses to different forms of antigens in mice that overexpress CD19 were the opposite of those in CD19-deﬁcient mice. In other words, CD19⫺/⫺ mice have decreased IgM and IgG responses to dinitrophenyl–keyhole limpet hemocyanin (DNP– KLH), and CD19tg mice have augmented responses to the same TD antigen. The opposite is observed when mice are immunized with the TI2 antigen DNP–Ficoll; namely, CD19⫺/⫺ mice generate augmented levels of antibody whereas CD19tg mice are hyporesponsive (Sato et al., 1995a). These data clearly demonstrate the importance of CD19 in B cell development and activation in response to different forms of antigen. To test directly the ability of CD19 to regulate the threshold of BCR signaling, Inaoki et al. (1997) crossed CD19tg mice with mice expressing a BCR speciﬁc for HEL (anti-HEL) and mice expressing soluble HEL (sHEL). Normally, mice that express both anti-HEL and sHEL are anergic and do not secrete anti-HEL antibodies (Goodnow et al., 1988). Overexpression of CD19 via a transgene led to signiﬁcant breakdown of autoantigen-induced silencing of B cells (Inaoki et al., 1997). Despite a reduction in peripheral B cell numbers, CD19tg/anti-HEL/sHEL triple-transgenic mice secreted, on average, more than 500-fold more autoantibody than did anti-HEL/sHEL double-transgenic mice (Inaoki et al., 1997). Interestingly, autoantigen-induced negative selection of the autoreactive B cells appears to be unperturbed by the overexpression of CD19. These ﬁndings contrast dramatically with the observation that the absence of the PTPase SHP-1 lowers the threshold for deletion of autoreactive B cells (Cyster and Goodnow, 1995). Future studies addressing the molecular mechanisms responsible for this breakdown in B cell silencing but normal negative selection should prove interesting.

120

ANDREW CRAXTON et al.

C. FC웂RIIB Coligation of the BCR with Fc애RIIB leads to abrogation of BCR signals and inhibition of proliferation and inﬂux of extracellular calcium (Phillips and Parker, 1983, 1984; Wilson et al., 1987; Choquet et al., 1993). CD19 phosphorylation is also inhibited when Fc웂RIIB is coligated with the BCR (Hippen et al., 1997). For these reasons, Fc웂RIIB is believed to function as a ‘‘negative’’ regulator of BCR signaling (Coggeshall, 1998). This is supported by the observation that Fc웂RIIB-deﬁcient mice generate elevated levels of antigen-speciﬁc IgG in response to immunization with TD and TI antigens (Takai et al., 1996). The cytoplasmic tail of Fc웂RIIB contains a 13-amino acid tyrosine-based motif responsible for its inhibitory properties (Muta et al., 1994). Because phosphorylation of a single tyrosine residue is critical for the function of Fc웂RIIB, it was proposed that a PTK associated with the BCR phosphorylates Fc웂RIIB and promotes the recruitment of an SH2 domain-containing signaling molecule, which subsequently down-regulates signaling cascades normally initiated via the BCR (Muta et al., 1994). Other data suggest that Lyn is a good candidate for this PTK. B cells from Lyn⫺/⫺ mice stimulated with whole anti-IgM, which coligates Fc웂RIIB with the BCR, have a signiﬁcant reduction in Fc웂RIIB phosphorylation and decreased inhibition of BCR-induced calcium inﬂux relative to B cells from wild-type mice (Nishizumi et al., 1998; Chan et al., 1998). Experiments designed to identify the molecules recruited to the cytoplasmic tail of phosphorylated Fc웂RIIB have led to the identiﬁcation of two SH2 domain-containing PTPases—SHP-1 and SHP-2—and the phosphatidylinositol-3,4,5-trisphosphate 5-phosphatase, SHIP (D’Ambrosio et al., 1995, 1996b; Damen et al., 1996; Ono et al., 1996). Initial reports demonstrated a role for SHP-1 in Fc웂RIIB-mediated inhibition of BCR-induced signaling cascades. The proliferative responses of B cells puriﬁed from SHP-1 me/me and SHP-1 mev/mev mice, unlike the responses of B cells puriﬁed from control mice, were not completely inhibited by coligation of Fc웂RIIB with the BCR (Pani et al., 1995; D’Ambrosio et al., 1996b). The observation that Fc웂RIIB was still capable of partially inhibiting BCR-induced proliferative responses in the absence of SHP-1 suggested that another molecule might be involved. Additional studies have questioned the involvement of SHP-1 in Fc웂RIIB-mediated inhibition of BCR-generated signals altogether. Nadler et al. (1997) used immortalized B cell lines generated from wild-type and SHP-1 me/me mice and found that SHP-1 is completely dispensable for Fc웂RIIB inhibition of BCRmediated calcium responses and CD19 phosphorylation. Furthermore, SHIP is still recruited to Fc웂RIIB in the absence of SHP-1 (Nadler et al., 1997).

B CELL SIGNAL TRANSDUCTION PATHWAYS

121

Ono et al. (1997) signiﬁcantly extended these observations using wildtype, SHIP-deﬁcient, and SHP-1-deﬁcient DT40 B cell lines expressing Fc웂RIIB. This study showed that SHIP is absolutely required for Fc웂RIIB inhibition of BCR-induced calcium inﬂux, whereas SHP-1 appears to be completely dispensable (Ono et al., 1997). In wild-type DT40 cells, coligation of Fc웂RIIB and the BCR did not inhibit BCR-induced apoptosis. Surprisingly, coligation in SHIP-deﬁcient cells led to a dramatic increase in sensitivity to antigen receptor-induced apoptosis, suggesting that an undeﬁned signaling domain within the tail of Fc웂RIIB promotes BCRinduced cell death and that this signal is attenuated by SHIP (Ono et al., 1997). This was further supported by the expression of a Y 씮 F mutant of Fc웂RIIB in wild-type DT40, which was unable to recruit SHIP. Coligation of this mutant Fc웂RIIB with the BCR led to a 10-fold increase in sensitivity to BCR-induced death. This apoptosis-enhancing domain and the speciﬁc signaling molecules responsible for these effects remain to be identiﬁed. D. CDW150 CDw150 was deﬁned initially with the mAb IPO-3, and was found to be localized in the cytoplasm of GC cells and on the surface of mantle zone B cells (Pinchouk et al., 1988; Sidorenko et al., 1992). IPO-3/CDw150 is up-regulated after CD40 ligation of B cells and expressed on immature thymocytes and CD45RO⫹ T cells, suggesting that it is a lymphocyte activation marker (Sidorenko and Clark, 1993). Cocks et al. (1995) described a cDNA encoding a surface molecule, known as SLAM, with properties strikingly similar to those of IPO-3. Transfectants expressing SLAM were bound by both IPO-3 and the A12 anti-SLAM mAb (Sidorenko, 1997); this allowed IPO-3/SLAM to be given a ‘‘CD’’ designation, CDw150 (Mason et al., 1997; Sidorenko, 1997). CDw150 is also expressed on DCs, and the expression of CDw150 is increased on activated T cells, B cells, or dendritic cells (Sidorenko and Clark, 1993; Cocks et al., 1995; Polacino et al., 1996). Ligating CDw150 on B cells augments proliferation induced by CD40 mAb and IL-4 (Sidorenko and Clark, 1993), and engaging CDw150 on activated T cells results in IL-2- and CD28-independent proliferation (Punnonen et al., 1997). Ligation of CDw150 also induces IFN-웂 production by CD4⫹ T cell clones and Ig production by activated B cells (Cocks et al., 1995; Punnonen et al., 1997; Aversa et al., 1997). Thus, CDw150 may be involved in expanding Th1 immune responses (Aversa et al., 1997). Ligating CDw150 on certain B and T cell lines also accelerates and increases CD95-mediated apoptosis (Mikhalap et al., 1999), implying that CDw150, like other ‘‘coreceptors,’’ has more than one signaling function.

122

ANDREW CRAXTON et al.

The mechanism by which CDw150 exerts these effects on lymphocytes is unknown. CDw150 is a sialylated phosphoglycoprotein of approximately 70 to 95 kDa in size and belongs to the CD2 subset of the Ig superfamily of type I transmembrane glycoproteins (Sidorenko and Clark, 1993; Cocks et al., 1995). It has an extracellular region with 209 amino acids arranged in one V and one C2 Ig-like domain containing eight potential glycosylation sites and a 77-residue cytoplasmic tail containing several tyrosines within motifs for SH2 domain binding sites. CDw150 has associated with it both tyrosine and serine/threonine kinase activities, and ligating CDw150 on B lymphocytes with mAb IPO-3 induces a rapid elevation of [Ca2⫹]i (Sidorenko and Clark, 1993). Mikhalap et al. (1999) have found that CDw150, like CD5, CD6, CD19, and CD22, is tyrosine phosphorylated after BCR ligation; it can associate with the Src family kinase Fgr, the inositol polyphosphate 5-phosphatase, SHIP, and also with the tyrosine phosphatase CD45. Ligation of CDw150 on B cells induces rapid tyrosine dephosphorylation of SHIP, suggesting that CDw150 may function to down-regulate SHIP activity and thereby increase [Ca2⫹]i levels, perhaps through increasing the cellular pool of PtdIns (3,4,5)P3. Sayos et al. (1998) used a yeast two-hybrid screen with the CDw150 tail as bait to identify a novel 15-kDa protein with one SH2 domain that can bind to CDw150 immunoprecipitated from activated T cells. This molecule, termed SAP (Nichols et al., 1998) or DSHP, is encoded by an X-linked gene, which causes X-linked lymphoproliferative disease (XLP). Patients with XLP have uncontrolled B cell proliferation following Epstein–Barr virus (EBV) infection (Seemayer et al., 1995; Nichols et al., 1998). Thus, it is possible to envision that SAP may interfere directly with the role of CDw150 in promoting B cell death (Mikhalap et al., 1999). Sayos et al. (1998) detected SAP in the Raji Burkitt’s lymphoma cell line, suggesting that SAP may be expressed in some B cells (Nichols et al., 1998). However, other B cell lines did not express SAP, whereas SAP was expressed in T cells. Furthermore, XLP is believed not to be a B cell defect but to be related to defective Th1 responses (Seemayer et al., 1995). Thus, Sayos et al. (1998) speculated that XLP is caused by the inability of SAP to regulate T cell responses. However, further studies investigating the potential role of SAP in CDw150⫹ B cell subsets should help clarify whether some XLP defects are manifested at the B cell level. IV. Regulation of BCR-Induced Responses by CD40

A. INTRODUCTION Since the identiﬁcation of the CD40 antigen using mAbs more than a decade ago, a plethora of studies have established critical roles for CD40 in the regulation of the immune response. In B lymphocytes at least, these

B CELL SIGNAL TRANSDUCTION PATHWAYS

123

functions include cell activation, proliferation, immunoglobulin isotype switching, germinal center B cell formation, and B cell maturation into a memory cell phenotype. Further information concerning the role of CD40 on other cell types in vitro and in vivo can be found in two review articles (van Kooten and Banchereau, 1996, 1997). The primary focus of this section is to provide a current synopsis of CD40-mediated signaling in B lymphocytes and to summarize and develop ideas on how signals generated after CD40 receptor engagement integrate with signals delivered via the BCR to regulate B cell responses. It is important to recognize that the combination of signal transduction pathways activated via CD40 may be unique at different stages of B cell maturation, because speciﬁc signaling molecules—e.g., kinases, phosphatases, G proteins, cyclins, caspases, and Bcl-2 homologues—may be absent or expressed differentially. Thus, how the CD40 and BCR signaling pathways integrate depends on the stage of B lymphocyte development. The CD40 receptor is a 45- to 50-kDa type I transmembrane glycoprotein member of the rapidly growing TNF receptor superfamily (Table II) (Smith et al., 1994; Banchereau et al., 1994). CD40 is expressed on B cells, follicular dendritic cells (FDCs), dendritic cells (DCs), activated monocytes, macrophages, endothelial and epithelial cells, and vascular smooth muscle cells (Clark and Ledbetter, 1986; Hart and MacKenzie, 1988; Schriever et al., 1989; Galy and Spits, 1992; Alderson et al., 1993; Scho¨nbeck et al., 1997). Receptor-mediated CD40 signaling is initiated by binding of CD40 ligand (CD40L or CD154) to its receptor, CD40. CD40L/ CD154 is a 33- to 35-kDa type 2 membrane glycoprotein expressed primarily on activated CD4⫹ T cells and some other cell types including macrophages, DCs, endothelial cells, and smooth muscle cells (Armitage et al., 1992; Pinchuk et al., 1996; Mach et al., 1997). Engagement of CD40 provides a potent costimulatory signal to resting mature B cells along with so-called competence signals via either mIgM or CD20 (Clark and Ledbetter, 1986). Similar to either CD22 or CD19 engagement, ligation of CD40 receptors can lower the threshold of signaling via the BCR required to induce cell proliferation (Wheeler et al., 1993). CD40 receptor ligation, in combination with BCR stimulation, may also induce a partial germinal center phenotype within resting B cells (Galibert et al., 1996). In certain human B cell lines (Valentine and Licciardi, 1992; Sumimoto et al., 1994; Lens et al., 1996) or the murine WEHI-231 line (Tsubata et al., 1993; Choi et al., 1995; Fang et al., 1995; Ishida et al., 1995; Wang et al., 1995), anti-Ig induces rather than prevents cell death, and this activationinduced cell death, rather than being enhanced, is blocked by CD40 ligation. However, it is not entirely clear when and where CD40 actually

124

ANDREW CRAXTON et al.

functions in vivo to protect immature B cells from BCR-induced cell death. Some investigators have found that some but not all mature B cells induced to die by extensive mIgM cross-linking can be rescued from death by CD40 mAb (Parry et al., 1994; Nomura et al., 1996); others, such as Marshall-Clarke et al. (1996), have found that CD40L could not block BCR-induced death of immature B cells. Monroe and co-workers have identiﬁed an immature B cell subset protected by CD40 ligation ( J. Monroe, personal communication, 1999). CD40 receptor ligation can also prevent spontaneous apoptosis of isolated human GC B cells (Liu et al., 1989; Billian et al., 1997). Such CD40-mediated rescue from apoptosis may be of critical importance in the development of high afﬁnity antibodygenerating memory B cells. In contrast, CD40 ligation can also lead to growth arrest or even cell death of B lineage cells. We ﬁrst observed this using the M12 early B cell line, which stops dividing and dies after CD40 ligation (Inui et al., 1990; Clark and Shu, 1990). The inhibitory effect of CD40 signaling is also observed in other mouse B cell lines (Heath et al., 1993). Funakoshi et al., (1994) found that ligating CD40 on certain lymphoma cells either with mAb or CD40L led to inhibition of proliferation both in vitro and in vivo. In addition, the establishment of EBV-induced human B cell lymphomas, but not normal B cells, is blocked by CD40 mAb in severe combined immunodeﬁcient (Scid ) mice (Murphy et al., 1995). Furthermore, CD40 ligation induces arrested cell growth and programmed cell death in the CD40⫹ myeloma line XG2 (Pellat-Deceunynck et al., 1996; Bergamo et al., 1997). The possibility that CD40 may also induce caspase-1 (ICE) expression and its activation in some B cells, similar to its role in vascular smooth muscle cells, should also be considered (Scho¨nbeck et al., 1997). It is not entirely clear when in B cell development CD40 may exert inhibitory effects; however, CD40 ligation can clearly block terminal B cell differentiation of GC B cells into plasma cells (Callard et al., 1995; Arpin et al., 1995). Arpin et al., (1997) suggested that once B cells are established as memory B cells, CD40 ligation is not as effective in blocking their differentiation as it is in inhibiting naive B cell differentiation. Miyashita et al., (1997) found that IgD⫺ memory B cells in particular were sensitive to inhibition by CD40 ligation. The molecular basis of this inhibition is not known, but it does appear that CD40 signaling pathways mediate different functions in naive, GC, and memory B cells. B. CD40-MEDIATED RESCUE FROM ANTI-IG-INDUCED APOPTOSIS As noted above, CD40 signaling may play an important role in both the rescue of immature B cells from BCR-induced cell death and prevention of spontaneous death of GC B cells (Liu et al., 1989; Tsubata et al., 1993).

B CELL SIGNAL TRANSDUCTION PATHWAYS

125

Such programmed cell death is a mechanism for regulating the clonal deletion of immature and mature B cells triggered by multivalent antigens interacting with the BCR. However, the mechanisms by which CD40 prevents anti-Ig-induced apoptosis are not well deﬁned. Induction of members of the Bcl-2 family of protooncogenes appears to play at least some role in regulating CD40-mediated survival in the WEHI-231 B cell lymphoma line. Although studies using this cell line have produced variable results Bcl-xL, but not Bcl-2, appears to play a role in CD40-dependent rescue from anti-Ig-induced apoptosis. Overexpression of Bcl-xL, but not Bcl-2, prevents BCR-mediated cell cycle arrest and apoptosis (Cuende et al., 1993; Choi et al., 1995; Ishida et al., 1995; Merino et al., 1995). Moreover, CD40 ligation up-regulates Bcl-xL but not Bcl-2 at both transcriptional and translational levels (Choi et al., 1995; Ishida et al., 1995; Wang et al., 1995). In addition, the endogenous levels of Bcl-2 do not correlate with the susceptibility of various WEHI-231 sublines to undergo anti-IgM-induced death (Hibner et al., 1993; Gottschalk et al., 1994). Also, CD40 receptor ligation can prevent B lymphocyte apoptosis in Bcl-2-deﬁcient mice (Nakayama et al., 1995). Further evidence of a role for Bcl-xL in CD40-mediated rescue comes from studies using a Bcl-xL antisense oligonucleotide. An antisense, but not a sense, Bcl-xL S-oligonucleotide partially inhibited CD40-dependent rescue from anti-Ig-triggered cell death (Wang et al., 1995). However, additional studies are needed to conﬁrm these ﬁndings because it was not apparent whether the Bcl-xL antisense oligonucleotide exerted its effects speciﬁcally by down-regulating Bcl-xL protein levels. We have found that another Bcl-2 family member may play a role in CD40-mediated rescue from anti-Ig-induced apoptosis in WEHI-231 cells. CD40 receptor engagement, in either the presence or absence of anti-Ig, strongly induces sustained increases in the antiapoptotic Bcl-2 homologue A1/Bﬂ-1, in addition to Bcl-xL, at least at the transcriptional level, whereas the levels of other Bcl-2 family members (Bcl-2, Bax, Bak, and Bad) are not signiﬁcantly changed (A. Craxton, P. I. Chuang, G. L. Shu, A. Alcher, D. M. Willerford, and E. A. Clark, in preparation). Interestingly, A1/ Bﬂ-1 is a relatively unique Bcl-2 family member in that it is induced by multiple cytokines and growth factors, including TNF-움, IL-1, LPS, GMCSF, all-trans-retinoic acid (ATRA), and vascular endothelial growth factor (VEGF) (Lin et al., 1993; Karsan et al., 1996; Moreb and Schweder, 1997; Chuang et al., 1998; Gerber et al., 1998). It is intriguing to speculate that the ability of TNF-움 to mimic CD40 ligand and prevent anti-Ig-induced apoptosis in the Ramos B cell line may reﬂect in part its propensity to upregulate A1/Bﬂ-1 expression (Lens et al., 1996; Karsan et al., 1996). In addition, A1/Bﬂ-1 mRNA levels are up-regulated 10-fold as B cells are

126

ANDREW CRAXTON et al.

recruited into the long-lived peripheral B cell pool, suggesting that A1 may be important in regulating B lymphocyte life-span (Tomayko and Cancro, 1998). More detailed studies are required to establish the precise role of A1/Bﬂ-1 in CD40-mediated rescue from anti-Ig-dependent cell death. Such studies should include an examination of the role of receptorregulated protein–protein interactions in A1/Bﬂ-1 function because Bcl2 family members may either homo- or heterodimerize with each other and other polypeptides, such as the CED-4 homologue Apaf-1 (Chao and Korsmeyer, 1998; Newton and Strasser, 1998). A role for sustained c-myc expression in CD40-mediated rescue of antiIg-induced apoptosis in WEHI-231 cells has been proposed on the basis of the ability of CD40 receptor ligation to maintain elevated c-myc levels following anti-Ig treatment (Schauer et al., 1996; Siebelt et al., 1997). In contrast, c-myc levels are transiently increased in response to anti-Ig stimulation but subsequently decline to subbasal levels as apoptosis occurs. Such effects appear to be mediated via differences in NF-␬B/Rel transcription factor binding, because NF-␬B/Rel plays a major role in the regulation of c-myc expression (Sonenshein, 1997). The tumor suppressor p53 and one of its transcriptional targets, the cyclin-dependent kinase inhibitor p21Waf1/Cip1, also appear to play roles in CD40-mediated rescue of BCR-induced cell death in WEHI-231 cells (Wu et al., 1998). Though BCR engagement increased p53 and p21Waf1/Cip1 protein levels, concurrent ligation of both the BCR and CD40 prevented p21Waf1/Cip1 expression, showing that the levels of p21Waf1/Cip1 correlated with the ability of CD40 to abrogate anti-IgM-induced apoptosis. Furthermore, microinjection of either a p21Waf1/Cip1 antisense plasmid vector or p21 antibodies signiﬁcantly reduced BCR-induced apoptosis, strongly suggesting that p21Waf1/Cip1 plays an important role in the regulation of cell fate in response to BCR signaling in WEHI-231 cells (Wu et al., 1998). Correlative studies have also suggested that the levels of the cyclindependent kinase inhibitor (CKI) p27Kip1, a member of the Cip/Kip family of CKIs, may play a role in CD40-mediated prevention of anti-Ig-induced cell cycle arrest in the late G1 phase in the WEHI-231 cell line (Ezhevsky et al., 1996; Han et al., 1996). Anti-Ig, but neither anti-CD40 nor a combination of both anti-Ig and anti-CD40, induced p27Kip1 expression and correlated with decreased levels of cyclin-dependent kinase, Cdk2 activity, which is required for cell cycle progression in the late G1 phase. A role for two other cyclin-dependent kinases, Cdk4 and Cdk6, in anti-Ig-induced cell cycle arrest is also possible, because anti-Ig alone, but not a combination of anti-Ig and anti-CD40, down-regulates both Cdk4 and Cdk6 protein levels (Ishida et al., 1995). Speciﬁc members of the IAP family of proteins (Table I) may also play a role in CD40-mediated B lymphocyte survival. CD40 receptor

B CELL SIGNAL TRANSDUCTION PATHWAYS

127

engagement up-regulated cIAP2 levels, but not its mammalian IAP family members cIAP1, XIAP, or NAIP (Craxton et al., 1998). This observation is particularly intriguing because cIAP2, in addition to cIAP1 and XIAP, binds directly to and inhibits caspase-3 and caspase-7 (Deveraux et al., 1997, 1998; Roy et al., 1997). Moreover, both cIAP1 and cIAP2 have been shown to associate indirectly with TNFR1 via direct interactions with TRAF1 and TRAF2 (Rothe et al., 1995), suggesting that these cIAPs may also bind indirectly to the cytoplasmic region of CD40 via TRAF2 and perhaps other TRAF family members. The possibility that such hypothetical cIAP interactions may modulate speciﬁc CD40 signaling pathways should also be considered, because one IAP member, XIAP, speciﬁcally activates JNK1, but not ERK2, p38 MAPK, or ERK5/BMK1 (Sanna et al., 1998). Further studies should investigate the potential role of the mammalian IAPs in CD40-induced rescue from anti-Ig-induced apoptosis. C. PROTEIN TYROSINE KINASES IN CD40-MEDIATED SIGNALING Several studies have suggested that CD40 receptor ligation induces changes in tyrosine phosphorylation of a subset of unidentiﬁed proteins, reﬂecting perhaps both activation and inactivation of speciﬁc PTKs and PTPases (Uckun et al., 1991; Faris et al., 1994). Although some reports suggest that CD40 engagement induces phosphorylation and concomitant activation of both the Src family PTKs (Lyn, Fgr, Blk) and ZAP-70 family PTK Syk (Ren et al., 1994; Gulbins et al., 1996), others suggest that both Lyn and Syk are rapidly and transiently dephosphorylated following CD40 receptor ligation (Faris et al., 1994). Further studies are required to clarify unequivocally whether Src family PTKs or Syk are activated following CD40 engagement, but it does appear that Lyn plays a role in CD40induced B cell proliferation (Nishizumi et al., 1995) and the induction of CD95/Fas on B cells (Wang et al., 1996). Intriguingly, Jak3, but not other members ( Jak1, Jak2, and Tyk2) of the Janus kinase ( Jak) family of receptor-associated PTKs, which typically bind to cytoplasmic regions of cytokine receptors, constitutively associates with the cytoplasmic domain of CD40 (Hanissian and Geha, 1997). Moreover, following CD40 ligation, Jak3 becomes rapidly autophosphorylated, correlating with STAT3 nuclear translocation and tyrosine phosphorylation and also the tyrosine phosphorylation of STAT6, two members of the family of signal transducers and activators of transcription, or STATs (Hanissian and Geha, 1997; Karras et al., 1997). Both Jak3 binding to CD40 and to CD40-induced CD23, ICAM-1 (CD54), and LT-움 expression require a proline-rich region within the membrane-proximal portion of the CD40 cytoplasmic tail, suggesting that Jak3 is required for CD40-induced CD23, CD54, and LT-움 (Hanissian and Geha, 1997). It is possible, however, that

128

ANDREW CRAXTON et al.

other CD40-mediated transcriptional events also share a requirement for the same proline-rich domain. Thus, it will be important to unequivocally identify Jak3-regulated CD40-responsive target genes. D. MAP KINASE FAMILY MEMBERS IN CD40 SIGNALING In common with other TNFR members, including TNFR, CD95/Fas, NGFR, and CD30, CD40 receptor ligation under speciﬁc circumstances stimulates each of the three major mammalian MAPK families: the stressactivated protein kinases (SAPKs) or c-jun amino-terminal kinases ( JNKs), p38 MAPK, and the extracellular-regulated kinases (ERKs) (Sakata et al., 1995; Berberich et al., 1996; Kashiwada et al., 1996; Li et al., 1996; Sutherland et al., 1996; Salmon et al., 1997; Craxton et al., 1998; Purkerson and Parker, 1998). Whereas CD40 ligation appears to stimulate both JNK and p38 MAPK under both mitogenic and non mitogenic conditions, CD40induced ERK activation seems to occur only when the CD40 signal delivered is mitogenic (Purkerson and Parker, 1998). Additional studies are required to determine whether the coupling of CD40 to ERK activation also depends on the stage of B cell maturation. In contrast to BCR-mediated stimulation of speciﬁc MAPK subfamilies that have distinct PTK requirements ( Jiang et al., 1998), little information exists regarding whether and which speciﬁc PTKs are required for CD40induced MAPK activation. Unlike BCR-induced ERK activation, which is PKA sensitive, CD40-dependent ERK stimulation appears to be regulated via a PKA-insensitive pathway (Purkerson and Parker, 1998). The ability of the MEK inhibitor PD98059 to block CD40-mediated ERK2 activation indicates that MEKs are required for CD40-induced ERK2 stimulation (Li et al., 1996), although the identities of the speciﬁc MEKs involved remain to be shown. CD40-induced MEK1 activation does appear to require Ras activation (Gulbins et al., 1996). Because CD40 activates both MKK4 and MKK7, two dual-speciﬁcity MAPK kinases that stimulate JNK, it is likely that both kinases contribute to CD40-mediated JNK activation (Foltz et al., 1998). However, a role for MKK4/SEK1 in CD40-mediated signaling remains to be delineated; MKK4/SEK1⫺/⫺ mice display normal proliferative responses to either antiCD40 alone or a combination of anti-CD40 and IL-4 (Nishina et al., 1997). Though p38 MAPK activation by other physiological stimuli requires the upstream kinases MKK3, MKK6, and perhaps MKK4 (Derijard et al., 1995; Enslen et al., 1998), it is unclear which of these kinases is required for CD40-induced p38 MAPK activation. The possibility that the small GTPase Rac1 may also play a role in CD40-mediated activation of JNK/ p38 MAPK merits further study because CD40 ligation stimulates Rac1 activity, and also because the low molecular mass GTPases Rac and Cd-

B CELL SIGNAL TRANSDUCTION PATHWAYS

129

c42Hs are required for JNK activation in response to other stimuli (Gulbins et al., 1996; Minden et al., 1995). Although a role for JNK in CD40-mediated signaling remains to be established, the discovery of a highly speciﬁc, cell-permeable pharmacological p38 MAPK inhibitor (Lee et al., 1994) has revealed that p38 MAPK is essential for CD40-induced B cell proliferation (Craxton et al., 1998). It seems plausible that p38 MAPK mediates at least some of its effects on B cell proliferation through NF-␬B-dependent gene expression, because the p38 MAPK inhibitor SB203580 partially prevented expression of both a minimal NF-␬B reporter gene and CD40-induced CD54/ICAM-1 expression, which may represent one example of a NF-␬B-responsive gene (Craxton et al., 1998). In contrast, p38 MAPK does not seem to play a role in CD40-mediated rescue from BCR-induced apoptosis in WEHI-231 cells (Salmon et al., 1997). E. CD40

AND THE

PI3K PATHWAY

A number of studies have shown that engagement of CD40 leads to both stimulation of PI3K activity and tyrosine phosphorylation of the p85 regulatory subunit in both normal B cells and B cell lines (Ren et al., 1994; Aagaard-Tillery and Jelinek, 1996; Gulbins et al., 1996; Yanagihara et al., 1997). However, it is not yet clear which PI3K isozymes (움, 웁, or 웂) are activated in response to CD40 ligation. The use of two potent and selective pharmacological inhibitors of PI3K, wortmannin and a structurally unrelated compound, LY294002, has enabled the role of PI3K in CD40mediated signaling to be evaluated (Arcaro and Wymann, 1993; Yano et al., 1993; Vlahos et al., 1994). PI3K activation appears to be required for both CD40-driven B cell proliferation and B cell differentiation into Ig-secreting cells (Aagaard-Tillery and Jelinek, 1996). However, the speciﬁc biochemical targets of CD40-mediated PI3K activation remain poorly characterized. One putative target is the atypical, Ca2⫹-independent and phorbol ester-insensitive PKC isozyme, PKC-␨, which may contribute to NF-␬B activation (Yanagihara et al., 1997). Another potential target is the serine/threonine protein kinase PKB/Akt/RAC-PK. However, whether PKB is a target for CD40-induced PI3K activation is presently unclear. F. TNF RECEPTOR-ASSOCIATED FACTORS IN CD40 SIGNALING Utilization of the yeast two-hybrid system with the cytoplasmic region of CD40 as bait led to the identiﬁcation of four members of the TNF receptor-associated factor (TRAF) family that physically associate with the cytoplasmic domain of CD40: TRAF2, TRAF3 (also known as CD40bp, CRAF1, CAP1 and LAP1), TRAF5, and TRAF6 (Hu et al., 1994; Cheng et al., 1995; Sato et al., 1995b; Ishida et al., 1996a,b). These TRAF family

130

ANDREW CRAXTON et al.

members are the primary candidates for initial mediators of CD40 signaling, particularly because there is presently relatively little convincing evidence that other molecules interact directly with the cytoplasmic domain of CD40. The TRAF proteins have several distinct structural features. These include a C-terminal TRAF domain of approximately 230 amino acids, which can be further divided into a TRAF-N subdomain, resembling a putative 움-helical coiled-coil region, and a highly conserved TRAF-C subdomain (Cheng et al., 1995). Second, an N-terminal RING ﬁnger and a series of zinc ﬁngerlike motifs have also been found within DNA-binding proteins and other transcriptional regulators. Indeed, a TRAF2 deletion mutant containing only the amino-terminal RING and multiple zinc ﬁnger domains appears to have the potential to function as a transcriptional activator or coactivator (Min et al., 1998). Deletion analysis using CD40 cytoplasmic domain fusion proteins showed that TRAFs 2, 3, and 5 bind to a domain of the cytoplasmic tail that includes a critical threonine residue required for CD40-mediated arrest of growth (Inui et al., 1990; Hu et al., 1994; Cheng and Baltimore, 1996; Ishida et al., 1996a). In contrast, TRAF6 apparently binds to a more N-terminal portion of the CD40 cytoplasmic tail (Ishida et al., 1996b). Rather intriguingly, TRAFs 2, 5, and 6, but not TRAF3, appear to play a role in NF-␬B activation (Fig. 8) (Rothe et al., 1995; Ishida et al., 1996a,b). In addition, TRAF6 may play a limited role in CD40-dependent ERK activation because dominant negative TRAF6 partially blocks CD40induced ERK stimulation, albeit in nonlymphoid cells (Kashiwada et al., 1998). The role of TRAF3 in CD40 signaling is, however, unclear because both CD40-induced proliferation of splenic B cells and CD23 and CD80 up-regulation were normal in TRAF3-deﬁcient mice (Xu et al., 1996). Whether individual or multiple TRAF family members are upstream mediators of CD40-induced JNK or p38 MAPK activation remains to be determined. Another protein, TANK (TRAF family member associated NF-␬B activator), heterodimerizes with at least TRAFs 1, 2, and 3, and synergizes with TRAF2 in the activation of NF-␬B (Cheng and Baltimore, 1996). TRAF5 and TRAF6 also mediate association, at least in vitro, of CD40 with RIP2, an NF-␬B- and apoptosis-inducing protein kinase that shares some homology with the serine/threonine kinase RIP (receptor-interacting protein), which is recruited to both TNFR1 and Fas/CD95 receptors (McCarthy et al., 1998). In addition to a possible role in CD40-mediated NF-␬B activation, RIP2 is a candidate for perhaps mediating CD40-induced cell death in some B cell lines (see Section IV,B), although whether CD40 associates with RIP2 in vivo is currently unknown. The possibility that other proteins such as TRAF1 and the speciﬁc caspase inhibitors, cIAP1 and cIAP2 may

B CELL SIGNAL TRANSDUCTION PATHWAYS

131

FIG. 8. A potential model for proximal CD40 signaling events. Prior to CD40 ligation, TRAF molecules (TRAFs 2, 3, 5, and 6) are preassociated with the cytoplasmic domain of CD40. TRAF2 may also bind to cIAP1 and/or cIAP2, whereas RIP2 may bind potentially with TRAF5 and/or TRAF6. Following CD40 ligand binding and receptor trimerization, release of the TRAF molecules via an unknown mechanism, which may involve an inhibitory signal to reduce CD40-TRAF binding, facilitates, for example, TRAF2 association with TANK, leading to synergistic NF-␬B activation. Release of TRAF5 also induces NF-␬B activation whereas free TRAF6 either activates NF-␬B or leads to partial ERK activation. The role of TRAF3 in CD40 signaling is currently unknown.

also associate indirectly with the CD40 cytoplasmic domain via TRAF 2, 3, 5, or 6 and thereby regulate CD40 signaling should also be considered. Despite a detailed knowledge of the TRAF proteins and some of their speciﬁc roles in CD40-mediated signaling, the regulation of CD40–TRAF protein interactions following receptor aggregation remains controversial. One model proposes that in the absence of CD40L binding, TRAFs are constitutively associated with CD40 (Fig. 8). Then, following receptor aggregation in response to ligand binding, TRAFs are released locally, resulting in, for example, the activation of NF-␬B (Cheng and Baltimore, 1996). Indeed, the release of constitutively associated TRAF2 from CD40 has been observed following receptor engagement (Chaudhuri et al., 1997).

132

ANDREW CRAXTON et al.

In contrast, other investigators found that CD40 ligation recruits TRAF family members to the receptor complex (Kuhne´ et al., 1997). Such differences may reﬂect the use of different reagents (CD40L or anti-CD40) to aggregate CD40 receptors or, alternatively, cell line variations. Clearly, further studies are required to reconcile such opposing models. V. CD95/Fas-Mediated Signaling and BCR-Mediated Resistance to CD95/Fas-Induced Death

A. CD95 SIGNALING CD95 (Fas/APO-1) is a member of the new subfamily of TNFRs referred to as the ‘‘death receptors’’ that have a cytoplasmic death domain (Table II). Binding of their respective ligands or speciﬁc agonistic antibodies induces apoptosis in many cell types. Stimulation of CD95 results in aggregation of its death domains, promoting recruitment of two signaling proteins to form the death-inducing signal complex (DISC) (Kischkel et al., 1995), FADD/MORT-1 (Fas-associating protein with death domain) via its C-terminal death domain (Boldin et al., 1995; Chinnaiyan et al., 1995; Chinnaiyan and Dixit, 1996), and FLICE/caspase-8/MACH/Mch5 (FADD-like ICE) (Boldin et al., 1996; Fernandes-Alnemri et al., 1996; Muzio et al., 1996) through its two N-terminal death effector domains (DEDs). Caspase-8 is activated by binding to the DISC and results in release of the active p18 and p10 subunits into the cytoplasm, which leads to activation of other caspases such as caspase-1, caspase-3, caspase-6, and caspase-7. For an excellent review of the biochemistry of CD95/Fasmediated signaling, see Nagata (1997). B. BCR-MEDIATED RESISTANCE TO CD95-INDUCED DEATH Activated B lymphocytes that express CD95 may either be sensitive or resistant to cytotoxicity induced by CD4⫹ Th1 effector cells that induce apoptosis in a CD95/Fas-dependent manner, depending on the nature of the speciﬁc stimulus ( Ju et al., 1994; Daniel and Krammer, 1994; Rothstein et al., 1995). CD40 receptor engagement induces expression of cell surface CD95 and increases sensitivity to CD95-dependent killing (Rothstein et al., 1995; Garrone et al., 1995; Schattner et al., 1995). In contrast, BCR engagement actively protects B cells from Th1 cell-mediated cytotoxicity (Th1 CMC), because B cells stimulated with both CD40L and anti-IgM or speciﬁc antigen are resistant to Th1 CMC (Rothstein et al., 1995). In addition, B lymphocytes are also protected from CD95-mediated cell death by IL-4 receptor engagement (Foote et al., 1996a). Neither anti-IgM nor IL-4-induced resistance to CD95-dependent apoptosis is mediated by a reduction in CD95 expression, and neither type of resistance involves

B CELL SIGNAL TRANSDUCTION PATHWAYS

133

differences in the interaction between the target B cell and Th1 effector T cells (Foote et al., 1996a,b). Anti-IgM-induced CD95 resistance requires both PKC activation and a Ca2⫹ signal, whereas IL-4-mediated CD95 resistance is PKC-independent (Foote et al., 1996a,b). The ability of the transcription and protein synthesis inhibitors, actinomycin D and cycloheximide, respectively, to block either anti-IgM-mediated or a combination of PMA and ionomycin-mediated CD95 resistance suggests that inducible gene expression and de novo protein synthesis are required for protection to occur (Foote et al., 1996b). A strong molecular candidate for mediating at least some of these effects is the Bcl-2 homologue, Bcl-xL (Schneider et al., 1997). Stimulation with anti-IgM in either the absence or presence of CD40L, but not with CD40L alone, strongly up-regulated Bcl-xL expression at both transcriptional and translational levels (Schneider et al., 1997). Moreover, primary splenic B cells from Bcl-xL-overexpressing transgenic mice were fourfold less sensitive to CD95-mediated apoptosis compared to wild-type B cells from control mice (Schneider et al., 1997). However, it remains a distinct possibility that other gene products may also contribute to BCR-induced CD95 resistance. The recent development of high-throughput cDNA and oligonucleotide expression arrays should facilitate a rapid increase in our knowledge of the gene products that regulate inducible CD95 resistance in B lymphocytes in the future. BCR ligation may protect B cells from CD95-mediated death by inactivating key death caspases: Bras et al. (1997) found that BCR cross-linking on a mouse B cell line prevents the cleavage of caspase-1, which is normally activated in response to CD95/Fas ligation. A PKC inhibitor abolished the ability of the BCR to protect against death, suggesting that one or more PKC isoforms may function to inactivate the death caspase cascade. In this respect it is interesting that CD95-induced death of Jurkat T cells is blocked by expression of a constitutively active form of MKK1, which activates the MAPK family member, ERK2 (Holmstrom et al., 1998). Thus, a PKC-induced pathway may activate ERK2, which in turn may inhibit the caspase pathway. Tolerant B cells in particular may be susceptible to CD95-mediated killing by FasL⫹ T cells if their antigen receptors are not engaged or have defective signaling pathways (Rothstein et al., 1995; Rothstein, 1996; Rathmell et al., 1996). The interplay between CD40L and FasL signaling is ﬁnely tuned: both CD40L and FasL are required for the elimination of self-Ag-reactive B cells, CD40L to up-regulate CD95, and FasL to stimulate death via CD95 (Rathmell et al., 1996). However, CD95 and CD40L are also essential for optimal B cell proliferation and antibody responses to foreign Ag such as HEL; apparently, CD95 and CD40 function together

134

ANDREW CRAXTON et al.

to exert a positive effect on B cell maturation via a BCR-induced pathway, whereas in the case of tolerant B cells when the BCR is not stimulated, the CD40/CD95 pathway leads to cell death (Rathmell et al., 1996). Although it is not clear how CD40 ligation makes B cells more susceptible to CD95induced death, it is clear that this is not simply due to up-regulation of more CD95/Fas on the B cell. In the Ramos cell line, CD40 protects from BCR-induced death (Valentine and Licciardi, 1992) while at the same time increasing susceptibility to CD95-induced death (Lens et al., 1996), thus demonstrating biochemically the truth of the old adage, ‘‘One man’s pleasure is another man’s poison.’’ VI. General Comments and Concluding Remarks

Recent insights into the ability of coreceptors to regulate signals initiated by the BCR have answered several questions but have raised many more. Most apparent in our minds is the question about whether it is appropriate to refer to a molecule strictly as a ‘‘negative’’ or ‘‘positive’’ regulator. B cells are capable of responding to signals initiated by the BCR in amazingly plastic ways (Goodnow, 1996). Coreceptors functioning within strict blackand-white ‘‘rules’’ would reduce the ﬂexibility B cells need to respond to the various situations they might encounter. An excellent example of a ﬂexible receptor is Fc웂RIIB, which has long been thought of as strictly a ‘‘negative’’ regulator (Coggeshall, 1998). Yet Ono et al. (1997) clearly showed that another previously uncharacterized signaling domain exists within the cytoplasmic tail of Fc웂RIIB, which functions in opposition to the domain that recruits SHIP. Similarly, another receptor, the granulocyte colony-stimulating factor receptor (G-CSFR), contains at least two signaling domains, which regulate growth signals both positively and negatively and are required for the recruitment of PI3K and SHIP, respectively (Hunter and Avalos, 1998). In addition, our studies indicate that CD22 also has two functional cytoplasmic domains. Surface receptors with opposing signaling domains may be a common theme in regulatory molecules. In the case of Fc웂RIIB and perhaps CD22, the tendency to dampen signals initiated via the BCR may often predominate over a signal-enhancing function. In other words, the B cell may err on the side of caution. What would be the purpose of having two signaling domains that oppose each other? Perhaps the two domains might function independently of each other in speciﬁc situations. One could envision that engagement of soluble or cell-bound ligands may inﬂuence the coreceptor’s accessibility to and activation of kinases that phosphorylate speciﬁc domains within the cytoplasmic tail of the coreceptor and promote recruitment of appropriate

B CELL SIGNAL TRANSDUCTION PATHWAYS

135

signaling molecules. Furthermore, the involvement of positive or negative signaling domains may themselves be regulated by other coreceptors. The use of genetically altered animals and B cell lines has enabled biologists to address interesting questions in new and elegant ways. This seems to be a trend that will continue far into the future. Crossing genetargeted mice to generate mice heterozygous at multiple loci in order to address the polygenic nature of complex traits, such as establishing the threshold for B cell anergy (Cornall et al., 1998), is a novel approach in mammalian biology that is certain to be implemented more frequently in the future. Another innovative approach is the use of inducible genetargeting strains of mice (Lam et al., 1997). These mice are born with normal expression of the molecule of interest and a wild-type phenotype, but are capable of disruption/alteration of speciﬁc genes on demand. This approach will enable researchers to determine the in vivo function of molecules at any stage of maturation without the caveat of possible alterations of cellular phenotype at previous stages of development. Many of the key players that determine the fate of B lymphocytes have now been identiﬁed, although the rules of the ‘‘game’’ of cell life and cell death are but poorly understood. The reason BCR signaling induces cell death in one context and promotes survival in another is still as perplexing as a cricket match is to non-Commonwealth observers! ACKNOWLEDGMENTS We wish to thank Drs. C. L. Law, S. Sidorenko, and T. Yun for providing unpublished manuscripts and Dr. Aaron Marshall and Ms. Sasha Solow for providing unpublished results. We also greatly appreciate the assistance of Marj Domenowske for illustrations and Kate Elias for editorial assistance. This work was supported by NIH grants GM37905, RR00166, AI44250, and DE08229 and a Howard Hughes predoctoral fellowship to K.L.O.

REFERENCES Aagaard-Tillery, K. M., and Jelinek, D. F. (1996). J. Immunol. 156, 4543–4554. Ahmed, R., and Gray, D. (1996). Science 272, 54–60. Alderson, M. R., Armitage, R. J., Tough, T. W., Strockbine, L., Fanslow, W. C., and Spriggs, M. K. (1993). J. Exp. Med. 178, 669–674. Ales-Martinez, J. E., Warner, G. L., and Scott, D. W. (1988). Proc. Natl. Acad. Sci. U.S.A. 85, 6919–6923. Alessi, D. R., and Cohen, P. (1998). Curr. Opin. Genet. Dev. 8, 55–62. Alessi, D. R., Andjelkovic, M., Caudwell, B., Cron, P., Morrice, N., Cohen, P., and Hemmings, B. A. (1996). EMBO J. 15, 6541–6551. Allman, D. M., Ferguson, S. E., Lentz, V. M., and Cancro, M. P. (1993). J. Immunol. 151, 4431–4444. Alnemri, E. S. (1997). J. Cell. Biochem. 64, 33–42. Alnemri, E. S., Livingston, D. J., Nicholson, D. W., Salvesen, G., Thornberry, N. A., Wong, W. W., and Yuan, J. (1996). Cell (Cambridge, Mass.) 87, 171.

136

ANDREW CRAXTON et al.

Anderson, D. M., Maraskovsky, E., Billingsley, W. L., Dougall, W. C., Tometsko, M. E., Roux, E. R., Teepe, M. C., DuBose, R. F., Cosman, D., and Galibert, L. (1997). Nature (London) 390, 175–179. Andersson, K. B., Draves, K. E., Magaletti, D. M., Fujioka, S., Holmes, K. L., Law, C. L., and Clark, E. A. (1996). Eur. J. Immunol. 26, 3170–3178. Andjelic, S., and Liou, H. C. (1998). Eur. J. Immunol. 28, 570–581. Andjelkovic, M., Alessi, D. R., Meier, R., Fernandez, A., Lamb, N. J., Frech, M., Cron, P., Cohen, P., Lucocq, J. M., and Hemmings, B. A. (1997). J. Biol. Chem. 272, 31515–31524. Arcaro, A., and Wymann, M. P. (1993). Biochem. J. 296, 297–301. Armitage, R. J., Fanslow, W. C., Strockbine, L., Sato, T. A., Clifford, K. N., Macduff, B. M., Anderson, D. M., Gimpel, S. D., Davis-Smith, T., Maliszewski, C. R., Clark, E. A., Smith, C. A., Grabstein, K. H., Cosman, D., and Spriggs, M. K. (1992). Nature (London) 357, 80–82. Arpin, C., Dechanet, J., Van Kooten, C., Merville, P., Grouard, G., Briere, F., Banchereau, J., and Liu, Y. J. (1995). Science 268, 720–722. Arpin, C., Banchereau, J., and Liu, Y. J. (1997). J. Exp. Med. 186, 931–940. Arsura, M., Wu, M., and Sonenshein, G. E. (1996). Immunity 5, 31–40. Aversa, G., Chang, C. C., Carballido, J. M., Cocks, B. G., and de Vries, J. E. (1997). J. Immunol. 158, 4036–4044. Baker, S. J., and Reddy, E. P. (1996). Oncogene 12, 1–9. Banchereau, J., Bazan, F., Blanchard, D., Briere, F., Galizzi, J. P., van Kooten, C., Liu, Y. J., Rousset, F., and Saeland, S. (1994). Annu. Rev. Immunol. 12, 881–922. Beckwith, M., Fenton, R. G., Katona, I. M., and Longo, D. L. (1996a). Blood 87, 202–210. Beckwith, M., Jorgensen, G., and Longo, D. L. (1996b). Blood 88, 3502–3507. Bellacosa, A., Chan, T. O., Ahmed, N. N., Datta, K., Malstrom, S., Stokoe, D., McCormick, F., Feng, J., and Tsichlis, P. (1998). Oncogene 17, 313–325. Berberich, I., Shu, G. L., Siebelt, F., Woodgett, J. R., Kyriakis, J. M., and Clark, E. A. (1996). EMBO J. 15, 92–101. Bergamo, A., Bataille, R., and Pellat-Deceunynck, C. (1997). Br. J. Haematol. 97, 652–655. Bignon, J. S., and Siminovitch, K. A. (1994). Clin. Immunol. Immunopathol. 73, 168–179. Billian, G., Mondiere, P., Berard, M., Bella, C., and Defrance, T. (1997). Eur. J. Immunol. 27, 405–414. Bireland, M. L., and Monroe, J. G. (1997). Crit. Rev. Immunol. 17, 353–385. Birge, R. B., Knudsen, B. S., Besser, D., and Hanafusa, H. (1996). Genes Cells. 1, 595–613. Boldin, M. P., Varfolomeev, E. E., Pancer, Z., Mett, I. L., Camonis, J. H., and Wallach, D. (1995). J. Biol. Chem. 270, 7795–7798. Boldin, M. P., Goncharov, T. M., Goltsev, Y. V., and Wallach, D. (1996). Cell (Cambridge, Mass.) 85, 803–815. Bolen, J. B. (1995). Curr. Opin. Immunol. 7, 306–311. Bolland, S., Pearse, R. N., Kurosaki, T., and Ravetch, J. V. (1998). Immunity 8, 509–516. Boussiotis, V. A., Freeman, G. J., Berezovskaya, A., Barber, D. L., and Nadler, L. M. (1997). Science 278, 124–128. Bradbury, L., Kansas, G. S., Levy, S., Evans, R. L., and Tedder, T. F. (1992). J. Immunol. 149, 2841–2850. Braesch-Andersen, S., Paulie, S., Koho, H., Nika, H., Aspenstrom, P., and Perlmann, P. (1989). J. Immunol. 142, 562–567. Bras, A., Martinez, A. C., and Baixeras, E. (1997). J. Immunol. 159, 3168–3177. Brink, R., Goodnow, C. C., Crosbie, J., Adams, E., Eris, J., Mason, D. Y., Hartley, S. B., and Basten, A. (1992). J. Exp. Med. 176, 991–1005.

B CELL SIGNAL TRANSDUCTION PATHWAYS

137

Buday, L., Egan, S. E., Rodriguez-Viciana, P., Cantrell, D. A., and Downward, J. (1994). J. Biol. Chem. 269, 9019–9023. Buday, L., Khwaja, A., Sipeki, S., Farago, A., and Downward, J. (1996). J. Biol. Chem. 271, 6159–6163. Buhl, A. M., Pleiman, C. M., Rickert, R. C., and Cambier, J. C. (1997). J. Exp. Med. 186, 1897–1910. Burgering, B. M., and Coffer, P. J. (1995). Nature (London) 376, 599–602. Burrows, P. D., and Cooper, M. D. (1997). Curr. Opin. Immunol. 9, 239–244. Callard, R. E., Herbert, J., Smith, S. H., Armitage, R. J., and Costelloe, K. E. (1995). Int. Immunol. 7, 1809–1815. Cambier, J. C. (1995). J. Immunol. 155, 3281–3285. Cambier, J. C., Pleiman, C. M., and Clark, M. R. (1994). Annu. Rev. Immunol. 12, 457–486. Campbell, M. A., and Klinman, N. R. (1995). Eur. J. Immunol. 25, 1573–1579. Campbell, K. S., Hager, E. J., and Cambier, J. C. (1991). J. Immunol. 147, 1575–1580. Carman, J. A., Wechsler-Reya, R. J., and Monroe, J. G. (1996). J. Immunol. 156, 4562–4569. Carsetti, R., Kohler, G., and Lamers, M. C. (1995). J. Exp. Med. 181, 2129–2140. Carter, R. H., and Fearon, D. T. (1992). Science 256, 105–107. Carter, R. H., Doody, G. M., Bolen, J. B., and Fearon, D. T. (1997). J. Immunol. 158, 3062– 3069. Chalupny, N. J., Kanner, S. B., Schieven, G. L., Wee, S. F., Gilliland, L. K., Aruffo, A., and Ledbetter, J. A. (1993). EMBO J. 12, 2691–2696. Chalupny, N. J., Aruffo, A., Esselstyn, J. M., Chan, P. Y., Bajorath, J., Blake, J., Gilliland, L. K., Ledbetter, J. A., and Tepper, M. A. (1995). Eur. J. Immunol. 25, 2978–2984. Chan, V. W. F., Meng, F., Soriano, P., DeFranco, A. L., and Lowell, C. A. (1997). Immunity 7, 69–81. Chan, V. W. F., Lowell, C. A., and DeFranco, A. L. (1998). Curr. Biol. 8, 545–553. Chao, D. T., and Korsmeyer, S. J. (1998). Annu. Rev. Immunol. 16, 395–419. Chaudhuri, A., Orme, S., Eilam, S., and Cherayil, B. J. (1997). J. Immunol. 159, 4244–4251. Chen, Y. R., Wang, X., Templeton, D., Davis, R. J., and Tan, T. H. (1996). J. Biol. Chem. 271, 31929–31936. Cheng, G., and Baltimore, D. (1996). Genes Dev. 10, 963–973. Cheng, G., Cleary, A. M., Ye, Z. S., Hong, D. I., Lederman, S., and Baltimore, D. (1995). Science 267, 1494–1498. Chinnaiyan, A. M., and Dixit, V. M. (1996). Curr. Biol. 6, 555–562. Chinnaiyan, A. M., and Dixit, V. M. (1997). Semin. Immunol. 9, 69–76. Chinnaiyan, A. M., O’Rourke, K., Tewari, M., and Dixit, V. M. (1995). Cell (Cambridge, Mass.) 81, 505–512. Chinnaiyan, A. M., O’Rourke, K., Lane, B. R., and Dixit, V. M. (1997). Science 275, 1122– 1126. Choi, M. S., Boise, L. H., Gottschalk. A. R., Quintans, J., Thompson, C. B., and Klaus, G. G. (1995). Eur. J. Immunol. 25, 1352–1357. Choquet, D., Partiseti, M., Amigorena, S., Bonnerot, C., Fridman, W. H., and Korn, H. (1993). J. Cell. Biol. 121, 355–363. Chuang, P. I., Yee, E., Karsan, A., Winn, R. K., and Harlan, J. M. (1998). Biochem. Biophys. Res. Commun. 249, 361–365. Clark, E. A., and Ledbetter, J. A. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 4494–4498. Clark, E. A., and Ledbetter, J. A. (1994). Nature (London) 367, 425–428. Clark, E. A., and Shu, G. L. (1990). J. Immunol. 145, 1400–1406. Cleveland, J. L., and Ihle, J. N. (1995). Cell (Cambridge, Mass.) 81, 479–482.

138

ANDREW CRAXTON et al.

Cocks, B. G., Chang, C. C., Carballido, J. M., Yssel, H., de Vries, J. E., and Aversa, G. (1995). Nature (London) 376, 260–263. Coggeshall, K. M. (1998). Curr. Opin. Immunol. 10, 306–312. Cook, M. C., Basten, A., and Fazekas de St. Groth, B. (1997). J. Exp. Med. 186, 631–643. Cornall, R. J., Cyster, J. G., Hibbs, M. L., Dunn, A. R., Otipoby, K. L., Clark, E. A., and Goodnow, C. C. (1998). Immunity 8, 497–508. Craxton, A., Shu, G. L., Graves, J. D., Saklatvala, J., Krebs, E. G., and Clark, E. A. (1998). J. Immunol. 161, 3225–3236. Craxton, A., Jiang, A., Kurosaki, T., and Clark, E. A. (1999). Submitted. Crespo, P., Schuebel, K. E., Ostrom, A. A., Gutkind, J. S., and Bustelo, X. R. (1997). Nature (London) 385, 169–172. Cuende, E., Ale´s-Martı´nez, J. E., Ding, L., Go´nzalez-Garcı´a, M., Martı´nez, A. C., and Nunez, G. (1993). EMBO J. 12, 1555–1560. Cyster, J. G., and Goodnow, C. C. (1995). Immunity 2, 13–24. Cyster, J. G., and Goodnow, C. C. (1997). Immunity 6, 509–517. Cyster, J. G., Hartley, S. B., and Goodnow, C. C. (1994). Nature (London) 371, 389–395. D’Ambrosio, D., Hippen, K. L., Minskoff, S. A., Mellman, I., Pani, G., Siminovitch, K. A., and Cambier, J. C. (1995). Science 268, 293–297. D’Ambrosio, D., Hippen, K. L., and Cambier, J. C. (1996a). Eur. J. Immunol. 26, 1960–1965. D’Ambrosio, D., Fong, D. C., and Cambier, J. C. (1996b). Immunol. Lett. 54, 77–82. Damen, J. E., Liu, L., Rosten, P., Humphries, R. K., Jefferson, A. B., Majerus, P. W., and Krystal, G. (1996). Proc. Natl. Acad. Sci. U.S.A. 93, 1689–1693. Daniel, P. T., and Krammer, P. H. (1994). J. Immunol. 152, 5624–5632. Datta, K., Bellacosa, A., Chan, T. O., and Tsichlis, P. N. (1996). J. Biol. Chem. 271, 30835– 30839. Datta, S. R., Dudek, H., Tao, X., Masters, S., Fu, H., Gotoh, Y., and Greenberg, M. E. (1997). Cell (Cambridge, Mass.) 91, 231–241. Davidson, W. F., Morse, H. C., III, Sharrow, S. O., and Chused, T. M. (1979). J. Immunol. 122, 884–891. Deckert, M., Tartare-Deckert, S., Couture, C., Mustelin, T., and Altman, A. (1996). Immunity 5, 591–604. DeFranco, A. L. (1997). Curr. Opin. Immunol. 9, 296–308. Degli-Esposti, M. A., Smolak, P. J., Walczak, H., Waugh, J., Huang, C. P., DuBose, R. F., Goodwin, R. G., and Smith, C. A. (1997a). J. Exp. Med. 186, 1165–1170. Degli-Esposti, M. A., Dougall, W. C., Smolak, P. J., Waugh, J. Y., Smith, C. A., and Goodwin, R. G. (1997b). Immunity 7, 813–820. del Peso, L., Gonzalez-Garcia, M., Page, C., Herrera, R., and Nunez, G. (1997). Science 278, 687–689. Derijard, B., Raingeaud, J., Barrett, T., Wu, I. H., Han, J., Ulevitch, R. J., and Davis, R. J. (1995). Science 267, 682–685. Devereaux, Q. L., Takahashi, R., Salvesen, G. S., and Reed, J. C. (1997). Nature (London) 388, 300–304. Devereaux, Q. L., Roy, N., Stennicke, H. R., Van Arsdale, T., Zhou, Q., Srinivasula, S. M., Alnemri, E. S., Salvesen, G. S., and Reed, J. C. (1998). EMBO J. 17, 2215–2223. Dhein, J., Walczak, H., Baumler, C., Debatin, K. M., and Krammer, P. H. (1995). Nature (London) 373, 438–441. Dickinson, J. L., Bates, E. J., Ferrante, A., and Antalis, T. M. (1995). J. Biol. Chem. 270, 27894–27904. Doody, G. M., Justement, L. B., Delibrias, C. C., Matthews, R. J., Lin, J., Thomas, M. L., and Fearon, D. T. (1995). Science 269, 242–244.

B CELL SIGNAL TRANSDUCTION PATHWAYS

139

Dorken, B., Moldenhauer, G., Pezzutto, A., Schwartz, R., Feller, A., Kiesel, S., and Nadler, L. M. (1986). J. Immunol. 135, 4470–4479. Downward, J. (1998). Curr. Opin. Cell. Biol. 10, 262–267. Dudek, H., Datta, S. R., Franke, T. F., Birnbaum, M. J., Yao, R., Cooper, G. M., Segal, R. A., Kaplan, D. R., and Greenberg, M. E. (1997). Science 275, 661–665. Duronio, V., Scheid, M. P., and Ettinger, S. (1998). Cell. Signal. 10, 233–239. Dynlacht, B. D. (1997). Nature (London) 389, 149–152. Ellis, H. M., and Horvitz, H. R. (1986). Cell (Cambridge, Mass.) 44, 817–829. Enari, M., Sakahira, H., Yokoyama, H., Okawa, K., Iwamatsu, A., and Nagata, S. (1998). Nature (London) 391, 43–48. Engel, P., Zhou, L. J., Ord, D. C., Sato, S., Koller, B., and Tedder, T. F. (1995). Immunity 3, 39–50. Enslen, H., Raingeaud, J., and Davis, R. J. (1998). J. Biol. Chem. 273, 1741–1748. Erickson, L. D., Tygrett, L. T., Bhatia, S. K., Grabstein, K. H., and Waldschmidt, T. J. (1996). Int. Immunol. 8, 1121–1129. Ezhevsky, S. A., Toyoshima, H., Hunter, T., and Scott, D. W. (1996). Mol. Biol. Cell. 7, 553–564. Fang, W., Rivard, J. J., Ganser, J. A., LeBien, T. W., Nath, K. A., Mueller, D. L., and Behrens, T. W. (1995). J. Immunol. 155, 66–75. Fang, N., Motto, D. G., Ross, S. E., and Koretzky, G. A. (1996). J. Immunol. 157, 3769–3773. Faris, M., Gaskin, F., Parsons, J. T., and Fu, S. M. (1994). J. Exp. Med. 179, 1923–1931. Feng, G. S., Ouyang, Y. B., Hu, D. P., Shi, Z. Q., Gentz, R., and Ni, J. (1996). J. Biol. Chem. 271, 12129–12132. Fernandes-Alnemri, T., Litwack, G., and Alnemri, E. S. (1994). J. Biol. Chem. 269, 30761– 30764. Fernandes-Alnemri, T., Armstrong, R. C., Krebs, J., Srinivasula, S. M., Wang, L., Bullrich, F., Fritz, L. C., Trapani, J. A., Tomaselli, K. J., Litwack, G., and Alnemri, E. S. (1996). Proc. Natl. Acad. Sci. U.S.A. 93, 7464–7469. Fischer, G., Kent, S. C., Joseph, L., Green, D. R., and Scott, D. W. (1994). J. Exp. Med. 179, 221–228. Fischer, K. D., Zmuldzinas, A., Gardner, S., Barbacid, M., Bernstein, A., and Guidos, C. (1995). Nature (London) 374, 474–477. Fischer, K. D., Kong, Y. Y., Nishina, H., Tedford, K., Marengere, L. E., Kozieradzki, I., Sasaki, T., Starr, M., Chan, G., Gardener, S., Nghiem, M. P., Bouchard, D., Barbacid, M., Bernstein, A., and Penninger, J. M. (1998). Curr. Biol. 8, 554–562. Fisher, R. P. (1997). Curr. Opin. Genet. Dev. 7, 32–38. Foltz, I. N., Gerl, R. E., Wieler, J. S., Luckach, M., Salmon, R. A., and Schrader, J. W. (1998). J. Biol. Chem. 273, 9344–9351. Foote, L. C., Howard, R. G., Marshak-Rothstein, A., and Rothstein, T. L. (1996a). J. Immunol. 157, 2749–2753. Foote, L. C., Schneider, T. J., Fischer, G. M., Wang, J. K. M., Rasmussen, B., Campbell, K. A., Lynch, D. H., Ju, S., Marshak-Rothstein, A., and Rothstein, T. L. (1996b). J. Immunol. 157, 1878–1885. Franke, T. F., Yang, S. I., Chan, T. O., Datta, K., Kazlauskas, A., Morrison, D. K., Kaplan, D. R., and Tsichlis, P. N. (1995). Cell (Cambridge, Mass.) 81, 727–736. Franke, T. F., Kaplan, D. R., and Cantley, L. C. (1997a). Cell (Cambridge, Mass.) 88, 435–437. Franke, T. F., Kaplan, D. R., Cantley, L. C., and Tokor, A. (1997b). Science 275, 665–668. Fu, C., and Chan, A. C. (1997). J. Biol. Chem. 272, 27362–27368. Fu, C., Turck, C. W., Kurosaki, T., and Chan, A. C. (1998). Immunity 9, 93–103.

140

ANDREW CRAXTON et al.

Fukazawa, T., Reedquist, K. A., Panchamoorthy, G., Soltoff, S., Trub, T., Druker, B., Cantley, L., Shoelson, S. E., and Band, H. (1995). J. Biol. Chem. 270, 20177–20182. Fulcher, D. A., Lyons, A. B., Korn, S. L., Cook, M. C., Koleda, C., Parish, C., Fazekas de St. Groth, B., and Basten, A. (1996). J. Exp. Med. 183, 2313–2328. Funakoshi, S., Longo, D. L., Beckwith, M., Conley, D. K., Tsarfaty, G., Tsarfaty, I., Armitage, R. J., Fanslow, W. C., Spriggs, M. K., and Murphy, W. J. (1994). Blood 83, 2787–2794. Galibert, L., Burdin, N., de Saint Vis, B., Garrone, P., Van Kooten, C., Banchereau, J., and Rousset, F. (1996). J. Exp. Med. 183, 77–85. Galy, A. H., and Spits, H. (1992). J. Immunol. 149, 775–782. Gan, H., Newman, G. W., and Remold, H. G. (1995). J. Immunol. 155, 1304–1315. Garrone, P., Neidhardt, E. M., Garcia, E., Galibert, L., van Kooten, C., and Banchereau, J. (1995). J. Exp. Med. 182, 1265–1273. Gerber, H. P., Dixit, V., and Ferrara, N. (1998). J. Biol. Chem. 273, 13313–13316. Gold, M. R., and Aebersold, R. (1994). J. Immunol. 152, 42–50. Gold, M. R., Matsuuchi, L., Kelly, R. B., and DeFranco, A. L. (1991). Proc. Natl. Acad. Sci. U.S.A. 88, 3436–3440. Gold, M. R., Chan, V. W., Turck, C. W., and DeFranco, A. L. (1992a). J. Immunol. 148, 2012–2022. Gold, M. R., Sanghera, J. S., Stewart, J., and Pelech, S. L. (1992b). Biochem. J. 287, 269–276. Gong, S., and Nussenzweig, M. C. (1996). Science 272, 411–414. Goodnow, C. C. (1996). Proc. Natl. Acad. Sci. U.S.A. 93, 2264–2271. Goodnow, C. C., Crosbie, J., Adelstein, S., Lavoie, T. B., Smith-Gill, S. J., Brink, R. A., Pritchard-Briscoe, H., Wotherspoon, J. S., Loblay, R. H., Raphael, K., Trent, R. J., and Basten, A. (1988). Nature (London) 334, 676–682. Gotoh, T., Hattori, S., Nakamura, S., Kitayama, H., Noda, M., Takai, Y., Kaibuchi, K., Matsui, H., Hatase, O., Takahashi, H., et al. (1995). Mol. Cell. Biol. 15, 6746–6753. Gottschalk, A. R., McShan, C. L., Merino, R., Nun˜ez, G., and Quinta´ns, J. (1994). Int. Immunol. 6, 121–130. Graves, J. D., Draves, K. E., Craxton, A., Saklatvala, J., Krebs, E. G., and Clark, E. A. (1996). Proc. Natl. Acad. Sci. U.S.A. 93, 13814–13818. Graves, J. D., Draves, K. E., Craxton, A., Krebs, E. G., and Clark, E. A. (1998a). J. Immunol. 161, 168–174. Graves, J. D., Gotoh, Y., Draves, K. E., Ambrose, D., Han, D., Wright, M., Chernoff, J., Clark, E. A., and Krebs, E. G. (1998b). EMBO J. 17, 2224–2234. Grawunder, U., Leu, T. M., Schatz, D. G., Werner, A., Rolink, A. G., Melchers, F., and Winkler, T. H. (1995). Immunity 3, 601–608. Green, D. R. (1998). Cell (Cambridge, Mass.) 94, 695–698. Grumont, R. J., Rourke, I. J., O’Reilly, L. A., Strasser, A., Miyake, K., Sha, W., and Gerondakism, S. (1998). J. Exp. Med. 187, 663–674. Gruss, H. J., and Dower, S. K. (1995). Blood 85, 3378–3404. Gu, H., Tarlinton, D., Muller, W., Rajewsky, K., and Forster, I. (1991). J. Exp. Med. 173, 1357–1371. Gulbins, E., Brenner, B., Schlottmann, K., Koppenhoefer, U., Linderkamp, O., Coggeshall, K. M., and Lang, F. (1996). J. Immunol. 157, 2844–2850. Haggerty, H. G., Wechsler, R. J., Lentz, V. M., and Monroe, J. G. (1993). J. Immunol. 151, 4681–4693. Han, H., Nomura, T., Honjo, T., and Tsubata, T. (1996). Eur. J. Immunol. 26, 2425–2432. Han, D. K., Chaudhary, P. M., Wright, M. E., Friedman, C., Trask, B. J., Riedel, R. T., Baskin, D. G., Schwartz, S. M., and Hood, L. (1997). Proc. Natl. Acad. Sci. U.S.A. 94, 11333–11338.

B CELL SIGNAL TRANSDUCTION PATHWAYS

141

Hanissian, S. H., and Geha, R. S. (1997). Immunity 6, 379–387. Hardy, R. R., Carmack, C. E., Shinton, S. A., Kemp, J. D., and Hayakawa, K. (1991). J. Exp. Med. 173, 1213–1225. Harnett, M. M., Holman, M. J., and Klaus, G. G. (1989). Eur. J. Immunol. 19, 1933–1939. Hart, D. N., and McKenzie, J. L. (1988). J. Exp. Med. 168, 157–170. Hartley, S. B., Cooke, M. P., Fulcher, D. A., Harris, A. W., Cory, S., Basten, A., and Goodnow, C. C. (1993). Cell (Cambridge, Mass.) 72, 325–335. Harwood, A. E., and Cambier, J. C. (1993). J. Immunol. 151, 4513–4522. Hashimoto, A., Okada, H., Jiang, A., Kurosaki, M., Greenberg, S., Clark, E. A., and Kurosaki, T. (1998). J. Exp. Med. 188, 1287–1295. Healy, J. I., Dolmetsch, R. E., Timmerman, L. A., Cyster, J. G., Thomas, M. L., Crabtree, G. R., Lewis, R. S., and Goodnow, C. C. (1997). Immunity 6, 419–428. Heath, A. W., Chang, R., Harada, N., Santos-Argumedo, L., Gordon, J., Hannum, C., Campbell, D., Shanafelt, A. B., Clark, E. A., Torres, R., and Howard, M. (1993). Cell. Immunol. 152, 468–475. Hedgecock, E. M., Sulston, J. E., and Thomson, J. N. (1983). Science 220, 1277–1279. Hengartner, M. O., and Horvitz, H. R. (1994). Cell (Cambridge, Mass.) 76, 665–676. Hengartner, M. O., Ellis, R. E., and Horvitz, H. R. (1992). Nature (London) 356, 494–499. Hibbs, M. L., Tarlinton, D. M., Armes, J., Grail, D., Hodgson, G., Maglitto, R., Stacker, S. A., and Dunn, A. R. (1995). Cell (Cambridge, Mass.) 83, 301–311. Hibner, U., Benhamou, L. E., Haury, M., Cazenave, P. A., and Sarthou, P. (1993). Eur. J. Immunol. 23, 2821–2825. Hippen, K. L., Buhl, A. M., D’Ambrosio, D., Nakamura, K., Persin, C., and Cambier, J. C. (1997). Immunity 7, 49–58. Holmstrom, T. H., Chow, S. C., Elo, I., Coffey, E. T., Orrenius, S., Sistonen, L., and Eriksson, J. E. (1998). J. Immunol. 160, 2626–2636. Holsinger, L. J., Graef, I. A., Swat, W., Chi, T., Bautista, D. M., Davidson, L., Lewis, R. S., Alt, F. W., and Crabtree, G. R. (1998). Curr. Biol. 8, 563–572. Horvitz, H. R., Sternberg, P. W., Greenwald, I. S., Fixsen, W., and Ellis, H. M. (1983). Cold Spring Harb. Symp. Quant. Biol. 48, 453–463. Hu, H. M., O’Rourke, K., Boguski, M. S., and Dixit, V. M. (1994). J. Biol. Chem. 269, 30069– 30072. Hunter, M. G., and Avalos, B. R. (1998) J. Immunol. 160, 4979–4987. Illera, V. A., Perandones, C. E., Stunz, L. L., Mower, D. A., Jr., and Ashman, R. F. (1993). J. Immunol. 151, 2965–2973. Inaoki, M., Sato, S., Weintraub, B. C., Goodnow, C. C., and Tedder, T. F. (1997). J. Exp. Med. 186, 1923–1931. Inui, S., Kaisho, T., Kikutani, H., Stamenkovic, I., Seed, B., Clark, E. A., and Kishimoto, T. (1990). Eur. J. Immunol. 20, 1747–1753. Ishiai, M., Kurosaki, M., Pappu, R., Okawa, K., Ronko, I., Fu, C., Shibata, M., Iwamatsu, A., Chan, A. C., and Kurosaki, T. (1999). Immunity 10, 117–125. Ishida, T., Kobayashi, N., Tojo, T., Ishida, S., Yamamoto, T., and Inoue, J. I. (1995). J. Immunol. 155, 5527–5535. Ishida, T., Tojo, T., Aoki, T., Kobayashi, N., Ohishi, T., Watanabe, T., Yamamoto, T., and Inoue, J. (1996a). Proc. Natl. Acad. Sci. U.S.A. 93, 9437–9442. Ishida, T., Mizushima, S., Azuma, S., Kobayashi, N., Tojo, T., Suzuki, K., Aizawa, S., Watanabe, T., Mosialos, G., Kieff, E., Yamamoto, T., and Inoue, J. (1996b). J. Biol. Chem. 271, 28745–28748. Ishigami, T., Kim, K. M., Horiguchi, Y. Higaki, Y., Hata, D., Heike, T., Katamura, K., Mayumi, M., and Mikawa, H. (1992). J. Immunol. 148, 360–368.

142

ANDREW CRAXTON et al.

Jackman, M. R., and Pines, J. N. (1997). Cancer Surv. 29, 47–73. Jackman, J. K., Motto, D. G., Sun, Q., Tanemoto, M., Turck, C. W., Peltz, G. A., Koretzky, G. A., and Findell, P. R. (1995). J. Biol. Chem. 270, 7029–7032. Jacobson, B. A., Rothstein, T. L., and Marshak-Rothstein, A. (1996). Behring. Inst. Mitt. 97, 185–199. James, C., Gschmeissner, S., Fraser, A., and Evan, G. I. (1997). Curr. Biol. 7, 246–252. Jiang, A., Craxton, A., Kurosaki, T., and Clark, E. A. (1998). J. Exp. Med. 188, 1297–1306. Ju, S. T., Cui, H., Panka, D. J., Ettinger, R., and Marshak-Rothstein, A. (1994). Proc. Natl. Acad. Sci. U.S.A. 91, 4185–4189. Kamitani, T., Nguyen, H. P., and Yeh, E. T. H. (1997). J. Biol. Chem. 272, 14001–14004. Kamitani, T., Kito, K., Nguyen, H. P., Fukuda-Kamitani, T., and Yeh, E. T. (1998). J. Biol. Chem. 273, 11349–11353. Kantor, A. B., and Herzenberg, L. A. (1993). Annu. Rev. Immunol. 11, 501–538. Karras, J. G., Wang, Z., Huo, L., Frank, D. A., and Rothstein, T. L. (1997). J. Immunol. 159, 4350–4355. Karsan, A., Yee, E., Kaushansky, K., and Harlan J. M. (1996). Blood 87, 3089–3096. Kashiwada, M., Kaneko, Y., Yagita, H., Okumura, K., and Takemori, T. (1996). Eur. J. Immunol. 26, 1451–1458. Kashiwada, M., Shirakata, Y., Inoue, J., Nakano, H., Okazaki, K., Okomura, K., Yamamoto, T., Nagaoka, H., and Takemori, T. (1998). J. Exp. Med. 187, 237–244. Kavanaugh, W. M., Pot, D. A., Chin, S. M., Deuter-Reinhard, M., Jefferson, A. B., Norris, F. A., Masiarz, F. R., Cousens, L. S., Majerus, P. W., and Williams, L. T. (1996). Curr. Biol. 6, 438–445. Kawauchi, K., Lazarus, A. H., Sanghera, J. S., Man, G. L., Pelech, S. L., and Delovitch, T. L. (1996). Mol. Immunol. 33, 287–296. Khan, W. N., Alt, F. W., Gerstein, R. M., Malynn, B. A., Larsson, I., Rathbun, G., Davidson, L., Muller, S., Kantor, A. B., Herzenberg, L. A., Rosen, F. S., and Sideras, P. (1995). Immunity 3, 283–299. Kim, K. M., Ishigami, T., Hata, D., Higaki, Y., Morita, M., Yamaoka, K., Mayumi, M., and Mikawa, H. (1992). J. Immunol. 148, 29–34. Kim, K. M., Adachi, T., Nielsen, P. J., Terashima, M., Lamers, M. C., Kohler, G., and Reth, M. (1994). EMBO J. 13, 3793–3800. Kim, T. J., Kim, Y. T., and Pillai, S. (1995). J. Biol. Chem. 270, 27504–27509. Kischkel, F. C., Hellbardt, S., Behrmann, I., Germer, M., Pawlita, M., Krammer, P. H., and Peter, M. E. (1995). EMBO J. 14, 5579–5588. Koretzky, G. A. (1997). Immunol. Today 18, 401–406. Kozlowski, M., Mlinaric, R. I., Feng, G. S., Shen, R., Pawson, T., and Siminovitch, K. A. (1993). J. Exp. Med. 178, 2157–2163. Krajewska, M., Wang, H. G., Krajewski, S., Zapata, J. M., Shabaik, A., Gascoyne, R., and Reed, J. C. (1997). Cancer Res. 57, 1605–1613. Krajewski, S., Gascoyne, R. D., Zapata, J. M., Krajewska, M., Kitada, S., Chhanabhai, M., Horsman, D., Berean, K., Piro, L. D., Fugier Vivier, I., Liu, Y. J., Wang, H. G., and Reed, J. C. (1997). Blood 89, 3817–3825. Kuhn, R., Schwenk, F., Aguet, M., and Rajewsky, K. (1995). Science 269, 1427–1429. Kuhne´, M. R., Robbins, M., Hambor, J. E., Mackey, M. F., Kosaka, Y., Nishimura, T., Gigley, J. P., Noelle, R. J., and Calderhead, D. M. (1997). J. Exp. Med. 186, 337–342. Kulik, G., Klippel, A., and Weber, M. J. (1997). Mol. Cell. Biol. 17, 1595–606. Kumar, S., and Baglioni, C. (1991). J. Biol. Chem. 266, 20960–20964. Kurosaki, T. (1997). Curr. Opin. Immunol. 9, 309–318.

B CELL SIGNAL TRANSDUCTION PATHWAYS

143

Kuwahara, K., Kawai, T., Mitsuyoshi, S., Matsuo, Y., Kikuchi, H., Imajoh, O. S., Hashimoto, E., Inui, S., Cooper, M. D., and Sakaguchi, N. (1996). Int. Immunol. 8, 1273–1285. Lam, K. P., Kuhn, R., and Rajewsky, K. (1997). Cell (Cambridge, Mass.) 90, 1073–1083. Lankester, A. C., van-Schijndel, G. M., Rood, P. M., Verhoeven, A. J., and van Lier, R. A. (1994). Eur. J. Immunol. 24, 2818–2825. Law, C. L., Torres, R. M., Sundberg, H. A., Parkhouse, R. M. E., Brannan, C. I., Copeland, N. G., Jenkins, N. A., and Clark, E. A. (1993). J. Immunol. 151, 175–187. Law, C. L., Sidorenko, S. P., and Clark, E. A. (1994). Immunol. Today 15, 442–449. Law, C. L., Craxton, A., Otipoby, K. L., Sidorenko, S. P., Klaus, S. J., and Clark, E. A. (1996a). Immunol. Rev. 153, 123–154. Law, C. L., Sidorenko, S. P., Chandran, K. A., Zhao, Z., Shen, S. H., Fischer, E. H., and Clark, E. A. (1996b). J. Exp. Med. 183, 547–560. Law, C. L., Chandran, K. A., Sidorenko, S. P., and Clark, E. A. (1996c). Mol. Cell. Biol. 16, 1305–1315. Law, C. L., Ewings, M. K., Chaudhary, P. M., Solow, S. A., Yun, T. J., Marshall, A. J., Hood, L., and Clark, E. A. (1999). J. Exp. Med. 189, in press. Lazarus, A. H., Kawauchi, K., Rapoport, M. J., and Delovitch, T. L. (1993). J. Exp. Med. 178, 1765–1769. LeBien, T. W. (1998). Curr. Opin. Immunol. 10, 188–195. Lee, J. R., and Koretzky, G. A. (1998). J. Immunol. 161, 1637–1644. Lee, J. C., Laydon, J. T., McDonnell, P. C., Gallagher, T. F., Kumar, S., Green, D., McNulty, D., Blumenthal, M. J., Heys, J. R., Landvatter, S. W., et al. (1994). Nature (London) 372, 739–746. Lemmon, M. A., and Ferguson, K. M. (1998). Curr. Top. Microbiol. Immunol. 228, 39–74. Lens, S. M. A., Tesselaar, K., den Drijver, B. F. A, van Oers, M. H., and van Lier, R. A. W. (1996). J. Immunol. 156, 507–514. Lens, S. M. A., den Drijver, B. F. A., Potgens, A. J., Tesselaar, K., van Oers, M. H., and van Lier, R. A. W. (1998). J. Immunol. 160, 6083–6092. Leprince, C., Draves, K. E., Geahlen, R. L., Ledbetter, J. A., and Clark, E. A. (1993). Proc. Natl. Acad. Sci. U.S.A. 90, 3236–3240. Levine, R. A., McCormack, J. E., Buckler, A., and sonenshein, G. E. (1986). Mol. Cell. Biol. 6, 4112–4116. Li, T., Tsukada, S., Satterthwaite, A., Havlik, M. H., Park, H., Takatsu, K., and Witte, O. N. (1995). Immunity 2, 451–460. Li, Y. Y., Baccam, M., Waters, S. B., Pessin, J. E., Bishop, G. A., and Koretzky, G. A. (1996). J. Immunol. 157, 1440–1447. Lin, L. C., and Putnam, F. W. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 504–508. Lin, E. Y., Orlofsky, A., Berger, M. S., and Prystowsky, M. B. (1993). J. Immunol. 151, 1979– 1988. Liu, Y. J. (1997). J. Exp. Med. 186, 625–629. Liu, Y. J., and Arpin, C. (1997). Immunol. Rev. 156, 111–126. Liu, Y. J., Joshua, D. E., Williams, G. T., Smith, C. A., Gordon, J., and MacLennan, I. C. (1989). Nature (London) 342, 929–931. Liu, S. K., Fang, N., Koretzky, G. A., and McGlade, C. J. (1999). Curr. Biol. 9, 67–75. Lowenstein, E. J., Daly, R. J., Batzer, A. G., Li, W., Margolis, B., Lammers, R., Ullrich, A., Skolnik, E. Y., Bar, Sagi. D., and Schlessinger, J. (1992). Cell (Cambridge, Mass.) 70, 431–442. Lu, L., and Osmond, D. G. (1997). J. Immunol. 158, 5136–5145. Mach, F., Scho¨nbeck, U., Sukhova, G. K., Bourcier, T., Bonnefoy, J. Y., Pober, J. S., and Libby, P. (1997). Proc. Natl. Acad. Sci. U.S.A. 94, 931–1936.

144

ANDREW CRAXTON et al.

MacLennan, I. C. M. (1994). Annu. Rev. Immunol. 12, 117–139. MacLennan, I. C., Gulbranson-Judge, A., Toellner, K. M., Casamayor-Palleja, M., Chan, E., Sze, D. M., Luther, S. A., and Orbea, H. A. (1997). Immunol. Rev. 156, 53–66. Marshall, A. J., Niiro, H., Yun, T. J., and Clark, E. A. (1999). Submitted. Marshall-Clarke, S., Owen, G., and Tasker, L. (1996). Immunology 87, 624–632. Marte, B. M., and Downward, J. (1997). Trends Biochem. Sci. 22, 355–358. Mason, D. Y., Jones, M., Hardie, D. L., van Schijndel, G., Johnson, G. D., van Lier, R., and MacLennan, I. C. M. (1997). In ‘‘Leukocyte Typing VI’’ (T. Kishimoto, ed.), pp. 206–229. Garland Publ., New York. Matthews, R. J., Browne, D. B., Flores, E., and Thomas, M. L. (1992). Mol. Cell. Biol. 12, 2396–2405. Mauri, D. N., Ebner, R., Montgomery, R. I., Kochel, K. D., Cheung, T. C., Yu, G. L., Ruben, S., Murphy, M., Eisenberg, R. J., Cohen, G. H., Spear, P. G., and Ware, C. F. (1998). Immunity 8, 21–30. Mayer, B. J., and Gupta, R. (1998). Curr. Top. Microbiol. Immunol. 228, 1–22. McCarthy, J. V., Ni, J., and Dixit, V. M. (1998). J. Biol. Chem. 273, 16968–16975. Meier, R., Alessi, D. R., Cron, P., Andjelkovic, M., and Hemmings, B. A. (1997). J. Biol. Chem. 272, 30491–30497. Melamed, D., and Nemazee D. (1997). Proc. Natl. Acad. Sci. U.S.A. 94, 9267–9272. Melamed, D., Benschop, R. J., Cambier, J. C., and Nemazee, D. (1998). Cell (Cambridge, Mass.) 92, 173–182. Merino, R., Grillot, D. A. M., Simonian, P. L., Muthukkumar, S., Fanslow, W. C., Bondada, S., and Nu´n˜ez, G. (1995). J. Immunol. 155, 3830–3838. Mikhalap, S. V., Shlapatska, L. M., Berdova, A. G., Zhao, Z., Law, C. L., Clark, E. A., and Sidorenko, S. P. (1999). J. Immunol. 162, in press. Min, W., Bradley, J. R., Galbraith, J. J., Jones, S. J., Ledgerwood, E. C., and Pober, J. S. (1998). J. Immunol. 161, 319–324. Minden, A., Lin, A., Claret, F. X., Abo, A., and Karin, M. (1995). Cell (Cambridge, Mass.) 81, 1147–1157. Miyashita, T., McIlraith, M. J., Grammer, A. C., Miura, Y., Attrep, J. F., Shimaoka, Y., and Lipsky, P. E. (1997). J. Immunol. 158, 4620–4633. Mond, J. J., Seghal, E., Kung, J., and Finkelman, F. D. (1981). J. Immunol. 127, 881–888. Mongini, P. K. A., Blessinger, C., Posnett, D. N., and Rudich, S. M. (1989). J. Immunol. 143, 1565–1574. Monroe, J. G. (1998). Immunol. Res. 17, 155–162. Moreb, J. S., and Schweder, M. (1997). Leukemia 11, 998–1004. Murphy, W. J., Funakoshi, S., Beckwith, M., Rushing, S. E., Conley, D. K., Armitage, R. J., Fanslow, W. C., Rager, H. C., Taub, D. D., Ruscetti, F. W., et al. (1995). Blood 86, 1946–1953. Murray, A., and Hunt, T. (1993). ‘‘The Cell Cycle, an Introduction’’ (W. H. Freeman, ed.), p. 251. New York. Musci, M. A., Motto, D. G., Ross, S. E., Fang, N., and Koretzky, G. A. (1997). J. Immunol. 159, 1639–1647. Muta, T., Kurosaki, T., Misulovin, Z., Sanchez, M., Nussenzweig, M. C., and Ravetch, J. V. (1994). Nature (London) 368, 70–73. Muthuk Kumar, S., Venkataraman, C., Woods, T., and Bondada, S. (1997). Mol. Immunol. 34, 865–875. Muzio, M., Chinnaiyan, A. M., Kischkel, F. C., O’Rourke, K., Shevchenko, A., Ni, J., Scafﬁdi, C., Bretz, J. D., Zhang, M., Gentz, R., et al. (1996) Cell (Cambridge, Mass.) 85, 817–827.

B CELL SIGNAL TRANSDUCTION PATHWAYS

145

Nadler, M. J. S., Chen, B., Anderson, J. S., Wortis, H. H., and Neel, B. G. (1997). J. Biol. Chem. 272, 20038–20043. Nagai, K., Takata, M., Yamamura, H., and Kurosaki, T. (1995). J. Biol. Chem. 270, 6824– 6829. Nagata, S. (1997). Cell (Cambridge, Mass.) 88, 355–365. Nagata, K., Nakamura, T., Kitamura, F., Kuramochi, S., Taki, S., Campbell, K. S., and Karasuyama, H. (1997). Immunity 7, 559–570. Nakanishi, H., Brewer, K. A., and Exton, J. H. (1993). J. Biol. Chem. 268, 13–16. Nakayama, K., Nakayama, K., Dustin, L. B., and Loh, D. Y. (1995). J. Exp. Med. 182 1101– 1109. Nel, A. E., Gupta, S., Lee, L., Ledbetter, J. A., and Kanner, S. B. (1995). J. Biol. Chem. 270 18428–18436. Neuberger, M. S. (1997). Cell (Cambridge, Mass.) 90, 971–973. Newton, K., and Strasser, A. (1998). Curr. Opin. Genet. Dev. 8, 68–75. Newton, K., Harris, A. W., Bath, M. L., Smith, K. G. C., and Strasser, A. (1998). EMBO J. 17, 706–718. Nichols, K. E., Harkin, D. P., Levitz, S., Krainer, M., Kolquist, K. A., Genovese, C., Bernard, A., Ferguson, M., Zuo, L., Snyder, E., et al. (1998). Proc. Natl. Acad. Sci. U.S.A. 95, 13765–13769. Nishina, H., Bachmann, M., Oliveria-dos-Santos, A. J., Kozieradzki, I., Fischer, K. D., Odermatt, B., Wakeham, A., Shahinian, A., Takimoto, H., Bernstein, A., Mak, T. W., Woodgett, J. R., Ohashi, P. S., and Penninger, J. M. (1997). J. Exp. Med. 186, 941–953. Nishizumi, H., Taniuchi, I., Yamanashi, Y., Kitamura, D., Ilic, D., Mori, S., Watanabe, T., and Yamamoto, T. (1995). Immunity 3, 549–560. Nishizumi, H., Horikawa, K., Mlinaric-Rascan, I., and Yamamoto, T. (1998). J. Exp. Med. 187, 1343–1348. Nitschke, L., Kosco, M. H., Kohler, G., and Lamers, M. C. (1993). Proc. Nat. Acad. Sci. U.S.A. 90, 1887–1891. Nitschke, L., Carsetti, R., Ocker, B., Kohler, G., and Lamers, M. C. (1997). Curr. Biol. 7, 133–143. Nomura, T., Han, H., Howard, M. C., Yagita, H., Yakura, H., Honjo, T., and Tsubata, T. (1996). Int. Immunol. 8, 867–875. Norvell, A., Monroe, J. G. (1996). J. Immunol. 156, 1328–1332. Norvell, A., Mandik, L., and Monroe, J. G. (1995). J. Immunol. 154, 4404–4413. O’Keefe, T. L., Williams, G. T., Davies, S. L., and Neuberger, M. S. (1996). Science, 274, 798–801. Okura, T., Gong, L., Kamitani, T., Wada, T., Okura, I., Wei, C. F., Chang, H. M., and Yeh, E. T. (1996). J. Immunol. 157, 4277–4281. Ono, M., Bolland, S., Tempst, P., and Ravetch, J. V. (1996). Nature (London) 383, 263–266. Ono, M., Okada, H., Bolland, S., Yanagi, S., Kurosaki, T., and Ravetch, J. V. (1997). Cell (Cambridge, Mass.) 90, 293–301. O’Rourke, L., Tooze, R., and Fearon, D. T. (1997). Curr. Opin. Immunol. 9, 324–329. Osmond, D. G. (1991). Curr. Opin. Immunol. 3, 179–185. Otipoby, K. L., Andersson, K. B., Draves, K. E., Klaus, S. J., Farr, A. G., Kerner, J. D., Perlmutter, R. M., Law, C. L., and Clark, E. A. (1996). Nature (London) 384, 634–637. Otipoby, K. L., Draves, K. E., and Clark, E. A. (1999). Submitted. Padmore, L., Radda, G. K., and Knox, K. A. (1996). Int. Immunol. 8, 585–594. Palmer, R. H., Dekker, L. V., Woscholski, R., Le, G. J. A., Gigg, R., and Parker, P. J. (1995). J. Biol. Chem. 270, 22412–22416.

146

ANDREW CRAXTON et al.

Pan, G., O’Rourke, K., Chinnaiyan, A. M., Gentz, R., Ebner, R., Ni, J., and Dixit, V. M. (1997a). Science 276, 111–113. Pan, G., Ni, J., Wei, Y. F., Yu, G., Gentz, R., and Dixit, V. M. (1997b). Science 277, 815–818. Pan, G., O’Rourke, K., and Dixit, V. M. (1998). J. Biol. Chem. 273, 5841–5845. Panchamoorthy, G., Fukazawa, T., Miyake, S., Soltoff, S., Reedquist, K., Druker, B., Shoelson, S., Cantley, L., and Band, H. (1996). J. Biol. Chem. 271, 3187–3194. Pani, G., Kozlowski, M., Cambier, J. C., Mills, G. B., and Siminovitch, K. A. (1995). J. Exp. Med. 181, 2077–2084. Parry, S. L., Hasbold, J., Holman, M., and Klaus, G. G. (1994). J. Immunol. 152, 2821–2829. Pawson, A. J., (ed.) (1998). Curr. Topics. Micro. Biol. Immunol. 228, 1–368. Pawson, T., and Scott, J. D. (1997). Science 278, 2075–2080. Peaker, C. J. G., and Neuberger, M. S. (1993). Eur. J. Immunol., 23, 1358–1363. Pellat-Deceunynck, C., Amiot, M., Robillard, N., Wijdenes, J., and Bataille, R. (1996). Cancer Res. 56, 1909–1916. Pezzutto, A., Dorken, B., Moldenhauer, G., and Clark, E. A. (1987). J. Immunol. 138, 98–103. Pezzutto, A., Rabinovitch, P. S., Dorken, B., Moldenhauer, G., and Clark, E. A. (1988). J. Immunol. 140, 1791–1795. Phillips, N. E., and Parker, D. C. (1983). J. Immunol. 130, 602–606. Phillips, N. E., and Parker, D. C. (1984). J. Immunol., 132, 627–632. Pinchouk, V.G., Sidorenko, S. P., Gluzman, D. F., Vetrova, E. P., Berdova, A. G., and Schlapatskaya, L. N. (1988). Anticancer Res. 8, 1377–1380. Pinchuk, L. M., Klaus, S. J., Magaletti, D. M., Pinchuk, G. V., Norsen, J. P., and Clark, E. A. (1996). J. Immunol. 157, 4363–4370. Pitti, R. M., Marsters, S. A., Ruppert, S., Donahue, C. J., Moore, A., and Ashkenazi, A. (1996). J. Biol. Chem. 271, 12687–12690. Pleiman, C. M., Hertz, W. M., and Cambier, J. C. (1994). Science 263, 1609–1612. Plutzky, J., Neel, B. G., and Rosenberg, R. D. (1992). Proc. Natl. Acad. Sci. U.S.A. 89, 1123– 1127. Polacino, P. S., Pinchuk, L. M., Sidorenko, S. P., and Clark, E. A. (1996). J. Med. Primatol. 25, 201–209. Przylepa, J., Himes, C., and Kelsoe, G. (1998). Curr. Top. Microbiol. Immunol. 229, 85–104. Punnonen, J., Cocks, B. G., Carballido, J. M., Bennett, B., Peterson, D., Aversa, G., and de Vries, J. E. (1997). J. Exp. Med. 185, 993–1004. Purkerson, J. M., and Parker, D. C. (1998). J. Immunol. 160, 2121–2129. Raab, M., da Silva, A. J., Findell, P. R., and Rudd, C. E. (1997). Immunity 6, 155–164. Racila, E., Hsueh, R., Marches, R., Tucker, T. F., Krammer, P. H., Scheuermann, R. H., and Uhr, J. W. (1996). Proc. Natl. Acad. Sci. U.S.A. 93, 2165–2168. Raff, M. C. (1996). Cell (Cambridge, Mass.) 86, 173–175. Rameh, L. E., Arvidsson, A. K., Carraway, K. L., 3rd., Couvillon, A. D., Rathbun, G., Crompton, A., VanRenterghem, B., Czech, M. P., Ravichandran, K. S., Burakoff, S. J., Wang, D. S., Chen, C. S., and Cantley, L. C. (1997). J. Biol. Chem. 272, 22059–22066. Ranger, A. M., Hodge, M. R., Gravallese, E. M., Oukka, M., Davidson, L., Alt, F. W., de la Brousse, F. C., Hoey, T., Grusby, M., and Glimcher, L. H. (1998). Immunity 8, 125–134. Rao, L., and White, E. (1997). Curr. Opin. Genet. Dev. 7, 52–58. Rathmell, J. C., Townsend, S. E., Xu, J. C., Flavell, R. A., and Goodnow, C. C. (1996). Cell (Cambridge, Mass.) 87, 319–329. Reed, J. C. (1996). Behring Inst. Mitt. 97, 72–100. Reed, J. C. (1997). Nature (London) 387, 773–776. Reed, S. I. (1997). Cancer Surv. 29, 7–23.

B CELL SIGNAL TRANSDUCTION PATHWAYS

147

Reif, K., and Cantrell, D. A. (1998). Immunity 8, 395–401. Ren, C. L., Morio, T., Fu, S. M., and Geha, R. S. (1994). J. Exp. Med. 179, 673–680. Reth, M., and Wienands, J. (1997). Annu. Rev. Immunol. 15, 453–479. Richards, J. D., Gold, M. R., Hourihane, S. L., DeFranco, A. L., and Matsuuchi, L. (1996). J. Biol. Chem. 271, 6458–6466. Rickert, R. C., Rajewsky, K., and Roes, J. (1995). Nature (London) 376, 352–355. Rodriguez-Viciana, P., Warne, P. H., Dhand, R., Vanhaesebroeck, B., Gout, I., Fry, M. J., Waterﬁeld, M. D., and Downward, J. (1994). Nature (London) 370, 527–532. Rodriguez-Viciana, P., Warne, P. H., Vanhaesebroeck, B., Waterﬁeld, M. D., and Downward, J. (1996). EMBO J. 15, 2442–2451. Roes, J., and Rajewsky, K. (1993). J. Exp. Med. 177, 45–55. Rothe, M., Sarma, V., Dixit, V. M., and Goeddel, D. V. (1995). Science 269, 1424–1427. Rothstein, T. L. (1996). Curr. Opin. Immunol. 8, 362–371. Rothstein, T. L., Wang, J. K., Panka, D. J., Foote, L. C., Wang, Z., Stanger, B., Cui, H., Ju, S. T., and Marshak-Rothstein, A. (1995). Nature (London) 374, 163–165. Roy, Deveraux, Q. L., Takahashi, R., Salvesen, G. S., and Reed, J. C. (1997). EMBO J. 16, 6914–6925. Sakahira, H., Enari, M., and Nagata, S. (1998). Nature (London) 391, 96–99. Sakata, N., Patel, H. R., Terada, N., Aruffo, A., Johnson, G. L., and Gelfand, E. W. (1995). J. Biol. Chem. 270, 30823–30828. Salim, K., Bottomley, M. J., Querfurth, E., Zvelebil, M. J., Gout, I., Scaife, R., Margolis, R. L., Gigg, R., Smith, C. I., Driscoll, P. C., Waterﬁeld, M. D., and Panayotou, G. (1996). EMBO J. 15, 6241–6250. Salmon, R. A., Foltz, I. N., Young, P. R., and Schrader, J. W. (1997). J. Immunol. 159, 5309– 5317. Salvesen, G. S., and Dixit, V. M. (1997). Cell (Cambridge, Mass.) 91, 443–446. Sanna, M. G., Duckett, C. S., Richter, B. W. M., Thompson, C. B., and Ulevitch, R. J. (1998). Proc. Natl. Acad. Sci. U.S.A. 95, 6015–6020. Sato, S., Steeber, D. A., and Tedder, T. F. (1995a). Proc. Natl. Acad. Sci. U.S.A. 92, 11558– 11562. Sato, T., Irie, S., and Reed, J. C. (1995b). FEBS. Lett. 358, 113–118. Sato, S., Miller, A. S., Inaoki, M., Bock, C. B., Janson, P. J., Tang, M. L., and Tedder, T. F. (1996a). Immunity 5, 551–562. Sato, S., Ono, N., Steeber, D. A., Pisetsky, D. S., and Tedder, T. F. (1996b). J. Immunol. 157, 4371–4378. Sato, S., Steeber, D. A., Jansen, P. J., and Tedder, T. F. (1997). J. Immunol. 158, 4662–4669. Saxton, T. M., van, O. I., Bowtell, D., Aebersold, R., and Gold, M. R. (1994). J. Immunol. 153, 623–636. Sayos, J., Wu, C., Morra, M., Wang, N., Zhang, X., Allen, D., Van Schaik, S., Notarangelo, L., Geha, R., Roncarolo, M. G., Oettgen, H., De Vries, J. E., Aversa, G., and Terhorst, C. (1998). Nature (London) 395, 462–469. Schaeffer, H. J., Catling, A. D., Eblen, S. T., Collier, L. S., Krauss, A., and Weber, M. J. (1998). Science 281, 1668–1671. Scharenberg, A. M., El, H. O., Fruman, D. A., Beitz, L. O., Li, Z., Lin, S., Gout, I., Cantley, L. C., Rawlings, D. J., and Kinet, J. P. (1998). EMBO J. 17, 1961–1972. Schattner, E. J., Elkon, K. B., Yoo, D. H., Tumang, J., Krammer, P. H., Crow, M. K., and Friedman, S. M. (1995). J. Exp. Med. 182, 1557–1565. Schauer, S. L., Wang, Z., Sonenshein, G. E., and Rothstein T. L. (1996). J. Immunol. 157, 81–86. Scheid, M. P., and Duronio, V. (1998). Proc. Natl. Acad. Sci. U.S.A. 95, 7439–44.

148

ANDREW CRAXTON et al.

Scheuermann, R. H., and Uhr, J. W. (1995). J. Exp. Med. 182, 903–906. Scheuermann, R. H., Racila, E., Tucker, T., Yefenof, E., Street, N. E., Vitetta, E. S., Picker, L. J., and Uhr, J. W. (1994). Proc. Natl. Acad. Sci. U.S.A. 91, 4048–4052. Schmidt, K. N., Hsu, C. W., Grifﬁn, C. T., Goodnow, C. C., and Cyster, J. G. (1998). J. Exp. Med. 187, 929–937. Schneider, T. J., Grillot, D., Foote, L. C., Nunez, G. E., and Rothstein, T. L. (1997). J. Immunol. 159, 4834–4839. Scho¨nbeck, U., Mach, M., Bonnefoy, J. Y., Loppnow, H., Flad, H. D., and Libby, P. (1997). J. Biol. Chem. 272, 19569–19574. Schriever, F., Freedman, A. S., Freeman, G., Messner, E., Lee, G., Daley J., and Nadler, L. M. (1989). J. Exp. Med. 169, 2043–2058. Schulte, R. J., Campbell, M. A., Fischer, W. H., and Sefton, B. M. (1992). Science 258, 1001– 1004. Scott, D. W., Grdina, T., and Shi, Y. (1996). J. Immunol. 156, 2352–2356. Seemayer, T. A., Gross, T. G., Egeler, R. M., Pirruccello, S. J., Davis, J. R., Kelly, C. M., Okano, M., Lanyi, A., and Sumegi, J. (1995). Pediatr. Res. 38, 471–478. Shen, S. H., Bastien, L., Posner, B. I., and Chretien, P. (1991). Nature (London) 352, 736–739. Sheridan, J. P., Marsters, S. A., Pitti, R. M., Gurney, A., Skubatch, M., Baldwin, D., Ramakrishnan, L., Gray, C. L., Baker, K., Wood, W. I., Goddard, A. D., Godowski, P., and Ashkenazi, A. (1997). Science 277, 818–821. Sherr, C. J. (1996). Science 274, 1672–1677. Shu, H. B., Halpin, D. R., and Goeddel, D. V. (1997). Immunity 6, 751–763. Shultz, L. D. (1988). Curr. Top. Microbiol. Immunol. 137, 216–222. Shultz, L. D., and Green, M. C. (1976). J. Immunol. 116, 936–943. Shultz, L. D., Coman, D. R., Bailey, C. L., Beamer, W. G., and Sidman, C. L. (1984). Am. J. Pathol. 116, 179–192. Shultz, L. D., Schweitzer, P. A., Rajan, T. V., Yi, T., Ihle, J. N., Matthews, R. J., Thomas, M. L., and Beier, D. R. (1993). Cell (Cambridge, Mass.) 73, 1445–1454. Sidman, C. L., Shultz, L. D., and Unanue, E. R. (1978a). J. Immunol. 121, 2392–2398. Sidman, C. L., Shultz, L. D., and Unanue, E. R. (1978b). J. Immunol. 121, 2399–2404. Sidman, C. L., Shultz, L. D., Hardy, R. R., Hayakawa, K., and Herzenberg, L. A. (1986). Science 232, 1423–1425. Sidman, C. L., Marshall, J. D., and Allen, R. D., (1989). Proc. Natl. Acad. Sci. U.S.A. 86, 6279–6282. Sidorenko, S. P. (1997). In ‘‘Leukocyte Typing VI’’ (T. Kishimoto, ed.) pp. 582–584. Garland Publ., New York. Sidorenko, S. P., and Clark, E. A. (1993). J. Immunol. 151, 4614–4624. Sidorenko, S. P., Vetrova, E. P., Yurchenko, O. V., Berdova, A. G., Shlapatskaya, L. N., and Gluzman, D. F. (1992). Neoplasma 39, 3–9. Sidorenko, S. P., Law, C. L., Chandran, K. A., and Clark, E. A. (1995). Proc. Natl. Acad. Sci. U.S.A. 92, 359–363. Sidorenko, S. P., Law, C. L., Klaus, S. J., Chandran, K. A., Takata, M., Kurosaki, T., and Clark, E. A. (1996). Immunity 5, 353–363. Siebelt, F., Berberich, I., Shu, G., Serﬂing, E., and Clark, E. A. (1997). Cell. Immunol. 181, 13–22. Sieh, M., Batzer, A., Schlessinger, J., and Weiss, A (1994). Mol. Cell. Biol. 14, 4435–4442. Sillman, A. L., and Monroe, J. G. (1995). J. Biol. Chem. 270, 11806–11811. Simonet, W. S., Lacey, D. L., Dunstan, C. R., Kelley, M., Chang, M. S., Luthy, R., Nguyen, H. Q., Wooden, S., Bennett, L., Boone, T., Shimamoto, G., DeRose, M., Elliott, R.,

B CELL SIGNAL TRANSDUCTION PATHWAYS

149

Colombero, A., Tan, H. L., Trail, G., Sullivan, J., Davy, E., Bucacy, N., Renshaw-Gegg, L., Hughes, T. M., Hill, D., Pattison, W., Campbell, P., Boyle, W. J., et al. (1997). Cell (Cambridge, Mass.) 89, 309–319. Smit, L., de Vries-Smits, A. M., Bos, J. L., and Borst, J. (1994). J. Biol. Chem. 269, 20209– 20212. Smit, L., van der Horst, G., and Borst, J. (1996a). J. Biol. Chem. 271, 8564–8569. Smit, L., van der Horst, G., and Borst, J. (1996b). Oncogene 13, 381–389. Smith, C. A., Farrah, T., and Goodwin, R. G. (1994). Cell (Cambridge, Mass.) 76, 959–962. Smith, K. G. C., Tarlinton, D. M., Doody, G. M., Hibbs, M. L., and Fearon, D. T. (1998). J. Exp. Med.187, 807–811. Sonenshein, G. E. (1997). J. Immunol. 158, 1994–1997. Spector, M. S., Desnoyers, S., Hoeppner, D. J., and Hengartner, M. O. (1997). Nature (London) 385, 653–656. Sprent, J. (1994). Curr. Opin. Immunol. 5, 433–438. Sprent, J. (1997). Curr. Opin. Immunol. 9, 371–379. Spurkland, A., Brinchmann, J. E., Markussen, G., Pedeutour, F., Munthe, E., Lea, T., Vartdal, F., and Aasheim, H. C. (1998). J. Biol. Chem. 273, 4539–4546. Stamenkovic, I., and Seed, B. (1990). Nature (London) 345, 74–77. Stamenkovic, I., Clark, E. A., and Seed, B. (1989)., EMBO J. 8, 1403–1410. Stokoe, D., Stephens, L. R., Copeland, T., Gaffney, P. R., Reese, C. B., Painter, G. F., Holmes, A. B., McCormick, F., and Hawkins, P. T. (1997). Science 277, 567–570. Su, B., and Karin, M. (1996). Curr. Opin. Immunol. 8, 402–411. Sumimoto, S., Heike, T., Kanazashi, S., Shintaku, N., Jung, E. Y., Hata, D., Katamura, K., and Mayumi, M. (1994). J. Immunol. 153, 2488–2496. Sun, J., Bird, C. H., Sutton, V., McDonald, L., Coughlin, P. B., De-Jong, T. A., Trapani, J. A., and Bird, P. I. (1996). J. Biol. Chem. 271, 27802–27809. Sun, J., Ooms, L., Bird, C. H., Sutton, V. R., Trapani, J. A., and Bird, P. I. (1997). J. Biol. Chem. 272, 15434–15441. Sutherland, C. L., Heath, A. W., Pelech, S. L., Young, P. R., and Gold, M. R. (1996). J. Immunol. 157, 3381–3390. Tailor, P., Jascur, T., Williams, S., von Willebrand, M., Couture, C., and Mustelin, T. (1996). Eur. J. Biochem. 237, 736–742. Takai, T., Ono, T., Hikida, M., Ohmori, H., and Ravetch, J. V. (1996). Nature (London) 379, 346–349. Takata, M., and Kurosaki, T. (1996). J. Exp. Med. 184, 31–40. Takata, M., Sabe, H., Hata, A., Inazu, T., Homma, Y., Nukada, T., Yamamura, H., and Kurosaki, T. (1994). EMBO J. 13, 1341–1349. Takata, M., Homma, Y., and Kurosaki, T. (1995). J. Exp. Med. 182, 907–914. Tangye, S. G., Liu, Y. J., Aversa, G., Phillips, J. H., and de Vries, J. E. (1998). J. Exp. Med. 188, 1691–1703. Tarakhovsky, A., Turner, M., Schaal, S., Mee, P. J., Duddy, L. P., Rajewsky, K., and Tybulewicz, V. L. (1995). Nature (London) 374, 467–470. Tedder, T. F., Inaoki, M., and Sato, S. (1997a). Immunity 6, 107–118. Tedder, T. F., Tuscano, J., Sato, S., and Kerl, J. H. (1997b). Annu. Rev. Immunol. 15, 481–504. ten Boekel, E., Melchers, F., and Rolink, A. G. (1998). Immunity 8, 199–207. Terashima, M., Kim, K. M., Adachi, T., Nielsen, P. J., Reth, M., Kohler, G., and Lamers, M. C. (1994). EMBO J. 13, 3782–3792. Thomas, M. L. (1995). J. Exp. Med. 181, 1953–1956.

150

ANDREW CRAXTON et al.

Tisch, R., Roifman, C. M., and Hozumi, N. (1988). Proc. Nat. Acad. Sci. U.S.A. 85, 6914– 6918. Toker, A., and Cantley, L. C. (1997). Nature (London) 387, 673–676. Toker, A., Meyer, M., Reddy, K. K., Falck, J. R., Aneja, R., Aneja, S., Parra, A., Burns, D. J., Ballas, L. M., and Cantley, L. C. (1994). J. Biol. Chem. 269, 32358–32367. Tomayko, M. M., and Cancro, M. P. (1998). J. Immunol. 160, 107–111. Tooze, R. M., Doody, G. M., and Fearon, D. T. (1997). Immunity 7, 59–67. Tordai, A., Franklin, R. A., Patel, H., Gardner, A. M., Johnson, G. L., and Gelfand, E. W. (1994). J. Biol. Chem. 269, 7538–7543. Torres, R. M., Law, C. L., Santos-Argumendo, L., Kirkman, P. A., Grabstein, K., Parkhouse, R. M. E., and Clark, E. A. (1992). J. Immunol. 149, 2641–2649. Torres, R. M., Flaswinkel, H., Reth, M., and Rajewsky, K. (1996). Science 272, 1804–1808. Treisman, R. (1996). Curr. Opin. Cell. Biol. 8, 205–215. Tridandapani, S., Chacko, G. W., Van Brocklyn, J. R., and Coggeshall, K. M. (1997a). J. Immunol. 158, 1125–1132. Tridandapani, S., Kelley, T., Cooney, D., Pradhan, M., and Coggeshall, K. M. (1997b). Immunol. Today 18, 424–427. Trub, T., Frantz, J. D., Miyazaki, M., Band, H., and Shoelson, S. E. (1997). J. Biol. Chem. 272, 894–902. Tseng, J., Eisfelder, B. J., and Clark, M. R. (1997). Blood 89, 1513–1520. Tsubata, T., Wu, J., and Honjo, T. (1993). Nature (London) 364, 645–648. Tsubata, T., Murakami, M., and Honjo, T. (1994). Curr. Biol. 4, 8–17. Tsui, H. W., Siminovitch, K. A., de Souza, L., and Tsui, F. W. (1993). Nature Genet. 4, 124–129. Turner, M., Mee, P. J., Costello, P. S., Williams, O., Price, A. A., Duddy, L. P., Furlong, M. T., Geahlen, R. L., and Tybulewicz, V. L. (1995). Nature (London) 378, 298–302. Turner, M., Gulbranson-Judge, A., Quinn, M. E., Walters, A. E., MacLennan, I. C., and Tybulewicz, V. L. (1997). J. Exp. Med. 186, 2013–2021. Tuscano, J. M., Engel, P., Tedder, T. F., Agarwal, A., and Kerl, J. H. (1996a). Eur. J. Immunol. 26, 1246–1252. Tuscano, J. M., Engel, P., Tedder, T. F., and Kerl, J. H. (1996b). Blood 87, 4723–4730. Tuveson, D. A., Carter, R. H., Soltoff, S. P., and Fearon, D. T. (1993). Science 260, 986–989. Uckun, F. M., Schieven, G. L., Dibirdik, I., Chandan-Langlie, M., Tuel-Ahlgren, L., and Ledbetter, J. A. (1991). J. Biol. Chem. 266, 17478–17485. Uckun, F. M., Burkhardt, A. L., Jarvis, L., Jun, X., Stealey, B., Dibirbik, I., Myers, D. E., Tuel-Ahlgren, L., and Bolen, J. B. (1993). J. Biol. Chem. 268, 21172–21184. Unkeless, J. C., and Jin, J. (1997). Curr. Opin. Immunol. 9, 338–343. Uren, A. G., Pakusch, M., Hawkins, C. J., Puls, K. L., and Vaux, D. L. (1996). Proc. Natl. Acad. Sci. U.S.A. 93, 4974–4978. Valentine, M. A., and Licciardi, K. A. (1992). Eur. J. Immunol. 22, 3141–3148. Vanhaesebroeck, B., Leevers, S. J., Panayotou, G., and Waterﬁeld, M. D. (1997). Trends Biochem. Sci. 22, 267–272. van Kooten, C., and Banchereau, J. (1996). Adv. Immunol. 61, 1–77. van Kooten, C., and Banchereau, J. (1997). Curr. Opin. Immunol. 9, 330–337. van Noesel, C. J., Lankester, A. C., van Schijndel, G. M., and van Lier, R. A. (1993). Int. Immunol. 5, 699–705. Vaux, D. L., Weissman, I. L., and Kim, S. K. (1992). Science 258, 1955–1957. Venkitaraman, A. R., Williams, G. T., Dariavach, P., and Neuberger, M. S. (1991). Nature (London) 352, 777–781.

B CELL SIGNAL TRANSDUCTION PATHWAYS

151

Vercammen, D., Beyaert, R., Denecker, G., Goossens, V., Van Loo, G., Declercq, W., Grooten, J., Fiers, W., and Vandenabeele, P. (1998). J. Exp. Med. 187, 1477–1485. Verheij, M., Bose, R., Lin, X. H., Yao, B., Jarvis, W. D., Grant, S., Birrer, M. J., Szabo, E., Zon, L. I., Kyriakis, J. M., Haimovitz, F. A., Fuks, Z., and Kolesnick, R. N. (1996). Nature (London) 380, 75–79. Vitetta, E., Pure, E., Isakson, P., Buck, L., and Uhr, J. (1980). Immunol. Rev. 52, 211–231. Vlahos, C. J., Matter, W. F., Hui, K. Y., and Brown, R. F. (1994). J. Biol. Chem. 269, 5241– 5248. Walczak, H., Degli-Esposti, M. A., Johnson, R. S., Smolak, P. J., Waugh, J. Y., Boiani, N., Timour, M. S., Gerhart, M. J., Schooley, K. A., Smith, C. A., Goodwin, R. G., and Rauch, C. T. (1997). EMBO J. 16, 5386–5397. Wallach, D. (1997). Nature (London) 388, 123–126. Wang, L., Miura, M., Bergeron, L., Zhu, H., and Yuan, J. (1994). Cell (Cambridge, Mass.) 78, 739–750. Wang, Z., Karras, J. G., Howard, R. G., and Rothstein, T. L. (1995). J. Immunol. 155, 3722– 3725. Wang, J., Koizumi, T., and Watanabe, T. (1996). J. Exp. Med. 184, 831–838. Wange, R. L., and Samelson, L. E. (1996). Immunity 5, 197–205. Ware, C. F., VanArsdale, S., and VanArsdale, T. L. (1996). J. Cell. Biochem. 60, 47–55. Wasserman, R., Li, Y. S., and Hardy, R. R. (1997). J. Immunol. 158, 1133–1138. Weng, W. K., Jarvis, L., and LeBien, T. W. (1994). J. Biol. Chem. 269, 32514–32521. Weng, Q. P., Andrabi, K., Klippel, A., Kozlowski, M. T., Williams, L. T., and Avruch, J. (1995). Proc. Natl. Acad. Sci. U.S.A. 92, 5744–5748. Wheeler, K., Pound, J. D., Gordon, J., and Jefferis, R. (1993). Eur. J. Immunol. 23, 1165– 1168. White, E. (1996). Genes Dev. 10, 1–15. Whitmarsh, A. J., Cavanagh, L., Tournier, C., Yasuda, L., and Davis, R. J. (1998). Science 281, 1671–1674. Wiley, S. R., Schooley, K., Smolak, P. J., Din, W. S., Huang, C. P., Nicholl, J. K., Sutherland, G. R., Smith, T. D., Rauch, C., Smith, C. A., and Goodwin, R. (1995). Immunity 3, 673–682. Wilson, H. A., Greenblatt, D., Taylor, C. W., Putney, J. W., Tsien, R. Y., Finkelman, F. D., and Chused, T. M. (1987). J. Immunol. 138, 1712–1718. Wilson, G. L., Fox, C. H., Fauci, A. S., and Kehrl, J. H. (1991). J. Exp. Med. 173, 137–146. Wong, B. R., Rho, J., Arron, J., Robinson, E., Orlinick, J., Chao, M., Kalachikov, S., Cayani, E., Bartlett, F. S., 3rd, Frankel, W. N., Lee, S. Y., and Choi, Y. (1997a). J. Biol. Chem. 272, 25190–25194. Wong, B. R., Josien, R., Lee, S. Y., Sauter, B., Li, H. L., Steinman, R. M., and Choi, Y. (1997b). J. Exp. Med. 186, 2075–2080. Wu, J., Motto, D. G., Koretzky, G. A., and Weiss, A. (1996). Immunity 4, 593–602. Wu, D., Wallen, H. D., and Nunez, G. (1997a). Science 275, 1126–1129. Wu, J., Zhao, Q., Kurosaki, T., and Weiss, A. (1997b). J. Exp. Med. 185, 1877–1882. Wu, M., Bellas, R. E., Shen, J., and Sonenshein, G. E. (1998). J. Exp. Med. 187, 1671–1679. Xia, Z., Dickens, M., Raingeaud, J., Davis, R. J., and Greenberg, M. E. (1995). Science 270, 1326–1331. Xu, Y., Cheng, G., and Baltimore, D. (1996). Immunity 5, 407–415. Yamanashi, Y., Kakiuchi, T., Mizuguchi, J., Yamamoto, T., and Toyoshima, K. (1991). Science 251, 192–194. Yamanashi, Y., Fukui, Y., Wongsasant, B., Kinoshita, Y., Ichimori, Y., Toyoshima, K., and Yamamoto, T. (1992). Proc. Natl. Acad. Sci. U.S.A. 89, 1118–1122.

152

ANDREW CRAXTON et al.

Yanagihara, Y., Basaki, Y., Ikizawa, K., and Kajiwara, K. (1997). Cell. Immunol. 176, 66–74. Yang, E., and Korsmeyer, S. J. (1996). Blood 88, 386–401 Yang, W., and Desiderio, S. (1997). Proc. Natl. Acad. Sci. U.S.A. 94, 604–609. Yang, S. H., Whitmarsh, A. J., Davis, R. J., and Sharrocks, A. D. (1998). EMBO J. 17, 1740– 1749. Yano, H., Nakanishi, S., Kimura, K., Hanai, N., Saitoh, Y., Fukui, Y., Nonomura, Y., and Matsuda, Y. (1993). J. Biol. Chem. 268, 25846–25856. Yao, X. R., Flaswinkel, H., Reth, M., and Scott, D. W. (1995). J. Immunol. 155, 652–661. Ye, Z. S., and Baltimore, D. (1994). Proc. Natl. Acad. Sci. U.S.A. 91, 12629–12633. Yi, T. L., Cleveland, J. L., and Ihle, J. N. (1992). Mol. Cell. Biol. 12, 836–846. Yi, A. K., Chang, M., Peckham, D. W., Krieg, A. M., and Ashman, R. F. (1998). J. Immunol. 160, 5898–5906. Yoshida, H., Nishina, H., Takimoto, H., Marengere, L. E., Wakeham, A. C., Bouchard, D., Kong, Y. Y., Ohteki, T., Shahinian, A., Bachmann, M., Ohashi, P. S., Penninger, J. M., Crabtree, G. R., and Mak, T. W. (1998). Immunity 8, 115–124. Yuan, J., Shaham, S., Ledoux, S., Ellis, H. M., and Horvitz, H. R. (1993). Cell (Cambridge, Mass.) 75, 641–652. Yun, T., Chaudhary, P. M., Shu, G. L., Ewings, M., Schwartz, S. M., Hood, L., and Clark, E. A. (1998). J. Immunol. 161, 6113–6121. Zhang, J., Cado, D., Chen, A., Kabra, N. H., and Winoto, A. (1998a). Nature (London) 392, 296–300. Zhang, W., Sloan-Lancaster, J., Kitchen, J., Trible, R. P., and Samelson, L. E. (1998b). Cell (Cambridge, Mass.) 92, 83–92. Zhao, Z., Shen, S. H., and Fischer, E. H. (1993). Proc. Natl. Acad. Sci. U.S.A. 90, 4251–4255. Zhou, L. J., Smith, H. M., Waldschmidt, T. J., Schwarting, R., Daley, J., and Tedder, T. F. (1994). Mol. Cell. Biol. 14, 3884–3894. Zou, H., Henzel, W. J., Liu, X., Lutschg, A., and Wang, X. (1997). Cell (Cambridge, Mass.) 90, 405–413.

ADVANCES IN IMMUNOLOGY, VOL. 73

Oral Tolerance: Mechanisms and Therapeutic Applications ANA M. C. FARIA AND HOWARD L. WEINER Center for Neurologic Diseases, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts 02115

I. Introduction

A majority of the contacts with foreign antigenic materials occur at mucosal surfaces. The total area of the mucosae is many hundredfold larger than the area of the skin. The mucosa of the small intestine alone is estimated to be 300 m2 in humans (Moog, 1981). Quite different from the skin, which is covered by keratin and a stratiﬁed squamous epithelium, the gut surface is a thin unicellular layer of epithelial cells characterized by absorptive properties. The mucosal surface is constantly and physiologically exposed to a large variety of antigenic materials (Cook and Olson, 1979; Dahl et al., 1984; Hemmings, 1978; Husby et al., 1985a,b; Kilshaw and Cant, 1984). Approximately 1 ton of food proteins reaches the human intestine during the course of 1 year and 130 to 190 g of these proteins is absorbed daily in the gut (Brandtzaeg, 1998). Although most dietary antigens are degraded by the time they reach the small intestine, studies in humans and rodents demonstrate that some undegraded or partially degraded antigen is absorbed into systemic circulation (Bruce and Ferguson, 1986a; Husby et al., 1985a). The bacterial ﬂora in the small intestine is also an additional source of natural antigenic stimulation in the gut and the number of bacteria colonizing the colonic mucosa in humans can reach 1012 microorganisms/g of stool (Macfarlane and Macfarlane, 1997). The most striking characteristic of the gut mucosa is the abundance of the lymphoid tissue associated with it. There are 1012 lymphoid cells per meter of human small intestine (Mestecky, 1987), and according to van der Heijden and Mestecky, the number of immunoglobulin-secreting cells located in the murine and human gut exceeds by severalfold the number of immunoglobulin-secreting cells from all other lymphoid organs together (Mestecky, 1987; van der Heijden et al., 1987). Thus, the mucosal surface is the major immunologic organ. A. IMMUNOLOGICAL CONSEQUENCES OF ORAL ANTIGEN ADMINISTRATION Three major immunological consequences of oral antigen administration have been described: (1) a local (noninﬂammatory) immune response resulting in the production of secretory immunoglobulin A (IgA), (2) a 153

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

154

ANA M. C. FARIA AND HOWARD L. WEINER

systemic inﬂammatory response with the generation of serum-speciﬁc antibodies, and (3) a systemic state of hyporesponsiveness to the antigen, termed ‘‘oral tolerance.’’ The ﬁrst two outcomes are often referred to as ‘‘oral immunization,’’ although they represent distinct phenomena (Bartholomeusz et al., 1990; Challacombe et al., 1992; Hag et al., 1995). The local production of secretory IgA is the result of antigen-induced activation, proliferation, and IgA commitment of B cells in Peyer’s patches. After their activation and partial differentiation, these cells migrate selectively to other mucosal sites, including the gut lamina propria, where they become IgA-producing plasma cells. The induction of IgA production is dependent on T helper cells but other cells may play a role. In humans and mice, transforming growth factor 웁 (TGF-웁) and interleukin 10 (IL-10) in concert with IL-4 and IL-5 have been shown to promote a B cell switch to IgA and differentiation into IgA-producing cells (Defrance et al., 1992; Kunimoto et al., 1988; Stavenzer et al., 1984). The high frequency of cells secreting IL-4, IL-10, and TGF-웁 in Peyer’s patches in addition to the ability of gut epithelial cells to produce TGF-웁 (Strober and Ehrhardt, 1994) results in the preferential production of IgA at the mucosal surfaces. Coupled to a secretory component, IgA is present in large amounts in all mucosal ﬂuids. IgA cannot activate complement and the secretory IgA response is noninﬂammatory and restricted to the mucosal sites (Strober and Ehrhardt, 1994). However, secretory IgA antibodies may bind to antigens that have been previously sampled into Peyer’s patches and prevent their entry via the gut, a phenomenon described as ‘‘immune exclusion’’ (Challacombe and Tomasi, 1980; Swarbrick et al., 1979). One of the aims of oral vaccines is to induce local immunity to prevent gut colonization by mucosal pathogens and to prevent the absorption of antigens potentially involved in food allergy. Normal individuals have very low, if any, levels of secretory IgA antibodies against common food antigens (O’Mahony et al., 1991) and inducing secretory IgA antibodies by immunizing animals orally with conventional antigens is very difﬁcult (Lycke and Holmgren, 1986; Muller, 1994; van der Heijden et al., 1991). Moreover, a substantial degree of subcompartimentalization exists within the mucosal lymphoid tissue regarding both homing and ﬁnal differentiation of plasma cell precursors. Thus, oral administration of antigens may induce a strong antibody response in the small intestine and in some distant exocrine glands whereas they may be inefﬁcient at evoking IgA secretion in the distal segments of the colon. Similar differential stimulation in the bronchial-associated lymphoid tissue is reported for intranasal administration of antigen (Czerkinsky and Holmgren, 1995). Systemic immunization via the oral route is a rare event and it is often associated with hypersensitivity reactions to food proteins, or inﬂammatory

ORAL TOLERANCE: MECHANISMS AND THERAPY

155

responses to bacterial invasion (Donnenberg and Kaper, 1992; Kantele and Makela, 1991; Muller, 1994; Sampson and Metcalfe, 1992; Van de Velde et al., 1991). Some antigens, such as cholera toxin and syncytial respiratory virus, are able to evoke the simultaneous appearance of serum IgG and secretory IgA in mucosal sites (Elson and Ealding, 1984a,b; Peri et al., 1982). In rabbits, ingestion of bovine serum albumin results in speciﬁc systemic immunization (IgG) without secretory IgA production (Peri et al., 1982). Speciﬁc protocols for oral immunization have been described to generate systemic immunization (Bartholomeusz et al., 1990; Faria et al., 1993; Peri et al., 1982), further attesting that special conditions are necessary for eliciting a systemic inﬂammatory response via the gut. The most common consequence of oral administration of an antigen is the development of a state of hyporesponsiveness, or oral tolerance, to subsequent challenges with the same antigen. This suppression can be measured by a decrease in cell-mediated responses such as delayed-type hypersensitivity (DTH) reactions in vivo (Gautam and Battisto, 1985; Kay and Ferguson, 1989; Miller and Hanson, 1979; Titus and Chiller, 1981) and lymphocyte proliferation in vitro (Challacombe and Tomasi, 1980; Higgins and Weiner, 1988; Lider et al., 1989; Titus and Chiller, 1981) and in humoral responses such as speciﬁc IgE and IgG production (Hanson et al., 1977; Mowat and Ferguson, 1982; Ngan and Kind, 1978; Richman et al., 1978; Swarbrick et al., 1979; Vaz et al., 1977). In this review we discuss our current understanding of the immunological mechanisms associated with oral tolerance, and the application of oral tolerance to treat disease states. B. HISTORY OF ORAL TOLERANCE The ﬁrst report related to oral tolerance appears to be in 1829 when the French physician R. Dakin reported an inﬂammatory cutaneous affection provoked by poison oak and poison ivy that could now be interpreted as a delayed-type hypersensitivity reaction. According to Dakin, ‘‘some good meaning, mystical, marvelous physicians, or favoured ladies with knowledge inherent, say the bane will prove the best antidote, and hence advice the forbidden leaves to be eaten, both as preventive and cure to the external disease’’ (Dakin, 1829). Nonetheless, he found that the ingestion of the poisonous leaves resulted not in the prevention of the skin rash, but instead the appearance of inﬂammation described as a ‘‘prominent swelling and eruption around the verge of the anus.’’ The ﬁrst experimental approach to a mucosal form of immune suppression was reported by the Russian scientist Alexandre Besredka, who joined the Pasteur Institute in Paris in 1893; in 1909 Besredka reported that guinea

156

ANA M. C. FARIA AND HOWARD L. WEINER

pigs treated with milk by either oral or rectal route became refractory to an anaphylaxis reaction induced by intracerebral injection of milk (Besredska, 1909). Besredka was a pioneer in mucosal immunology. In 1919, he developed the concept of a local humoral response operating at the gut independent of systemic immunity (Besredka, 1919). Working with rabbits, he demonstrated that oral immunization with Shigella dysenteriae protected the animals against fatal dysentery, irrespective of their serum antibody titer. His early experiments along with his concept of mucosal immunity and the principle of oral vaccination are contained in a treatise he published in 1927 (Besredka, 1927). In most reviews, the American scientist H. G. Wells is usually quoted as the ﬁrst reference for oral tolerance experiments. In 1911 Wells published a large series of experiments in guinea pigs showing that anaphylaxis to hen’s egg as well as to vegetable proteins could be prevented by prior feeding of these proteins to the animals (Wells, 1911; Wells and Osborne, 1911). Moreover, he reported that the duration of antigen feeding is one of the important factors affecting anaphylaxis inhibition. In his experiments with hen’s egg, absorption of the hen’s egg protein at ﬁrst rendered guineapigs hypersensitive, whereas if the feeding was maintained for a long enough period the animals become refractory to subsequent challenge with the same protein. Wells was focused on the analysis of suitable proteins for triggering the anaphylaxis phenomenon and he attributed the hyporesponsiveness observed to a special property of the antigen, an inhibitory compound in the hen’s egg preparation. Merrill W. Chase was the ﬁrst to address directly oral tolerance as an immunological phenomenon. In 1946, he compared the effect of oral, parenteral, and cutaneous administration of dinitrochlorobenzene on subsequent cutaneous challenge with the chemical and demonstrated that feeding had a substantial blocking effect in the hypersensitivity reaction (Chase, 1946). Moreover, he established the speciﬁc nature of the inhibition and showed that the suppression was easily induced as a prevention for future immunization but worked poorly as a therapy for already sensitized animals. Desensitization by prior intravenous injection of the immunizing substance had been reported in 1930 by Marion Sulzberger and the speciﬁc suppression of an immunological reaction was designated as the Chase– Sulzberger phenomenon. In spite of its physiological importance and potential applications, the Chase–Sulzberger phenomenon remained a footnote in the medical literature until the 1970s. It came under study again after the publication of the clonal selection theory (Burnet, 1959) and the discovery of T lymphocytes in the 1960s in association with investigation of immunological tolerance to self-components. Orally administered antigens, primarily ovalbumin and

ORAL TOLERANCE: MECHANISMS AND THERAPY

157

sheep red blood cells, were studied by a number of investigators and the phenomenon gained attention, having been being renamed ‘‘oral tolerance’’ (Andre et al., 1975; Saklayen et al., 1983; Thomas and Parrott, 1974; Titus and Chiller, 1981; Vaz et al., 1977). In the 1980s, oral tolerance began to attract considerable attention when it was shown to be an effective means of inhibiting immune responses to antigens of immunopathological importance in animals, including type II collagen (Nagler-Anderson et al., 1986; Thompson and Staines, 1986a,b) and myelin basic protein (Bitar and Whitacre, 1988; Higgins and Weiner, 1988). In the 1990s oral tolerance was then applied for the treatment of human diseases (Weiner, 1997). In addition, the mechanisms of oral tolerance have provided important insight into the basic function and workings of the immune system (Weiner and Mayer, 1996). C. DEFINITION OF ORAL AND IMMUNOLOGIC TOLERANCE Oral tolerance has classically been deﬁned as the speciﬁc suppression of cellular and/or humoral immune responses to an antigen by prior administration of the antigen by the oral route. It presumably evolved to prevent hypersensitivity reactions to food proteins and bacterial antigens present in the mucosal ﬂora. Immunologic tolerance has often been deﬁned as a mechanism by which the immune system avoids pathologic autoreactivity against self and thus prevents autoimmune diseases. The term ‘‘tolerance’’ was ﬁrst used by Burnet (Burnet, 1959) and three assumptions are implicit in Burnet’s concept of tolerance: (1) the primary function of the immune system is to defend the organism against pathogens or, in a broader sense, against nonself materials; (2) in order to perform such a function, the major immunologic response is an inﬂammatory class of response; and (3) because the operation of the immune system is driven by its reactions to foreign pathogens, tolerance is a negative counterpart of the immune system accomplished by neonatal deletion of ‘‘forbidden clones.’’ With a better understanding of the immune system, it is now clear that tolerance is a much more complicated and diverse process. Autoreactive cells, such as those reacting with brain antigens, thyroglobulin, serum albumin, collagen, and other autoantigens, are, in fact, not deleted and are present in all individuals (Avrameas, 1991; Zhang et al., 1993). They not only remain harmless under normal conditions, but cells autoreactive with self may have an important function in maintaining tissue homeostasis (Moalem et al., 1999) and may be differentially focused, depending on the tissue and the autoantigen (Cohen and Young, 1991). Furthermore, the basis of immunologic tolerance does not appear to simply be distinguishing between self and nonself, but reacting to danger signals that confront the

158

ANA M. C. FARIA AND HOWARD L. WEINER

immune system (Matzinger, 1994). Thus, immunological tolerance cannot rely solely on neonatal deletional events, but requires an active process that functions during the entire life of the organism. Tolerance has been deﬁned as a lack of response to self but a more appropriate deﬁnition of tolerance is ‘‘any mechanism by which a potentially injurious immune response is prevented, suppressed, or shifted to a noninjurious class of immune response.’’ The mechanisms by which immune tolerance is achieved are listed in Table I. Thus, tolerance is related to productive self-recognition rather than blindness of the immune system to its autocomponents. Oral tolerance, in this sense, is of unique immunologic importance because it is a continuous natural immunologic event driven by exogenous antigen. Due to their privileged access to the internal milieu, antigens that continuously contact the mucosa represent a frontier between foreign and self components. Thus, oral tolerance is an immunologic mechanism that evolved to treat external agents that gain access to the body via a natural route as internal components, which then become part of self. Given this, it would seem logical that autoimmune diseases caused by an inappropriate response to self-antigens might ultimately be treated by presenting such autoantigens to the mucosal surface, where they can be dealt with in a noninjurous (noninﬂammatory) immunologic environment. II. Mechanisms of Oral Tolerance

A. ORAL TOLERANCE AS AN ACTIVE IMMUNOLOGIC EVENT It is now clear that oral tolerance is an active immunologic process and is mediated by more than one mechanism. Low doses of antigen administration favor the induction of active cellular regulation whereas higher doses favor the induction of anergy or deletion. Although important principles regarding oral tolerance were described in the 1970s and 1980s, most of these early studies of oral tolerance did not distinguish dose effects. Thus, the feeding protocols (dose and frequency) used in the early studies TABLE I MECHANISMS OF IMMUNE TOLERANCE Ignorance: lack of critical mass of T cells or antigen Deletion: activation-induced cell death Anergy Receptor down-regulation Cellular regulation (suppression)

ORAL TOLERANCE: MECHANISMS AND THERAPY

159

must be carefully delineated when interpreting their results. In many respects the term ‘‘oral tolerance’’ is an inadequate immunologic term, because it has primarily been used to deﬁne the occurrence of systemic hyporesponsiveness when an animal is immunized after oral antigen administration. We now realize that oral tolerance is a complex process that involves suppression of some immune responses and induction of others. Thus, an understanding of oral tolerance and its use for the treatment of autoimmune or inﬂammatory diseases involve deﬁning the basic immunologic events that occur when antigen encounters the gut-associated lymphoid tissue (GALT). Antigen may act directly at the level of the GALT or have an effect following absorption (discussed below). In this regard, oral tolerance and mucosal immunization are part of one immunologic continuum and are ultimately explained in the context of how an antigenpresenting cell interacts with a T cell in the GALT and the factors that modulate and regulate this response. Thus, in addition to antigen dose, the nature of the antigen, the innate immune system, the genetic background and immunological status of the host, and mucosal adjuvants inﬂuence the immunologic outcome following oral antigen administration. B. THE INDUCTIVE PHASE OF ORAL TOLERANCE Orally administered antigens encounter the gut-associated lymphoid tissues, a well-developed immune network consisting of lymphoid nodules termed Peyer’s patches, villi containing epithelial cells, intraepithelial lymphocytes (IELs), and lymphocytes scattered throughout the lamina propria (Fig. 1). Although dietary antigens are degraded by the time they reach the small intestine, studies in humans and rodents have indicated that degradation is partial and that some intact antigen is absorbed, especially when large doses of antigen are fed (Bruce and Ferguson, 1986a,b, 1987). High-dose oral antigen may result in systemic antigen presentation, which induces hyporesponsiveness either via clonal T cell anergy or clonal deletion (see below). It is generally believed that in animals fed low-dose antigen, oral tolerance is induced in the gut-associated lymphoid tissue. Several cells capable of antigen presentation exist in the GALT. These include macrophages, dendritic cells, B cells, and epithelial cells. Dendritic cells (DCs) have been shown to be the major intestinal antigen-presenting cell (APC), which can acquire and process orally administered antigen (Liu and MacPherson, 1993). Epithelial cells may preferentially trigger the activation of CD8⫹ regulatory T cells. In the rat, these epithelial cell-induced CD8⫹ T cells are antigen speciﬁc (Bland and Warren, 1986), whereas in the human they were found to be antigen nonspeciﬁc (Mayer and Shlien, 1987). MHC class II-positive intestinal epithelial cells from 1-chloro-2,4-dinitrochloro-

160

ANA M. C. FARIA AND HOWARD L. WEINER

FIG. 1. Antigen and T cell trafﬁc in the gut-associated lymphoid tissue. Antigen is taken up either via M cells into lymphoid nodules termed Peyer’s patches or into the villus epithelium. Particulate antigen is preferentially taken up by M cells and soluble antigen by the villus epithelium. Antigen presentation results in the induction of cells that trafﬁc to the systemic circulation via the mesenteric lymph node and thoracic duct, and then migrate bac to the lamina propria and to other mucosal and nonmucosal sites. The villi contain intraepithelial lymphocytes (IELs), which are CD8⫹ T cells unique to the gut. T cells in the lamina propria are in a different state of activation compared to those in Peyer’s patches. Peyer’s patches also contain B cell-rich, poorly formed germinal centers where induction of antibody responses occurs.

benzene (DNCB)-fed mice could induce anergy of DNCB-primed T cells (Galliaerde et al., 1995). It has been shown that lamina propria cells (LPCs) may serve as antigen-presenting cells for oral tolerance (Harper et al., 1996). Presentation of Ag by LPCs stimulated high levels of interferon 웂 (IFN-웂) and TGF-웁, and adoptive transfer of Ag-pulsed LPCs induced oral tolerance to that antigen in the recipients (Harper et al., 1996). However, the type of APC responsible for the effect of LPCs was not resolved. T helper 2 (Th2) cells are preferentially generated in the GALT (Daynes et al., 1990; Xu-Amano et al., 1992), and their differentation depends on the cytokine microenvironment to which Th precursor cells are exposed during their activation (Abbas et al., 1996). If IL-12 is present during activation, Th1 cells differentiation occurs, whereas IL-4 induces Th2 cell differentiation. The intestinal mucosa has high basal levels of IL-4, IL-10, and TGF-웁 expression, and shortly after oral administration of antigen, their expression is up-regulated (Gonnella et al., 1998). This cytokine

ORAL TOLERANCE: MECHANISMS AND THERAPY

161

microenvironment may be crucial for the induction of Th2 or Th3 (TGF웁-secreting cells) in the gut. DCs from Peyer’s patches preferentially stimulate Th0 clones to produce large amounts of IL-4, whereas DCs from spleen induce high IFN-웂 production (Everson et al., 1997). There is also evidence that DCs may be involved in oral tolerance induction because the expansion of DCs in vivo with Flt3 ligand enhances oral tolerance (Viney et al., 1998). It is possible that dendritic cells, the most potent APCs in activating resting T cells, under the inﬂuence of the gut cytokine milieu, present antigen for Th2 or Th3 cell differentiation. Antigen-presenting cells provide costimulatory signals needed for T cell activation. B7.1 and B7.2 are the most important costimulatory molecules. B7.2 has been shown to be critical for Th2-type cell differentiation (Freeman et al., 1995). In vivo in the experimental autoimmune encephalomyelitis (EAE) model, injection of anti-B7.2 but not anti-B7.1 inhibited the induction of oral tolerance to low-dose (0.5 mg) but not high-dose myelin basic protein (MBP) (20 mg) (Liu et al., 1998b). On the other hand, we have also found that CTLA-4 stimulation is important for high-dose oral tolerance (Samoilova et al., 1998). CD40 ligand-CD40 interactions are also important for high dose oral tolerance (Kweon et al., 1999). Class II molecules on the APC also appear to be critical for the induction of oral tolerance in that oral tolerance cannot be induced in class II-deﬁcient mice (Desvignes et al., 1996). B cell-deﬁcient mice can be orally tolerized (R. Maron, unpublished observations) even though there is a defect in cytokine secretion in the gut of MT애/Mt애 mice (P. A. Gonnella, unpublished results). The mechanisms involved with oral tolerance in these mice have not yet been established. Of note is that studies from our laboratory suggest that B cells may be required for the induction of active TGF-웁 in splenocytes (Komagata et al., 1998). C. THE EFFECTOR PHASE OF ORAL TOLERANCE Based on ﬁndings in our laboratory and others it has become clear that there are two primary effector mechanisms of oral tolerance: the induction of regulatory T cells that mediate active suppression and the induction of clonal anergy or deletion. The primary factor that determines which form of peripheral tolerance develops following oral administration of antigen is the dose of antigen fed. Low doses of antigen favor the generation of active suppression or regulatory cell-driven tolerance whereas high doses of antigen favor anergy-driven tolerance (Fig. 2). Although these forms of oral tolerance are not mutually exclusive and may occur simultaneously, they are distinct, and the use of oral tolerance to treat autoimmune diseases is critically dependent on which of these two mechanisms is triggered.

162

ANA M. C. FARIA AND HOWARD L. WEINER

FIG. 2. The different mechanisms of oral tolerance are determined by the dose of fed antigen. Abbreviations: GALT, gut-associated lymphoid tissue; IL, interleukin; TGF-웁, transforming growth factor 웁; Th, T helper.

The delineation of these two mechanisms of oral tolerance was based on the following work: (1) investigations in our laboratory in which low doses of orally administered autoantigens were shown to suppress experimental autoimmune diseases via the generation of regulatory cells that suppressed in vitro and in vivo via the secretion of down-regulatory cytokines such as TGF-웁 (Miller et al., 1992b), (2) investigations from other laboratories demonstrating clonal anergy following oral administration of large doses of antigen with no evidence of active suppression (Melamed and Friedman, 1993a,b; Whitacre et al., 1991), (3) a large series of investigations demonstrating transferable suppression following oral tolerance (reviewed in Mowat, 1987), including work that showed two components of oral tolerance, one that was abrogated by treatment with low-dose cyclophoshamide and one that was not, a difference that was dose dependent (Mowat et al., 1982), and (4) direct comparison in our laboratory demonstrating that the two mechanisms depend on the dose (Friedman and Weiner, 1994). As shown in Figs. 1 and 2, low doses of antigen result in the generation of antigen-speciﬁc regulatory cells following presentation of antigen by gut-

ORAL TOLERANCE: MECHANISMS AND THERAPY

163

associated antigen-presenting cells. Such presentation preferentially induces regulatory cells, which, on subsequent recognition of antigen in vivo, or in vitro, secrete the suppressive cytokine TGF-웁. In addition, Th2 responses are preferentially generated in the gut, resulting in cells that secrete IL-4 and IL-10. These antigen-speciﬁc regulatory cells migrate to lymphoid organs, suppressing immune responses by inhibiting the generation of effector cells, and to the target organ, suppressing disease by releasing antigennonspeciﬁc cytokines (bystander suppression). Several factors can affect the generation of regulatory cells, including costimulation requirements, the cytokine milieu in which the immune response is generated, and differential epitope presentation, which preferentially may trigger regulatory cells. High doses of orally administered antigen result in anergy/deletion in the gut and in systemic antigen presentation after antigen passes through the gut and enters the systemic circulation either as intact protein or antigen fragments. High doses of antigen induce unresponsiveness of Th1 cell function, primarily via clonal anergy/deletion. The degree to which clonal anergy/deletion following high doses of antigen merely represents the direct passage of small amounts of antigen into the systemic or portal circulation or is dependent on ﬁltration by the gut is unknown. Why there is reduced active suppression with high doses of orally administered antigen is unclear, but could relate to anergizing cells involved in the generation of active suppression. In addition, it is not known the degree to which costimulatory requirements, cytokine milieu, and differential epitope recognition may preferentially favor the generation of anergy in Th1 cells. Anergy and deletion may not be the only mechanisms involved in highdose oral tolerance, because in some instances we have observed secretion of TGF-웁 by spleen cells of mice fed high doses of ovalbumin (OVA) (A. Faria and H.L. Weiner, unpublished results). Similarly, OVA T cell receptor (TCR) transgenic mice fed 500 mg OVA show a reduction of 10–20% of CD4⫹ T cells in spleen, thymus, and all other lymphoid organs examined, but TGF-웁-producing T cells are not deleted (Chen et al., 1995a). These results suggest that under certain circumstances TGF-웁 secretion may participate in high-dose oral tolerance induction. 1. Active Suppression Many early studies demonstrated that active suppression is an important mechanism for oral tolerance (Cowdery and Johlin, 1984; Gautam et al., 1990; MacDonald, 1982; Mattingly and Waksman, 1978; Miller and Hanson, 1979; Mowat, 1987; Richman et al., 1978; Strobel et al., 1983). After feeding antigens such as ovalbumin or sheep red blood cells, transferable suppression to cell-mediated immune responses was demonstrated using T cells from Peyer’s patches, mesenteric lymph node, and spleen as sources of cells for adoptive transfer experiments. Investigators have also reported

164

ANA M. C. FARIA AND HOWARD L. WEINER

initial sensitization prior to the appearance of suppression (Gautam et al., 1990). Further characterization of active suppression in these systems did not occur, which is most probably related to the difﬁculties in deﬁning the biology of suppression (Bloom et al., 1992; Sercarz and Krzych, 1991). Nonetheless, the demonstration of transferable cellular suppression associated with oral tolerance is a recurrent theme reported by many investigators (Mowat, 1987). Of note are the studies of Mowat (Mowat et al., 1982) demonstrating that high doses of ovalbumin induced tolerance not abrogated by cyclophosphamide and such tolerance affected antibody responses. Low doses of ovalbumin induced a state of tolerance that could be reversed by cyclophosphamide and primarily affected cell-mediated responses. Cyclophosphamide is believed to abrogate active suppression. It appears that these early studies were delineating components of active suppression vs. anergy, depending on dose. In addition, Hanson and Miller (1982) reported two components of oral tolerance following oral administration of ovalbumin. They found tolerance was observed both in cyclophosphamide-treated and -untreated animals, but they were unable to transfer tolerance from cyclophosphamide-treated animals. Our studies of oral tolerance in autoimmune models have found active suppression to be a primary mechanism and we have identiﬁed regulatory cells generated following oral tolerance that act via the secretion of antigennonspeciﬁc down-regulatory cytokines following triggering by the fed antigen (Weiner, 1997). Such cells were ﬁrst characterized in the Lewis rat model of EAE orally tolerized to low doses of guinea pig myelin basic protein. The regulatory cells identiﬁed in that model were CD8⫹ (Lider et al., 1989) and acted via the secretion of TGF-웁 following antigen-speciﬁc triggering (Miller et al., 1992b). They transfer suppression in vivo and can suppress in vitro. Additional studies have demonstrated that the epitopes of guinea pig MBP triggering CD8⫹ regulatory cells following orally administered MBP are different than the encephalitogenic determinant (Miller et al., 1993a). In addition, TGF-웁-secreting regulatory cells can be found in Peyer’s patches 24–48 hr after one feeding of low doses of MBP (Santos et al., 1994). Of note is that cells from Peyer’s patches removed after one feeding of MBP do not proliferate in response to 1 mg MBP even though they release TGF-웁 on in vitro stimulation. When similar studies were extended to a mouse system, it was found that not only CD8⫹ but also CD4⫹ cells were responsible for active suppression, both in vivo and in vitro (Chen et al., 1994, 1995b). Thus, when a low dose of MBP was administered orally to SJL/J mice, Th1 but not Th2 immune responses were suppressed. In fact, Th2 cytokines (IL-4, IL-10) and TGF-웁 were signiﬁcantly increased in mice fed with low doses of MBP. Furthermore, if animals are fed MBP and then immunized intraperi-

ORAL TOLERANCE: MECHANISMS AND THERAPY

165

toneally with the same antigen, one enhances the production of IL-4 and IL-10, as well as TGF-웁 (Chen et al., 1994). McGhee and colleagues have also found that exposure of soluble antigens to the gut preferentially generates Th2-type responses as judged by increased IL-4 and IL5 production (Xu-Amano et al., 1992). Thus, in the gut, immune responses to soluble antigen are preferentially of a Th2 type and involve the generation of cells that secrete TGF-웁 (Th3 cells). Why a Th2-type response is preferentially generated in the gut is unknown but may relate to antigen presentation or the cytokine milieu. When cells from mice fed and immunized with MBP were further studied in vitro, it was found that both CD4 and CD8 cells secreted TGF-웁, whereas only CD4 cells secreted IL-4 and IL-10 (Chen et al., 1994). This is consistent with our studies in the Lewis rat in which CD8 cell populations secreted TGF-웁. However, it was clear in the SJL model that a population of TGF-웁-secreting CD4⫹ cells was also generated and ampliﬁed in the gut following the feeding and subsequent immunization with MBP. After cloning regulatory CD4⫹ T cells from the mesenteric lymph nodes of MBP-fed mice, it was found that the majority of T cell clones produced active TGF-웁 in addition to varying amounts of one or the other of Th2type cytokines (IL-4 and IL-10). However, it appeared that the TGF-웁 clones were different from classic Th2-type cells, in that there was a general correlation between the secretion of IL-4 and IL-10 in an individual clone, whereas this was dissociated for TGF-웁 and IL-4/IL10. Mucosally derived CD4⫹ clones were further characterized for their epitope speciﬁcity, MHC restriction, and TCR usage. Sequence analysis of their cDNA revealed that they used either V움1 or V움3, and V웁4 or V웁17, all of which were also used by encephalitogenic Th1 cells. Most interestingly, one of these mucosal Th2 clones and one of the encephalitogenic Th1 clones used identical TCR V움 and V웁 chains. The regulatory T cell clones generated in this study were identical to encephalitogenic CD4⫹ Th1 cell clones regarding their speciﬁcity, TCR usage, and MHC restriction. However, they can be distinguished from the latter by the fact that they produce suppressive cytokines (TGF-웁, IL-4/IL-10) following antigen-speciﬁc activation. These clones inhibit the proliferation and cytokine production of MBP-speciﬁc Th1 cells, and they suppress the development of MBPinduced EAE and also proteolipid (PLP)-induced EAE, and this suppression was abrogated by in vivo injection of anti-TGF-웁 antibodies. This suggests these clones were able to mediate bystander suppression in vivo mediated by TGF-웁 production (Chen et al., 1994). The mechanism by which these regulatory cells are induced remains unknown. It is also not known the degree to which the generation of regulatory cells is related to unique antigen-presenting cells in the gut,

166

ANA M. C. FARIA AND HOWARD L. WEINER

the cytokine milieu, or other factors. The experiments reported above indicate that at least some of the regulatory cells generated in the gut are not dependent on unique T cell receptors, or MHC restriction, but are conventional T cells whose major difference relates to their cytokine proﬁle. The proﬁle appears stable and is uniquely generated depending on the environment in which they are induced. Daynes has reported that the lymphoid tissue microenvironment can determine the pattern of T cell responses in different lymphoid organs (Daynes et al., 1990). T cells in lymphoid organs drained by nonmucosal sites such as axillary or inguinal lymph nodes secreted IL-2 as the primary T cell growth factor after activation, whereas T cells from mucosal sites such as Peyer’s patches produced IL-4. Xu-Amano et al. (1993) have reported the selective induction of Th2 cells [IL-5-secreting cells as measured by enzyme-linked immunospot (ELI SPOT) assay] in murine Peyer’s patches following oral immunization with sheep red blood cells as compared to systemic immunization in which IFN-웂-producing cells predominated. On the other hand, IL-10 is known to stimulate Th2 cell differentiation (De Vries, 1995) and IL-4 and TGF-웁 were shown to be involved in the differentiation of TGF-웁-producing cells from naive CD4⫹ cells (Seder et al., 1998). Thus, the gut microenvironment, rich in TGF-웁 and Th2 cytokines, may function as a promoter of the preferential differentiation of T cells stimulated there. Furthermore, TGF웁 promotes Th2- and Th3-type immune deviation by altering accessory signals on antigen-presenting cells by impairing the ability of macrophages to produce IL-12 and express CD40 (Takeuchi et al., 1997, 1998). TGF-웁-secreting CD4⫹ cells were also cloned from MBP–TCR transgenic mice by culturing in the presence of IL-4 but not IL-2. These clones did not secrete IL-2, IFN-웂, IL-4, or IL-10 (Inobe et al., 1998). Thus, CD4⫹ cells that primarily produce TGF-웁 appear to be a unique T cell subset that includes mucosal helper T cell function and down-regulatory properties for Th1 and other immune cells, and thus have been termed Th3 cells. In contrast to Th1 and Th2 cells, Th3 cells provide help for IgA production and primarily secrete TGF-웁 (Table II). TGF-웁 can enhance the differentiation of Th3 cells, a process that can be enhanced by IL-10 or anti-IL-12 (Seder et al., 1998). Th3-type cells appear distinct from Th2 cells because CD4⫹ TGF-웁-secreting cells that suppress a form of colitis have been generated from IL-4-deﬁcient mice (Powrie et al., 1996). TGF웁-secreting regulatory cells have been reported to play a critical role in donor transfusion-induced allograft tolerance ( Josien et al., 1998) and TGF-웁 secreted by CD4⫹ cells and non-T cells may play an important role in tolerance induced by injecting soluble protein in the anterior chamber of the eye (Kosiewicz et al., 1998). Gene therapy with MBP-speciﬁc T cells engineered to express latent TGF-웁 suppresses EAE (Chen et al., 1998a).

167

ORAL TOLERANCE: MECHANISMS AND THERAPY

TABLE II CHARACTERISTICS OF T CELL SUBSETSa Cytokine Proﬁle IFN-웂 IL-4 TGF-웁 IL-10 Growth/differentiation factors Help Suppression

Th1

Th2

Th3

Tr1

⫹⫹⫹⫹ ⫺ ⫹/⫺ ⫺ IL-2

⫺ ⫹⫹⫹⫹ ⫹/⫺ ⫹⫹ IL-2/IL-4

⫹/⫺ ⫹/⫺ ⫹⫹⫹⫹ ⫹/⫺ IL-4/TGF-웁

⫹ ⫺ ⫹⫹ ⫹⫹⫹⫹ IL-10

DTH/IgG2a Th2

IgG1/IgE Th1

IgA Th1/Th2

? Th1

a Abbreviations: DTH, delayed-type hypersensitivity; IFN-웂, interferon 웂; IL, interluekin; TGF-웁, transforming growth factor 웁; Th, T helper; Tr, T regulatory.

It is of note that CTLA-4 negatively regulates T cell activation, and engagement of CTLA-4 induces TGF-웁 production by murine CD4⫹ cells (Chen et al., 1998c). Studies on rats have also demonstrated an essential role for TGF-웁 and IL-4 in the prevention of autoimmune thyroiditis by peripheral CD4⫹ CD45RC⫺ cells and CD4⫹ CD8⫺ thymocytes (Seddon and Mason, 1999). TGF-웁-secreting cells have been observed in multiple sclerosis (MS) patients following oral administration of myelin proteins (Fukaura et al., 1996). Another type of regulatory T cell is driven by IL-10 and secretes both IL-10 and TGF-웁 and has been termed a Tr1 cell (Groux et al., 1997). In some situations, especially in high antigen dose, oral feeding has also been reported to be capable of suppressing Th2 responses (Mowat et al., 1996; Russo et al., 1998). Feeding allergen could also suppress subsequent Th2-driven allergen-speciﬁc IgE and IgG1 responses (van Halteren et al., 1997). Data from our laboratory suggest that, although T cells can secrete latent TGF-웁, in some instances B cells may be required to activate the TGF-웁 secreted (Komagata et al., 1998). Of note is that highand low-dose oral tolerance can be induced in TGF-웁 null mice. In these animals, IL-4 and IL-10 are also decreased after low-dose feeding, indicating other mechanisms may be involved in the induction of tolerance (Barone et al., 1998). a. Bystander Suppression. Bystander suppression is a concept that regulatory cells induced by a fed antigen can suppress immune responses stimulated by different antigen, as long as the fed antigen is present in the anatomic vicinity (Miller et al., 1991). Bystander suppression was demonstrated in vitro when it was shown that cells from animals fed low doses of MBP could suppress proliferation of an ovalbumin line across a transwell

168

ANA M. C. FARIA AND HOWARD L. WEINER

(Miller et al., 1991). The cells from MBP-fed animals suppressed across the transwell only when triggered by the fed antigen. In an analogous fashion, cells from OVA-fed animals suppressed an MBP line across the transwell when stimulated with OVA. The soluble factor shown to be responsible for the suppression was TGF-웁. Bystander suppression was then demonstrated in vivo. Feeding ovalbumin has no effect on MBPinduced EAE in the Lewis rat. However, if animals are fed ovalbumin and then given aqueous ovalbumin in the footpad following immunization in the footpad with MBP/complete Freund’s adjuvant, EAE is suppressed. Suppression is mediated by OVA-speciﬁc regulatory cells that migrate to the draining lymph node and secrete TGF-웁 on encountering OVA, thus inhibiting the generation of the MBP-speciﬁc immune response being generated in the lymph node (Miller et al., 1991). Bystander suppression is speciﬁc to the fed antigen and is transferable. Further demonstration of bystander or tissue-speciﬁc suppression in vivo was obtained using MBP peptides (Miller et al., 1993a). In the Lewis rat, MBP peptide 21–40 is a nonencephalitogenic epitope whereas peptide 71–90 is the encephalitogenic epitope. Peptide 21–40 triggers TGF-웁 release following oral tolerization and orally administered peptide 21–40 suppresses peptide 71–90induced EAE in the Lewis rat. Furthermore, in peptide 71–90-immunized animals protected by oral administration of peptide 21–40, DTH responses in the ear to peptide 71–90 are not suppressed, whereas DTH responses to whole MBP are suppressed and suppression occurs because the peptide 21–40 epitope is present in whole MBP to trigger TGF-웁-secreting cells. Another example of tissue-speciﬁc bystander suppression is the suppression of PLP peptide-induced disease in the SJL mouse by feeding low doses of myelin basic protein or MBP peptides (al-Sabbagh et al., 1994). In addition to its role in EAE (Racke and Lovett-Racke, 1998), bystander suppression has also been demonstrated in several autoimmune and other models (Table III). Bystander suppression solves a major conceptual problem in the design of antigen- or T cell-speciﬁc therapy for inﬂammatory autoimmune diseases such as multiple sclerosis, type 1 diabetes, and rheumatoid arthritis (RA), in which the autoantigen is unknown or there are reactivities to multiple autoantigens in the target tissue. During the course of chronic inﬂammatory autoimmune processes in animals, there is intra- and interantigenic spread of autoreactivity at the target organ (Cross et al., 1993). In human autoimmune diseases, there are reactivities to multiple autoantigens in the target tissue. For example, in MS there is immune reactivity to at least three myelin antigens: MBP, PLP, and myelin oligodendrocyte glycoprotein (MOG) (Kerlero de Rosbo et al., 1993; Zhang et al., 1993). In type 1 diabetes, there are multiple islet cell antigens that could be the target of

ORAL TOLERANCE: MECHANISMS AND THERAPY

169

TABLE III MODELS OF AUTOIMMUNE AND OTHER DISEASES THAT DEMONSTRATE BYSTANDER SUPPRESSIONa Autoimmune Disease

Immunizing Antigen

Arthritis EAE EAE

BSA, mycobacteria PLP MBP peptide 71–90

EAE Diabetes IBD Stroke

MBP LCMV CD4⫹CD45RBhi None

Oral Antigen Type II collagen MBP MBP peptide 21–40 OVA Insulin OVA MBP

Target Organ Joint Brain Brain Lymph node, DTH response Pancreatic islets Intestine Brain

a Abbreviations: BSA, bovine serum albumin; DTH, delayed-type hypersensitivity; EAE, experimental allergic encephalomyelitis; LCMV, lymphocytic choriomeningitis virus; MBP, myelin basic protein; OVA, ovalbumin; PLP, proteolipid; IBD, inﬂammatory bowl disease.

autoreactivity, including glutamic acid decarboxylase (GAD), insulin, and heat-shock proteins (Harrison, 1992). Thus for a human organ-speciﬁc inﬂammatory disease, it is not necessary to know the speciﬁc antigen that is the target of an autoimmune response, but only to administer orally an antigen capable of inducing regulatory cells, which then migrate to the target tissue and suppress inﬂammation. Bystander suppression has been shown by IL-10-secreting Tr1 cells in which an OVA-speciﬁc Tr1 clone could suppress a murine model of inﬂammatory bowel disease in vivo when fed OVA (Groux et al., 1997). In arthritis, bystander suppression was demonstrated by feeding type II collagen in the antigen arthritis model (Yoshino, 1995), the adjuvant arthritis model (Zhang et al., 1990), and the streptococcal cell wall arthritis model (Chen et al., 1998b). Oral insulin suppresses lymphocytic choriomeningitis virus (LCMV)induced diabetes (von Herrath et al., 1996). Bystander suppression induced by nonself antigens have been demonstrated by Bloom in the EAE model (Falcone and Bloom, 1997) and by Pullerits (Pullerits et al., 1998) in the response to the hapten trimellitic anhydride (TMA), a cause of occupational asthma. In the later model, DTH responses but not antibody levels were suppressed by feeding and coimmunization with ovalbumin. Bystander suppression may be more difﬁcult to induce in uveitis models (Wildner and Thurau, 1995), where it may occur at the site of induction of autoimmune cells but not in the target organ. A contributing factor could by a difference in the ability of Th1 and Th2 cells to migrate into

170

ANA M. C. FARIA AND HOWARD L. WEINER

the target organ. Differential migration has been mentioned by Miossec and van den Berg (1997) as a potential limitation of the oral tolerance approach, speciﬁcally in reference to arthritis. Age may also play a role in bystander suppression. Feeding high doses of OVA has been reported to induce bystander suppression in adult rats, but anergy in young animals (Lundin et al., 1996). In theory, bystander suppression could be applied to the treatment of organ-speciﬁc inﬂammatory conditions that are not classic autoimmune diseases, such as psoriasis, or could be used to target antiinﬂammatory cytokines to an organ where inﬂammation may play a role in disease pathogenesis even if the disease is not primarily autoimmune in nature. For example, oral MBP decreased stroke size in a rat stroke model, presumably by decreasing inﬂammation associated with ischemic injury (Becker et al., 1997). Although bystander suppression was initially described in association with regulatory cells induced by oral antigen, the process could in principle be induced by any immune manipulation that induces Th2- or Th3-type regulatory cells. Bystander suppression mediated by TGF-웁 secretion was also reported in a mouse model of transplantation tolerance (Teng et al., 1998). A form of bystander suppression in oral tolerance was reported by Vaz and co-workers (Vaz et al., 1981), when they showed that BDF1 mice rendered tolerant to high doses of OVA by previous oral exposure do not respond to human gamma globulin (HGG) after intraperitoneal immunization with HGG and OVA in adjuvant. The same group has suggested other mechanisms for the cross-suppression observed in oral tolerance to unrelated proteins. Carvalho (Carvalho et al., 1994) described a model showing that the cross-suppressive effect obtained feeding high doses of antigen does not require the simultaneous injection of the two antigens and the effect is still present 72 hr after an injection of OVA in OVAtolerant mice, but does not occur if the unrelated protein is injected 24 hr before the tolerated protein. In addition, the two proteins can be injected by different routes (subcutaneous and intraperitoneal) and the cross-suppressive effects do not block secondary responses to unrelated proteins if the primary immunization is made in the absence of the tolerated protein. The authors suggest that the indirect effect they describe in their model is not mediated by cytokines but it is related to the idotypic connectivity of the immune system, where the two unrelated proteins presented together are incorporated into the same suppressive circuits (Carvalho et al., 1997). b. Role of 웂␦ T Cells. Studies have suggested that T cells bearing the 웂␦ TCR may also play an important regulatory role in speciﬁc mucosal responses and in oral tolerance. Aerosol-induced tolerance to OVA in rats

ORAL TOLERANCE: MECHANISMS AND THERAPY

171

can be transferred with very small numbers of splenic CD8⫹ 웂␦ T cells (McMenamin et al., 1991, 1994, 1995). They are able to suppress IgE but not IgG responses, apparently by the secretion of IFN-웂. Diabetes in nonobese diabetic (NOD) mice can also be prevented by CD8⫹ 웂␦ T cells induced by aerosol administration of insulin (Harrison et al., 1996). In mice it has been reported that oral tolerance can be abrogated by injection of anti-웂␦ monoclonal antibodies (Ke et al., 1997; Mengel et al., 1995). ␦ TCR knockout mice that lack 웂␦ T cells have impaired mucosal immunoglobulin A responses and they show a defective tolerance induction when fed a low dose of antigen (Ke et al., 1997; Kapp and Ke, 1997) but not a high dose (Spahn and Weiner, 1998). Moreover, investigators have demonstrated that 웂␦ T cells from fed mice are able to transfer tolerance to subsequent immunization (Mengel et al., 1995) and to prevent the induction of autoimmune disease (Wildner et al., 1996). The source of these 웂␦ T cells and their mechanism of suppression is still unclear. Transfer of systemic tolerance to proteins and allografts, induced either by feeding or intravenous injection of the antigen, have been reported with spleen (Mengel et al., 1995) as well as hepatic 웂␦ T cells (Gorczynski, 1994). On the other hand, it was reported that 웂␦ cells from the intraepithelial compartment abrogate oral tolerance when adoptively transferred (Fujihashi et al., 1992). These differing results may reﬂect differences in the populations of cells studied. IEL 웂␦ T cells are extrathymically derived and they show repertoire and functional differences compared to thymusderived spleen cells. Intestinal 웂␦ IELs appear to arrive from cryptopatches that are part of the murine intestinal immune compartment (Saito et al., 1998). The origin of the hepatic 웂␦ T cell population is still unknown but the cells show similarities to thymic-derived cells. Interestingly, Cardillo et al. (1998) have demonstrated that splenic 웂␦ T cells from euthymic young mice exhibit suppressor activity during experimental murine Trypanosoma cruzi infection. Splenic 웂␦ T cells from aged or athymic mice have a diminished (or completely lack) suppressor activity, suggesting that thymic emigrants are important for peripheral tolerance (Cardillo et al., 1998). c. Idiotypic and Other Regulatory Cells. The degree to which other forms of active suppression following oral tolerization are induced will depend on better characterization of mechanisms of active suppression (Bloom et al., 1992, Sercarz and Krzych, 1991). Some data have suggested that the 움 chain of the T cell receptor may be important in some forms of antigen-speciﬁc suppression (Kuchroo et al., 1991). In addition, anti-T cell receptor responses may actively suppress (Bloom et al., 1992; Sercarz and Krzych, 1991). Thus it is theoretically possible that orally administered

172

ANA M. C. FARIA AND HOWARD L. WEINER

antigen could lead to the generation of antiidiotypic regulatory cells speciﬁc for CD4⫹ cells. Interestingly, mouse strains such as NZB and NZW, which display abnormal patterns of idiotypic connectivity when young, are susceptible to autoimmune pathology and tend to be refractory to oral tolerance induction (Dighiero et al., 1987; Staples and Talal, 1969). Regulatory cells of the CD4⫹ CD25⫹ phenotype have been described and appear to act in a non-cytokine-dependent fashion (Suri-Payer et al., 1998). Their potential role in oral tolerance has not been examined. 2. Clonal Deletion Using OVA–TCR transgenic mice fed increasing doses of OVA (0.5–500 mg) (Chen et al., 1995a), we showed that depending on the dosage and the frequency of feeding, there was either a decrease or an increase in the percentage of CD4⫹ V웁8.2⫹ T cells in the Peyer’s patches of fed mice. In addition, in the Peyer’s patches of animals fed 500 mg OVA, there was a marked increase in the percentage of apoptotic cells, demonstrating that clonal deletion occurs after feeding a high dose of antigen. In mice fed 500 mg OVA, a reduction of 10–20% of CD4⫹ V웁8.2⫹ T cells in the spleen, thymus, and all other lymphoid tissues was observed. Feeding antigen to normal animals also seems to increase the susceptibility of their lymphocytes to die by apoptosis after systemic challenge with antigen in adjuvant. Mowat and co-workers showed that lymphocytes from these animals die rapidly when cultured in vitro in the absence of antigen (Mowat et al., 1996). Other investigators did not ﬁnd deletion in wild-type mice transferred with T cells from OVA–TCR transgenic mice when they fed the mice 25 mg OVA (Van Houten and Blake, 1996). Oral tolerance to high doses of ovalbumin is reported to be normal in fas-deﬁcient lpr mice (Miller et al., 1984; Mowat, 1998). Thus, clonal deletion occurs in transgenic mice fed a very high dose of antigen, but its role in high-dose tolerance in normal animals is unclear. A recent study in the rat EAE model showed that RNA expression of V웁 8.2⫹ TCR (expressed by encephalitogenic T cells) is decreased in lymph nodes draining the site of encephalitogenic challenge in high-dose orally tolerized rats. No change of V웁 8.2⫹ TCR RNA expression was detected in mesenteric lymph nodes. These results suggest that the mechanism of oral tolerance involved local clonal deletion or migration of pathogenic T cells (Goldman-Brezinski et al., 1998). 3. Anergy Anergy has also been proposed as a possible mechanism for oral tolerance (Melamed and Friedman, 1993; Whitacre et al., 1991). Anergy is deﬁned as a state of T lymphocyte unresponsiveness characterized by absence of proliferation and IL-2 production, and diminished expression of IL-2R

ORAL TOLERANCE: MECHANISMS AND THERAPY

173

(Schwartz, 1990). Anergy may be experimentally differentiated from clonal deletion by demonstrating the presence of antigen-speciﬁc TCR clonotypes, or by release from the anergic state, which is accomplished by preculture of cells in IL-2 (DeSilva et al., 1991). Delivery of antigenic signals without corresponding costimulatory signals induces anergy (Mueller et al., 1989). In certain experimental systems, anergy is also induced by partial or antagonistic signals delivered through the TCR/CD3 complex (Sloan-Lancaster et al., 1993, 1994). Bitar and Whitacre (1988) ﬁrst reported anergy as a mechanism for oral tolerance when they found diminished IL-2 and IFN-웂 production in rats orally fed 5 mg MBP in the presence of the soybean protease inhibitor; Melamed and Freedman (1993a) showed that a single feeding of 20 mg OVA induced a state of anergy in OVA-speciﬁc T lymphocytes: cells did not respond to OVA by proliferation, IL-2 production, or IL-2R expression, and the nonresponsive state was reversed by preculture of tolerized cells in IL-2. T cell clones derived from high-dose MBP-fed rats were characterized, and following several cell divisions in the presence of IL-2, they undergo a reversal of unresponsiveness ( Jewell et al., 1998). Anergy as a mechanism has also been suggested in a transfer system with OVA–TCR transgenic T cells (Van Houten and Blake, 1996). Clonal anergy has been demonstrated following the oral administration of staphylococcal enterotoxin B (Migita and Ochi, 1994). Since its proposal by Schwartz in 1989 (Schwartz et al., 1989), anergy has been used to explain tolerance induction in situations where evidence of active suppression could not be found. However, studies on the cells rendered anergic have raised the possibility that these cells do not function in a totally passive fashion in the tolerance they evoke. First, IL-2 production and proliferation are only partial measures of T cell activity because T cells can secrete a range of other potent cytokines without proliferative responses. Naturally activated self-reactive T cells isolated ex vivo from normal animals perform very efﬁciently in effector assays for helper or suppressor activities, and yet neither proliferates in vitro nor produces IL2 (Bandeira et al., 1987), although they are excellent producers of IL-4 (Ben-Sasson et al., 1990). Second, reports have suggested that anergic cells can actively suppress T cell responses either through modulation of the T cell-activating capacity of the APC (APC/T cell interaction) (Taams et al., 1998) or by inhibition of T cells recognizing their ligand in close proximity on the same APC (‘‘linked suppression’’ through T/T cell interactions) (Hoyne and Lamb, 1997). Mannie and co-workers (Mannie et al., 1996) suggest that anergic T cells can affect APC signaling to other T cells by inhibiting their function and inducing them to be anergic cells as well. These mechanisms may explain the genesis and spread of ‘‘infectious toler-

174

ANA M. C. FARIA AND HOWARD L. WEINER

ance’’ (Qin et al., 1993) via cell contact between anergic and other T cells. In all of the cases described above, the anergic cells serve as regulatory cells that mediate tolerance via an active mechanism. III. Immune Functions Affected by Oral Tolerance

A. CELLULAR VERSUS HUMORAL RESPONSES Virtually all manifestations of speciﬁc immune responsiveness tested can be suppressed by oral antigen administration. This includes in vivo responses such as formation of immunoglobulin of different isotypes (Ngan and Kind, 1978; Vaz et al., 1977), delayed hypersensitivity reactions (Miller and Hanson, 1979; Mowat et al., 1982), and changes in the rate of antigen clearance from the circulation (Hanson et al., 1979), as well as in vitro assays such as speciﬁc plaque-forming cells (Richman et al., 1978; Titus and Chiller, 1981), lymphocyte proliferation (Hanson and Miller, 1982; Higgins and Weiner, 1988; Lider et al., 1989; Titus and Chiller, 1981), and cytokine production (Fishman-Lobell et al., 1994; Weiner, 1997). Of note, the suppression of TGF-웁 response following oral antigen has not been reported for any dose or regimen of orally administered antigen. Nonetheless, these immunological parameters are differentially affected by oral antigen. Delayed-type hypersensitivity is more easily suppressed by oral antigen than is antibody formation in mice (Kay and Ferguson, 1989; Ke and Kapp, 1996; Mowat, 1986; Mowat et al., 1982), guinea pigs (Heppell and Kilshaw, 1982) and humans (Husby et al., 1994). Suppression of DTH requires lower doses of oral antigen and tolerance persists longer (Strobel and Ferguson, 1987). Also of note, suppression of DTH induction, but not of antibody responses, may be induced in neonates reconstituted with adult spleen cells (Hanson and Morimoto, 1987; Peng et al., 1989). On the other hand, oral tolerance to DTH is more readily broken by treatment with cyclophosphamide (Mowat et al., 1982) or estradiol (Mowat and Parrot, 1983) and is more difﬁcult to induce in protein-deprived mice than is oral tolerance to humoral responses (Lamont et al., 1987). A possible explanation for the in vivo differences in susceptibility to oral tolerance induction may relate to the susceptibility of distinct subsets of CD4⫹ T cells, namely Th1 and Th2 subpopulations, to oral tolerance induction. Th1-dependent cytokines, such as IL-2 and IFN-웂, are readily inhibited by feeding multiple low doses of antigen whereas Th2-dependent cytokines, such as IL-4, IL-5, and IL-10, require high doses of antigen to be suppressed (Fishman-Lobell et al., 1994; Gregerson et al., 1993; Melamed et al., 1996; Melamed and Friedman, 1994; Weiner, 1997). This would explain also why IL-4-dependent IgG1 responses are more resistant

ORAL TOLERANCE: MECHANISMS AND THERAPY

175

to oral tolerance induction than are IFN-웂-dependent IgG2a responses (Melamed and Friedman, 1994; Mowat et al., 1996). Other studies, however, argue against such a simple interpretation. It has been known since 1977 that the IL-4-dependent IgE response is highly susceptible to oral tolerance induction (Ngan and Kind, 1978; Saklayen et al., 1984; Vaz et al., 1977). At the same time, IL-5 and IFN-웂 may exhibit similar susceptibilities to feeding over a wide range of single doses of OVA (Mowat et al., 1996). A number of authors have reported that the development of oral tolerance is preceded by antigen-speciﬁc IFN-웂 production (Chen et al., 1997; Gautam et al., 1990; Hoyne and Thomas, 1995; Marth et al., 1996; Mowat et al., 1996), raising the theoretical possibility of a role for IFN-웂 in oral tolerance. Of note is that nasal tolerance induction to IgE responses seems to be dependent on 웂␦ T cells and IFN웂 production (McMenamin et al., 1995). Investigators working with selfantigens and superantigens also suggest that IFN-웂 may play a role in tolerance induction (Cauley et al., 1997). Liu and Janeway (1990) reported a reversal of tolerance by depletion of IFN-웂 and Cauley et al., (1997) have shown secretion of IFN-웂 by regulatory cells able to transfer suppression to staphylococcal enterotoxin A (SEA) to naive recipients. Some investigators have shown intact oral tolerance induction in IFN-웂 receptor-deﬁcient mice (Kjerrulf et al., 1997), whereas others have been unable to induce suppression by feeding antigen to these mice (Kweon et al., 1998). On the other hand, administration of IFN-웂 as well as processes such as graft-versus-host (GVH) disease that induce IFN-웂 inhibit oral tolerance and there is increased expression of class II in the gut after IFN-웂 administration (Mowat and Parrot, 1983; Strobel et al., 1985; Zhang and Michael, 1990). Consistent with these results, anti-IL-12 antibody augments TGF-웁 production and oral tolerance (Marth et al., 1996). It is quite possible that if there is a role for IFN-웂 in oral tolerance it may be limited to the very early stages of tolerance induction and be inhibitory in the later stages. IFN-웂 may have paradoxical effects in Th1-type autoimmune models. Many investigators have studied the effect of oral antigen on IgA responses. The pioneering experiments comparing the effects of oral administration of antigen on systemic responsiveness and mucosal immunity were published by Challacombe and Tomasi in 1980. They showed that oral administration of high doses of either soluble ovalbumin or Streptococcus mutans cell wall antigen led to systemic unresponsiveness associated with mucosal IgA response to the antigens (Challacombe and Tomasi, 1980). A complementary set of experiments demonstrated that parenteral immunization caused antigen-speciﬁc suppression of mucosal IgA response (Koster and Pierce, 1983), suggesting that mucosal and systemic contacts with

176

ANA M. C. FARIA AND HOWARD L. WEINER

antigens may have distinct effects on different sites of the immune system. Other studies conducted later conﬁrmed that feeding antigen yields a split regulation at mucosal versus nonmucosal sites. Fuller and co-workers demonstrated that large doses of orally administered MBP (25 mg) suppressed MBP-speciﬁc serum IgG and IgA antibody levels while augmenting the levels of salivary IgA speciﬁc for MBP (Fuller et al., 1990). Fujihashi and co-workers reported a booster in IgA responses, measured by gastrointestinal secretion and serum levels, after the oral administration of diphtheria toxoid in a 12-dose protocol (Fujihashi et al., 1996). IgG production measured both in spleen and in serum was reduced. The exemption of mucosal IgA response from the suppressive effect of antigen feeding was ﬁrst described in experiments showing that feeding 20 mg ovalbumin to mice led to the simultaneous generation, in Peyer’s patches, of Th cells that supported IgA responses and suppressor cells that migrated to the spleen and down-regulated antigen-speciﬁc IgM and IgG responses (Richman et al., 1981a,b). It is now well established that one of the most important cytokines present in the gut is TGF-웁. TGF-웁 has been shown to play an important role in oral tolerance and to suppress IgM and IgG B cell differentiation while serving as a switching factor for IgA production. We did not ﬁnd suppression of TGF-웁 at any dose fed to OVA–TCR transgenic mice (Chen et al., 1995a). In this regard, studies by Saklaylen reported that feeding and immunizing mice with ovalbumin resulted in differential effects on serum IgG, IgE, and IgA antibody levels. Suppression of IgG and IgE was readily demonstrated, whereas IgA levels were minimally affected by any feeding protocol (Saklayen et al., 1984). Grdic and co-workers reported in studies of CD8⫺/⫺ mice that antigen feeding induces systemic but not local IgA hyporesponsiveness, suggesting that CD8⫹ T cells in the normal gut mucosa exert an important down-regulatory function (Grdic et al., 1998). Other investigators (Oliver and Silbart, 1998) showed that orally administered dinitrochlorobenzene at higher doses induces both systemic and local tolerance, as indicated by reduced fecal IgA and IgG anti-DNP responses. Conversely, oral treatment with low doses of DNCB induced only local tolerance. B. KINETICS OF ORAL TOLERANCE Oral tolerance has been observed as early as 24 hr after an oral administration of protein (Challacombe and Tomasi, 1980; Melamed and Friedman, 1993a), and other studies in TCR transgenic mice demonstrate that the circuits of oral tolerance may be triggered much faster. At 6 hr after feeding 0.5 mg ovalbumin, there is a marked up-regulation in IL-4, IL-10, and TGF-웁 production in the gut-associated lymphoid tissue of OVA–TCR transgenic mice. Increases in IL-4 and IL-10 were found in the domes

ORAL TOLERANCE: MECHANISMS AND THERAPY

177

of Peyer’s patches whereas increases in TGF-웁 were observed in the interfollicular region and the villi (Gonnella et al., 1998). In addition, 6 hr after an oral dose of 0.5 mg of cytochrome c to cytochrome c–TCR transgenic mice, activated speciﬁc T cells in spleen and mesenteric lymph nodes can be observed (Gu¨tgemann et al., 1998). Strobel and co-workers have shown that feeding 25 mg ovalbumin to mice can induce the appearance of a tolerogenic moiety in the serum that can transfer oral tolerance to naive recipients 1 hr after feeding but not after 24 hr (Peng et al., 1990). These data suggest that the induction of oral tolerance is a sequence of events that begins during the ﬁrst 24 hr after the oral exposure to the antigen, although optimal induction of systemic suppression usually occurs 7–14 days after feeding (Mowat, 1994). Oral tolerance may persist for long periods of time. DTH responses have been reported to be inhibited for 17 months after a single feeding of large doses of antigen (Strobel and Ferguson, 1987), and proliferative responses are reported to be inhibited for as long as 6 months (Melamed and Friedman, 1993b). The length of persistence may relate to the mechanism induced. High-dose induction of anergy/deletion may persist longer than low-dose induction of regulatory cells. In the EAE model, protection against disease by multiple low-dose feedings lasted 2–3 months (Higgins and Weiner, 1988). Earlier reports on oral tolerance induction demonstrate that the suppression of humoral responses lasts about 2 months after feeding high doses of either ovalbumin or sheep red blood cells (SRBCs) (Challacombe and Tomasi, 1980; Ngan and Kind, 1978; Vaz et al., 1997). We have found that a decrease in systemic IgG response may persist for 1 year after one oral administration of 20 mg OVA in B6D2F1 mice (Faria et al., 1998). Moreau and Gaboriau-Routhiau (1996) reported that tolerance of serum IgG antibody responses lasts for 3–6 months after a single feeding of 20 mg OVA in C3H/He mice. Use of different strains of mice could account for these differences. Strobel and Ferguson (1987) also reported that suppression of antibody responses is present at 3 but not after 6 months of the oral treatment with a high dose of OVA, although there was a 25% suppression of antibody levels after 12 months of oral treatment. In experimental systems, it is not known whether the presence of antigen in the immunization regimen affects tolerance persistence. Melamed and Friedman (1993b) demonstrated that oral tolerance lasts longer in mice challenged with antigen in complete Freund’s adjuvant after oral treatment than in mice exposed to the antigen only by oral route. Long-lasting oral tolerance was also observed in mice challenged with antigen in alum after feeding 20 mg OVA (Faria et al., 1998). Although these data suggest the depot effect of the adjuvant could play a role, it is unlikely that this

178

ANA M. C. FARIA AND HOWARD L. WEINER

effect would account for 12 months of tolerance maintenance. More likely, mechanisms that preserve memory for immune responses relate to oral tolerance persistence. These mechanisms may include changes in the repertoire and life span of lymphoid cells and the induction of regulatory cells. Oral tolerance induction to particulate antigens shows a different kinetics. Suppression of cellular responses to SRBCs require a regimen of multiple feedings over a period of 2 weeks and the duration of the suppression lasts for only 6 weeks after the last feeding (Kagnoff, 1978). This difference for particulate antigens may relate to antigen processing and the fact that SRBCs gain access solely to the M cells on Peyer’s patches, as opposed to epithelial cells. IV. Factors Affecting Oral Tolerance Induction and Maintenance

A. ANIMALS Oral tolerance has been observed in a number of animal species, including mice, guinea pigs, rats, hamsters, and rabbits. Besides rodents, investigators have reported that oral tolerance can be readily induced in humans (discussed later). There is a report in the literature of a failed attempt to induce oral tolerance in horses. Nevertheless, only one dose of one antigen was tested (10 g ovalbumin) for 14 days (Fitzpatrick et al., 1992). 1. Age Age at the ﬁrst oral encounter with the antigen is an important factor determining whether tolerance or immunization occurs after feeding (Strobel and Ferguson, 1984). Animals fed OVA, MBP, or myelin oligodendrocyte glycoprotein (MOG) on the ﬁrst day or two of life do not develop oral tolerance and show evidence of systemic priming when challenged parenterally as adults. This priming effect has been demonstrated for DTH responses to OVA (Hanson, 1981; Strobel et al., 1981) as well as for disease induction in the EAE model with MOG and MBP (Maron et al., 1998a; Miller et al., 1994a). An exception is the ability to neonatally tolerize NOD mice with oral insulin (Maron and Weiner, 1998). The general lack of oral tolerance in neonatal mice is ascribed to two main differences between neonates and adults: the special characteristics of antigen absorption in neonates and the lack of a fully competent immune system. The development of intestinal permeability to macromolecules during the ﬁrst few days of life is peculiar in mice. The columnar absorptive cells of the small intestine are fully permeable to proteins and colloidal materials up to 15–17 days of age, when a phenomenon known as ‘‘closure’’ occurs (Clark, 1979; Hormatz et al., 1989). Before closure, there are few microvilli, there is little secretion of proteolytic enzymes, and the gastric pH is nearly neutral. Marked changes occur at day 16 after birth: microvilli appear,

ORAL TOLERANCE: MECHANISMS AND THERAPY

179

trypsin and HCl acid are secreted, and the absorption of proteins into the blood falls to adult levels (Clark, 1979). A study by Maron and Weiner (1998) using peptides for oral tolerance induction showed the prevention of spontaneous diabetes in NOD mice fed insulin B-chain 10–24 peptide during their ﬁrst 5 days of life. However, feeding of MOG peptide to 1- to 5-day-old NOD mice failed to suppress the development of EAE induced by immunization with MOG peptide, suggesting that poor enzymatic cleavage of the antigen is not the only factor involved in the phenomenon. Furthermore, beginning at day 2 after birth, young mice become progressively more susceptible to oral tolerance induction, although they are not able to degrade the antigen enzymatically in the gut during this period (Hanson, 1981). It is likely that oral tolerance induction for native proteins requires a fully competent immune system to develop. Gut-associated lymphoid tissues are poorly developed in young mice (Ogra, 1980) and this might account for the temporary failure to become tolerant. In addition, neonatal mice become more susceptible to oral tolerance induction when transferred with adult spleen cells (Hanson, 1981; Peng et al., 1989), suggesting that systemic aspects of immune competence are also needed. An interesting aspect of this transient lack in oral tolerance in neonate animals is the fact that it is not associated with development of allergic reactions to any of the proteins in milk. A possible explanation for this may be that human milk contains small amounts of cytokines such as IL1 and IL-6 and large amounts of TGF-웁. Oral administration of human colostrum and TGF-웁 in mice strongly inhibits anti-SRBC responses obtained following oral immunization with SRBCs (Ishizaka et al., 1994). It is possible that mouse colostrum is also rich in TGF-웁 and that TGF-웁 inhibits potential inﬂammatory responses elicited by milk proteins during the transient lack of oral tolerance in neonates. Although susceptibility to oral tolerance induction gradually improves with time, starting 48 hr after birth, there is a transient period of refractoriness around the weaning. This observation was reported for mice, piglets, and calves (Hanson, 1981; Miller et al., 1994b; Strobel et al., 1983). The defect seems to be related to the weaning process rather than to age, because delaying the weaning is followed by an equivalent delay in the refractory period. Weaning is associated with a rapid change of morphological parameters, ﬂuctuations in hormone levels, and alterations in intestinal ﬂora. Histologically, one can observe a lengthening of crypts, increase in epithelial cell kinetics, intraepithelial lymphocytes, mucosal mast cells, and eosinophils, and an increase in jejunal goblet cell numbers. Moreover, the amount of antigen uptake increases dramatically when breast milk is replaced by other forms of food (Strobel, 1996). Thus, several factors are likely responsible for the absence of oral tolerance in the neonatal period.

180

ANA M. C. FARIA AND HOWARD L. WEINER

Of note is that maternal Ig from colostrum and milk seems to play a role not only in the formation of the B cell repertoire but also in the induction of T cell tolerance. Oral Ig efﬁciently induces tolerance of ␬-speciﬁc CD8⫹ T cells for several weeks after birth in ␬⫺/⫺ pups breast-fed by ␬⫹/⫹ mothers. Tolerance appears mediated by suppression rather than deletion (RueffJuy et al., 1998). Age seems to interfere not only with susceptibility to oral tolerance induction but also with the immunological circuits triggered by fed antigen. Lundin and co-workers (Lundin et al., 1996) reported that young (4-week-old) and adult (12-week-old) rats can both be rendered tolerant to ovalbumin after a 4-week regimen of feeding OVA-containing pellets (0.8 g/day intake) but that the two age groups display differences in the mechanisms of tolerance induced. In adult rats induced regulatory cells are able to inhibit primed T cells in vitro and to mediate bystander suppression to another antigen. Young rats, on the contrary, showed no evidence of active suppression in vitro or bystander suppression in vivo and tolerance appears to be induced by anergy/deletion. In the young rats cell-mediated responses (DTH and T cell proliferation) were inhibited without suppression of the antibody response measured by serum levels of anti-OVA IgG and IgE antibodies. It is possible that the absence or lower levels of regulatory cells in the young fed rats could be due to the immaturity and lack of antigen presentation capabilities of the gut of these animals. The development of immune function in the intestine of humans has also been related to neonatal enteric disease (Insoft et al., 1996). Senescence has also been shown to affect oral tolerance. Studies in several strains of mice fed a single dose of 20 mg ovalbumin have demonstrated that susceptibility to oral tolerance induction for antibody response wanes as early as 25 weeks of age (Faria et al., 1993; Rios et al., 1988) and may be totally lacking in senile animals (Faria et al., 1998a,b; Lahmann et al., 1992). Interestingly, the age-associated impairment in the susceptibility to oral tolerance induction is relative and can be overcome by prolonged continuous feeding of the antigen (Faria et al., 1998a,b). Moreau and Gaboriau-Routhiau (1996) found induction of oral tolerance for IgG and IgE response in 20-month-old C3H/He mice fed 20 mg ovalbumin. These old mice, however, remain tolerant for a much shorter period of time than do young control animals. In studies using the parenteral route for tolerance induction, a decrease in susceptibility of tolerance induction in aged BALB/ c, C57BL/6, and CBA/CaJ mice (Habicht, 1980; Weigle et al., 1988) and a relative inability to induce tolerance in adult NZB and (NZB ⫻ NZW)F1 mice (Staples and Talal, 1969) has been described. The impairment of tolerance induction seems to affect both B and T cell populations in 6month-old BALB/c mice (DeKruyff et al., 1980; Doken et al., 1980).

ORAL TOLERANCE: MECHANISMS AND THERAPY

181

Mucosal-associated lymphoid tissue is also affected by aging. Levels of IgA, the number of plasma cells in the gut lamina propria, and mucosalassociated T cell functions are severely reduced in old animals (Kawanishi and Kiely, 1987; Kawanishi et al., 1989; Koyama et al., 1990). Furthermore, studies on the antibody repertiore in young and old mice show that aging is associated with a rise in auto-antiidiotypic antibody production (Goidl et al., 1980; Huetz et al., 1990; Weksler et al., 1989). Oral tolerance seems to be preserved in old mice fed antigen when they were young (Faria et al., 1998b), and transfer of spleen cells from 7- to 70-week-old syngeneic recipients can restore the ability of the latter to be rendered tolerant even to novel antigens (Lahmann et al., 1992). 2. Genetic Background The inﬂuence of the genetic background of oral tolerance has not been systematically studied using our latest understanding of the mechanisms and dose dependency of oral tolerance. Most mouse strains are readily tolerized by feeding a single 20-mg dose of ovalbumin, the antigen that has been most frequently used to study oral tolerance (Lamont et al., 1988b; Vaz et al., 1987). Mice carrying the H-2d MHC haplotype (such as BALB/c, and DBA/2) are particularly susceptible whereas mice carrying the H-2b haplotype (such as BALB/B, C57BL/6) are usually less susceptible. Other genes, however, also seem to play a role, because strains bearing the same H-2 haplotype in different genetic backgrounds show major differences in susceptibility to oral tolerance. The H-2k haplotype is illustrative—C3H/He mice are susceptible whereas C3H.SW and CBA mice are relatively resistant to oral tolerance induction using 20 mg of ovalbumin (Vaz et al., 1987). In studies using bovine gamma globulin (BGG) and HGG as antigens, Vaz and co-workers showed a similar inﬂuence of both MHC-linked and background genes in the susceptibility to oral tolerance (Rios et al., 1988). Immunologically relevant genes other than MHC related, such as those encoding the idiotypic regions of immunoglobulins, failed to show a signiﬁcant inﬂuence in oral tolerance sensitivity (Ishii et al., 1993; Lamont et al., 1988b). The nature of the inﬂuence exerted by genes other than the ones linked to MHC has not been explored. 3. Immunological Status Oral tolerance is easily induced in naive animals but is less inducible in animals already primed (Yoshino, 1995). In primed animals, oral administration of antigen may result in serum (Conde et al., 1998; Hanson et al., 1979) and mucosal (Elson and Ealding, 1984b) secondary responses. A single gastric administration of 20 mg ovalbumin can reduce by 25% the T cell lymphocyte proliferative response when introduced simultaneously

182

ANA M. C. FARIA AND HOWARD L. WEINER

or up to 4 days after immunization with ovalbumin and complete Freund’s adjuvant in the footpad (Melamed and Friedman, 1993b). These results indicate clearly that tolerogenic signals delivered by oral route are dominant as compared to immunogenic signals and this may reﬂect the ontogeny of the immune system wherein tolerance is a natural and precedent event. However, the degree of inhibition is usually inversely related to the interval between parenteral immunization (Melamed and Friedman, 1993b). The authors hypothesize that this short window of time for oral tolerance induction after priming may be related to the appearance of serum antibodies, which may bind to the antigen because administration of speciﬁc antibodies simultaneously with oral exposure to an antigen abrogates oral tolerance induction (Hanson et al., 1979a,b). Nonetheless, prolonged oral exposure instead of a single gavage is able to suppress ongoing immune responses (Bloch et al., 1983), including the formation of IgE (Lafont et al., 1982; Maia et al., 1974), DTH reactions (Lamont et al., 1988a,b), and T cell proliferation (Melamed and Friedman, 1993a,b). Investigators have also shown inhibition of ongoing immune responses by all effector T cell subsets, although antibody levels were not affected (Leishman et al., 1998). More importantly, multiple feeding regimens were reported to suppress ongoing disease in several T dependent autoimmune disease models (Higgins and Weiner, 1988; Lider et al., 1989; Thompson and Staines, 1990). Ongoing T cell responses are more easily suppressed than are established antibody responses. Antibody formation can be aborted by prolonged and continuous feeding of high doses of antigen only during the ﬁrst 2 weeks after primary immunization in mice (Conde et al., 1998). Again, differently from T cell responses, concomitant exposure to antigen by oral and parenteral route abrogates oral tolerance to antibody responses, suggesting that antibody formation may interfere with oral tolerance induction (Hanson et al., 1979a,b). The use of the B subunit of cholera toxin (CTB) coupled to antigens enhances the ability to induce oral tolerance in primed animals. Czerkinsky and co-workers (Czerkinsky et al., 1996) have shown that the oral administration of CTB-coupled antigens was effective at doses 100- to 1000-fold lower than the dose of corresponding unconjugated antigens required to induce similar levels of inhibition of DTH responses. This effect extends to a variety of antigens [SRBCs, trinitrophenyl (TNP), MBP] and still manifests in primed animals. A single dose of CTB conjugated to 25 애g myelin basic protein fed to Lewis rats 7 days after the induction of experimental autoimmune encephalomyelites abrogates clinical disease, whereas administration of MBP alone is effective only at 20-fold higher doses in a multiple-feeding protocol. The special binding properties of CTB to a ganglioside present in the epithelial cells are believed to play a

ORAL TOLERANCE: MECHANISMS AND THERAPY

183

major role in its ability to augment systemic tolerance and mucosal IgA responses simultaneously (discussed later). 4. Intestinal Flora The bacterial ﬂora of the intestine is an important modulator of oral tolerance, although its presence is not essential for tolerance induction. The ﬁrst report of feeding protein to germ-free animals was published by Wannemuehler and co-workers (Wannemuehler et al., 1982). They showed that germ-free mice could not be orally tolerized to SRBCs. Other investigators have shown normal oral tolerance induction to soluble ovalbumin in germ-free mice (Furrie et al., 1995a; Moreau and Corthier, 1988; Moreau and Gaboriau-Routhiau, 1996). Moreover, conventional C3H/HeJ, LPSunresponsive animals cannot be orally tolerized to SRBCs, unlike normal congenic conventional C3H. When fed SRBCs, C3H/HeJ mice develop systemic antibody and DTH responses as well as mucosal IgA antibodies. In these animals, oral tolerance, to soluble antigens (Mowat et al., 1986), components from the autologous ﬂora (Duchmann et al., 1996), and haptens (Elson et al., 1996) can be readily induced. Thus, the inability of germ-free mice to be tolerized to SRBCs may be related to two factors: (1) bacterial lipopolysaccharide (LPS), a major component of some bacterial walls, is a modulator of oral tolerance; (2) particulate antigens, such as SRBCs, are poor inducers of oral tolerance and they may require the presence of positive modulators such as LPS in the gut ﬂora. Several studies suggest LPS may play a modulatory role in oral tolerance (Khoury et al., 1992; Kim and Ohsawa, 1995; Mowat et al., 1986; Saklayen et al., 1983). A study on oral administration of bacterial stimulants such as LPS and Escherichia coli extract OM-89 indicated that these products induce a Th2 shift in gut cytokine gene expression. This shift may explain their boosting effect in oral tolerance induction (Bellmann et al., 1997). The presence of gut ﬂora is reported to interfere with the persistence of oral tolerance to ovalbumin (Moreau and Corthier, 1988). The duration of IgG anti-OVA antibody hyporesponsiveness is decreased in germ-free animals but not the inhibition of IgE- or DTH-speciﬁc responses (Moreau and Gaboriau-Routhiau, 1996). Moreover, LPS can enhance the induction of oral tolerance to ovalbumin (Kim and Ohsawa, 1995) and MBP (Khoury et al., 1990, 1992) when administered at the time of feeding. B. ANTIGENS It appears that all soluble proteins in the absence of adjuvants may induce oral tolerance, as can contact-sensitizing agents such as heavy metals, haptens, oxazolone, and dextran sulfate (Asherson et al., 1977; Elson et al., 1996; Gautam and Battisto, 1985; Neurath et al., 1996; Newby et

184

ANA M. C. FARIA AND HOWARD L. WEINER

al., 1980; van Hoogstraten et al., 1993; Waldo et al., 1994). Although these chemical compounds are not antigenic, they react to body proteins and become part of the antigenic structure of the latter. They probably have an effect similar to the haptenated self-components (Neurath et al., 1996) that are also able to induce oral tolerance. Several proteins of immunopathological importance have been successfully used in oral tolerance protocols. Some of them are potent allergens such as extracts of pollen or Dermatophagoides mites (Suko et al., 1995); others are antigens involved in autoimmune disease models. It was not until the late 1980s that oral administration of autoantigens was applied as a strategy to treat autoimmune diseases. In 1986, two reports appeared on the effects of oral tolerance in animal arthritis. Thompson and Staines (1986a,b) showed that oral treatment with type II collagen delayed the onset and decreased the severity of collagen-induced arthritis in rats. Nagler-Anderson et al., (1986) demonstrated suppression of type II collagen-induced arthritis by orally administered type II collagen in mice. In 1988, two independent groups (Bitar and Whitacre, 1988; Higgins and Weiner, 1988) reported that experimental autoimmune encephalomyelitis in rats could be suppressed by oral treatment with guinea pig myelin basic protein. Since then, many other autoimmune disease models have been suppressed by oral autoantigen administration (discussed below). Other antigens able to induce oral tolerance include superantigens, heterologous red blood cells (Kagnoff, 1978a,b; Kiyono et al., 1982; Mattingly and Waksman, 1978), allogeneic cells (Sayegh et al., 1992a,b), and inactivated viruses and killed bacteria (Bersani Amado et al., 1991; Challacombe and Tomasi, 1980; Rubin et al., 1981; Stokes et al., 1979). As expected, the immune system is tolerant to autologous intestinal ﬂora (Foo and Lee, 1972) and breakdown in this tolerant state may lead to development of inﬂammatory bowel disease. This general assumption is supported by studies in humans showing that mononuclear peripheral cells do not proliferate in response to bacterial sonicates derived from autologous intestine but do proliferate in response to bacterial sonicates derived from heterologous intestine (Duchmann et al., 1996). In mice, mononuclear cells from the spleen and small and large bowel show no proliferative response to bacterial products from their own ﬂora but display a strong response against bacterial sonicates from heterologous intestine of syngeneic mice. Furthermore, experimental induction of chronic intestinal inﬂammation with trinitrobenzene sulfonic acid (TNBS) abrogates both systemic and local tolerance to intestinal bacterial ﬂora (Duchmann et al., 1996). Several groups have reported the induction of oral tolerance to peptides containing T cell epitopes (al-Sabbagh et al., 1994; Gregorian et al., 1993;

ORAL TOLERANCE: MECHANISMS AND THERAPY

185

Hoyne et al., 1994; Javed et al., 1995; Khare et al., 1995; Miller et al., 1993a,b). In recent years, the use of intact and altered forms of imunogenic peptides has allowed a more detailed analysis of T cell activation in disease and tolerance induction. Oral administration of peptide sequences has yielded variable results, suggesting that intestinal uptake and handling of such small structures may differ in some respects from normal digestion and absorption of whole proteins. Knowledge of the inﬂuence of structural antigenic differences following antigen feeding is important for using oral tolerance to treat autoimmune conditions. Three classes of antigens fail to induce oral tolerance: polysaccharide antigens, particulate antigens, and proteins that bind to the intestinal epithelium. Polysaccharides are T cell-independent antigens that interact directly with B cells. Oral tolerance, on the other hand, is known to be a thymus-dependent phenomenon. As demonstrated years ago, a suppressed antibody response to a hapten resulting from feeding the haptenconjugated protein can be restored if a second protein carrier coupled to the hapten is used for the immunization (Hanson et al., 1977; Reese and Cebra, 1975, Richman et al., 1978). These classic experiments clearly show that although antibody production is ultimately affected by oral tolerance induction, suppression of antibody responses is T cell dependent. This suggests that by activating B cells directly, these antigens would avoid the gut T cell circuits especially prone to tolerance induction (Stockes et al., 1979). Second, certain particulate antigens, aggregated proteins, and invasive bacteria are often unable to induce oral tolerance. Instead they are strong immunogens for mucosal IgA responses. Several bacterial products have been shown to induce secrectory as well as systemic immune responses in mice when administered by the oral route (Stockes et al., 1979). Administration of protein antigens as an oil-in-water emulsion also failed to induce tolerance and elicited a signiﬁcant serum IgG antibody response (Kaneko et al., 1998). These components were either T-independent or particulate antigens. For the latter, it is speculated that their exclusive entry through the M cells of Peyer’s patches may facilitate induction of a mucosal immune response and may bypass some of the suppressor mechanisms triggered in the villi and also involved in oral tolerance. As discussed later, we have achieved oral tolerization of animals deﬁcient in Peyer’s patches, indicating that PPs are not the exclusive site of oral tolerance induction. Experiments with rats colonized from birth with an E. coli genetically manipulated to produce ovalbumin showed that, at 2 months of age, these animals have an increased DTH reaction to OVA and they produce higher levels of anti-OVA antibodies after immunization with OVA in complete Freund’s adjuvant than do noncolonized control animals (Dahlman et al., 1992).

186

ANA M. C. FARIA AND HOWARD L. WEINER

Interestingly, colonized rats eating an OVA-containing diet showed a 73% decrease in DTH reaction to OVA and signiﬁcantly lower levels of IgE and IgG antibodies to OVA compared to colonized rats fed a normal diet. Thus, feeding the antigen abrogated the intestinally induced hypersensitivity (Dahlman et al., 1994). The OVA produced by the E. coli is retained in the bacterial periplasmic space and not secreted, so there is probably a limited amount of free OVA in the intestine of colonized animals. Thus, two main differences exist between the OVA present in the diet and the bacterial particulate OVA. First, the antigen in the diet is exposed to enzymatic digestion in the intestinal lumen. Second, soluble OVA is taken up over the whole surface of the intestine, including villi and M cells, whereas in the colonized animals, OVA is predominantly taken up in a bacterial context in Peyer’s patches. Presumably aggregated antigens would follow the same route of entry, M cells of Peyer’s patches, and this would also explain their poor performance as oral tolerance inducers. Nonetheless, some isolated components of bacteria, such as streptococcal cell wall, (SCW), are able to induce circulating TGF-웁 and suppress the development of SCW-induced erosive polyarthritis in rats (Chen et al., 1998b). It is unclear whether these antigens are able to enter through the same route as soluble antigens or whether they are strong stimulators of TGF웁 secretion in the Peyer’s patches, being able to elicit both IgA responses and oral tolerance. A third class of antigens that fail to elicit oral tolerance are proteins with the ability to bind to intestinal epithelial cells. In this category are included bacterial lectins such as cholera toxin (CT) and E. coli heat-stable enterotoxin (LT). Instead, these antigens abrogate oral tolerance induction to other antigens (Elson and Ealding, 1984a; Gaboriau-Routhiau and Moreau, 1996) and elicit high levels of secretory IgA. Thus, they are being widely used as adjuvants for oral immunization. The usual explanation for the well-known refractoriness to tolerance induction displayed by these materials relies on their abnormal entrance via the columnar gut epithelium rather than through the M cells present in Peyer’s patches. It is often argued that Peyer’s patches are the major sites for oral tolerance induction and bypassing this route would trigger a mucosal immune response. However, Peyer’s patches are also the privileged sites for priming IgA responses in mucosal surfaces, a type of immunity these antigens are known to evoke. Also, as mentioned before, we observed oral tolerance induction in mice lacking Peyer’s patches. Moreover, data on the use of isolated subunit B of cholera toxin have shown it is a strong adjuvant for oral tolerance induction, yet it retains the binding ability of the whole molecule (Czerkinsky et al., 1999). The delicate balance between mucosal immunity and oral tolerance induction suggests that antigen presentation in the gut is a crucial

ORAL TOLERANCE: MECHANISMS AND THERAPY

187

event in determining which class of response will prevail. In this regard, Agren and co-workers (Agren et al., 1997) have shown that cholera toxin retains its adjuvant ability if the A subunit is linked to the Ig-binding domains of Staphylococcus protein A, a substance that binds to lymphoid cells rather than epithelial cells. On the other hand, the boosting effect of multiple emulsion systems in oral tolerance induction to proteins is belived to be due to the preferential delivery of the antigen into the lymphoid follicles instead of into the epithelial layer (Elson et al., 1996). Oral administration of the bacterial superantigen staphylococcal enterotoxin B activates mucosal T cells with induction of IL-2 and IFN-웂 within 4 hr of feeding, suggesting a role for the GALT in the gastrointestinal manifestation of enteric poisoning (Spiekermann and Nagler-Anderson, 1998). Whether antigen presentation by epithelial cells versus presentation by professional APCs in Peyer’s patches can explain the dramatic shift in the immune mechanisms triggered by different forms of antigen entrance into the gut remains to be established. It is also possible that tolerance induction and mucosal responses are simultaneous nonexcludable events, as shown by several authors, and the mechanisms for induction of both phenomena are overlapping. Of note is that CT, which prevents oral tolerance induction when coadministered with the antigen, is unable to break established tolerance to DNCB (Oliver and Silbart, 1998) and aﬂatoxin B1 (Oliver et al., 1997) orally administered. At mucosal sites, however, coadministration of CT and staphylococcal entertoxin B (SEB) is able to prevent anergy of Peyer’s patch T cells induced by oral SEB administration alone (Iijima et al., 1998). 1. Structure of the Antigen As a T cell-dependent phenomenon, oral tolerance is not expected to depend on the native conﬁguration of the antigen. However, because T cells only recognize antigens presented by other cells, the type of APC and the context of antigen presentation in the gut are major determinants for oral tolerance development. This is most likely the best explanation for the observation that modiﬁcations in the structure of the antigen fed led to profound effects in the immunological outcome of the feeding. Changes in structure may affect the uptake and handling by antigenpresenting cells in the gut. Stimulating dendritic cells with Flt3 ligand enhances oral tolerance (Viney et al., 1998). Although soluble proteins easily induce oral tolerance, heat- or chemically denatured proteins are less effective (Peng et al., 1995, 1998a,b; Stransky et al., 1998). It is possible that denaturation facilitates protein enzymatic digestion and that extended degradation may destroy relevant T cell epitopes. Similarly, cationization of proteins by side chain substitution

188

ANA M. C. FARIA AND HOWARD L. WEINER

abrogates oral tolerance induction (Domen et al., 1987; Jacobs et al., 1993). Three possibilities were suggested to explain the effects observed in cationized antigens. First, cationized proteins are more resistant to enzymatic degradation (Domen et al., 1987). Second, these antigens tend to adhere better to slightly anionic cell membranes (Muckerheide et al., 1987). They are ﬁrmly retained by glomerular basement and synovial membranes, being able to trigger glomerulonephritis and arthritis in animals (Border et al., 1982; Schalkwijk et al., 1985). Thus, they probably have inﬂammatory abilities through complement ﬁxation. Third, this same ability to bind to anionic cell membranes may facilitate their uptake by antigen-presenting cells. Indeed, these proteins are shown to enter the cell by a nonspeciﬁc adsortive mechanism (Apple et al., 1988). In addition, because they interact directly with the APC membrane, the need for a degradative processing event might be circumvented. Interestingly, all three properties described here are reported to interfere with the induction of oral tolerance in other circumstances. 2. Enzymatic Digestion and ‘‘Biologic Filtration’’ of Antigen Although extensive enzymatic degradation of antigen has been demonstrated to interfere with oral tolerance induction, some degradation of antigen seems to be a crucial step. Bypassing gastric digestion by injection of antigen into the ileum in mice diminishes the degree of suppression observed after parenteral challenge with antigen (Michael, 1989; Stransky et al., 1998) and allows for the generation of immune responsiveness to OVA and bovine serum albumin (BSA). Oral administration of ovalbumin ‘‘enterecoated’’ with an acid-resistant acrylic polymer induces a strong humoral immune response in BDF1 mice measured by increased OVAspeciﬁc IgA, IgG1, and IgE ( Jain et al., 1996). In addition, studies analyzing the physicochemical properties of important food allergens such as peanut lectin, soybean, milk 웁-lactoglobulin, and mustard show that in contrast to nonallergenic components of the diet, these substances are very stable to acid gastric digestion (Astwood et al., 1996). Hanson and co-workers also showed that the in vivo inhibition of enzymatic degradation of antigen by administration of aprotinin, a protease inhibitor, abrogates oral tolerance induction (Hanson et al., 1993). Nonetheless, others have shown oral tolerance to MBP in animals fed antigen with protease inhibitor (Whitacre et al., 1991). ‘‘Biologic ﬁltration’’ of antigen may be important in oral tolerance and was ﬁrst described in 1983 by Strobel and co-workers (Strobel et al., 1983). They observed that serum collected 60 min after feeding ovalbumin or BSA to BALB/c mice transfers speciﬁc tolerance to naive recipients. Serum collected 5 min after feeding OVA, although containing signiﬁcant levels

ORAL TOLERANCE: MECHANISMS AND THERAPY

189

of immunoreactive OVA, was unable to transfer tolerance. This was true even after dose adjustment to levels of antigen equivalent to those measured in serum collected 1 hr after feeding (Peng et al., 1990). The mechanism involved in the generation of the ‘‘tolerogenic form’’ of the antigen was termed ‘‘biological ﬁltration’’ of the protein. It is known that intravenously injected proteins left to circulate for a certain period before transfer to recipient mice can also induce tolerance, and thus biological ﬁltration was postulated to remove aggregated immunogenic portions of the antigen, leaving in the serum a deaggregated tolerogenic fraction (Das et al., 1973, Das and Leskowitz, 1974). However, this does not appear to be the mechanism after feeding (Bruce and Ferguson, 1986b), because serum collected from irradiated or severe combined immune-deﬁcient (SCID) mice, which lack a functioning GALT, is unable to induce suppression of DTH following transfer to naive mice. Tolerance induction by intravenous injection of antigen was not impaired in these mice. Moreover, the ability to induce the tolerogenic moiety in the serum was fully recovered in irradiated mice after their reconstitution with spleen cells from syngeneic normal donors, suggesting that the lymphoid population rather than the epithelial cells in the gut may be involved in the processing of the antigen. We tested the ability of serum from C57BL/6 nu/nu mice as well as C57BL/6 B celldeﬁcient mice fed 20 mg OVA 1 hr before bleeding to transfer tolerance to C57BL/6 naive mice. Both strains were able to transfer tolerance, suggesting that they can generate the tolerogenic moiety when fed and that neither thymic-derived T cells nor B cells are important for the putative antigen processing in the gut lymphoid tissue (R. Maron, unpublished). The exact mechanism responsible for this processing event is still unclear, but subsequent pieces of evidence from Strobel, Ferguson, and co-workers suggest that the putative tolerogen is some altered form of the antigen (Bruce and Ferguson, 1986a,b; Furrie et al., 1994, 1995b; Peng et al., 1990). Afﬁnity absorption of immunoreactive OVA from tolerogenic serum rendered that serum nontolerogenic in recipient mice. Furrie and coworkers were able to demonstrate by immunoblotting using anti-OVA antibodies the presence of two bands of 21 and 24 kDa in the serum of BALB/c but not SCID mice fed ovalbumin 1 hr but not 5 min before (Furrie et al., 1995b). It is postulated that these moieties or tolerogens are generated in some way by passage of the originial antigen through the intestine. 3. Intestinal Absorption and Antigen Presentation in the Gut There are two major routes in the gut by which antigens can gain access to the body: (1) through the specialized epithelial cells or M cells overlying Peyer’s patches and (2) through the epithelial cells of the villi. Under

190

ANA M. C. FARIA AND HOWARD L. WEINER

normal circumstances, water, electrolytes, and small molecules can cross the epithelial barrier through the space between the enterocytes. Macromolecules, however, very seldom trafﬁc through this route (Sanderson and Walker, 1993). Of note is the observation that intestinal absorption of small but signiﬁcant amounts of macromolecular proteins into the bloodstream has been described in several mammalian species (Peng et al., 1990). Fed antigen interacts with the immune system at a number of levels and anatomic sites (Fig. 3). Antigen ﬁrst enters the GALT via Peyer’s patches or the villi. Conventional antigen presentation occurs in Peyer’s patches where antigen is presented by either macrophages, dendritic cells, or B cells to conventional T cells. In the villi, epithelial cells and dendritic cells are present as well as 웂␦ T cells and NK1.1 T cells, but the details of antigen presentation to these nonconventional populations of T cells are still unclear. These interactions occur after administration of both low and high doses of antigen. Antigen is drained by the lymphatic network to the mesenteric lymph nodes, where further antigen presentation may occur. After high-dose administration, some of the antigen passes into the liver via portal circulation and then to the bloodstream. In the liver, 웂␦ T cells and NK1.1 T cells may play an important role in tolerance induction. Antigen that enters the bloodstream can affect the immune system in the spleen or even distal sites such as the thymus. In all instances, antigen is processed by the immune system and presented to T cells and other cells to induce active immunologic events that result in generation of regulatorytype cells or anergy/deletion. Penetration of macromolecular antigens into the circulation occurs physiologically primarily as a result of the activity of special absorptive M cells. Not only macromolecules but much larger particles, such as viruses, microspheres, or pollen grains, may gain access to the circulation of adult animals through these cells (Volkheimer, 1975). As cells highly specialized in absorption, M cells have a large apical surface, low enzymatic activity, and can express binding receptors for microorganisms. Living organisms such as poliovirus seem to be absorbed in a receptor-speciﬁc fashion (Mayer, 1996; Sicinki et al., 1990), but usually macromolecules can cross the cytoplasm of these cells in pynocytotic vesicles (Neutra and Kraehenbuhl, 1996). Insolubility or globular antigens may favor uptake by M cells. Antigen penetration through the M cells is a very fast event, taking no more than 30–60 min for the antigen to reach the Peyer’s patches (Neutra and Kraehenbuhl, 1996). Thus M cells can clearly absorb antigen form the lumen in healthy adults and this process may be more intensive in younger animals (Halsey and Benjamin, 1976). The data supporting the ability of M cells to act as antigen-presenting cells are less compelling (Allan et al., 1993). Most likely antigens sampled through the M cells are either taken

ORAL TOLERANCE: MECHANISMS AND THERAPY

191

FIG. 3. Antigen presentation after oral administration of antigen. Fed antigen interacts with the immune system at a number of levels and anatomic sites. Antigen ﬁrst enters the GALT via Peyer’s patches or the villi. Conventional antigen presentation occurs in Peyer’s patches, where antigen is presented by either macrophages (Mo), dendritic cells (DC), or B cells (B) to conventional T cells (T). In the villi, epithelial cells (EC) and dendritic cells are present as well as 웂␦ T cells (IEL) and NK1.1 T cells (TNK ). These interactions occur after administration of both low and high doses of antigen. Antigen is drained by a lymphatic network to the mesenteric lymph nodes, where further antigen presentation may occur. After high-dose administration, some of the antigen passes into the liver via portal circulation and then to the bloodstream. In the liver, 웂␦ T cells and NK1.1 T cells may play an important role in tolerance induction. Antigen that enters the bloodstream can affect the immune system in the spleen or even distal sites such as the thymus.

up by conventional APCs (dendritic cells, macrophages, and B cells) underlying the epithelium or collected by the lymphatic and blood vessels, which reach the mesenteric lymph nodes and the liver, respectively. There are several cells capable of antigen presentation in the gutassociated lymphoid tissue (Bland and Kambarage, 1991; Mahida et al., 1988, Panja and Mayer, 1994; Sminia and Jeurissen, 1986; Spalding et al., 1983, 1984). These include macrophages and dendritic cells in Peyer’s

192

ANA M. C. FARIA AND HOWARD L. WEINER

patches and lamina propria, B lymphocytes, and epithelial cells. Macrophage-enriched cells obtained from mice fed high doses of OVA are able to stimulate antigen-primed lymph node T cells in vitro in an antigen-speciﬁc fashion without further exposure to antigen (Richman et al., 1981a). Dendritic cells also appear to acquire antigen fragments following oral or intraintestinal injection (Liu and MacPherson, 1993). Nonprofessional APCs exist in the gut and they have been studied as potential regulators of mucosal immune responses (Panja et al., 1994). Intestinal epithelial cells have the ability to process and present soluble antigens to immunocompetent T cells in vitro (Mayer, 1996). B cells also have been reported to act as antigen-presenting cells (Chesnut et al., 1982; Chesnut and Grey, 1981; Rock et al., 1984). There are two possibilities to account for the mechanism by which antigen presentation in the gut generates oral tolerance. The ﬁrst is that professional APCs present in Peyer’s patches and in the lamina propria present antigen sampled by the M cells and the epithelia of the villi in a conventional fashion (Mahida et al., 1988). The induction of Th2 and Th3 cells is secondary to the special milieu of the intestine, which is rich in IL-4, IL-10, and TGF-웁. The second is that APCs in the gut are structurally unique from APCs in the other parts of the immune system. The possibility that antigen presentation in the gut by nonconventional APCs may be involved in oral tolerance induction has been suggested by Mayer et al. (1996). Regulatory cells are generated in Peyer’s patches after feeding antigen (Asherson et al., 1977; Mattingly and Waksman, 1978; Ngan and Kind, 1978; Richman et al., 1978; Santos et al., 1994; Thomas et al., 1976), and may appear later in other organs such as the spleen as a result of selective migration. L-Selectin and B7 integrin molecules are important for migration of cells to mesenteric lymph nodes and Peyer’s patches (Steeber et al., 1998; Wagner et al., 1998). Regulatory cells may be either CD4⫹ or CD8⫹, depending on the system. Recent studies in knockout mice show that CD8-deﬁcient but not CD4-deﬁcient animals can be rendered tolerant by feeding (Desvignes et al., 1996; Tada et al., 1996). Antigen presentation in Peyer’s patches is mainly performed by dendritic cells and macrophages. Flt3L, a growth factor that expands dendritic cells in vivo, markedly enhances oral tolerance (Viney et al., 1998). These results emphasize the importance of antigen presentation and that oral tolerance is an active immunologic event. Th2 and Th3 cells are preferentially generated in the GALT (Daynes et al., 1990; Xu-Amano et al., 1992). Th2 differentiation depends on the cytokine microenvironment (Abbas et al., 1996). IL-12 drives Th1 cell differentiation whereas IL-4 induces Th2 differentiation. Both IL-4 and TGF-웁 enhance Th3 cell differentiation. The intestinal mucosa has high

ORAL TOLERANCE: MECHANISMS AND THERAPY

193

basal levels of IL-4, IL-10, and TGF-웁 expression, and shortly after oral administration of antigen, their expression is up-regulated (Gonnella et al., 1998). We have shown that dendritic cells, when exposed to IL-10, can drive Th2 cell differentiation of naive OVA TCR CD4⫹ transgenic T cells (Liu et al., 1998c). The inﬂuence of the cytokine milieu on the antigen presentation by DCs has also been demonstrated in vivo. Thus, DCs exposed to IL-10 in vitro, when injected into the footpad of mice, can prime for Th2-type responses (De Smedt et al., 1997). It has been shown in humans that prostaglandin E2 (PGE2)-treated DCs can produce IL-10 and can prime naive human T cells for Th2 differentiation (Kalinski et al., 1997). DCs from PP preferentially stimulate Th0 clones to produce large amounts of IL-4 and IL-10 while DCs from spleen induce high IFN-웂 production (Everson et al., 1997; Iwasaki and Kelsall, 1999). It is possible, then, that DCs in the gut, under the inﬂuence of the gut cytokine milieu, preferentially present antigen for Th2 and Th3 differentiation. In addition, IL-10 enhances the generation of regulatory cells that secrete IL-10 and TGF-웁. The inﬂuence of APCs on the preferential Th2 and Th3 cytokine proﬁle exhibited by mucosal T cells most probably also involves costimulatory signals required for the activation of T cells. B7.1 and B7.2 are two major costimulatory molecules. As mentioned before, B7.2 has been shown to be critical for Th2-type cell differentiation (Freeman et al., 1995). To determine the role of costimulatory molecules in the induction of oral tolerance, we have tested the effect of anti-B7.1 or anti-B7.2 mAb on the induction of tolerance by both high- and low-dose antigen feeding (Liu et al., 1998b). In the EAE model, injection of anti-B7.2 but not anti-B7.1 inhibited the induction of oral tolerance to low-dose (0.5 mg) but not highdose (20 mg) MBP. The receptors for B7.1 and B7.2 are CD28 and CTLA4. CD28 delivers positive signals that are required for T cell activation whereas CTLA-4 delivers negative signals that are crucial for immune down-regulation. Deﬁciency in CTLA-4 led to systemic hyperactivation of lymphocytes. We have found blocking both CD28 and CTLA-4 partially prevented T cell tolerance, whereas selective blockade of CTLA-4 abrogated oral tolerance induced by high-dose antigen (Samoilova et al., 1998). These results suggest that CTLA-4 is directly involved in the induction of high-dose oral tolerance. Investigation of costimulatory requirements for the generation of mucosal cytotoxic CD8⫹ T cells indicated a role for B7.1, which was not required for generation of peripheral CD8 T cells (Kim et al., 1998). It has been suggested that lamina propria cells (LPCs) may be involved in antigen presentation for oral tolerance induction (Harper et al., 1996). Antigen-pulsed lamina propria APCs stimulated high levels of IFN-웂 and TGF-웁 in the absence of other cytokines. Adoptive transfer of Ag-pulsed

194

ANA M. C. FARIA AND HOWARD L. WEINER

LPCs to syngeneic animals abrogated the induction of DTH responsiveness in the recipient mice. The type of APC responsible for this effect, however, is not yet clear. Mayer and co-workers have shown that gut epithelial cells can present antigens (Bland and Warren, 1986; Kaiserlian et al., 1993; Mayer and Shlien, 1987) and they selectively activate CD8⫹ cells (Li et al., 1995; Mayer and Shlien, 1987). Cumulative ﬁndings of confocal and electron microscopy demonstrate that antigens taken up by epithelial cells gain access to the cytoplasm, where they are transported to the endoplasmic reticulum and allowed to associate with class I molecules (Mayer, 1996). Interestingly, blocking monoclonal antibodies to class I and class II fail to inhibit CD8⫹ T cell activation induced by normal epithelial cells. Antibodies to the class Ib molecule CD1d are capable of inhibiting proliferation in these cocultures (Panja et al., 1993). The uptake of antigen by these cells is slower, suggesting that they are inefﬁcient antigen-processing cells, and that antigen processing and presentation by epithelial cells are not classical in any sense. Soluble antigens sampled by epithelial cells are, according to this model, likely to evoke suppressor mechanisms such as the one involved in oral tolerance induction. This would provide an explanation for the impressive boosting effect in oral tolerance exhibited by coupling cholera toxin B subunit to soluble antigens, because epithelial cells present a ganglioside receptor that binds CTB, focusing the uptake of these antigens through the villi. However, invasive bacteria can circumvent the intracellular trafﬁcking pattern described. They are instead transported in separate vesicles from the apical to the basolateral surface. Epithelial cells probably contribute to both tolerance and active immunity, depending on the antigen trafﬁcking pathway within the cell. A similar mechanism may exist for M cells. Epithelial cells are capable of stimulating CD4⫹ T cells as well (Kaiserlian et al., 1993; Mayer and Eisenhardt, 1990) via class II-mediated pathways, because class II antigens are also expressed by these cells (CerfBensussan et al., 1984; Mayer and Shlien, 1987). In this scenario, tolerogens might be turned into immunogens by altering the structure of antigen (therefore modifying its route of entry), the cytoplasmic trafﬁcking within potential APCs, the restriction element they are presented in, and ﬁnally the subpopulation and the pattern of T cell activation they trigger. 4. Inﬂammatory Processes in the Gut and the Abrogation of Oral Tolerance Inﬂammatory processes in the gut are known to alter intestinal permeability, change the cytokine balance in the gut mucosa, and hinder oral tolerance induction. Induction of inﬂammatory bowel disease in animals inhibits oral tolerance (Duchmann et al., 1996; Mowat and Parrot, 1983).

ORAL TOLERANCE: MECHANISMS AND THERAPY

195

Similarly, IL-10 knockout mice develop chronic inﬂammatory bowel disease affecting multiple sites in the intestinal tract if housed under conventional conditions (Kuhn et al., 1993). Animals held in speciﬁc pathogen free (SPF) conditions do not have small bowel lesions (Kuhn et al., 1993) and mice raised under germ-free conditions do not develop any intestional lesions at all (Sellon et al., 1997), suggesting that reactivity to intestinal ﬂora is required for colitis. Other procedures that induce inﬂammatory bowel disease in mice, such as graft-versus-host reaction or interarectal administration of TNBS (TNBS-induced colitis), are also shown to inhibit oral tolerance induction to proteins (Mowat and Parrot, 1983) and bacterial ﬂora (Duchmann et al., 1996). Tolerance can be restored by administration of IL-10 or administration of anti-IL-12 antibody (Duchmann et al., 1996). Conversely, in a model of Crohn’s disease induced in SCID mice by infusion of naive CD45RBhi T cells, colonic inﬂammation is blocked by the coinfusion of mature CD45RBlo T cells (Powrie et al., 1994). The protective effect was shown to be mediated by secretion of TGF-웁 (Powrie et al., 1996). TNBS colitis can also be prevented by feeding TNBShaptenated colonic protein at the time of intrarectal hapten administration. Such feeding leads to the appearance of TGF-웁-secreting cells both in Peyer’s patches and at the inﬂammatory site, and coadministration of antiTGF-웁 antibody at the time of the feeding abrogates its protective effect (Neurath et al., 1996). There is evidence that increased intestinal permeability to macromolecules, either in human Crohn’s disease and celiac disease (Walker, 1986) or following treatment with indomethacin in mice (Louis et al., 1996), is associated with high levels of food-speciﬁc antibodies and abrogation of oral tolerance. It is probable that in these abnormal conditions, the increased rate of antigen uptake does not occur through the M cells but instead through the damaged epithelia of the villi. Increased absorption per se may not determine the induction of immunity, but again the cytokine milieu in which the antigens are processed and presented to T cells is important for the outcome of feeding. 5. Dose of Antigen Oral tolerance can be induced by a wide range of antigen doses, and the exact dose required depends on the protein being administered. The inﬂuence of dose has been described earlier. Effective suppression of immune response to ovalbumin can be achieved with doses ranging from 1 to 25 mg. Doses below the tolerogenic range have been shown to induce priming in mice (Lamont et al., 1989; Meyer et al., 1996b) and humans ( Jarrett, 1984). Usually, DTH responses can be easily tolerized by low doses of antigen whereas antibody responses require higher doses. Of note is that although these different mechanisms may be preferentially triggered,

196

ANA M. C. FARIA AND HOWARD L. WEINER

depending on dose, they may overlap in many cases. This may be especially true for situations spanning the two extremes of very high and very low dosages of antigen. C. REGIMENS OF FEEDING The frequency and form of administration of the antigen fed is a crucial determinant factor for the induction of oral tolerance. Experiments in mice (Saklayen et al., 1984) and guinea pigs (Heppell and Kilshaw, 1982) show that multiple feedings were more effective than a single feeding of the antigen, and this latter regimen can even supress ongoing IgE immune responses. A number of investigators have shown that multiple feeding is a more effective regimen to induce tolerance to naive animals (Faria et al., 1998a,b) and is able to suppress ongoing immune responses and autoimmune disease development (Higgins and Weiner, 1988; Melamed and Friedman, 1993a,b; Meyer et al., 1996b; Peng et al., 1989; Thompson and Staines, 1990). Multiple feeding of high doses of antigen seems to be more effective in inhibiting IgG responses as well as DTH responses. In addition, Wu and co-workers (Wu et al., 1998) have shown in OVA–TCR transgenic mice that a regimen of multiple feeding of microdoses (100 애g/day) of OVA is able to suppress selectively Th2 responses (measured by speciﬁc IgE and IL-4 production). The population of CD4⫹ clonotypic T cells and antigen-presenting cells expressing B7.2 on the surface was decreased in the spleen of the mice that underwent the feeding regimen, suggesting that T cell anergy is a mechanism for the oral tolerance induction in this model. On the other hand, Hoyne and Thomas (1995) reported that in mice fed multiple low doses (3 mg) of OVA there is no depletion of antigenreactive T cells following oral tolerance induction. OVA-reactive T cells can be detected in the mesenteric lymph nodes, Peyer’s patches, and spleen and they produce large amounts of granulocyte–macrophage colony-stimulating factor (GM-CSF) and IFN-웂 but no IL-2 following activation in vitro. We also studied cytokine production after multiple feeding of low doses of MBP to SJL mice (Chen et al., 1994). We found that MBP T cells are present in the draining lymph nodes and spleens of MBP-fed mice and they secrete Th2 cytokines (IL-4 and IL-10) in addition to TGF웁. In experiments with OVA–TCR transgenic mice fed low doses (500 애g) of OVA, we also found an increase in the frequency of clonotypic CD4⫹ T cells in the Peyer’s patches. Thus, our results and the results of others do not suggest that signiﬁcant amounts of anergy or deletion after multiple lowdose feeding occur in mice. When compared to single-dose regimens, the multiple low-dose feeding protocol is more efﬁcacious because a single gavage of doses below 100 애g of OVA actually primes the mice.

ORAL TOLERANCE: MECHANISMS AND THERAPY

197

The continuous administration of antigen in drinking water or in the diet produces even more profound tolerance. The suppression achieved by continuous feeding of the antigen affects both Th1 and Th2 responses. Melamed and co-workers (Melamed et al., 1996) have reported that continuous feeding of OVA is able to inhibit selectively Th2 responses whereas a single feeding suppresses only Th1 responses. We have also shown that continuous feeding of OVA is more effective at tolerizing for IgG1 responses compared to multiple gavages, and a single day of continuous feeding is able to tolerize aged mice otherwise refractory to oral tolerance induction for IgG1 responses (Faria et al., 1998a,b). In addition, continuous feeding of high doses of OVA was also able to suppress ongoing IgG1 responses in mice up to 10 days after immunization, whereas a gavage protocol was ineffective (Conde et al., 1998). We also found that continuous feeding of low doses of OVA in C57BL/6, BALB/c, and OVA–TCR transgenic mice enhanced the production of active TGF-웁 by spleen cells as compared to multiple feedings of the identical amount of antigen (Faria et al., 1999). The same effect was observed by serially feeding ﬁve times every two hours over one day. Interestingly, the production of TGF-웁 was completely abrogated in splenic lymphocytes from continuously fed OVA–TCR transgenic mice when B cells were depleted in vitro. The total TGF-웁 was not affected by B cell depletion, indicating that these cells are important for the activation of the TGF-웁 produced by other cells. We did not observe active TGF-웁 production by spleen cells from C57BL/6 B cell-deﬁcient mice continuously fed OVA. However, IL-10 production was up-regulated. Thus, our data suggest that continuous feeding of antigen up-regulates the production of TGF-웁 by spleen cells and that the active form of TGF-웁 produced requires the presence of B cells. They also indicate that in the absence of B cells in vivo, other regulatory cytokines such as IL-10 may be induced. The mechanisms underlying the effect of continuous or multiple serial feeding of antigen remain to be determined, but they most probably represent the effect of multiple exposures to low doses antigen. The ability of these regimens to up-regulate regulatory cytokines and to improve the suppressive effect of orally administered antigen may have clinical implications for therapy of autoimmune disease. D. ROLE OF PEYER’S PATCHES Absorption by M cells and lymphocyte activation in Peyer’s patches clearly lead to induction of mucosal immunity. Since the classical experiments of Craig and Cebra (1971), who showed that adoptively transferred Peyer’s patch cell populations could repopulate irradiated mice with IgAproducing cells, we have learned that these lymphoid organs are privileged

198

ANA M. C. FARIA AND HOWARD L. WEINER

sites for the induction of secretory IgA production. It was subsequently shown that this process involves T cells also activated in Peyer’s patches, cytokines such as TGF-웁, IL-4, and IL-5, and the migration of IgAcommitted B cells to the effector sites in the mucosal surfaces (Strober and Ehrhardt, 1994). Several studies have demonstrated that Peyer’s patches are important lymphoid organs for oral tolerance induction. Experiments reported in the 1970s have shown that speciﬁc T cells with suppressor abilities originate in Peyer’s patches after antigen feeding, and that only later do they appear in other tissues (Asherson et al., 1977; Mattingly and Waksman, 1978; Ngan and Kind, 1978; Richman et al., 1978; Thomas et al., 1976). Moreover, adoptive transfer to naive recipients of cells obtained from Peyer’s patches of antigen-fed donor mice rendered the recipients tolerant to subsequent parenteral administration of the antigen (Santos et al., 1994). Additional studies also support the importance previously ascribed to Peyer’s patches in oral tolerance development. Studies in OVA transgenic mice show dramatic differences in percentage of CD4⫹, V웁8.2⫹ (OVA reactive) T cells in Peyer’s patches after feeding these mice increasing doses of OVA. Depending on the dosage and the frequency of feeding, there was either an increase or a decrease in the number of OVA-reactive T cells in the Peyer’s patches. Higher doses induce deletion and lower doses augment the number of T cells (Chen et al., 1994, 1995a). Marth and co-workers also showed induction of IFN-웂 in Peyer’s patches of OVA transgenic mice fed tolerogenic doses of OVA (Marth et al., 1996). Coadministration of anti-IL-12 monoclonal antibodies at the time of antigen feeding suppresses IFN-웂 production and increases TGF-웁 secretion without altering Th2 cytokines in Peyer’s patches of these animals. Gonnella and co-workers (Gonnella et al., 1998) also showed an increased expression of regulatory cytokines such as IL-10, IL-4, and TGF-웁 in Peyer’s patches of OVA transgenic mice fed low and high doses of OVA. In the TNBS-induced colitis model (Duchmann et al., 1996), administration of a haptenating agent per rectum, a route that bypasses the Peyer’s patches, induces a local inﬂammatory response in mice. Feeding of TNBS-haptenated colonic proteins, on the other hand, led to the appearance of TGF-웁-secreting cells in the Peyer’s patches and at the site of inﬂammation, protecting the animal from colitis. The authors explain their results by the different routes followed by the haptenated antigens administered: one reaches the circulation, bypassing the intestine (and the Peyer’s patches), whereas the other gains access to the organism via normal intestinal absorption. However, another interpretation for the data is that the structure of the antigen used in each kind of administration, rectal versus oral, is distinct.

ORAL TOLERANCE: MECHANISMS AND THERAPY

199

Feeding an already haptenized protein may allow its degradation by the digestive enzymes before it reaches the sites of intestinal absorption, whereas haptenization of colonic proteins in situ, by rectal administration of TNBS, would circumvent that enzymatic process, leading to a different handling of the antigen by the APCs present in the gut. Furthermore, although several studies have demonstrated the participation of cells from Peyer’s patches in oral tolerance development and their ability in transferring tolerance, no study has established an absolute requirement for Peyer’s patches for oral tolerance to develop. We studied oral tolerance in TNF/LT웁⫺/⫺ mice that lack Peyer’s patches and mesenteric lymph nodes (Spahn et al., 1999). Oral tolerance to low (0.4 mg) and high (20 mg) doses of OVA could not be induced in these animals but they are readily tolerized by intraperitoneal injection of OVA. TNF⫺/⫺ mice as well as mice treated with anti-TNF/LT웁 antibody could be successfully tolerized by feeding. These results suggest that Peyer’s patches and mesenteric lymph nodes are important for oral tolerance induction. Mice treated in utero with the fusion protein mLT웁–hIgG1 do not develop Peyer’s patches. Feeding these animals high and low doses of OVA when 7 weeks old led to a suppression of DTH, proliferative response, and IL-2 production comparable to normal mice when they were immunized with OVA/complete Freund’s adjuvant. These results indicate that although T cells from Peyer’s patches may play a role in oral tolerance, Peyer’s patches are not essential for oral tolerance induction. Earlier studies have also shown that the presence of Peyer’s patches is not essential for oral tolerance induction. After oral administration of SRBCs to rats that had their Peyer’s patches surgically removed, the rats displayed normal susceptibility to oral tolerance induction (Enders et al., 1986). Experiments using proteins coupled to the subunit B of cholera toxin or to IgG demonstrate that this coupling procedure strongly enhances oral tolerance efﬁciency. Because both CTB and IgG bind to receptors in the gut epithelial cells, the preferential route of antigen entry in this case is directly to the lamina propria of the villi rather than through the M cells of Peyer’s patches. In addition, a number of investigators have reported that a state of unresponsiveness resembling oral tolerance can be induced by injection of antigen into the portal vein, but not into other veins, a route that bypasses the Peyer’s patches and the entire intestinal route (Chung and Gorczynski, 1995; Fujiwara et al., 1986; Gorczynski et al., 1996; Qian et al., 1985; Sato et al., 1988). Thus, it appears that although Peyer’s patches are a site where suppressor mechanisms are triggered, other related mucosal and nonmucosal sites may also be involved in its induction.

200

E. ROLE

ANA M. C. FARIA AND HOWARD L. WEINER

OF THE

LIVER

The majority of blood drained from the intestine circulates through the liver via the portal venous system, and large quantities of antigens enter the liver shortly after feeding. Because the liver contains a major part of the reticuloendothelial system, it may be an important site of processing and presentation events determining the outcome of antigens that gain access to the body by the oral route. Although the immunological functions of the liver are still elusive, it has been known for several years that administration of soluble antigen directly into the portal vein induces a state of systemic hyporesponsiveness similar to oral tolerance in many ways, including prevention of EAE development ( Jewell et al., 1998), preferential suppression of DTH and IFN-웂 responses, and concomitant production of IL-4 and IL-10 (Chung and Gorczynski, 1995; Gorczynski, 1994, Gorczynski et al., 1995, 1996). In addition, tolerance induction by this route can be transferred by T cells expressing the 웂␦ TCR and it seems to involve mechanisms distinct from the ones triggered by other forms of intravenously induced tolerance (Gorczynski, 1994; Gorczynski et al., 1996). Nevertheless, the role of the liver in oral tolerance induction is still unclear. Conﬂicting reports have been based on studies in which the liver is bypassed by a porto-caval shunt (Thomas et al., 1976; Yang et al., 1994), and the local effects of this type of surgery blur the interpretation of the results. Several authors suggest that aggregated large antigens, in contrast to soluble antigens, are absorbed preferentially via M cells in Peyer’s patches and from there they reach the lymphatic system, avoiding the circulation and passage through the liver. This would explain their inability to induce tolerance when fed. The ability of the liver to induce tolerance is supported by reports dating back to Calne et al. (1969). Transplantation of allogeneic liver requires little or no immunosupression, and an allogeneic liver graft imposes alloantigenspeciﬁc tolerance, allowing the transplantation of other organs from the same donor. This form of tolerance is so powerful that it overrides priming, and it is dependent on a population of intrahepatic cells that undergo rapid turnover (Kamada et al., 1981). The ﬁrst report on the importance of the liver for oral tolerance induction was in 1963, when Battisto and Miller injected picryl chloride into mesenteric veins of guinea pigs and produced tolerogenic effects similar to feeding the antigen (Battisto and Miller, 1963). Later, Cantor and Dumont (1967) showed that feeding 1-chloro-2,4-dinitrobenzene to dogs before subcutaneous injection suppresses antibody response to DNCB. In dogs in which portal blood was diverted from the liver by a porto-cava shunt, suppression was abolished.

ORAL TOLERANCE: MECHANISMS AND THERAPY

201

There are several characteristics shared by the intestine and the liver besides their ability to induce tolerance. They both contain a population of nonconventional lymphocytes, 웂␦ T cells, and NK1.1 T cells (Crispe and Mehal, 1996), shown to be involved in suppressive and regulatory functions. They are sites for intrathymic maturation of T lymphocytes and present a prominent population of antigen-presenting cells. Moreover, the liver seems to be a site for extensive apoptosis (Huang et al., 1994). Thus, two types of suppressive mechanisms may be triggered in the liver, leading to tolerance: regulatory cells and cell depletion. Liver transplantation is followed by active lymphocyte inﬁltration of the graft, which peaks at around 2 weeks and subsequently disappears (Kamada et al., 1988). This suggests ingress, followed by apoptosis of a population of alloreactive T cells. In transgenic mice expressing an alloantigen speciﬁcally in the liver, low levels of expression result in the loss of TCR and CD8 from T cells, but a high level of antigen expression results in T cell depletion (Ferber et al., 1994), suggesting that both apoptosis and cell inactivation may play a suppressor role in the liver. It thus appears that the liver plays an important role in the generation of tolerance following oral antigen, given its functional similarities to and anatomical relationship with the intestine. F. ROLE OF THE SPLEEN Although the spleen has not been extensively investigated in oral tolerance, there are reports that splenectomy abrogates oral tolerance in uveitis (Suh et al., 1993). A possible mechanism might involve the generation in the spleen of regulatory cells that are dependent on dendritic cells or macrophages that have acquired antigen in the gut and then traveled to the spleen. This scenario is analogous to the anterior chamber-associated immune deviation (ACAID) phenomenon that occurs when antigen is injected into the anterior chamber of the eye and suppression is dependent on the spleen (Streilein et al., 1993). Both the anterior chamber of the eye and the gut are rich in TGF-웁, and DCs that acquire antigen in the context of TGF-웁 may have special suppressive properties. It is well known that oral tolerance can be transferred to naive recipients by spleen cells. In this regard, a report by Yoshida and co-workers (Yoshida et al., 1998) showed that oral tolerance to 움-S1-casein can be restored in SCID mice reconstituted with either whole splenocytes, T cells, or CD4⫹ cells, but not B cells or CD8 T cells, from normal BALB/c mice. V. Modulation of Oral Tolerance

A number of factors have been reported to modulate oral tolerance. Because oral tolerance has usually been deﬁned in terms of Th1 responses,

202

ANA M. C. FARIA AND HOWARD L. WEINER

anything that suppress Th1 and/or enhances Th2 or Th3 cell development would enhance oral tolerance (Table IV). Th3 cells appear to use IL-4 as one of its growth/differentiation factors (Inobe et al., 1998). Seder et al,. (1998) have also found that IL-4 and TGF-웁 may serve to promote growth of TGF-웁-secreting cells. Thus, IL-4 intraperitoneal administration enhances low-dose oral tolerance to MBP in the EAE model and is associated with increased fecal IgA anti-MBP antibodies (Inobe et al., 1998). Oral IL-10 and IL-4 can also enhace oral tolerance when coadministered with antigen (Slavin et al., 1998) and cytokines have also been administered by the nasal route (Xiao et al., 1998a,b). Large doses of IFN-웂 given intraperitoneally abrogate oral tolerance (Zhang and Michael, 1990), anti-IL-12 enhances oral tolerance and is associated both with increased TGF-웁 production and T cell apotosis (Marth et al., 1996), and subcutaneous administration of IL-12 reverses mucosal tolerance (Claessen et al., 1996). In the uveitis model, intraperitoneal IL-2 potentiates oral tolerance and is associated with increased production of TGF-웁, IL-10, and IL-4 (Rizzo et al., 1994). Oral but not subcutaneous lipopolysaccaride enhances oral tolerance to MBP (Khoury et al., 1990) and is associated with increased expression of IL-4 in the brain. Oral IFN-웁 synergizes with the induction of oral tolerance in SJL /PLJ mice fed low doses of MBP (Nelson et al., 1996) as does oral IFN-␶ (Soos et al., 1999). Cholera toxin (CT) is one of the most potent mucosal adjuvants, and feeding CT abrogates oral tolerance when fed with an unrelated protein antigen (Elson and Ealding, 1984a,b). TABLE IV MODULATION OF ORAL TOLERANCEa Augments IL-2 IL-4 IL-10 Anti-IL-12 Ab TGF-웁 IFN-웁 IFN-␶ CTB F1t3 ligand

Decreases IF-웂 IL-12 CT Anti-MCP-1 Anti-애␦ Ab GVH response Anti-B7.2 mAb (lowdose tolerance) Anti-CTLA4 (high-dose tolerance)

LPS Multiple emulsions a Abbreviations: Ab, antibody; CT, cholera toxin; CTB, choleran toxin B subunit; GVH, graft versus host; IFN, interferon; IL, interleukin; LPS, lipopolysacchride; mAb, monoclonal antibody; MCP, monocyte chemotactic protein 1; TGF, transforming growth factor.

ORAL TOLERANCE: MECHANISMS AND THERAPY

203

However, when a protein is coupled to recombinant cholera toxin B subunit and given orally, there is enhancement of peripheral immune tolerance (Sun et al., 1994). Oral administration of corneal epithelial cells coupled to CTB markedly enhanced the corneal allograft survival (Ma et al., 1997a, 1998). Antibody to chemokine monocyte chemotactic protein 1 (MCP-1) abrogates oral tolerance (Karpus et al., 1998). Oral antigen delivery using a multiple emulsion system also enhances oral tolerance (Elson et al., 1996). 웂␦ T cells may have an important role in oral tolerance induction because it is more difﬁcult to induce oral tolerance in animals depleted of such cells (Ke et al., 1997; Mengel et al., 1995) or in ␦-chain-deﬁcient animals (Spahn and Weiner, 1998). The steroid hormone dehydroepiandrosterone (DHEA) breaks intranasally induced tolerance (Wolvers et al., 1998) and diesel exhaust particles block induction of oral tolerance in mice (Yoshino et al., 1998). In the arthritis model, intraperitoneal administration of TGF-웁 or dimaprid (a histamine type 2 receptor agonist), both of which are believed to promote the development of immunoregulatory cells, enhances the induction of oral tolerance to collagen II even after the onset of arthritis (Thorbecke et al., 1999). VI. Nasal Tolerance

Similar to the gut mucosa, the respiratory tract is continually exposed to a wide variety of antigens. Moreover, the bronchial-associated lymphoid tissue (BALT) is a well-developed mucosal surface in the respiratory tract. A well-developed lymphoid tissue also surrounds the nasal cavity with its own distinctive environment (Hiroi et al., 1998; Kuper et al., 1992). The bronchial mucosa resembles the mucosa of the intestine in several ways. A network of dendritic antigen-presenting cells within the airway epithelium traps inhaled antigen. 웂␦ T cells are also present in the bronchial epithelium and they express CD8 (Holt and Sedgwick, 1987). Presumably some of the antigen that reaches the BALT is processed and presented locally to T cells, thus initiating an immune response. Also, a small degree of antigen leakage across the intact lung occurs, thus providing a direct route for the penetration of antigen into the peripheral blood (Kaltreider, 1976; Sedgwick and Holt, 1983). This may be particularly relevant in relation to encounters with allergens. These proteins are normally small, highly soluble, and have enzymatic activity, and thus may diffuse across the epithelium more readily than larger proteins (Lowrey et al., 1998). Such an event, however, seems to evoke in most circumstances tolerance instead of inﬂammatory allergic responses. The precise mechanisms by which immunological tolerance to daily sampled antigens in the bronchial epithelia is induced are still unclear, although insights are being obtained in this area.

204

ANA M. C. FARIA AND HOWARD L. WEINER

Induction of tolerance to IgE responses to inhaled antigens via the nasal route was ﬁrst reported in 1981 by Holt and co-workers (Holt et al., 1981). This suppression was an active process involving T cells that could transfer suppression. Tolerance developed most rapidly for IgE antibody responses and was followed by suppression of DTH and IgG responses. Suppression was followed by a compensatory rise in IgG2a responses (McMenamin et al., 1994). The surface phenotype of the regulatory cells that mediated the tolerance described in this system was CD8⫹ 웂␦ T cells and they produced IFN-웂 when stimulated in vitro (McMenamin et al., 1995; Sedgwick and Holt, 1983, 1985). In a similar study using nebulized OVA, it was observed that two different populations of CD4⫹ T cells are activated by inhaled antigen. CD4⫹ T cells expressing a V웁8.2 TCR appear to be important for the induction of IgE synthesis, whereas CD4⫹ T cells expressing V웁2 TCR inhibit the production of IgE in vivo (Renz et al., 1993). CD8⫹ T cells are also activated in this model, which can inhibit IgE synthesis when adoptively transferred into naive animals (Renz et al., 1994). Studies in which mice were exposed to various concentrations of aerosolized OVA demonstrated prolonged loss of IgE and eosinophil responsiveness, which did not require CD8⫹ cells, 웂␦ T cells, or IFN-웂 (Seymour et al., 1998). Other ivestigators have reported that Th2 responses are not required to suppress Th1 responses by nasal-induced tolerance (Wolvers et al., 1997). Since these ﬁrst reports on nasal-induced tolerance to soluble proteins, a number of self-proteins have been tested for inhibition of experimental autoimmune disease. Daniel and Wegmann (1996a,b) demonstrated that intranasal administration of 40 애g of B chain peptide 9–23 during 3 days in prediabetic NOD mice resulted in a marked delay in the onset and a decrease in the incidence of diabetes. The protective effect was associated with a reduced T cell proliferative response to the peptide and the secretion of Th2 cytokines. Nasal administration of another autoantigen (GAD65) in the NOD mice at the dose of 200 애g reduced insulinitis, long-term insulin-dependent diabetes mellitus (IDDM), inhibited IFN-웂 secretion, and enhanced the secretion of IL-4 and IL-5 (Tian et al., 1996). Splenic CD4⫹ (but not CD8⫹ ) T cells from GAD65-treated mice inhibited the adoptive transfer of IDDM to NOD-scid/scid mice. These results suggest immune deviation plays a major role in the tolerance induced. Harrison and co-workers (Harrison et al., 1996) reported similar effects of inhaled aerosol insulin in NOD mice. Insulin-treated mice also showed no proliferative responses to B chain peptide 9–23 and up-regulation of IL-4 and IL10 production by spleen cells. In this model, the cells responsible for adoptive transfer of tolerance were CD8⫹ 웂␦ T cells. We have also found that nasal administration of insulin B chain peptide 10–24 suppresses

ORAL TOLERANCE: MECHANISMS AND THERAPY

205

diabetes in NOD mice and instillation of 50 애g MBP by nasal route over 3 days prevents EAE induction (R. Maron and H. L. Weiner, unpublished). Nasal administration of collagen has been shown to suppress arthritis (Myers et al., 1997; Staines et al., 1996). Nasal type II and type IX collagens (CII and CIX) suppressed pristane-induced arthritis (Lu and Holmdahl, 1999). Interestingly, a single dose of CII given prior to disease suppressed whereas three doses given prior to disease exacerbated. We also found suppression of collagen-induced arthritis by nasal administration of 30 애g CII. T cells from mice treated nasally or orally with CII showed a decrease in IFN-웂 production and T cell lines secreted IL-4, IL-10, and TGF웁. Moreover, suppression of CIA by nasal collagen was associated with diminished levels of TNF-움 and IL-6 mRNA in the joints of tolerized mice (Garcia et al., 1999). In the H-2u-mouse model of EAE, inhalation but not oral administration of encephalitogenic peptide inhibited disease induction (Metzler and Wraith, 1993). Oral administration of the encephalitogenic peptide failed to induce oral tolerance to EAE. In contrast, a single intranasal dose of peptide profoundly inhibited EAE induced by subcutaneous injection peptide or a complex mixture of myelin antigens contained in spinal cord homogenate. Inhibition of EAE in rats can also be achieved by nasal administration of a mixture of MBP peptides 68–86 and 87–99 (Liu et al., 1998a), but not of synthetic peptides of acetylcholine receptor (AchR) (Zhang et al., 1998). Nasal administration of AchR in rats induces effective tolerance to experimental autoimmune myasthenia gravis (EAMG), and the dose required is 1/1000 of the amount of antigen used for oral tolerance induction. Coadministration of minute amounts of IFN웂 nasally blocks tolerance induction (Li et al., 1998b). Nasally administered AchR has a protective effect against EAMG even in primed animals (Shi et al., 1998). Although the mechanism of nasal tolerance induction in the EAMG model is still unclear, there are data indicating it involves active suppression. Xiao and co-workers (Xiao et al., 1998b) showed that tolerance to EAMG by nasal administration of AchR is associated with a decrease in LFA-1 expression on CD4⫹ T cells and up-regulation of TGF-웁 production. Decreased expression of LFA-1 may contribute to reduction of the inﬁltration of inﬂammatory CD4⫹ T cells, whereas up-regulated TGF-beta may inhibit lymphocyte functions. The nasal route is also being explored as a route for immunization against infectious agents such as human immunodeﬁciency virus (Imaoka et al., 1998). Although most of the work on nasal tolerance indicates a role of immune deviation in this type of mucosal tolerance, Hoyne and co-workers (Hoyne et al., 1993, 1996, 1997) demonstrated the presence of anergic cells in mice rendered tolerant by nasal route to Der p1 peptide. Naive mice treated intranasally with the immunodominant peptide p111–139 could

206

ANA M. C. FARIA AND HOWARD L. WEINER

be rendered profoundly unresponsive to an immunogenic challenge with the whole Der p1 protein. The suppression lasts for a period longer than 6 months. Lymph node cells from tolerized mice secreted very low levels of IL-2 and proliferated poorly when restimulated in vitro as compared to the untreated control mice. The authors suggest antigen-speciﬁc cells were anergic in this system and they could mediate ‘‘linked suppression’’ to other epitopes in the Der p1 molecule. In summary, nasal administration appears to induce tolerance by many of the same mechanisms as oral antigen, although differences exist and remain to be deﬁned. Of note is that Li and co-workers reported dose-dependent mechanisms related to nasal tolerance induction and protection against EAE in rats. Only low-dose (30 애g) MBP-tolerized rats had high numbers of IL-4 mRNA-expressing lymph node cells, and adoptive transfer revealed that only spleen cells from rats pretreated with a low dose, but not from those pretreated with a high (600 애g) dose, of MBP transferred protection against EAE to naive recipients (Li et al., 1998a). VII. Other Forms of Antigen-Driven Tolerance

Since the description of antigen-driven adult tolerance, many protocols have been used to induce speciﬁc T cell unresponsiveness to foreign protein antigens or peptides, including intraperitoneal injection with (Aichele et al., 1994; Vidard et al., 1994) or without incomplete Freund’s adjuvant (IFA) (Burstein and Abbas, 1993; Romball and Weigle, 1993). Soluble proteins have been reported to be particularly suitable for tolerance induction by parenteral routes. Self-proteins can also induce tolerance through the intravenous route (Gammon and Sercaz, 1989). These studies have provided new information on the similarities and differences in the mechanisms of tolerance induction by parenteral and oral routes. We have found that intravenously administered MBP suppresses EAE in the Lewis rat in an antigen-speciﬁc fashion with no evidence of active suppression. Intravenously administered MBP 21–40 does not suppress MBP 71–90-induced EAE, whereas orally administered MBP 21–40 does (Miller et al., 1993a,b). Because MBP 71–90 is an encephalitogenic peptide, but not MBP 21–40, it implies that intravenous tolerance probably involves anergy or clonal deletion of encephalitogenic T cells, whereas oral tolerance relies on the activation of regulatory T cells that can mediate bystander suppression. Moreover, we and other investigators (Swierkiosz and Swanborg, 1977) have demonstrated that protection against EAE by intravenous injection of MBP could not be adoptively transferred, whereas protection induced by oral MBP could. Similarly, in proteolipid-induced

ORAL TOLERANCE: MECHANISMS AND THERAPY

207

EAE, both intravenous and orally administered PLP suppress disease; however, orally administered MBP suppresses PLP-induced disease and intravenously administered MBP does not (al-Sabbagh et al., 1994). In the diabetes model, in NOD mice, parenteral administration of insulin B chain and B9–23 peptide in IFA has different effects, depending on the route of administration: subcutaneous injection protected (either partially or completely) from IDDM whereas intraperitoneal injection failed to modify the disease (Hutchings, 1998). Differential effects of intravenous versus oral tolerization were also observed in the allograft model. Orally administered alloantigen suppressed inﬂammation in the allograft and prevented accelerated allograft rejection in association with IL-4 expression in the graft. In contrast, intravenously administered alloantigen suppressed graft rejection without intragraft IL-4 expression (Hancock et al., 1993). Moreover, treatment with monoclonal antibodies to CD4 abrogated the oral effect but had no effect on the tolerance induced by spleen cells injected intravenously. Hence, the mechanism mediating the intravenous tolerance in this model was distinct from oral tolerance and more analogous to the form of tolerance seen in tolerance induction in rat allograft recipients by intravenous injections of spleen cells (Dallman et al., 1991; Salom et al., 1992). This supports a role for Th2-type responses in association with oral tolerance and indicates that an active suppressive mechanism is not generally implicated in intravenous tolerance. In the EAE model, IL-4-secreting cells may also play a role following oral tolerization. In the Lewis rat EAE model, orally fed MBP is associated with increased expression of TGF-웁 in the brain without IL-4, whereas MBP plus LPS is associated with IL-4 plus TGF-웁 (Khoury et al., 1992). Additional studies in the Lewis rat EAE model have shown suppression of disease by intrathymic injection of MBP or its peptides (Khoury et al., 1993). Such suppression is associated with a decrease in inﬂammatory cytokines in the brain, but not with expression of IL-4 or TGF-웁, as is seen with oral tolerization. Furthermore, thymic injection of the encephalitogenic peptide 71–90 but not peptide 21–40 suppresses EAE, suggesting that this mechanism of antigen-driven tolerance is different from oral tolerization in that it does not involve up-regulation of antiinﬂammatory cytokines. Wilson and co-workers (Wilson et al., 1998) also reported induction of tolerance to EAE by intrathymic injection of MPB peptide 68–86. Their study suggests that this type of tolerance is not due to inactivation of speciﬁc T cells in the thymus, but rather to an active mechanism of suppression. Singer and Abbas were able to correlate directly the deletion of transgenic T cells in peripheral lymphoid organs after systemic antigen adminis-

208

ANA M. C. FARIA AND HOWARD L. WEINER

tration with the expression of Fas antigen, a TNF family receptor that transduces signals mediating programmed cell death (Singer and Abbas, 1994). However, Elliot and co-workers (Elliott et al., 1996) showed that EAE can be completely prevented by intravenous administration of a fusion protein between the 21.5-kDa isoform of MBP and a genetically engineered form of PLP to H-2s Fas-deﬁcient mice. The absence of disease in the Fas-deﬁcient animals could be accounted for by a Fas-independent pathway of antigen-driven clonal deletion, such as the TNF-mediated induction of apoptosis in Fas-deﬁcient T cell populations. The possibility that other mechanisms such as anergy or immune deviation were involved is not excluded, although most reports on intravenous tolerance suggest that active suppression does not play a major role. Similar results were reported by Gaur and co-workers (Gaur et al., 1992) on intraperitoneal tolerance. They showed prevention of EAE in PL /J mice by intraperitoneal injection of synthetic peptides of immunodominant determinants of MBP emulsiﬁed in IFA. The proliferative responses of lymph node cells of treated mice to MBP was restored by treatment with IL-2, suggesting that anergy was a mechanism involved in the suppression. An interesting question that arises from these experiments on intravenous and intrathymic induced tolerance is their relationship with oral tolerance induction. Absorption of intact proteins from the gut into the circulation occurs physiologically mainly as a result of the activity of special absorptive M cells overlying the Peyer’s patches (Neutra et al., 1996). Once in the circulation, these antigens can reach all organs and tissues, including the thymus, and they can induce tolerance via the same mechanisms triggered by intravenous or intrathymic injected antigen. For antigens administered by oral route at high doses, it is likely that intravenouslike mechanisms of tolerance may exist in addition to the suppressive circuits triggered in the gut. VIII. Treatment of Autoimmune and Inﬂammatory Diseases in Animals

Several studies have demonstrated the effectiveness of mucosally (oral, nasal, aerosol) administered autoantigens in animal models of autoimmune and inﬂammatory diseases. A. EXPERIMENTAL ALLERGIC ENCEPHALOMYELITIS 1. Lewis Rat (MBP) The ﬁrst studies to show that orally administered myelin antigens could suppress EAE were performed in the Lewis rat (Table V). EAE was suppressed by low doses of oral MBP and MBP fragments (Higgins and Weiner, 1988) and by high doses of MBP given in bicarbonate (Bitar and

209

ORAL TOLERANCE: MECHANISMS AND THERAPY

TABLE V MUCOSAL TOLERANCE AND EXPERIMENTAL AUTOIMMUNE ENCEPHALOMYELITISa Study/Results Oral MBP in the Lewis rat Suppression by MBP and MBP fragments Suppression by high doses of MBP CD8⫹ T cells transfer protection Serum IgA/IgG but not IgM suppressed; increased salivary IgA Suppression enhanced by oral LPS Evidence for clonal anergy Suppression of relapsing EAE Homologous MBP most effective TGF-웁 up-regulated and TNF/IFN-웂 down\regulated in CNS MBP-speciﬁc suppressor T cells act via TGF-웁 Different epitopes of MBP given orally trigger TGF-웁 Differential effects of oral vs. intravenous tolerization Oral MBP in neonates enhances EAE in adults Differential effects of MBP and rat 68–88 Suppression of EAE and EAMG by oral MBP ⫹ AchR Reversal of suppressed IgA by IL-4/IL-5 Modulation of spinal cord inﬂammation Effect of oral MBP on V웁8⫹ T cells Oral MBP in the SJL mouse Suppression by recombinant human MBP TGF-웁-secreting regulatory cells in Peyer’s patches Oral MBP suppresses PLP-induced EAE MBP regulatory clones suppress EAE CD4⫹ and CD8⫹ cells mediate active suppression Suppression of chronic relapsing EAE Oral IL-4 enhances protection Oral MBP in TCR transgenic mice Suppression of disease and dose-dependent induction of regulatory cells Multiple mechanisms of tolerance following highdose feeding Oral PLP Suppression of PLP disease with PLP 140–159 Suppression of relapsing–remitting PLP-induced EAE mice with high-dose PLP 139–151 Nasal/aerosol MBP Inhibition of EAE by nasal but not oral MBP peptide Aerosol MBP in the Lewis rat

Ref. Higgins and Weiner (1988) Bitar and Whitacre (1988) Lider et al. (1989) Fuller et al. (1990) Khoury et al. (1990) Whitacre et al. (1991) Brod et al. (1991) Miller et al. (1992a,b) Khoury et al. (1992) Miller et al. (1992a) Miller et al. (1993a,b) Miller et al. (1993a,b) Miller et al. (1994a) Javed et al. (1995) Wang et al. (1995a,b) Kelly and Whitacre (1996) Popovich et al. (1997) Goldman-Brezinski et al. (1998) Oettinger et al. (1993) Santos et al. (1994) al-Sabbagh et al. (1994) Chen et al. (1994) Chen et al. (1995a,b) Meyer et al. (1996a,b) Inobe et al. (1998) Chen et al. (1996) Meyer et al. (1996a,b) al-Sabbagh et al. (1994) Karpus and Lukacs (1996) Metzler and Wraith (1993) al-Sabbagh et al. (1996a) (continues)

210

ANA M. C. FARIA AND HOWARD L. WEINER

TABLE V MUCOSAL TOLERANCE AND EXPERIMENTAL AUTOIMMUNE ENCEPHALOMYELITISa (Continued) Study/Results Nasal MBP suppresses relapsing EAE in DA rats Hierarchy in suppressive properties of nasal myelin antigens Synergistic effect of nasal MBP peptides in the Lewis rat Nasal MBP in protracted–relapsing EAE Oral myelin Suppression of relapsing EAE Long-term (6-month) administration is beneﬁcial Oral MBP is more effective than oral myelin Oral glatiramer acetate Suppression of EAE in both rat and murine models Suppression of EAE in MBP TCR transgenic and SJL mice

Ref. Bai et al. (1997) Anderton and Wraith (1998) Liu et al. (1998a) Bai et al. (1998) Brod et al. (1991) al-Sabbagh et al. (1996b) Benson et al. (1999) Teitelbaum et al. (1998, 1999) Maron et al. (1998b)

a Abbreviations: EAE, experimental autoimmune encephalomyelitis; EAMG, experimental autoimmune myasthenia gravis; IFN, interferon; IL, interleukin; LPS, lipopolysaccharide; MBP, myelin basic protein; PLP, phospholipid; TCR, T cell receptor; TGF, transforming growth factor; TNF, tumor necrosis factor.

Whitacre, 1988). High doses of MBP can suppress EAE via the mechanism of T cell clonal anergy ( Javed et al., 1995), whereas multiple lower doses prevent EAE by transferable active cellular suppression (Miller et al., 1993 a,b). In the nervous system of low-dose-fed animals, inﬂammatory cytokines such as TNF and IFN-웂 are down-regulated and TGF-웁 is upregulated (Khoury et al., 1992). Oral MBP partially suppresses serum antibody responses, especially at higher doses. Administration of myelin to sensitized animals in the chronic guinea pig model or larger doses of MBP in the murine EAE model is protective and does not exacerbate disease (Brod et al., 1991; Meyer et al., 1996a,b) and long-term (6-month) administration of myelin in the chronic EAE model was beneﬁcial (alSabbagh et al., 1996b). 2. Murine Models (MBP/PLP) A number of studies have demonstrated suppression of EAE in murine models (see Table V). Both conventional and T cell receptor transgenic animals have been used and both oral MBP and oral PLP have been administered, although the majority of studies have used MBP. In these models, MBP regulatory clones have been described and such cells have also been induced in MBP T cell receptor transgenic mice. Both CD4

ORAL TOLERANCE: MECHANISMS AND THERAPY

211

and CD8 cells have been shown to mediate active suppression and all mechanisms of tolerance, including anergy and deletion, have also been demonstrated. 3. Nasal/Aerosol MBP MBP given by either the nasal or the oral route has been shown to suppress EAE both in murine and in rat models. In some instances, investigators have reported suppression by nasal but not by oral MBP (Metzler and Wraith, 1993). 4. Glatiramer Acetate The latest approach in animal models has been to utilize glatiramer acetate (Cop1, Copaxone), a drug approved for therapy of multiple sclerosis, which is given to patients by injection. Teitelbaum et al., (1998, 1999) have found that oral glatiramer acetate suppresses EAE in both mouse and rat models and we have found that oral glatiramer acetate suppresses EAE in MBP T cell receptor transgenic animals and induces the up-regulation of TGF-웁 when given orally (Maron et al., 1998b). These effects were not seen in OVA–TCR transgenic mice. Our working hypothesis is that glatiramer acetate is acting as an altered peptide ligand and is immunologically active in the gut (Weiner, 1999). Trials are currently planned for the oral administration of glatiramer acetate in MS patients (discussed below). B. ARTHRITIS MODELS There are several animal models of arthritis, including collagen-induced arthritis, adjuvant arthritis, pristane-induced arthritis, antigen-induced arthritis, silicone-induced arthritis, and streptococcal cell wall arthritis (Table VI). One of the ﬁrst studies to demonstrate that an orally administered autoantigen can suppress an autoimmune disease was the use of oral type II collagen in collagen-induced arthritis (Thompson and Staines, 1986a,b). In addition, oral administration of type II collagen and other antigens have been shown to be effective for suppression of other arthritis models. 1. Collagen-Induced Arthritis Immunization with heterologous or homologous species of type II collagen produces autoimmune responses to CII that lead to development of arthritis in susceptible mouse strains (Courtenay et al., 1980). Collagen-induced arthritis (CIA) has been used as an animal model for rheumatoid arthritis (RA) and is characterized by chronic inﬂammation within the joints, associated with synovitis and erosion of cartilage and bone (Trentham et al., 1977). One of the ﬁrst experiments of oral tolerance using rat CIA was done by Thompson and Staines (1986b). They immunized WA/KIR rats with CII in incomplete

212

ANA M. C. FARIA AND HOWARD L. WEINER

TABLE VI MUCOSAL TOLERANCE AND ARTHRITISa Study/Results Collagen-induced arthritis Oral CII in rat model Oral CII in DBA mice Oral CII and active peripheral suppression Oral human CII peptide Nasal CII Nasal CII and CII peptide in DBA mice Aerosol CII Anti-IL-4 reverses CII oral tolerance Nasal CII Oral recombinant CII Immunogenic nasal CII peptides Adjuvant arthritis CII suppression at lower but not higher doses Oral 65-kDa HSP Nasal HSP60 peptide Pristane-induced arthritis Oral CII Nasal CIX Antigen-induced arthritis Bystander suppression at lower doses of CII Clonal anergy Anti-IL-4 reverses protection by oral antigen Streptococcal cell wall arthritis Oral streptococcal cell wall and CII suppression Silicone-induced arthritis Oral CII Avridine-induced arthritis Nasal HSP60 peptide a

Ref. Thompson and Staines (1986a,b) Nagler-Anderson et al. (1986) Thompson et al. (1993a) Khare et al. (1995) Staines et al. (1996) Myers et al. (1997) al-Sabbagh et al. (1996a) Yoshino (1998) Lu and Holmdahl (1999) Myers et al. (1998) Chu and Londei (1999) Zhang et al. (1990) Haque et al. (1996) Prakken et al. (1997) Thompson et al. (1993b) Lu and Holmdahl (1999) Yoshino et al. (1995a,b) Inada et al. (1997) Yoshino and Yoshino (1998) Chen et al. (1998b) Yoshino (1995b) Prakken et al. (1997)

Abbreviations: CII, CIX, types II and IX collagens; HSP, heat-shock protein.

Freund’s adjuvant following oral administration of 2.5 or 25 애g CII per gram of weight. They found the disease onset was delayed and the severity was reduced at both dosages. At the same time, Nagler-Anderson et al. induced CIA in DBA/1 mice by immunizing with 300 애g CII in complete Freund’s adjuvant (Nagler-Anderson et al., 1986). Oral administration of CII prior to immunization suppressed the incidence of CIA. There was a tendency toward reduced IgG2 responses in the CII-fed mice. Collagen peptides are also capable of inducing CIA. Immunodominant collagen peptides have been used to suppress CIA by oral or nasal administration. Khare et al. induced CIA in DBA/1 mice by immunizing with human CII peptide (250–270) in CFA (Khare et al., 1995). Human peptide

ORAL TOLERANCE: MECHANISMS AND THERAPY

213

CII (250–270)-tolerized mice showed diminished T cell proliferation. Oral tolerance with human peptide CII (250–270) abolished antihuman and antimouse CII Ab, and markedly reduced the disease severity both at early and effector phases. Recombinant CII was shown to be effective in CIA (Meyers et al., 1998). Staines et al. showed CIA induced by whole bovine CII was suppressed by nasal administration of bovine CII peptide (184– 198) (Staines et al., 1996). This nasal peptide administration was found to delay the onset of disease, reduce the severity, and shift the anti-CII antibody response from IgG2b to IgG1 isotype. Other groups have shown that nasally administered CII effectively suppressed CIA in the mouse or rat. Myers et al. reported that nasal administration of either intact CII or synthetic peptide reduced the incidence and severity of arthritis (Myers et al., 1997). They also showed that lymph node and spleen cells from treated mice secreted more IL-4 or IL-10 in response to CII than did nontolerized mice. Chu and Londei (1999) have found differential effects of CII peptides given nasally to suppress CIA. al-Sabbagh et al. found that aerosolization of CII suppressed CIA in Wister/Furth rats to a degree similar to that seen with oral administration (al-Sabbagh et al., 1996a). Recent analysis of cytokine expression in joint tissue of arthritis showed that proinﬂammatory cytokines such as TNF-움, IL-1, IL-6, GM-CSF, and chemokines such as IL-8 are abundant (Feldman et al., 1996). We have found that mice treated with CII nasally showed diminished expression of mRNA of TNF-움 and IL-6 locally (G. Garcia et al., unpublished). We also established cell lines producing antiinﬂammatory cytokines such as IL-4 and IL-10 from mice treated with CII orally or nasally. It was reported that systemic anti-IL-4 treatment during oral administration of CII blocked the suppression of CIA by oral tolerance (Yoshino, 1998). This blockade of suppression of CIA was associated with the blockade of IL-4 secretion, decreases in anti-CII IgG2a Ab, and proliferation of lymphoid cells to CII in CII-fed mice. 2. Adjuvant Arthritis Another major model for RA is adjuvant arthritis (AA), which is a wellcharacterized and fulminant form of experimental arthritis. Oral administration of chicken CII consistently suppressed the development of AA in Lewis rats (Zhang et al., 1990). A decrease in DTH responses to CII was observed in tolerized animals. Of note is that oral type I collagen was also shown to suppress AA. Suppression of AA could be adoptively transferred by T cells from CII-fed animals. Suppression was observed at doses of 3 and 30 애g, but not at 300 or 1000 애g, suggesting that the mechanism involved the generation of suppressive regulatory T cells rather than clonal anergy because active suppression may be lost at higher doses.

214

ANA M. C. FARIA AND HOWARD L. WEINER

Heat-shock proteins (HSPs) play an important role in the AA model (van Eden et al., 1988). It was reported that oral administration of mycobacterial 65-kDa HSP suppressed the development of AA in rats (Haque et al., 1996). Suppression of AA was adoptively transferred by spleen cells from orally tolerized rats. Higher amounts of TGF-웁 were detected in the culture supernatant of spleen cells from HSP-fed rats when cultured with HSP. Prakken et al., showed that HSP peptide (176–190), which includes the arthritogenic epitope, could induce nasal tolerance for AA (Prakken et al., 1997). Of note is that this peptide could also inhibit nonmicrobially induced experimental arthritis (avridine-induced arthritis) (Prakken et al., 1997). 3. Other Arthritis Models Oral tolerance is also effective in pristane-induced arthritis (PIA) (Thompson et al., 1993b). Immunizing mice twice with 2,6,10,14-tetramethylpentadecane (pristane) leads to arthritis after 앑100–200 days. Increasing doses of orally administered CII lowered both the incidence and severity of PIA. Conversely, increasing the dose of intraperitoneally administered pristane worsened PIA. Nasal CIX was also effective in PIA (Lu and Holmdahl, 1999). Yoshino et al., (1995) produced antigen-induced arthritis (AIA) in Lewis rats by systemic immunization with methylated bovine serum albumin (mBSA) in complete Freund’s adjuvant, followed by intraarticular injection of mBSA 2 weeks later. Orally administered CII at lower but not higher doses signiﬁcantly reduced joint swelling, whereas oral keyhole limpet hemocyanin (KLH) had no effect. These results demonstrate the biologic relevance of bystander suppression associated with oral tolerance. It was reported by same group that oral tolerance in AIA might also be mediated by clonal anergy (Inada et al., 1997). A report by Yoshino and Yoshino (1998), however, showed that suppression of AIA by oral administration of bovine serum albumin was associated with IL-4 secretion, and in vivo treatment with monoclonal antibodies to IL-4 abolished the suppression. Other animal models of arthritis that have been successfully treated by mucosal tolerance include streptococcal cell wall arthritis (Chen et al., 1998b), silicone-induced arthritis (Yoshino, 1995b), and avridine-induced arthritis (Prakken et al., 1997). Methotrexate is a widely used drug in rheumatoid arthritis, thus we tested the effect of oral collagen on oral tolerance in animals treated with methotrexate. We previously found that there is a synergistic effect between methotrexate and orally administered antigens such as MBP (al-Sabbagh et al., 1997) and a synergistic effect with oral methotrexate was also observed in the adjuvant arthritis model (Weiner and Komagata, 1998).

215

ORAL TOLERANCE: MECHANISMS AND THERAPY

C. DIABETES Oral insulin has been shown to delay and, in some instances, prevent diabetes in the nonobese diabetic mouse model. Such suppression is transferable (Zhang et al., 1991), primarily with CD4⫹ cells (Bergerot et al., 1994) (Table VII). Immunohistochemistry of pancreatic islets of Langerhans isolated from insulin-fed animals demonstrates decreased insulitis associated with decreased IFN-웂, as well as increased expression of tumor necrosis factor (TNF), IL-4, IL-10, TGF-웁, and prostaglandin E2 (Hancock et al., 1995). It was also reported that nasal administration of the insulin B chain or GAD and aerosol insulin suppresses diabetes in the NOD TABLE VII MUCOSAL TOLERANCE AND DIABETESa Study/Results Oral insulin Suppression of diabetes in the NOD mouse Induction of regulatory CD4⫹ T cells Cytokine shifts in pancreatic islets Active suppression is determined by antigen dose Suppression by insulin and adjuvants Suppression of LCMV-induced diabetes Insulin B chain suppresses diabetes Enhanced protection by cholera toxin–insulin Enhanced protection by bacterial adjuvant Induction of IL-4-secreting regulatory T cells Protection by a plant-based CTB–insulin fusion protein Suppression by mucosally derived Th2 liner Nasal/aerosol insulin Nasal insulin peptide suppresses diabetes Aerosol insulin suppresses diabetes and induces CD8 웂␦ T cells Oral glutamic acid decarboxylase Oral GAD in transgenic plants suppresses diabetes Nasal glutamic acid decarboxylase Nasal GAD suppresses diabetes Transgenic models Oral ovalbumin induces diabetes in OVA transgenic model BB rat Oral insulin is not protective in the BB rat Oral insulin plus adjuvant may exacerbate disease

Ref. Zhang et al. (1991) Bergerot et al. (1994) Hancock et al. (1994) Bergerot et al. (1996) Sai and Rivereau (1996) von Herrath et al. (1996) Polanski et al. (1997) Bergerot et al. (1997) Hartmann et al. (1997) Ploix et al. (1998) Arakawa et al. (1998) Maron et al. (1999) Daniel and Wegmann (1996a,b) Harrison et al. (1996) Ma et al. (1997b) Tian et al. (1996) Blanas et al. (1996) Mordes et al. (1996) Bellman et al. (1998)

a Abbreviations: CTB, cholera toxin B subunit; GAD, glutamic acid decarboxylase; LCMV, lymphocytic choriomeningitis virus; NOD, nonobese diabetic; OVA, ovalbumin.

216

ANA M. C. FARIA AND HOWARD L. WEINER

mouse (Daniel and Wegmann, 1996a,b; Harrison et al., 1996; Tian et al., 1996). Oral insulin suppressed diabetes in a viral-induced model of diabetes in which LCMV was expressed under the insulin promoter and animals infected with LCMV to induce diabetes (von Herrath et al., 1996). Protection was associated with protective cytokine shifts (IL-4/IL-10, TGF-웁) in the islets. It has also been shown that expression of TGF-웁 in the pancreatic islets protects the NOD mouse from diabetes and that TGF-웁 appears to alter the APC preference, polarizing islet antigen responses toward a Th2 phenotype (King et al., 1998). Oral administration of the B chain of insulin, a 30-amino acid peptide, slowed the development of diabetes and prevented disease in some animals (Polanski et al., 1997). This effect was associated with a decrease in IFN-웂 and an increase in IL-4, TGF-웁, and IL-10 expression. Oral dosing with bacterial stimulants such as LPS and E. coli extract OM-89 in NOD mice induces a Th2 shift in the gut cytokine gene expression and, concomitantly, improves diabetes prevention by oral insulin administration (Bellmann, 1997). Oral administration of recombinant GAD from plant sources suppresses the development of diabetes in NOD mice (Ma et al., 1997b), as does oral administration of a plant-based CTB–insulin fusion protein (Arakawa et al., 1998). Of note is that oral insulin is not very effective in the BB rat (Bellman et al., 1998; Mordes et al., 1996), perhaps due to a defect in regulatory cells in these animals. Studies on the pathogenesis of diabetes in NOD mice (Herbelin et al., 1998) showed that in prediabetic mice peripheral CD4⫹ T lymphocytes were highly effective at preventing disease transfer by autoreactive T lymphocytes. These suppressor cells were TCR-움/웁⫹ and CD62L⫹ and they seemed to control peripheral pathogenic autoimmune effectors through an active mechanism. It is possible that oral administration of an autoantigen may potentiate the effect of these cells. Regulatory Th2 type T cell lines against insulin and GAD peptides derived from orally and nasally treated NOD mice suppress diabetes (Maron et al., 1999). Under special experimental conditions, large doses of OVA given to OVA double transgenic mice resulted in diabetes mediated by OVA-speciﬁc CTLs (Blanas et al., 1996). These animals expressed OVA on the islets under the rat insulin promoter and were made chimeric to enrich for OVA-speciﬁc transgenic TCR CTLs. D. UVEITIS Oral administration of S antigen (S-Ag), a retinal autoantigen that induces experimental autoimmune uveitis (EAU), or S-Ag peptides prevents or markedly diminishes the clinical appearance of S-Ag-induced disease as measured by ocular inﬂammation (Nussenblatt et al., 1990) (Table VIII). S-Ag-induced EAU can also be suppressed by feeding an HLA peptide (Wildner and Thurau, 1994). Feeding interphotoreceptor binding protein

217

ORAL TOLERANCE: MECHANISMS AND THERAPY

TABLE VIII MUCOSAL TOLERANCE AND UVEITISa Study/Results S-antigen Suppression of EAU with oral S-Ag Suppression of EAU by a uveitogenic 20-mer peptide Peptide 35 suppresses EAU with suppression of IgA but not IgG or IgM Cross-reactive homolog of S-Ag can suppress EAU Inhibition of EAU by oral S-Ag and peptides Dose dependency of mechanism of oral tolerance to S-Ag peptides Splenectomy abrogates S-Ag-induced EAU Suppression of EAU by recombinant E. coli expressing S-Ag CD8 cells are not essential for low-dose tolerance to S-Ag S-Ag peptides more effective prior to immunization IRBP IL-2 potentiates oral tolerance induction to IRBP IL-4 and IL-10 are required for oral tolerance HLA peptide Feeding HLA peptide suppresses S-Ag-induced disease Other Bystander suppression in EAU occurs in periphery, not in the eye Orally induced peptide-speciﬁc 웂␦ cells suppress EAU EAU induced by oral or nasal HSP-derived peptide

Ref. Nussenblatt et al. (1990) Thurau et al. (1991) Thurau et al. (1991) Singh et al. (1992) Vrabec et al. (1992) Gregerson et al. (1993) Suh et al. (1993) Singh et al. (1996) Vistica et al. (1996) Torseth and Gregerson (1998) Rizzo et al. (1994) Rizzo et al. (1999) Wildner and Thurau (1994) Wildner and Thurau (1995) Wildner et al. (1996) Hu et al. (1998)

a Abbreviations: EAU, experimental autoimmune uveitis; HSP, heat-shock protein; IRBP, interphotoreceptor binding protein.

(IRBP) suppresses IRBP-induced disease and is potentiated by IL-2 (Rizzo et al., 1994). IL-4 and IL-10 are both required for induction of oral tolerance in this model (Rizzo et al., 1999). Oral feeding of retinal antigen not only can prevent acute disease but also can effectively suppress a second attack in chronic-relapsing EAU, demonstrating that oral tolerance may have practical clinical implications in uveitis, which is predominantly a chronic-relapsing condition in humans (Thurau et al., 1997 a,b). Other investigators (Torseth and Gregerson, 1998) have found that oral administration of bovine S-Ag peptides is very efﬁcient in preventing EAU but could only inhibit mild disease if feeding was delayed until after immunization, and relatively high feeding doses were required.

218

ANA M. C. FARIA AND HOWARD L. WEINER

E. MYASTHENIA Although myasthenia gravis is an antibody-mediated disease, oral and nasal (Ma et al., 1995) administration of the Torpedo acetylcholine receptor (AchR) to Lewis rats prevented or delayed the onset of myasthenia gravis (Wang et al., 1993a,b) (Table IX). The levels of anti-AchR antibodies in the serum were lower in orally tolerized animals than in control animals. The effect was dose dependent and large doses of antigen (at least 5 mg of AchR) plus soybean tripsin inhibitor (STI) (Wang et al., 1993 a,b) were required, suggesting that anergy may be the primary mechanism. Puriﬁed AchR was found more effective than an unpuriﬁed mixture (Okumura et al., 1994). Experimental autoimmune myasthenia gravis can also be suppressed by nasally administered AchR (Ma et al., 1995, 1996) and AchR peptides (Karachunski et al., 1997). F. AUTOIMMUNE DISEASES AND INFLAMMATORY CONDITIONS 1. Allergy Both oral and nasal administration of allergens have been shown to be effective ways to suppress IgE responses and intestinal mast cell responses (van Halteren et al., 1997) (Table X). The allergens used include the Der p1 epitope and Dp extract of house mite allergen and pollen extract (Aramaki et al., 1994; Hoyne et al., 1996; Sato et al., 1998). Low and high TABLE IX MUCOSAL TOLERANCE AND MYASTHENIAa Study/Results Oral acetylcholine receptor Suppression of EAMG with oral AchR in Lewis rats Puriﬁed AchR is more protective than unpuriﬁed AchR Induction of IL-4, TGF-웁, and IFN-웂 following oral AchR Oral AchR suppresses AchR-speciﬁc B cell responses Oral immunodominant AchR epitope suppresses T cell responses in mice Nasal acetylcholine receptor Suppression of EAMG by nasal AchR Up-regulation of TGF-웁 and decreased IFN-웂 in tolerized rats Nasal AchR peptides can suppress EAMG Decreased LFA-1 and increased TGF-웁 in clones from nasally tolerized rats Reversal of nasal tolerance by IFN-웂 Synthetic peptides failed to suppress EAMG a

Ref. Wang et al. (1993a,b) Okumura et al. (1994) Wang et al. (1994) Wang et al. (1995a,b) Antozzi et al. (1998) Ma et al. (1995) Ma et al. (1996) Karachunski et al. (1997) Xiao et al. (1998b) Li et al. (1998a) Zhang et al. (1998)

Abbreviations: AchR, acetylcholine receptors; EAMG, experimental autoimmune myasthenia gravis.

219

ORAL TOLERANCE: MECHANISMS AND THERAPY

TABLE X MUCOSAL TREATMENT OF OTHER AUTOIMMUNE DISEASES AND INFLAMMATORY CONDITIONSa Treatment/Results Allergy Sublingual allergen administration suppresses allergen-speciﬁc IgE responses Inhibition of T cell and Ab responses to house mite allergen by inhalation of dominant T cell epitope Der p1 Oral tolerance to pollen extract in BALB/c mice suppresses IgE responses Oral Der p1 cryptic epitopes suppress responses of other epitopes Intranasal Der p1 peptide transiently activates CD4 cells prior to in vivo tolerance Oral and nasal tolerance to ingested allergen suppresses IgE and intestinal mast cell responses Intranasal Der p1 peptides induce T cell tolerance and linked suppression Oral Dp extract inhibits IgE responses Antiphospholipid syndrome TGF-웁-mediated suppression by low-dose oral 웁2-glycoprotein Colitis TGF-웁-dependent suppression of TNBS colitis by oral tolerance Suppression of colitis by orally induced bystander suppression Experimental allergic neuritis Oral bovine peripheral nerve myelin or P2 protein suppresses clinical and histological EAN Gau Nasal administration of P2 57–81 suppresses EAN Immune complex disease Antigen in drinking water modiﬁes antigeninduced immune complex disease Single high-dose antigen suppresses immune complex nephritis Defective oral tolerance promotes experimental IgA nephropathy Multiple autoimmune diseases Oral AchR plus oral MBP suppresses EAMG and EAE Nasal administration of multiple autoantigen antigens suppresses EAMG, EAE, and EAN

Ref. Holt et al. (1988) Hoyne et al. (1993) Aramaki et al. (1994) Hoyne et al. (1994) Hoyne et al. (1996) van Halteren et al. (1997) Hoyne and Lamb (1997) Sato et al. (1998) Blank et al. (1998) Neurath et al. (1996) Groux et al. (1997) Gaupp et al. (1997) Zou et al. (1998) Devey and Bleasdale (1984) Browning and Parrott (1987) Gesualdo et al. (1990) Wang et al. (1995a) Shi et al. (1998) (continues)

220

ANA M. C. FARIA AND HOWARD L. WEINER

TABLE X MUCOSAL TREATMENT OF OTHER AUTOIMMUNE DISEASES AND INFLAMMATORY CONDITIONSa (Continued) Treatment/Results Nickel sensitization Oral tolerance to nickel and chromium in guinea pigs Oral tolerance to nickel in mice Thyroiditis Oral porcine thyroglobulin suppression of murine autoimmune thyroiditis. Oral human thyroglobulin suppresses autoimmune thyroiditis TGF-웁 and IL-4 regulatory cells following oral thyroglobulin Tracheal eosinophilia TGF-웁 induced by oral tolerance ameliorates experimental tracheal eosinophilia High-dose Th1 and Th2 oral tolerance prevents antigen-induced eosinophilia recruitment Prevention of Th2 lung eosinophil inﬂammation by antigen in drinking water Transplantation Allogeneic cells prevent sensitization by skin grafts Oral allopeptides suppress DTH in the Lewis rat Oral but not intravenous alloantigen upregulates intragraft Th2 cells Differential effects of oral vs. intrathymic class I peptide Corneal allograft survival enhanced by oral alloantigen CTB-coupled allogeneic cells enhance survival of corneal grafts Other Oral bacterial extract enhances alveolar macrophage activity Oral adenoviral antigen permits long-term gene expression Oral casein differentially suppresses certain B and T cell epitopes Oral tolerance to milk whey protein

Ref. van Hoogstraten et al. (1992) van Hoogstraten et al. (1993) Peterson and Braley-Mullen (1995) Guimaraes et al. (1995) Guimaraes et al. (1996) Haneda et al. (1997) Nakao et al. (1998) Russo et al. (1998) Sayegh et al. (1992b) Sayegh et al. (1992a) Hancock et al. (1993) Hancock et al. (1994) He et al. (1996) Ma et al. (1997a, 1998) Broug-Holub et al. (1995) IIan et al. (1997) Hachimura et al. (1994), Kim et al. (1993) Enomoto et al. (1993)

a Abbreviations: AchR, acetylcholine receptor; DTH, delayed-type hypersensitivity; EAN, experimental allergic neuritis; TNBS, trinitrobenzene sulforic acid.

ORAL TOLERANCE: MECHANISMS AND THERAPY

221

doses of Dp extract given orally are able to inhibit IgE responses in both naive and sensitized mice (Sato et al., 1998). An increase of anti-IgE autoantibodies in orally tolerant mice may be involved in the modulation of the allergic response. In some instances, linked suppression has been demonstrated in which a single Der p1 peptide suppresses other epitopes (Hoyne and Lamb, 1997). 2. Antiphospholipid Syndrome The antiphospholipid syndrome is characterized by the presence of high titers of IgG anticardiolipin antibodies and /or lupus anticoagulant antibodies. Oral administration in BALB/c mice of low doses of 웁2-glycoprotein prevented the serologic and clinical manifestation of experimental antiphospholipid syndrome on immunization with the autoantigen (Blank et al., 1998). Decreased T cell responses and antibody responses, and increased expression of TGF-웁, which mediated the suppression, were demonstrated. Tolerance was transferred by CD8⫹ class I-restricted TGF-웁-secreting cells. 3. Colitis TGF-웁 appears to play a crucial role in the development of animal models of colitis, including TNBS colitis, colitis in the IL-2-deﬁcient animal model following systemic administration of TNP–KLH in adjuvant, and in the model of Th1 colitis in SCID mice. It has been shown that TNBS colitis can be prevented by oral administration of TNBS, which acts via the induction of TNBS-speciﬁc TGF-웁 responses (Neurath et al., 1996). In addition, colitis can be suppressed by orally administered ovalbumin, which acts via bystander suppression in an OVA-induced colitis model (Groux et al., 1997). 4. Experimental Allergic Neuritis Experimental allergic neuritis (EAN) is the counterpart of EAE and can be suppressed both by nasal and oral administration of peripheral nerve proteins (Gaupp et al., 1997; Zou et al., 1998). 5. Immune Complex Disease Immune complex disease can be suppressed following administration of a single large dose of antigen (Browning and Parrott, 1987) or by placing antigen in drinking water (Devey and Bleasdale, 1984). These studies were performed before oral tolerance was applied to autoimmune diseases. It has also been questioned whether defective oral tolerance may be associated with experimental IgA nephropathy (Gesualdo et al., 1990). 6. Multiple Autoimmune Diseases Link and co-workers have experimented with administering antigens associated with more than one autoimmune disease and immunizing with

222

ANA M. C. FARIA AND HOWARD L. WEINER

a mixture of antigens. They have demonstrated that oral administration of AchR plus MBP suppresses EAMG and EAE and that nasal administration of three autoantigens suppresses EAMG, EAE, and EAN (Shi et al., 1998; Wang et al., 1995a,b). Thus it appears there is no interference by one autoantigen versus another when they are from different target organs. 7. Nickel Sensitization Metals such as nickel and chromium can induce contact sensitivity responses and oral tolerance to nickel has been demonstrated both in the mouse and in the guinea pig (van Hoogstraten et al., 1992, 1993). Based on these animal ﬁndings, oral administration of nickel is being applied to the treatment of allergy in humans (discussed below). 8. Thyroiditis Thyroiditis has been effectively suppressed following oral administration of either porcine or human thyroglobulin (Guimaraes et al., 1995; Peterson and Braley-Mullen, 1995). In the murine model, CD8⫹ regulatory cells that produce IL-4 and TGF-웁 mediated the suppression (Guimaraes et al., 1996). These cells also appear to induce bystander suppression on triggering with the fed antigen. Other investigators (Zhang and Kong, 1998) reported that tolerance induced by intravenous administration of deaggregated thyroglobulin in experimental autoimmune thyroiditis (EAT) is dependent on CD4⫹ T cells but independent of IL-4 and IL-10. 9. Tracheal Eosinophilia Airway inﬂammation plays a major role in human asthma and increasing evidence points to a correlation between eosinophilic inﬁltration and allergic lung disease. In murine models of tracheal eosinophilia, suppression of disease has been achieved by antigen given in high doses (Nakao et al., 1998) or placed in drinking water (Russo et al., 1998). In these instances, both Th1 and Th2 responses were suppressed. The TGF-웁 induced by oral tolerance appears to be the factor that ameliorates the experimental condition (Haneda et al., 1997). 10. Transplantation Oral administration of allogeneic cells prevents sensitization by skin grafts and changes accelerated rejection of vascularized cardiac allografts to an acute form typical of unsensitized recipients (Sayegh et al., 1992a,b). Orally administered allopeptides in the Lewis rat reduce DTH responses to the peptide (Sayegh et al., 1992a). Oral, but not intravenous, alloantigen was accompanied by elevation of intragraft levels of IL-4 (Hancock et al., 1993). Oral alloantigen enhanced corneal allograft survival even in

ORAL TOLERANCE: MECHANISMS AND THERAPY

223

preimmune hosts (Ma et al., 1997a). When orally administered cells are conjugated with CTB and administered before corneal transplantation, the tolerance effect is optimized (Ma et al., 1998). 11. Other Recombinant adenoviruses have been used by many investigators for somatic gene therapy. The duration of transgene expression is limited by the host immune response, which precludes gene expression on readministration of the virus. It has been shown that the immune response can be abrogated by oral tolerization with protein extracts of recombinant adenovirus (Ilan et al., 1997). Lymphocytes from tolerized rats showed increased expression of TGF-웁, IL-2, and IL-4 on exposure to viral antigens, whereas IFN-웂 expression became undetectable. Thus, oral tolerization of adenoviral antigens has a potential to prevent immune responses associated with repeated injections of recombinant adenoviruses. Investigators have studied food allergy using 움S1-casein, a major protein in cow’s milk. Oral tolerance can be induced by feeding such a protein although antibody responses can occur following immunization. Investigators have found differential suppression in terms of B cell and T cell determinants following oral administration of the casein (Hachimura et al., 1994; Kim et al., 1993). Thus, orally administered antigen may not induce tolerance to some portions of the repertoire, which could lead to food hypersensitivity. G. WORSENING OF AUTOIMMUNE DISEASES Although mucosally administered antigens have been used successfully to treat a wide variety of autoimmune and inﬂammatory conditions, under certain experimental conditions worsening of autoimmune diseases in animals by oral antigen has been reported. We have observed that neonatal administration of guinea pig MBP in the Lewis rat does not induce oral tolerance and in fact makes animals more susceptible to EAE induction as adults (Miller et al., 1994a). This raises the possibility that neonatal exposure to antigen may be a factor in the development of autoimmune disease in adulthood. This indeed has been postulated for diabetes in terms of exposure to cow’s milk as a risk factor for development of diabetes. Interestingly, we fed insulin to neonatal NOD mice and did not ﬁnd an enhancement of diabetes, but better protection against the development of diabetes (Maron and Weiner, 1998). We did not observe protection with oral MOG, given to neonates, but it did not exacerbate EAE. Thus worsening of autoimmunity by neonatal exposure may depend on strain of animal and antigen fed. Meyer et al., (1996a,b) reported that a single feeding of a small dose of guinea pig MBP exacerbated the clinical course of disease in the B10.PL

224

ANA M. C. FARIA AND HOWARD L. WEINER

model of chronic relapsing EAE. Larger single doses suppressed the disease. These investigators also found that multiple oral doses of MBP were required to suppress clinical disease once it was established. We and others have found that there may be initial sensitization of Th1 responses when small doses of antigen are fed, followed by suppression as additional doses are given (Chen et al., 1997; Gautam et al., 1990). Thus, in animal models wherein immunization immediately follows feeding, exacerbation may be observed if only one feeding of a small dose is given. Nasal MBP has been associated with a trend of disease worsening in a protracted relapsing EAE model in DA rats (Bai et al., 1998). In diabetes models, Blanas et al., (1996) were able to induce diabetes by orally administering large doses of OVA (30 mg) to OVA double transgenic mice. The animals expressed OVA on the islets under the rat insulin promoter and were made chimeric to enrich for OVA-speciﬁc transgenic TCR CTLs. Although diabetes was induced, it is not clear the degree to which this transgenic model applies to conventional animals. Nonetheless, these studies demonstrate that orally administered antigen is active immunologically when it encounters the GALT and under special circumstances the immune response generated can have a harmful effect. It has been demonstrated that insulin given with an adjuvant may exacerbate diabetes in the BB rat model (Bellman et al., 1998). It has been difﬁcult to protect the BB rat from diabetes (Mordes et al., 1996) by oral administration of insulin, which may relate to defects the animals have in regulatory cell populations and could account for the ability to enhance diabetes by coadministered adjuvant. In results reported to date from human trials in which new-onset diabetics were fed 1 or 10 mg of insulin, there was no evidence of disease worsening (Coutant et al., 1998). There have been two instances in which investigators have reported the induction of autoimmune disease in animals after oral antigen administration. Terato et al. (1996) reported the induction of collagen-induced arthritis in animals fed chicken type II collagen for intervals of 2–3 weeks over a 15-week period or fed collagen plus LPS. Hu et al. (1998) studied uveitis induced by HSP peptide 336–351 in the Lewis rat and were unable to protect against disease by mucosal administration of the peptide prior to immunization. They were, however, able to induce uveitis with the HSP peptide given orally or nasally in doses ranging from 2.5 to 250 애g. The induction of uveitis could be reversed by treatment with anti-CD4 antibody and suppressed by administration of IL-4, suggesting that CD4⫹ Th1-type cells were responsible for the uveitis, although disease was not transferred from mucosally treated rats. These results suggest that in certain instances, peptides may induce Th1-type autoimmune disease. We have observed that it has been difﬁcult to tolerize orally with the immunodominant PLP

ORAL TOLERANCE: MECHANISMS AND THERAPY

225

peptide 139–151 in SJL mice, and in some instances enhancement of disease has been observed when low doses were fed (250 애g) (A. Slavin and H. L. Weiner, unpublished observations). Protection was observed, however, with the whole molecule or other peptide fragments of PLP (121–122) (al-Sabbagh et al., 1994). Oral tolerance to PLP peptide 139–151 and suppression of EAE in SJL mice were obtained by feeding large doses (2 mg) (Karpus et al., 1996). In the Lewis rat we have identiﬁed encephalitogenic and tolerogenic epitopes of guinea pig MBP (Miller et al., 1993a) and found that nonencephalitogenic fragments may be more tolerogenic (Higgins and Weiner, 1988). Careful examination of the work of Metzler and Wraith (1993) shows a trend toward worsening of EAE relapse in PL /J animals fed the wild-type Ac1–11 peptide, though the differences were not statistically signiﬁcant. Taken together, the aforementioned results suggest that there may be peptide fragments of a protein, perhaps the primary immunodominant ones, that have properties resulting in the induction of autoimmunity when given mucosally in appropriate animal strains, something that is not observed when the whole molecule is given. However, this is not true for all peptides, because we have not observed this with the immunodominant peptide of the insulin B chain, MOG, or MBP. In summary, animal data suggest that worsening of autoimmune disease could theoretically occur in humans treated with oral or mucosal antigen. This was a major concern of our group when we initiated pilot trials of oral myelin and oral type II collagen in MS and RA patients. However, no worsening was observed in pilot trials and no worsening has subsequently been observed in 250 MS patients treated with oral bovine myelin or in over 1200 RA patients treated with collagen II, some of whom have been taking these oral preparations for over 3 years. Others have also not reported worsening of RA in patients treated with oral bovine collagen (Sieper et al., 1996). In a small uveitis trial, however, there was a suggestion (not statistically signiﬁcant) that oral administration of a retinal mixture appeared to have worsened disease, whereas the puriﬁed protein appeared to ameliorate the disease (Nussenblatt et al., 1997). Thus it is clear that mucosally administered antigens for the treatment of autoimmune or inﬂammatory conditions must ﬁrst be tested in dose-ranging phase I trials for potential toxic effects and patients must be carefully monitored when larger trials are performed. IX. Treatment of Autoimmune Diseases in Humans

Based on the long history of oral tolerance and the safety of the approach, human trials have been initiated in autoimmune diseases, MS, RA, uveitis, and diabetes (Table XI). These initial trials suggest that there has been no

226

ANA M. C. FARIA AND HOWARD L. WEINER

TABLE XI HUMAN STUDIES OF MUCOSAL TOLERANCEa Study/Results Allergy Oral immunotherapy for birch pollen hay fever Oral encapsulated ragweed extract Double-blind study of encapsulated ragweed Oral immunotherapy for allergic rhinoconjunctivitis HIV patients allergic to sulfonamides desensitized by oral trimethoprim–sulfamethoxazole Double-blind study of oral immunotherapy for respiratory allergy Contact sensitivity Oral and mucosal suppression of contact sensitivity to DNCB Diabetes Preliminary report: Increased 웁-cell function and decreased antiinsulin antibodies after 7.5 mg oral insulin in new-onset diabetics over 20 years old Keyhole limpet hemocyanin Oral KLH suppresses T cell but not B cell responses to immunization Nasal KLH suppresses antibody and DTH responses to immunization Oral KLH decreases KLH-reactive precursor cell frequency Maternal-donor renal allografts Breast-feeding in infancy may beneﬁt maternaldonor renal transplants Multiple sclerosis Oral myelin decreased MRI lesions in DR2⫹ males; no effect on clinical relapse Increased TGF-웁-secreting myelin cells after oral myelin Nickel Reduced nickel allergy following oral nickel contact at early age Sublingual treatment effective for nickel sulﬁte dermatitis Oral desensitization in nickel allergy decreases nickel-speciﬁc T cells Rheumatoid arthritis Oral collagen ameliorates rheumatoid arthritis Oral collagen beneﬁts juvenile RA in open-label trial 20 애g best in double-blind oral dosing trial of type II collagen

Ref. Taudorf et al. (1987) Litwin et al. (1996) Litwin et al. (1997) Taudorf (1992) Kalanadhabhatta et al. (1996) Ariano et al. (1998) Lowney (1968, 1971, 1973, 1974) Coutant et al. (1998)

Husby et al. (1994) Waldo et al. (1994) Matsui et al. (1996) Campbell et al. (1984) Weiner (1997) Fukaura et al. (1996) van Hoogstraten et al. (1991) Morris (1998) Bagot et al. (1995) Trentham et al. (1993) Barnett et al. (1996) Barnett et al. (1998) (continues)

227

ORAL TOLERANCE: MECHANISMS AND THERAPY

TABLE XI HUMAN STUDIES OF MUCOSAL TOLERANCEa (Continued) Study/Results Oral bovine collagen at higher doses without positive effect 60 애g oral collagen best in composite analysis of dosing trials Rhinitis Sublingual therapy with house dust mite beneﬁts patients with perennial rhinitis Thyroid disease Decreased cellular immunity to thyroglobulin in patients receiving oral thyroglobulin Uveitis Oral S-Ag appeared to allow medication taper; retinal mixture appeared to worsen uveitis Oral HLA peptide allowed steroid taper Other Decreased immune responses after oral BSA Oral desensitization in Rh disease Oral spirochetes suppress lymphocyte responses IgA antibody-producing cells after oral Streptococcus mutans Oral type I collagen beneﬁts systemic sclerosis (pilot study)

Ref. Sieper et al. (1996) Weiner and Komagata (1998) Scadding and Brostoff (1986) Lee et al. (1998) Nussenblatt et al. (1997) Thurau et al. (1997a,b) Korenblatt et al. (1968) Gold et al. (1983) Shenker et al. (1984) Czerkinsky et al. (1987) McKown et al. (1997)

a Abbreviations: BSA, bovine serum albumin; DNCB, 1-chloro-2,4-dinitrobenzene; DTH, delayedtype hypersensitivity.

systemic toxicity or exacerbation of disease, although clinical efﬁcacy resulting in an approved drug has not yet been demonstrated. Results in humans, however, have paralleled several aspects of what has been observed in animals. In addition, mucosal tolerance has been applied to other conditions. A. ALLERGY Mucosal tolerance has been used for the treatment of some allergic conditions in humans. In 1987, Taudorf et al. demonstrated that capsules of birch pollen extract were effective in decreasing eye symptoms and allergy scores in conjunctival sensitivity to birch pollen (Taudorf et al., 1987). In 1992, these same investigators had treated allergic rhinoconjunctivitis with oral grass pollen and birch pollen (Taudorf, 1992). Scadding and Brostoff (1986) treated patients with sublingual house dust mite and found a decrease in allergic rhinitis. Litwin et al. (1996, 1997) treated patients with short ragweed extract and found that patients appeared to do better in ragweed season compared to untreated patients. A double-

228

ANA M. C. FARIA AND HOWARD L. WEINER

blind study of oral tolerance in pollen respiratory allergy showed statistically signiﬁcant positive effects in the treated versus the placebo groups (Ariano et al., 1998). B. CONTACT SENSITIVITY Investigators have shown that exposure of a contact-sensitizing agent via the mucosa prior to subsequent skin challenge led to unresponsiveness in a portion of patients studied (Lowney, 1968). C. DIABETES Oral insulin is attractive as a treatment to prevent diabetes because of ease of administration to children and lack of toxicity. Six different trials are currently underway testing mucosally administered recombinant human insulin as a tolerizing agent in type 1 diabetes: 1. A multicenter double-blind study in the United States evaluating oral insulin therapy versus placebo in adults and children with new-onset disease. Preliminary analysis demonstrated preserved 웁 cell function as measured by endogenous C-peptide insulin responses in adults fed 10 mg versus placebo (Coutant et al., 1998). 2. A double-blind study in France to compare oral insulin therapy and parenteral insulin therapy versus placebo in patients during the remission phase. Evaluation criteria include duration of remission, measures of insulin secretion /sensitivity, and immunological parameters. 3. A multicenter double-blind study in Italy to evaluate whether the addition of oral insulin is able to improve the integrated parameters of metabolic control and modify immunological ﬁndings compared to placebo in patients with recent-onset disease treated with intensive insulin therapy (IMDIAB VI). 4. A multicenter double-blind study in the United States to determine if diabetes can be prevented by subcutaneous insulin therapy or oral insulin therapy in subjects at risk for diabetes (DPT-1). 5. A double-blind study in Australia evaluating aerosolized insulin versus placebo in patients with new-onset disease. 6. A double-blind study in Finland evaluating nasally administered insulin versus placebo in patients with new-onset disease. D. KLH KLH administered orally to human subjects has been reported to decrease subsequent cell-mediated immune responses, although antibody responses were not affected (Husby et al., 1994), and to decrease KLHprecursor frequency (Matsui et al., 1996). Nasal KLH has also been reported to induce tolerance in humans (Waldo et al., 1994).

ORAL TOLERANCE: MECHANISMS AND THERAPY

229

E. MATERNAL DONOR ALLOGRAFTS It is known that a large number of maternal lymphocytes are present in breast milk and investigators asked the question whether exposure of an infant to maternal lymphocytes during breast feeding would have an effect on subsequent reactivity of a patient to a maternal donor-related renal transplant. In a study reported in 1984, there was a suggestion that breast-fed recipients may have done better than non-breast-fed recipients (Campbell et al., 1984). Of note is that it has been shown in animal models that transfusion tolerance involving cells may be mediated by TGF-웁 ( Josien et al., 1998). F. MULTIPLE SCLEROSIS In MS patients, MBP- and PLP-speciﬁc TGF-웁-secreting Th3-type cells have been observed in the peripheral blood of patients treated with an oral bovine myelin preparation and not in patients who were untreated (Fukaura et al., 1996). There was no increase in MBP- or PLP-speciﬁc IFN-웂-secreting cells in treated patients. These results demonstrate that it is possible to immunize via the gut for autoantigen-speciﬁc TGF-웁secreting cells in a human autoimmune disease by oral administration of the autoantigen. However, a completed 515-patient, placebo-controlled, double-blind phase III trial of single-dose bovine myelin in relapsing– remitting MS did not show differences between placebo and treated groups in the number of relapses; a large placebo effect was observed (AutoImmune Inc., Lexington, Massachusetts). The dose of myelin was 300 mg given in capsule form and contained 8 mg MBP and 15 mg PLP. Preliminary analysis of magnetic resonance imaging data showed signiﬁcant changes favoring oral myelin in certain patient subgroups. Based on the results of oral tolerance in uveitis in humans (Nussenblatt et al., 1997) and in animal models (Okumura et al., 1994; Benson et al., 1999), it appears that protein mixtures may not be as effective oral tolerogens as puriﬁed proteins. New trials of oral tolerance in MS are being planned with the MBP analog, glatiramer acetate, which is currently given by injection to MS patients and has been shown to be effective orally in animals and to induce regulatory cells that mediate bystander suppression (Teitelbaum et al., 1999; Weiner, 1999). G. NICKEL Oral desensitization to nickel allergy in humans induces a decrease in nickel-speciﬁc T cells and affects cutaneous eczema (Bagot et al., 1995). Sublingual treatment is effective for nickel sulﬁte dermatitis (Morris, 1998) and there is reduced nickel allergy in those exposed to nickel at an early age (van Hoogstraten et al., 1991).

230

ANA M. C. FARIA AND HOWARD L. WEINER

H. RHEUMATOID ARTHRITIS In RA, a 280-patient double-blind phase II dosing trial of type II collagen in liquid doses ranging from 20 to 2500 애g per day for 6 months demonstrated statistically signiﬁcant positive effects in the group treated with the lowest dose (Barnett et al., 1998). Oral administration of larger doses of bovine type II collagen (1–10 mg) did not show a signiﬁcant difference between tested and placebo groups, although a higher prevalence of responders was reported for the groups treated with type II collagen (Sieper et al., 1996). These results are consistent with animal studies of orally administered type II collagen in which protection against adjuvant- and antigen-induced arthritis and bystander suppression was observed only at the lower doses (Yoshino et al., 1995; Zhang et al., 1990). An open-label pilot study of oral collagen in juvenile RA gave positive results with no toxicity (Barnett et al., 1996). This lack of systemic toxicity is an important feature for the clinical use of oral tolerance, especially in children for whom the long-term effects of immunosuppressive drugs are unknown. Five phase II randomized studies of oral type II collagen have been performed, and based on the results obtained, a multicenter double-blind phase III trial study of oral type II collagen (Colloral) is underway (AutoImmune Inc.). In the ﬁve double-blind phase II studies 805 patients were treated with oral type II collagen and 296 patients were treated with placebo. Two of the studies have been published (Barnett et al., 1998; Trentham et al., 1993). The other three studies were included in a integrated analysis that led to the decision to carry out a phase III trial. A dose reﬁnement study tested doses of 5, 20, and 60 애g. Colloral at 60 애g was found to be the most signiﬁcant dose compared to other doses. Weighted averages for the Paulus 20 and Paulus 50 responses were calculated for the 60 애g dose and placebo. A signiﬁcant effect favoring 60 애g was observed for both the Paulus 20 and the Paulus 50 response. Integrated efﬁcacy analyses of predictors of response, including HLA, phenotype, rheumatoid factor, CII antibodies, duration of disease, and tender and swollen joint count, were performed, but no statistically signiﬁcant predictors were identiﬁed. Nonsteroidal antiinﬂammatory drugs (NSAIDs) did not appear to affect the clinical response of RA patients to oral type II collagen. Safety analyses demonstrated that Colloral was extraordinarily safe with no side effects. The magnitude of the clinical responses of Colloral appear to be on the same level as NSAIDs for the majority of patients. However, there is a subgroup of patients that appear to have a more signiﬁcant response to the medication. Based on these data, a 760-patient phase III trial has been initiated comparing 60 애g of Colloral to placebo. The studies to date are encouraging for the potential use of oral collagen in the treatment of rheumatoid arthritis in the future.

ORAL TOLERANCE: MECHANISMS AND THERAPY

231

I. RHINITIS Low-dose sublingual therapy with house dust mite was reported to be effective in relieving symptoms in 72% of a group of patients with perennial rhinitis due to house dust mite in a double-blind placebo-controlled crossover trial (Scadding and Brostoff, 1986). J. THYROID Thirteen patients receiving thyroid hormone replacement with synthetic thyroxin were randomly assigned to receive oral porcine thyroid or remain on synthetic T4 (Lee et al., 1998). Humoral and cellular immune responses were measured over the course of a year. A decrease in cellular immunity to thyroid peptides was observed in the fed versus the control group. No changes between groups were observed in autoantibody levels. K. UVEITIS In uveitis, a pilot trial of S-Ag and an S-Ag mixture was conducted at the National Eye Institute (Bethesda, Maryland) and showed positive trends with oral bovine S-Ag but not the retinal mixture (Nussenblatt et al., 1997). Feeding of peptide derived from a patient’s own HLA antigen appeared to have an effect on uveitis in that patients could discontinue their steroids because of reduced intraocular inﬂammation mediated by oral tolerance (Thurau et al., 1997b). L. OTHER Positive effects were reported in an open-label pilot study of oral type I collagen in patients with systemic sclerosis (McKown et al., 1997). A pilot immunological study of oral major histocompatability peptides has been initiated in transplantation patients in our institution. In terms of immune response to food antigens in humans, Husby and co-workers studied humoral immunity to dietary antigens in healthy adults and found 90% of subjects had antibodies to ovalbumin and 24% had antibodies to 움-lactalbumin (Husby et al., 1985a,b). Peripheral and intestinal lymphocyte activation after in vitro exposure to cow’s milk antigens was observed in normal subjects and in patients with Crohn’s disease (Biancone et al., 1987). In another study, humoral responses to food antigens were reported to be of the IgG subclass (Quinti et al., 1989). Immune reactions induced in infants by intestinal absorption of incompletely digested cow’s milk protein have also been reported (Lippard et al., 1936). Zivny and coworkers reported multiple mechanisms of oral tolerance to food antigens in humans, including anergy and active suppression (Zivny et al., 1996). Of note, a study reported that grapefruit juice increases felodipine oral

232

ANA M. C. FARIA AND HOWARD L. WEINER

availability in humans by decreasing intestinal CYP3A protein expression (Lown et al., 1997). In summary, based on results to date in humans, it appears that the clinical application of oral antigen for the treatment of human conditions will depend on the speciﬁc disease and the nature and dosages of proteins administered, and might beneﬁt from the use of synergists or mucosal adjuvants to enhance biologic effects. Also, recombinant human proteins may be more efﬁcacious than animals proteins (Miller et al., 1992a). X. Future Directions

Although it is clear that oral antigen can suppress autoimmunity and inﬂammatory diseases in animals, much remains to be learned. Cell surface molecules and cytokines associated with inductive events in the gut that generate and modulate oral tolerance are not completely understood. Important areas of investigation include the cytokine milieu, antigen presentation and costimulation requirements, routes of antigen processing, form of the antigen, role of the liver, the effect of oral antigens on antibody and IgE responses and on CTLs, and the role of 웂␦ T cells. As the molecular events associated with the generation and modulation of oral tolerance are better understood, the ability to apply mucosal tolerance successfully for the treatment of human autoimmune and other diseases will be further enhanced. ACKNOWLEDGMENTS This work was supported by NIH grant R01AI43458. Ana M. C. Faria is supported by a scholarship from CAPES, Brazil.

REFERENCES Abbas, A. K., Murphy, K. M., and Sher, A. (1996). Functional diversity of helper T lymphocytes. Nature (London) 383, 787–793. Agren, L. C., Ekman, L., Lowenadler, B., and Lycke, N. Y. (1997). Genetically engineered nontoxic vaccine adjuvant that combines B cell targeting with immunomodulation by cholera toxin A1 subunit. J. Immunol. 158, 3936–3946. Aichele, P., Kyburz, D., Ohashi, P. S., Odermatt, B., Zinkernagel, R. M., Hengartner, H., and Pircher, H. (1994). Peptide-induced T-cell tolerance to prevent autoimmune diabetes in a transgenic mouse model. Proc. Natl. Acad. Sci. U.S.A. 91, 444–448. Allan, C. H., Mendrick, D. L., and Trier, J. S. (1993). Rat intestinal M cells contain acidic endosomal-lysosomal compartments and express class II major histocompatibility complex determinants. Gastroenterology 104, 698–708. al-Sabbagh, A., Miller, A., Santos, L. M. B., and Weiner, H. L. (1994). Antigen-driven tissue-speciﬁc suppression following oral tolerance: Orally administered myelin basic protein suppresses proteolipid induced experimental autoimmune encephalomyelitis in the SJL mouse. Eur. J. Immunol. 24, 2104–2109.

ORAL TOLERANCE: MECHANISMS AND THERAPY

233

al-Sabbagh, A., Nelson, P. A., Akselband, Y., Sobel, R. A., and Weiner, H. L. (1996a). Antigen-driven peripheral immune tolerance: Suppression of experimental autoimmmune encephalomyelitis and collagen-induced arthritis by aerosol administration of myelin basic protein or type II collagen. Cell. Immunol. 171, 111–119. al-Sabbagh, A. M., Goad, E. P., Weiner H. L., and Nelson, P. A. (1996b). Decreased CNS inﬂammation and absence of clinical exacerbation of disease after six months oral administration of bovine myelin in diseased SJL/J mice with chronic relapsing experimental autoimmune encephalomyelitis. J. Neurosci. Res. 45, 424–429. al-Sabbagh, A. M., Garcia, G., Slavin, A. J., Weiner, H. L., and Nelson P. A. (1997). Combination therapy with oral myelin basic protein and oral methotrexate enhances suppression of experimental autoimmune encephalomyelitis. Neurology 48, A421. Anderton, S., and Wraith, D. (1998). Hierarchy in the ability of T-cell epitopes to induce peripheral tolerance to antigens from myelin. Eur. J. Immunol. 28, 1251–2161. Andre, C., Heremans, J. F., Vaerman, J. P., and Cambiaso, C. L. (1975). A mechanism for the induction of immunological tolerance by antigen feeding: Antigen-antibody complexes. J. Exp. Med. 142, 1509–1519. Antozzi, C., Baggi, F., Andreetta, F., Simoncini, O., Longhi, R., Cornelio, F., and Mantegazza, R. (1998). Oral administration of an immunodominant TAChR epitope modulates antigen-speciﬁc T-cell responses in mice. Ann. N.Y. Acad. Sci. 841, 568–571. Apple, R. J., Domen, P. L., Muckerheide, A., and Michael, J. G. (1988). Cationization of proteins antigens. IV. Increased antigen uptake by antigen-presenting cells. J. Immunol. 140, 3290–3295. Arakawa, T., Yu, J., Chong, D. K., Hough, J., Engen, P. C., and Langridge, W. H. (1998). A plant-based cholera toxin B subunit–insulin fusion protein protects against the development of autoimmune diabetes. Nature Biotechnol. 16, 934–938. Aramaki, Y., Fujii, Y., Suda, H., Suzuki, I., Yadomae, T., and Tsuchiya, S. (1994). Induction of oral tolerance after feeding of ragweed pollen extract in mice. Immunol. Lett. 40, 21–25. Ariano, R., Panzani, R. C., and Augeri, G. (1998). Efﬁcacy and safety of oral immunotherapy in respiratory allergy to Parietaria judaica pollen. A double-blind study. J. Invest. Allerg. Clin. Immunol. 8, 155–160. Asherson, G. L., Zembala, M., Perera, M. A. C. C., Mayhew, B., and Thomas, W. R. (1977). Production of immunity and unresponsiveness in the mouse by feeding contact sensitizing agents and the role of suppressor cells in the Peyer’s patches, mesenteric lymph nodes and other lymphoid tissues. Cell. Immunol. 33, 145–155. Astwood, J. D., Leach, J. N., and Fuchs, R. L. (1996). Stability of food allergens to digestion in vitro. Cell. Immunol. 14, 1269–1273. Avrameas, S. (1991). Natural autoantibodies: From ‘horror autotoxicus’ to ‘gnothi seauton.’ Immunol. Today 12, 154–159. Bagot, M., Charue, D., Flechet, M. L., Terki, N., Toma, A., and Revuz, J. (1995). Oral desensitization in nickel allergy induces a decrease in nickel-speciﬁc T-cells. Eur. J. Dermatol. 5, 614–617. Bai, X. F., Shi, F. D., Xiao, B. G., Li, H. L., van der Meide, P. H., and Link, H. (1997). Nasal administration of myelin basic protein prevents relapsing experimental autoimmune encephalomyelitis in DA rats by activating regulatory cells expressing IL-4 and TGF-웁 mRNA. J. Neuroimmunol. 80, 65–75. Bai, X. F., Li, H. L., Shi, F. D., Liu, J. Q., Xiao, B. G., Van der Meide, P. H., and Link, H. (1998). Complexities of applying nasal tolerance induction as a therapy for ongoing relapsing experimental autoimmune encephalomyelitis (EAE) in DA rats. Clin. Exp. Immunol. 111, 205–210.

234

ANA M. C. FARIA AND HOWARD L. WEINER

Bandeira, A., Larsson, E. L., Forni, L., Pereira, P., and Coutinho, A. (1987). In vivo activated splenic T-cells are refractory to interleukin 2 growth in vitro. Eur. J. Immunol. 17, 901–909. Barnett, M. L., Combitchi, D., and Trentham, D. E. (1996). A pilot trial of oral type II collagen in the treatment of juvenile rheumatoid arthritis. Arthritis Rheum. 39, 623–628. Barnett, M. L., Kremer, J. M., St. Clair, E. W., Clegg, D. O., Furst, D., Weisman, M., Fletcher, M. J. F., Lavin, P. T., Finger, E., Morales, A., Le, C. H., and Trentham, D. E. (1998). Treatment of rheumatoid arthritis with oral type II collagen: Results of a multicenter, double-blind, placebo-controlled trial. Arthritis Rheum. 41, 290–297. Barone, K. S., Tolarova, D. D., Ormsby, I., Doetschman, T., and Michael J. G. (1998). Induction of oral tolerance in TGF-beta 1 null mice. J. Immunol. 161, 154–160. Bartholomeusz, R. C. A., Forrest, B. D., Labroov, J. T., Ey, P. L., Pyle, D., Shearman, D. J. C., and Rowley, D. (1990). The serum polymeric IgA antibody response to typhoid vaccination; its relationship to intestinal IgA response. Immunology 69, 190–194. Battisto, J. R., and Miller, J. (1963). Mesenteric vein administration of picryl chloride produces effects similar to oral feeding. Proc. Soc. Exp. Biol. Med. 111, 111–115. Becker, K. J., McCarron, R. M., Ruetzler, C., Laban, O., Sternberg, E., Flanders, K. C., and Hallenbeck, J. M. (1997). Immunologic tolerance to myelin basic protein decreases stroke size after transient focal cerebral ischemia. Proc. Natl. Acad. Sci. U.S.A. 94, 10873– 10878. Bellmann, K., Kolb, H., Hartmann, B., Rothe, H., Rowsell, P., Rastegar, S., Burghardt, K., and Scott, F. W. (1997). Intervention in autoimmune diabetes by targeting the gut immune system. Int. J. Immunopharmacol. 19, 573–577. Bellman, K., Kolb, H., Rastegar, S., Jee, P., and Scott, F. W. (1998). Potential risk of oral insulin with adjuvant for the prevention of Type I diabetes: A protocol effective in NOD mice may exacerbate disease in BB rats. Diabetologia 41, 844–847. Ben-Sasson, S. Z., Le Gros, G., Conrad, E. H., Finkelman, F. D., and Paul, W. E. (1990). IL-4 production by T-cells from naive donors. J. Immunol. 145, 1127–1136. Benson, J. M., Stuckman, S. S., Cox, K. L., Wardrop, R. M., Gienapp, I. E., Cross, A. H., Trotter, J. L., and Whitacre, C. C. (1999). Oral administration of MBP is superior to myelin in suppressing established relapsing experimental autoimmune encephalomyelitis. J. Immunol., in press. Bergerot, J., Fabien, N., Maguer, V., and Thivolet, C. (1994). Oral administration of human insulin to NOD mice generates CD4⫹ T-cells that suppress adoptive transfer of diabetes. J. Autoimmun. 7, 655–663. Bergerot, I., Fabien N., Mayer, A., and Thivolet, C. (1996). Active suppression of diabetes after oral administration of insulin is determined by antigen dosage. Ann. N.Y. Acad. Sci. 778, 362–367. Bergerot, I., Fioix, C., Peterson, J., Moulin, V., Rask, C., Fabien, N., Lindblad, M., Mayer, A., Czerkinsky, C., Holmgren, J., and Thivolet, C. (1997). A cholera toxoid-insulin conjugate as an oral vaccine against spontaneous autoimmune diabetes. Proc. Natl. Acad. Sci. U.S.A. 94, 4610–4614. Bersani Amado, C. A., Barbuto, J. A. M., and Jancar, S. (1991). Comparative study of adjuvant induced arthritis in susceptible and resistant strains of rats. II. Effects of oral administration of BCG and PPD. J. Rheumatol. 17, 738–742. Besredka, A. (1909). De l’anaphylaxie. Sixie´me memoire de l’anaphylaxie lactique. Ann. Inst. Pasteur 23, 166–174. Besredka, A. (1919). De la vaccination contre les e´tats typhoides par la voie buccale. Ann. Inst. Pasteur 33, 882–903. Besredka, A. (1927). ‘‘Local Immunization.’’ Balliere, Tindal-Cox, London.

ORAL TOLERANCE: MECHANISMS AND THERAPY

235

Biancone, L., Paganelli, R., Fais, S., Squarcia, O., D’Ofﬁzi, G., and Pallaone, F. (1987). Peripheral and intestinal lymphocyte activation after in vitro exposure to cow’s milk antigens in normal subjects and in patients with Crohn’s disease. Clin. Immunol. Immunopathol. 45, 491–498. Bitar, D. M., and Whitacre, C. C. (1988). Suppression of experimental autoimmune encephalomyelitis by the oral administration of myelin basic protein. Cell. Immunol. 112, 364–370. Blanas, E., Carbone, F. R., Allison, J., Miller, J. F. A. P., and Heath, W. R. (1996). Induction of autoimmune diabetes by oral administration of autoantigen. Science 274, 1707–1709. Bland, P. W., and Kambarage, D. M. (1991). Antigen handling by the epithelium and lamina propria macrophages. Gastroenterol. Clin. North Am. 20, 577–596. Bland, P. W., and Warren, L. G. (1986). Antigen presentation by epithelial cells of the rat small intestine. II. Selective induction of suppressor T-cells. Immunology 58, 9–14. Blank, M., George, J., Barak, V., Tincani, A., Koike, T., Shoenfeld, Y. (1998). Oral tolerance to low dose B2-Glycoprotein. I: Immunomodulation of experimental antiphospholipid syndrome. J. Immunol. 161, 5303–5312. Bloch, K. J., Perry, R., Bloch, M., and Walker, A. W. (1983). Induction of (partial) systemic tolerance in primed rats subjected to prolonged oral administration of antigen. Ann. N.Y. Acad. Sci. 409, 787–788. Bloom, B. R., Modlin, R. L., and Salgame, P. (1992). Stigma variations: Observations on suppressor T-cells and leprosy. Annu. Rev. Immunol. 10, 453–488. Border, W. A., Ward, H. J., Kamil, E. S., and Cohen, A. H. (1982). Induction of membranous nephropathy in rabbits by administration of an exogenous cationic antigen. J. Clin. Invest. 69, 451–458. Brandtzaeg, P. (1998). Development and basic mechanisms of human gut immunity. Nutr. Rev. 56, S5–18. Brod, S. A., al-Sabbagh, A., Sobel, R. A., Haﬂer, D. A., and Weiner, H. L. (1991). Suppression of experimental autoimmune encephalomyelitis by oral administration of myelin antigens. IV. Suppression of chronic relapsing disease in the Lewis rat and strain 13 guinea pig. Ann. Neurol. 29, 615–622. Broug-Holub, E., Persoons, J. H., Schornagel, K., and Kraal, G. (1995). Changes in cytokine and nitric oxide secretion by rat alveolar macrophages after oral administration of bacterial extracts. Clin. Exp. Immunol. 101, 302–307. Browning, M. J., and Parrott, D. M. (1987). Protection from chronic immune complex nephritis by a single dose of antigen administered by the intragastric route. Adv. Exp. Med. Biol. 216B, 1619–1625. Bruce, M. G., and Ferguson, A. (1986a). The inﬂuence of intestinal processing on the immunogenicity and molecular size of absorbed, circulating ovalbumin in mice. Immunology, 59, 295–300. Bruce, M. G., and Ferguson, A. (1986b). Oral tolerance to ovalbumin in mice: Studies of chemically modiﬁed and ‘biologically ﬁltered’ antigen. Immunology 57, 627–630. Bruce, M. G., and Ferguson, A. (1987). Oral tolerance induced by gut-processed antigen. Adv. Exp. Med. Biol. 216A, 721–731. Burnet, M. (1959). ‘‘The Clonal Selection Theory of Acquired Immunity.’’ Vanderbilt Univ. Press, Nashville, Tennessee. Burstein, H. J., and Abbas, A. K. (1993). In vivo role of interleukin 4 in T-cell tolerance induced by aqueous protein antigen. J. Exp. Med. 177, 457–463. Calne, R. Y., Sells, R. A., and Pena, J. R. (1969). Nature (London) 223, 472–476. Campbell, D. A., Jr., Lorber, M. I., Sweeton, J. C., Turcotte, J. G., Niederhuber, J. E., and Beer, A. E. (1984). Breast feeding and maternal-donor renal allografts. Possibly the original donor-speciﬁc transfusion. Transplantation 37, 340–344.

236

ANA M. C. FARIA AND HOWARD L. WEINER

Cantor, H. M., and Dumont, A. E. (1967). Hepatic suppression of sensitization to antigen absorbed into the portal system. Nature, (London) 215, 744–745. Cardillo, F., Nomizo, A., and Mengel, J. (1998). The role of the thymus in modulating 웂␦ T-cell suppressor activity during experimental Trypanosoma cruzi infection. Int. Immunol. 10, 107–116. Carvalho, C. R., Verdolin, B. A., de Souza, A. V., and Vaz, N. M. (1994). Indirect effects of oral tolerance in mice. Scand. J. Immunol. 39, 533–538. Carvalho, C. R., Verdolin, B. A., and Vaz, N. M. (1997). Indirect effects of oral tolerance cannot be ascribed to bystander suppression. Scand. J. Immunol. 45, 276–281. Cauley, L. S., Cauley, K. A., Shub, F., Huston, G., and Swain, S. L. (1997). Transferable anergy: Superantigen treatment induces CD4⫹ T-cell tolerance that is reversible and requires CD–CD8 cells and IFN-웂. J. Exp. Med. 186, 71–81. Cerf-Bensussan, N., Quaroni, A., Kurnick, J., and Bhan, A. (1984). Intraepithelial lymphocytes modulate 1a expression by intestinal epithelial cells. J. Immunol. 132, 2244–2251. Challacombe, S. J., and Tomasi, T. B. (1980). Systemic tolerance and secretory immunity after oral immunization. J. Exp. Med. 152, 1459–1472. Challacombe, S. J., Rahman, D., Jeffrey, H., Davis, S. S., and O’Hagan, D. T. (1992). Enhanced secretory IgA and systemic IgG antibody responses after oral immunization with biodegradable microparticles containing antigen. Immunology 76, 164–168. Chase, M. (1946). Inhibition of experimental drug allergy by prior feeding of the sensitizing agent. Proc. Soc. Exp. Biol. Med. 61, 257–259. Chen, Y., Kuchroo, V. K., Inobe, J.-I., Haﬂer, D. A., and Weiner, H. L. (1994). Regulatory T-cell clones induced by oral tolerance: Suppression of autoimmune encephalomyelitis. Science 265, 1237–1240. Chen, Y., Inobe, J., Marks, R., Gonnella, P., Kuchroo, V. K., and Weiner, H. L. (1995a). Peripheral deletion of antigen-reactive T-cells in oral tolerance. Nature (London) 376, 177–180. Chen, Y., Inobe, J., and Weiner, H. L. (1995b). Induction of oral tolerance to myelin basic protein in CD8-depleted mice: Both CD4⫹ and CD8⫹ cells mediate active suppression. J. Immunol. 155, 910–916. Chen, Y., Inobe, J., Kuchroo, V. K., Baron, J. L., Janeway, C. A., Jr., and Weiner, H. L. (1996). Oral tolerance in myelin basic protein T-cell receptor transgenic mice: Suppression of autoimmune encephalomyelitis and dose-dependent induction of regulatory cells. Proc. Natl. Acad. Sci. U.S.A. 93, 388–391. Chen, Y., Inobe, J.-I., and Weiner, H. L. (1997). Inductive events in oral tolerance in the TCR transgenic adoptive transfer model. Cell. Immunol. 178, 62–68. Chen, L. Z., Hochwald, G. M., Huang, C., Dakin, G., Tao, H., Cheng, C., Simmons, W. J., Dranoff, G., and Thorbecke, G. J. (1998a). Gene therapy in allergic encephalomyelitis using myelin basic protein-speciﬁc T-cells engineered to express latent TGF-웁1. Proc. Natl. Acad. Sci. U.S.A. 95, 12516–12521. Chen, W., Jin, W., Cook, M., Weiner, W. L., and Wahl, S. M. (1998b). Oral delivery of group A streptococcal cells walls augments circulating TGF-웁 and suppresses streptococcal cell wall arthritis. J. Immunol. 161, 6297–6304. Chen, W., Jin, W., and Wahl, S. (1998c). Engagement of cytoxic T lymphocyte-associated antigen 4 (CTLA-4) induces transforming growth factor beta (TGF-웁) production by murine CD4⫹ T cells. J. Exp. Med. 188, 1849–1857. Chesnut, R. W., and Grey, H. M. (1981). Studies on the capacity of B cells to serve as antigen presenting cells. J. Immunol. 126, 1075–1079. Chesnut, R. W., Colon, S. M., and Grey, H. M. (1982). Antigen presentation by normal B cells, B cell tumors, and macrophages: Functional and biochemical comparison. J. Immunol. 128, 1764–1768.

ORAL TOLERANCE: MECHANISMS AND THERAPY

237

Chu, C. Q., and Londei, M. (1999). Differential activities of immunogenic collagen type II peptides in the induction of nasal tolerance to collagen-induced arthritis. J. Autoimmun. 12(1), 35–42. Chung, S. and Gorczynski, R. (1995). Induction of T-cell hyporesponsiveness by intrahepatic modulation of donor antigen-presenting cells. Immunology 85, 582–590. Claessen, A. M., von Blomberg, B. M., De Groot, J., Wolvers, D. A., Kraal, G., and Scheper, R. J. (1996). Reversal of mucosal tolerance by subcutaneous administration of interleukin12 at the site of attempted sensitization. Immunology 88, 363–367. Clark, S. L. (1979). The ingestion of proteins and colloidal materials by columnar absorptive cells in the small intestine in suckling rats and mice. J. Biophys. Biochem. Cytol. 5, 41–49. Cohen, I., and Young, D. B. (1991). Autoimmunity, microbial immunity and the immunological homunculus. Immunol. Today 12, 105–110. Conde, A. A., Stransky, B., Faria, A. M., and Vaz, N. M. (1998). Interruption of recently induced immune responses by oral administration of antigen. Braz. J. Med. Biol. Res. 31, 377–380. Cook, P. M., and Olson, G. F. (1979). Ingested mineral ﬁbers: Elimination in human urine. Science 204, 195–199. Courtenay, J. S., Dallman, M. S., Dayan, A. D., Martin, A., and Mosedale, B. (1980). Immunization against heterologous type II collagen induces arthritis in mice. Nature (London) 283, 666–668. Coutant, R., Zeidler, A., Rappaport, R., Schatz, D., Schwartz, S., Raskin, P., Rogers, D., Bode, B., Crockett, S., Marks, J., Deeb, L., Chalew, S., and MacLaren, N. (1998). Oral insulin therapy in newly-diagnosed immune mediated (type I) diabetes. Preliminary analysis of a randomized double blind placebo controlled study. Diabetes 47 (Suppl. 1), A97. Cowdery, J. S., and Johlin, B. J. (1984). Regulation of the primary in vitro response to TNP-polymerized ovalbumin by T suppressor cells induced by ovalbumin feeding. J. Immunol. 132, 2783–2789. Craig, S. W., and Cebra, J. J. (1971). Peyer’s patches: An enriched source of precursors for IgA-producing immunocytes in the rabbit. J. Exp. Med. 134, 188–200. Crispe, N., and Mehal, W. (1996). Strange brew: T-cells in the liver. Immunol. Today 17, 522–525. Cross, A. H., Tuohy, V. K., and Raine, C. S. (1993). Development of reactivity to new myelin antigens during chronic relapsing autoimmune demyelination. Cell. Immunol. 146, 261–270. Czerkinsky, C., and Holmgren, J. (1995). The mucosal immune system and prospects for anti-infectious and anti-inﬂammatory vaccines. Immunologist 3, 97–193. Czerkinsky, C., Prince, S. J., Michalek, S. M., Jackson, S., Russell, M. W., Moldoveanu, Z., McGhee, J. R., and Mestecky, J. (1987). IgA antibody-producing cells in peripheral blood after antigen ingestion: Evidence for a common mucosal immune system in humans. Proc. Natl. Acad. Sci. U.S.A. 84, 2449–2453. Czerkinsky, C., Sun, J. B., Lebens, M., Li, B. L., Rask, C., Lindblad, M., and Holmgren, J. (1996). Cholera toxin B subunit as transmucosal carrier-delivery and immunomodulating system for induction of antiinfectious and antipathological immunity. Ann. N.Y. Acad. Sci. 13, 185–193. Czerkinsky, C., Sun, J., and Holmgren, J. (1999). Oral tolerance and anti-pathological vaccines. Curr. Top. Microbiol. Immunol. 236, 79–91. Dahl, G. M., Telemo, E., Westrom, B. R., Jakobsson, I., Lindberg, T., and Karlsson, B. W. (1984). The passage of orally-fed proteins from mother to foetus in the rat. Comp. Biochem. Physiol. A. 77, 199–201.

238

ANA M. C. FARIA AND HOWARD L. WEINER

Dahlman, A., Ahlstedt, S., Hanson, L. A., Telemo, E., Wold, A. E., and Dahlgren, U. I. (1992). Induction of IgE antibodies and T-cell reactivity to ovalbumin in rats colonized with Escherichia coli genetically manipulated to produce ovalbumin. Immunology 76, 225–228. Dahlman, A., Telemo, E., Ahlstedt, S., Hanson, L. A., Wold, A. E., and Dahlgren, U. I. (1994). Immune response against ovalbumin in rats colonized with an ovalbumin-producing Escherichia coli and the inﬂeunce of feeding ovalbumin. Int. Arch. Allergy Immunol. 105, 381–385. Dakin, R. (1829). Remarks on a cutaneous affection produced by certain poisonous vegetables. Am. J. Med. Sci. 4, 98–100. Dallman, M. J., Shiho, O., Page, T. H., Wood, K. J., and Morris, P. J. (1991). Peripheral tolerance to alloantigen results from altered regulation of the interleukin 2 pathway. J. Exp. Med. 173, 79–87. Daniel, D., and Wegmann, D. R. (1996a). Intranasal administration of insulin peptide B:9–23 protects NOD mice from diabetes. Ann. N.Y. Acad. Sci. 778, 371–372. Daniel, D., and Wegmann, D. R. (1996b). Protection of nonobese diabetic mice from diabetics by intranasal or subcutaneous administration of insulin peptide B-(9–23). Proc. Natl. Acad. Sci. U.S.A. 93, 956–960. Das, S., and Leskowitz, S. (1974). The cellular basis for tolerance or immunity to bovine gamma globulin in mice. J. Immunol. 112, 107–114. Das, S., Di Padua, D., and Leskowitz, S. (1973). Tolerance induction with aggregated bovine gamma globulin in C57BL/6 mice. Proc. Soc. Exp. Biol. Med. 144, 373–375. Daynes, R., Araneo, B., Dowell, T., Huang, K., and Dudley, D. (1990). Regulation of murine lymphokine production in vivo. III. The lymphoid tissue microenvironment exerts regulatory inﬂuences over T helper cell function. J. Exp. Med. 171, 979–996. Defrance, T., Vanbervliet, B., Brie`re, F., Durand, I., Rousset, F., and Banchereau, J. (1992). Interleukin 10 and transforming growth factor 웁 cooperate to induce anti-CD40-activated naive human B cells to secrete immunoglobulin A. J. Exp. Med. 175, 671–682. DeKruyff, R. H., Rinnooy Kan, E. A., Weksler, M. E., and Siskind, G. W. (1980). Effect of aging on T-cell tolerance induction. Cell. Immunol. 56, 58–67. DeSilva, D. R., Urdahl, K. B., and Jenkins, M. K. (1991). Clonal anergy is induced in vitro by T-cell receptor occupancy in the absence of proliferation. J. Immunol. 147, 3261–3267. De Smedt, T., Van Mechelen, M., De Becker, G., Urbain, J., Leo, O., and Moser, M. (1997). Effect of interleukin-10 on dendritic cell maturation and function. Eur. J. Immunol. 27, 1229–1235. Desvignes, C., Bour, H., Nicolas, J. F., and Kaiserlian, D. (1996). Lack of oral tolerance but oral priming for contact sensitivity to dinitroﬂuorobenzene in major histocompatibility complex class II-deﬁcient mice and in CD4⫹ T-cell-depleted mice. Eur. J. Immunol., 26, 1756–1761. De Vries, J. E. (1995). Immunosuppressive and anti-inﬂammatory properties of interleukin 10. Ann. Med. 27, 537–547. Devey, M. E., and Bleasdale, K. (1984). Antigen feeding modiﬁes the course of antigeninduced immune complex disease. Clin. Exp. Immunol. 56, 637–644. Dighiero, G., Lim, A., Poncet, P., Kaushik, A., Ge, X. R., and Mazie, J. C. (1987). Agerelated natural antibody speciﬁcities among hybridoma clones originating from NZB spleen. Immunology 62, 341–347. Doken, J., Weksler, M. E., and Siskind, G. W. (1980). Effect of age on ease of B-cell tolerance induction. Cell. Immunol. 55, 66–73. Domen, P. L., Muckerheide, A., and Michael, J. G. (1987). Cationization of protein antigens. III. Abrogation of oral tolerance. J. Immunol. 139, 3195–3198.

ORAL TOLERANCE: MECHANISMS AND THERAPY

239

Donnenberg, M. S., and Kaper, J. B. (1992). Enteropathogenic Escherichia coli. Infect. Immun. 60, 3953–3961. Duchmann, R., Schmitt, E., Knolle, P., Meyer zum Buschenfelde, K. H., and Neurath, M. (1996). Tolerance towards resident intestinal ﬂora in mice is abrogated in experimental colitis and restored by treatment with interleukin-10 or antibodies to interleukin-12. Eur. J. Immunol. 26, 934–938. Elliott, E. A., McFarland, H. I., Nye, S. H., Coﬁell, R., Wilson, T. M., Wilkins, J. A., Squinto, S. P., Matis, L. A., and Mueller, J. P. (1996). Treatment of experimental encephalomyelitis with a novel chimeric fusion protein of myelin basic protein and proteolipid protein. J. Clin. Invest. 98, 1602–1612. Elson, C. O., and Ealding, W. (1984a). Cholera toxin feeding did not induce oral tolerance in mice and abrogated oral tolerance to an unrelated protein antigen. J. Immunol. 133, 2892–2897. Elson, C. O., and Ealding, W. (1984b). Generalized systemic and mucosal immunity in mice after mucosal stimulation with cholera toxin. J. Immunol. 132, 2736–2741. Elson, C. O., Tomasi, M., Dertzbaugh, M. T., Thaggard, G., Hunter, R., and Weaver, C. (1996). Oral antigen delivery by way of a multiple emulsion system enhances oral tolerance. Ann. N.Y. Acad. Sci. 778, 156–162. Enders, G., Gottwald, T., and Brendel, W. (1986). Induction of oral tolerance in rats without Peyer’s patches. Immunology 58, 311–314. Enomoto, A., Konishi, M., Hachimura, S., and Kaminogawa, S. (1993). Milk whey protein fed as a constituent of the diet induced both oral tolerance and a systemic humoral response, while heat-denatured whey protein induced only oral tolerance. Clin. Immunol. Immunopathol. 66, 136–142. Everson, M. P., Lemak, D. G., McGhee, J. R., and Beagley, K. W. (1997). FACS-sorted spleen and Peyer’s patch dendritic cells induce different responses in Th0 clones. Adv. Exp. Med. Biol. 417, 357–362. Falcone, M., and Bloom, B. R. (1997). A T helper cell 2 (Th2) immune response against non-self antigens modiﬁes the cytokine proﬁle of autoimmune T-cells and protects against experimental allergic encephalomyelitis. J. Exp. Med. 185, 901–907. Faria, A. M. C., Maron, R., Komagata, Y., and Weiner, H. L. (1999). Continuous feeding of low dose ovalbumin enhances TGF-beta production by spleen cells of Balb/c, C57BL/6, and OVA TCR transgenic mice. FASEB J. 13(4), A608. Faria, A. M. C., Ficker, S. M., Speziali, E., Menezes, J. S., Stransky, B., Verdolin, B. A., Lahmann, W. M., Rodrigues, V. S., and Vaz, N. M. (1998b). Aging and immunoglobulin isotype patterns in oral tolerance. Braz. J. Med. Biol. Res. 31, 35–48. Faria, A. M., Ficker, S. M. Speziali, E., Menezes, J. S., Stransky B., Silva Rodrigues, V., and Vaz, N. M. (1998a). Aging affects oral tolerance induction but not its maintenance in mice. Mech. Ageing Dev. 102, 67–80. Faria, A. M., Garcia, G., Rios, M. J., Michalaros, C. L., and Vaz, N. M. (1993). Decrease in susceptibility to oral tolerance induction and occurrence of oral immunization to ovalbumin in 20–38-week-old mice. The effect of interval between oral exposures and rate of antigen intake in the oral immunization. Immunology 78, 147–151. Feldman, M., Brennan, F. M., and Maini, R. N. (1996). Role of cytokines in rheumatoid arthritis. Annu. Rev. Immunol. 14, 397–440. Ferber, I., Schonrich, G., Schenke. J., Mellor, A. L., Hammerling, G., and Arnold, B. (1994). Science 263, 674–676. Fishman-Lobell, J., Friedman, A., and Weiner, H. L. (1994). Different kinetic patterns of cytokine gene expression in vivo in orally tolerant mice. Eur. J. Immunol. 24, 2720–2724.

240

ANA M. C. FARIA AND HOWARD L. WEINER

Fitzpatrick, J. L., Bailey, M., Harbour, D. A., and Stokes, C. R. (1992). Comparison of antibody and cell-mediated immune responses in horses following feeding of a novel dietary antigen, ovalbumin, and rotavirus. Vet. Immunol. Immunopathol. 34, 245–257. Foo, M. C., and Lee, A. (1972). Immunological response of mice to members of the autochthonous intestinal microﬂora. Infect. Immun. 6, 525–532. Freeman, G. J., Boussiotis, V. A., Anumanthan, A., Bernstein, G. M. Ke, K.-Y., Rennert, P. D., Gray, G. S., Gribben, J. G., and Nadler, L. M. (1995). B7-1 and B7-2 do not deliver identical costimulatory signals, since B7-2 but not B7-1 preferentially costimulates the initial production of IL-4. Immunity 2, 523–532. Friedman, A., and Weiner, H. (1994). Induction of anergy or active suppression following oral tolerance is determined by antigen dosage. Proc. Natl. Acad. Sci. U.S.A. 91, 6688–6692. Fujihashi, K., Taguchi, T., Aicher, W. K., McGhee, J. R., Bluestone, J. A., Eldridge, J. H., and Kiyono, H. (1992). Immunoregulatory functions for murine intraepithelial lymphocytes: 웂/␦ T-cell receptor-positive (TCR⫹ ) T-cell abrogate oral tolerance while 움/웁 TCR⫹ Tcells provide B cell help. J. Exp. Med. 175, 695–707. Fujihashi, K., McGhee, J. R., Yamamoto, M., Hiroi, T. and Kiyono, H. (1996). Role of gamma delta T-cells in the regulation of mucosal IgA response and oral tolerance. Ann. N.Y. Acad. Sci. 778, 55–63. Fujiwara, H., Qian, J. H., Satoh, S., Kokudo, S., Ikegami, R., and Hamaoka, T. (1996). Studies on the induction of tolerance to alloantigens. II. The generation of serum factor(s) able to transfer alloantigen-speciﬁc tolerance for delayed-type hypersensitivity by portal venous inoculation with allogeneic cells. J. Immunol. 136, 2763–2768. Fukaura, H., Kent, S. C., Pietrusewicz, M. J., Khoury, S. J., Weiner, H. L., and Haﬂer, D. A. (1996). Induction of circulating myelin basic protein and proteolipid protein-speciﬁc transforming growth factor-beta1-secreting Th3 T-cells by oral administration of myelin in multiple sclerosis patients. J. Clin. Invest. 98, 70–77. Fuller, K. A., Pearl, D., and Whitacre, C. C. (1990). Oral tolerance in experimental autoimmune encephalomyelitis: Serum and salivary antibody responses. J. Neuroimmunol. 28, 15–26. Furrie, E., Turner, M. W., and Storbel, S. (1994). Failure of SCID mice to generate an oral tolerogen after a feed of ovalbumin: A role for a functioning gut-associated lymphoid system. Immunology 83, 562–567. Furrie, E., Turner, M., and Strobel, S. (1995a). The absence of gut ﬂora has no effect on the induction of oral tolerance to ovalbumin. Adv. Exp. Med. Biol. 1239–1241. Furrie, E., Turner, M. W., and Strobel, S. (1995b). Partial characterization of a circulating tolerogenic moiety which, after a feed of ovalbumin, suppress delayed-type hypersensitivity in recipient mice. Immunology 86, 480–486. Gaboriau-Routhiau, V., and Moreau, M. (1996). Gut ﬂora allows recovery of oral tolerance to ovalbumin in mice after transient breakdown mediated by cholera toxin and Escherichia coli heat-labile enterotoxin. Pediatr. Res. 39, 625–629. Galliaerde, V., Desvignes, C., Peyron, E., and Kaiserlian, D. (1995). Oral tolerance to haptens: Intestinal epithelial cells from 2,4-dinitrochlorobenzene-fed mice inhibit haptenspeciﬁc T-cell activation in vitro. Eur. J. Immunol. 25, 1385–1390. Gammon, G., and Sercaz, E. (1989). Nature (London) 342, 183–185. Gaupp, S., Hartung, H. P., Toyka, K., and Jung, S. (1997). Modulation of experimental autoimmune neuritis in Lewis rats by Oral application of myelin antigens. J. Neuroimmunol. 79, 129–137. Gaur, A., Wiers, B., Liu, A., Rothbard, J., and Fathman, C. G. (1992). Amelioration of autoimmune encephalomyelitis by myelin basic protein synthetic peptide-induced anergy. Science 258, 1491–1494.

ORAL TOLERANCE: MECHANISMS AND THERAPY

241

Gautam, S. C., and Battisto, J. R. (1985). Orally induced tolerance generates an efferently acting suppressor T-cell and an acceptor T-cell that together down-regulate contact sensitivty. J. Immunol. 135, 2975–2983. Gautam, S. C., Chikkala, N. F., and Battisto, J. R. (1990). Oral administration of the contact sensitizer trinitrochlorobenzene: Initial sensitization and subsequent appearance of a suppressor population. Cell Immunol. 125, 437–448. Gesualdo, L., Lamm, M. E., and Emancipator, S. N. (1990). Defective oral tolerance promotes nephritogenesis in experimental IgA nephropathy induced by oral immunization. J. Immunol. 145, 3684–3691. Goidl, E. A., Thorbecke, G. J., Weksler, M. E., and Siskind, G. W. (1980). Production of auto-anti-idiotypic antibody during the normal immune response: Changes in the autoanti-idiotypic antibody response and the idiotype repertoire associated with aging. Proc. Natl. Acad. Sci. U.S.A., 77, 6788–6792. Gold, W. R., Jr., Queenan, J. T., Woody, J., and Sacher, R. A. (1983). Oral desensitization in Rh disease. Am. J. Obstet. Gynecol. 146, 980–981. Goldman-Brezinski, S., Brezinski, K., Zhang, X. M., Gienapp, I., Cox, K., Heber-Katz, E., and Whitacre, C. (1998). Effects of oral tolerance induction by myelin basic protein on Vbeta8⫹ Lewis rat T-cells. J. Neurosci. Res. 51, 67–75. Gonnella, P. A., Chen, Y., Inobe, J.-I., Quartulli, M., and Weiner, H. L. (1998). In situ immune response in gut associated lymphoid tissue (GALT) following oral antigen in TcR transgenic mice. J. Immunol. 160, 4708–4718. Gorczynski, R. M. (1994). Adoptive transfer of unresponsiveness to allogeneic skin grafts with hepatic gamma delta⫹ T-cells. Immunology 81, 27–35. Gorczynski, R. M., Hozumi, N., Wolf, S., and Chen, Z. (1995). Interleukin 12 in combination with anti-interleukin 10 reverses graft prolongation after portal venous immunization. Transplantion 60, 1337–1341. Gorczynski, R. M., Chen, Z., Hoang, Y., and Rossi-Bergman, B. (1996). A subset of gamma delta T-cell receptor-positive cells produce T-helper type-2 cytokines and regulate mouse skin graft rejection following portal venous pretransplant preimmunization. Immunology 87, 381–389. Grdic, D., Hornquist, E., Kjerrulf, M., and Lycke, N. Y. (1998). Lack of local suppression in orally tolerant CD8-deﬁcient mice reveals a critical regulatory role of CD8⫹T-cells in the normal gut mucosa. J. Immunol. 160, 754–762. Gregerson, D. S., Obritsch, W. F., and Donoso, L. A. (1993). Oral tolerance in experimental autoimmune uveoretinitis. Distinct mechanisms of resistance are induced by low dose vs high dose feeding protocols. J. Immunol. 151, 5751–5761. Gregorian, S. K., Clark, L., Heber-Katz, E., Amento, E. P., and Rostami, A. (1993). Induction of peripheral tolerance with peptide-speciﬁc anergy in experimental autoimmune neuritis. Cell. Immunol. 150, 298–310. Groux, H., O’Garra, A., Bigler, M., Rouleau, M., Antonenko, S., de Vries, J. E., and Roncarolo, M. G. (1997). A CD4⫹ T-cell subset inhibits antigen-speciﬁc T-cell responses and prevents colitis. Nature (London) 389, 737–742. Guimaraes, V. C., Quintans, J., Fisfalen, M.-E., Straus, F. H., Wilhelm, K., Medeiros-Neto, G. A., and DeGroot, L. J. (1995). Suppression of experimental autoimmune thyroiditis by oral administration of thyroglobulin. Endocrinology 136, 3353–3359. Guimaraes, V. C., Quintans, J., Fisfalen, M. E., Straus, F. H., Fields, P. E., MedeirosNeto, G., and DeGroot, L. J. (1996). Immunosuppression of thyroiditis. Endocrinology 137, 2199–2204. Gu¨tgemann, I., Fahrer, A. M., Altman, J. D., Davis, M. M., and Chien, Y. (1998). Induction of rapid T-cell activation and tolerance by systemic presentation of an orally administered antigen. Immunity 8, 667–673.

242

ANA M. C. FARIA AND HOWARD L. WEINER

Habicht, G. S. (1980). Tolerance to human IgG in aged C57BL/6 and BALB/c mice. Fed. Proc. 39, 2118–2124. Hachimura, S., Fujikawa, Y., Enomoto, A., Kim, S. M., Ametani, A., and Kaminogawa, S. (1994). Differential inhibition of T and B cell responses to individual antigenic determinants in orally tolerized mice. Int. Immunol. 6, 1791–1797. Hag, T. A., Mason, H. S., Clements, J. D., and Arntzen, C. J. (1995). Oral immunization with a recombinant bacterial antigen produced in transgenic plants. Science 268, 714–716. Halsey, J. F., and Benjamin, D. C. (1976). Induction of immunologic tolerance in nursing neonates by absorption of tolerogen from colostrum. J. Immunol. 116, 1204–1207. Hancock, W., Sayegh, M., Kwok, C., Weiner, H., and Carpenter, C. (1993). Oral, but not intravenous, alloantigen prevents accelerated allograft rejection by selective intragraft Th2 cell activation. Transplantion 55, 1112–1118. Hancock, W. W., Khoury, S. J., Carpenter, C. B., and Sayegh, M. H. (1994). Differential effects of oral versus intrathymic administration of polymorphic major histocompatibility complex class II peptides on mononuclear and endothelial cell activation and cytokine expression during a delayed-type hypersensitivity response. Am. J. Pathol. 144, 1149–1158. Hancock, W. W., Polanski, M., Zhang, Z. J., Blogg, N., and Weiner, H. L. (1995). Suppression of insulitis in NOD mice by oral insulin administration is associated with selective expression of IL-4, IL-10, TGF-웁 and prostaglandin-E. Am. J. Pathol. 147, 1193–1199. Haneda, K., Sano, K., Tamura, G., Sato, T., Habu, S., and Shirato, K. (1997). TGF-웁 induced by oral tolerance ameliorates experimental tracheal eosinophilia. J. Immunol. 159, 4484–4490. Hanson, D. G. (1981). Ontogeny of orally induced tolerance to soluble proteins in mice. I. Priming and tolerance in newborns. J. Immunol. 127, 1518–1524. Hanson, D. G., and Miller, S. D. (1982). Inhibition of speciﬁc immune responses by feeding protein antigens. V. Induction of the tolerant state in the absence of speciﬁc suppressor T-cells. J. Immunol. 128, 2378–2381. Hanson, D. G., and Morimoto, T. (1987). Delayed recovery of orally-induced tolerance to proteins in irradiated and spleen-cell reconstituted mice. Adv. Exp. Med. Biol. 216A, 733–738. Hanson, D. G., Vaz, N. M., Maia, L. C., Hornbrook, M. M., Lynch, J. M., and Roy, C. A. (1977). Inhibition of speciﬁc immune responses by feeding protein antigens. Int. Arch. Allergy Appl. Immunol. 55, 526–532. Hanson, D. G., Vaz, N. M., Maia, L. C., and Lynch, J. M. (1979a). Inhibition of speciﬁc immune responses by feeding protein antigens. III. Evidence against maintenance of tolerance to ovalbumin by orally induced antibodies. J. Immunol. 123, 2337–2343. Hanson, D. G., Vaz, N. M., Rawlings, L. A., and Lynch, J. L. (1979). Inhibition of speciﬁc immune responses by feeding protein antigens. II. Effects of prior passive and active immunization. J. Immunol. 122, 2261–2266. Hanson, D. G., Roy, M. J., Green, G. M., and Miller, S. D. (1993). Inhibition of orally-induced immune tolerance in mice by prefeeding an endopeptidase inhibitor. Reg. Immunol. 5, 76–84. Haque, M. A., Yoshino, S., Inada, S., Nomaguchi, H., Tokunaga, O., and Kohashi, O. (1996). Suppression of adjuvant arthritis in rats by induction of oral tolerance to mycobacterial 65-kDa heat shock protein. Eur. J. Immunol. 26, 2650–2656. Harper, H. M., Cochrane, L., and Williams, N. A. (1996). The role of small intestinal antigen-presenting cells in the induction of T-cell reactivity to soluble protein antigens: Association between aberrant presentation in the lamina propria and oral tolerance. Immunology 89, 449–456.

ORAL TOLERANCE: MECHANISMS AND THERAPY

243

Harrison, L. C. (1992). Islet cell antigens in insulin-dependent diabetes: Pandora’s box revisited. Immunol. Today 13, 348–352. Harrison, L. C., Dempsey-Collier, M., Kramer, D. R., and Takahashi, K. (1996). Aerosol insulin induces regulatory CD8 웂␦ T-cells that prevent murine insulin-dependent diabetes. J. Exp. Med. 184, 2167–2174. Hartmann, B., Bellman, K., Ghiea, I., Kleeman, R., and Kolb, H. (1997). Oral insulin for diabetes prevention in NOD mice: Potentiation by enhancing Th2 cytokine expression in the gut through bacterial adjuvant. Diabetology 40, 902–909. He, Y. G., Mellon, J., and Niederkorn, J. Y. (1996). The effect of oral immunization on corneal allograft survival. Transplantion 61, 920–926. Hemmings, W. A. (1978). ‘‘Antigen Absorption by the Gut.’’ MTP Press, Lancaster. Heppell, L. M. J., and Kilshaw, P. J. (1982). Immune responses in guinea pigs to dietary protein. Int. Arch. Allergy Appl. Immunol. 68, 54–59. Herbelin, A., Gombert, J. M., Lepault, F., Bach, J. F., and Chatenoud, L. (1998). Mature mainstream TCR alpha beta⫹CD4⫹ thymocytes expressing L-selectin mediate active tolerance in the nonobese diabetic mouse. J. Immunol. 161, 2620–2628. Higgins, P., and Weiner, H. L. (1988). Suppression of experimental autoimmune encephalomyelitis by oral administration of myelin basic protein and its fragments. J. Immunol. 140, 440–445. Hiroi, T., Iwatani, K., Iijima, H., Kodama, S., Yanagita, M., and Kiyono, H. (1998). Nasal immune system: Distinctive Th0 and Th1/Th2 type environments in murine nasalassociated lymphoid tissues and nasal passage, respectively. Eur. J. Immunol. 28, 3346– 3353. Holt, P. G., and Sedgwick, J. D. (1987). Suppression of IgE responses following inhalation of antigen. Immunol. Today 8, 14–15. Holt, P. G., Batty, J. E., and Turner, K. J. (1981). Inhibition of speciﬁc IgE response in mice by pre-exposure to inhaled antigen. Immunology 42, 409–417. Holt, P. G., Vines, J., and Britten, D. (1988). Sublingual allergen administration. I. Selective suppression of IgE production in rats by high allergen doses. Clin. Allergy 18, 229–234. Hormatz, P. R., Bloch, K. J., Brown, M., Walker, W. A., and Kleinam, R. E. (1989). Intestinal adaptation during lactation in the mouse. 1. Enhanced intestinal uptake of dietary protein antigens. Immunology 67, 551–555. Hoyne, G. F., and Lamb, J. R. (1997). Regulation of T-cell function in mucosal tolerance. Immunol. Cell Biol. 75, 197–201. Hoyne, G. F., and Thomas, W. R. (1995). T-Cell responses to orally administered antigens. Study of the kinetics of lymphokine production after single and multiple feeding. Immunology 84, 304–309. Hoyne, G. F., O’Hehir, R. E., Wraith, D. C., Thomas, W. R., and Lamb, J. R. (1993). Inhibition of T-cell and antibody responses to house dust mite allergen by inhalation of the dominant T-cell epitope in naive and sensitized mice. J. Exp. Med. 178, 1783–1788. Hoyne, G. F., Callow, M. G., Kuo, M. C., and Thomas, W. R. (1994). Differences in epitopes recognized by T-cells during oral tolerance and priming. Immunol. Cell Biol. 72, 29–33. Hoyne, G. F., Askinas, B. A., Hetzel, C., Thomas, W. R., and Lamb, J. R. (1996). Regulation of house dust mite responses by intranasally administered peptide: Transient activation of CD4⫹ T-cells precedes the development of tolerance in vivo. Int. Immunol. 8, 335–342. Hoyne, G. F., Jarnicki, A. G., Thomas, W. R., and Lamb, J. R. (1997). Characterization of the speciﬁcity and duration of T-cell tolerance to intranasally administered peptides in mice: A role for intramolecular epitope suppression. Int. Immunol. 9, 1165–1173.

244

ANA M. C. FARIA AND HOWARD L. WEINER

Hu, W., Hasan, A., Wilson, A., Stanford, M., Li-Yang, Y., Todryk, S., Whiston, R., Shinnick, T., Mizushima, Y., van der Zee, R., and Lehner, T. (1998). Experimental mucosal induction of uveitis with the 60-kDa heat shock protein-derived peptide 336–351. J. Immunol. 28, 2444–2455. Huang, L., Soldevila, G., Leeker, M., Flavell, R., and Crispe, I. N. (1994). The liver eliminates T-cells undergoing antigen-triggered apoptosis in vivo. Immunity 1, 741–749. Huetz, F., Tai Kom, Y., and Coutinho, A. (1990). Cellular basis for the age-associated increase in autoimmune reactions. Int. Immunol. 2, 329–335. Husby, S., Jensenius, J. C., and Svehag, S. E. (1985a). Passage of undergraded dietary antigen into the blood of healthy adults. Scand. J. Immunol. 22, 83–92. Husby, S., Oxelius, V. A., Teisner, B., Jensenius, J. C., and Svehag, S. E. (1985b). Humoral immunity to dietary antigens in healthy adults. Int. Arch. Allergy Immunol. 77, 416–422. Husby, S., Mestecky, J., Moldoveanu, Z., Holland, S., and Elson, C. O. (1994). Oral tolerance in humans. T-cell but not B cell tolerance after antigen feeding. J. Immunol. 152, 4663– 4670. Iijima, H., Takahashi, I., Hiroi, T., Shimaoka, M., Kawano, S., Nagano, K., Hori, M., and Kiyono, H. (1998). Orally administered cholera toxin prevents murine intestinal T cells from staphylococcal enterotoxin B-induced anergy. Gastroenterology 115, 1197–1204. Ilan, Y., Prakash, R., Davidson, A., Jona, V., Droguett, G., Horwitz, M. S., Chowdhury, N. R., and Chowdhury, J. R. (1997). Oral tolerization to adenoviral antigens permits longterm gene expression using recombinant adenoviral vectors. J. Clin. Invest. 99, 1098–1106. Imaoka, K., Miller, C. J., Kubota, M., McChesney, M. B., Lohman, B., Yamamoto, M., Fujihashi, K., Someya, K., Honda, M., McGhee, J. R., and Kiyono, H. (1998). Nasal immunization of nonhuman primates with simian immunodeﬁciency virus p55gag and cholera toxin adjuvant induces Th1/Th2 help for virus-speciﬁc immune responses in reproductive tissues. J. Immunol. 161, 5952–5958. Inada, S., Yoshino, S., Haque, M. A., Ogata, Y., and Kohashi, O. (1997). Clonal anergy is a potent mechanism of oral tolerance in the suppression of acute antigen-induced arthritis in rats by oral administration of the inducing antigen. Cell. Immunol. 175, 67–75. Inobe, J., Slavin, A. J., Komagata, Y., Chen, Y., Liu, L., and Weiner, H. L. (1998). IL-4 is a differentiation factor for transforming growth factor-beta secreting Th3 cells and oral administration of IL-4 enhances oral tolerance in experimental allergic encephalomyelitis. Eur. J. Immunol. 28, 2780–2790. Insoft, R. M., Sanderson, I. R., and Walker, W. A. (1996). Development of immune function in the intestine and its role in neonatal diseases. Pediatr. Clin. North Am. 43, 551–571. Ishii, N., Moriguchi, N., Sugita, Y., Nakajima, H., Tanaka, S., and Aoki, I. (1993). Analysis of responsive cells in tolerance by the oral administration of ovalbumin. Immunol. Invest. 22, 451–462. Ishizaka, S., Kimoto, M., Tsujii, T., and Saito, S. (1994). Antibody production system modulated by oral administration of human milk and TGF-웁. Cell. Immunol. 159, 77–84. Iwasaki, A., and Kelsall, B. L. (1999). Distinct T cell priming by freshly isolated dendritic cells from Peyer’s patch and spleen. FASEB J. 13(4), A606. Jacobs, M. J. M., Van Den Hoek, A. E. M., Van De Putte, L. B. A., and Van Den Berg, W. B. (1993). Aberrant tolerance induction with cationic antigens. Scand. J. Immunol. 37, 97–103. Jain, S. L., Barone, K. S., and Michael, J. G. (1996). Activation patterns of murine T-cells after oral administration of an enterocoated soluble antigen. Cell. Immunol. 167, 170–175. Jarrett, E. E. (1984). Perinatal inﬂuences on IgE responses. Lancet 2, 797–799. Javed, N. H., Gienapp, I. E., Cox, K. L., and Whitacre, C. C. (1995). Exquisite peptide speciﬁcity of oral tolerance in experimental autoimmune encephalomyelitis. J. Immunol. 155, 1599–1605.

ORAL TOLERANCE: MECHANISMS AND THERAPY

245

Jewell, S., Dierksheide, J., Curry, A., Shrestha, A., and Waldman, J. (1998). Suppression of experimental autoimmune encephalomyelitis (EAE) by portal vein (PV) injection of myelin basic protein (MBP). FASEB J. 12, A600. Josien, R., Douillard, P., Guillot, C., Mu¨schen, M., Anegon, I., Chetritt, J., Menoret, S., Vignes, C., Soulillou, J. P., and Cuturi, M. C. (1998). A critical role for transforming growth factor-웁 in donor transfusion-induced allograft tolerance. J. Clin. Invest. 102, 1920–1926. Kagnoff, M. F. (1978a). Effects of antigen-feeding on intestinal and systemic immune responses. II. Suppression of delayed-type hypersensitivity reactions. J. Immunol. 120, 1509–1513. Kagnoff, M. F. (1978b). Effects of antigen-feeding on intestinal and systemic immune responses. III. Antigen-speciﬁc serum-mediated suppression of humoral antibody responses after antigen feeding. Cell. Immunol. 40, 186–203. Kaiserlian, D., Vidal, K., MacDonald, H. R., and Grosjean, I. (1993). Mouse intestinal epithelial cells express the self superantigen MIs1a. Eur. J. Immunol. 23, 2717–2720. Kalanadhabhatta, V., Muppidi, D., Sahni, H., Robles, A., and Kramer, M. (1996). Successful oral desensitization to trimethoprim–sulfamethoxazole in acquired immune deﬁciency syndrome. Ann. Allergy Asthma Immunol. 77, 394–400. Kalinski, P., Hilkens, C. M., Snijders, A., Snijdewint, F. G., and Kapsenberg, M. L. (1997). IL-12 deﬁcient dendritic cells, generated in the presence of prostaglandin E2, promote type 2 cytokine production in maturing human naive T helper cells. J. Immunol. 159, 28–35. Kaltreider, H. B. (1976). Expression of immune mechanisms in the lung. Am. Rev. Respir. Dis. 113, 347–379. Kamada, N., Davies, H. S., and Roser, B. (1981). Nature (London) 292, 840–842. Kamada, N., Teramoto, K., Baguerizo, A., Ishikawa, M., Sumimoto, R., and Ohkouchi, Y. (1988). Cellular basis of transplantation induced by liver grafting in the rat. Extent of clonal deletion among thoracic duct lymphocytes, spleen, and lymph node cells. Transplantion 46, 165–167. Kaneko, T., Terasawa, Y., Senou, Y., Nagata, M., and Kuwata, T. (1998). Oral tolerance is not inﬂuenced by oral application of oil-emulsiﬁed proteins. Int. Arch. Allergy Immunol. 116, 81–82. Kantele, A., and Makela, P. H. (1991). Different proﬁles of the human immune response to primary and secondary immunization with an oral Salmonella typhi Ty21a vaccine. Vaccine 9, 423–427. Kapp, J. A., and Ke, Y. (1997). The role of gamma delta TCR-bearing T cells in oral tolerance. Res. Immunol. 148, 561–567. Karachunski, P. I., Ostlie, N. S., Okita, D. K., and Conti-Fine, B. M. (1997). Prevention of experimental myasthenia gravis by nasal administration of synthetic acetylcholine receptor T epitope sequences. J. Clin. Invest. 100, 3027–3035. Karpus, W. J., and Lukacs, N. W. (1996). The role of chemokines in oral tolerance: Abrogation of nonresponsiveness by treatment with antimonocyte chemotactic protein-1. Ann. N.Y. Acad. Sci. 778, 133–144. Karpus, W., Kennedy, K., Smith, W., and Miller, S. (1996). Inhibition of relapsing experimental autoimmune encephalomyelitis in SJL mice by feeding the immunodominant PLP139– 151 peptide. J. Neurosci. Res. 45, 410–423. Karpus, W. J., Kennedy, K. J., Kunkel, S. L., and Lukacs, N. W. (1998). Monocyte chemotactic protein 1 regulates oral tolerance induction by inhibition of T helper cell-related cytokines. J. Exp. Med. 187, 733–741. Kawanishi, H., and Kiely, J. (1987). Immunoregulatory defects in murine aged Peyer’s patches. Eur. J. Immunol. 17, 1223–1228.

246

ANA M. C. FARIA AND HOWARD L. WEINER

Kawanishi, H., Senda, S., and Ajitsu, S. (1989). Aging-associated intrinsic defects in IgA production by murine Peyer’s patch B cells stimulated by autoreactive Peyer’s patch Tcell hybridoma-derived B cell stimulatory factors (BSF). Mech. Aging Dev. 49, 61–78. Kay, R. A., and Ferguson, A. (1989). The immunological consequences of feeding cholera toxin. I. Feeding cholera toxin suppresses the induction of systemic delayed-type hypersensitivity but not humoral immunity. Immunology 66, 410–415. Ke, Y., and Kapp, J. A. (1996). Oral antigen inhibits priming of CD8⫹ CTL, CD4⫹ Tcells, and antibody responses while activating CD8⫹ suppressor T-cells. J. Immunol. 156, 916–921. Ke, Y., Pearce, K., Lake, J. P., Ziegler, H. K., and Kapp, J. A. (1997). Gamma delta T lymphocytes regulate the induction and maintenance of oral tolerance. J. Immunol. 158, 3610–3618. Kelly, K. A., and Whitacre, C. C. (1996). Oral tolerance in EAE: Reversal of tolerance by T helper cell cytokines. J. Neuroimmunol. 66, 77–84. Kerlero de Rosbo, N., Milo, R., Lees, M. B., Burger, D., Bernard, C. C. A., and Ben-Nun, A. (1993). Reactivity to myelin antigens in multiple sclerosis: Peripheral blood lymphocytes respond predominantly to myelin oligodendrocyte glycoprotein. J. Clin. Invest. 92, 2602– 2608. Khare, S. D., Krco, C. J., Grifﬁths, M. M., Luthra, H. S., and David, C. S. (1995). Oral administration of an immunodominant human collagen peptide modulates collageninduced arthritis. J. Immunol. 155, 3653–3659. Khoury, S. J., Lider, O., al-Sabbagh, A., and Weiner, H. L. (1990). Suppression of experimental autoimmune encephalomyelitis by oral administration of myelin basic protein. III. Synergistic effect of lipopolysaccharide. Cell. Immunol. 131, 302–310. Khoury, S. J., Hancock, W. W., and Weiner, H. L. (1992). Oral tolerance to myelin basic protein and natural recovery from experimental autoimmune encephalomyelitis as associated with downregulation of inﬂammatory cytokines and differential upregulation of transforming growth factor 웁, interleukin 4, and prostaglandin E expression in the brain. J. Exp. Med. 176, 1355–1364. Khoury, S. J., Sayegh, M. H., Hancock, W. W., Gallon, L., Carpenter, C. B., and Weiner, H. L. (1993). Acquired tolerance to experimental autoimmune encephalomyelitis by intrathymic injection of myelin basic protein or its major encephalitogenic peptide. J. Exp. Med. 178, 559–566. Kilshaw, P. J., and Cant, A. J. (1984). The passage of maternal dietary proteins into human breast milk. Int. Arch. Allergy Appl. Immunol. 75, 8–15. Kim, J. H., and Ohsawa, M. (1995). Oral tolerance to ovalbumin in mice as a model for detecting modulators of the immunologic tolerance to a speciﬁc antigen. Biol. Pharm. Bull. 18, 854–858. Kim, S. M., Enomoto, A., Hachimura, S., Yamauchi, K., and Kaminogawa, S. (1993). Serum antibody response elicited by a casein diet is directed to only limited determinants of alpha s1-casein. Int. Arch. Allergy Immunol. 101, 260–265. Kim, S. K., Reed, D. S., Olson, S., Schnell, M. J., Rose, J. K., Morton, P. A., and Lefrancois, L. (1998). Generation of mucosal cytotoxic T-cells against soluble protein by tissue-speciﬁc environmental and costimulatory signals. Proc. Natl. Acad. Sci. U.S.A. 95, 10814–10819. King, C., Davies, J., Mueller, R., Lee, M. S., Krahl, T., Yeung, B., O’Connor, E., and Sarvetnick, N. (1998). TGF-웁-1 alters APC preference, polarizing islet antigen responses toward a Th2 phenotype. Immunity 8, 601–613. Kiyono, H., McGhee, J. R., Wannemuehler, M. J., and Michalek, S. M. (1982). Lack of oral tolerance in C3H/HeJ mice. J. Exp. Med. 155, 605–610.

ORAL TOLERANCE: MECHANISMS AND THERAPY

247

Kjerrulf, M., Grdic, D., Ekman, L., Schon, K., Vajdy, M., and Lycke, N. Y. (1997). Interferongamma receptor-deﬁcient mice exhibit impaired gut mucosal immune responses but intact oral tolerance. Immunology 92, 60–68. Komagata, Y., Liu, L. M., and Weiner, H. L. (1998). B cells are important for the production of active TGF-웁. FASEB J. I, A308. Korenblatt, P. E., Rothberg, R. M., Minden, P., and Farr, R. S. (1968). Immune responses of human adults after oral and parenteral exposure to bovine serum albumin. J. Allergy Clin. Immunol. 41, 226–235. Kosiewicz, M. M., Alard, P., and Streilein, J. W. (1998). Alterations in cytokine production following intraocular injection of soluble protein antigen: Impairment in IFN-gamma and induction of TGF-beta and IL-4 production. J. Immunol. 161, 5382–5390. Koster, F. T., and Pierce, N. F. (1983). Parenteral immunization causes antigen-speciﬁc cell-mediated suppression of an intestinal IgA response. J. Immunol. 131, 115–118. Koyama, K., Hosokawa, T., and Aoike, A. (1990). Aging effect on the immune function of murine gut-associated lymphoid tissues. Dev. Comp. Immunol. 14, 465–473. Kuchroo, V. K., Byrne, M. C., Atsumi, Y., Greenﬁeld, E., Connolly, J. B., Whitters, M. J., O’Hara, R., Jr., Collins, M., and Dorf, M. E. (1991). T-cell receptor alpha chain plays a critical role in antigen-speciﬁc suppressor cell function. Proc. Natl. Acad. Sci. U.S.A. 88, 8700–8704. Kuhn, R., Lohler, J., Rennick, D., Rajewsky, K., and Muller, W. (1993). Interleukin-10 deﬁcient mice develop chronic enterocolitis. Cell 75, 263–274. Kunimoto, D. Y., Harriman, G. R., and Strober, W. (1988). Regulation of IgA differentiation in CH12LX B cells by lymphokines. IL-4 induces membrane IgM-positive CH12LX cells to express membrane IgA and IL-5 induces membrane IgA-positive CH12LX cells to secrete IgA. J. Immunol. 141, 713–718. Kuper, C. F., Koornstra, P. J., Hameleers, D. M., Biewenga, J., Spit, B. J., Duivestijn, A. M., van Breda Vriesman, P. J., and Sminia, T. (1992). The role of nasopharyngeal lymphoid tissue. Immunol. Today 13, 219–224. Kweon, M. N., Fujihashi, K., Wakatsuki, Y., Koga, T., Yamamoto, M., McGhee, J. R., and Kiyono, H. (1999). Mucosally induced systemic T cell unresponsiveness to ovalbumin requires CD40 ligand-CD40 interactions. J. Immunol. 162(4), 1904–1909. Kweon, M., Fujihashi, K., VanCott, J., Higuchi, K., Yamamoto, M., McGhee, J., and Kiyono, H. (1998). Lack of orally induced systemic unresponsiveness in IFN-gamma knockout mice. J. Immunol. 160, 1687–1693. Lafont, S., Andre, C., Andre, F., Gillon, J., and Fargier, M.-C. (1982). Abrogation by subsequent feeding of antibody response, including IgE, in parenterally immunized mice. J. Exp. Med. 155, 1573–1578. Lahmann, W. M., Menezes, J. S., Verdolin, B. A., and Vaz, N. M. (1992). Inﬂuence of age on the induction of oral tolerance in mice and its adoptive transfer by spleen cells. Braz. J. Med. Biol. Res. 25, 813–821. Lamont, A. G., Gordon, M., and Ferguson, A. (1987). Oral tolerance in protein-deprived mice. I. Profound antibody tolerance but impaired DTH tolerance after antigen feeding. Immunology 61, 333–337. Lamont, A. G., Bruce, M. G., Watret, K. C., and Ferguson, A. (1988a). Suppression of an established DTH response to ovalbumin in mice by feeding antigen after immunization. Immunology 64, 135–139. Lamont, A. G., Mowat, A. M., Browning, M. J., and Parrott, D. M. (1988b). Genetic control of oral tolerance to ovalbumin in mice. Immunology 63, 737–739. Lamont, A. G., Mowat, A. M., and Parrot, D. M. V. (1989). Priming of systemic and local delayed-type hypersensitivity responses by feeding low doses of ovalbumin to mice. Immunology 66, 595–599.

248

ANA M. C. FARIA AND HOWARD L. WEINER

Lee, S., Scherberg, N., and De Groot, L. J. (1998). Induction of oral tolerance in human autoimmune thyroid disease. Thyroid 8, 229–234. Leishman, A. J., Garside, P., and Mowat, A. M. (1998). Immunological consequences of intervention in established immune responses by feeding protein antigens. Cell. Immunol. 183, 137–148. Li, Y., Yio, X. Y., and Mayer, L. (1995). Human intestinal epithelial cell-induced CD8⫹ Tcell activation is mediated through CD8 and the activation of CD8-associated p56lck. J. Exp. Med. 182, 1079–1088. Li, H. L., Liu, J. Q., Bai, X. F., van der Meide, P. H., and Link, H. (1998a). Dose-dependent mechanisms relate to nasal tolerance induction and protection against experimental autoimmune encephalomyelitis in Lewis rats. Immunology 94, 431–437. Li, H. L., Shi, F. D., Bai, X. F., Huang, Y. M., van der Meide, P. H., Xiao, B. G., and Link, H. (1998b). Nasal tolerance to experimental autoimmune myasthenia gravis: Tolerance reversal by nasal administration of minute amounts of interferon-gamma. Clin. Immunol. Immunopathol. 87, 15–22. Lider, O., Santos, L. M., Lee, C. S., Higgins, P. J., and Weiner, H. L. (1989). Suppression of experimental autoimmune encephalomyelitis by oral administration of myelin basic protein. II. Suppression of disease and in vitro immune responses is mediated by antigenspeciﬁc CD8⫹ T lymphocytes. J. Immunol. 142, 748–752. Lippard, V. W., Schloss, O. M., and Johnson, P. A. (1936). Immune reactions induced in infants by intestinal absorption of incompletely digested cow’s milk protein. Am. J. Dis. Child. 52. Litwin, A., Flanagan, M., Entis, G., Gottschlich, G., Esch, R., Gartside, P., and Michael, J. G. (1996). Immunologic effects of encapsulated short ragweed extract: A potent new agent for oral immunotherapy. Ann. Allergy Asthma Immunol. 77, 132–138. Litwin, A., Flanagan, M., Entis, G., Gottschlich, G., Esch, R., Gartside, P., and Michael, J. G. (1997). Oral immunotherapy with short ragweed extract in a novel encapsulated preparation: A double-blind study. J. Allergy Clin. Immunol. 100, 30–38. Liu, Y., and Janeway, C. A. (1990). Interferon 웂 plays a critical role in induced cell death of effector cell: A possible third mechanism of self-tolerance. J. Exp. Med. 172, 1735–1739. Liu, L. M., and MacPherson, G. G. (1993). Antigen acquisition by dendritic cells: Intestinal dendritic cells acquire antigen administered orally and can prime naive T-cells in vivo. J. Exp. Med. 177, 1299–1307. Liu, J. Q., Bai, X. F., Shi, F. D., Xiao, B. G., Li, H. L., Levi, M., Mustafa, M., Wahren, B., and Link, H. (1998a). Inhibition of experimental autoimmune encephalomyelitis in Lewis rats by nasal administration of encephalitogenic MBP peptides: Synergistic effects of MBP 68–86 and 87–99. Int. Immunol. 10, 1139–1148. Liu, L., Kuchroo, V. K., and Weiner, H. L. (1998b). B7.2 but not B7.1 costimulation is required for the induction of low dose oral tolerance. FASEB J. I, A597. Liu, L., Rich, B. E., Inobe, J.-I., Chen, W., and Weiner, H. L. (1998c). Induction of T helper 2 cell differentiation in the primary immune response: Dendritic cells isolated from adherent cell culture treated with interleukin-10 prime naive CD4⫹ T-cells to secrete interleukin-4. Int. Immunol. 10, 1017–1026. Louis, E., Franchimont, D., Deprez, M., Lamproye, A., Schaaf, N., Mahieu, P., and Belaiche, J. (1996). Decrease in systemic tolerance to fed ovalbumin in indomethacin-treated mice. Int. Arch. Allergy Immunol. 109, 21–26. Lown, K. S., Bailey, D. G., Fontana, R. J., Janardan, S. K., Adair, C. H., Fortlage, L. A., Brown, M. B., Guo, W., and Watkins, P. B. (1997). Grapefruit juice increases felodipine oral availability in humans by decreasing intestinal CYP3A protein expression. J. Clin. Invest. 99, 2545–2553.

ORAL TOLERANCE: MECHANISMS AND THERAPY

249

Lowney, E. D. (1968). Immunologic unresponsiveness to a contact sensitizer in man. J. Invest. Dermatol. 51, 411–417. Lowney, E. D. (1971). Tolerance of dinitrochlorobenzene, a contact sensitizer, in man. J. Allergy Clin. Immunol. 48, 28–35. Lowney, E. D. (1973). Suppression of contact sensitization in man by prior feeding of antigen. J. Invest. Dermatol. 61, 90–93. Lowney, E. D. (1974). A single-step procedure for inducing partial tolerance of DNCB in human subjects. J. Invest. Dermatol. 63, 260–261. Lowrey, J. A., Savage, N. D. L., Palliser, D., Corsin-Jimenez, M., Forsyth, L. M. G., Hall, G., Lindey, S., Stewart, G. A., Tan, K. A. L., Hoyne, G. F., and Lamb, J. R. (1998). Induction of tolerance via the respiratory mucosa. Int. Arch. Allergy Immunol. 116, 93–102. Lu, S., and Holmdahl, R. (1999). Different therapeutic and bystander effects by intranasal administration of homologous type II and type IX collagens on the collagen-induced arthritis and pristance-induced arthritis in rats. Clin Immunol. 90, 119–127. Lundin, B. S., Dahlgren, U. I. H., Hanson, L. A., and Telemo, E. (1996). Oral tolerization leads to active suppression and bystander tolerance in adult rats while anergy dominates in young rats. Scand. J. Immunol. 43, 56–63. Lycke, N., and Holmgren, J. (1986). Strong adjuvant properties of cholera toxin on gut mucosal immune responses to orally presented antigens. Immunology 59, 301–308. Ma, C.-G., Zhang, G.-X., Xiao, B.-G., Link, J., Olsson, T., and Link, H. (1995). Suppression of experimental autoimmune myasthenia gravis by nasal administration of acetylcholine receptor. J. Neuroimmunol. 58, 51–60. Ma, C. G., Zhang, G. X., Xiao, B. G., and Link, H. (1996). Cellular mRNA expression of interferon-gamma (IFN-웂), IL-4 and transforming growth factor-beta (TGF-웁) in rats nasally tolerized against experimental autoimmune myasthenia gravis (EAMG). Clin. Exp. Immunol. 104, 509–516. Ma, D., Mellon, J., and Niederkorn, J. Y. (1997a). Oral administration as a strategy for enhancing corneal allograft survival. Br. J. Ophthalmol. 81, 778–784. Ma, S. W., Zhao, D. L., Yin, Z. Q., Mukherjee, R., Singh, B., Qin, H. Y., Stiller, C. R., and Jevnikar, A. M. (1997b). Transgenic plants expressing autoantigens fed to mice to induce oral immune tolerance. Nature Med. 3, 793–796. Ma, D., Mellon, J., and Niederkorn, J. Y. (1998). Conditions affecting enhanced corneal allograft survival by oral immunization. Invest. Ophthalmol. Vis. Sci. 39, 1835–1846. MacDonald, T. T. (1982). Immunosuppression caused by antigen feeding. I. Evidence for the activation of a feedback suppressor pathway in the spleens of antigen-fed mice. Eur. J. Immunol. 12, 767–773. Macfarlane, G. T., and Macfarlane, S. (1997). Human colonic microbiota: ecology, physiology and metabolic potential of intestinal bacteria. Scand. J. Gastroenterol. (Suppl.) 222, 3–9. Mahida, Y. R., Wu, K. C., and Jewell, D. P. (1988). Characterization of antigen-presenting activity of intestinal mononuclear cells isolated from normal and inﬂammatory bowel disease colon and ileum. Immunology 65, 543–549. Maia, L. C. S., Vaz, N. M., and Vaz, E. M. (1974). Effect of soluble antigen on IgE responses in the mouse. Int. Arch. Allergy Appl. Immunol. 46, 339–346. Mannie, M. D., Rendall, S. K., Arnold, P. Y., Nardella, J. P., and White, G. A. (1996). Anergy-associated T-cell antigen presentation. J. Immunol. 157, 1062–1070. Maron, R., and Weiner, H. L. (1998). Oral tolerance in neonatal NOD mice to autoantigens: Oral insulin suppresses diabetes but oral myelin oligodendrocyte glycoprotein (MOG) has no effect on experimental autoimmune encephalomyelitis (EAE). Diabetes 47, A210.

250

ANA M. C. FARIA AND HOWARD L. WEINER

Maron, R., Melican, N. S., and Weiner, H. L. (1999). Regulatory Th2 type T cell lines against insulin and GAD peptides derived from orally and nasally treated NOD mice suppress diabetes. J. Autoimmun., in press. Maron, R., Palanivel, V., Weiner, H. L., and Harn, D. A. (1998a). Oral administration of insulin B-chain and schistosome egg antigens generate and enhances Th2-type responses in NOD mice. Clin. Immunol. Immunopathol. 87, 85–92. Maron, R., Slavin, A., and Weiner, H. L. (1998b). Oral tolerance to glatiramer acetate (Cop1, Copaxone) in MBPT cell receptor transgenic mice. J. Neuroimmunol. 90, 82. Marth, T., Strober, W., and Kelsall, B. L. (1996). High dose oral tolerance in ovalbumin TCR-transgenic mice: Systemic neutralization of IL-12 augments TGF-웁 secretion and T-cell apoptosis. J. Immunol. 157, 2348–2357. Matsui, M., Haﬂer, D. A., and Weiner, H. L. (1996). Pilot study of oral tolerance to keyhole limpet hemocyanin in humans: Down-regulation of KLH-reactive precursor-cell frequency. Ann. N.Y. Acad. Sci. 778, 398–404. Mattingly, J. A., and Waksman, B. H. (1978). Immunologic suppression after oral administration of antigen. I. Speciﬁc suppressor cells formed in rat Peyer’s patches after oral administration of sheep erythrocytes and their systemic migration. J. Immunol. 121, 1878– 1882. Matzinger, P. (1994). Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12, 991–1045. Mayer, L. (1996). Oral tolerance: Boom or bust for autoimmunity. Inﬂamm. Bowel Dis. 2, 211–213. Mayer, L., and Eisenhardt, D. (1990). Lack of induction of suppressor T-cells by intestinal epithelial cells from patients with inﬂammatory bowel disease. J. Clin. Invest. 86, 1255– 1260. Mayer, L., and Shlien, R. (1987). Evidence for function of Ia molecules on gut epithelial cells in man. J. Exp. Med. 166, 1471–1483. Mayer, L., So, L. P., Yio, X. Y., and Small, G. (1996). Antigen trafﬁcking in the intestine. Ann. N.Y. Acad. Sci. 778, 28–35. McKown, K. M., Carbone, L. D., Bustillo, J., Sever, J. M., Kang, A. H., and Postlethwaite, A. E. (1997). Open trial of oral type I collagen in patients with systemic sclerosis. Arthritis Rheum. 40, S100. McMenamin, C., Oliver, J., Girn, B., Holt, B. J., Kees, U. R., Thomas, W. R., and Holt, P. G. (1991). Regulation of T-cell sensitization at epithelial surfaces in the respiratory tract: Suppression of IgE responses to inhaled antigens by CD3⫹ Tcr alpha-/beta-lymphocytes (putative gamma/delta T-cells). Immunology 74, 234–239. McMenamin, C., Pimm, C., McKersey, M, and Holt, P. G. (1994). Regulation of IgE responses to inhaled antigen in mice by antigen-speciﬁc 애␦ T-cells. Science 265, 1869– 1871. McMenamin, C., McKersey, M., Kuhnlein, P., Hunig, T., and Holt, P. G. (1995). 웂␦ Tcells down-regulate primary IgE responses in rats to inhaled soluble protein antigens. J. Immunol. 154, 4390–4394. Melamed, D., and Friedman, A. (1993a). Direct evidence for anergy in T lymphocytes tolerized by oral administration of ovalbumin. Eur. J. Immunol. 23, 935–942. Melamed, D., and Friedman, A. (1993b). Modiﬁcation of the immune response by oral tolerance: Antigen requirements and interaction with immunogenic stimuli. Cell. Immunol. 146, 412–420. Melamed, D., and Friedman, A. (1994). In vivo tolerization of Th1 lymphocytes following a single feeding with ovalbumin: Anergy in the absence of suppression. Eur. J. Immunol. 24, 1974–1981.

ORAL TOLERANCE: MECHANISMS AND THERAPY

251

Melamed, D., Fishman-Lovell, J., Uni, Z., Weiner, H. L., and Friedman, A. (1996). Peripheral tolerance of Th2 lymphocytes induced by continuous feeding of ovalbumin. Int. Immunol. 8, 717–724. Mengel, J., Cardillo, F., Aroeira, L. S., Williams, O., Russo, M., and Vaz, N. M. (1995). Anti웂␦ T-cell antibody blocks the induction and maintenance of oral tolerance to ovalbumin in mice. Immunol. Lett. 48, 97–102. Mestecky, J. (1987). The common mucosal immune system and current strategies for induction of immune responses in external secretions. J. Clin. Immunol. 7, 265–276. Metzler, B., and Wraith, D. C. (1993). Inhibition of experimental autoimmune encephalomyelitis by inhalation but not oral administration of the encephalitogenic peptide: Inﬂuence of MHC binding afﬁnity. Int. Immunol. 5, 1159–1165. Meyer, A., Gienapp, I., Cox, K., Goverman, J., Hood, L., and Whitacre, C. (1996a). Oral tolerance in myelin basic protein TCR transgenic mice. Ann. N.Y. Acad. Sci. 778, 412–413. Meyer, A. L., Benson, J. M., Gienapp, I. E., Cox, K. L., and Whitacre, C. C. (1996b). Suppression of murine chronic relapsing experimental autoimmune encephalomyelitis by the oral administration of myelin basic protein. J. Immunol. 157, 4230–4238. Meyers, L. K., Brand, D. D., Ye, X. J., Cremer, M. A., Rosloniec, E. F., Bodo, M., Myllyharju, J., Helaakoski, T., Nokelainen, M. Pihlajaniemi, T., Kivirikko, K., Yang, C. L., Ala-Kokko, L., Prockop, D. J., Notbohm, H., Fietzek, P., Suart, J. M., and Kang, A. H. (1998). Characterization of recombinant type II collagen: Arthritogenicity and tolerogenicity in DBA/1 mice. Immunology 95, 631–9. Michael, J. (1989). The role of digestive enzymes in orally induced immune tolerance. Immunol. Invest. 18, 1049–1054. Migita, K., and Ochi, A. (1994). Induction of clonal anergy by oral administration of staphylococcal enterotoxin B. Eur. J. Immunol. 24, 2081–2086. Miller, S., and Hanson, D. (1979). Inhibition of speciﬁc immune responses by feeding protein antigens. IV. Evidence for tolerance and speciﬁc active suppression of cellmediated immune responses to ovalbumin. J. Immunol. 123, 2344–2350. Miller, M. L., Cowdery, J. S., Laskin, C. A., Curtin, M., Jr., and Steinberg, A. D. (1984). Heterogeneity of oral tolerance defects in autoimmune mice. Clin. Immunol. Immunopathol. 31, 231–240. Miller, A., Lider, O., and Weiner, H. L. (1991). Antigen-driven bystander suppression following oral administration of antigens. J. Exp. Med. 174, 791–798. Miller, A., Lider, O., al-Sabbagh, A., and Weiner, H. (1992a). Suppression of experimental autoimmune encephalomyelitis by oral administration of myelin basic protein. V. Hierarchy of suppression by myelin basic protein from different species. J. Neuroimmunol. 39, 243–250. Miller, A., Lider, O., Roberts, A. B., Sporn, M. B., and Weiner, H. L. (1992b). Suppressor T-cells generated by oral tolerization to myelin basic protein suppress both in vitro and in vivo immune responses by the release of TGF-웁 following antigen speciﬁc triggering. Proc. Natl. Acad. Sci. U.S.A. 89, 421–425. Miller, A., al-Sabbagh, A., Santos, L., Das, M. P., and Weiner, H. L. (1993a). Epitopes of myelin basic protein that trigger TGF-웁 release following oral tolerization are distinct from encephalitogenic epitopes and mediate epitope driven bystander suppression. J. Immunol. 151, 7307–7315. Miller, A., Zhang, Z. J., Sobel, R. A., al-Sabbagh, A., and Weiner, H. L. (1993b). Suppression of experimental autoimmune encephalomyelitis by oral administration of myelin basic protein. VI. Suppression of adoptively transferred disease and differential effects of oral vs. intravenous tolerization. J. Neuroimmunol. 46, 73–82.

252

ANA M. C. FARIA AND HOWARD L. WEINER

Miller, A., Lider, O., Abramsky, O., and Weiner, H. L. (1994a). Orally administered myelin basic protein in neonates primes for immune responses and enhances experimental autoimmune encephalomyelitis in adult animals. Eur. J. Immunol. 24, 1026–1032. Miller, B. G., Whittemore, C. T., Stokes, C. R., and Telemo, E. (1994b). The effects of delayed weaning on the development of oral tolerance to soya-bean protein in pigs. Br. J. Nutr. 71, 615–625. Miossec, P., and van den Berg, W. (1997). Th1/Th2 cytokine balance in arthritis [see comments]. Arthritis Rheum. 40, 2105–2115. Moalem, G., Leibowitz-Amit, R., Yoles, E., Mor, F., Cohen, I. R., and Schwartz, M. (1999). Autoimmune T cells protect neurons from secondary degeneration after central nervous system axotomy. Nat. Med. 5(1), 49–55. Moog, F. (1981). The lining of the small intestine. Sci. Am. 245, 116–125. Mordes, J. P., Schirf, B., Roipko, D., Grenier, D. L., Weiner, H. L., Nelson, P., and Rossini, A. A. (1996). Oral insulin does not prevent insulin-dependent diabetes mellitus in BB rats. Ann. N.Y. Acad. Sci. 778, 418–421. Moreau, M. C., and Corthier, G. (1988). Effect of the gastrointestinal microﬂora on induction and maintenance of oral tolerance to ovalbumin in C3H/HeJ mice. Infect. Immun. 56, 2766–2768. Moreau, M. C., and Gaboriau-Routhiau, V. (1996). The absence of gut ﬂora, the doses of antigen ingested and aging affect the long-term peripheral tolerance induced by ovalbumin feeding in mice. Res. Immunol. 147, 49–59. Morris, D. L. (1998). Intradermal testing and sublingual desensitization for nickel. Cutis 61, 129–132. Mowat, A. M. (1986). Depletion of suppressor T-cells by 2⬘-deoxyguanosine abrogates tolerance in mice fed ovalbumin and permits the induction of intestinal delayed-type hypersensitivity. Immunology 58, 179–184. Mowat, A. M. (1987). The regulation of immune responses to dietary protein antigens. Immunol. Today 8, 93–98. Mowat, A. M. (1994). Oral tolerance and regulation of immunity to dietary antigens. In ‘‘Handbook of Mucosal Immunology’’ (P. L. Ogra, M. E. Lamm, J. R. McGhee, J. Mestecky, W. Strobe., and J. Bienensta eds.), Vol. 16, pp. 185–201. Academic Press, San Diego. Mowat, A. (1998). Putative role of p55 TNF receptor, but not fas in oral tolerance. FASEB J. 12, A598. Mowat, A., and Ferguson, A. (1982). Migration inhibition of lymph nodes lymphocytes as an assay for regional cell-mediated immunity in the intestinal lymphoid tissues of mice immunized orally with ovalbumin. Immunology 47, 365–370. Mowat, A. M., and Parrot, D. M. (1983). Immunological responses to fed protein antigens in mice. IV. Effects of stimulating the reticuloendothelial system on oral tolerance and intestinal immunity to ovalbumin. Immunology 50, 547–554. Mowat, A. M., Strobel, S., Drummond, H. E., and Ferguson, A. (1982). Immunological response to fed protein antigens in mice: I. Reversal of oral tolerance to ovalbumin by cyclophosphamide. Immunology 45, 105–113. Mowat, A. M., Thomas, M. J., MacKenzie, S., and Parrott, D. M. (1986). Divergent effects of bacterial lipopolysaccharide on immunity to orally administered protein and particulate antigens in mice. Immunology 58, 677–683. Mowat, A. M., Steel, M., Worthy, E. A., Kewin, P. J., and Garside, P. (1996). Inactivation of Th1 and Th2 cells by feeding ovalbumin. Ann. N.Y. Acad. Sci. 778, 122–132. Muckerheide, A., Apple, R. J., Pesce, A. J., and Michael, J. G. (1987). Cationization of protein antigens II. Alteration of regulatory properties. J. Immunol. 138, 833–837.

ORAL TOLERANCE: MECHANISMS AND THERAPY

253

Mueller, D. L., Jenkins, M. K., and Schwartz, R. H. (1989). Clonal expansion versus functional clonal inactivation: A costimulatory signalling pathway determines the outcome of T-cell antigen receptor occupancy. Annu. Rev. Immunol. 7, 445–480. Muller, C. P. (1994). Mucosal vaccines: High-tech reagents for low-tech immunization. Immunol. Today 15, 458–459. Myers, L. K., Seyer, J. M., Stuart, J. M., and Kang, A. H. (1997). Suppression of murine collagen-induced arthritis by nasal administration of collagen. Immunology 90, 161–164. Myers, L. K., Brand, D. D., Ye, X. J., Cremer, M. A., Rosloniec, E. F., Bodo, M., Myllyharju, J., Helaakoski, T., Nokelainen, M., Pihlajaniemi, T., Kivirikko, K., Yang, C. L., Ala-Kokko, L., Prockop, D. J., Notbohm, H., Fietzek, P., Stuart, J. M., and Kang, A. H. (1998). Characterization of recombinant type II collagen: Arthritogenicity and tolerogenicity in DBA/1 mice. Immunology 95, 631–639. Nagler-Anderson, C., Bober, L. A., Robinson, M. E., Siskind, G. W., and Thorbeke, F. J. (1986). Suppression of type II collagen-induced arthritis by intragastric administration of soluble type II collagen. Proc. Natl. Acad. Sci. U.S.A. 83, 7443–7446. Nakao, A., Kasai, M., Kumano, K., Nakajima, H., Kurasawa, K., and Iwamoto, I. (1998). High-dose oral tolerance prevents antigen-induced eosinophil recruitment into the mouse airways. Int. Immunol. 10, 387–394. Nelson, P. A., Akselband, Y., Dearborn, S., al-Sabbagh, A., Tian, Z. J., Gonnella, P., Zamvil, S., Chen, Y., and Weiner, H. L. (1996). Effect of oral beta interferon on subsequent immune responsiveness. Ann. N.Y. Acad. Sci. 778, 145–155. Neurath, M. F., Fuss, I., Kelsall, B. L., Presky, D. H., Waegell, W., and Strober, W. (1996). Experimental granulomatous colitis in mice is abrogated by induction of TGF-웁-mediated oral tolerance. J. Exp. Med. 183, 2605–2616. Neutra, R. M., and Kraehenbuhl, J.-P. (1996). M cells as a pathway for antigen uptake and processing. In ‘‘Essentials of Mucosal Immunology’’ (M. F. Kagnoff and H. Kiyono, eds.). Academic Press, San Diego. Neutra, M. R., Frey, A., and Kraehenbuhl, J. P. (1996). Epithelial M cells: Gateways for mucosal infection and immunization. Cell 86, 345–348. Newby, T. J., Stokes, C. R., and Bourne, F. J. (1980). Effects of feeding bacterial lipopolysaccharide and dextran sulphate on the development of oral tolerance to contact sensitizing agents. Immunology 41, 617–621. Ngan, J., and Kind, L. S. (1978). Suppressor T-cells for IgE and IgG in Peyer’s patches of mice made tolerant by the oral administration of ovalbumin. J. Immunol. 120, 861–865. Nussenblatt, R. B., Caspi, R. R., Mahdi, R., Chan, C. C., Roberge, F., Lider, O., and Weiner, H. L. (1990). Inhibition of S-antigen induced experimental autoimmune uveoretinitis by oral induction of tolerance with S-antigen. J. Immunol. 144, 1689–1695. Nussenblatt, R. B., Gery, I., Weiner, H. L., Ferris, F., Shiloach, J., Ramaley, N., Perry, C., Caspi, R., Haﬂer, D. A., Foster, S., and Whitcup, S. M. (1997). Treatment of uveitis by oral administration of retinal antigens: Results of a phase I/II randomized masked trial. Am. J. Ophthalmol. 123, 583–592. Oettinger, H. F., al-Sabbagh, A., Jingwu, Z., LaSalle, J. M., Weiner, H. L., and Haﬂer, D. A. (1993). Biological activity of recombinant human myelin basic protein. J. Neuroimmunol. 44, 157–162. Ogra, P. (1980). Ontogeny of the local immune system. Pediatrics 64, 675–687. Okumura, S., McIntosh, K., and Drachman, D. B. (1994). Oral administration of acetylcholine receptor: Effects on experimental myasthenia gravis. Ann. Neurol. 36, 704–713. Oliver, A. R., and Silbart, L. K. (1998). Local and systemic tolerance to orally administered dinitrochlorobenzene is not broken by cholera toxin. Int. Arch. Allergy Immunol. 116, 318–324.

254

ANA M. C. FARIA AND HOWARD L. WEINER

Oliver, A. R., Silbart, L. K., Keren, D. F., Van Kruiningen, H. J., Miller, B. F., and Rearick, C. (1997). Mucosal unresponsiveness to aﬂatoxin B1 is not broken by cholera toxin. Immunol. Cell Biol. 75, 47–53. O’Mahony, S., Arranz, E., Barton, J. R., and Ferguson, A. (1991). Dissociation between systemic and mucosal humoral immune responses in coeliac disease. Gut 32, 29–35. Panja, A., and Mayer, L. (1994). Diversity and function of antigen-presenting cells in mucosal tissues. In ‘‘Handbook of Mucosal Immunology’’ P. L. Ogra, M. E. Lamm, J. R. McGhee, J. Mestecky, W. Strober, and J. Bienenstock, eds., Vol. 16, pp. 177-181. Academic Press, San Diego. Panja, A., Blumberg, R. S., Balk, S. P., and Mayer, L. (1993). CD1d is involved in Tcell–intestinal epithelial cell interactions. J. Exp. Med. 178, 1115–1119. Panja, A., Barone, A., and Mayer, L. (1994). Stimulation of lamina propria lymphocytes by intestinal epithelial cells: Evidence for recognition of nonclassical restriction elements. J. Exp. Med. 179, 943–950. Peng, H. J., Turner, M. W., and Strobel, S. (1989). Failure to induce oral tolerance to protein antigens in neonatal mice can be corrected by transfer of adult spleen cells. Pediatr. Res. 26, 486–490. Peng, H. J., Turner, M. W., and Strobel, S. (1990). The generation of a ‘‘tolerogen’’ after the ingestion of ovalbumin is time-dependent and unrelated to serum levels of immunoreactive antigen. Clin. Exp. Immunol. 81, 510–515. Peng, H. J., Chang, Z. N., Han, S. H., Won, M. H., and Huang, B. T. (1995). Chemical denaturation of ovalbumin abrogates the induction of oral tolerance of speciﬁc IgG antibody and DTH responses in mice. Scand. J. Immunol. 42, 297–304. Peng, H. J., Chang, Z. N., Lin, S. Y., Han, S. H., and Chang, C. H. (1998a). Chemical denaturation of ovalbumin abrogates the induction of oral tolerance of mouse reaginic antibody responses. Scand. J. Immunol. 47, 475–480. Peng, H. J., Chang, Z. N., Tsai, L. C., Su, S. N., Shen, H. D., and Chang, C. H. (1998b). Heat denaturation of egg-white proteins abrogates the induction of oral tolerance of speciﬁc Th2 immune responses in mice. Scand. J. Immunol. 48, 491–496. Peri, B. A., Theodore, C. M., Losonsky, G. A., Fishaut, J. M., Rothberg, R. M., and Ogra, P. L. (1982). Antibody content of rabbit milk and serum following inhalation or ingestion of respiratory syncytial virus and bovine serum albumin. Clin. Exp. Immunol. 48, 91–97. Peterson, K. E., and Braley-Mullen, H. (1995). Suppression of murine experimental autoimmune thyroiditis by oral administration of porcine thyroglobulin. Cell. Immunol. 166, 123–130. Ploix, C., Bergerot, I., Fabien, N., Moulin, V., and Thivolet, C. (1998). Protection against autoimmune diabetes with oral insulin is associated with the homing of IL-4 T helper type 2 cells to the pancreas and pancreatic lymph nodes. Diabetes 47, 39–44. Polanski, M., Blogg, N. S., Zhang, J., and Weiner, H. L. (1997). Oral administration of the immunodominant B-chain of insulin suppresses diabetes in NOD mice and is associated with a switch from Th1 to Th2 cytokines. J. Autoimmun. 10, 339–346. Popovich, P. G., Yu, J. Y., and Whitacre, C. C. (1997). Spinal cord neuropathology in rat experimental autoimmune encephalomyelitis: Modulation by oral administration of myelin basic protein. J. Neuropathol. Exp. Neurol. 56, 1323–1338. Powrie, F., Correa-Oliveira, R., Mauze, S., and Coffman, R. L. (1994). Regulatory interactions between CD45RBhigh and CD45RBlow CD4⫹ T-cells are important for the balance between protective and pathogenic cell-mediated immunity. J. Exp. Med. 179, 589–600. Powrie, F., Carlino, J., Leach, M. W., Mauze, S., and Coffman, R. L. (1996). A critical role for transforming growth factor-beta but not interleukin 4 in the suppression of T helper type 1-mediated colitis by CD45RB (low) CD4⫹ T-cells. J. Exp. Med. 183, 2669–2674.

ORAL TOLERANCE: MECHANISMS AND THERAPY

255

Prakken, B. J., van der Zee, R., Anderton, S. M., van Kooten, P. J. S., Kuis, W., and van Eden W. (1997). Peptide-induced nasal tolerance for a mycobacterial heat shock protein 60 T-cell epitope in rats suppresses both adjuvant arthritis and nonmicrobially induced experimental arthritis. Proc. Natl. Acad. Sci. U.S.A. 94, 3284–3289. Pullerits, T., Lundin, S., Cui, Z. H., Dahlgren, U., Telemo, E., and Lotvall, J. (1998). Bystander suppression of occupational hapten sensitization in rats made tolerant to ovalbumin. Eur. Respir. J. 12, 889–894. Qian, J., Hashimoto, T., Fujiwara, H., and Hamaoka, T. (1985). Studies on the induction of tolerance to alloantigens. I. The abrogation of potentials for delayed-type-hypersensitivity response to alloantigens by portal venous inoculation with allogeneic cells. J. Immunol. 134, 3656–3661. Qin, S., Cobbold, S. P., Pope, H., Elliott, J., Kioussis, D., Davies, J., and Waldmann, H. (1993). ‘Infectious’ transplantation tolerance. Science 259, 974–977. Quinti, I., Paganelli, R., Scala, E., Guerra, E., and Aiuti, F. (1989). Humoral response to food antigens. Allergy 44, 59–64. Racke, M. K., and Lovett-Racke, A. E. (1998). Bystander suppression in experimental autoimmune encephalomyelitis: Where and how does it occur? Res. Immunol. 149, 820–827. Reese, R. T., and Cebra, J. J. (1975). Anti-dinitrophenyl antibody production in strain 13 guinea pigs fed or sensitized with dinitrochlorobenzene. J. Immunol. 114, 863–871. Renz, H., Bradley, K., Saloga, J., Loader, J., Larsen, G. L., and Gelfand, E. W. (1993). Tcells expressing speciﬁc V beta elements regulate immunoglobulin E production and airways responsiveness in vivo. J. Exp. Med. 177, 1175–1180. Renz, H., Lack, G., Saloga, J., Schwinzer, R., Bradley, K., Loader, J., Kupfer, A., Larsen, G. L., and Gelfand, E. W. (1994). Inhibition of IgE production and normalization of airways responsiveness by sensitized CD8 T-cells in a mouse model of allergen-induced sensitization. J. Immunol. 152, 351–360. Richman, L. K., Chiller, J. M., Brown, W. R., Hanson, D. G., and Vaz, N. M. (1978). Enterically induced immunological tolerance. I. Induction of suppressor T lymphocytes by intragastric administration of soluble proteins. J. Immunol. 121, 2429–2433. Richman, L. K., Graeff, A. S., and Strober, W. (1981a). Antigen presentation by macrophageenriched cells from the mouse Peyer’s patch. Cell. Immunol. 62, 110–118. Richman, L. K., Graeff, A. S., Yarchoan, R., and Strober, W. (1981b). Simultaneous induction of antigen-speciﬁc IgA helper T-cells and IgA suppressor T-cells in the murine Peyer’s patch after protein feeding. J. Immunol. 126, 2079–2083. Rios, M. J. C., Pereira, M. A. C., Lopes, L. M., Faria, A. M. C., Gontijo, C. M., Castanheira, E. B., and Vaz, N. M. (1988). Tolerance induction and immunological priming initiated by mucosal contacts with protein antigens in inbred strains of mice. Braz. J. Med. Biol. Res. 21, 825–831. Rizzo, L. V., Morawetz, R. A., Miller-Rivero, N. E., Choi, R., Wiggert, B., Chan, C. C., Morse, H. C., III, Nussenblatt, R. B., and Caspi, R. R. (1999). IL-4 and IL-10 are both required for the induction of oral tolerance. J. Immunol. 162(5), 2613–2622. Rizzo, L. V., Miller-Rivero, N. E., Chan, C.-C., Wiggert, B., Nussenblatt, R. B., and Caspi, R. R. (1994). Interleukin-2 treatment potentiates induction of oral tolerance in a murine model of autoimmunity. J. Clin. Invest. 94, 1668–1672. Rock, K. L., Benacerraf, B., and Abbas, A. K. (1984). Antigen presentation by haptenspeciﬁc B-lymphocytes. I. Role of surface immunoglobulin receptors. J. Exp. Med. 160, 1120–1113. Romball, C. G., and Weigle, W. O. (1993). In vivo induction of tolerance in murine CD4⫹ cell subsets. J. Exp. Med. 178, 1637–1644.

256

ANA M. C. FARIA AND HOWARD L. WEINER

Rubin, D., Weiner, H., Fields, B., and Greene, M. (1981). Immunologic tolerance after oral administration of reovirus: Requirement for two viral gene products for tolerance induction. J. Immunol. 127, 1697–1701. Rueff-Juy, D., Faure, M., Drapier, A. M., and Cazenave, P. A. (1998). Role of maternal Ig in the induction of C kappa-speciﬁc CD8⫹ T cell tolerance. J. Immunol. 161, 721–728. Russo, M., Jancar, S., Siqueria, A. L. P., Mengel, J., Gomes, E., Ficker, S. M., and Faria, A. M. C. (1998). Prevention of lung eosinophilic inﬂammation by oral tolerance. Immunol. Lett. 61, 15–23. Sai, P., and Rivereau, A. S. (1996). Prevention of diabetes in the nonobese diabetic mouse by oral immunological treatments. Comparative efﬁciency of human insulin and two bacterial antigens, lipopolysacharide from Escherichia coli and glycoprotein extract from Klebsiella pneumoniae. Diabetes Metab. 22, 341–348. Saito, H., Kanamori, Y., Takemori, T., Nariuchi, H., Kubota, E., Takahashi-Iwanaga, H., Iwanaga, T., and Ishikawa, H. (1998). Generation of intestinal T-cells from progenitors residing in gut cryptopatches. Science 280, 275–278. Saklayen, M. G., Pesce, A. J., Pollak, V. E., and Michael, J. G. (1983). Induction of oral tolerance in mice unresponsive to bacterial lipopolysaccharide. Infect. Immun. 41, 1383– 1385. Saklayen, M. G., Pesce, A. J., Pollak, V. E., and Michael, J. G. (1984). Kinetics of oral tolerance: study of variables affecting tolerance induced by oral administration of antigen. Int. Arch. Allergy Appl. Immunol. 73, 5–9. Salom, R. N., Maguire, J. A., and Hancock, W. W. (1992). Mechanism of a clinically relevant protocol to induce tolerance: Peritransplant spleen cells plus cyclosporine suppress IL2 and IFN-gamma production. Transplant. Proc. 24, 2307–2308. Samoilova, E. B., Horton, J. L., Zhang, H., Khoury, S. J., Weiner, H. L., and Chen, Y. (1998). CTLA4 is required for the induction of high dose oral tolerance. Int. Immunol. 10, 491–498. Sampson, H. A., and Metcalfe, D. D. (1992). Food allergies. JAMA 268, 2840–2844. Sanderson, I. R., and Walker, W. A. (1993). Uptake and transport of macromolecules by the intestine: Possible role in clinical disorders (an update). Gastroenterology 104, 622–639. Santos, L. M. B., al-Sabbagh, A., Londono, A., and Weiner, H. L. (1994). Oral tolerance to myelin basic protein induces regulatory TGF-웁-secreting T-cells in Peyer’s patches of SJL mice. Cell. Immunol. 157, 439–447. Sato, S., Qian, J. H., Kokudo, S., Ikegami, R., Suda, T., Hamaoka, T., and Fujiwara, H. (1988). Studies on the induction of tolerance to alloantigens. III. Induction of antibodies directed against alloantigen-speciﬁc delayed-type hypersensitivity T-cells by a single injection of allogeneic lymphocytes via portal venous route. J. Immunol. 140, 717–722. Sato, M. N., Carvalho, A. F., Silva, A. O., MacIel, Jr., M., Fusaro, A. E., and Duarte, A. J. (1998). Oral tolerance induced to house dust mite extract in naive and sensitized mice: Evaluation of immunoglobulin G, anti-immunoglobulin G, anti-immunoglobulin E autoantibodies and IgC–IgE complexes. Immunology 95, 193–199. Sayegh, M. H., Khoury, S. J., Hancock, W. H., Weiner, H. L., and Carpenter, C. B. (1992a). Induction of immunity and oral tolerance with polymorphic class II major histocompatability complex allopeptides in the rat. Proc. Natl. Acad. Sci. U.S.A. 89, 7762–7766. Sayegh, M. H., Zhang, Z. J., Hancock, W. W., Kwok, C. A., Carpenter, C. B., and Weiner, H. L. (1992b). Down-regulation of the immune response to histocompatibility antigen and prevention of sensitization by skin allografts by orally administered alloantigen. Transplantion 53, 163–166. Scadding, G. K., and Brostoff, J. (1986). Low dose sublingual therapy in patients with allergic rhinitis due to house dust mite. Clin. Allergy 16, 483–491.

ORAL TOLERANCE: MECHANISMS AND THERAPY

257

Schalkwijk, J., van den Berg, W. B., van de Putte, L. B. A., Joosten, L. A. B., and van den Bersselaar, L. (1985). Cationization of catalase, peroxidase, and superoxide dismutase. Effect of improved intraarticular retention on experimental arthritis in mice. J. Clin. Invest. 76, 198–205. Schwartz, R. H. (1990). A cell culture model for T lymphocyte clonal anergy. Science 248, 1349–1356. Schwartz, R. H., Mueller, D. L., Jenkins, M. K., and Quill, H. (1989). T-cell clonal anergy. Cold Spring Harb. Symp. Quant. Biol. 54, 605–610. Seddon, B., and Mason, D. (1999). Regulatory T cells in the control of autoimmunity: The essential role of transforming growth factor beta and interleukin 4 in the prevention of autoimmune thyroiditis in rats by peripheral CD4(⫹)CD45RC⫺ cells and CD4(⫹)CD8(⫺) thymocytes. J. Exp. Med. 189, 279–288. Seder, R. A., Marth, T., Sieve, M. C., Strober, W., Letterio, J. J., Roberts, A. B., and Kelsall, B. (1998). Factors involved in the differentiation of TGF-웁-producing cells from naive CD4⫹ T-cells: IL-4 and IFN-웂 have opposing effects, while TGF-웁 positively regulates its own production. J. Immunol. 160, 5719–5728. Sedgwick, J. D., and Holt, P. G. (1983). Induction of IgE-isotype speciﬁc tolerance by passive antigenic stimulation of the respiratory mucosa. Immunology 50, 625–630. Sedgwick, J. D., and Holt, P. G. (1985). Down-regulation of immune responses to inhaled antigen: Studies on the mechanism of induced suppression. Immunology 56, 635–642. Sellon, R. K., Tonkonogy, S. L., Grable, H., Kwon, J., Schultz, M., Balish, E., Rennick, D., and Sartor, R. B. (1997). Development of spontaneous colitis in IL-10 knockout mice requires normal enteric bacterial ﬂora. Gastroenterology 112, A1088. Sercarz, E., and Krzych, U. (1991). The distinctive speciﬁcity of antigen-speciﬁc suppressor T-cells. Immunol. Today 12, 111–118. Seymour, B. W. P., Gershwin, L. J., and Coffman, R. L. (1998). Aerosol-induced immunoglobulin (Ig)-E unresponsiveness to ovalbumin does not require CD8⫹ or T-cell receptor (TCR)-gamma/delta⫹ T-cells or interferon (IFN)-gamma in a murine model of allergen sensitization. J. Exp. Med. 187, 721–731. Shenker, B. J., Listgarten, M. A., and Taichman, N. S. (1984). Suppression of human lymphocyte responses by oral spirochetes: A monocyte-dependent phenomenon. J. Immunol. 132, 2039–2045. Shi, F. D., Bai, X. F., Xiao, B. G., van der Meide, P. H., and Link, H. (1998). Nasal administration of multiple antigens suppresses experimental autoimmune myasthenia gravis, encephalomyelitis and neuritis. J. Neurol. Sci. 155, 1–12. Sicinki, P. J., Rowinski, J. B., Warchol, A., Jarzabek, Z., Gut, W., Szcygiel, B., Bielecki, K., and Koch, G. (1990). Poliovirus type I enters the human host through intestinal M cells. Gastroenterology 98, 56–58. Sieper, J., Kary, S., So¨rensen, H., Alten, R., Eggens, U., Hu¨ge, W., Hiepe, F., Ku¨hne, A., Listing, J., Ulbrich, N., Braun, J., Zink, A., and Mitchison, N. A. (1996). Oral type II collagen treatment in early rheumatoid arthritis. Arthritis Rheum. 39, 41–51. Singer, G., and Abbas, A. (1994). The Fas antigen is involved in peripheral but not thymic deletion of T lymphocytes in T-cell receptor transgenic mice. Immunity 1, 365–371. Singh, V. K., Kalra, H. K., Yamaki, K., and Shinohara, T. (1992). Suppression of experimental autoimmune uveitis in rats by the oral administration of the uveitopathogenic S-antigen fragment and a cross-reactive homologous peptide. Cell. Immunol. 139, 81–90. Singh, V. K., Anand, R., Sharma, K., and Agarwal, S. S. (1996). Suppression of experimental autoimmune uveitis in Lewis rats by oral administration of recombinant Escherichia coli expressing retinal S-antigen. Cell. Immunol. 172, 158–162.

258

ANA M. C. FARIA AND HOWARD L. WEINER

Slavin, A. J., Maron, R., Garcia, G., Gonnella, P., and Weiner, H. L. (1998). Oral administration of IL-4 and IL-10 enhance the induction of low dose oral tolerance. FASEB J. II, A599. Sloan-Lancaster, J., Evavold, B. D., and Allen, P. M. (1993). Induction of T-cell anergy by altered T-cell receptor ligand on live antigen presenting cells. Nature (London) 363, 156–159. Sloan-Lancaster, J., Evavold, B. D., and Allen, P. M. (1994). Th2 cell clonal anergy as a consequence of partial activation. J. Exp. Med. 180, 1195–1199. Sminia, T., and Jeurissen, S. H. (1986). The macrophage population of the gastrointestinal tract of the rat. Immunobiology 172, 72–80. Soos, J. M., Johnson, H. M., Weiner, H. L., and Zamvill, S. S. (1999). Combination therapy with oral IFN-tau and oral myelin basic protein results in increased IL-4 and IL-10 production and enhanced suppression of experimental allergic encephalomyelitis. FASEB J. 13(4), A607. Spahn, T. W., and Weiner, H. L. (1998). 웂␦ T-cells are necessary for low dose but not high dose oral tolerance. FASEB J. 12, A597, 3464. Spahn, T. W., Slavin, A. J., Zhang, X., Marino, M. W., Old, L., Fontana, A., and Weiner, H. L. (1999). Evidence for a critical role of the gut associated lymphatic system in the induction of oral tolerance. FASEB J. 13(4), A607. Spalding, D. M., Koopman, W. J., Eldridge, J. H., McGhee, J. R., and Steinman, R. (1983). Accessory cells in murine Peyer’s patch. I. Identiﬁcation and enrichment of a functional dendritic cell. J. Exp. Med. 157, 1646–1659. Spalding, D. M., Williamson, S. I., Koopman, W. J., and McGhee, J. R. (1984). Preferential induction of polyclonal IgA secretion by murine Peyer’s patch dendritic cell-T-cell mixtures. J. Exp. Med. 160, 941–946. Spiekermann, G. M., and Nagler-Anderson, C. (1998). Oral administration of the bacterial superantigen staphylococcal enterotoxin B induces activation and cytokine production by T-cells in murine gut-associated lymphoid tissue. J. Immunol. 161, 5825–5831. Staines, N. A., Harper, N., Ward, F. J., Malmstro¨m, V., Holmdahl, R., and Bansal, S. (1996). Mucosal tolerance and suppression of collagen-induced arthritis (CIA) induced by nasal inhalation of synthetic peptide 184–198 of bovine type II collagen (CII) expressing a dominant T-cell epitope. Clin. Exp. Immunol. 103, 368–375. Staples, P. J., and Talal, N. (1969). Relative inability to induce tolerance in adult NZB and (NZB ⫻ NZW)F1 mice. J. Exp. Med. 130, 123–130. Stavnezer, J., Abbot, J., and Sirlin, S. (1984). Immunoglobulin heavy chain switching in cultured I.29 murine B cell lymphoma cells: commitment to an IgA or IgE switch. Curr. Top. Microbiol. Immunol. 113, 109–116. Steeber, D. A., Tang, M. L., Zhang, X. Q., Muller, W., Wagner, N., and Tedder, T. F. (1998). Efﬁcient lymphocyte migration across high endothelial venules of mouse Peyer’s patches requires overlapping expression of L-selectin and beta7 integrin. J. Immunol. 161, 6638–6647. Stokes, R., Newby, T. J., Huntley, J. H., Patel, D., and Bourne, F. J. (1979). The immune response of mice to bacterial antigens given by mouth. Immunology 38, 497–502. Stransky, B., Faria, A. M. C., and Vaz, N. M. (1998). Oral tolerance induction with altered forms of ovalbumin. Braz. J. Med. Biomed. Res. 31, 381–386. Streilein, J., Bradley, D., Eichorn, M., Horneber, M., and Lutjen-Drecoll, E. (1993). Evidence that the anterior chamber of primate eyes is immune privileged. Invest. Ophthalmol. Vis. Sci. 34(Suppl.), 902. Strobel, S. (1996). Neonatal oral tolerance. Ann. N.Y. Acad. Sci. 778, 88–102. Strobel, S., and Ferguson, A. (1984). Immune responses to fed protein antigens in mice. 3. Systemic tolerance or priming is related to age at which antigen is ﬁrst encountered. Pediatr. Res. 18, 588–594.

ORAL TOLERANCE: MECHANISMS AND THERAPY

259

Strobel, S., and Ferguson, A. (1987). Persistence of oral tolerance in mice fed ovalbumin is different for humoral and cell-mediated immune responses. Immunology 60, 317–318. Strobel, S., Mowat, A. M., Drummond, H. E., and Ferguson, A. (1981). Age at ﬁrst feed inﬂuences the immune responses to fed antigens in mice. Pediatr. Res. 15, 1193–1199. Strobel, S., Mowat, A. M., Drummond, H. E., Pickering, M. G., and Ferguson, A. (1983). Immunological responses to fed protein antigens in mice. II. Oral tolerance for CMI is due to activation of cyclophosphamide-sensitive cells by gut-processed antigen. Immunology 49, 451–456. Strobel, S., Mowat, A. M., and Ferguson, A. (1985). Prevention of oral tolerance induction to ovalbumin and enhanced antigen presentation during a graft-versus-host reaction in mice. Immunology 56, 57–64. Strober, W., and Ehrhardt, R. O. (1994) Regulation of IgA B cell development. In ‘‘Handbook of Mucosal Immunology’’ (P. L. Ogra, J. Mestecky, M. E. Lamm, W. Strober, J. R. McGhee, and J. Bienenstock, eds.), Vol. 16, pp. 159–176. Academic Press, San Diego. Suh, E. D., Vistica, B. P., Chan, C.-C., Raber, J. M., Gery, I., and Nussenblatt, R. B. (1993). Splenectomy abrogates the induction of oral tolerance in experimental. Curr. Eye Res. 12, 833–839. Suko, M., Mori, A., Ito, K., and Okudaira, H. (1995). Oral immunotherapy may induce Tcell anergy. Int. Arch. Allergy Immunol. 107, 278–281. Sun, J.-B., Holmgren, C., and Czerkinsky, C. (1994). Cholera toxin B subunit: An efﬁcient transmucosal carrier-delivery system for induction of peripheral immunological tolerance. Proc. Natl. Acad. Sci. U.S.A. 91, 10795–10799. Suri-Payer, E., Amar, A. Z., Thornton, A. M., and Shevach, E. M. (1998). CD4⫹CD25⫹ Tcells inhibit both the induction and effector function of autoreactive T-cells and represent a unique lineage of immunoregulatory cells. J. Immunol. 160, 1212–1218. Swarbrick, E. T., Stokes, C. R., and Soothill, J. F. (1979). Absorption of antigens after oral immunization and the simultaneous induction of speciﬁc systemic tolerance. Gut 20, 121–125. Swierkiosz, J. E., and Swanborg, R. H. (1977). Immunoregulation of experimental allergic encephalomyelitis: Conditions for induction of suppressor cells and analysis of mechanism. J. Immunol. 119, 1501–1506. Taams, L. S., van Rensen, A. J. M. L., Poelen, M. C. M., van Els, C. A. C. M., Besseling, A. C., Wagenaar, J. P. A., van Eden, W., and Wauben, M. H. M. (1998). Anergic Tcells actively suppress T-cell responses via the antigen-presenting cell. Eur. J. Immunol. 28, 2902–2912. Tada, Y., Ho, A., Koh, D. R., and Mak, T. W. (1996). Collagen-induced arthritis in CD4or CD8-deﬁcient mice. CD8⫺ T-cells play a role in initiation and regulate recovery phase of collagen-induced arthritis. J. Immunol. 156, 4520–4526. Takeuchi, M., Kosiewicz, M. M., Alard, P., and Streilein, J. W. (1997). On the mechanisms by which transforming growth factor-beta 2 alters antigen-presenting abilities of macrophages on T-cell activation. Eur. J. Immunol. 27, 1648–1656. Takeuchi, M., Alard, P., and Streilein, J. W. (1998). TGF-웁 promotes immune deviation by altering accessory signals of antigen-presenting cells. J. Immunol. 160, 1589–1597. Taudorf, E. (1992). Oral immunotherapy of adults with allergic rhinoconjunctivitis: Clinical effects in birch and grass pollinosis. Dan. Med. Bull. 39, 542–560. Taudorf, E., Laursen, L. C., Lanner, A., Bjorksten, B., Dreborg, S., Seborg, M., and Weeke, B. (1987). Oral immunotherapy in birch pollen hay fever. J. Allergy Clin. Immunol. 80, 153–161. Teitelbaum, D., Arnon, R., and Sela, M. (1998). Immunomodulation of experimental allergic encephalomyelitis by oral administration of copolymer 1 (Copaxone威). J. Neuroimmunol. 90, 85.

260

ANA M. C. FARIA AND HOWARD L. WEINER

Teitelbaum, D., Arnon, R., and Sela, M. (1999). Immunomodulation of experimental autoimmune encephalomyelitis by oral administration of copolymer 1. Proc. Natl. Acad. Sci. U.S.A. 96, 3842–3847. Teng, Y., Gorczynski, R., and Hozumi, N. (1998). The function of TGF-beta-mediated innocent bystander suppression associated with physiological self-tolerance in vivo. Cell. Immunol. 190, 51–60. Terato, K., Xiu, J. Y., Miyahara, H., Cremer, M. A., and Grifﬁths, M. M. (1996). Induction by chronic autoimmune arthritis in DBA/1 mice by oral administration of type II collagen and Escherichia coli lipopolysaccharide. Br. J. Rheumatol. 35, 828–838. Thomas, H. C., and Parrott, D. M. V. (1974). The induction of tolerance to a soluble protein antigen by oral administration. Immunology 27, 631–639. Thomas, H. C., Ryan, C. J., Benjamin, I. S., Blumgart, L. H., and MacSween, R. N. M. (1976). The immune system response in cirrhotic rats. The induction of tolerance to orally administered protein antigens. Gastroenterology 71, 114–117. Thompson, H. S., and Staines, N. A. (1986a). Suppression of collagen-induced arthritis with pregastrically or intravenously administered type II collagen. Agents Actions 19, 318–319. Thompson, H. S. G., and Staines, N. A. (1986b). Gastric administration of type II collagen delays the onset and severity of collagen-induced arthritis in rats. Clin. Exp. Immunol. 64, 581–586. Thompson, H. S. and Staines, N. A. (1990). Could speciﬁc oral tolerance be a therapy for autoimmune disease? Immunol. Today 11, 396–399. Thompson, H. S. G., Harper, N., Bevan, D. J., and Staines, N. A. (1993a). Suppression of collagen induced arthritis by oral administration of type II collagen: Changes in immune and arthritic responses mediated by active peripheral suppression. Autoimmunity 16, 189–199. Thompson, S. J., Thompson, H. S. G., Harper, N., Day, M. J., Coad, A. J., Elson, C. J., and Staines, N. A. (1993b). Prevention of pristane-induced arthritis by the oral administration of type II collagen. Immunology 79, 152–157. Thorbecke, G. J., Schwarcz, R., Leu, J., Huang, C., and Simmons, W. J. (1999). Modulation by cytokines of induction of oral tolerance to type II collagen. Arthritis Rheum. 42, 110–118. Thurau, S. R., Chan, C. C., Suh, E., and Nussenblatt, R. B. (1991). Induction of oral tolerance to S-antigen induced experimental autoimmune uveitis by a uveitogenic 20mer peptide. J. Autoimmun. 4, 507–516. Thurau, S. R., Chan, C. C., Nussenblatt, R. B., and Caspi, R. R. (1997a). Oral tolerance in a murine model of relapsing experimental autoimmune uveoretinitis (EAU): Induction of protective tolerance in primed animals. Clin. Exp. Immunol. 109, 370–376. Thurau, S. R., Diedrichs-Mohring, M., Fricke, H., Arbogast, S., and Wildner, G. (1997b). Molecular mimicry as a therapeutic approach for an autoimmune disease: Oral treatment of uveitis-patients with an MHC-peptide crossreactive with autoantigen—ﬁrst results. Immunol. Lett. 57, 193–201. Tian, J., Atkinson, M. A., Clare-Salzler, M., Herschenfeld, A., Forsthuber, T., Lehmann, P. V., and Kaufman, D. L. (1996). Nasal administration of glutamate decarboxylase (GAD65) peptides induces Th2 responses and prevents murine insulin-dependent diabetes. J. Exp. Med. 183, 1561–1567. Titus, R. G., and Chiller, J. M. (1981). Orally induced tolerance. Deﬁnition at the cellular level. Int. Arch. Allergy Appl. Immunol. 65, 323–338. Torseth, J. W., and Gregerson, D. S. (1998). Oral tolerance in experimental autoimmune uveoretinitis: Feeding after disease induction is less protective than prefeeding. Clin. Immunol. Immunopathol. 88, 297–304.

ORAL TOLERANCE: MECHANISMS AND THERAPY

261

Trentham, D. E., Townes, A. S., and Kang, A. S. (1977). Autoimmunity to type II collagen: An experimental model of arthritis. J. Exp. Med. 146, 857–868. Trentham, D., Dynesius-Trentham, R., Orav, E., Combitchi, D., Lorenzo, C., Sewell, K., Haﬂer, D., and Weiner, H. (1993). Effects of oral administration of type II collagen on rheumatoid arthritis. Science 261, 1727–1730. van der Heijden, P. J., Bianchi, A. T. J., Dol, M., Pals, J. W., Stok, W., and Bokhout, B. A. (1991). Manipulation of intestinal immune responses against ovalbumin by cholera toxin and its B subunit in mice. Immunology 72, 89–93. van der Heijden, P. J., Stock, W., and Bianchi, A. T. J. (1987). Contribution of immunoglobulin-secreting cells in the murine small intestine to the total ‘background’ immunoglobulin production. Immunology 62, 551–555. Van de Velde, H., von Hoegen, I., Luo, W., Parnes, J. R., and Thielemans, K. (1991). The B-cell surface protein CD72/Lyb-2 is the ligand for CD5. Nature (London) 351, 662–665. van Eden, W., Thole, J. E., van der Zee, R., Noordzij, A., van Embden, J. D., Hensen, E. J., and Cohen, I. R. (1988). Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature (London) 331, 171–173. van Halteren, A. G., van der Cammen, M. J., Cooper, D., Savelkoul, H. F., Kraal, G., and Holt, P. G. (1997). Regulation of antigen-speciﬁc IgE, IgG1, and mast cell responses to ingested allergen by mucosal tolerance induction. J. Immunol. 159, 3009–3015. van Hoogstraten, I. M., Andersen, K. E., von Blomberg, B. M., Boden, D., Bruynzeel, D. P., Burrows, D., Camarasa, J. G., Dooms-Goossens, A., Kraal, G., and Lahti, A. (1991). Reduced frequency of nickel allergy upon oral nickel contact at an early age. Clin. Exp. Immunol. 85, 441–445. van Hoogstraten, I. M., Boden, D., von Blomberg, M. E., Kraal, G., and Scheper, R. J. (1992). Persistent immune tolerance to nickel and chromium by oral administration prior to cutaneous sensitization. J. Invest. Dermatol. 99, 608–616. van Hoogstraten, I. M., Boos, C., Boden, D., Von Blomberg, M. E., Scheper, R. J., and Kraal, G. (1993). Oral induction of tolerance to nickel sensitization in mice. J. Invest. Dermatol. 101, 26–31. Van Houten, N., and Blake, S. F. (1996). Direct measurement of anergy of antigen-speciﬁc T-cells following oral tolerance induction. J. Immunol. 157, 1337–1341. Vaz, N. M., Maia, L. C. S., Hanson, D. G., and Lynch, J. M. (1977). Inhibition of homocytotropic antibody responses in adult inbred mice by previous feeding of the speciﬁc antigen. J. Allergy Clin. Immunol. 60, 110–115. Vaz, N. M., Maia, L. C., Hanson, D. G., and Lynch, J. M. (1981). Cross-suppression of speciﬁc immune responses after oral tolerance. Mem. Inst. Oswaldo Cruz 76, 83–91. Vaz, N. M., Rios, M. J., Lopes, L. M., Gontijo, C. M., Castanheira, E. B., Jacquemart, F., and Andrade, L. A. (1987). Genetics of susceptibility to oral tolerance to ovalbumin. Braz. J. Med. Biol. Res. 20, 785–790. Vaz, N., Faria, A. M. C., Verdolon, B. A., and Carvalho, C. R. (1997). Immaturity, ageing and oral tolerance. Scand. J. Immunol. 46, 225–229. Vidard, L. J., Colarusso, L., and Benacerraf, B. (1994). Speciﬁc T-cell tolerance may be preceded by a primary response. Proc. Natl. Acad. Sci. U.S.A. 91, 5627–5631. Viney, J. L., Mowat, A. M., O’Malley, J. M., Williamson, E., and Fanger, N. A. (1998). Expanding dendritic cells in vivo enhances the induction of oral tolerance. J. Immunol. 160, 5815–5825. Vistica, B. P., Chanaud, N. P. R., Felix, N., Caspi, R. R., Rizzo, L. V., Nussenblatt, R. B., and Gery, I. (1996). CD8 T-cells are not essential for the induction of ‘‘low-dose’’ oral tolerance. Clin. Immunol. Immunopathol. 78, 196–202.

262

ANA M. C. FARIA AND HOWARD L. WEINER

Volkheimer, G. (1975). Hematogenous dissemination of ingested polyvinyl chloride particles. Ann. N.Y. Acad. Sci. 246, 165–171. von Herrath, M. G., Dyrberg, T., and Oldstone, M. B. A. (1996). Oral insulin treatment suppresses virus-induced antigen-speciﬁc destruction of beta cells and prevents autoimmune diabetes in transgenic mice. J. Clin. Invest. 98, 1324–1331. Vrabec, T. R., Gregerson, D. S., Dua, H. S., and Donoso, L. A. (1992). Inhibition of experimental autoimmune uveoretinitis by oral administration of S-antigen and synthetic peptides. Autoimmunity 12, 175–184. Wagner, N., Lohler, J., Tedder, T. F., Rajewsky, K., Muller, W., and Steeber, D. A. (1998). L-Selectin and beta7 integrin synergistically mediate lymphocyte migration to mesenteric lymph nodes. Eur. J. Immunol. 28, 3832–3839. Waldo, F. B., Van Den Wall Bake, A. W. L., Mestecky, J., and Husby, S. (1994). Suppression of the immune response by nasal immunization. Clin. Immunol. Immunopathol. 72, 30–34. Walker, W. A. (1986). Antigen handling by the small intestine. Clin. Gastroenterol. 15, 1–20. Wang, Z. Y., Qiao, J., and Link, H. (1993a). Suppression of experimental autoimmune myasthenia gravis by oral administration of acetylcholine receptor. J. Neuroimmunol. 44, 209–214. Wang, Z. Y., Qiao, J., Melms, A., and Link, H. (1993b). T-cell reactivity to acetylcholine receptor in rats orally tolerized against experimental autoimmune myasthenia gravis. Cell. Immunol. 152, 394–404. Wang, Z. Y., Link, H., Ljungdahl, A˚., Ho¨jeberg, B., Link, J., He, B., Qiao, J., Melms, A., and Olsson, T. (1994). Induction of interferon-웂, interleukin-4, and transforming growth factor-웁 in rats orally tolerized agaisnt experimental autoimmune myasthenia gravis. Cell. Immunol. 157, 353–368. Wang, Z.-Y., He, B., Qiao, J., and Link, H. (1995a). Suppression of experimental autoimmune myasthenia gravis and experimental allergic encephalomyelitis by oral administration of acetylcholine receptor and myelin basic protein: Double tolerance. J. Neuroimmunol. 63, 79–86. Wang, Z. Y., Huang, J., Olsson, T., He, B., and Link, H. (1995b). B cell responses to acetylcholine receptor in rats orally tolerized against experimental autoimmune myasthenia gravis. J. Neurol. Sci. 128, 167–174. Wannemuehler, M. J., Kiyono, H., Babb, J. L., Michalek, S. M., and McGhee, J. R. (1982). Lipopolysaccharide (LPS) regulation of the immune response: LPS converts germfree mice to sensitivity to oral tolerance induction. J. Immunol. 129, 959–965. Weigle, W. O., Thoman, M. L., and Goodman, M. G. (1988). The effect of aging on the induction of tolerance in a subpopulation of B lymphocytes. Cell. Immunol. 111, 253–257. Weiner, H. L. (1997). Oral tolerance: Immune mechanisms and treatment of autoimmune diseases. Immunol. Today 18, 335–343. Weiner, H. L. (1999). Oral tolerance with cop 1 for the treatment of multiple sclerosis. Proc. Natl. Acad. Sci. U.S.A. 96, 3333–3335. Weiner, H. L., and Komagata, Y. (1998). Oral tolerance and the treatment of rheumatoid arthritis. Springer Semin. Immunopathol. 20, 289–308. Weiner, H. L., and Mayer, L. F. (1996). Oral tolerance: mechanisms and applications. Introduction. Ann. N.Y. Acad. Sci. 13, xiii–xviii. Weksler, M. E., Russo, C., and Siskind, G. W. (1989). Peripheral T-cells select the B-cell repertoire in old mice. Immunol. Rev. 110, 173–185. Wells, H. G. (1911). Studies on the chemistry of anaphylaxis (III). Experiments with isolated proteins, especially those of the hen’s egg. J. Infect. Dis. 8, 147–171. Wells, H. G., and Osborne, T. B. (1911). Biological reactions of the vegetable proteins. I. Anaphylaxis. J. Infect. Dis. 8, 66–123.

ORAL TOLERANCE: MECHANISMS AND THERAPY

263

Whitacre, C. C., Gienapp, I. E., Orosz, C. G., and Bitar, D. (1991). Oral tolerance in experimental autoimmune encephalomyelitis. III. Evidence for clonal anergy. J. Immunol. 147, 2155–2163. Wildner, G., and Thurau, S. R. (1994). Cross-reactivity between an HLA-B27-derived peptide and a retinal autoantigen peptide: A clue to major histocompatibility complex association with autoimmune disease. Eur. J. Immunol. 24, 2579–2585. Wildner, G., and Thurau, S. R. (1995). Orally induced bystander suppression in experimental autoimmune uveoretinitis occurs only in the periphery and not in the eye. Eur. J. Immunol. 25, 1292–1297. Wildner, G., Hunig, T., and Thurau, S. R. (1996). Orally induced, peptide-speciﬁc 웂/␦ TCR⫹ cells suppress experimental autoimmune uveitis. Eur. J. Immunol. 26, 2140–2148. Wilson, D. B., Wilson, D. H., and Schroder, K. (1998). Acquired thymic tolerance and experimental allergic encephalomyelitis in the rat. I. Parameters and analysis of possible mechanisms. Eur. J. Immunol. 28, 2770–2779. Wolvers, D. A., van der Cammen, M. J., and Kraal, G. (1997). Mucosal tolerance is associated with, but independent of, up-regulation Th2 responses. Immunology 92, 328–333. Wolvers, D. A., Bakker, J. M., Bagchus, W. M., and Kraal, G. (1998). The steroid hormone dehydroepiandrosterone (DHEA) breaks intranasally induced tolerance, when administered at time of systemic immunization. J. Immunol. 89, 19–25. Wu, X.-M., Nakashima, M., and Watanabe, T. (1998). Selective suppression of antigenspeciﬁc Th2 cells by continuous micro-dose oral tolerance. Eur. J. Immunol. 28, 134–142. Xiao, B. G., Bai, X. F., Zhang, G. X., and Link, H. (1998a). Suppression of acute and protracted–relapsing experimental allergic encephalomyelitis by nasal administration of low-dose IL-10 in rats. J. Neuroimmunol. 84, 230–237. Xiao, B. G., Zhang, G. X., Shi, F. D., Ma, C. G., and Link, H. (1998b). Decrease of LFA1 is associated with upregulation of TGF-웁 in CD4(⫹) T-cell clones derived from rats nasally tolerized against experimental autoimmune myasthenia gravis. Clin. Immunol. Immunopathol. 89, 196–204. Xu-Amano, J., Aicher, W. K., Taguchi, T., Kiyono, H., and McGhee, J. R. (1992). Selective induction of Th2 cells in murine Peyer’s patches by oral immunization. Int. Immunol. 4, 433–445. Xu-Amano, J., Kiyono, H., Jackson, R. J., Staats, H. F., Fujihashi, K., Burrows, P. D., Elson, C. O., Pillai, S., and McGhee, J. R. (1993). Helper T-cell subsets for immunoglobulin A responses: Oral immunization with tetanus toxoid and cholera toxin as adjuvant selectively induces Th2 cells in mucossa associated tissues. J. Exp. Med. 178, 1309–1320. Yang, R., Liu, Q., Grosfeld, J., and Pescovitz, M. (1994). Intestinal drainage through liver is prerequisite for oral tolerance induction. J. Pediatr. Surg. 29, 1145–1148. Yoshida, H., Hachimura, S., Hirahara, K., Hisatsune, T., Nishijima, K., Shiraishi, A., and Kaminogawa, S. (1998). Induction of oral tolerance in splenocyte-reconstituted SCID mice. Clin. Immunol. Immunopathol. 87, 282–291. Yoshino, S. (1995a). Antigen-induced arthritis in rats is suppressed by the inducing antigen administered orally before, but not after immunization. Cell. Immunol. 163, 55–58. Yoshino, S. (1995b). Downregulation of silicone-induced chronic arthritis by gastric administration of type II collagen. Immunopharmacology 31, 103–108. Yoshino, S. (1998). Treatment with an anti-IL-4 antibody blocks suppression of collageninduced arthritis in mice by oral administration of type II collagen. J. Immunol. 160, 3067– 3071. Yoshino, S., and Yoshino, J. (1998). Effect of a monoclonal antibody against interleukin-4 on suppression of antigen-induced arthritis in mice by oral administration of the inducing antigen. Cell. Immunol. 187, 139–144.

264

ANA M. C. FARIA AND HOWARD L. WEINER

Yoshino, S., Quattrocchi, E., and Weiner, H. L. (1995). Oral administration of type II collagen suppresses antigen-induced arthritis in Lewis rats. Arthritis Rheum 38, 1092– 1096. Yoshino, S., Ohsawa, M., and Sagai, M. (1998). Diesel exhaust particles block induction of oral tolerance in mice. J. Pharmacol. Exp. Ther. 287, 679–683. Zhang, W., and Kong, Y. C. (1998). Noninvolvement of IL-4 and IL-10 in tolerance induction to experimental autoimmune thyroiditis. Cell. Immunol. 187, 95–102. Zhang, Z., and Michael, J. G. (1990). Orally inducible immune unresponsiveness is abrogated by IFN-애 treatment. J. Immunol. 144, 4163–4165. Zhang, J. Z., Lee, C. S. Y., Lider, O., and Weiner, H. L. (1990). Suppression of adjuvant arthritis in Lewis rats by oral administration of type II collagen. J. Immunol. 145, 2489– 2493. Zhang, J. Z., Davidson, L., Eisenbarth, G., and Weiner, H. L. (1991). Suppression of diabetes in NOD mice by oral administration of porcine insulin. Proc. Natl. Acad. Sci. U.S.A. 88, 10252–10256. Zhang, J., Markovic, S., Raus, J., Lacet, B., Weiner, H. L., and Haﬂer, D. A. (1993). Increased frequency of IL-2 responsive T-cells speciﬁc for myelin basic protein and proteolipid protein in peripheral blood and cerebrospinal ﬂuid of patients with multiple sclerosis. J. Exp. Med. 179, 973–984. Zhang, G.-X., Shi, F.-D., Zhu, J., Xiao, B.-G., Levi, M., Wahren, B., Yu, L.-Y. and Link, H. (1998). Synthetic peptides fail to induce nasal tolerance to experimental autoimmune myasthenia gravis. J. Neuroimmunol. 85, 96–101. Zivny, J. H., Russell, M. W., Vu, H. L., Moldoveanu, J., Mestecky, J., and Elson, C. O. (1996). Multiple mechanisms of oral tolerance to food antigens in humans. FASEB J. 10, A1116. Zou, L.-P., Zhu, J., Deng, G.-M., Levi, M., Wahren, B., Diab, A., Hillert, J., and Link, H. (1998). Treatment with P2 protein peptide 57–81 by nasal route is effective in Lewis rat experimental autoimmune neuritis. J. Neuroimmunol. 85, 137–145.

ADVANCES IN IMMUNOLOGY, VOL. 73

Caspases and Cytokines: Roles in Inﬂammation and Autoimmunity JOHN C. REED The Burnham Institute, La Jolla, California 92037

I. Introduction

Caspases are a family of intracellular cysteine proteases that cleave their speciﬁc substrates at aspartic acid residues. These proteases have been implicated in two aspects of cytokine biology: (1) proteolytic processing and bioactivation of the proforms of certain interleukins (ILs), such as pro-IL-1웁 and pro-IL-18, and (2) signaling by apoptosis-inducing members of the tumor necrosis factor (TNF) receptor family, such as tumor necrosis factor receptor-1 (TNFR-1) (CD120a) and Fas (CD95). Abundant evidence, derived from molecular biological, genetic, and pharmacological studies, has demonstrated a critical role for speciﬁc caspases in the bioactivation or bioactivity of particular cytokines. Moreover, these caspases appear to play nonredundant roles in a wide variety of cytokine-mediated inﬂammatory and autoimmune diseases. The involvement of caspases in cytokine biology is examined, with emphasis on relevance to infectious diseases, inﬂammation, and autoimmunity. II. The Caspase Family

To date, 10 members of the caspase family of cysteine proteases in humans have been described (Alnemri, 1997; Salvesen and Dixit, 1997). These proteins are initially synthesized as single polypeptide zymogens that undergo proteolytic processing at speciﬁc aspartic acid residues to produce the active enzymes. The active, processed caspases are composed of heterotetramers, containing two large and two small subunits, which form two active sites per molecule. Caspases represent a unique family of intracellular cysteine proteases, with absolute speciﬁcity for aspartic acid in the P-1 position of substrates. This fact, together with the observation that procaspase zymogens are processed at aspartic acids to produce the active proteases, creates opportunities both for autoactivation of caspases (probably through either cis or transmechanisms) and for cascades of caspase activation in which one active member cleaves and activates another. The N-terminal region of the zymogens is typically but not always removed from the protein during processing. The length and sequence of this so-called prodomain varies widely among caspases, but can be used to help deﬁne subcategories of more closely related family members. 265

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

266

JOHN C. REED

Importantly, the prodomains of some caspases function as protein interaction motifs that allow them either to oligomerize with each other or to interact with other types of proteins that can either promote or inhibit zymogen activation (see Alnemri, 1997; Salvesen and Dixit, 1997; Wallach et al., 1997). Caspases can be grouped into two broad categories: (a) the caspase-1like family and (b) the caspase-3-like family. The categorization of caspases into subfamilies is based on differences in their substrate preferences, comparisons of their primary amino acids sequences, and depending on whether their primary function is most relevant to cytokine processing versus apoptosis regulation. The human caspase-1 subfamily consists of caspase-1 [formerly known as interleukin-1웁-converting enzyme (ICE)], caspase-4 (ICE-rel II, TX, ICH-2), and caspase-5 (ICE-rel III, TY). The murine caspase-11 protein is probably the orthologue or a close homologue of human caspase-4 (Wang et al., 1998). These caspases share extensive amino acid sequence identity and are intimately involved in proinﬂammatory cytokine processing, as described below. Their substrate preference as deﬁned by combinatorial peptide library screening methods and analysis of substrate cleavage sites is (W/L)EHD, where a bulky hydrophobic group is found in the P-4 position (Rano et al., 1997; Thornberry et al., 1997). The human caspase-3 subfamily consists of caspase-2 (ICH-1), caspase-3 (CPP32/Yama/apopain), caspase-6 (Mch-2), caspase-7 (Mch-3, ICE-LAP3, CMH-1), caspase-8 (FLICE, MACH-1, Mch-5), caspase-9 (Mch-6, Apaf3, ICE-LAP6), and caspase-10 (Mch-4, FLICE-2). Compared to the caspase-1 (ICE) subfamily, these caspases generally share greater primary amino acid sequence homology in their catalytic domain regions with the cell death protease CED-3 of Caenorhabditis elegans (Yuan et al., 1993). Like CED-3 of C. elegans, the members of the caspase-3 subfamily are intimately involved in apoptosis, functioning either as upstream initiators or downstream effectors of this cell death process. Based mostly on differences in their prodomains, the caspase-3 subfamily can be further divided into three additional subgroups, as depicted in Fig. 1. The ‘‘initiator’’ caspases (caspase-2, -8, -9, and -10) contain large prodomains that allow them to interact with other proteins, which trigger conversion of the zymogens into active protease (Boldin et al., 1996; Duan and Dixit, 1997; Li et al., 1997b; Muzio et al., 1996). Procaspase-8 and -10 contain death effector domains (DEDs) in their N-terminal prodomains, whereas procaspase-2 and -9 contain caspase recruitment domains (CARDs). It is thought that the primary substrates of these proteases are other downstream procaspases. The ‘‘effector’’ caspases (caspase-3, -6, and -7) contain short prodomains. They appear to become activated primarily through cleavage by upstream caspases or by ampliﬁcation loops in which activated effector

CASPASES AND CYTOKINES

267

FIG. 1. Human caspases. The known members of the human caspase family of proteases is summarized, indicating the four subgroups (A–D), their preferred tetrapeptide substrate speciﬁcities, type of N-terminal prodomain, and apparent biological functions. The caspase1 subfamily (A) is primarily involved in cytokine proprotein processing, and appears to play a less direct role in apoptosis. The caspse-8/10 subfamily (B) proteases contain death effector domains (DEDs) that allow them to bind FADD or possibly other adaptor proteins responsible for their recruitment to TNF family death receptors. The CARD domaincontaining caspases (C) include caspase-2, which binds to RAIDD, an adaptor protein that interacts with TNFR1 complexes, and caspase-9, which binds the CED-4 homologue Apaf1, thus liniking mitochondrial cytochrome c release to caspase activation. The downstream effort caspases (D) contain short prodomains.

caspases cleave and activate their own zymogens. These downstream caspases represent the ﬁnal effectors of apoptosis, cleaving a variety of structural and regulatory proteins at speciﬁc aspartic acid residues, thereby either directly or indirectly inducing apoptosis. The substrate preference of executioner caspases is DEXD. Comparison of the X-ray crystallographic structures of caspase-1 with caspase-3 reveals the presence of an additional loop near the active site of caspase-3, as well as other structural differences, accounting for the differences in the substrate speciﬁcities of these two prototypical members of the caspase family (Rotonda et al., 1996; Wilson et al., 1994). These differences in the geometry of the active sites of caspase-1 and -3 have permitted the synthesis of speciﬁc peptidyl and nonpeptidyl inhibitors, which distinguish the two major branches of the caspase family. III. Caspases and Cytokines

Abundant evidence implicates caspases in inﬂammatory and autoimmune conditions mediated by cytokines (Fig. 2). The production of bioactive interleukin-1 (IL-1), a critical mediator of endotoxic shock, is absolutely

268

JOHN C. REED

FIG. 2. Role of caspase-1 in cytokine processing. Among the identiﬁed activators of caspase-1 are murine caspase-11 (probably equivalent to human caspase-4), granzyme B (Gra B), and the Shiegella protein ipa. Once activated, caspase-1 can also cleave and activate additional procaspase-1 molecules. Caspase-1 is directly responsible for processing of proIL-1웁 and pro-IL-18, and appears to be required for processing of IL-1 alpha. IL-18 plays a major role in the in vivo regulation of interferon-웂 production.

dependent on caspase-1 (thus the former nomenclature IL-1웁-converting enzyme) in vivo. Similarly, proteolytic processing of pro-IL-18, a major inducer of interferon-웂 in vivo, appears to be mediated exclusively by caspase-1, as revealed by caspase-1 gene knockout experiments in mice. Cytolytic Tcells and natural killer (NK) cells also induce caspase-1 activation by introduction into their target cells of granzyme B, a serine protease that directly cleaves and activates procaspase-1 and several other procaspases. The central importance of caspase-1 as a mediator of inﬂammatory responses is further supported by the observation that some types of bacteria produce proteins that directly bind to and activate caspase-1, whereas some viruses express proteins that directly inhibit this protease. Several members of the tumor necrosis factor family of cytokines, including TNF-움, Fas ligand (CD95 ligand), the DR3 ligand, and TRAIL, induce processing and activation of caspase-8 (and in some cases its closely related homologue caspase-10) (Fig. 3). Activation of caspase-8 (or caspase-10)

CASPASES AND CYTOKINES

269

FIG. 3. Caspase activation by TNF family receptors. In humans and mice, ﬁve members of the TNF cytokine receptor family have been identiﬁed to contain cytosolic death domains. On ligand-induced clustering, these cystolic domains recruit the adaptor protein FADD, and probably others that have yet to be discovered. FADD then binds to DED-containing procaspases, particularly procaspase-8 and perhaps procaspase-10, resulting in activation of these caspases and triggering the apoptotic protease cascade. Downstream protease within this pathway can cleave and activate procaspase-1 (resulting in pro-IL-1웁 and pro-IL-1웁 processing in cells that express these cytokine precursors) and procaspase-3, which reportedly cleaves and activates pro-IL-16 in some types of cells.

often results in apoptotic elimination of cells, as well as processing of caspase-1 in some circumstances. Fas-dependent cell death mechanisms have been directly linked via animal models, and circumstantially implicated through studies of human tissues, to the autoimmune destruction

270

JOHN C. REED

of pancreatic 웁-cells in insulin-dependent diabetes mellitus (IDDM), of thyroid epithelial cells in Hashimoto’s thyroiditis, and of oligodendrocytes in multiple sclerosis. Because Fas and other cytotoxic members of the TNF receptor family rely primarily if not exclusively on caspase-8 or its close homologue caspase-10 for inducing apoptosis, inhibitors of these caspases may spare cells from autoimmune attack. A summary of the known effects of caspases on cytokine processing and their relevance to cytokine signal transduction follows. Reference is made, where available, to information concerning the physiological and pathophysiological roles of these cytokines and the corresponding caspases to immune and inﬂammatory cell regulation. The involvement of caspases in cerebral ischemia, which can involve the participation of inﬂammatory cytokines, is not addressed here. For that, the reader is referred to the work of Bergeron et al. (1998), Chen et al. (1998), Endres et al. (1998), Friedlander et al. (1997), Hara et al. (1997a,b), and Yaoita et al. (1997). A. INTERLEUKIN-1 Two forms of interleukin-1, IL-1움 and IL-1웁, are encoded by separate genes. Though the primary amino acids sequences of IL-1움 and IL-1웁 are only remotely similar, both of these cytokines bind to the same receptor(s), explaining their essentially superimpossible bioactivity proﬁles (Dinarello, 1994; Dinarello and Wolff, 1993). Both IL-1움 and IL-1웁 are produced initially as intracellular proproteins which require proteolytic processing to result in secretion from cells. These cytokines are principally produced by activated macrophages, though some other cells types can also produce IL-1웁, including microglial cells, endothelial cells, vascular smooth muscles cells, and epidermal Langerhans cells. IL-1웁 stimulates proinﬂammatory responses in neutrophils, endothelial cells, synovial cells, osteoclasts, and other cell types (Dinarello and Wolff, 1993). IL-1웁 also functions as a pyrogen, inducing febrile responses through a central mechanism. IL-1 has been implicated in a wide variety of inﬂammatory conditions in vivo. Multiple animal model experiments performed with a naturally occurring IL-1 receptor antagonist (IL-1RA), neutralizing anti-IL-1 antibodies, and soluble IL-1 receptor have demonstrated a critical role of IL1 for inﬂammatory conditions, including lipopolysaccharide (LPS)- and bacteria-induced sepsis, arthritis, inﬂammatory bowel disease, insulindependent diabetes mellitus, glomerulonephritis, allograph rejection, and graft-versus-host disease (Alexander et al., 1991; Arend, 1993; Fanslow et al., 1990; Lewthwaite et al., 1994; McCarthy et al., 1991; Ohlsson et al., 1990; Schwab et al., 1991; Wakabayashi et al., 1991; Wooley et al., 1993; Jacobs et al., 1991; Thompson et al., 1992; Van Den Berg et al., 1994). IL-1RA has also been reported to reduce osteoporotic bone loss (Kimble

CASPASES AND CYTOKINES

271

et al., 1994) and ischemia-induced neuronal cell death in rats (Relton and Rothwell, 1992). Clinical studies in patients with rheumatoid arthritis suggest that soluble IL-1R or IL-1RA may reduce joint tenderness and inﬂammatory symptoms (Campion et al., 1996; Drevlow et al., 1993, 1996; Lebsack et al., 1991, 1993). Recombinant IL-1RA, however, had no measurable beneﬁt in sepsis trials in humans (Opal et al., 1997), but it is unknown to what extent this and other shortcomings of IL-1RA and soluble IL-1R therapy can be attributed to the general pharmacodynamic and biodistribution problems associated with the administration of recombinant proteins ( Jacobs et al., 1993), as opposed to small-molecule drugs. 1. Caspase-1 and Proteolytic Processing of IL-1웁 Proteolytic processing of pro-IL-1웁 is required for its secretion and bioactivity. In vivo, the maturation of pro-IL-1웁 is mediated predominantly if not exclusively by caspase-1. LPS is a potent inducer of a signaling pathway that results in procaspase-1 processing in macrophages and microglial and endothelial cells (Schumann et al., 1998; Yao and Johnson, 1997). CD40 ligand also induces procaspase-1 processing and IL-1웁 production in vascular smooth muscle cells and endothelial cells, which could be relevant to the inﬂammatory aspects of atherosclerosis (Scho¨nbeck et al., 1997). In addition, interferon produced in the course of virus infections is an inducer of procaspase-1 synthesis. The molecular mechanisms responsible for procaspase-1 processing and activation remain largely enigmatic. At least one cellular protein (CARDIAK) has been described that binds the N-terminal prodomain of pro-caspase-1 and appears to trigger its autocatalytic processing (Burns et al., 1998). Some types of bacteria may also indirectly induce caspase-1 activation. For example, Shigella, an intracellular bacterium that infects macrophages in the gut, causing a severe form of dysentery, produces a protein ipaB that binds directly to procaspase-1, inducing its processing and activation, and thus promoting inﬂammation through production of IL-1웁 (Chen et al., 1996). In addition, granzyme-B is a potent activator of caspase-1 (Shi et al., 1996). Granzyme B is an aspartyl-speciﬁc serine protease that is stored in the cytotoxic granules of killer T cells and NK cells. This protease is ‘‘injected’’ into target cells via a perforin-dependent mechanism, resulting in processing and activation of procaspase-1 and several other members of the caspase family. Regardless of the mechanisms involved in activation of procaspase-1, once some active caspase-1 is produced, this creates opportunities for an ampliﬁcation process, because active caspase-1 efﬁciently cleaves and activates procaspase-1. Experiments involving caspase-1 gene knockout mice have revealed that this protease is absolutely required for production of IL-1웁 in vivo (Kuida

272

JOHN C. REED

et al., 1995; Li et al., 1995b). For reasons that are unclear, caspase-1 knockout mice also fail to produce interleukin-1움. The importance of this particular caspase for endotoxic shock has been demonstrated by the failure of even high doses of LPS to have a lethal effect in caspase-1-deﬁcient mice (Li et al., 1995b, 1997a). Production of IL-1움 and IL-1웁 induced by propionibacteria is also absent in caspase-1 knockout animals (Gu et al., 1997). The in vivo importance of caspase-1 for inﬂammatory diseases has also been conﬁrmed by animal model experiments using peptidyl inhibitors of this protease. Even though relatively feeble inhibitors of caspase-1 in vitro and in cell culture, carbobenzyloxy-valine-alanine-aspartate-CH2O (IC50 against caspase-1 in vitro 앑 1 애M ) and carbobenzyloxy-valine-alanineaspartate-CH2O-dichlrobenzoate (IC50 in vitro 앑 10 nM ) have been reported to reduce plasma levels of IL-1웁 and IL-1움 in mice to about onethird of control when administered 1 hr after LPS (Ku et al., 1996). These irreversible peptidyl inhibitors of caspase-1 also signiﬁcantly reduced collagen-induced arthritis in mice, both when administered prophylactically to animals before injection with Type II collagen and in animals with established disease. Moreover, caspase-1 inhibitors were superior to indomethacin and prednisone in this animal model of arthritis (Ku et al., 1996). Similarly, the caspase-1 antagonist benzyloxy-valinyl-alanyl-3(S )3-amino-4-oxo-5-(2,6-dichlorobenzoyl-oxy acid) ethyl ester, which inhibits LPS-induced IL-1웁 secretion from a monocyte cell line with an IC50 of 앑 0.24 애M, suppressed carrageenin-induced paw edema in rats by up to 60% when administered 1 hr after the inﬂammatory agent (Elford et al., 1995). This compound was also effective at preventing febrile responses in rats, presumably reﬂecting the role of IL-1 as a major endogenous pyrogen. Caspase-1 inhibitors also markedly reduce IL-1웁 production in mouse models of endotoxic shock (LPS injection) and prevent LPS-induced processing of procaspase-1 in circulating leukocytes in vivo (Fletcher et al., 1995). In vitro studies of caspase inhibitory compounds have provided further evidence that suppression this protease can prevent IL-1웁 production under conditions potentially relevant to human disease. For example, the caspase-1 inhibitor, acetyl-tyrosinyl-valinyl-alaninyl-aspartyl-CHO (AcYVAD-CHO) greatly reduces IL-1웁 production by macrophages recovered from the resected bowels of patients with inﬂammatory bowel disease (IBD) (McAlindon et al., 1998). Macrophages within the gut of IBD patients also contain processed p20/p10 caspase-1, whereas macrophages of control specimens contain only the p45 procaspase-1 zymogen, thus implying that caspase-1 is indeed active in these cells in vivo. Induction of procaspase-1 processing and IL-1 웁 secretion by LPS, the principal

CASPASES AND CYTOKINES

273

inducer of gram-negative sepsis, is also potently inhibited by YVAD-CHO or YVAD-carboxymethyl ketone (YVAD-CMK) in cultured monocytes (Schumann et al., 1998). Shigella-induced processing of procaspase-1 and apoptosis in macrophages are also inhibited effectively by YVAD-CHO (Chen et al., 1996). Though overexpression of caspase-1 can induce apoptosis (Miura et al., 1993), lack of caspase-1 does not interfere with the normal cell turnover required for tissue homeostasis; caspase-1 knockout mice develop normally and have no apparent increased incidence of tumors or other anomalies (Kuida et al., 1995; Li et al., 1995a). Though thymocytes from caspase-1deﬁcient mice exhibit partial resistance to CD95 (Fas)-induced apoptosis, these animals have no increased incidence of autoimmune or lymphoproliferative disorders (Kuida et al., 1995). Moreover, absence of caspase-1 does not interfere with any immune system functions. The T cell proliferative responses to Listeria monocytogenes and delayed-type hypersensitivity reactions, for example, are reportedly normal in caspase-1-deﬁcient mice, suggesting that the inﬂammatory cytokines generated by caspase-1dependent processing may be more relevant to macrophage-mediated inﬂammatory responses than to T cell-dependent immune reactions (Gu et al., 1997) (see below). 2. Role of Caspase-4/Caspase-11 in pro-IL-1웁 Processing Human caspase-4 and its apparent homologue in mice, caspase-11, can induce processing of pro-IL-1웁 in vitro, but with 앑250⫻ lower efﬁciency than caspase-1 (Wang et al., 1996b). Instead, caspase-4 (caspase-11) appears to induce processing of pro-IL-1웁 indirectly by dimerizing/oligomerizing with procaspase-1, resulting in activation of caspase-1 (Wang et al., 1998). In caspase-11 knockout mice, LPS fails to induce production of IL1움 and IL-1웁. Caspase-11-deﬁcient mice are also markedly resistant to lethal endotoxic shock induced by high-dose LPS. Though these animals do experience shivering, fever, and lethargy, these symptoms are less severe than in wild-type mice and 앑90% of animals recover, whereas 90% of wild-type mice die (Wang et al., 1998). It thus appears that caspase-11 (human caspase-4) is a critical upstream activator of caspase-1 in at least some in vivo contexts. B. INTERLEUKIN-6 LPS-induced production of IL-6 is somewhat reduced in caspase-1 and caspase-11 knockout mice (Li et al., 1995b; Wang et al., 1998). IL-6 production induced by IL-1 is markedly reduced in caspase-1-deﬁcient mice, probably because of a lack of pro-IL-18 processing (Ghayur et al., 1997; Gu et al., 1997).

274

JOHN C. REED

C. INTERLEUKIN-16 IL-16, a pleiotrophic cytokine produced predominantly by CD8⫹ T cells, exerts chemoattractant, growth stimulatory, and other effects on lymphocytes, monocytes, and eosinophils (Center et al., 1996). IL-16 is synthesized as a 앑56-kDa protein precursor that must undergo proteolytic processing to 앑20 kDa to attain bioactivity. The N-terminal sequencing of the mature IL-16 molecule revealed cleavage after aspartic acid residue 253, thus implicating a caspaselike protease (Zhang et al., 1998b). In vitro, pro-IL16 is cleaved at Asp-253 by caspase-3 but not by caspase-1 or -2. Processing of pro-IL-16 induced by lysates derived from activated CD8⫹ T cells is inhibited by DEVD-CHO but not by YVAD-cmk, again consistent with cleavage by a caspase-3/CED-3-like protease rather than a caspase-1-like protease (Zhang et al., 1998b). Taken together, however, these data do not exclude the possibility of processing and activation of pro-IL-16 by any of the several caspases that efﬁciently cleave DEVD, including caspase3, -7, -8, and -10 (Stennicke and Salvesen, 1997). Thus, a caspase-3-like protease rather than caspase-3 might be the physiologically relevant inducer of IL-16 processing in vivo. It is speculated that activation of caspases at subapoptotic levels may be sufﬁcient to trigger processing and secretion of IL-16 by CD8⫹ T cells, thus amplifying immune responses through its ability to attract CD4⫹ T cells (Zhang et al., 1998b). D. INTERLEUKIN-18 AND INTERFERON-웂 IL-18 is closely related to the IL-1 family of cytokines and is produced as a 앑24-kDa proprotein, which undergoes proteolytic processing at Asp36 to produce the mature 앑18-kDa cytokine that is ultimately secreted from cells (Akita et al., 1997; Gu et al., 1997). IL-18 [also known as interferon-웂-inducing factor (IGIF)] plays a major role in induction of interferon-웂 production by T cells and NK cells (Micallef et al., 1996; Ushio et al., 1996), and thus shares functional similarity with IL-12. Interferon-웂 promotes activation of macrophages and NK cells, and also contributes to the regulation of helper T cell (Th1) immune responses (Ushio et al., 1996). Interferon-웂 has been linked to the pathogens of colitis, insulindependent diabetes mellitis, and other inﬂammatory and autoimmune diseases (Campbell et al., 1991; Neurath et al., 1995). Caspase-1 and its closely related family members appear to be absolutely required for proteolytic processing of pro-IL-18. In vitro, pro-IL-18 is processed only by caspase-1, -4, and -5 (the ICE/caspase-1 subfamily) but not by other caspases typically associated with apoptosis (caspase-2, -3, -6, and -7). However, cleavage of pro-IL-18 by caspase-4 is approximately two orders of magnitude less efﬁcient than that by caspase-1 (Gu et al.,

CASPASES AND CYTOKINES

275

1997), implying that caspase-1 is the more important of these proteases for processing of pro-IL-18. Caspase-1 gene knockout mice fail to produce signiﬁcant amounts of either IL-18 or interferon-웂 in response to LPS or Propionibacterium acnes (Ghayur et al., 1997; Gu et al., 1997), demonstrating the critical role of caspase-1 not only for generation of bioactive IL18 but also indirectly for interferon-웂 production. The indirect mechanism of caspase-1-mediated production of interferon-웂 has been demonstrated by experiments in which exogenously supplied IL-18 was shown to restore interferon-웂 production in caspase-1 knockout mice (Gu et al., 1997). Kupffer cells and adherent splenocytes derived from caspase-1 knockout mice also fail to secrete detectable amounts of either IL-18 or interferon웂 (Fantuzzi et al., 1998; Gu et al., 1997). Cells from caspase-1-deﬁcient mice appear to contain normal levels of pro-IL-18 but do not process this proprotein in response to LPS or other stimuli that normally result in caspase-1 activation. In addition to evidence of the importance of caspase1 for interferon-웂 production from knockout mouse experiments, the caspase-1 inhibitory peptide, YVAD-CHO, has been reported to completely block LPS-stimulated production of interferon-웂 by peripheral blood mononuclear cells (PBMCs) in vitro at concentrations of 앑20 애M, as well as ablating production of IL-1웁 (Ghayur et al., 1997). The effects of caspase-1 gene deﬁciency on interferon-웂 production are similar to those described for IL-12 knockout mice (Magram et al., 1996). These observations, together with in vitro experiments demonstrating synergy between IL-12 and IL-18 in stimulating interferon-웂 production by leukocytes (Micallef et al., 1996; Okamura et al., 1995), have prompted speculations that the combination of both IL-12 and IL-18 is needed for production of physiologically signiﬁcant levels of interferon-웂 in vivo (Ghayur et al., 1997). Thus, ablating function of caspase-1 (1) directly eliminates production of IL-1웁 and IL-18 due to a lack of proteolytic processing, (2) indirectly and profoundly prevents production of IL-1움 and interferon-웂, and (3) may also indirectly negate some of the biological actions of IL-12 in vivo. Taken together, therefore, caspase-1 represents an attractive target for treatment of autoimmune, inﬂammatory, and other diseases when these cytokines are implicated. E. TUMOR NECROSIS FACTOR-움 TNF-움 has been implicated at some level in most inﬂammatory and autoimmune diseases, as well as in graft rejection, heart failure, and other conditions (Bazzoni and Beutler, 1996; Beutler and Cerami, 1989). LPSinduced production of TNF-움 is somewhat reduced in caspase-1 and caspase-11 knockout mice (Li et al., 1995b; Wang et al., 1998), but probably not enough to be of physiological relevance. TNF-움 exerts its effects on

276

JOHN C. REED

cells through two independent receptors: TNFR1 (CD120움), which is expressed on most types of cells, and TNFR2 (CD120b), which is found principally on lymphocytes (Tartaglia and Goeddel, 1992). TNFR1 can recruit caspase-2 and -8 to its cytosolic domain after ligand binding through interactions with the adaptor proteins, TRADD, RAIDD/CRADD, and FADD/MORT-1 (Boldin et al., 1995; Chinnaiyan et al., 1995; Duan and Dixit, 1997; Hsu et al., 1995). Once activated, both caspase-2 and -8 can induce apoptosis in some circumstances, but in general a parallel pathway involving activation of NF-␬B precludes an apoptotic response under most circumstances (Antwerp et al., 1996; Beg and Baltimore, 1996; Liu et al., 1996; Wang et al., 1996a). Thus, it is only when inhibitors of macromolecular synthesis such as cycloheximide or actinomycin D are employed that TNF-움 typically induces apoptosis in normal endothelial cells, ﬁbroblasts, keratinocytes, or other noncancerous cell types. When TNF-움 does induce apoptosis, broad-speciﬁcity caspase inhibitors such as z-VAD-fmk as well as the caspase-inhibiting proteins CrmA from cowpox (see below) or p35 from baculovirus reproducibly inhibit it (Beidler et al., 1995; Cahill et al., 1996; Orth et al., 1996; Srinivasan et al., 1998; Talley et al., 1995; Tewari et al., 1995b). However, in occasional cell types, namely L929 cells, suppression of caspases has been reported to lead instead to necrotic cell death through an oxidative mechanism following TNF treatment (Vercammen et al., 1998). Caspase-8 gene ablation in murine embryo ﬁbroblasts abrogates TNF-움-mediated apoptosis, but caspase-2 does not (Bergeron et al., 1998; Varfolomeev et al., 1998). TNF-induced production of ceramide is caspase dependent, unlike its induction of NF-␬B (Dbaibo et al., 1997; Games et al., 1998; Genestier et al., 1998; Sillence and Allan, 1997). Ceramide is a lipid second messenger that has been implicated mostly by correlative studies in a number of inﬂammatory conditions (Hannun, 1996; Smyth et al., 1997). Studies with peptidyl inhibitors of caspases and the cowpox virus protein CrmA have demonstrated that accumulation of ceramide in TNF-움-stimulated cells occurs as a downstream consequence of caspase activation. The CrmA protein binds tightly to and inhibits caspase-1 and -8. Though CrmA is capable of suppressing some other caspases at high concentrations, quantitative studies of the kinetics of inhibition suggest that only caspase-1 and -8 (not caspase-3, -6, -7, or -10) are the physiologically relevant targets of this viral protease inhibitor (Zhou et al., 1997). Thus, presumably either caspase-8 or caspase-1 is responsible for the generation of ceramide seen in TNF-움-stimulated cells. Caspase-8 triggers processing and activation of other caspases in cells, including caspase-1 under at least some circumstances, thus potentially contributing indirectly to proinﬂammatory responses (Kuida et al., 1995).

CASPASES AND CYTOKINES

277

The importance of caspase-8 in inﬂammatory diseases, however, awaits production and characterization of caspase-8 knockout mice. F. FAS LIGAND The CD95 system plays an important role in several aspects of immune system regulation. The receptor, Fas (CD95), is expressed widely throughout the body, but in a highly dynamic manner depending on the physiological or pathophysiological context (Leithauser et al., 1993). The ligand for Fas (FasL) is constitutively expressed on a few speciﬁc cell types, including neutrophils, some kinds of neurons, thyrocytes, stroma cells of the retina, acinar cells of the salivary gland, corneal epithelium, vascular endothelial cells, Paneth cells of the small intestine, and Sertoli cells of the testes. FasL can be induced on CD4⫹ and CD8⫹ T cells as well as on NK cells after activation, on macrophages infected with HIV, and on hepatocytes treated with ethanol (Badley et al., 1997; French and Tschopp, 1997). Fas is a potent inducer of apoptosis. Among its most important normal functions is as an effector of cytolytic T cell- and NK cell-mediated killing of virus-infected cells, allogeneic cells, tumors, or other cellular targets (Golstein, 1997). Fas also performs an essential function as a downregulator of peripheral T cell and B cell responses following clearance of pathogens, when it is appropriate to cease immune responses (Abbas, 1996; Van Parjis and Abbas, 1996, 1998). The coexpression of FasL and Fas on T cells is used as an autocrine or paracrine mechanism for down-regulation of T cell numbers after immune responses and also helps to eradicate potentially autoreactive lymphocytes that may arise during the postthymic life of T cells (Brunner et al., 1995; Dhein et al., 1995). B cells are also peripherally eliminated through Fas-dependent mechanisms that downregulate normal immune responses and help to ensure elimination of any autoantibody-producing cells that might arise (Rothstein et al., 1995; Schattner et al., 1995). Mutant strains of mice that carry inactivating mutations in either Fas (lpr/lpr strain) or FasL ( gld/gld strain) have been described (Takahashi et al., 1994; Watanabe-Fukunaga et al., 1992). These animals develop an age-dependent lymphoproliferative autoimmune syndrome (Abbas, 1996; Van Parjis and Abbas, 1996, 1998). Similar defects have been described in humans with autoimmune lymphoproliferative syndrome, a rare autosomal disease attributed to mutations in Fas (Fisher et al., 1995; Sneller et al., 1997). The predominant signaling mechanism used by Fas involves recruitment of procaspase-8 to the cytosolic domain of this TNFR1 homologue via its ligand-dependent interaction with the procaspase-8 binding protein FADD/MORT-1 (Wallach et al., 1997). The closely related zymogen,

278

JOHN C. REED

caspase-10, may also be recruited to the Fas cytosolic domain in some types of cells (Vincenz and Dixit, 1997). Unlike TNFR1, there is little evidence to support a role for NF-␬B in Fas-dependent signaling. Consequently, the apoptotic effects of caspase-8 are unopposed by NF-␬Binduced expression of antiapoptotic genes, and as a result Fas ligation commonly triggers apoptosis rather than a proinﬂammatory response. However, expression of Fas on the surface of cells does not guarantee sensitivity to FasL-induced apoptosis. Rather, mechanisms exist for rendering various types of cells resistant to Fas (Inohara et al., 1998; Irmler et al., 1997; Medema et al., 1998; Sato et al., 1995; Srinivasula et al., 1997; Thome et al., 1997; Torigoe et al., 1994). Indeed, the approach of creating a differential sensitivity of Fas is exploited commonly within the immune system for preventing Fas-mediated elimination of activated T and B cells that have been appropriately triggered to respond to non-self antigens, while simultaneously ensuring apoptotic elimination of aberrantly activated lymphocytes that might cause problems with self-reactivity (Lagresle et al., 1996; Peter et al., 1997; Rothstein et al., 1995). A wide variety of studies have linked caspases to apoptosis induced by Fas (Wallach et. al., 1997). Ceramide production as a result of Fas stimulation is also blocked efﬁciently by caspase inhibitors (Games et al., 1998; Genestier et al., 1998; Sillence and Allan, 1997), thus functionally placing this event downstream of caspase activation. Studies of the effects of caspase inhibitors on Fas signaling have relied primarily on either peptidyl inhibitors such as z-VAD-fmkor DEVD-CHO, which potently inhibit caspase-8 and -10, as well as many other members of the caspase family. Ectopic expression of the CrmA protein has also been employed as a technique for interfering with Fas-induced apoptosis in a wide variety of cellular contexts (Srinivasula et al., 1996; Tewari et al., 1995a,b). Because CrmA is a potent inhibitor of caspase-8 but not caspase-10 (Zhou et al., 1997; G. Salvesen, personal communication), this ﬁnding strongly suggests that caspase-8 is the predominant protease that become recruited to the cytosolic domain of Fas on ligation of this receptor has revealed only caspase-8, except under experimental conditions in which caspase-10 is overexpressed (Medema et al., 1997; Muzio et al., 1996). Though CrmA is also a potent inhibitor of caspase-1, processing of caspase-1 does not uniformly accompany Fasinduced apoptosis, thus suggesting that it is caspase-8 rather than caspase1 that is the physiologically relevant target of CrmA when used as a suppressor of Fas-induced apoptosis. Moreover, because Fas-induced apoptosis is largely intact in cells derived from caspase-1 knockout mice, it is clear that caspase-1 activation is not essential for Fas-induced apoptosis, though it may play a contributory role in some limited circumstances (Kuida et al., 1995).

CASPASES AND CYTOKINES

279

Recent gene knockout experiments indicate that caspase-8 plays an essential role in Fas-induced apoptosis, at least in embryo ﬁbroblasts (Varfolomeev et al., 1998). It must be emphasized, however, that because caspase-8 gene disruption results in embryonic lethality, at present it remains unproved that caspase-8 is essential for Fas-induced apoptosis in all cell types. Caspase-8 knockout ﬁbroblasts are also resistant to apoptosis induction by two other TNF-family cytokine receptors, TNFR1 and DR3, demonstrating that this particular caspase plays an indispensible role in apoptosis indication by these death receptors as well. The ablation of caspase-8 expression, however, does not prevent TNF-움 or DR3 induction of NF-␬B or activation of the kinase JNK, conﬁrming that caspaseindependent pathways are responsible for these signal transduction events. It has been suggested that a caspase-independent pathway can be activated by Fas in some cellular contexts, resulting Jun N-terminal kinase ( JNK) activation and facilitating cell death (Toyoshima et al., 1997; Verheij et al., 1996). Indeed, a protein, Daxx, which putatively binds to the cytosolic domain of Fas and which induces JNK activation (Yang et al., 1997), has been described. However, the evidence that this or related ‘‘stress kinase’’ pathways plays a substantial role in Fas-induced cytotoxicity is highly controversial (Kiriakidou et al., 1997) and most studies put JNK activation down stream of caspases (Cahill et al., 1996; Liu et al., Nishina et al., 1997). On balance, therefore, the preponderance of the data available to date strongly implicate caspase-8 as the initiator of Fas-induced apoptosis. Nevertheless, Fas may have other functions besides induction of apoptosis. In this regard, Fas has been shown to provide costimulatory signals that promote T cell proliferation or lymphokine production in some scenarios (Alderson et al., 1993; Yeh et al., 1998; Zhang et al., 1998a). Moreover, gene knockout experiments that accomplished inactivation of the adaptor protein FADD/MORT-1, which binds to the cytosolic domain of Fas on liganding with Fas L or cross-linking anti-Fas antibodies, have demonstrated an unexpected requirement for this protein for proliferative responses of T cells, in addition to the anticipated requirement for Fasinduced apoptosis (Yeh et al., 1998; Zhang et al., 1998a). Thus, FADD/ MORT-1 may have two functions: (1) binding to and activating pro caspase8 (and possibly pro caspase-10) and (2) stimulating T cell proliferation through as yet unknown mechanisms. Thus, though very often informative for analysis of the in vivo roles of Fas and Fas L, the conclusion that must be reached is that predicting the in vivo consequences of caspase-8 inhibition based on observations from autoimmune strains of mice with defects in Fas (lpr/lpr) or Fas L ( gld/gld) or based on FADD knock outs therefore is not without its caveats. In essence, caspase-8 is not synonymous

280

JOHN C. REED

with Fas or FADD, though the collaboration of these three proteins is probably essential for Fas-induced apoptosis. An alternative source of information about what suppression of caspase8 might do in vivo comes from studies of transgenic mice that express the CrmA protein in various tissues. For example, expression of CrmA in thymocytes and T cells under the control of the CD2 gene enhancer (a T-lineage-speciﬁc transcriptional element) resulted in protection from Fasinduced apoptosis but not cell death induced by other stimuli as 웂irradation or glucocorticoids (Smith et al., 1996). Unlike Fas-defective lpr/ lpr mice, the CrmA transgenics also did not develop autoimmune disease, lymphadenopathy, or T cell hyperplasia. Also, unlike FADD knockout mice, the T cells of these CrmA transgenic mice proliferate normally in vitro in response to mitogens, thus further demonstrating the functional differences between elimination of FADD and suppression of caspases. Finally, CrmA transgenics show no evidence of problems with thymic education against self-antigens and do not develop autoantibodies, unlike lpr/lpr mice (Smith et al., 1996). As mentioned above, when interpreting these data obtained with CrmA transgenic mice, it is important to recall that this viral serpin like protein is a suppressor of both caspase-1 and caspase-8. However, because T cells are not involved in IL-1 production, it is unlikely that expression of CrmA selectively in T cells has any inhibitory inﬂuence on IL-1 production that might account for the ﬁndings. Taken together, the available data therefore suggest that selective inhibition of caspase-8 is unlikely to result in the autoimmune and lymphoproliferative problems seen in Fas- or FasL-deﬁcient animals. 1. Fas and Disease Alterations in the regulation of the FasL/Fas system have been implicated in several diseases, including autoimmune disorders, immunodeﬁciency associated with HIV infection, mycobacterium tuberculosis, viral and chemical hepatitis, inﬂammatory bowel disease, allograph rejection, and graft-versus-host disease. A summary of some of the evidence implicating Fas-based cytotoxic mechanisms in some of these disorders follows. a. Insulin-Dependent Diabetes Mellitus. Insulin-dependent diabetes mellitus (IDDM) is an autoimmune disease resulting from immune cellmediated eradication of insulin-producing 웁cells within the pancreatic islets (Benoist and Mathis, 1997; Tisch and McDevitt, 1996). A variety of animal experiments, including adoptive cell transfer studies, indicate that IDDM is a T cell-dependent autoimmune disorder (reviewed in (Chervonsky et al., 1997; De Maria and Testi, 1998)). Though controversial, antecedent viral infection in individuals with an HLA-determined genetic predispo-

CASPASES AND CYTOKINES

281

sition may initiate an inﬂammatory reaction that affects the islets (Benoist and Mathis, 1997). Normally, islet cells do not express Fas. However, in the setting of insulinitis that precedes IDDM in animal models (non obese mice and diabetes-prone rats), Fas expression is found on 웁 cells but not on the 움 or ␦ cells (Stassi et al., 1997). Moreover, IL-1웁, which is believed to be produced locally by inﬂammatory cells in the islets, has been shown to induce Fas expression selectively on 웁 cells but not on 움 cells or ␦ cells in vitro (Stassi et al., 1997). Thus, caspase-1-mediated processing of proIL-1웁 may contribute to the pathogenesis of IDDM by inducing Fas expression on insulin-producing 웁 cells. During the insulinitis process, T cells found within the inﬂamed islets express FasL and adjacent islet cells undergoing apoptosis are readily detectable (Stassi et al., 1997). In vivo evidence of an essential role for Fas in the autoimmune destruction of 웁 cells has come from studies of lpr/lpr nonobese diabetic mice that lack functional Fas and that are completely resistant to both spontaneous diabetes and diabetes induced by adoptive transfer of islet reactive T cells (Chervonsky et al., 1997; Itoh et al., 1997). b. Multiple Sclerosis. Multiple sclerosis (MS) is an incurable and inevitably fatal demyelinating disease of the central nervous system; it is caused by autoimmune destruction of myelin-producing oligodendrocytes. Much of the effort to understand the pathogenesis of MS has relied on experimental induction of experimental allergic encephalomyelitis (EAE) in various animal species, but immunophenotyping of MS lesions from humans has also provided insights. In animal models, MS can be adoptively transferred by myelin-reactive CD4⫹ T cells (Martin et al., 1992; Steinman, 1996). Thus, T cells initiate the disease, though other types of cells are known to contribute subsequently, including IL-1웁-producing macrophages and glial cells. Normally, oligodendrocytes do not express Fas. However, in both acute and chronic MS plaques from humans, intense Fas immunostaining is present on these cells (D’Souza et al., 1996; Dowling et al., 1996). In addition, FasL-expressing cells are typically found adjacent to Fasexpressing oligodendrocytes, with the highest and most reproducible FasL immunostaining appearing on resident microglial cells and inﬁltrating T cells (D’Souza et al., 1996; Dowling et al., 1996). Direct evidence of an important role for Fas in MS has come from studies of Fas-deﬁcient (lpr/ lpr) and FasL-deﬁcient ( gld/gld ) mice, which were shown to be markedly resistant to EAE induction (Sabelko et al., 1997; Waldner et al., 1997). c. Hashimoto’s Thyroiditis. Hashimoto’s thyroiditis (HT) is an inﬂammatory disorder involving destruction of thyroid epithelial cells. Interestingly, thyroid epithelial cells constitutively express FasL (Giordano et

282

JOHN C. REED

al., 1997). Current models of HT propose that during the idiopathic inﬂammatory process that initiates HT, local production of IL-1웁 (caspase1) by macrophages induces thyroid epithelial cells to express high levels of Fas. The coexpression of FasL and Fas on these cells then results in an autocrine destruction of the thyroid epithelial cells (De Maria and Testi, 1998; Arscott and Baker, 1998). Thus, as in IDDM and MS, the pathogenesis of this disorder may involve the participation of both caspase1 in its capacity as a producer of bioactive IL-1웁, which induces expression of Fas, followed by caspase-8-mediated apoptosis of the Fas-expressing cells. d. Acquired Immunodeﬁciency Syndrome. Hyperactivity of the Fas/ FasL system has been implicated in the T cell depletion that accompanies human immunodeﬁciency virus (HIV) infection and that culminates in acquired immunodeﬁciency syndrome (AIDS). It is estimated that ⬎99% of the T cells that die in persons with HIV are not infected by HIV, implicating host mechanisms in most of the immune cell depletion that occurs (Finkel and Banda, 1994; Finkel et al., 1995; Gougeon et al., 1993; Gougeon and Montagnier, 1993; Lewis et al., 1994; Li et al., 1995a). Furthermore, though HIV infects only CD4⫹ T cells, equivalent loss of both CD4⫹ and CD8⫹ cells is observed in HIV-infected persons. The CD4⫹ as well as CD8⫹ T cells circulating in the peripheral blood of asymptomatic HIV carriers have been shown to express elevated levels of Fas and exhibit increased sensitivity to anti-Fas antibody-induced apoptosis compared to lymphocytes from uninfected individuals (Boudet et al., 1996; Debatin et al., 1994; Estaquier et al., 1996; Gougeon et al., 1996; Katsikis et al., 1995). Levels of surface Fas antigen expression on CD4⫹ T cells from HIV-infected persons have also been reported to increase with clinical progression (Aries et al., 1995). HIV infection of monocyte/macrophage cells induces expression of FasL, and allows these cells to kill Fas-expressing T cells in vitro (Badley et al., 1996). Moreover, abundant surface expression of FasL has been detected by immunostaining on macrophages and dendritic cells within nodes of HIV-infected persons. Anti-gp120 antibodies found in the plasma of HIV-infected patients also may cross-react with Fas, thus triggering apoptosis of T cells (Stricker et al., 1998). Additional connections between HIV and Fas-induced apoptosis have been suggested by the observations that the HIV tat protein and gp120 envelope protein sensitize activated T cells to Fas-induced apoptosis (Westendorp et al., 1995). The broad-spectrum caspase inhibitor z-VAD-fmk can prevent HIVinduced apoptosis of primary T cells and T cell lines in vitro (Chinnaiyan et al., 1997; Glynn et al., 1996). However, suppression of HIV-induced

CASPASES AND CYTOKINES

283

apoptosis is also associated with enhanced viral replication, at least during acute infection in vitro (Chinnaiyan et al., 1997). Thus, it remains unclear whether suppression of caspases would be overall helpful versus counterproductive in the setting of HIV infection. However, the clear evidence that the vast preponderance of T cell death in the setting of HIV involves uninfected cells, together with the possibility of using caspase inhibitors in combination with HIV reverse transcriptase and HIV protease inhibitors, suggests that this approach requires further examination. e. Graft-versus-Host Disease. Total body irradiation followed by allogenic bone marrow transplant is a highly effective and potentially curative therapy for several hematological malignancies. Unfortunately, most patients that receive allogenic bone marrow transplants ultimately suffer from graft-versus-host disease (GVHD), which can be fatal in many cases. Transfer of allogenic T cells that contaminate bone marrow preparations into irradiated recipients induces depletion of host lymphocytes, destruction of intestinal epithelial cells, skin inﬂammation, hepatitis, and other symptoms. Fas has been strongly implicated in GVHD. For example, transfer of allogenic T cells from gld/gld mice that lack functional FasL induces signiﬁcantly less antihost lymphoid depletion and other manifestations of GVHD (Baker et al., 1997; Braun et al., 1996; Via et al., 1996). FasL-expressing T cells have also been detected at markedly elevated levels in vivo in the intestinal submucosa in mouse models of acute GVHD, adjacent to intestinal epithelial cells that constitutively express Fas (Sakai et al., 1997), and within the livers of donor mice that develop GVHDassociated hepatitis (Bobe et al., 1997). Moreover, intestinal epithelial cells and Fas-expressing hepatocytes are killed in vitro by donor T cells recovered from the intestine or livers of mice suffering from GVHD, through a mechanism that is largely suppressible by Fas/Fc chimeric protein (Bobe et al., 1997; Sakai et al., 1997). Conversely, transfer of normal allogeneic donor T cells into sublethally irradiated Fas-deﬁcient lpr/lpr recipients results in signiﬁcantly less intestinal epithelial cell apoptosis in vivo (Lin et al., 1998), further suggesting an important role for the Fas/ FasL system in GVHD. In a mouse model of GVHD, neutralizing antiFas and anti-TNF antibodies each individually reduced symptoms, but their combination was superior (Hattori et al., 1998). Given that Fas and TNFR1 activate common caspases, therefore, it might be speculated that the superiority of combined anti-Fas and anti-TNF therapy is attributable to redundancy in the mechanisms used by these cytokines receptors. If true, this hypothesis suggests that caspase inhibitors should be particularly effective at inhibiting GVHD. However, it is also possible that TNF-

284

JOHN C. REED

induced activation of NF-␬B or other caspase-independent signaling pathways plays a role in GVHD. It has been suggested, based on indirect correlative evidence in humans, that Fas-induced apoptosis may also be helpful for limiting the expansion of donor-derived alloreactive T cells, in which case suppression of Fasinduced apoptosis might be counterproductive under some circumstances in patients with GVHD (Dey et al., 1998). In direct opposition, however, Fas ligand has been associated with expansion of donor T cells in vivo in mouse models of GVHD; mice lacking FasL ( gld/gld ) reportedly have less accumulation of donor T cells (Via et al., 1996). The controversial effects of the Fas/FasL system on expansion of donor T cells not withstanding, it is clear that Fas is an important effector of cytotoxicity of host immune and non immune cells (Via et al., 1996), and thus interfering with Fas-induced apoptosis should signiﬁcantly spare host tissues from destruction during GVHD. f. Cardiovascular Disease. Studies of the role of Fas in cardiovascular diseases are limited but provocative (Bromme and Holt, 1996; Umansky and Tomei, 1997). For example, survival of cardiac allographs is substantially prolonged when FasL-deﬁcient ( gld/gld ) or Fas-deﬁcient (lpr/lpr) mice are used as the recipients, suggesting that host immune cells rely on the Fas pathway for rejection of histoincompatable heart transplants (Seino et al., 1996). Up-regulation of Fas levels on cardiac myocytes has also been reported following ischemic injury and by stretching muscle in vitro to simulate chronic heart failure conditions (Cheng et al., 1995). Changes in the plasma levels of soluble Fas have also been associated with myocarditis and chronic congestive heart failure in humans (Nishigaki et al., 1997; Toyozaki et al., 1998). g. Hepatic Diseases. The notion that Fas could play an important role in liver diseases was ﬁrst suggested by studies of mice injected with agonistic anti-Fas antibodies. Injection with Fas-triggering antibodies triggers massive hepatocyte apoptosis, with ensuing liver hemorrhage and death (Ogasawara et al., 1993). Evidence that Fas may be involved in hepatic diseases in humans comes from several lines of investigation, including in vitro studies of primary cultured hepatocytes, mouse models of sepsis and other diseases, and analysis of human liver biopsy materials. Wilson’s disease results from hereditary deﬁciency of a copper-carrier protein, and causes accumulation of this metal in the liver with resultant hepatic failure. In Wilson’s disease patients experiencing fulminant hepatic failure, elevated levels of Fas and FasL mRNA are found in the liver as well as accompanying apoptosis (Strand et al., 1998). In vitro, treatment

CASPASES AND CYTOKINES

285

of cultured hepatoma cells with copper results in apoptosis, which is largely blockable with neutralizing Fas/Fc chimeric antibodies. Fulimnant hepatic failure associated with acute toxic exposure to various chemicals or pharmaceuticals has also been associated with increased immunostaining for Fas on the surface of hepatocytes and concominant FasL expression on inﬁltrating mononuclear cells (Galle et al., 1995). LPS induces expression of FasL on hepatic Kupffer cells and sinusoidal endothelial cells, but not parencymal hepatocytes in primary cultures. Though not inducing FasL, LPS does induce marked increases in Fas surface expression on hepatocytes (Muschen et al., 1998). Moreover, primary cultured rat and human hepatocytes are readily induced to undergo apoptosis by agonistic anti-Fas antibodies (Fladmark et al., 1997; Galle et al., 1995). The LPS induction of Fas on hepatocytes is not ameliorated by dexamethasone treatment; rather, glucocorticoids appear to further enhance the LPS-induced expression of Fas on these cells (Muschen et al., 1998). Administration of endotoxin to mice results in liver failure and parechymal cell apoptosis followed by secondary inﬁltration of neutrophils. In mouse models, endotoxin has been shown to induce 앑17-fold elevations in the levels of DEVD-cleaving caspases in liver, with no apparent change in caspase-1-like activity ( Jaeschke et al., 1998). In contrast, the livers of endotoxin-resistant C3H/HeJ mice do not accumulate caspase activity and parenchymal hepatocytes fail to undergo apoptosis in vivo following endotoxin administration. Direct evidence of caspase involvement in sepsisassociated liver failure has come from experiments involving treatment of mice with the caspase inhibitor z-Val-Ala-Asp-CH2F (10 mg/kg) at 3 hr after administration of endotoxin (and galactosamine), which resulted in attenuation of liver apoptosis by ⬎80%and reductions in liver necrosis (area involved) by ⬎95% ( Jaeschke et al., 1998). An effective treatment for chronic hepatitis caused by hepatitis B or C virus infection represents one of the largest unmet medical needs worldwide. Immunohistochemistry and in situ hybridization analysis of human liver biospies have revealed up-regulation of Fas on hepatocytes and expression of Fas-L mRNA in inﬁltrating lymphocytes in patients with HBVrelated cirrhosis (Galle et al., 1995). TGF-웁, which is produced by activated T cells and various immune or inﬂammatory cells, is known to induce apoptosis of primary hepatocytes. In a hepatoma model, TGF-웁-induced apoptosis was nearly completely suppressed by expression of the caspase-inhibitory CrmA proteins or by the peptidyl caspase inhibitor z-VAD-fmk (Chen and Chang, 1997). TGF웁-induced proteolysis of intracellular proteins known to be substrates of caspases was also blocked by these caspase-inhibiting agents. Oxidative stress, a common accompaniment of inﬂammatory liver diseases, has also

286

JOHN C. REED

been reported to induce expression of FasL on the surface of cultured hepatoma cells (Hug et al., 1997). Alcohol abuse ranks second only to viral hepatitis among causes of cirrhosis and liver failiure worldwide. In situ hybridization studies of liver specimens from persons suffering from alcoholic cirrhosis suggest that FasL mRNA levels are increased in hepatocytes (Galle et al., 1995). Because hepatocytes constitutively express Fas, the receptor for FasL, this Finding raises the possibility that Fas-mediated hepatocyte cell death may contribute to cirrhosis in the setting of chronic ethanol abuse. h. Inﬂammatory Bowel Disease. Fas is constitutively expressed on colonic and intestinal epithelial cells. In vitro, Fas ligation induces rapid apoptosis of cultured colonic epithelial cells (Strater et al., 1997). Though few FasL-expressing cells are normally found in the lamina propria (submucosa) of the bowel, in ulcerative colitis FasL-expressing mononuclear cells are plentiful and epithelial cell apoptosis is readily detectable in biopsy specimens (Strater et al., 1997). Involvement of Fas in intestinal epithelial cell apoptosis induced by T cells has also been documented in the setting of GVHD (see above). Though further work is needed, these observations suggest a potentially prominent role for the Fas/FasL system in some types of inﬂammatory bowel diseases. G. DR3 LIGAND The DR3 legand (Apo-3L) is a ligand for the TNF-R1 homologue known as DR3, Apo-3, or Weasle (Chinnaiyan et al., 1996; Kitson et al., 1996; Marsters et al., 1998). DR3 shares many similarities with TNF R1 in its signal transduction apparatus (Chinnaiyan et al., 1996), including binding to its cytosolic domain of TRADD, and recruitment of the caspase-8 activator FADD/MORT-1 and the NF-␬B-activating protein TRAF-2. Little is known about the physiological roles of Apo-3L or its receptor in normal immune responses or inﬂammatory conditions. As mentioned above, caspase-8 gene disruption abrogates DR3-induced apoptosis in embryonic ﬁbroblasts (Varfolomeev et al., 1998). H. TRAIL The TNF family cytokine TRAIL binds at least four receptors, including DR4 and DR5, which are capable of inducing apoptosis through mechanisms resembling those used by Fas (MacFarlane et al., 1997; Pan et al., 1997 a, b; Sheridan et al., 1997). As with Fas, peptidyl inhibitors of caspases are efﬁcient inhibitors of TRAIL-induced apoptosis (Mariani et al., 1997). At present, essentially nothing is known about the role of TRAIL and its receptors in inﬂammatory and autoimmune diseases. However, a recent

CASPASES AND CYTOKINES

287

report that examined mechanisms of HIV-induced apoptosis of T cells in vitro provided evidence that both Fas and TRAIL-dependent mechanisms can be involved, both of which are blockable by the caspase inhibitor zVAD-fmk (Katsikis et al., 1997). Also, expression of both FasL and TRAIL (Apo-2L) is up-regulated on T cells after activation in vitro (Mariani and Krammer, 1998), implying that both of these cytotoxic molecules are commonly available for mediating T cell-based killing and suggesting that it may, therefore, be necessary to interfere with shared postreceptor signaling events, such as caspase-8 activation, for effective obviation of detrimental T cell responses. Further support for this idea of redundancy created by co-expression of FasL and TRAIL comes from studies showing that Fasresistant cell lines may still be killed by TRAIL, though both Fas- and TRAIL-induced apoptosis are effectively blocked by caspase inhibitors (Mariani et al., 1997). IV. Conclusions

Caspases are absolutely required for the production of several proinﬂammatory cytokines, including IL-1움, IL-1웁, IL-18, and probably interferon-웂. These cysteine proteases also play a requisite role in the signaling mechanisms by which several TNF family receptors, including those for TNF움, FasL, Apo-3L, and TRAIL, induce cellular apoptosis and tissue destruction. Human diseases involving these cytokines therefore are potentially amenable to modulation by caspase inhibitors. Further investigations of the effects of caspase gene inactivation in mice and evaluation of both selective and broad-speciﬁcity caspase inhibitory smallmolecule compounds in preclinical animals models of inﬂammatory, autoimmune, and infectious diseases are now required to probe in greater detail the in vivo roles of this family of intracellular proteases. From this analysis, clinical indications for caspase-inhibiting drugs can be deduced and appropriate clinical trials initiated to test their efﬁcacy in humans. REFERENCES Abbas, A. K. (1996). Die and let live: Eliminating dangerous lymphocytes. Cell 84, 655–657. Akita, K., Ohtsuki, T., Nukada, Y., Tanimoto, T., Namba, M., Okura, T., Takakura-Yamamoto, R., Torigoe, K., Gu, Y., Su, M. S. S., Fujii, M., Satoh-Itoh, M., Yamamoto, K., Kohno, K., Ikeda, M., and Kurimoto, M. (1997). Involvement of caspase-1 and caspase-3 in the production and processing of mature human interleukin 18 in monocytic THP.1 cells. J. Biol. Chem. 272, 26595–26603. Alderson, M. R., Armitage, R. J., Maraskovsky, E., Tough, T. W., Roux, E., Schooley, K., Ramsdell, F., and Lynch, D. H. (1993). Fas transduces activation signals in normal human T lymphocytes. J. Exp. Med. 178, 2231–2235.

288

JOHN C. REED

Alexander, H., Doherty, G., Burcsh, C., Venzon, D., and Norton, J. (1991). A recombinant human receptor anatagonist for interleukin-1 improves survival after lethal endotoxemia in mice J. Exp. Med. 173, 1029–1032. Alnemri, E. S. (1997). Mammalian cell death proteases: A family of highly conserved aspartate speciﬁc cysteine proteases. J. Cell Biochem. 64, 33–42. Antwerp, D. J., Martin, S. J., Kafri, T., Green, D. R., and Verma, I. M. (1996). Suppression of TNF-움-induced apoptosis by NF-␬B, Science 274, 787–789. Arend, W. (1993). Interleukin-1 receptor antagonist, Adv. Immunol. 54, 167–227. Aries, S. P., Schaaf, B., Muller, C., Dennin, R. H., and Dalhoff, K. (1995). Fas (CD95) expression on CD4⫹ T cells from HIV-infected patients increases with disease progression. J. Mol. Med. 73, 591–593. Arscott, P. L., and Baker, J. R. (1998). Apoptosis and thyroiditis. Clin. Immunol. Immunopathol. 87, 207,–217. Badley, A. D., McElhinny, J. A., Leibson, P. J., Lynch, D. H., Alderson, M. R., and Paya, C. V. (1996). Upregulation of Fas ligand expression by human immunodeﬁciency virus in human macrohpages mediates apoptosis of uninfected T lymphocytes. J. Virol. 70, 199–206. Badley, A., Dockrell, D., Simpson, M., Schut, R., Lynch, D., Leibson, P., and Paya, C. (1997). Macrophage-dependent apoptosis of CD4⫹ lymphocytes from HIV-infected individuals is mediated by FasL and tumor necrosis factor. J. Exp. Med. 185, 55–64. Baker, M., Rilet, R., Podack, L., and Levy, R. (1997). Graft versus host-disease-associated lymphoid hypoplasia and B-cell disfunction is dependent upon donor T-cell-mediated Fas-ligand function, but not perforin function. Proc. Natl. Acad. Sci. U.S.A. 94, 1366–1371. Bazzoni, F., and Beutler, B. (1996). The tumor necrosis factor ligand and receptor families. N. Engl. J. Med. 334, 1717–1725. Beg, A. A., and Baltimore, D. (1996). An essential role for NF-␬B in preventing TNF움induced cell death. Science 274, 782–784 Beidler, D. R., Tewari, M., Friesen, P. D., Poirier, G., and Dixit, V. M. (1995). The baculovirus p35 protein inhibits Fas and tumor necrosis factor-induced apoptosis. J. Biol. Chem. 270, 1–3. Benoist, C., and Mathis, D. (1997). Cell death mediators in autoimmune diabetes—No shortage of suspects. Cell 89, 1–3. Bergeron, L., Perez, G. I., Macdonald, G., Shi, L., Sun, Y., Jurisicova, A., Varmuza, S., Latham, K. E., Flaws, J. A., Salter, J. C. M., Hara, H., Moskowitz, M. A., Li, E., Greenberg, A., Tilly, J. L., and Yuan, J. (1998). Defects in regulation of apoptosis in caspase-2deﬁcient mice. Genes Dev. 12, 1304–1314. Beutler, B., and Cerami, A. (1989). The biology of cachectin/TNF—A primary mediator of the host response Annu. Rev. Immunol. 7, 625–655. Bobe, P., Bonardelle, D., Reynes, M., Godeau, F., Mahiou, J., Joulin, V., and Kiger, N. (1997). Fas mediated liver damage in mrl hemopoietic chimeras undergoing lpr-mediated graft-versus-host disease. J. Immunol. 159, 4197–4204. Boldin, M. P., Varfolomeev, E. E., Pancer, Z., Mett, I. L., Camonis, J. H., and Wallach, D. (1995). A novel protein that interacts with the death domain of Fas/APO1 contains a sequence motif related to the death domain. J. Biol. Chem. 270, 7795–7798. Boldin, M. P., Goncharov, T. M., Goltsev, Y. V., and Wallach, D. (1996). Involvement of MACH, a novel MORT1/FADD-interacting protease, in Fas/APO-1 and TNF receptorinduced cell death. Cell 85, 803–815. Boudet, F., Lecoeur, H., and Gougeon, M.-L. (1996). Apoptosis associated with ex vivo down regulation of Bcl-2 and up-regulation of Fas in potential cytotoxic CD8⫹ T lymphocytes during HIV infection. J. Immunol. 156, 2282–2293.

CASPASES AND CYTOKINES

289

Braun, M. Y., Lowin, B., French, I., Acha-Orbea, H., and Tschopp, J. (1996). Gytotoxic Tcells deﬁcient in both functional fas ligand and perforion show residual cytolytic activity yet lose their capacity to induce lethal acute graft-versus-host disease. J. Exp. Med. 183, 657–661. Bromme, H., and Holt, J. (1996). Apoptosis in the heart: When and why? Mol. Cell. Biochem. 163–164, 261-275. 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 (London) 373, 441–444. Burns, K., Martinon, F., Esslinger, C., Pahl, H., Schneider, P., Bodmer, J.-L., Di Marco, F., French, L., and Tschopp, J. (1998). MyD88, an adapter protein involved in interleukin 1 signaling. The J. Biol. Chem. 273, 12203–12209. Cahill, M. A., Peter, M. E., Kischkel, F. C., Chinnaiyan, A. M., Dixit, V. M., Krammer, P. H., and Nordheim, A. (1996). CD95 (APO 1/Fas) induces activation of SAP kinases downstream of ICE-like proteases. Oncogene 13, 2087–2096. Campbell, I. L., Kay, T. W., Oxbrow, L., and Harrison, L. C. (1991). Essential role for interferon-gamma and interleukin-6 in autoimmune insulin-dependent diabetes in NOD/ Wehi mice. J. Clin. Invest. 87, 739–742. Campion, G. V., Lebsack, M. E., Lookabaugh, J., Gordon, G., and Catalano, M. (1996). Dose-range and dose-frequency study of recombinant human interleukin-1 receptor antagonist in patients with rheumatoid arthritis. The IL-1Ra Arthritis Study Group. Arthritis Rheum. 39, 1092–1101. Center, D., Kornfeld, H., and Cruikshank, W. (1996). Interleukin 16 and its function as a CD4 ligand. Immunol. Today 17, 476–481. Chen, R. H., and Chang, T. Y. (1997). Involvement of caspase family proteases in transforming growth factor-beta-induced apoptosis. Cell Growth Diff. 8, 821–827. Chen, Y., Smith, M. R., Thirumalai, K., and Zychlinsky, A. (1996). A bacterial invasin induces macrophage apoptosis by binding directly to ice. Embo J. 15, 3853–3860 Chen, J., Nagayama, T., Jin, K., Stetler, R., Zhu, R., Graham, S., and Simon, R. (1998). Induction of caspase-3-like protease may mediate delayed neuronal death in the hippocampus after transient cerebral ischemia. J. Neurosci. 18, 4914–4928. Cheng, W., Li, B., Kajstura, J., Li, P., Wolin, M. S., Sonnenblick, E. H., Hintze, T. H., Olivetti, G., and Anversa, P. (1995). Stretch-induced programmed myocyte cell death. J. Clin. Invest. 96, 2247–2259. Chervonsky, A. V., Wang, Y., Wong, F. S., Visintin, I., Flavell, R. A., Janeway, J. C. A., and Matis, L. A. (1997). The role of Fas in autoimmune diabetes. Cell 89, 17–24. Chinnaiyan, A. M., O’Rourke, K., Tewari, M., and Dixit, V. M. (1995). FADD, a novel death domain-containing protein, interacts with the death, domain of Fas and initiates apoptosis. Cell 81, 505–512. Chinnaiyan, A. M., O’Rourke, K., Yu, G.-L., Lyons, R. H., Garg, M., Duan, R., Xing, L., Gentz, R., Ni, J., and Dixit, V. M. (1996). Signal transduction by DR3, a death domaincontaining receptor related to TNFR-1 and CD95. Science 274, 990–992. Chinnaiyan, A. M., Woffendin, C., Dixit, V. M., and Nabel, G. J. (1997). The inhibition of pro-apoptotic ICE-like proteases enhances HIV replication, Nature Med. 3, 333–337. Dbaibo, G. S., Perry, D. K., Garnard, C. J., Platt, R., Poirier, G. G., Obeid, L. M., and Hannun, Y. A. (1997). Gytokine response modiﬁer A (CrmA) inhibits ceramide formation in response to tumor necrosis factor (TNF)-움: CrmA and Bcl-2 target distinct components in the apoptotic pathway J. Exp. Med. 185, 481–490.

290

JOHN C. REED

Debatin, K.-M., Fahrig-Faissner, A., Enenkel-Stoodt, S., Kreuz, W., Benner, A., and Krammer, P. H. (1994). High expression of APO-1 (CD95) on T lymphocytes from human immunodeﬁciency virus-1-infected children, Blood 3, 3101–3103. De Maria, R., and Testi, R. (1998). Fas–FasL interactions: A common pathogenetic mechanism in organ speciﬁc autoimmunity. Immunol. Today 19, 122–126. Dey, B., Yang, Y.-G., Szot, G., Pearson, D., and Sykes, M. (1998). Interleukin-12 inhibits graft-versus-host disease through a fas-mediated mechanism associated with alterations in donor T-cell activation and expansion. Blood 91, 3315–3322. Dhein, J., Walczak, H., Ba¨umler, C., Debatin, K.-M., and Krammer, P. H. (1995). Autocrine T-cell suicide mediated by APO-1 /(Fas/CD95). Nature (London) 373, 438–441. Dinarello, C. (1994). The interleukin-1 family: 10-years of discovery FASEB J. 8, 1314–1325. Dinarello, C., and Wolff, S. (1993). The role of interleukin-1 in disease. N. Engl. J. Med. 328, 106–113. Dowling, P., Shang, G., Raval, S., Menonna, J., Cook, S., and Husar, W. (1996). Involvement of the CD95 (APO-1/Fas) receptor/ligand system in multiple sclerosis brain. J. Exp. Med. 184, 1513–1518. Drevlow, B., Capezio, J., Lovis, R., Jacobs, C., Landay, A., and Pope, R. (1993). Phase I study of recombinant human interleukin-1 receptor administered intra-articularly in active rheumatoid arthritis. Arthritis Rheum. 36, S39. Drevlow, B. E., Lovis, R., Haag, M. A., Sinacore, J. M., Jacobs, C., Blosche, C., Landay, A., Moreland, L. W., and Pope, R. M. (1996). Recombinant human interleukin-1 receptor type I in the treatment of patients with active rheumatoid arthritis. Arthritis Rheum. 39, 257–265. D’Souza, S., Bonetti, B., Balasingam, V., Cashman, N., Barker, P., Troutt, A., Raine, C., and Antel, J. (1996). Multiple sclerosis: Fas signaling in oligodendrocyte cell death. J. Exp. Med. 184, 2361–2370. Duan, H., and Dixit, V. M. (1997). RAIDD is a new; ‘‘death’’ adaptor molecule. Nature (London) 385, 86–89. Elford, P. R., Heng, R., Revesz, L., and MacKenzie, A. R. (1995). Reduction of inﬂammation and pyrexia in the rat by oral administration of SDZ-224-015, an inhibitor of the interleukin1 beta converting enzyme. Br. J. Pharmacol. 115, 601–606. Endres, M., Namura, S., Shimizu-Sasamata, M., Waeber, C., Zhang, L., Gomez-Isla, T., Hyman, B., and Moskowitz, M. (1998). Attentuation of delayed neuronal death after mild focal ischemia in mice by inhibition of the caspase Family. J. Cereb. Blood Flow Metab. 18, 238–247. Estaquier, J., Tanaka, M., Suda, T., Nagata, S., Golstein, P., and Ameisen, J. C. (1996). Fas-mediated apoptosis of CD4⫹ and CD8⫹ T cells from human immunodeﬁciency virusinfected persons: Differential in vitro preventive effect of cytokines and protease antagonists. Blood 87, 4959–4966. Fanslow, W., Sims, J., Sassenﬁeld, H., Morrissey, P., Gillis, S., Dower, S., and Widmer, M. (1990). Regulation of alloreactivity in vivo by a soluble form of the interleukin 1 receptor. Science 248, 739–742. Fantuzzi, G., Puren, A. J., Harding, M. W., Livingston, D. J., and Dinarello, C. A. (1998). Interleukin-18 regulation of interferon gamma production and cell proliferation as shown in interleukin-1beta-converting enzyme (caspase-1)-deﬁcient mice. Blood 91, 2118–2125. Finkel, T. H., and Banda, N. K. (1994). Indirect mechanisms of HIV pathogenesis: How does HIV kill T cells? Curr. Opin. Immun 6, 605–615. Finkel, T. H., Tudor-Williams, G., Banda, N. K., Cotton, M. F., Curiel, T., Monks, C., Baba, T. W., Ruprecht, R. M., and Kupfer, A. (1995). Apoptosis occurs predominantly

CASPASES AND CYTOKINES

291

in bystander cells and not in productively infected cells of HIV- and SIV-infected lymph nodes. Nature Med. 1, 129–134. Fisher, G. H., Rosenberg, F. J., Straus, S. E., Dale, J. K., Middelton, L. A., Lin, A. Y., Strober, W., Lenardo, M. J., and Puck, J. M. (1995). Dominant interfering fas gene mutations impair apoptosis in a human autoimmune lymphoproliferative syndrome. Cell 81, 935–946. Fladmark, K. E., Gjertsen, B. T., Doskeland, S. O., and Vintermyr, O. K. (1997). Fas/ APO-1(CD95)-induced apoptosis of primary hepatocytes is inhibited by cAMP. Biochem. Biophys. Res. Commun. 232, 20–25. Fletcher, D. S., Agarwal, L., Chapman, K. T., Chin, J., Egger, L. A., Limjuco, G., Luell, S., MacIntyre, D. E., Peterson, E. P., and Thornberry, N. A. (1995). A synthetic inhibitor of interleuking-1 beta converting enzyme prevents endotoxin-induced interleukin-1 beta production in vitro and in vivo. J. Interferon Cytokine Res. 15, 243–248. French, L., and Tschopp, J. (1997). Thyroiditis and hepatitis: Fas on the road to disease. Nature Med. 3, 387–388. Friedlander, R. M., Gagliardini, V., Hara, H., Fink, K. B., Li, W., MacDonald, G., Fishman, M. C., Greenberg, A. H., Moskowitz, M. A., and Yuan, J. (1997). Expression of a dominant negative mutant of interleukin-1 beta converting enzyme in transgenic mice prevents neuronal cell death induced by trophic factor withdrawal and ischemic brain injury. J. Exp. Med. 185, 933–40. Galle, P. R., Hofmann, W. J., Walczak, H., Schaller, H., Otto, G., Stremmel, W., Krammer, P. H., and Runkel, L. (1995). Involvement of the CD95 (APPO-1/Fas) receptor and ligand in liver damage. J. Exp. Med. 182, 1223–1230. Games, S., Anel, A., Pineiro, A., and Naval, J. (1998). Caspases are the main executioners of Fas-mediated apoptosis, irrespective of the ceramid signalling pathway. Cell Death Diff. 5, 241–249. Genestier, L., Prigent, A.-F., Paillot, R., Quemeneur, L., Durand, I., Banchereau, J., Revillard, J., and Bonnefoy-Berard, N. (1998). Caspase-dependent ceramide production in Fas- and HLA class I-mediated peripheral T cell apoptosis. J. Biol. Chem. 273, 5060–5066. Ghayur, T., Banerjee, S., Hugunin, M., Butler, D., Herzog, L., Carter, A., Quintal, L., Sekut, L., Talanian, R., Paskind, M., Wong, W., Kamen, R., Tracey, D., and Allen, H. (1997). Caspase-1 processes IFN-웂-inducing factor and regulates LPS-induced IFN-웂 production. Nature (London) 386, 619–622. Giordano, C., Stassi, G., De Maria, R., Todaro, M., Richiusa, P., Papoff, G., Ruberti, G., Bagnasco, M., Testi, R., and Galluzzo, A. (1997). Potential involvement of Fas and its ligand in the pathogenesis of Hashimoto’s thyroiditis. Science 275, 960–963. Glynn, J. M., McElligott, D. L., and Mosier, D. E. (1996). Apoptosis-induced by HIV infection in H9 T cells is blocked by ICE-family protease, inhibition but not by a Fas (CD95) antagonist. J. Immunol. 157, 2754–2758. Golstein, P. (1997). Controlling cell death. Science 275, 1081–1082. Gougeon, M. L., and Montagnier, L. (1993). Apoptosis in AIDS. Science 260, 1269–1270. Gougeon, M. L., Garcia, S., Heeney, J., Tschopp, R., Lecoeur, H., Guetard, D., Rame, V., Dauguet, C., and Montagnier, L. (1993). Programmed cell death in AIDS-related HIV and SIV infections. AIDS Res. Hum. Retroviruses 9, 553–563. Gougeon, M.-L., Lecoeur, H., Dulioust, A., Enouf, M.-G., Crouvoisier, M., Goujard, C., Debord, T., and Montagnier, L. (1996). Programmed cell death in peripheral lymphocytes from HIV-infected persons. J. Immunol. 156, 3509–3520. Gu, Y., Kuida, K., Tsutsui, H., Ku, G., Hsiao, K., Fleming, M. A., Hayashi, N., Higashino, K., Okamura, H., Nakanishi, K., Kurimoto, M., Tanimoto, T., Flavell, R. A., Sato, V.,

292

JOHN C. REED

Harding, M. W., Livingston, D. J., and Su, M. S.-S. (1997). Activation of interferon-웂 inducing factor mediated by interleukin-1웁 converting enzyme. Science 275, 206–209. Hannun, Y. A. (1996). Functions of ceramide in coordinating cellular responses to stress. Science 274, 1855–1859. Hara, H., Fink, K., Endres, M., Friedlander, R. M., Gagliardini, V., Yuan, J., and Moskowitz, M. A. (1997a). Attenuation of transient focal cerebral ischemic injury in transgenic mice expressing a mutant ICE inhibitory protein. J. Cereb. Blood Flow Metab. 17, 370–5. Hara, H., Friedlander, R. M., Gagliardini, V., Ayata, C., Fink, K., Huang, Z., ShimizuSasamata, M., Yuan, J., and Moskowitz, M. A. (1997b). Inhibition of interleukin 1beta converting enzyme family proteases reduces ischemic and excitotoxic neuronal damage. Proc. Natl. Acad. Sci. U.S.A. 94, 2007–2012. Hattori, K., Hirano, T., Miyajima, H., Yamakawa, N., Tateno, M., Oshimi, K., Kayagaki, N., Yagita, H., and Okumura, K. (1998). Differential effects of anti-Fas ligand and antitumor necorosis factor 움 antibodies on acute graft-versus-host disease pathologies. Blood 91, 4051-4055. Hsu, H., Xiong, J., and Goeddel, D. V. (1995). The TNF receptor 1-associated protein TRADD signals cell death and NF-␬ B activation Cell 81, 495–504. Hug, H., Strand, S., Grambihler, A., Galle, J., Hack, V., Stremmel, W., Krammer, P. H., and Galle, P. R. (1997). Reactive oxygen intermediates are involved in the induction of CD95 ligand mRNA expression by cytostatic drugs in hepatoma cells: J. Biol. Chem. 272, 28191–28193. Inohara, N., del Peso, L., Koseki, T., Chen, S., and Nunez, G. (1998). RICK, a novel protein kinase containing a caspase recruitment domain, interacts with CLARP and regulates CD95-mediated apoptosis. J. Biol. Chem. 273, 12296–12300. Irmler, M., Thome, M., Hahne, M., Schneider, P., Hofmann, K., Steiner, V., Bodmer, J.-L., Schro¨ter, M., Burns, K., Mattmann, C., Rimoldi, D., French, L. E., and Tschopp, J. (1997). Inhibition of death receptor signals by cellular FLIP. Nature (London) 388, 190–195. Itoh, N., Imagawa, A., Hanafusa, T., Waguri, M., Yamamoto, K., Iwahashi, H., Moriwaki, M., Nakajima, H., Miyagawa, J., Namba, M., Makino, S., Nagata, S., Kono, N., and Matsuzawa, Y. (1997). Requirement of Fas for the development of autoimmune diabetes in nonobese diabetic mice. J. Exp. Med. 186, 613–618. Jacobs, C., Backer, P., Roux, E., Picna, K., Toivola, B., Waugh, S., and Kennedy, M. (1991). Experimental autoimmune encaphalomyelitis is exacerbated by IL-1 alpha and suppressed by soluable IL-1 receptor. J. Immunol. 146, 2983–2989. Jacobs, C. A., Beckmann, M. P., Mohler, K., Maliszewski, C. R., Fanslow, W. C., and Lynch, D. H. (1993). Pharmacokinetic parameters and biodistribution of soluble cytokine receptor. Int. Rev. Exp. Pathol. 34, (Pt. B), 123–135. Jaeschke, H., Fisher, M. A., Lawson, J. A., Simmons, C. A., Farhood, A., and Jones, D. A. (1998). Activation of caspase 3 (CPP32)-like proteases is essential for TNF-alpha-induced hepatic parenchymal cell apoptosis and neutrophil-mediated necrosis in a murine endotoxin shock model. J. Immunol. 160, 3480–3486. Katsikis, P. D., Wunderlich, E. S., Smith, C. A., and Herzenberg, L. A. (1995). Fas antigen stimulation induces marked apoptosis of T-lymphocytes in human immunodeﬁciency virus-infected individuals. J. Exp. Med. 181, 2029–2036. Katsikis, P. D., Garcia-Ojeda, M. E., Torres-Roca, J. F., Tijoe, I. M., Smith, C. A., Herzenberg, L. A., and Herzenberg, L. A. (1997). Interleukin-1 beta converting enzyme-like protease involvement in Fas-induced and activation induced peripheral blood T cell apoptosis in HIV infection. TNF-related apoptosis-inducing ligand can mediate activationinduced T cell death in HIV infection. J. Exp. Med. 186, 1365–1372.

CASPASES AND CYTOKINES

293

Kimble, R., Vannice, J., Bloedow, D., Thompson, R., Hopper, W., Kung, V., Brownﬁeld, C., and Paciﬁci, R. (1994). Interleukin-1 receptor antagonist decreases bone loss and bone resorption in ovariectomized rats. J. Clin. Invest. 93, 1959–1967. Kiriakidou, M., Driscoll, D. A., Guisa, J. M. L., and Strauss III, J. F. (1997). Cloning and expression of primate Daxx cDNAs and mapping of the human gene to chromosome 6p21.3 in the MHC region. DNA Cell Biol 16, 1289–1298. Kitson, J., Raven, T., Jiang, Y.-P., Goeddel, D. V., Giles, K. M., Pun, K.-T., Grinham, C. J., Brown, R., and Farrow, S. N. (1996). A death-domain containing receptor that mediates apoptosis. Nature (London) 384, 372–375. Ku, G., Faust, T., Lauffer, L. L., Livingston, D. J., and Harding, M. W. (1996). Interleukin1 beta converting enzyme inhibition blocks progression of type II collagen-induced arthritis in mice. Cytokine 8, 377–386. Kuida, K., Lippke, J., Ku, G., Harding, M., Livingston, D., Su, M., and Flavell, R. (1995). Altered cytokine export and apoptosis in mice deﬁcient in interleukin-1웁 converting enzyme, Science 267, 2000–2002. Lagresle, C., Mondie`re, P., Bella, C., Krammer, P. H., and Defrance, T. (1996). Concurrent engagement of CD40 and the antigen receptor protects naive and memory human B cells from APO-1/Fas-mediated apoptosis. J. Exp. Med. 183, 1377–1388. Lebsack, M., Paul, C., Blocdow, D., Burch, F., Sack, M., Chase, W., and Catalano, M. (1991). Subcutaneous II-1 receptor antagonist in patients with rheumatoid arthritis. Arthritis Rheum. 36, S45. Lebsack, M., Paul, C., Martindale, J., and Catalano, M. (1993). A dose and regimen ranging study of IL-1 receptor antagonist in patients with rheumatoid arthritis. Arthritis Rheum. 36, S39. Leithauser, F., Dhein, J., and Mechtercheimer, G. (1993). Constitutive and induced expression of APO-1, a new member of the nerve growth factor/tumor necrosis factor receptor superfamily, in normal and neoplastic cells. Lab. Invest. 69, 415–429. Lewis, D. E., Ng Tang, D. S., Adu-Oppong, A., Schober, W., and Rodgers, J. R. (1994). Anergy and apoptosis in CD8⫹ T cells from HIV-infected persons. J. Immunol. 153, 412–420. Lewthwaite, J., Blake, S., Hardingham, T., Warden, P., and Henderson, B. (1994). The effect of recombinant human interleukin-1 receptor antagonist on the induction phase of antigen induced arthritis in the rabbit. J. Rheumatol. 21. Li, C. J., Friedman, D. J., Wang, C., Metelev, V., and Pardee, A. B. (1995a). Induction of apoptosis in uninfected lymphocytes by HIV-1 Tat protein. Science 268, 429–431. 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. (1995b). Mice deﬁcient in IL-1웁-converting enzyme are defective in production of mature IL-1웁 and resistant to endotoxic shock. Cell 80, 401–411. Li, P., Allen, H., Banerjee, S., and Seshadri, T. (1997a). Characterization of mice deﬁcient in interleukin-1 beta converting enzyme. J. Cell. Biochem. 64, 27–32. Li, P., Nijhawan, D., Budihardjo, I., Srinivasula, S., Ahmad, M., Alnemri, E., and Wang, X. (1997b). Cytochrome c and dATP-dependent formation of Apaf-1/caspase-9 complex initiates an apoptotic protease cascade. Cell 91, 479–489. Lin, T., Brunner, T., Tietz, B., Madsen, J., Bonfoco, E., Reaves, M., Huﬂeit, M., and Green, D. R. (1998). Fas ligand-mediated killing by intestinal intraepithelial lymphocytes. Participation in intestinal graft-versus-host disease. J. Clin. Invest. 101, 570–577. Liu, Z.-G., Hsu, H., Goeddel, D. V., and Karin, M. (1996). Dissection of TNF receptor 1 effector functions: JNK activation is not linked to apoptosis while NF-␬B activation prevents cell death. Cell 87, 565–576.

294

JOHN C. REED

MacFarlane, M., Ahmad, M., Srinivasula, S., Fernandes-Alnemri, T., Cohen, G., and Alnemri, E. (1997). Identiﬁcation and molecular cloning of two novel receptors for the cytotoxic ligand TRAIL. J. Biol. Chem. 272, 25417–25420. Magram, J., Connaughton, S. E., Warrier, R. R., Carvajal, D. M., Wu, C. Y., Ferrante, J., Stewart, C., Sarmiento, U., Faherty, D. A. and Gately, M. K. (1996). IL-12-deﬁcient mice are defective in IFN gamma production and type-1 cytokine responses. Immunity 4, 471–481. Mariani, S. M., and Krammer, P. H. (1998). Surface expression of TRAIL/Apo-2 ligand in activated mouse T and B cells. Eur. J. Immunol. 28, 1492–1498. Mariani, S. M., Matiba, B., Armandola, E. A., and Krammer, P. H. (1997). Interleukin 1 beta-converting enzyme related proteases/caspases are involved in TRAIL-induced apoptosis of myeloma and leukemia cells. J. Cell Biol. 137, 221–229. Marsters, S., Sheridan, J., Pitti, R., Brush, J., Goddard, A., and Ashkenazi, A. (1998). Identiﬁcation of a ligand for the death-domain-containing receptor Apo3. Curr. Biol. 8, 525–528. Martin, R., McFarland, H., and McFarlin, D. (1992). Immunological aspects of demyelinating diseases. Annu. Rev. Immunol. 10, 153–187. McAlindon, M. E., Hawkey, C. J., and Manida, Y. R. (1998). Expression of interleukin 1 beta and interleukin 1 beta converting enzyme by intestinal macrophages in health and inﬂammatory bowel disease. Gut 42, 214–219. McCarthy, P., Abhyankar, S., and Neben, S. (1991). Inhibition of interleukin-1 by interleukin1 receptor antagonist presents graft vs. host disease. Blood 78, 1915–1918. Medema, J. P., Scafﬁdi, C., Kischkel, F. C., Shevdhenko, A., Mann, M., Krammer, P. H., and Peter, M. E. (1997). FLICE is activated by association with the CD95 death-inducing signaling complex (DISC). EMBO J. 16, 2794–2804. Medema, J., Scafﬁdi, C., Krammer, P., and Peter, M. (1998). Bcl-XL acts downstream of FLICE/caspase-8. J. Biol. Chem. 273, 3388–3393. Micallef, M. J., Ohtsuki, T., Kohno, K., Tanabe, F., Ushio, S., Namba, M., Tanimoto, T., Torigoe, K., Fujii, M., Ikeda, M., Fukada, S., and Kurimoto, M. (1996). Interferongamma-inducing factor enhances T helper 1 cytokine production by stimulated human T cells: Synergism with interleukin-12 for interferon-gamma production. Eur. J. Immunol. 26, 1647–1651. Miura, M., Zhu, H., Rotello, R., Hartwieg, E. A., and Yuan, J. (1993). Induction of apoptosis in ﬁbroblasts by IL-1 beta-converting enzyme, a mammalian homolog of the C. elegans cell death gene ced-3 Cell 75, 653–660. Muschen, M., Warskulat, U., Douillard, P., Gilbert, E., and Haussinger, D. (1998). Regulation of CD95 (APO-1/Fas) receptor and ligand expression by lipopoolysaccharide and dexamethasone in parenchymal and nonparenchymal rat liver cells. Hepatology 27, 200–208. Muzio, M., Chinnaiyan, A. M., Kischkel, F. C., O’ Rourke, K., Shevchenko, A., Ni, J., Scafﬁdi, C., Bretz, J. D., Zhang, M., Gentz, R., Mann, M., Krammer, P. H., Peter, M. E., and Dixit, V. M. (1996). Flice, a novel FADD-homologous ICE/CED-3-like protease, is recruited to the CD95 (Fas/APO-1) death-inducing signaling complex. Cell 85, 817–827. Neurath, M. F., Fuss, I., Kelsall, B. L., Stuber, E., and Strober, W. (1995). Antibodies to interleukin 12 abrogate established experimental colitis in mice. J. Exp. Med. 182, 1281– 1290. Nishigaki, K., Minatoguchi, S., Seishima, M., Asano, K., Noda, T., Yasuda, N., Sano, H., Kumada, H., Takemura, M., Noma, A., Tanaka, T., Watanabe, S., and Fujiwara, H. (1997). Plasma Fas ligand, an inducer of apoptosis, and plasma soluable Fas, an inhibitor of

CASPASES AND CYTOKINES

295

apoptosis; in patients with chronic congestive heart failure. J. Am. Coll. Cardiol. 29, 1214– 1220. Nishina, H., Fischer, K. D., Radvanyl, L., Shahinian, A., Hakem, R., Rubie, E. A., Bernstein, A., Mak, T. W., Woodgett, J. R., and Penninger, J. M. (1997). Stress-signalling kinase Sek1 protects thymocytes from apoptosis mediated by CD95 and CD3. Nature (London) 385, 350–353. Ogasawara, J., Watanabe-Fukunaga, R., Adachi, M., Matsuzawa, A., Kasugai, T., Kitamura, Y., Itoh, N., Suda, T., and Nagata, S. (1993). Lethal effect of the anti-Fas antibody in mice. Nature (London) 364, 806–809. Ohlsson, K., Bjork, P., Bergenfeldt, M., Hageman, R., and Thompson, R. (1990). Interleukin1 receptor antagonist reduces mortality from endotoxin shock. Nature (London) 348, 550–552. Okamura, H., Tsutsi, H., Komatsu, T., Yutsudo, M., Hakura, A., Tanimoto, T., Torigoe, K., Okura, T., Nukada Hattori, K., et al. (1995). Cloning of a new cytokine that induces IFNgamma production by T cells. Nature (London) 378, 88–91. Opal, S. M., Fisher, C. J., Jr., Dhainaut, J. F., Vincent, J. L., Brase, R., Lowry, S. F., Sadoff, J. C., Slotman, G. J., Levy, H., Balk, R. A., Shelly, M. P., Pribble, J. P., LaBrecque, J. F., Lookabaugh, J., Donovan, H., Dubin, H., Baughman, R., Norman, J., DeMaria, E., Matzel, K., Abraham, E., and Seneff, M. (1997). Conﬁrmatory interleukin-1 receptor antagonist trial in severe sepsis: A phase III, randomized, double-blind, placebo-controlled, multicenter trial. The Interleukin-1 Receptor Antagonist Sepsis Investigator Group. Crit. Care Med. 25, 1115–1124. Orth, K., O’Rourke, K., Salvesen, G. S., and Dixit, V. M. (1996). Molecular ordering of apoptotic mammalian CED-3/ICE-like proteases. J. Biol. Chem. 271, 20977–20980. Pan, G., Ni, J., Wei, Y. F., Yu, G., Gentz, R., and Dixit, V. M. (1997a). An antagonist decoy receptor and a death domain-containing receptor for TRAIL. Science 277, 815–818. Pan, G., O’Rourke, K., Chinnaiyan, A. M., Gentz, R., Ebner, R., Ni, J., and Dixit, V. M. (1997b). The receptor of the cytotoxic ligand TRAIL. Science 276, 111–113. Peter, M. E., Kischkel, F. C., Scheuerpﬂug, C. G., Medema, J. P., Debatin, K. M., and Krammer, P. H. (1997). Resistance of cultured peripheral T cells towards activationinduced cell death involves a lack of recruitment of FLICE (MACH/caspase 8) to the CD95 death-inducing signaling complex. Eur. J. Immunol. 27, 1207–1212. Rano, T. A., Timkey, T., Peterson, E. P., Rotonda, J., Nicholson, D. W., Becker, J. W., Chapman, K. T., and Thornberry, N. A. (1997). A combinatorial approach for determining protease speciﬁcities: Application to interleukin-1beta converting enzyme (ICE). Chem. Biol. 4, 149–155. Relton, J., and Rothwell, N. (1992). Interleukin-1 receptor antagonist inhibits ischaemic and excitotoxic neuronal damage in the rat. Brain Res. Bull. 29, 243–246. Rothstein, T. L., Wang, J. K. M., Panka, D. J., Foote, L. C., Wang, Z., Stanger, B., Cui, H., Ju, S.-T., and Marshak-Rothstein, A. (1995). Protection against fas-dependent Th1mediated apoptosis by antigen receptor engagement in B cells. Nature (London) 374, 163–165. Rotonda, J., Nicholson, D. W., Fazil, K. M., Gallant, M., Gareau, Y., Labelle, M., Peterson, E. P., Rasper, D. M., Ruel, R., Vaillancourt, J. P., Thornberry, N. A., and Becker, J. W. (1996). The three-dimensional structure of apopain/CPP32, a key mediator of apoptosis. Nature Struct. Biol. 3, 619–625. Sabelko, K., Kelly, K., Nahm, M., Cross, A., and Russell, J. (1997). Fas and Fas ligand enhance the pathogenesis of experimental allergi encephalomyelitis, but are not essential for immune privilege in the central nervous system. J. Immunol. 159, 3096–3099.

296

JOHN C. REED

Sakai, T., Kimura, Y., Inagaki-Ohara, K., Kusugami, K., Lynch, D. H., and Yoshikai, Y. (1997). Fas-mediated cytotoxicity by intestinal intraepithelial lymphocytes during acute graft-versus-host disease in mice. Gastroenterology 113, 168–174. Salvesen, G. S., and Dixit, V. M. (1997). Caspases: Intracellular signaling by proteolysis. Cell 91, 443–446. Sato, T., Irie, S., Kitada, S., and Reed, J. C. (1995). FAP-1: A protein tyrosine phosphatase that associates with Fas. Science 268, 411–415. Schattner, E. J., Elkon, K. B., Yoo, D., Tumang, J., Krammer, P. H., Crow, M. K., and Friedman, S. M. (1995). CD40 ligation induces Apo-1/Fas expression on human B lymphocytes and facilitates apoptosis through the Apo-1/Fas pathway. J. Exp. Med. 182, 1557– 1565. Scho¨nbeck, U., Mach, F., Bonnefoy, J.-Y., Loppnow, H., Flad, H.-D., and Libby, P. (1997). Ligation of CD40 activates interleukin 1웁-converting enzyme (caspase-1) activity in vascular smooth muscle and endothelial cells and promotes elaboration of active interleukin 1웁. J. Biol. Chem. 272, 19569–19574. Schumann, R. R., Belka, C., Reuter, D., Lamping, N., Kirschning, C. J., Weber, J. R., and Pfeil, D. (1998). Lipopolysaccharide activates caspase-1 (interleukin-1-converting enzyme) in cultured monocytic and endothelial cells. Blood 91, 577–584. Schwab, J., Anderle, S., Brown, R., Dalldorf, F., and Thompson, R. (1991). Pro- and antiinﬂammatory roles of interleukin-1 in recurrence of bacterial cell wall-induced arthritis in rats. Infect. Immunol. 59, 4436–4442. Seino, K., Kayagaki, N., Bashuda, H., Okumura, K., and Yagita, H. (1996). Contribution of Fas ligand to cardiac allograft rejection. Int. Immunol. 8, 1347–1354. Sheridan, J. P., Marsters, S. A., Pitti, R. M., Gurney, A., Skubatch, M., Baldwin, D., Ramakrishnan, L., Gray, C. L., Baker, K., Wood, W. I., Goddard, A. D., Godowski, P., and Ashkenazi, A. (1997). Control of TRAIL-induced apoptosis by a family of signaling and decoy receptors [see comments]. Science 277, 818–21. Shi, L., Chen, G., MacDonald, G., Bergeron, L., Li, H., Miura, M., Rotello, R. J., Miller, D. K., Li, P., Seshadri, T., Yuan, J., and Greenberg, A. (1996). Activation of an interleukin 1 converting enzyme-dependent apoptosis pathway by granzyme B. Proc. Natl. Acad. Sci. U.S.A. 93, 11002–11007. Sillence, D. J., and Allan, D. (1997). Evidence against an early signalling role for ceramide in Fas-mediated apoptosis. Biochem. J. 324, 29–32. Smith, K. G. C., Strasser, A., and Vaux, D. L. (1996). CrmA expression in T lymphocytes of transgenic mice inhibits CD95 (FAS/APO-1)-transduced apoptosis, but does not cause lymphadenopathy or autoimmune disease. EMBO J. 15, 5167–5176. Smyth, M. J., Obeid, L. M., and Hannun, Y. A. (1997). Ceramide: A novel lipid mediator of apoptosis. In ‘‘Apoptosis: Pharmacological Implications and Therapeutic Opportunities’’ (S. H. Kaufmann, ed.), pp. 133–154. Academic Press, San Diego. Sneller, M. C., Wang, J., Dale, J. K., Strober, W., Middleton, L. A., Choi, Y., Fleisher, T. A., Lim, M. S., Jaffe, E. S., Puck, J. M., Lenardo, M. J., and Straus, S. E. (1997). Clinical, immunologic, and genetic features of an autoimmune lymphoproliferative syndrome associated with abnormal lymphocyte apoptosis. Blood 89, 1341–1348. Srinivasan, A. F. L., Wong, A., Kodandapani, L., Smidt, R., Krebs, J., Fritz, L., Wu, J., and Tomaselli, K. (1998). Bcl-xL functions downstream of caspase-8 to inhibit Fas- and tumor necrosis factor receptor 1-induced apoptosis of MCF7 breast carcinoma cells. J. Biol. Chem. 273, 4523–4529. Srinivasula, S., Ahmad, M., Fernandes-Alnemri, T., Litwack, G., and Alnemri, E. S. (1996). Molecular ordering of the fas-apoptotic pathway: The fas/APO-1 protease Mch5 is a

CASPASES AND CYTOKINES

297

CrmA-inhibitableprotease that activates multiple Ced-3/ICE-like cysteine proteases. Proc. Natl. Acad. Sci. U.S.A. 93, 14486–14491. Srinivasula, S. M., Ahmad, M., Ottilie, S., Bullrich, F., Banks, S., Wang, Y., FernandesAlnemri, T., Croce, C. M., Litwack, G., Tomaselli, K. J., Armstrong, R. C., and Alnemri, E. S. (1997). FLAME-1, a novel FADD-like antiapoptotic molecule that regulates Fas/ TNFR1-induced apoptosis. J. Biol. Chem. 272, 18542–18545. Stassi, G., Maria, R., Trucco, G., Rudert, W., Testi, R., Galluzzo, A., Giordano, C., and Trucco, M. (1997). Nitric oxide primes pancreatic beta cells for Fas-mediated destruction in insulin-dependent diabetes mellitus. J. Exp. Med. 186, 1193–2000. Steinman, L. (1996). Multiple sclerosis: A coordinated immunological attack against myolin in the central nervous system. Cell, 85, 299–302. Stennicke, H. R., and Salvesen, G. S. (1997). Biochemical characteristics of caspase-3, -6, -7, and -8. J. Biol. Chem. 272, 25719–25723. Strand, S., Hofmann, W. J., Grambihler, A., Hug, H., Volkmann, M., Otto, G., Wesch, H., Mariani, S. M., Hack, V., Stremmel, W., Krammer, P. H., and Galle, P. R. (1998). Hepatic failure and liver cell damage in acute Wilson’s disease involve CD95 (APO-1/Fas)-mediated apoptosis. Nature Med. 4, 588–593. Strater, J., Wellisch, I., Riedl, S., Walczak, H., Koretz, K., Tandara, A., Krammer, P. H., and Moller, P. (1997). CD95 (APO-1/Fas)-mediated apoptosis in colon-epithelial cells: A possible role in ulcerative colitis. Gastroenterology 113, 160–167. Stricker, K., Knipping, E., Bohler, T., Benner, A., Krammer, P., and Debatin, K. (1998). Anti-CD95 (APO-1/Fas) autoantibodies and T cell depletion in human immunodeﬁciency virus type 1 (HIV-1)-infected children. Cell Death Differ. 5, 222–230. Takahashi, T., Tanaka, M., Brannan, C. I., Jenkins, N. A., Copeland, N. G., Suda, T., and Nagata, S. (1994). Generalized lymphoproliferative disease in mice, caused by a point mutation in the Fas ligand. Cell 76, 969–976. Talley, A. K., Dewhurst, S., Perry, S. W., Dollard, S. C., Gummuluru, S., Fine, S. M., New, D., Epstein, L. G., Gendelman, H. E., and Gelbard, H. A. (1995). Tumor necrosis factor alpha-induced apoptosis in human neuronal cells: Protection by the antioxidant N-acetylcysteine and the genes bcl-2 and crmA. Mol. Cell. Biol. 15, 2359–2366. Tartaglia, L. A., and Goeddel, D. V. (1992). Two TNF receptors. Immunol. Today 13, 151–153. Tewari, M., Beidler, D. R., and Dixit, V. M. (1995a). CrmA-inhibitable cleavage of the 70kDa protein component of the U1 small nuclear ribonucleoprotein during Fas- and tumor necrosis factor-induced apoptosis. J. Biol. Chem. 270, 18738–18741. Tewari, M., Telford, W. G., Miller, R. A., and Dixit, V. M. (1995b). CrmA, a poxvirus-encoded serpin, inhibits cytotoxic T-lymphocyte-mediated apoptosis. J. Biol. Chem. 270, 1–4. Thome, M., Schneider, P., Hofmann, K., Fickenscher, H., Meinl, E., Neipel, F., Mattmann, C., Burns, K., Bodmer, J.-L., Schroter, M., Scafﬁdi, C., Krammer, P. H., Peter, M. E., and Tschopp, J. (1997). Viral FLICE-inhibitory proteins (FLIPs) prevent apoptosis induced by death receptors. Nature (London) 386, 517–521. Thompson, R., Dripps, D., and Eisenberg, S. (1992). Interleukin-1 receptor anatagonist as a probe and as a treatment for IL-1 mediated disease. Int. J. Immunopharmacol. 14, 475–480. Thornberry, N., Rano, T., Peterson, E., Rasper, D., Timkey, T., Garcia-Calvo, M., Houtzager, V., Nordstrom, P., Roy, S., Vaillancourt, J., Chapman, K., and Nicholson, D. (1997). A combinatorial approach deﬁnes speciﬁcities of members of the caspase family and granzyme B. J. Biol. Chem. 272, 17907–17911. Tisch, R., and McDevitt, H. (1996). Insulin-dependent diabetes mellitus. Cell 85, 291–297.

298

JOHN C. REED

Torigoe, T., Millan, J. A., Takayama, S., Taichman, R., Miyashita, T., and Reed, J. C. (1994). Bcl-2 inhibits T-cell-mediated cytolysis of leukemia cell line. Cancer Res. 54, 4851–4854. Toyoshima, F., Moriguchi, T., and Nishida, E. (1997). Fas induces cytoplasmic apoptotic responses and activation of the MKK7-JNK/SAPK and MKK6-p38 pathways independent of CPP32-like proteases. J. Cell. Biol. 139, 1005–1015. Toyozaki, T., Hiroe, M., Saito, T., Ijima, Y., Takano, H., Hiroshima, K., Kohno, H., Ishiyama, S., Marumo, F., Masuda, Y., and Ohwada, H. (1998). Levels of soluble Fas in patients with myocarditis, heart failure of unknown origin and in healthy volunteers. Am. J. Cardiol. 81, 798–800. Umansky, S. R., and Tomei, L. D. (1997). Apoptosis in the heart. In ‘‘Apoptosis: Pharmacological Implications and Therapeutic Opportunities’’ (S. H. Kaufmann, ed.), pp. 383–407. Academic Press San Diego. Ushio, S., Namba, M., Okura, T., Hattori, K., Nukada, Y., Akita, K., Tanabe, F., Konishi, K., et al., (1996). Cloning of the cDNA for human IFN-gamma inducing factor, expression in Escherichia coli, and studies on the biologic activities of the protein. J. Immunol. 156, 4274–4279. Van Den Berg, W., Joosten, L., Helsen, M., and Van De Loo, F. (1994). Amelioration of established murine collagen-induced arthritis with anti-IL-1 treatment. Clin. Exp. Immunol. 95, 237–243. Van Parjis, L., and Abbas, A. K. (1996). Role of Fas-mediated cell death in the regulation of immune responses. Curr. Opin. Immunol. 8, 355–361. Van Parjis, L., and Abbas, A. (1998). Homeostasis and self tolerance in the immune system: Turning lymphocytes off. Science 280, 243–248. Varfolomeev, E. E., Schuchmann, M., Luria, V., Chiannilkulchai, N., Beckmann, J. S., Mett, I. L., Rebrikov, D., Brodianski, V. M., Kemper, O. C., Kollet, O., Lapidot, T., Soffer, D., Sobe, T., Avraham, K. B., Goncharov, T., Holtmann, H., Lonai, P., and Wallach, D. (1998). Targeted disruption of the mouse capsase 8 gene ablates cell death inducation by the TNF receptors, Fas/Apo 1, and DR3 and is lethal prenatlally Immunity 9, 267–276. Vercammen, D., Beyaert, R., Denecker, G., Goossens, V., Van Loo, G., Declercq, W., Grooten, J., Fiers, W., and Vandenabeele, P. (1998). Inhibition of caspases increases the sensitivity of L929 cells to necrosis mediated by tumor necrosis factor. J. Exp. Med. 187, 1477–1485. Verheij, M., Bose, R., Lin, X. H., Yao, B., Jarvis, W. D., Grant, S., Birrer, M. J., Szabo, E., Zon, L. I., Kyriakis, J. M., Haimovitz-Friedman, A., Fuks, Z., and Kolesnick, R. N. (1996). Requirement for ceramide-initiated SAPK/JNK signalling in stress-induced apoptosis. Nature (London) 380, 75–79. Via, C. S., Nguyen, P., Shustov, A., Drappa, J., and Elkon, K. B. (1996). A major role for the Fas pathway in acute graft-versus-host disease. J. Immunol. 157, 5387–5393. Vincenz, C., and Dixit, V. M (1997). Fas-associated death domain protein interleukin-1betaconverting enzyme 2 (FLICE2), an ICE/Ced-3 homologue, is proximally involved in CD95- and p55-mediate death-signaling. J. Biol. Chem. 272, 6578–6583. Wakabayashi, G., Gelfand, J., Burke, J., Thompson, R., and Dinarello, C. (1997). A speciﬁc receptor antagonist for interleukin-2 prevents Escherichia coli-induced shock. FASEB J. 5, 338–343. Waldner, H., Sobel, R., Howard, E., and Kuchroo, V. (1997). Fas- and FasL-deﬁcient mice are resistant to induction of autoimmune encephalomyelitis. J. Immunol. 159, 3100–3103. Wallach, D., Boldin, M., Varfolomeev, E., Beyaert, R., Vandenabeele, P., and Fiers, W. (1997). Cell death induction by receptors of he TNF family: Towards a molecular understanding. FEBS Lett. 410, 96–106.

CASPASES AND CYTOKINES

299

Wang, C.-Y., Mayo, M. W., and Baldwin, A. S. (1996a). TNF- and cancer therapy-induced apoptosis: Potentation by inhibition of NF-␬B. Science 274, 784–787. Wang, S., Miura, M., Zhu, H., Gagliardini, V., and Yuan, J. (1996b). Ideniﬁcation and characterization of caspase-11, a member of the ICE/Ced-3 family and an upstream regulator of ICE. J. Biol. Chem. 271, 20580–20587. Wang, S., Miura, M., Jung, Y. K., Zhu, H., Li, E., and Yuan, J. (1998). Murine caspase-11, an ICE-interacting protease, is essential for the activation of ICE. Cell 92, 501–509. Watanabe-Fukunaga, R., Brannan, C. I., Copeland, N. G., Jenkins, N. A., and Nagata, S. (1992). Lymphoproliferation disorder in mice explained by defects in Fas antigen that mediates apoptosis. Nature (London) 356, 314–317. Westendorp, M. O., Frank, R., Ochsenbauer, C., Stricker, K., Dhein, J., Waleczak, H., Debatin, K.-M., and Krammer, P. H. (1995). Sensitization of T cells to CD 95 mediated apoptosis by HIV-1 Tat and gp120. Nature (London) 375, 497–500. Wilson, K. P., Black, J. F., 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 Livingston, D. J. (1994). Structure and mechanism and interleukin-1웁converting enzyme. Nature (London) 370, 270–275. Wooley, P., Whalen, J., Chapman, D., Berger, A., Richard, K., Aspar, D., and Staite, N. (1993). The effect of an interleukin-1 receptor antagonist proteins on type II collageninduced arthritis and antigen-induced in mice. Arthritis Rheum. 36, 1305–1314. Yang, X., Khosravi-Far, R., Chang, H. Y., and Baltimore, D. (1997). Daxx a novel Fasbinding protein that activates JNK and apoptosis. Cell 89, 1067–1076. Yao, J., and Johnson, R. W. (1997). Induction of interleukin-1 beta-converting enzyme (ICE) in murine microglia by lipopolysaccharide. Mol. Brain Res. 51, 170–178. Yaoita, H., Ogawa, K., Maehara, K., and Maruyama, Y. (1997). ‘‘Attenuation of Ischemia/ Reperfusion Injury in Rats by a caspase Inhibitor.’’ Basic Science Reports, pp. 276–281. Yeh, W., De La Pompa, J., El-Deiry, W., Lowe, S., Goeddel, D., and Mak, T. (1998). FADD: Essential for embryo development and signaling from some, but not all, inducers of apoptosis. Science 279, 1954–1958. 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. Zhang, J., Cado, C., Chen, A., Kabra, N., and Winota, A. (1998a). Fas-mediated apoptosis and activation-induced T cell proliferation are defective in mice lacing FADD/Mort Nature (London) 392, 296–299. Zhang, Y., Center, D. M., Wu, M. H., Cruikshank, W. W., Yuan, J., Andrews, D. W., and Kornfeld, H. (1998b). Processsing and activation of pro-interleukin-16 by caspase-3. J. Biol. Chem. 273, 1144–1149. Zhou, Q., Snipas, S., Orth, K., Muzio, M., Dixit, V. M., and Salvesen, G. S. (1997). Target protease speciﬁcity of the viral serpin CrmA: Analysis of ﬁve caspases. J. Biol. Chem. 272, 7797–7800.

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 73

T Cell Dynamics in HIV-1 Infection DAWN R. CLARK,* ROB J. DE BOER,† KATJA C. WOLTHERS,* AND FRANK MIEDEMA*,‡ *CLB, Sanquin Blood Supply Foundation, Department of Clinical Viro-Immunology, Laboratory for Experimental and Clinical Immunology, Academic Medical Center, University of Amsterdam, Amsterdam, The Netherlands; †Department of Theoretical Biology, University of Utrecht, Utrecht, The Netherlands; and ‡Department of Human Retrovirology, Academic Medical Center, Amsterdam, The Netherlands.

I. Introduction

One of the most prominent features of HIV-1 infection is CD4⫹ T cell depletion. This statement is widely used in papers on HIV-1 research; however, while true, it is deceptively simplistic in that it fails to describe what is actually a complex change in the representation of T cell subsets during HIV-1 infection. Figure 1, adapted from Roederer (1995), shows this complex pattern of T cell subset composition with disease progression. Overall, CD4⫹ T cells decline in number while the CD8⫹ T cell population increases over time. The increase in the CD8⫹ T cell pool is the result of massive peripheral expansion of memory cells. This subset only begins to decline shortly preceding AIDS diagnosis (Margolick et al., 1995). The CD4⫹ memory compartment also initially expands due to peripheral expansion, but memory cells are then progressively lost. Interestingly, in both CD4⫹ and the CD8⫹ subsets, the naive compartment begins to decline soon after infection (Rabin et al., 1995; Roederer et al., 1995). Thus, the T cell depletion observed in HIV-1 infection consists of naive cells of both CD4⫹ and CD8⫹ subsets, and memory cells of the CD4⫹ subset. The majority of research has focused on the depletion of CD4⫹ T cells and the bulk of the discussion that follows will concentrate on this subset. Although the observation of T cell depletion in HIV-1 infection was made early, the mechanism for this decline is still not properly understood. Over the past 10 years of AIDS research, investigators have discussed several possible mechanisms for CD4⫹ T cell depletion: virus-related killing, activation-induced apoptosis, and disturbed renewal mechanisms (Ho et al., 1989; Ameisen and Capron, 1991; Meyaard et al., 1992; Groux et al., 1992; Gougeon and Montagnier, 1993; Bonyhadi et al., 1993; Schnittman et al., 1990; Kaneshima et al., 1994; McLean and Michie, 1993). A major breakthrough in our understanding of HIV-1 infection came in 1994 with the introduction of protease inhibitors, a new generation of antiretroviral drugs that effectively blocks the replication of the virus in previously infected cells. Combination of these new drugs with other antiretroviral drugs made it possible to treat HIV-1-infected individuals more successfully 301

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

302

DAWN R. CLARK et al.

FIG. 1. Representation of T cell subset composition of blood during progression to AIDS. Adapted from Roederer (1995), by courtesy of the author and permission of Nature Medicine.

than before (Danner et al., 1995; Markowitz et al., 1995). In addition to the tremendous clinical beneﬁt, new possibilities for research were created. For the ﬁrst time, viral replication could be effectively blocked for a prolonged time, and the consequences for immune recovery could be studied. The advent of potent antiretroviral therapy gave renewed impetus to the debate over viral and T cell turnover. The application of mathematical models to viral load reduction and CD4⫹ T cell increase after treatment, as was done by Ho et al. (1995), Perelson et al. (1996), and Wei et al. (1995) has changed our view on HIV pathogenesis dramatically. By calculating that 1010 virions were produced and destroyed per day (Wei et al., 1995; Ho et al., 1995; Perelson et al., 1996), these investigators showed that HIV-1 infection is a highly dynamic process with much more virus ‘‘turn-

T CELL DYNAMICS IN HIV-1 INFECTION

303

over’’ than was previously anticipated. From the same data, rapid turnover of CD4⫹ T cells in HIV-1 was proposed as the mechanism leading to CD4⫹ T cell depletion because the immune system would not be able to keep up with high rates of renewal inﬁnitely. This view of rapid CD4⫹ T cell turnover was challenged by studies on T cell telomere length, showing no evidence for increased CD4⫹ T cell turnover or exhaustion (Wolthers et al., 1996). These data provided the focal point for renewed debate over this issue (Mosier et al., 1995; Grossman and Herberman, 1997; Hellerstein and McCune, 1997). Since then, several groups have tried to further elucidate the magnitude of T cell turnover in HIV-1 infection. The issue of turnover involves cell death, proliferation of existing cells, and development of new cells from progenitors. Here, we give an overview of what is known about normal T cell dynamics, T cell dynamics in HIV-1 infection, how the different studies relate to each other, and what insight they provide to explain T cell depletion in HIV-1 infection. II. Normal T Cell Renewal from Progenitors

The process of T cell renewal in the maintenance of the T cell population can involve two separate mechanisms: proliferation of mature cells in the periphery and development of new T cells from a progenitor source. The relative contribution of each of these mechanisms to overall T cell renewal is dependent on the age of the individual and the proﬁle of the remaining T cell pool. The common view is that early T cell development requires a thymus and involves a high degree of development from progenitor sources, but that as the organism ages the T cell population is maintained primarily through peripheral expansion of dividing mature cells (Mackall and Gress, 1997b; Rocha et al., 1989; Tough and Sprent, 1994; Sprent and Tough, 1995). Here we summarize what is known about the mechanism of regeneration of T cells in a depleted individual, the effect of HIV-1 infection on T cell renewal, and the impact of therapy on regeneration of T cells. A. T CELL RENEWAL IN MICE Our understanding of the importance of thymus-dependent and thymusindependent mechanisms for T cell regeneration comes initially from murine studies. Several studies compare the T cell regeneration from thymus-bearing and thymectomized animals, which were irradiated and repopulated with bone marrow. In the ﬁrst studies, adult animals were given syngeneic T cell-depleted bone marrow as a source of progenitor cells and congenic lymph node cells as a source of mature T cells (Mackall and Gress, 1997b). In euthymic animals, T cell renewal occurred via

304

DAWN R. CLARK et al.

thymus-dependent development from progenitors with little expansion of the lymph node cells. In thymectomized animals, very little development from progenitors was detected while substantial expansion of mature lymph node cells occurred. Repopulation from T cell-depleted and undepleted bone marrow was compared (Dulude et al., 1997). In the case where T cells were not present in the inoculum, regeneration occurred through extrathymic development of progenitor cells. However, when mature T cells were present, regeneration occurred primarily through expansion of these cells. It is important to note that the thymus-independent development was not able to restore normal T cell numbers (Mackall and Gress, 1997b; Dulude et al., 1997). In addition, T cell renewal through peripheral expansion gave rise to cells bearing a memory phenotype based on expression of CD45RO and CD44 (Mackall and Gress, 1997b; Mackall et al., 1993; Tanchot and Rocha, 1995). B. T CELL RENEWAL IN HUMANS Similar studies were conducted in humans that had been treated with T cell-depleting doses of chemotherapy. In these patients, there was an age-related regeneration of naive CD4⫹ T cells (Mackall and Gress, 1997b; Mackall et al., 1995; Weinberg et al., 1995). Younger individuals showed regeneration of substantial numbers of CD4⫹ CD45RA⫹ T cells whereas older patients still showed CD4⫹ depletion 6 months after therapy, though, after longer periods of time, the naive CD4⫹ T cells did rise in these patients. The degree of regeneration of CD45RA⫹ CD4⫹ T cells was directly related to thymic function (Mackall et al., 1995; Weinberg et al., 1995; Heitger et al., 1997), underscoring the importance of the thymus in the generation of naive CD4⫹ T cells. In adult patients given doses of chemotherapy that only moderately depleted the T cell subset, the majority of regenerating T cells expressed a memory phenotype indicating they were derived by expansion of previously existing T cells (Hakim et al., 1997). These results indicate that, in the case of CD4⫹ T cells, regeneration from a progenitor source depends on thymic function and occurs at low rates in adults. The slow regeneration of naive CD4⫹ cells can also be seen in adult patients undergoing monoclonal antibody (mAb) therapy for rheumatoid arthritis (Moreland et al., 1994) or multiple sclerosis (Rep et al., 1997) and in normal aged individuals (Mackall and Gress, 1997b). Interestingly, requirements for the regeneration of CD8⫹ T cells differ from those for CD4⫹ T cells. In the same experiments described above, by 3 months after chemotherapy most individuals had recovered normal CD8⫹ T cell numbers (Mackall, 1997). In addition, the age-related recovery rate seen for CD4⫹ T cells was not observed for CD8⫹ T cells. In one individual thymectomized before chemotherapy, the CD45RA⫹CD8⫹ T

T CELL DYNAMICS IN HIV-1 INFECTION

305

cells were recovered though the CD45RA⫹CD4⫹ T cells were not (Heitger et al., 1997). The results show that generation of naive CD8⫹ T cells does not require a thymus and is faster than that of CD4⫹ T cells (Mackall, 1997). One possible caveat to this conclusion is that the naive and memory CD8⫹ subsets cannot be distinguished by the CD45 isoforms alone (Hamann et al., 1997; Rabin et al., 1995). The recovery of cells with a surface molecule expression pattern that does deﬁne these subsets in CD8⫹ cells has not been measured. It is still possible that truly naive CD8⫹ T cells do require a thymus for their development as well. The ﬁnal conclusion of these studies is that maintenance of the T cell population in adults involves both mechanisms, but that the primary mechanism is peripheral expansion of previously existing T cells. The age-related decline in thymic function and the slow rate of repopulation in depleted individuals has led to the proposal that T cell regeneration in adults normally occurs near maximum capacity to maintain normal T cell numbers and that this rate cannot be substantially increased (Zhang et al., 1998). In mice depleted of CD4⫹ T cells by mAb treatment and subsequently thymically injected with ﬂuorescein to tag thymic immigrants, there was no change in thymic output compared to undepleted animals (Gabor et al., 1997). In another study, mice were either oversupplied with thymic emigrants by grafting of additional neonatal thymi under the kidney capsule, or undersupplied by neonatal thymectomy. The thymic export rate was constant from both the intact and the graft thymi, regardless of whether the mouse was oversupplied or undersupplied, and the peripheral pool remained the same (Berzins et al., 1998). These data suggest that the thymus-dependent regeneration rate is not altered by the size of the peripheral pool, or by the number of recent thymic emigrants in the periphery of adults. However, it remains to be determined whether the rate of peripheral mechanisms of regeneration can increase as a result of depletion. Combining the data from these studies, maintenance of the overall T cell pool in the adult can be said to result from proliferation of existing cells. However, renewal of the naive pools may differ for the two subsets; CD4⫹ cells are renewed through thymus-dependent mechanisms whereas CD8⫹ cells can be renewed by both thymus-dependent and independent mechanisms. III. T Cell Renewal from Progenitors in HIV-1 Infection

The impact of HIV-1 infection on T cell renewal from a progenitor source was originally thought to be minor based on the lack of infection of bone marrow progenitor cells (Stanley et al., 1992; Davis et al., 1991).

306

DAWN R. CLARK et al.

However, with the advent of potent antiretroviral therapies more interest has focused on parameters of immune reconstitution in infected individuals. For an adult infected with HIV-1 even the low thymic regeneration found in uninfected adults is likely to be lacking. Determination of the level of regeneration in humans is complicated by a lack of suitable markers for distinguishing between cells that developed from a progenitor and those arising by proliferation of existing cells. Because the naive compartment is thought to proliferate little, if at all, without alteration of the CD45 phenotype, increases in this pool are considered to result from development of new cells. By contrast, increases in the memory pool must be via proliferation either of existing memory cells or of activated naive cells, which will then change their CD45 phenotype. For newly developed cells, it is difﬁcult to determine whether cells developed within the thymus or outside the thymus. Despite these difﬁculties, researchers have used a variety of means to try to assess the level of regeneration in HIV-1-infected individuals. A. REGENERATION ASSESSMENT BY BONE MARROW FUNCTION There are several pieces of evidence that the bone marrow of HIV-1infected persons displays developmental abnormalities. This observation can be explained by changes in the progenitors and/or by alterations in the stroma that support development of the progenitors (Moses et al., 1996). The ﬁrst observations suggesting diminished development of bone marrow were of the multiple cytopenias experienced by many infected individuals (Scadden et al., 1989). Enumeration of progenitor subsets by mAb staining of surface molecules showed the presence of all progenitor subsets in bone marrow from HIV-1-infected individuals (Weichold et al., 1998), though some reports suggest that certain subsets have been reduced in HIV-1 infection (Marandin et al., 1996; Bagnara et al., 1990). No alteration in the number of very primitive progenitors in infected persons could be detected in assays that support the development of these cells (Weichold et al., 1998). Therefore, there is no clear evidence that the number of progenitors has been adversely affected by HIV-1 infection. Bone marrow stroma from HIV-1 infected persons is able to promote the development of progenitors from an uninfected individual (Sloand et al., 1997). However, the progenitors from the infected individual were unable to develop on bone marrow stroma from the uninfected individual. Bone marrow from HIV-1-infected individuals was shown to have diminished capacity to develop cells of the granulocyte, erythrocyte, and megakaryocyte lineages in in vitro colony assays (Zauli et al., 1992, 1996; Steinberg et al., 1991). Because the T cell lineage is derived from the same common precursor as these lineages, it was reasonable to assume that development of this lineage would also be affected. These data add

T CELL DYNAMICS IN HIV-1 INFECTION

307

to the accumulating evidence that the bone marrow progenitors of HIV1-infected individuals are impaired in their ability to develop to mature hematopoietic cells of multiple lineages. B. PROGENITOR DEVELOPMENT CAPACITY MEASUREMENTS Progenitor development to the T cell lineage can be measured by in vitro T cell development systems, such as fetal thymus organ culture or thymic monolayer cultures, or by development in SCID-hu mice. HIV-1infected SCID-hu mice show a thymic pathology, similar to that of infected children, with a lack of thymic subsets beyond the very primitive triplenegative stage and alterations in stromal architecture (Aldrovandi et al., 1993). Interpretation of these results in the context of progenitor cell function is difﬁcult because the thymic tissue is also of human origin and a potential target for infection. Additionally, the thymic and liver tissues used are of fetal origin, which may have a different pattern of response than adult thymic tissue and bone marrow. Progenitor function has been measured by fetal thymus organ culture (FTOC). This technique uses a murine fetal thymus to support the development of human progenitors into mature T cells. This xenogeneic system eliminates the confounding factor of the thymic tissue as a target for infection. In the initial cross-sectional study, the ability of HIV-1-infected individuals to develop T cells in FTOC was shown to be signiﬁcantly lower than that of uninfected individuals (Clark et al., 1997). This impairment in T cell development capacity was seen in all individuals, even those with normal numbers of CD4⫹ T cells. A longitudinal study comparing individuals who progressed to AIDS and long-term nonprogressors (LTNPs) showed that progressors lost T cell development capacity very early in infection but LTNPs still retained signiﬁcant capacity after 8 years of infection (Clark et al., 1998). C. THYMIC FUNCTION ASSESSMENT Another way of measuring the effect of HIV-1 infection on T cell development is by assessing thymic function, particularly because the naive CD4⫹ T cell compartment has been shown to require a functioning thymus for its maintenance. Thymic mass of HIV-1-infected adults has been measured by computerized tomography (CT) scan (McCune et al., 1998). A number of individuals over 40 years of age were found to have larger thymic mass than the same aged uninfected individuals. The number of naive cells in the periphery could be correlated to the amount of thymic mass. These results can be interpreted two ways. The authors concluded that depletion in the periphery due to HIV-1 infection caused a response in the size of the thymus; the thymus got bigger to compensate for the loss of cells.

308

DAWN R. CLARK et al.

However, there is no evidence that this occurs in other cases of peripheral depletion and the size of the thymus in these individuals, prior to infection, is not known. Clearly, however, these results show that the key to maintaining the naive pool is development of naive cells, which requires a functional thymus. Another technique has been used to measure thymic function. Excision circle polymerase chain reaction (PCR) detects T cells that have recently recombined their T cell receptor (TCR) genes. When a progenitor cell develops into a mature T cell, it must recombine gene segments to form a functional T cell receptor. During this process in 움웁-TCR-bearing cells (the majority of T cells in the body), the entire ␦ locus is excised and forms a stable circle in the nucleus. These excision circles, detected by speciﬁc PCR, are only found in CD45RA⫹ cells (Douek et al., 1998). This technique has been described as a measure of thymic function. In fact, it is a measurement of development of T cells irrespective of thymic function, because any developing T cell must recombine TCR genes. In the case of CD4⫹ cells, however, this technique would measure thymic function because development of this subset is thought to occur only in a thymus. Using this technique, Koup and co-workers (Douek et al., 1998) have shown that the number of cells in the periphery expressing excision circles declines with HIV-1 infection. Taken together, results on thymic function support the contention that HIV-1 infection alters the T cell renewal capability of infected individuals. These data show that HIV-1 infection has a direct inhibitory effect on the already low thymic-dependent development in adults. This translates to a reduced development of CD4⫹ cells with a naive phenotype and places even more emphasis on the peripheral expansion of CD4⫹ cells to maintain this population as it is being depleted by the virus. Immunodeﬁciency may be directly related to this thymus-independent T cell renewal. The cells that result from peripheral expansion have been shown to be less capable of responding to neoantigen (Mackall and Gress, 1997b), to have a skewed TCR repertoire (Mackall et al., 1993), and to be prone to apoptosis (Hakim et al., 1997). These data also suggest that the eventual depletion of CD4⫹ cells is, at least in part, due to a block in the development of T cells from a progenitor source, which prevents the complete regeneration of the naive CD4⫹ T cell compartment. In addition, these data explain the observation that naive CD8⫹ T cells are also depleted in HIV-1 infection. Whether the rate of peripheral mechanisms of regeneration are incapable of increasing to make up for the loss of T cells, or whether HIV-1 infection directly interferes with these peripheral mechanisms as well, remains to be determined.

T CELL DYNAMICS IN HIV-1 INFECTION

309

D. T CELL RENEWAL AFTER POTENT ANTIRETROVIRAL THERAPY The effect of antiretroviral therapy on T cell renewal can provide additional information on the effect of HIV-1 infection on this parameter. Patients experience a rise in the number of cells in the periphery after initiation of therapy. Initially, the majority of this increase consists of cells with a memory phenotype; however there is a slow, consistent rise in the number of naive cells in most individuals (Kostense et al., 1998; Pakker et al., 1998; Gorochov et al., 1998). Data from FTOC has shown that the T cell development capacity of progenitors increases after therapy (Clark et al., 1998). This increase could be correlated with the number of naive cells in the periphery. In addition, the number of cells bearing TCR excision circles goes up after therapy (Douek et al., 1998). In children infected with HIV-1, those with more thymic volume prior to therapy had the largest increase in number of CD4⫹ T cells, in CD4⫹ CD45RA/RO ratio, and in TCR repertoire (Vigano et al., 1998). These data show that functional progenitors, in combination with functional thymic tissue, are required for reconstitution of treated individuals. They also support the contention that immune depletion in HIV-1 infection is, at least partially, the result of nonfunctional progenitors and/or nonfunctional thymic tissue. IV. Getting Quantitative on CD4ⴙ T Cell Production

Ho et al. (1995) and Wei et al. (1995), based on the slopes of CD4⫹ T cell increase in the ﬁrst 30 days after strong antiretroviral therapy, concluded that there was high turnover of CD4⫹ T cells in HIV-1-infected individuals. It was calculated that 2 ⫻ 109 CD4⫹ T cells were destroyed and replaced per day. A high turnover of CD4⫹ T cells in HIV-1 infection was compatible with the observed high viral replication, likely to lead to high numbers of infected cells with a short life span (Perelson et al., 1996). At the time, few estimates of normal rates of CD4⫹ T cell production were available. It was therefore assumed that normal production would be found in those infected patients with little increase in number of CD4⫹ cells after start of therapy. Therefore, it was concluded that in other patients the CD4⫹ T cell production was increased up to 70-fold as reﬂected by the enormous increase in CD4⫹ T cell numbers in the blood after start of therapy. This conclusion was predicated on the basic assumption that the cells that appeared in the blood were newly produced and spared from killing by the virus due to the therapy. Apart from the debate on the origin of the CD4⫹ cells that repopulate the blood, the issue of T cell turnover in HIV infection was hampered by the lack of proper data on normal T cell turnover in humans. To date, new studies have provided more insight

310

DAWN R. CLARK et al.

into the amount of ongoing T cell proliferation, both in HIV-1-infected persons and in healthy humans. The dynamics of lymphocyte populations have largely been studied in mice, and for these small rodents it was typically concluded that the normal turnover is high (Freitas and Rocha, 1993; Freitas et al., 1986; Sprent and Tough, 1994). The consensus is that murine naive T cells are relatively long-lived and are mostly produced by progenitor renewal in the thymus (Dutton et al., 1998). In mice, thymectomy typically leads to loss of naive CD4⫹ T cells on a time scale of weeks, whereas memory CD4⫹ T cell numbers remain unaffected (Swain et al., 1990). Indeed total body counts of naive and memory CD8⫹ T lymphocytes seem to be regulated independently (Tanchot and Rocha, 1995). Memory T cells can be long or shortlived, are generated during immune reactions, and are partly maintained by low-level proliferation (Dutton et al., 1998). There is strong evidence in favor of a homeostatic regulation of total body lymphocyte counts (Freitas et al., 1996), and this homeostatic control involves lymphocyte speciﬁcity and repertoire diversity (Freitas et al., 1996; Mclean et al., 1997; DeBoer and Perelson, 1997). Finally, naive and memory CD8⫹ T cells in transgenic mice have different survival and renewal requirements (Tanchot and Rocha, 1997); naive cells require the correct major histocompatability complex (MHC) restriction element for survival and the correct MHC with antigen for expansion. Memory CD8⫹ T cells, on the other hand, require any MHC class I molecule for survival and the correct MHC restriction element for expansion (Tanchot and Rocha, 1997). It is not known how these rodent data translate to the human system. We will here focus on human CD4⫹ T cell production and turnover and will suggest that the normal human lymphocyte turnover is low, about 1% per day, corresponding to a production of about 2.5 ⫻ 109 CD4⫹ T cells per day. A. TOTAL BODY NUMBERS OF CD4⫹ AND CD8⫹ T CELLS Estimates for the total body numbers of CD4⫹ T lymphocytes are calculated either by extrapolating from peripheral blood counts, or from small samples of lymphoid tissue. One should be careful, however, with extrapolations from blood measurements, because only a small fraction of the T lymphocytes resides in the peripheral blood. Such estimates are extremely sensitive to small changes in the distribution of T cells over the blood and lymphoid tissue. It is conventionally assumed that in a healthy human adult 2% of the lymphocytes resides in the blood (Westermann and Pabst, 1990). Considering that human adults have 5 liters of blood, with a typical CD4⫹ T cell count of a 1000 cells/애l, one obtains a total body estimate of 2.5 ⫻ 1011 CD4⫹ T cells (Ho et al., 1995). Although the value 2% for the percentage of lymphocytes residing in the blood is used

T CELL DYNAMICS IN HIV-1 INFECTION

311

throughout the literature, the different subsets of lymphocytes probably have somewhat different percentages. In the peripheral blood the CD4 : CD8 ratio is approximately one, but in the lymphoid tissue CD4⫹ T cells are overrepresented (Westermann and Pabst, 1990; Fleury et al., 1998). Assuming that the conventionally used 2% of lymphocytes in the blood remains valid as an average over the different lymphocyte subsets, an average of 1.6% of the CD4⫹ cells residing in the blood was calculated from measurements in lymph nodes and peripheral blood (Fleury et al., 1998). This is reassuringly close to the conventional estimate of 2%. Similar studies (Zhang et al., 1998) document the numbers of CD4⫹ T cells in peripheral blood and lymphoid tissue (mostly tonsils). Assuming 700 g of lymphoid tissue in 70-kg individuals, the data from ﬁve HIV-negative subjects on average yield a total of 2.24 ⫻ 1011 CD4⫹ T cells in the lymphoid tissue. The peripheral blood counts in the same group of subjects yield an average of 4.85 ⫻ 109 CD4⫹ T cells in the peripheral blood. Thus, 2.1% of the CD4⫹ T cells resides in the blood (Zhang et al., 1998). Both this total body estimate and the distribution of CD4⫹ T cells over the blood and lymphoid compartments are in close agreement with the earlier extrapolation from the peripheral blood. The CD4⫹ T cells in a typical human adult are composed of naive CD45RA⫹ and memory CD45RO⫹ cells in an approximately 1 : 1 ratio (Cossarizza et al., 1996; DePaoli et al., 1988). This should correspond to an order of magnitude of 1011 cells in each subclass of CD4⫹ T cells. For CD8⫹ T cells the situation is somewhat different, however, because they are underrepresented in the lymphoid tissue. Estimates for the CD4 : CD8 ratio in lymphoid tissue vary between 2.5 (Westermann and Pabst, 1992), 3.7 (Tenner-Racz et al., 1998), and 5 (Fleury et al., 1998). Based on the assumption outlined above, an average of 5.6% of the CD8⫹ cells residing in the blood was calculated in HIV-negative subjects (Fleury et al., 1998). Thus, a CD8⫹ T cell count of a 1000 cells/애l yields a body total of approximately 1011 cells, which is 2.5-fold less than that of the CD4⫹ T cells. In HIV-1-infected persons, the distribution of CD4⫹ and CD8⫹ lymphocytes differs in blood and lymphoid tissue (Fleury et al., 1998; Zhang et al., 1998). This distribution seems to normalize during highly active antiretroviral therapy (HAART) (Mosier et al., 1995; Pakker et al., 1998; Gorochov et al., 1998; Zhang et al., 1998; Hellerstein and McCune, 1997; Sprent and Tough, 1995). In a group of HIV-positive patients in early stage of disease (Fleury et al., 1998), the total body numbers of CD8⫹ T cells in the peripheral blood increase approximately 2-fold, and in the lymphoid tissue approximately 3-fold. As a consequence, the percentage of CD8⫹ T cells residing in the blood decreases from 5.6% to approximately 3.2%. Thus, during HIV-1 infection there seems to be an increased trapping

312

DAWN R. CLARK et al.

of CD8⫹ T cells in the lymphoid tissue. For the CD4⫹ T cells in the same set of data (Fleury et al., 1998), numbers in the peripheral blood and lymphoid tissue both decrease approximately 0.8-fold, such that the percentage of CD4⫹ T cells residing in the blood is 1.8% (Westermann and Pabst, 1990). For a group of patients at later stages of disease, however, the distribution of CD4⫹ T cells over the peripheral blood and the lymphoid tissue does suggest CD4⫹ T cell trapping in the lymphoid tissue (Zhang et al., 1998). Before the onset of HAART these patients have total body counts of 1.2 ⫻ 109 CD4⫹ T cells in the blood, and 9.8 ⫻ 1010 CD4⫹ T cells in the lymphoid tissue, which corresponds to 1.2% CD4⫹ T cells in the blood. After 3 weeks of treatment this percentage has normalized to the conventional 2% in the blood (Zhang et al., 1998). A 1% change in the distribution of CD4⫹ T cells is more than sufﬁcient to explain the marked increase in the CD4⫹ T cell counts in the peripheral blood during HAART (Pakker et al., 1998). B. MEASURING DIVISION RATES BY THE LOSS OF CHROMOSOME DAMAGE During radiotherapy, part of the lymphocytes become ‘‘marked’’ by chromosome damage. Radiation induces two types of microscopically detectable damage. Stable damage consists of breaks in the chromatid and is passed on to one daughter cell during cell division. Unstable damage consists of dicentric rings and leads to death of the cell in the next mitosis (Michie et al., 1992). Because cells marked by the unstable chromosome damage die on cell division, the loss rate of such cells is a measure for their division rate (Michie et al., 1992). Conversely, cells marked by stable chromosome damage should only disappear at their normal death rate (Mclean and Michie, 1995). The latter assumption is probably inaccurate, however, because cells with stable chromosome damage may continue to be produced, which probably accounts for the very long estimated average lifetime of both naive and memory T cells (Mclean and Michie, 1995). Thus, only data on unstable chromosome damage are discussed (Michie et al., 1992). In a group of 19 patients treated with radiotherapy the lymphocyte count in the peripheral blood falls and slowly recovers. During the recovery period the number of CD45RA⫹ cells marked by their unstable chromosome damage ﬁrst increases and then declines. The marked CD45RO⫹ T cells decline continuously (Michie et al., 1992). To allow for an increase in marked CD45RA⫹ T cells, the data are ﬁtted to a mathematical model allowing for a reversion from the CD45RO to the CD45RA phenotype, and for different (division-associated) death rates for the CD45RA⫹ and CD45RO⫹ T cells. Assuming that most of the death is indeed associated with division, the results showed that naive T cells on

T CELL DYNAMICS IN HIV-1 INFECTION

313

average divide every 1000 days, that memory T cells divide every 263 days, and that memory cells revert to the CD45RA⫹ phenotype every 278 days (Michie et al., 1992). This classic study established that memory T cells divide three times more frequently than naive T cells do, implying that long-lived memory is maintained by long-lived clones rather than by longlived memory cells (Michie et al., 1992). The fact that CD45RO⫹ divide more frequently than naive T cells is now conﬁrmed by a study using Ki67 as a marker for dividing cells (Sachsenberg et al., 1998). One should keep in mind that these estimated division rates are the sum of the natural death and the true division rate, that this study lumps CD4⫹ with CD8⫹ T cells, and that the immune systems of these patients are recuperating from radiation therapy and are not at steady state. We expect that the CD8⫹ T cell population is largely responsible for the estimated CD45RO to CD45RA reversion (Mackall and Gress, 1997a). Finally, one may employ these estimated division rates for calculating the total body T cell production. Assuming a total body count of 2.5 ⫻ 1011 CD4⫹ T cells and 1011 CD8⫹ T cells, a total of 3.5 ⫻ 1011 T cells, one obtains a production of about 3.5 ⫻ 108 naive T cells, and 109 memory T cells per day. Both ﬁgures are in good agreement with the results of studies employing Ki67 that are discussed below. C. RECOVERY RATES FOLLOWING CD4⫹ T CELL DEPLETION Adult human patients typically recover slowly from depletion of CD4⫹ T cells by chemotherapy, radiotherapy, or CD4 monoclonal antibody treatment (Mackall et al., 1995; Moreland et al., 1995; Rep et al., 1997). In most cases naive CD4⫹ T cells were depleted more strongly than the memory CD4⫹ T cells. CD4⫹ T cell recovery correlated strongly with age and thymic function in a group of patients treated with chemotherapy, and this is especially true for the CD4⫹ CD45RA⫹ subset (Mackall et al., 1995). In children the typical recovery rates are much better. Following treatment with CD4 mAb, the decrease in the number of CD4 T cells in the peripheral blood of two patients with juvenile chronic arthritis was only short-lasting, and numbers returned to normal values within 1 to 8 weeks (Horneff et al., 1995). One can employ the recovery rates following lymphocyte depletion for estimating the total body production rate of CD4⫹ T cells. Immune systems recuperating from CD4⫹ T cell depletion by chemotherapy, radiotherapy, or CD4 mAb treatment are not at a natural steady state, however. Thus the estimated production rates could either be too high, when density-dependent effects increase CD4⫹ T cell production at low total body counts, or too low, when only part of the TCR repertoire is recovering. A study of patients recovering from chemotherapy demonstrated the importance of thymic function (Mackall et al., 1995). Here we estimate recovery rates for some of the adult patients in this

314

DAWN R. CLARK et al.

study. The three 24-year-old patients in this set recovered their CD4⫹ T cell counts in the peripheral blood with linear recovery rates of 0.54, 0.26, and 0.63 cells/애l per day (Mackall et al., 1995). For the total body such a recovery rate of about 0.5 cells/애l per day amounts to a production of a magnitude of 108 CD4⫹ T cells/day. For a 23-year-old patient the study (Mackall et al., 1995) provides more detailed data on the CD4⫹ T cell recovery. Fitting the data by linear regression, both a linear growth and an exponential growth model ﬁt reasonably well. This ﬁrst yields an estimated growth rate of 1 cell/애l per day, the second a growth rate of 0.008 per day. By order of magnitude this amounts to a total production of about 108 CD4⫹ T cells day. Note that these two estimates are about 10-fold lower than the estimates reviewed above. Treatment of rheumatoid arthritis patients with CD4 mAb resulted in a severe CD4⫹ T cell depletion to counts of about 300 CD4⫹ T cells/애l (Moreland et al., 1995). Figure 2 depicts these data and shows that following CD4⫹ T cell depletion by this means, the recovery is slow. Six months after treatment, CD45RO⫹ memory cells numbers had returned to normal, but the counts of CD45RA⫹ naive CD4⫹ T cells were only 50% of the normal count. By linear regression one ﬁnds an average daily increase of

FIG. 2. Recovery of CD4⫹ T cells in rheumatoid arthritis patients treated with CD4 mAb. Symbols indicate the mean CD4 count/애l peripheral blood: 䊉 total CD4⫹ T cells; 䊏, CD4⫹ CD45RO⫹ memory T cells; 䉬, CD4⫹ CD45RA⫹ naive T cells. The three regression lines have slopes of 2.5, 1.5, and 1.1 cells/애l/day, respectively.

T CELL DYNAMICS IN HIV-1 INFECTION

315

2.5, 1.5, and 1.1 CD4⫹ T cells/애l, for the total, the memory, and the naive CD4⫹ T cells, respectively. Because the production of memory cells is only marginally higher than that of the naive cells, the last two estimates amount to a total body production of approximately 108 naive CD4⫹ T cells and 108 memory CD4⫹ T cells per day. These production rates are very similar to those of the patients treated with chemotherapy (Mackall et al., 1995). The production rate of 108 naive CD4⫹ T cells per day corresponds closely to the similar estimate based on the unstable chromosome damage data (Michie et al., 1992). For the memory CD4⫹ T cells, however, the current estimate of 108 cells/day equals that of the chemotherapy patients (Mackall et al., 1995), which was 10-fold lower than our other estimates. Similar data on CD4⫹ T cell depletion come from a study of multiple sclerosis patients that had been administered CD4 mAb (Rep et al., 1997). These patients received antibody until the CD4⫹ T cell count in the peripheral blood dropped to 200 cells/애l. The subsequent recovery of the CD4⫹ T cell count was exceedingly slow. Over the time course of 1 year, the average CD45RA⫹ naive CD4⫹ T cell count in the blood increased approximately linearly from 75 to 110 cells/애l. This increase in naive cells was paralleled by a similar increase in the CD4⫹ memory cells, from 115 to 150 cells/애l (Rep et al., 1997). Both correspond to a daily increase of only 0.1 cell/애l per day. For the total body production these ﬁgures amount to approximately 2.5 ⫻ 107 naive and only 2.5 ⫻ 107 memory CD4⫹ T cells per day. Compared to the data reviewed below, both estimates are 10-fold and 100-fold lower. D. BrdU LABELING OF PROLIFERATING CELLS The most direct method for estimating T cell production is counting the numbers of dividing cells. The DNA precursor bromodeoxyuridine (BrdU) (Tough and Sprent, 1994) was applied to study turnover of lymphocytes in mice. Labeling simian immunodeﬁciency virus (SIV)-infected and noninfected macaques with BrdU showed faster labeling with, and elimination of, BrdU in lymphocytes and natural killer (NK) cells in SIV-infected monkeys (Mohri et al., 1998; Rosenzweig et al., 1998). Labeling and elimination rates of both CD4⫹ and CD8⫹ T cells were reported to increase 2to 3-fold in SIV-infected animals compared to healthy control animals. Conversion of the elimination rates into estimates of average life spans showed that naive T cells live about 16 weeks and memory cells about 7 weeks in normal animals, which is much shorter than previous estimates of T cell life span in humans (Mclean and Michie, 1995). By mathematical modeling, Mohri et al. (1998) calculated a proliferation rate of 3.6 ⫻ 107 CD4⫹ T cells per day and a death rate of 3.6 ⫻ 108 CD4⫹ T cells per day in normal animals, and in SIV-infected animals a proliferation rate of 4 ⫻

316

DAWN R. CLARK et al.

108 T cells and a death rate of 7.5 ⫻ 108 CD4⫹ T cells per day. Therefore, the production rates in infected monkeys were two to three times increased. The same increase in turnover rates was found in CD8⫹ T lymphocytes, NK cells, and B cells from infected animals. BrdU labeling studies as described for mice and macaques cannot be performed in humans. Another approach to evaluate proliferation in peripheral blood lymphocytes (PBLs) is to incubate freshly isolated PBLs ex vivo with BrdU (Tissot et al., 1998). In this study, increased DNA synthesis could be observed only in patients with less than 100 CD4⫹ T cells/mm3 who had opportunistic infections. Thus, labeling peripheral blood T cells ex vivo provided no evidence for extensive proliferation in these cells in asymptomatic HIV-1 infection. V. Measuring Cell Division with the Ki67 mAb

The Ki67 antigen is expressed by human proliferating cells during the late G1, S, G2, or M phases of the cell cycle (Bruno and Darzynkiewicz, 1992; Schwarting et al., 1986; Tsurusawa et al., 1992), and the number of cells can be quantiﬁed with the Ki67 mAb by ﬂow cytometry. Because the vast majority of lymphocyte cell divisions takes place in the lymphoid tissue, measurements within the lymphoid tissue seem most reliable (Zhang et al., 1998; Fleury et al., 1998; Tenner-Racz et al., 1998). In HIV-infected patients there is the additional problem that dividing CD4⫹ T cells are targets for HIV-1 infection. If Ki67⫹ CD4⫹ T cells die by HIV-1 infection, the total body numbers of such cells can only be a lower bound for the true CD4⫹ T cell production. A. HEALTHY HUMAN ADULTS Few groups have published data on Ki67 expression in the peripheral blood and lymphoid tissue of HIV-negative subjects. The percentages of CD4⫹ cells expressing Ki67 and the calculations of cell production based on these results are presented in Table I. Two studies (Zhang et al., 1998; Fleury et al., 1998) report that about 0.5% of the CD4⫹ T cells in lymphoid tissues express the Ki67 antigen. Having about 2.5 ⫻ 1011 CD4⫹ T cells in lymphoid tissue, and assuming that the Ki67 antigen is expressed over most of the cell cycle of about 24 hr, these percentages correspond to a total body production of about 109 CD4⫹ T cells per day. Another group distinguished between the T cell zone and the germinal center area in lymph nodes, and reports somewhat higher percentages (Tenner-Racz et al., 1998). In the T cell zone 2.6% of the CD4⫹ T cells expresses the Ki67 antigen, whereas in germinal centers 1.4% of the CD4⫹ T cells is Ki67⫹. Recalculating these percentages in terms of total body production is more

317

T CELL DYNAMICS IN HIV-1 INFECTION

TABLE I CD4⫹ T CELL PRODUCTION RATES ESTIMATED BY THE Ki67 mAb Ki67⫹ Sourcea

Statusb

%

Foldc

Number

Fold

PBMCe PBMCe PBMCe PBMCe PBMC f PBMC f CD45RO⫹f CD45RO⫹f CD45RA⫹f CD45RA⫹f LNe LNe LNe LNg LNg LNg LNg LNg GCh GCh T cell zoneh T cell zoneh

HIV⫺ HIV⫹ 4–8 wk H 30 wk H HIV⫺ HIV⫹ HIV⫺ HIV⫹ HIV⫺ HIV⫹ HIV⫺ HIV⫹ 30 wk H HIV⫺ HIV⫹ 2 day H 3 wk H 24 wk H HIV⫺ HIV⫹ HIV⫺ HIV⫹

0.60 0.80

1.3

3.9 ⫻ 109 3.4 ⫻ 109

0.89

1.1 6.5 2.8 6.7 0.8 2.7 0.54 0.39 0.4 1.2 1.0 1.0 0.4 1.4 18.7 2.6 3.7

Productiond

Estimated total

5.9 2.4 3.4

4.2 2.0 1.8 9.1

⫻ ⫻ ⫻ ⫻

109 109 109 108

0.46 0.52

0.72

2.5 ⫻ 1011 1.9 ⫻ 1011

0.78

3 2.5 2.5 1.0

2.2 9.5 9.0 1.1 1.2

⫻ ⫻ ⫻ ⫻ ⫻

0.47 0.45 0.55 0.60

1011 1010 1010 1011 1011

Number 2.2 2.7 1.0 6.9 4.4 8.3 4.4 7.0

⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻

107 107 107 107 107 107 107 107

1.3 6.5 3.5 8.0 1.1 9.0 1.1 4.8

⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻ ⫻

109 108 109 108 109 108 109 108

Fold 1.2 0.45 3.14 1.9 1.6

0.51 2.7 1.4 1.1 1.4 0.6

13.4 1.4

a

PBMC, peripheral blood mononuclear cells; LN, lymph node; GC, germinal center. H, highly active antiretroviral therapy. Fold increase compared to the HIV-negative value. d Estimated daily production in the compartment (assuming Ki67 is expressed for most of the 24-hr cycle). e Fleury et al. (1998). f Sachsenberg et al. (1998). g Zhang et al. (1998). h Tenner-Racz et al. (1998). b c

difﬁcult because it is not known what fraction of CD4⫹ T cell population resides in T cell zones, and in the germinal centers, respectively. In germinal centers CD4⫹ T cell numbers tend to be low (McHeyzer-Williams and Davis, 1995). Ki67 measurements in the peripheral blood also suggest low division rates of the CD4⫹ T cells. Comparing peripheral blood with lymphoid tissue, in HIV-negative subjects the percentage of Ki67⫹ CD4⫹ T cells was similar in both compartments (Fleury et al., 1998). Sachsenberg et al. (1998) reported a very similar percentage of Ki67⫹ CD4⫹ T cells in the

318

DAWN R. CLARK et al.

blood of healthy adults. This study also distinguished CD4⫹ CD45RA⫹ naive from CD45RO⫹ memory T cells, and reports that in the blood, four times more CD45RO⫹ CD4⫹ T cells were Ki67⫹ than naive CD45RA⫹ CD4⫹ T cells (Sachsenberg et al., 1998). Although it remains difﬁcult to extrapolate total body estimates for CD4⫹ T cell production from data in peripheral blood, these conﬁrm that, on average, memory T cells divide more frequently than naive T cells (Michie et al., 1992; Mclean and Michie, 1995; DeBoer and Noest, 1998). Data regarding Ki67 expression in CD8⫹ T cells are summarized in Table II. In the normal situation the total production of CD8⫹ T cells seems to be almost 10-fold lower than CD4⫹ T cell production (Fleury et al., 1998). In the lymphoid tissue one ﬁnds that only 0.2% of the CD8⫹ T cells expresses the Ki67 antigen. This would correspond to a total body production of about 108 CD8⫹ T cells per day. Measurements in the T cell zone of lymph nodes of healthy human adults, however, suggest a more than 10-fold higher percentage Ki67⫹ CD8⫹ T cells (Tenner-Racz et al., 1998), which would correspond to a more than 10-fold higher producTABLE II CD8⫹ T CELL PRODUCTION RATES ESTIMATED BY THE Ki67 mAb Ki67⫹ Sourcea

Status b

%

PBMC⫹e PBMCe PBMC f PBMC f CD45RO⫹⫹f CD45RO⫹⫹f CD45RA⫹⫹f CD45RA⫹f LNe LNe LNe T cell zoneg T cell zoneg

HIV⫺ HIV⫹ HIV⫺ HIV⫹ HIV⫺ HIV⫹ HIV⫺ HIV⫹ HIV⫺ HIV⫹ 30 wk H HIV⫺ HIV⫹

0.41 1.2 1 4.3 2 6.8 0.9 1.7 0.17 0.38

2.2

2.9 5.5

1.9

a

Foldc 3 4.3 3.4 1.9

Productiond

Estimated total Number 3.2 5.7 2.2 3.3 4.7 1.0

⫻ ⫻ ⫻ ⫻ ⫻ ⫻

109 109 109 109 109 109

5.4 ⫻ 1010 1.7 ⫻ 1011

Fold 1.8 1.5 2

3.1

Number 1.2 7.1 2.1 1.3 8.5 4.8

⫻ ⫻ ⫻ ⫻ ⫻ ⫻

107 107 107 108 106 107

9.5 ⫻ 107 6.5 ⫻ 108 9.1 ⫻ 108

Fold 5.8 6.2 6

6.8 9.6

PBMC, peripheral blood mononuclear cells; LN, lymph node. H, highly active retroviral therapy. c Fold increase compared to the HIV-negative value. d Estimated daily production in the compartment (assuming Ki67 is expressed for most of the 24-hr cycle). e Fleury et al. (1998). f Sachsenberg et al. (1998). g Tenner-Racz et al. (1998). b

T CELL DYNAMICS IN HIV-1 INFECTION

319

tion of CD8⫹ T cells per day. The reasons for this discrepancy remain unclear. For the CD8⫹ T cells in the peripheral blood one ﬁnds smaller discrepancies. Fleury et al. (1998) report that 0.41% of the CD8⫹ T cells in the blood expresses the Ki67 antigen, whereas Sachsenberg et al. (1998) reported that 1.0% of the CD8⫹ cells is Ki67⫹. Distinguishing CD45 subsets, the latter study also reports that in blood CD8⫹ T cells 0.9% of the CD45RA⫹ cells, and 2.0% of the CD45RO⫹ cells, express the Ki67 antigen (Sachsenberg et al., 1998). This again demonstrates that CD45RO⫹ T cells divide more frequently than CD45RA⫹ T cells. B. HIV-INFECTED PATIENTS In HIV-positive patients the percentages of Ki67⫹ T cells tend to increase in both CD4⫹ and CD8⫹ T cell compartments (Fleury et al., 1998; Sachsenberg et al., 1998; Tenner-Racz et al., 1998). This is at least partly due to generalized immune activation that is associated with HIV-1 infection (Mohri et al., 1998). A mere increase in the percentages of Ki67⫹ T cells need not imply, however, that the total production in the CD4 and CD8 compartments is proportionally increased. In the lymphoid tissue of a group of early-stage HIV-patients, the percentage of Ki67⫹ cells has changed 0.7-fold for CD4⫹ cells, and 2.2-fold for CD8⫹ cells, compared to healthy controls (Fleury et al., 1998). However, because the total numbers of CD4⫹ T cells have decreased in these patients, the total production of CD4⫹ T cells seems somewhat less than normal (Fleury et al., 1998). The total numbers of CD8⫹ T cells have increased, such that the total production of CD8⫹ T cells is also increased (Fleury et al., 1998). Note that in the peripheral blood of these patients both the percentage of Ki67⫹ CD4⫹ T cells and the percentage of Ki67⫹ CD8⫹ T cells have increased (Fleury et al., 1998). In HIV-positive patients, dividing (Ki67⫹ ) T cells are apparently overrepresented in the blood, thus one cannot simply extrapolate from peripheral blood measurements to total body estimates. In a group of HIV patients at a later stage of disease Zhang et al. (1998) report that an average of 1.2% of the CD4⫹ T cells in the lymphoid tissue expresses the Ki67 antigen. This is 3-fold higher than their control value (Zhang et al., 1998), and 3-fold higher than the percentages found in earlystage patients (Fleury et al., 1998). The total body CD4⫹ T cell counts are substantially depleted in these patients, however, on average 0.44-fold lower (Zhang et al., 1998). The 3-fold higher percentage of dividing CD4⫹ T cells could therefore be due to a more generalized hyper-activation at this stage of disease, and/or due to an increased CD4⫹ T cell growth rate by density mechanisms. Although the percentage of dividing CD4⫹ T cells has increased 3-fold because of the HIV-1 infection, the total body production of CD4⫹ T cells has increased only 1.27-fold because of the

320

DAWN R. CLARK et al.

depletion of total CD4⫹ T cell numbers. Over 24 weeks of HAART the percentage of Ki67⫹ CD4⫹ T cells in the lymphoid tissue normalizes, whereas the number of CD4⫹ T cells in lymphoid tissue increases only slightly (Zhang et al., 1998). Thus, CD4⫹ production decreases with HAART. Similar results were reported by Fleury et al. (1998). This drop in Ki67 expression during HAART has two important implications. First, it demonstrates that HIV-1 infection is involved with a level of hyperactivation that falls when the viremia drops. Second, it suggests that there is little masking of cells expressing Ki67 by HIV-1 killing of dividing cells. In a group of 12 HIV patients with CD4⫹ T cell counts ⬎500 cells/애l, much higher percentages of Ki67⫹ T cells were found in the T cell zone and in germinal centers of lymph nodes (Tenner-Racz et al., 1998). We have already mentioned that the healthy control data from this group are also higher than those of other studies (Zhang et al., 1998; Fleury et al., 1998). For the T cell zone the fold increases remain comparable to those of others (Zhang et al., 1998; Fleury et al., 1998). The reason for the high percentage in the germinal centers remains unclear, however. Sachsenberg et al. (1998) documented Ki67 percentages in the peripheral blood and showed that for both CD4⫹ and CD8⫹ T cells the percentages increase with HIV-1 infection. The majority of the production was in the CD45RO⫹ subset. Moreover, it was shown that the percentage of Ki67⫹ T cells increased with decreasing CD4⫹ T cell counts. Again, this could be due to density-dependent mechanisms, or to stronger immune activation due to higher viral loads that are associated with lower CD4⫹ T cell counts (Sachsenberg et al., 1998). Finally, note that if dividing cells are indeed over represented in the peripheral blood (Fleury et al., 1998), all these percentages should be regarded as overestimates. VI. What Is the Cause of CD4ⴙ T Cell Depletion in HIV-1 Infection?

With respect to CD4⫹ T cell depletion it should be noted that this is, in general, measured in the blood. It is now thought that the steep increase in CD4⫹ and CD8⫹ T cell numbers in the blood immediately after initiation of antiviral therapy should be interpreted as redistribution of T cells that were retained in the lymphoid tissues and at other inﬂammatory sites (Pakker et al., 1998; Autran et al., 1997). Interestingly, the magnitude of this redistribution is inversely correlated with baseline CD4⫹ T cell counts and is very limited in patients with baseline CD4⫹ cell counts greater than 400 mm3 (Fleury et al., 1998). This indicates that, in patients with CD4⫹ cell counts below 400 mm3, but rarely in patients with higher CD4⫹ T cell counts, CD4⫹ T cell depletion is overestimated based on the CD4⫹ T cell counts in the blood. In patients with very low CD4⫹ T cell counts, in fact,

T CELL DYNAMICS IN HIV-1 INFECTION

321

CD4⫹ T cell counts sometimes do increase 6-fold within days of the start of therapy due to redistribution of cells that, apparently, were present in tissue before therapy (Pakker et al., 1998; J. M. Prins, personal communication). In line with these observations, it has been reported that in lymphoid tissue the depletion of CD4⫹ T cells is less severe compared to that in blood (Zhang et al., 1998; Rosok et al., 1996). Experimental results on the number of cells productively infected with HIV have shown that these numbers are in fact very low, in particular given the enormous numbers of viral particles produced per day. Chun et al. (1997) reported a total of 5 ⫻ 107 CD4 cells with integrated proviral HIV and Haase and co-workers (Embretson et al., 1993) reported a total of 5 ⫻ 108 cells that are positive for viral DNA. CD4⫹ T cell death may also involve uninfected cells that die through activation-related apoptosis, although this is reportedly much higher for CD8⫹ T cells (Finkel et al., 1995; Meyaard et al., 1992). Zhang et al. (1998) report a twofold increase in CD4⫹ T cell apoptosis in lympoid tissue. Taken together, the number of productively infected cells and the number of cells involved in activationinduced apoptosis are indicative of a modestly increased destruction of CD4⫹ cells in HIV-1-infected patients. Moreover, the fact that production of CD4⫹ T cells is not immediately increased after start of HAART, but increases only later during HAART (Fleury et al., 1998), indicates that in HIV-1 infection there is not a masking of high cell production by efﬁcient killing of dividing cells. The question is whether this modestly increased destruction of CD4⫹ T cells could still be the main cause of the gradual CD4⫹ T cell depletion. This could be the case if CD4⫹ T cell renewal is very limited and is incapable of meeting the higher demand, as has been argued by Haase and colleagues (Zhang et al., 1998). As pointed out previously, much of the controversy on the magnitude of T cell production and destruction in HIV-1 infection relates to what the investigators regard as high turnover or high CD4⫹ T cell production. The calculated production of 2 ⫻ 109 cells/day, based on the initial rise of CD4⫹ T cell numbers in the blood, was believed to be very high and was suggested to be at least an order of magnitude higher than normal (Ho et al., 1995). It was argued that the system was highly stressed trying to keep up with the increased demand due to high levels of CD4⫹ T cell death caused by HIV-1. To account for CD4⫹ T cell depletion, this highly increased level of CD4⫹ T cell production was believed to exhaust the renewal potential. As we have discussed, new experimental evidence on T cell production numbers in normal controls and HIV-infected patients suggests that CD4⫹ T cell production is about two- to threefold increased with HIV-1 infection and is on the order of 109 cells per day (Table I). Studies that provided a mathematical interpretation of telomere length

322

DAWN R. CLARK et al.

dynamics in CD45RA⫹ and CD45RO⫹ CD4⫹ cells demonstrated that telomere shortening rates comparable to uninfected controls are compatible with a small increase in production of CD4⫹ T cells (Wolthers et al., 1996, 1998). Therefore, the data suggest that there is a modest increase in the amount of proliferation of T cells in HIV-1-infected individuals. This low increase in production of CD4⫹ T cells is not likely to result in exhaustion of the CD4⫹ renewal potential, however, which is also indicated by the lack of signiﬁcant telomere shortening. There is now considerable data suggesting that the low level of thymic development that occurs in adults may be further reduced with HIV-1 infection. Studies of development of bone marrow from HIV-infected persons all show a lower capacity to develop cells of multiple hematopoietic lineages (Steinberg et al., 1991; Zauli et al., 1992, 1996). The data from FTOC show that changes in T cell developmental capacity occur rapidly in individuals who progress to AIDS but not in long-term nonprogressors (Clark et al., 1998). In addition, studies in SCID-hu mice and measurement of thymic tissue have shown that the thymus is also affected by HIV-1 infection ( Joshi and Oleske, 1985; Aldrovandi et al., 1993; Mosier et al., 1991; Vigano et al., 1998). The export of new cells from the thymus, as measured by excision circle PCR, have shown that the number of new naive cells in the periphery is reduced after infection (Douek et al., 1998). Many of these deﬁciencies improve with potent antiretroviral therapy (Douek et al., 1998; Clark et al., 1998; Vigano et al., 1998). Therefore, there is now good evidence that in HIV-1 infection the capacity to develop new cells from progenitors is perturbed. Finally, we can propose a coherent model for CD4⫹ cell depletion based on the data we have summarized. During the course of asymptomatic infection an increasing number of naive cells will be activated and become memory cells (Roederer, 1995). If these cells cannot be replaced by development of new naive cells, as described above, the naive pool will be effectively depleted over time. Loss of naive cells could also be accelerated in those individuals harboring syncitium-inducing virus, because these viruses can infect naive cells (H. Blaak, personal communication). The memory pool is gradually depleted by activation-induced cell death and, in the case of CD4⫹ T cells, by HIV-1-related death. Because development is inhibited, there are few naive cells to feed into the memory population, so it must be maintained by increased proliferation of already existing cells. As discussed, the increase in proliferation does not appear to be vast and perhaps, though not directly shown, is insufﬁcient to keep up with cell loss. In any case, the cells that result from this proliferation would be more likely to die sooner (Mackall and Gress, 1997b). The result is the eventual loss of the memory CD4⫹ cells, and concomitant with AIDS diagnosis, the

T CELL DYNAMICS IN HIV-1 INFECTION

323

loss of the memory CD8⫹ T cells. Thus CD4⫹ T cell depletion is likely to be due to the combination of inhibition of development of new cells, and the increase in proliferation of already existing cells. VII. Appendix : Summarizing in Terms of a Mathematical Model

Because most studies suggest that the production of naive CD45RA⫹ cells requires a functional thymus (Mackall et al., 1995; Mackall and Gress, 1997a; Hellerstein and McCune, 1997), we ignore cell division in the naive compartment, and write for the naive CD4⫹ T cells N, dN/dt ⫽ ␴ ⫺ ␦NN ⫺ 움N ⫺ ␧NN2,

(1)

where the ␴ term is the source of naive CD4⫹ T cells from the thymus, the ␦N N term represents death, the 움N term represents activation of naive cells due to priming by antigen, and the ␧NN2 term reﬂects a possible additional density-dependent death rate by competition within the naive compartment (Tanchot and Rocha, 1995). The best current parameter estimate seems to be a production ␴ ⫽ 108 cells/day in human adults. The average life span and the priming rate remain uncertain, however. Additionally, there are no data allowing any direct estimate of the densitydependent additional death rate, ␧N. One can estimate the average life span of a naive CD4⫹ T cell, however, from the estimated production ␴ and the steady-state total body count of 2.5 ⫻ 1011 cells. With a steady-state source of ␴ ⫽ 108 cells/day, the average life span in the naive compartment is 2.5 ⫻ 1011/108 ⫽ 2500 days (or approximately 7 years). For the memory CD4⫹ T cells M we allow for a source from the naive compartment, for renewal, and death. Thus we write that dM/dt ⫽ c움N ⫹ ␳M/(1 ⫹ M/K) ⫺ ␦MM ⫺ ␧MM2,

(2)

where the ␦ and ␧ parameters again allow for a normal and for a densitydependent death rate. The c움N term represents the clonal expansion of the activated naive CD4⫹ T cells (where c is the clonal expansion factor). The renewal term allows for an expansion/proliferation rate of ␳ cell divisions per day when memory cell numbers are low (i.e., when M Ⰶ K ), and a total production of ␳K cells/day when memory cell numbers are high (i.e., when M Ⰷ K ). This proliferation rate ␳ should reﬂect that typically only a small fraction of the memory cells is dividing (␳ should be small). Most studies reviewed above suggest a production of 108 to 109 memory CD4⫹ T cells per day, which seems fairly independent of the total body

324

DAWN R. CLARK et al.

lymphocyte counts. This suggests that the memory CD4⫹ T cell density K at which the renewal levels off should be fairly low, or K Ⰶ 2.5 ⫻ 1011. If memory CD4⫹ T cell production is indeed 10-fold higher than that of their naive counterparts, their average life span at steady state should also be 10-fold lower, about 250 days. What fraction of the total memory production is due to the activation and clonal expansion of naive CD4⫹ T cells, due to the c움N term, and what fraction is due to renewal, remains the major open question, however. Data establishing cell division rates of CD4⫹ memory T cells (Sachsenberg et al., 1998; Michie et al., 1992; McLean and Michie, 1995) unfortunately fail to distinguish between the two possibilities. REFERENCES Aldrovandi, G. M., Feuer, G., Gao, L., Jamieson, B. D., Kristeva, M., Chen, I. S. Y., and Zack, J. A. (1993). Nature (London) 363, 732–736. Ameisen, J. C., and Capron, A. (1991). Immunol. Today 12, 102–105. Autran, B., Carcelain, G., Li, T. S., Blanc, C., Mathez, D., Tubiana, R., Katlama, C., Debre´, P., and Leibowitch, J. (1997). Science 277, 112–116. Bagnara, G. P., Zauli, G., Giovannini, M., Re, M. C., Furlini, G., and La Placa, M. (1990). Exp. Hematol. 18, 426–430. Berzins, S. P., Boyd, R. L., and Miller, J. F. A. P. (1998). J. Exp. Med 187, 1839–1848. Bonyhadi, M. L., Rabin, L., Salimi, S., Brown, D. A., Kosek, J., McCune, J. M., and Kaneshima, H. (1993). Nature (London) 363, 728–732. Bruno, S., and Darzynkiewicz, Z. (1992). Cell Prolif. 25, 31–40. Chun, T.-W., Carruth, L., Finzi, D., Shen, X., DiGiuseppe, J. A., Taylor, H., Hermankova, M., Chadwick, K., Margolick, J., Quinn, T. C., Kuo, Y.-H., Brookmeyer, R., Zeiger, M. A., Bardltch-Crovo, P., and Siliciano, R. F. (1997). Nature (London) 387, 183–187. Clark, D. R., Ampel, N. M., Hallet, C. A., Yedavalli, V. R. K., Ahmad, N., and DeLuca, D. (1997). J. Infect. Dis. 176, 649–654. Clark, D. R., Repping, S., Pakker, N. G., Prins, J. M., Notermans, D. W., Wit, F. W. N. M., Reiss, P., Danner, S. A., Coutinho, R. A., Lange, J. M. A., and Miedema, F. (1998). Diminished T cell renewal in HIV-1 infection contributes to CD4⫹ T cell depletion and is reversed by antiretroviral therapy. Submitted. Cossarizza, A., Ortolani, C., Paganelli, R., Barbieri, D., Monati, D., Sansoni, P., Fagiolo, U., Castellani, G., Bersani, F., Londei, M., and Franceschi, C. (1996). Mech. Ageing Dev. 86, 173–195. Danner, S. A., Carr, A., Leonard, J. M., Lehman, L. M., Gudiol, F., Gonzales, J., Raventos, A., Rubio, R., Bouza, E., Pintado, V., Gil Aguado, A., de Lomas, J. G., Delgado, R., Borleffs, J. C. C., Hsu, A., Valdes, J. M., Boucher, C. A. B., and Cooper, D. A. (1995). N. Engl. J. Med. 333, 1528–1533. Davis, B. R., Schwartz, D. H., Marx, J. C., Johnson, C. E., Berry, J. M., Lyding, J., Merigan, T. C., and Zander, A. (1991). J. Virol. 65, 1985–1990. DeBoer, R. J., and Noest, A. J. (1998). J. Immunol. 160, 5832–5837. DeBoer, R. J., and Perelson, A. S. (1997). Int. Immunol. 9, 779–790. DePaoli, P., Battistin, S., and Santini, G. F. (1988). Clin. Immunol. Immunopathol. 48, 290–296. Douek, D. C., McFarland, R. D., Keiser, P. H., Gage, E. A., Massey, J. M., Haynes, B. F., Polis, M. A., Haase, A. T., Feinberg, M. B., Sullivan, J. L., Jamieson, B. D., Zack, J. A., Picker, L. J., and Koup, R. A. (1998). Nature (London) 396, 690–695.

T CELL DYNAMICS IN HIV-1 INFECTION

325

Dulude, G., Brochu, S., Fontaine, P., Baron, C., Gyger, M., Roy, D.-C., and Perrault, C. (1997). Exp. Hematol. 25, 992–1004. Dutton, R. W., Bradley, L. M., and Swain, S. L. (1998). Annu. Rev. Immunol. 16, 201–203. Embretson, J., Zupancic, M., Beneke, J., Till, M., Wolinsky, S., Ribas, J. L., Burke, A., and Haase, A. T. (1993). Proc. Natl. Acad. Sci. U.S.A. 90, 357–361. Finkel, T. H., Tudor-Williams, G., Banda, N. K., Cotton, M. F., Curiel, T., Monks, C., Baba, T. W., Ruprecht, R. M., and Kupfer, A. (1995). Nature Med. 1, 129–134. Fleury, S., De Boer, R. J., Rizzardi, G. P., Wolthers, K. C., Otto, S. A., Welbon, C. C., Graziosi, C., Knabenhans, C., Soudeyns, H., Bart, P.-A., Gallant, S., Corpataux, J.-M., Gillet, M., Meylan, P., Schnyder, P., Meuwly, J. Y., Spreen, W., Glauser, M. P., Miedema, F., and Pantaleo, G. (1998). Nature Med. 4, 794–801. Freitas, A. A., and Rocha, B. B. (1993). Immunol. Today 14, 25–29. Freitas, A. A., Rocha, B. B., and Coutinho, A. A. (1986). Immunol. Rev. 91, 5–37. Freitas, A. A., Agenes, F., and Coutinho, G. C. (1996). Eur. J. Immunol. 26, 2640–2649. Gabor, M. J., Scollay, R., and Godfrey, D. I. (1997). Eur. J. Immunol. 27, 2986–2993. Gorochov, G., Neumann, A. U., Kereveur, A., Parizot, C., Li, T., Katlama, C., Karmochkine, M., Raguin, G., Autran B, and Debre´, P. (1998). Nature Med. 4, 215–221. Gougeon, M., and Montagnier, L. (1993). Science 260, 1269–1270. Grossman, Z., and Herberman, R. B. (1997). Nature Med. 3, 486–490. Groux, H., Torpier, G., Monte´, D., Mouton, Y., Capron, A., and Ameisen, J. C. (1992). J. Exp. Med. 175, 331–340. Hakim, F. T., Cepeda, R., Kaimei, S., Mackall, C. L., McAtee, N., Zujewski, J., Cowan, K., and Gress, R. E. (1997). Blood 90, 3789–3798. Hamann, D., Baars, P. A., Rep, M. H. G., Hooibrink, B., Kerkhof-Garde, S. R., Klein, M. R., and Van Lier, R. A. W. (1997). J. Exp. Med. 186, 1407–1418. Heitger, A., Neu, N., Kern, H., Panzer-Gru¨mayer, E.-R., Greinix, H., Nachbaur, D., Niederwieser, D., and Fink, F. M. (1997). Blood 90, 850–857. Hellerstein, M. K., and McCune, J. M. (1997). Immunity 7, 583–589. Ho, D. D., Moudgil, T., and Alam, M. (1989). N. Engl. J. Med. 321, 1621–1625. Ho, D. D., Neumann, A. U., Perelson, A. S., Chen, W., Leonard, J. M., and Markowitz, M. (1995). Nature (London) 373, 123–126. Horneff, G., Dirksen, U., Schulze-Koops, H., Emmrich, F., and Whan, V. (1995). Ann. Rheum. Dis. 54, 846–849. Joshi, V. V., and Oleske, J. M. (1985). Arch. Pathol. Lab. Med. 109, 142–146. Kaneshima, H., Su, L., Bonyhadi, M. L., Connor, R. I., Ho, D. D., and McCune, J. M. (1994). J. Virol. 68, 8188–8192. Kostense, S., Raaphorst, F. M., Notermans, D. W., Joling, J., Hooibrink, B., Pakker, N. G., Danner, S. A., Teale, J. M., and Miedema, F. (1998). AIDS 12, F235–F240. Mackall, C. L. (1997). Blood 89, 3700–3707. Mackall, C. L., and Gress, R. E. (1997a). Immunol. Rev. 157, 61–72. Mackall, C. L., and Gress, R. E. (1997b). Immunol. Rev. 160, 91–102. Mackall, C. L., Granger, L., Sheard, M. A., Cepeda, R., and Gress, R. E. (1993). Blood 82 (No. 8), 2585–2594. Mackall, C. L., Fleisher, T. A., Brown, M., Andrich, M. P., Chen, C., Feuerstein, I. M., Horowitz, M. E., Magrath, I. T., Shad, A. T., Steinberg, S. M., Wexler, L. H., and Gress, R. E. (1995). N. Engl. J. Med. 332, 143–149. Marandin, A., Katz, A., Oksenhendler, E., Tulliez, M., Picard, F., Vainchenker, W., and Louache, F. (1996). Blood 88, 4568–4578. Margolick, J. B., Mun˜oz, A., Donnenberg, A. D., Park, L. P., Galai, N., Giorgi, J. V., O’Gorman, M. R. G., and Ferbas, J. (1995). Nature Med. 1, 674–680.

326

DAWN R. CLARK et al.

Markowitz, M., Saag, M. S., Powderly, W. G., Hurley, A. M., Hsu, A., Valdes, J. M., Henry, D., Sattler, F., La Marca, A., Leonard, J. M., and Ho, D. D. (1995). N. Engl. J. Med. 333, 1534–1540. McCune, J. M., Loftus, R., Schmidt, D. K., Carroll, P., Webster, D., Swor-Yim, L. B., Francis, I. R., Gross, B. H., and Grant, R. M. (1998). J. Clin. Invest. 101, 2301–2308. McHeyzer-Williams, M. G., and Davis, M. M. (1995). Science 268, 106–111. McLean, A., and Michie, C. (1993). Nature (London) 365, 301–301. McLean, A. R., and Michie, C. A. (1995). Proc. Natl. Acad. Sci. U.S.A. 92, 3707–3711. McLean, A. R., Rosado, M. M., Agenes, F., Vasconcellos, R., and Freitas, A. A. (1997). Proc. Natl. Acad. Sci. U.S.A. 94, 5792–5797. Meyaard, L., Otto, S. A., Jonker, R. R., Mijnster, M. J., Keet, R. P. M., and Miedema, F. (1992). Science 257, 217–219. Michie, C. A., McLean, A., Alcock, C., and Beverley, P. C. (1992). Science 360, 264–265. Mohri, H., Bonhoeffer, S., Monard, S., Perelson, A. S., and Ho, D. D. (1998). Science 279, 1223–1227. Moreland, L. W., Pratt, P. W., Bucy, R. P., Jackson, B. S., Feldman, J. W., and Koopman, W. J. (1994). Arthritis Rheum. 37, 834–838. Moreland, L. W., Bucy, R. P., and Koopman, W. J. (1995). N. Engl. J. Med. 332, 1651–1652. Moses, A. V., Williams, S., Heneveld, M. L., Strussenberg, J., Rarick, M., Loveless, M., Bagby, G., and Nelson, J. A. (1996). Blood 87, 919–925. Mosier, D. E., Gulizia, R. J., Baird, S. M., Wilson, D. B., Spector, D. H., and Spector, S. A. (1991). Science 251, 791–794. Mosier, D. E., Sprent, J., Tough, D., Dimitrov, D. S., and Martin, M. A. (1995). Nature (London) 375, 193–195. Pakker, N. G., Notermans, D. W., De Boer, R. J., Roos, M. T. L., Wolf, F., Hill, A., Leonard, J. M., Danner, S. A., Miedema, F., and Schellekens, P. T. A. (1998). Nature Med. 4, 208–214. Perelson, A. S., Neumann, A. U., Markowitz, M., Leonard, J. M., and Ho D. D. (1996). Science 271, 1582–1586. Rabin, R. L., Roederer, M., Maldonado, Y., Petru, A., and Herzenberg, L. A. (1995). J. Clin. Invest. 95, 2054–2060. Rep, M., Van Oosten, B. W., Roos, M. T. L., Ade`r, H. J., Polman, C. H., and Van Lier, R. (1997). J. Clin. Invest. 99, 2225–2231. Rocha, B., Dautigny, N., and Pereira, P. (1989). Eur. J. Immunol. 19, 905–911. Roederer, M. (1995). Nature Med. 1, 621–622. Roederer, M., Gregson Dubs, J., Anderson, M. T., Raju, P. A., Herzenberg, L. A., and Herzenberg, L. (1995). J. Clin. Invest. 95, 2061–2066. Rosenzweig, M., DeMaria, M. A., Harper, D. M., Friedrich, S., Jain, R. K., and Johnson, R. P. (1998). Proc. Natl. Acad. Sci. U.S.A. 95, 6388–6393. Rosok, B. I., Bostad, L., Voltersvik, P., Bjerknes, R., Olofsson, J., Asjo, B., and Brinchmann, J. E. (1996). AIDS 10, F35–F38 Sachsenberg, N., Perelson, A. S., Yerly, S., Schokmel, G. A., Leduc, D., Hirschel, B., and Perrin, L. (1998). J. Exp. Med. 187, 1295–1303. Scadden, D. T., Zon, L. I., and Groopman, J. E. (1989). Blood 74, 1455–1463. Schnittman, S. M., Denning, S. M., Greenhouse, J. J., Justement, J. S., Baseler, M., Kurtzberg, J., Haynes, B. F., and Fauci, A. S. (1990). Proc. Natl. Acad. Sci. U.S.A. 87, 7727–7731. Schwarting, R., Gerdes, J., Niehus, J., Jaeschke, L., and Stein, H. (1986). J. Immunol. Methods 90, 65–70. Sloand, E. M., Young, N. S., Sato, T., Kumar, P., Kim, S., Weichold, F. F., and Maciejewski, J. P. (1997). AIDS 11, 1547–1552.

T CELL DYNAMICS IN HIV-1 INFECTION

327

Sprent, J., and Tough, D. F. (1994). Science 265, 1395–1400. Sprent, J., and Tough, D. (1995). Nature (London) 375, 194–194. Stanley, S. K., Kessler, S. W., Justement, J. S., Schnittman, S. M., Greenhouse, J. J., Brown, C. C., Musongela, L., Musey, K., Kapita, B., and Fauci, A. S. (1992). J. Immunol. 149, 689–697. Steinberg, H. N., Crumpacker, C. S., and Chatis, P. A. (1991). J. Virol. 65, 1765–1769. Swain, S. L., Weinberg, A. D., and English, M. (1990). Science 265, 1395–1400. Tanchot, C., and Rocha, B. (1995). Eur. J. Immunol. 25, 2127–2136. Tanchot, C., and Rocha, B. (1997). J. Exp. Med. 186, 1099–1106. Tenner-Racz, K., Stellbrink, H.-J., Van Lunzen, J., Schneider, C., Jacobs, J.-P., Raschdorff, B., Groszschupff, G., Steinman, R. M., and Racz, P. (1998). J. Exp. Med 187, 949–959. Tissot, O., Viard, J.-P., Rabian, C., Ngo, H. S., Burgard, M., Rouzioux, C., and Penit, C. (1998). AIDS 12, 879–884. Tough, D. F., and Sprent, J. (1994). J. Exp. Med. 179, 1127–1135. Tsurusawa, M., Ito, M., Zha, Z., Kawai, S., Takasaki, Y., and Fujimoto, T. (1992). Leukemia 6, 669–674. Vigano, A., Clerici, M., Bricalli, D., Saresella, M., Difabio, S., Principi, N., and Vella, S. (1998). Immune reconstitution and role of the thymus during potent antiretroviral therapy in vertically HIV-1 infected children. 12th World AIDS Conference, Geneva, Switzerland (Abstract). Wei, X., Ghosh, S. K., Taylor, M. E., Johnson, V. A., Emini, E. A., Deutsch, P., Lifson, J. D., Bonhoeffer, S., Nowak, M. A., Hahn, B. H., Saag, M. S., and Shaw, G. M. (1995). Nature (London) 373, 117–122. Weichold, F. F., Zella, D., Barabitskaja, O., Maciejewski, J. P., Dunn, D. E., Sloand, E. M., and Young, N. S. (1998). Blood 91, 907–915. Weinberg, K., Annett, G., Kashyap, A., Lenarsky, C., Forman, S. J., and Parkman, R. (1995). Biol. Blood Marrow Transplant. 1, 18–23. Westermann, J., and Pabst, R. (1990). Immunol. Today 11, 406–410. Westermann, J., and Pabst, R. (1992). Clin. Investig. 70, 539–544. Wolthers, K. C., Wisman, G. B. A., Otto, S. A., De Roda Husman, A. M., Schaft, N., De Wolf, F., Goudsmit, J., Coutinho, R. A., Van der Zee, A. G. J., Meyaard, L., and Miedema, F. (1996). Science 274, 1543–1547. Wolthers, K. C., Noest, A. J., Otto, S. A., Miedema, F., and DeBoer, R. J. (1998). Normal telomere lengths in naive and memory CD4⫹ T cells in HIV-1 infection: A mathematical interpretation. (Submitted). Zauli, G., Re, M. C., Visani, G., Furlini, G., Mazza, P., Vignoli, M., and La Placa, M. (1992). J. Infect. Dis. 166, 710–716. Zauli, G., Vitale, M., Gibellini, M., and Capitani, S. (1996). J. Exp. Med. 183, 99–108. Zhang, Z. Q., Notermans, D. W., Sedgewick, G., Cavert, W., Wietgrefe, S., Zupancic, M., Gebhard, K., Henry, K., Boies, L., Chen, Z., Jenkins, M., Mills, R., McDade, H., Goodwin, G., Schuwirth, C. M., Danner, S. A., and Haase, A. T. (1998). Proc. Natl. Acad. Sci. U.S.A. 95, 1154–1159.

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 73

Bacterial CpG DNA Activates Immune Cells to Signal Infectious Danger HERMANN WAGNER Institute of Medical Microbiology, Immunology and Hygiene, Technical University of Munich, D-81675 Munich, Germany

I. Introduction

Following the discovery of the double helix, DNA was primarily viewed as a blueprint for genetic information but essentially as immunologically inert. The immune responses to DNA that were eventually realized appeared only to be harmful, because they often were associated with autoimmunity, for example, in systemic lupus erythematosis (SLE) (Emlen et al. 1986). This view has evolved, however—compelling evidence now shows that prokaryotic DNA, through pattern recognition, is sensed by cells of the immune system, thereby signaling ‘‘infectious danger’’ (Pisetsky, 1996; Krieg, 1996; Krieg et al. 1998c; Lipford et al. 1998). In contrast to the codon-based nucleotide triplets used to translate amino acid sequences, immune cells sense as a ‘‘danger code’’ the CpG dinucleotides, in a particular base sequence context, termed ‘‘CpG motifs.’’ The frequency of these CpG motifs differs in eukaryotic and prokaryotic DNA; furthermore, in mammalian DNA the cytosine is commonly methylated at the C-5 position. There has also been a change of our view of the interplay between the innate and the adaptive immune system. The latter was thought of as the ﬁnest accomplishment of vertebrate immunity because it creates, via somatic recombination, an abundance of antigen receptors able to evaluate all possible antigenic structures (Thompson, 1995; Gellert, 1997). In contrast, the former has at its disposal only a limited number of germline-encoded receptors, selected during evolution, to discriminate pathogens via pattern recognition receptors (PRRs) (Dempsey et al. 1996). The strength of this system, however, is that via pattern recognition it evaluates biologically relevant pathogens and subsequently communicates this evaluation to the adaptive immune system (Fearon and Locksley, 1996; Medzhitov and Janeway, Jr., 1996). A prototypic example is the PRR recognition of lipopolysaccharide (LPS) from gram-negative bacterial cell walls. LPS is recognized by the signaling-incompetent, glycosylphosphatidyl inositol-anchored plasma membrane glycoprotein, CD14 (Ulevitch and Tobias, 1995), which appears to transfer LPS to the signalingcompetent receptor toll-2 (Yang et al., 1998). LPS-mediated signaling in turn causes cytokine secretion by macrophages and dendritic cells (release 329

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

330

HERMANN WAGNER

of IL-12 and TNF-움) and by natural killer cells (release of IFN-웂), and up-regulation of expression of membrane proteins, including CD80 and CD86 on antigen-presenting cells (APCs) (Unanue, 1997; Lenschow et al. 1996). By interacting with CD28 on antigen-reactive T cells, activated APCs deliver via costimulatory molecules and the proper cytokines the second signal required for productive T cell activation (Matzinger, 1994). Viewed from this perspective, APCs sensing infectious danger are not only at the center of the communication network linking innate and acquired immunity but in fact control via cytokines and costimulatory molecules (i.e., the second signal) the magnitude as well as the quality (i.e., Th1 versus Th2 polarization) of emanating T cell responses (Mosmann and Coffman, 1989; Abbas et al., 1996). The term ‘‘second signal,’’ which was originally coined in the context of self–nonself discrimination (Bretscher and Cohn, 1970), now combines the sensing of infectious danger via innate immunity with the structural speciﬁties of acquired immunity (Fearon and Locksley, 1996). Bacterial products such as LPS or killed mycobacteria often function as natural adjuvants by enhancing acquired immune responses to admixed soluble antigens ( Janeway, Jr., 1989). APCs are likely to recognize via germ-line-encoded receptors dangerous bacterial products, and to become activated by this interaction. This activation enables them to present, in addition to antigenic peptides, a second signal to antigen-reactive T cells. Overreaction to bacterial products such as LPS represents a potentially sinister side of the immune system. The worst scenario of this reaction is the so-called septic shock (Parillo, 1993). Septic shock reﬂects multifactorial events, but overproduction by APCs of proinﬂammatory cytokines, including TNF-움, appears key (Rietschel and Wagner, 1996). Furthermore, LPS can induce the secretion of hemopoetically active growth factors and thus cause extramedullary hemopoesis (Staber and Metcalf, 1980). If one views LPS as prototypic for microbial products transmitting infectious danger to innate immunity, one might anticipate the immunobiology of bacterial CpG motifs if they were also to signal infectious danger. This may include adjuvanticity, polarization of Th1 responses, ampliﬁcation of cytotoxic T cell responses, and, in its extreme, septic shock. This review deﬁnes bacterial DNA and ‘‘mimicking’’ synthetic CpGoligonucleotides (ODNs) as the signal for infectious danger. Molecular, cellular, and in vivo studies will be discussed in the context of an increasing body of literature, examining the effects on cells of the immune system. The premise is that the unique adjuvant-like functions of CpG DNA can guide future investigations aimed at correlating the effects observed in animal model systems to the effects in humans.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

331

II. Bacterial DNA and CpG Motifs: History of Unraveling Immunobiology

The ability of bacterial products/toxins to augment immune reactions represents, in retrospect, the basis of Coley’s attempts to cause tumor regression in cancer patients (Starnes, 1992). In the 1970s, the apparently successful treatment of cancer in experimental model systems (Zbar et al., 1971, 1974) and in humans (Terry and Yamamura, 1979; Bast and Bast, 1976) with Mycobacterium bovis (bacillus of Calmette and Guerin, or BCG) attracted attention. Tokunaga and associates (Shimada et al., 1986; Tokunaga et al. 1992), among others, attempted to identify the active components within BCG. Their demonstration that a DNA-rich fraction represented the active component able to inhibit growth of syngeneic tumors in mice and guinea pigs (Shimada et al., 1985) in essence pioneered the ﬁeld of the immunobiology of bacterial DNA. The DNA-rich fraction, termed MY-1, not only caused tumor regression in vivo but also induced mouse spleen cells and human peripheral blood mononuclear cells (PBMCs) to produce interferons (IFN-움/웁, IFN-웂) and augment natural killer (NK) cell activity in vitro (Tokunaga et al., 1984). Of note, spleen cells from LPS-resistant C3H/HeJ mice also responded to bacterial DNA and some viral DNA, but not to vertebrate DNA. Because 45-mer singlestranded ODNs randomly selected from BCG cDNA encoding BCG proteins triggered in vitro IFNs and NK cell activity in mouse splenocytes, the minimal and essential sequence(s) responsible for IFN secretion and NK cell activation could be assessed. These experiments led to the conclusion that 45-mer ODNs containing a 6-base palindromic motif centered on a CpG dinucleotide were most active (Yamamoto et al. 1992 a,b; Tokunaga et al., 1984). In the autoimmune disease SLE, DNA represents a major autoantigen and a target for autoantibodies (Plescia et al., 1965). Although in mice, immunization with murine DNA usually fails to elicit signiﬁcant antibody responses, sera of healthy humans often contain high titers of antibacterial DNA antibodies, presumably as part of the host immune response to infections (Karounous et al., 1968). Indeed antigen DNA isolated from immune complexes in plasma of SLE patients appears to be of bacterial origin (Terada et al., 1991). Pisetzky and associates were ﬁrst to report that naked bacterial DNA, but not eukaryotic DNA, was mitogenic for murine B cells in vitro. In addition, bacterial DNA augmented polyclonal IgM responses (Messina et al., 1991). DNA-mediated mitogenicity appeared to be sequence dependent because methylation of synthetic poly(dC⭈dG)ODN caused complete loss of B cell mitogenicity (Messina et al., 1993). In retrospect, these are landmark observations. Gene inhibition by using antisense ODNs allows speciﬁc manipulation of gene expression (Wagner, 1994; Milligan et al., 1993). Inhibition of gene

332

HERMANN WAGNER

expression by antisense ODNs relies on their ability to bind complementary messenger RNA sequences and thus to prevent translation (Milligan et al., 1993), although a good antisense molecule is hard to ﬁnd (Branch, 1998). ODNs containing natural phosphodiester linkages ([P]ODN) become rapidly degraded, with a typical half-life of around 20 min (Fisher et al., 1993). Unless constant intracellular concentrations can be maintained by continuous delivery, [P]ODNs need to be made resistant to nucleases by chemical modiﬁcations, such as phosphorothioate linkages ([S]ODN) (Fisher et al., 1993). However, the [S]-ODN sequence independently may induce Sp1 transcription factor activity (Perez et al., 1994) or may provoke side effects due to charged polyanionic effects when used at high concentrations (Barker, Jr., et al., 1996; Ramasamy et al., 1996). Another problem with antisense ODNs is the low efﬁciency with which they cross cellular membranes. When added to the culture medium, they preferentially accumulate and become degraded in cytoplasmic granules (presumably endosomes/lysosomes) and may not reach the nucleus. Because of this low efﬁciency, permeabilization techniques such as cationic liposomes are often required for cytoplasmic translocation. In applying antisense technology, several investigators have noticed unexpected side effects. For example, antisense ODNs complementary to an endogeneous retroviral envelope (env) gene caused in vitro proliferation of murine splenocytes and up-regulation of lymphocyte MHC class II and immunoglobulin (Ig) membrane expression (Krieg et al., 1989). With the modiﬁcation of [S]ODNs, this sequence caused marked splenomegaly primarily due to B cell proliferation, increased class II MHC expression on B cells, and increased numbers of Igproducing cells (Mojcik et al., 1993). Pisetzky and associates had reported that bacterial DNA as well as certain synthetic ODNs were mitogenic for murine B cells. The mitogenic activity was thought to result from structural determinants of bacterial DNA that are rarely present in mouse DNA (Pisetsky et al., 1990). Using an antisense [S]ODN complementary to the UL13 open reading frame of herpes simplex virus (HSV), it was shown that this compound caused proliferation and antibody production by mouse spleen cells (Pisetsky and Reich, 1994). Immune cell stimulation was also observed with an antisense [S]ODN to the rev gene of HIV-1 (Branda et al., 1993). On intravenous (i.v.) application, mice developed splenomegaly, polyclonal hypergammaglobulin, and B cell proliferation. The anti-rev ODN was in vitro mitogenic for mouse B cells and for human PBMCs (Branda et al., 1996a). Thus, until 1994, independent lines of evidence had converged to suggest sequencedependent immunomodulating effects of bacterial and viral DNA.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

333

III. Binding and Cellular Uptake of ODNs

Because of their polyanionic nature, [P]ODNs as well as [S]ODNs cross lipophilic cell membranes only inefﬁciently (Stein and Cohen, 1988). Yet gene inhibition via antisense ODNs is operative (Stein and Cheng, 1993). Although the existence of cell membrane-bound DNA had been recognized earlier, a 30-kDa protein that binds double-stranded (ds) DNA was described subsequently (Bennett et al., 1985). dsDNA binding was heparin sensitive, but could not be inhibited by sequence-unrelated ODNs. An 80-kDa cell surface protein was identiﬁed that may mediate receptormediated endocytosis (Loke et al., 1989). ODNs of any length were competitive inhibitors for this binding. However, none of these putative DNAbinding proteins were further characterized. In HL60 cells, internalization of 10- to 15-mer DNA, via ﬂuid-phase endocytosis, has been suggested to lead, via PKC inhibition and stimulation of exocytosis, to a diminished intracellular ODN concentration (Stein et al., 1993). [S]ODNs have been shown to bind to a large number of heparin-binding proteins that in fact often are signaling-competent receptors for growth factors (Guvakova et al., 1995). Of note, BCG-derived MY-1 and synthetic palindromic [P]ODNs mimicking MY-1 only weakly stimulated IFN production by spleen cells depleted of murine macrophages. On the other hand, ligands for scavenger receptors on macrophages, such as dextran sulfate and 30-mer poly(G)ODNs [but not 30-mer poly(A)ODNs], competitively blocked IFN induction by the palindromic sequence AACGTT attached to a 30-mer poly(G) sequence (Kimura et al., 1994). This observation raised the question whether the poly(G) tail was targeting the immunostimulating palindrome to scavenger receptors. Mac1 (C11b/CD18; 움M/웁2), a heparinbinding integrin expressed on macrophages, NK cells, and polymorphonuclear leukocytes (PMNs), is likely to represent a prime receptor for both [P]ODNs and [S]ODNs (Benimetskaya et al., 1997b). Soluble ﬁbrinogen, its natural ligand, effectively competed for ODN binding, whereas antiMac1 mAbs blocked. Increase in the cell surface expression of Mac1 correlated with an increase in ODN internalization, which in turn triggered production of reactive oxygen species (Benimetskaya et al., 1997b). Overall, these data implicated ‘‘outside-in’’ signaling. IV. Sequence-Independent Effects of the Backbone

As already pointed out, DNA and synthetic ODNs are polyanions and nonspeciﬁc effects can be observed with [S]ODNs (Stein and Cheng, 1993). At very high concentrations (⬎100 애M ), [S]ODNs nonspeciﬁcally stimulated cytokine secretion from keratinocytes (Crooke et al., 1996) and

334

HERMANN WAGNER

were mitogenic for B cells, which paralled splenomegaly in vivo (Monteith et al., 1997). At low concentrations (2 애M ), [S]ODNs nonspeciﬁcally activated the transcription factor Sp1 in human and mouse lymphoid cell lines in vitro and in vivo (Perez et al., 1994). [S]ODNs can also inﬂuence other immune signal pathways. For example, [S]ODNs synergized at low concentrations with LPS to increase TNF-움 production in human PBMCs (Hartmann et al., 1996). The latter data might mean that antisense [S]ODN application during sepsis could be harmful. To beneﬁt from their nuclease resistance, [S]ODNs are often used in antisense strategies. However, they bind avidly to many proteins, forming complexes with dissociation constants one to three orders of magnitude lower than those of [P]ODNs. This may cause unexpected non-antisense side effects (Stein, 1995, 1996). For example, inhibition of leukemic cell proliferation by targeting the oncogenic fusion protein BCR/ABL in antisense therapies for chronic myeloid leukemia turned out to be due to cytotoxic ODN breakdown products (Vaerman et al., 1997). Interestingly, in a test for B cell mitogenicity [S]ODNs were 2 log more potent than [P]ODNs of the same sequence (Krieg et al., 1996), whereas in another study [S] modiﬁcations inactivated immunostimulating CpG[P]ODN sequences (Lapatschek et al., 1998). Accordingly, direct activation of macrophages and B lymphocytes by CpG-ODNs derived from acutely pathogenic simian immunodeﬁciency virus was observed only with [P]ODN sequences. V. CpG DNA Sequence-Speciﬁc Effects on B Cells

A landmark discovery galvanizing interest in the immunobiology of bacterial DNA dates back to 1995. To deﬁne the base sequences conferring mitogenicity for murine B cells (Messina et al., 1993; Krieg et al., 1989; Mojcik et al., 1993; Pisetsky et al., 1990; Branda et al., 1993), Krieg and associates iteratively changed the sequences of immunostimulating singlestranded (ss) CpG motifs containing [S]ODNs 8–20 bases in length and tested for their ability to activate murine B cells in vitro (Krieg et al., 1995). It was found that optimal B cell activation required a consensus DNA motif in which an unmethylated CpG dinucleotide is ﬂanked by two 5⬘ purines and two 3⬘ pyrimidines. Sequence speciﬁcity was exquisite because CG dinucleotide inversion to GC or methylation of cytosine abolished B cell mitogenicity. These data were important because they suggested a link between immune defense based on the recognition of CGrich microbial DNA and of CpG suppression (Bird et al., 1987) in vertebrate DNA. Additionally, CpG motifs exerting B cell mitogenicity were virtually identical to the ones that had previously been described to trigger IFN production and NK cell activation (Yamamoto et al., 1992a), as if the

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

335

hexamer 5⬘Pu-Pu-CpG-Pyr-Pyr3⬘ (Krieg et al., 1995) were active toward distinct subsets of immune cells. Although the CpG motif caused B cell proliferation as well as Ig secretion (Krieg et al., 1995), an antisense ODN (containing this motif ) targeting sequences in Ig2b was mitogenic for B cells, but inhibited Ig secretion (Tanaka et al., 1992; Mojcik et al., 1993; McIntyre et al., 1993). As will be discussed below, there may be several mechanisms through which CpG-ODNs initiate cell signaling, such a signal-competent cell surface receptors, or chloroquine-sensitive endosomal pathways. Interestingly, at low concentrations of mitogenic CpG-ODNs, B cell activation was strongly augmented by B cell receptor (BCR) crosslinking (Krieg et al., 1995), as if CpG motifs had the potential to costimulate antigen-reactive B cells by providing the second signal. Using ELI-spot assays to probe for cytokine-secreting cells triggered by CpG[S]ODNs, immune cells were shown to produce IL-6, IL-12, and IFN-웂 in vitro and in vivo (Klinman et al., 1996). Curiously, these studies did not attribute IL-12 secretion to monocytes. CpG[S]ODNinduced IL-6 production by B cells appeared to be initiated by reactive oxygen-dependent pathways, and B cell IgM secretion was found to be dependent on secreted IL-6 (Yi et al., 1996c, 1998b). CpG[S]ODNs induced NK cells to produce IFN-웂 (Ballas et al., 1996), although later studies proved that IFN-웂 production by NK cells was a consequence of IL-12 production by macrophages (Chace et al., 1997). Furthermore, IFN-웂 promoted IL-6 and IgM secretion by B cells in response to CpG motifs (Yi et al., 1996a). Rather unexpectedly, DNA released from dying Drosophila cells expressing MHC class I molecules was found to activate and thus to up-regulate costimulatory molecules on B cells, and once activated, CD80- and CD86-positive B cells provided bystander costimulation for the primary activation of peptide-reactive cytotoxic T cells. Again, CpG methylation abolished the mitogenic activity of Drosophila DNA (Sun et al., 1996). The strong activation effect of CpG[S]ODNs toward B cells also affected B cell apoptosis. Similar to LPS and CD4OL stimulation (Tsubata et al., 1993), CpG[S]ODNs protected cells of the WEHI-231 B cell line from anti-IgM-mediated apoptosis (Yi et al., 1996b). Protection was associated with down-regulation of c-myc and expression and up-regulation of bcl-2 and bcl-x1 mRNA. Subsequently, these studies have been extended to mature peripheral B cells (Yi et al., 1998a). Although a CpG[S]ODNdriven B cell cycle entry inhibited spontaneous apoptosis, partial apoptosis protection was also observed in G0-arrested B cells, as if CpG-ODNs trigger resting B cells to enter G1-inhibiting apoptosis and allow B cells to proceed to S phase. Viewed from the two-signal:death/survival model (Green and Scott 1994), these studies suggested that CpG-ODNs can

336

HERMANN WAGNER

inﬂuence both signals at once. CpG-ODN-driven modulation of I␬B움 and I␬B웁 may be important because it led to sustained activation of NF␬B, which in turn activated the protooncogene c-myc (Yi and Krieg, 1998a). VI. CpG DNA Sequence-Speciﬁc Effects on APCs

Apart from their mitogenic effects on B cells, CpG-ODN sequences speciﬁcally activate antigen-presenting cells, the sentinels bridging innate and adaptive immunity (Bendelac and Fearon, 1997; Medzhitov and Janeway, Jr., 1997). Uptake of bacterial plasmid DNA as well as palindromic ACCGGT-ODN (Yamamoto et al., 1992a) by macrophages was associated with NF␬B activation and up-regulation of TNF-움 mRNA (Stacey et al., 1996). Of note, DNA prepared from gram-positive or gram-negative bacteria was found to trigger NF␬B activation, accumulation of TNF-움 mRNA, and subsequent fulminant TNF-움 production from macrophages, both in vitro and in vivo. These events could be mimicked by unmethylated CpG[S]ODN (Sparwasser et al., 1997a). Mice challenged intraperitoneally (i.p.) with active CpG[S]ODN had transiently high serum levels of TNF-움, IL-6, and IL-12, similar in magnitude to those provoked by LPS. As a consequence D-galactosamine (D-GaIN)-sensitized mice succumbed to acute cytokine release syndrome (Sparwasser et al., 1997b). Macrophagederived TNF-움 was shown to cause, via TNF receptor p55, acute apoptosis of liver cells (Sparwasser et al., 1997b). These studies revealed that CpG[S]ODN sequences speciﬁcally activate macrophages to express a full complement of costimulatory molecules, including CD40, CD80, and CD86, and to produce an array of proinﬂammatory cytokines, including IL-12 and TNF-움. On the one hand, they showed that CpG[S]ODNdriven secretion of high levels of TNF-움 may be harmful in that it can promote toxic shock. On the other hand, they were ﬁrst to indicate that CpG-driven induction of IL-12 may be beneﬁcial (Lipford et al., 1997a,b), promoting Th1-polarized T cell responses. Indeed, sequence modiﬁcations of CpG[S]ODN on the basis of information obtained in gene sequence database screens for canonical 5⬘Pu-Pu-CpG-Pyr-Pyr3⬘ motif-related sequences arrived at a CpG motif that effectively induced IL-12 but not TNF-움 in macrophages, both in vitro and in vivo (Lipford et al., 1997b). This CpG motif may be prototypic for useful CpG-DNA motifs that do not cause harmful side effects. Cyclosporin A (CsA)-mediated immune suppressive effects have been attributed to binding of the compound to cyclophilin. This complex inhibited the phosphatase calcineurin, thus blocking downstream activation of NF-AT-driven transcription of IL-2 in T cells (Clipstone and Crabtree,

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

337

1992). CsA enhanced CpG[s]ODN-triggered IL-12 production by suppressing secretion of IL-10 in macrophages and B cells (Redford et al., 1998; Anitescu et al., 1997). These results suggested that IL-10 acts as negative feedback regulator of CpG[S]ODN-induced IL-12 production. Dendritic cells (DCs) guide and control via their antigen presentation machinery, via expression of costimulatory molecules, and by cytokine secretion antigen-speciﬁc T cell activation (Steinman et al., 1997; Banchereau and Steinman, 1998). The DCs that are ubiquitously distributed throughout nonlymphoid tissue are immature. At this differentiation stage DCs are efﬁcient in cellular uptake and processing of soluble and particulate antigens, including apoptotic cells. Yet they are poor in MHC class I and II expression, cytokine production, and costimulatory molecules (Cella et al., 1997; Schuler et al., 1997; Albert et al., 1998). On maturation, DCs translocate to T cell areas of secondary lymphoid tissues, where, as ‘‘professional’’ APCs, they activate naive antigen-reactive T cells (Cella et al., 1997; Schuler et al., 1997). Microbial stimuli such as LPS (De Smedt et al., 1996; Sousa et al., 1997) and bacterial DNA, as well as mimicking CpG[S]ODN (Sparwasser et al., 1998; Jakob et al., 1998), were particularly efﬁcient in triggering DC maturation. For immature DCs grown from bone marrow cells in GM-CSFconditioned medium, bacterial DNA and CpG[S]DNA turned out to be as effective as LPS in stimulating maturation of DCs, as determined by upregulation of MHC class II, CD40, and CD86. In parallel to phenotypical maturation, CpG[S]ODNs caused production of intermediate levels of IL-6 and TNF-움, but high concentrations of IL-12. On maturation, DCs were also able to act as ‘‘professional’’ APCs in mixed lymphocyte cultures (Sparwasser et al., 1998). CpG[S]ODNs exerted similar effects on Langerhans cell (LC)-like murine fetal skin-derived DCs in vitro ( Jakob et al., 1998). In vivo challenge with CpG[S]ODNs into murine dermis led to enhanced expression of MHC class II, CD86, and IL-12 by LCs. Taken together, these data suggest that IFN-웂-independent up-regulation of IL-12 production in splenic DCs and subsequent DC translocation to T cell areas within central lymphoid tissue may not only be a capacity of Toxoplasma gondii extracts (Sousa et al., 1997) but also of bacterial DNA and CpG[S]ODN. As a consequence, IL-12-producing DCs may act as key regulators for initiation of Th1 differentiation/expansion. After subcutaneous challenge, CpG[S]ODNs were found to induce lymphadenopathy in draining lymph nodes (LNs) (Sparwasser et al., 1999b). Within the ﬁrst 24 hr, DCs containing intracytoplasmatic IL-12 could be identiﬁed within T cell areas of draining LNs. Surprisingly, LNs subsequently increased in cellularity 20- to 50-fold with a peak on days 8–10. A dramatic inﬂux of T cells, B cells, and in particular DCs was

338

HERMANN WAGNER

observed due to as yet ill-deﬁned chemokine activities. In addition, elevated serum levels of IL-12 persisted. Interestingly, in this time window (up to 14 days) the affected draining LN remained in a Th1-polarized state. The sustained Th1 polarization of T cells within draining LNs may be associated with, and perhaps explain, a sustained state of local resistance to infection, for example, infection with Listeria monocytogenes (Krieg et al., 1998a), or infection with Leishmania major in susceptible BALB/c mice (Lipford et al., 1999b). VII. CpG DNA Effects on T Cells

In contrast to B cells and APCs, bacterial DNA and CpG[S]ODNs did not directly activate T cells. However, CpG[S]ODNs costimulated antigenreactive T cells that had received a stimulus. On cross-linking of T cell receptors (TCRs) via plastic-bound anti-CD3 mAb, CpG[S]ODNs costimulated T cells to produce IL-2, to express IL-2 receptors, and to proliferate and differentiate into CTLs (Bendings et al., 1999). CpG[S]ODNmediated costimulation of T cells was, at least in part, operative in T cells from CD28⫺/⫺ mutant mice, as if CpG[S]ODN-substituted CD28 mediated signal 2 in T cells activated by TCR cross-linking (signal 1). Of note, CpG[S]ODN-mediated costimulation of T cells activated via TCR occupation violated several rules deﬁned for CpG[S]ODN-mediated stimulation of APCs and B cells. First, T cell costimulation turned out to be chloroquine resistant (see below). Second, neither cytosine methylation, CG inversion to GC, nor ﬂanking canonical CpG motifs with [P]poly(G)stretches affected T cell costimulation. However, all of these modiﬁcations abrogated direct activation of APCs (Lipford et al., 1999a). These data imply that T cell costimulation is mediated by signal pathways distinct from those characterizing CpG[S]ODN-mediated signaling via the chloroquine-sensitive endosomal pathway (see below). Furthermore [S]poly(G) but not [P]poly(G) stretches added onto CpG[S]ODNs abolished their ability to activate APCs in vitro yet did not affect their adjuvanticity in CTL responses to ovalbumin (Lipford et al., 1999a). [S]poly(G)-extended CpG-ODNs may therefore be useful to target CpG-ODNs to T cells. This may be advantageous to avoid potentially harmful side effects such as induction in DC/macrophages of toxic amounts of TNF-움. Although the two-signal concept (Bretscher and Cohn, 1970; Lafferty and Woolnough, 1977) is probably a simplistic view of primary T cell activation, the ability of CpG[S]ODN to costimulate T cells may have implications. In the case of tumor cells expressing tumor antigens but lacking costimulatory molecules, T cell costimulation via CpG[S]ODNs

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

339

may be useful to trigger productive T cell activation. On the other hand, use of CpG[S]ODNs as adjuvant may be harmful by facilitating activation of autoreactive cells (Segal et al., 1997). LPS known to be mitogenic for murine B cells also activates murine T cells in the apparent absence of TCR ligation (Tough et al., 1996), possibly due to bystander effects of cytokines such as IFN-움/웁. Poly(I : C), a known inducer of IFN-움/웁 (Sen and Lengyel, 1992), as well as puriﬁed IFN-움/ 웁, mimicked this effect (Tough et al., 1997). IFN-움/웁 also induced via APC-delivered IL-15 bystander proliferation of memory-type T cells, but not of naive T cells (Zhang et al., 1998). Similar to LPS, CpG[S]ODN also activated T cells in the absence of TCR ligation both in vitro, and in vivo (G. B. Lipford, unpublished data). Within 20 hr subcutaneous application of CpG[S]ODNs caused up-regulation of CD69 and CD86 on T cells within the draining LNs, but not that of IL-2 receptors (IL-2Rs). Puriﬁed T cells up-regulated CD69 in response to recombinant IFN-움/웁. Therefore, it is likely that this type of CpG-ODN-mediated T cell activation represents a TCR-independent direct effect of IFN-움/웁. In view of these data it may be interesting to reanalyze the B cell mitogenicity (Krieg et al., 1995) of CpG[S]ODNs. VIII. CpG[S]ODN Effects on NK Cells

Lytic activity of NK cells and secretion of IFN-웂 were rapidly induced by CpG[S]ODNs in vivo (Shimada et al., 1986; Tokunaga et al., 1984; Cowdery et al., 1996). In vitro CpG[S]ODN-mediated activation of NK cells appeared to be indirect and required costimulation by cytokines from CpG[S]ODN-activated APCs, such as IL-12 and TNF-움 (Halpern et al., 1996). CpG[S]ODNs failed to trigger directly IFN-웂 production by NK cells, yet it synergized with low concentrations of IL-12 to induce secretion of high levels of IFN-웂 (Chace et al., 1997). A role of B and T cells in NK responses to CpG-ODNs was not apparent (Ballas et al., 1996). Whether CpG[S]ODNs trigger NK-like murine NK1.1 T cells has not yet been studied. NK1.1 T cells recognize the products of the conserved family of CD1 genes such as lipid antigens (Beckman et al., 1994) and have the unique potential to secrete very rapidly high amounts of Th1- and Th2type cytokines (Bendelac et al., 1997). Because CpG[S]ODNs trigger DCs to produce rapidly large amounts of IL-12 (Sparwasser et al., 1998; Jakob et al., 1998) and because IL-12 rapidly activates NK1.1 T cells to display lytic activity (Hashimoto et al., 1995) and to produce IFN-웂 (Cella et al., 1997), a search for effects of CpG[S]ODN on NK1.1 T cells may be rewarding.

340

HERMANN WAGNER

IX. CpG Motifs Affect Plasmid DNA Biology in Gene Vaccination

Immunization against microbial/viral infections and cancer represents an attractive alternative to chemotherapy. Advances in recombinant DNA technology have led to a new generation of recombinant subunit and synthetic peptide vaccines. However, without proper adjuvants their immunogenicity appears rather poor (Gluck et al., 1992). On the other hand, the landmark observation that immunization with naked plasmid DNA leads to strong and persistent cellular and humoral immune responses toward the antigen encoded had a great impact on vaccination strategies (Fynan et al., 1993; Tang et al., 1992). Surprisingly, vaccination of mice with naked DNA preferentially induced Th1-polarized responses, whereas vaccination with protein antigens predominantly induced Th2 responses (Raz et al., 1994, 1996; Manickan et al., 1995; Donnelly et al., 1997). Furthermore, expression vectors for 웁-galactosidase (웁-Gal) containing the bacterial ampicillin resistance gene (ampR) induced stronger anti-웁-Gal responses compared to expression vectors containing the kanamycin resistance gene (KanR) (Sato et al., 1996). It turned out that the ability to enhance immune responses was not related to the amount of antigen produced, but to the number of palindromic hexamer 5⬘AACGTT3⬘ CpG motifs in the ampR-containing plasmid (Sato et al., 1996). Methylation of bacterial plasmids used for DNA vaccination reduced immunogenicity (Klinman et al., 1997). Yet coinjection of plasmid vaccine with CpG[S]ODNs augmented immune responses in one study (Klinman et al., 1997) but not in another analysis (Weeratna et al., 1998). CpG motifs also augmented efﬁcacy of dengue 2 DNA vaccines (Porter et al., 1998). Because plasmid DNA-based immunization is thought to be initiated, in vivo, by transfected DCs (Condon et al., 1996; Manickan et al., 1997), the ability of CpG motifs to cause DC maturation as well as acute cytokine secretion (Sparwasser et al., 1998; Jakob et al., 1998) may explain the intriguing adjuvanticity of plasmid DNA. In summary, these results suggest that DNA vaccine consists operationally of two components. First, the plasmid insert encoding the protein antigen provides speciﬁcity by being transcribed and translated to produce at least a small amount of protein. Second, CpG motifs in the plasmid DNA backbone directly induce maturation and activation of transfected DCs. The above data could also explain the observation that gene-gun administration of DNA vaccines preferentially triggered Th2 responses, whereas subcutaneous or intramuscular (i.m.) application of DNA vaccines preferentially caused Th1 responses (Raz et al., 1996). In general, s.c. or i.m. challenge protocols use larger amounts of plasmid DNA, which may provide more CpG motifs and thus stronger signals, biasing for Th1 responses. Immunostimulating CpG motifs within

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

341

plasmids may be also biologically active in humans. When complexed with cationic lipids to enhance cellular uptake, CpG motifs within plasmids encoding the cystic ﬁbrosis transmembrane receptor could be responsible for side effects seen in clinical trials (Freimark et al., 1998). X. Sequence-Speciﬁc Effects of Poly(G) Motifs

Due to their polyanionic properties, ODNs interacted in a sequencenonspeciﬁc manner with the integrin Mac1 (Benimetskaya et al., 1997b). Of note, scavenger receptors on macrophages exhibited unusually broad, but circumscribed, polyanionic ligand-binding speciﬁcities in that G-quartet-stabilized four-stranded helixes appeared as a structural prerequisite for ODN binding (Pearson et al., 1993). Such G-quartet structures may also be involved in the telomeric replication machinery (Williamson et al., 1989). Macrophage scavenger receptors are trimeric integral membrane glycoproteins with remarkably well-conserved cationic cysteine-rich protein domains, and they bind LPS (Krieger, 1992). Targeting of poly[P]GODNs or CpG[P]ODNs extended with poly[P] G stretches to scavenger receptors not only increased cellular translocation (Kimura et al., 1994), but in another study triggered DNA synthesis (Sasaki et al., 1996). The efﬁcacy of antisense ODNs targeted to c-myc (Saijo et al., 1997), c-myb (Castier et al., 1998), RelA (Benimetskaya et al., 1997a), and ras GAP (White et al., 1996) appeared also to be independent of antisense mechanisms, but dependent on the presence of G-rich blocks in the ODN sequence. Poly[S]G-ODN, but not its variant poly[P]G-ODN, blocked IFN-웂 production in splenocytes by the mitogen concanavalin A (ConA) (Halpern and Pisetsky, 1995), and binding of IFN-웂 to its receptor (Ramanathan et al., 1994; Hertl et al., 1995). Poly[S]G regions as well as poly[P]G regions within ODNs blocked, via competitive inhibition, cellular uptake of CpG[S]ODN (Ballas et al., 1996; Hacker et al., 1998). However, extension of CpG[S]ODN with poly[S]G (⬎4) abolished immunobiology in that these ODN variants failed to activate APCs via the endosomal pathway (Lipford et al., 1999a). Of note, such G-rich CpG[S]ODN variants were still effective in costimulating T cells, probably due to high-afﬁnity binding to cell surface receptors. CpG-ODNs containing poly(G[S]) stretches (⬎4) sequence speciﬁcally caused outgrowth of macrophage precursors from bone marrow cultures (Lang et al., 1999). Electrophoretic mobility shift assays (EMSAs) revealed speciﬁc binding of the G-rich [S]ODNs to nonnuclear proteins. The ability of G-rich [S]ODNs to bind selectively cellular proteins (as assayed in EMSAs) correlated with their ability to cause outgrowth of macrophage precursors in vitro (Lang et al., 1999).

342

HERMANN WAGNER

Poly[S]G-ODNs failed to induce TNF-움 production, but inhibited TNF-움 production induced by canonical CpG-ODNs, in macrophages (K. Heeg, personal communication). Such poly[S]G-ODNs competitively blocked cellular uptake (translocation) of bacterial DNA or CpG-ODNs in macrophages, thus curtailing downstream signaling. In vivo, these poly[S]G-ODNs prevented the lethal shock triggered by bacterial DNA in a dose-dependent fashion, yet did not interfere with superantigen or LPS-induced lethal shock. In vivo, poly[S]G-ODNs acted as adjuvants for the generation of ovalbumin-speciﬁc cytotoxic T cells, although they were inert toward DCs (Lipford et al., 1999a). XI. Immunosuppressive CpG DNA Motifs

Not only are they suppressed in frequency, the remaining CpG motifs within vertebrate DNA lack immunostimulatory effects. This has been attributed, at least in part, to the high degree of methylation of cytosine at position C-5 (Bird et al., 1987). However, vertebrate DNA was still nonstimulatory when completely demethylated (Sun et al., 1997). This puzzling observation might ﬁnd its explanation in the recently demonstrated existence of immunosuppressive CpG motifs, as deﬁned in adenoviral DNA (Krieg et al., 1998b). Adenovirus serotypes 2, 5, and 12 exhibit no CpG suppression (Karlin et al., 1994). Of these adenoviruses, type 12 DNA activated murine and human immune cells, whereas types 2 and 5 adenoviral DNA did not (Krieg et al., 1998b). Furthermore, types 2 and 5 adenoviral DNA suppressed in trans immunostimulatory activity of bacterial genomic DNA. Genomic analysis of hexamer CpG motifs revealed that types 2 and 5 adenoviral DNA contained a 15- to 30-fold excess of clusters of CG dinucleotides, or of C on the 5⬘ or G on the 3⬘ side of CG dinucleotiodes, but low frequencies of canonical CpG motifs. Indeed, CpG motifs consisting of various combinations of CGCG, CCG, or CGG were found to be immunosuppressive. Interestingly, vertebrate DNA has an about 5-fold excess of such suppressive CpG motifs (Krieg et al., 1998b). XII. CpG-ODN-Mediated Signaling

Even though it is not completely understood how CpG-ODN sequences speciﬁcally mediate cell activation, some rate-limiting steps in signaling have been unraveled in APCs (Hacker et al., 1998) and in B cells (Yi and Krieg, 1998b). The rate-limiting steps include cellular binding of CpGODN, translocation into early endosomes, sensitivity to lysosomotropic compounds, and activation of the mitogen-activated protein kinase (MAPK)

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

343

pathways, known to control transcriptional activities of AP-1 and of NF␬B (Fig. 1). A. CELLULAR UPTAKE OF CpG-ODN ODN bound to Mac1 (CD11b/CD18) triggers the generation of reactive oxygen species (ROS) in TNF-움-treated human PMNs (Benimetskaya et al., 1997b), implying outside-in signaling via Mac1. However, binding to Mac1 appears not to be CpG-ODN sequence speciﬁc. An antisense ODN to rev of HIV, which was mitogenic for murine splenocytes (Branda et al.,

FIG. 1. A postulated mechanism for bacterial DNA or CpG-ODN-mediated signal transduction in DCs. (a) Sequence-nonspeciﬁc binding of bacterial DNA or CpG-ODNs to cellular surface receptors is followed by translocation into early endosomes (b). Acidiﬁcation and endosomal maturation are essential because chloroquine blocks downstream signaling. Translation of the CpG-ODN sequence into signaling via a postulated CpG-ODN-speciﬁc receptor (c) must occur upstream of the stress kinase pathway. Strees kinase pathway activation (SAPK/p38) yields transcriptional activity of AP-1 (d) and NF␬B (e). Nuclear translocation (f ) of the active transcription factors and binding to their respective DNA binding sites (g) promote activation and cytokine gene transcription. Modiﬁed from Trends Microbiol, 6(12), Lipford, G. B., Heeg, K., and Wagner, H., Bacterial DNA as immune cell activator, 496–500, copyright 1998, with permission from Elsevier Science.

344

HERMANN WAGNER

1993), exhibited marked mitogenicity to human B cells when coupled covalently to sepharose beads (Liang et al., 1996). On the other hand, canonical CpG[S]ODNs linked to a solid support were not mitogenic for murine B cells (Krieg et al., 1995). Of note, the ODN antisense to rev of HIV [termed DSP30 by Liang et al. (1996)] lacked the canonical 5⬘PuPu-CpG-Pyr-Pyr3⬘ motif that stimulated murine B cells (Krieg et al., 1995), as if distinct sequence requirements exist in mouse and human DNA. Alternatively, and perhaps also dependent on ODN sequences, distinct signal pathways may be initiated. For example, lipofection of 22-mer [P]ODNs displaying an AACGTT palindrome CpG motif strongly enhanced IFN-움/웁 production of murine splenocytes, as if translocation were rate limiting (Yamamoto et al., 1994b). Subsequent studies revealed that the minimal essential structure required for IFN-움/웁 induction was a hexamer palindromic CG motif, provided the CpG motif was translocated by cationic liposomes (Sonehara et al., 1996). These data suggest that base sequences neighboring the ‘‘essential’’ hexamer CpG motif were critical in cell surface receptor-mediated ODN translocation. For example, poly[P]G stretches added to hexamer CG motifs targeted the CpG[P]ODN to scavenger receptors and enhanced translocation (Kimura et al., 1994). As is the case for antisense ODNs (Tonkinson and Stein, 1994), CpG[S]ODNs were taken up (translocated) by murine APCs into acidic vesicles of the endosomal–lysosomal compartment (Hacker et al., 1998). Uptake was not sequence speciﬁc because non-CpG-ODNs became comparably well translocated, and non-CpG-ODNs competitively blocked uptake of CpG-ODNs (Hacker et al., 1998). Poly[S]G-ODNs and to a lesser extent poly[P]G-ODNs (see also Kimura et al., 1994) competed best, as if poly[S]G sequences (within ODNs) provided high-afﬁnity binding to cell surface receptors (Hacker et al., 1998). Translocation into endosomes represented a prerequisite for CpG-ODNmediated activation of APCs (Hacker et al., 1998). However, poly[S]GODN-mediated T cell costimulation did not require cellular uptake (translocation) (Lipford et al., 1999a), as was the case for activation of human B cells by ODN DSP30 (Liang et al., 1996). These observations suggest that CpG-ODNs displaying canonical 5⬘Pu-Pu-CpG-Pyr-Pyr3⬘ hexamer motifs activate immune cells after translocation to endosomal compartments. However, CpG-ODN variants displaying poly[S]G sequences (⬎4) may costimulate via signal-competent cell surface receptors. B. ANTAGONISTIC EFFECT OF LYSOSOMOTROPIC COMPOUNDS To activate APCs, canonical CpG-ODNs required endosomal translocation (Hacker et al., 1998). Lysosomotropic compounds known to inhibit endosomal maturation, such as chloroquine, baﬁlomycin A, or monensin

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

345

(Yoshimori et al., 1991; Fox, 1993), completely inhibited CpG-ODN-driven antiapoptotic effects in WEHI-231 B cells (Macfarlane and Manzel, 1998; Macfarlane et al., 1997) as well as CpG-ODN-driven activation of APCs (Hacker et al., 1998). The lysosomotropic compounds did not affect cellular uptake of CpG-ODNs (Hacker et al., 1998), but completely destroyed CpG-ODN-driven downstream events, such as activation of stress kinases and of the transcription factors such as AP-1 (Hacker et al., 1998). Similar ﬁndings were described for B cells (Yi and Krieg, 1998b). C. ACTIVATION OF THE STRESS KINASE PATHWAY Activation of APCs (Hacker et al., 1998) or B cells (Yi and Krieg, 1998b) was mediated, at least in part, by rapid induction of mitogen-activated protein kinases (the ‘‘stress pathway’’) (Karin, 1995; Kyriakis et al., 1994). CpG-DNA triggered severalfold the kinase activity of JNK, as did LPS. Antiphosphoantibodies speciﬁc for the phosphorylation site of JNKK1 (Hibi et al., 1993) identiﬁed this kinase as an upstream target of CpG[S]ODN-initiated signaling. p38, another MAPK originally identiﬁed as a kinase activated by LPS (Han et al., 1994), is an additional immediateearly target of CpG[S]ODN. JNK activation and p38 activation were associated with subsequent phosphorylation of c-Jun and ATF2, components of the transcription factor AP-1 (Ray et al., 1989). Transient and stable transfection of macrophages with reporter constructs under control of AP-1 binding sites revealed that CpG[S]ODNs induced sequence-speciﬁc transcriptional activity of AP-1, as did LPS (Hacker et al., 1998). Furthermore, CpG[S]ODN-triggered IL-12 and TNF-움 secretion was suppressed by selective inhibition of p38 kinase activity, as if p38 controlled cytokine release (Hacker et al., 1998). The lysosomotropic compounds baﬁlomycin A or chloroquine effectively blocked cytokine release, TNF-움 and IL-12p40 promoter activities, and immediate-early activation of the stress kinase p38 (Hacker et al., 1998). These data deﬁned endosomal maturation as an essential step for ‘‘translation’’ of CpG[S]ODN sequences into cellular signaling. Studies of CpG[S]ODN-activated signal pathways in B cells led to similar conclusions (Yi and Krieg, 1998b). CpG[S]ODNs activated JNK and p38 kinase but not the extracellular receptor kinase (ERK) pathway. This contrasts to APCs, in which CpG-ODNs triggered the ERK pathway (H. Hacker, personal communication). Downstream effects of CpG-ODNs in B cells included phosphorylation of ATF2 and c-Jun and activation of AP1. Again, selective inhibition of p38 suppressed cytokine production. Chloroquine blocked JNK and p38 kinase activation but not activation via CD40 cross-linking (Yi and Krieg, 1998b). Activation of NF␬B in CpG[S]ODN-treated B cells was preceded by and may be dependent on

346

HERMANN WAGNER

the generation of ROS. However, the lysosomotropic compound chloroquine blocked CpG DNA-mediated activation of NF␬B, but not when triggered by LPS, anti-CD30, or anti-IgM. Because MEKK1, an upstream kinase of the stress kinase pathway in HeLa cells, liberated NF␬B via activation of the I␬B움 kinase complex (Lee et al., 1997), chloroquine may block all of the CpG[S]ODN-triggered immediate-early signal steps identiﬁed so far in B cells. D. CHLOROQUINE-SENSITIVE AND -RESISTANT SIGNALING Chloroquine or baﬁlomycin A sensitivity of CpG DNA-mediated cell activation (Hacker et al., 1998; Yi and Krieg, 1998b) may be viewed as a strict requirement for endosomal acidiﬁcation. Such pH changes could either trigger dissociation of CpG DNA from sequence-nonspeciﬁc cell surface receptors mediating translocation, or enable binding of CpG DNA to sequence-speciﬁc endosomal DNA-binding receptors. Alternatively, a less well-deﬁned molecular mechanism downstream of acidiﬁcation might be responsible. A case of chloroquine/baﬁlomycin resistance is the CpG-ODN-mediated costimulation of T cells partially activated by TCR cross-linking. T cell costimulation using either canonical CpG-ODNs or poly[S]G-ODNs was found to be insensitive to lysosomotropic compounds (Lipford et al., 1999a). XIII. Sensing of Pathogen DNA: Evolutionary Vestige of Foreign DNA

CpG dinucleotides are rare in vertebrate genomes (CpG suppression) and constitute only about 25% of the level expected from base composition (Bird, 1980; Jones et al., 1992). CpG suppression may be due to the mutability associated with cytosine methylation (Coulondre et al., 1978). Cytosine residues in the remaining CpG sequences become postsynthetically methylated. CpG methylation is involved in long-term silencing of genes during mammalian development (Li et al., 1993). The methyl-CpGbinding proteins MeCP1/2 (Meehan et al., 1989) speciﬁcally interact with methylated DNA and mediate transcriptional repression (Nan et al., 1997). Low-density genomic methylation (about 98%) has been thought of as silencing tissue-speciﬁc genes as well as background transcriptional activities (Prestridge and Burks, 1993). High-density methylation under certain conditions is seen in CpG-rich islands and is thought of as leakproof inactivation of, for example, X-linked genes (Bird, 1993). However, the majority of CpG-rich islands in the promoters of mammalian genes are spared from methylation, presumably because they represent initiation sites for both transcription and DNA replication (Delgado et al., 1998). The relative use of CpG dinucleotides is normal in almost all bacteria,

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

347

fungi, and vertebrate and large eukaryotic viruses, but is suppressed in vertebrate genomes and in virtually all small eukaryotic viruses (Burge et al., 1992; Karlin et al., 1994). Due to CpG suppression, mammalian DNA has about one-quarter of the CpG dinucleotides of bacterial DNA. Furthermore, eukaryotic 5⬘-CpG-3⬘ motifs are preferentially methylated at the C-5 position of cytosine. Interestingly, most phages show no CpG suppression (Burge et al., 1992), as if CpG suppression is rooted in vertebrate evolution. It has been suggested that vertebrates have retained the ancestral cellular defense function of DNA methylation to inactivate dangerous foreign DNA when genomically integrated (Doerﬂer, 1991). Vertebrates may have adapted this process to serve also as a repressor of endogenous promoters (Bird, 1993). In support of the ‘‘defense’’ argument, foreign DNA can be taken up from the gut and subsequently retrieved from genomic DNA of, for example, circulating murine lymphocytes (Schubbert et al., 1994, 1997). The observation that microbial DNA activates NK cells and B cells (Tokunaga et al., 1992; Messina et al., 1991) suggested that structural characteristics of bacterial DNA that may be responsible for this are not present in mammalian DNA. Because of the exquisite sequence speciﬁcity of CpG[S]ODN-mediated B cell stimulation, the concept emerged that this type of immune cell activation has evolved as an immune defense mechanism capable of recognizing pathogens on the basis of CpG motifs (Krieg et al., 1995; Pisetsky, 1996; Krieg et al., 1998c; Lipford et al., 1998) In support of this, methylation of bacterial DNA and CpG inversion to GpC abolished not only mitogenicity toward B cells (Krieg et al., 1995), but also activation of APCs (Sparwasser et al., 1997a,b; Hacker et al., 1998) and blocked the immunobiology of plasmid DNA (Klinman et al., 1997; Sato et al., 1996). Silencing of integrated and thus genotoxic foreign DNA, via methylation, may represent a key mechanism of cellular defense (Doerﬂer, 1991). Prevention of infections before they cause accumulation of genotoxic foreign DNA might be another strategy adopted during evolution. Accordingly, cells of the innate immune system may have adapted DNA-binding proteins to recognize sequence-speciﬁc bacterial CpG DNA, causing activation of innate immune cells. Once activated, immune cells may serve to prevent and to combat infections. In support of this, DNA-damaging environmental stimuli (Kyriakis and Avruch, 1996) as well as bacterial DNA cause immediate-early activation of the p38 and JNK MAPK cascades (Hacker et al., 1998; Yi and Krieg, 1998b). Whether GADD45-like proteins (Fornace, Jr., et al., 1988; Takekawa and Saito, 1998) are involved in initiating CpG-ODN-triggered stress responses is not yet known.

348

HERMANN WAGNER

Biological evaluation of pathogens, via cell surface-bound pattern recognition receptors, is thought of as the prime task of cells of the innate immune system (Fearon, 1997). Cell surface receptors such as Mac1 may be responsible for DNA uptake into the cells (Benimetskaya et al., 1997b), but they were not sequence speciﬁc. Lysosomotropic compounds impaired CpG[S]ODN-driven immediate-early activation signals (Hacker et al., 1998; Yi and Krieg, 1998b), implying that the postulated PRRs, which recognize and translate CpG-ODN sequences into signaling, act downstream of endosomal maturation. Despite ongoing efforts, CpG-ODN sequence-discriminating DNA-binding proteins acting as PRRs have not yet been identiﬁed. XIV. CpG DNA Acts as Adjuvant for Th1 Responses

The anti-rev ODN sequence (Branda et al., 1993), when used as adjuvant for antibody production to tetanus toxoid, was partially active in mice (Branda et al., 1996b). Several subsequent reports imply that CpG[S]ODNs represent a new class of adjuvants able to promote Th1 responses to proteinaceous antigens. For example, mice challenged subcutaneously with liposome-encapsulated ovalbumin (Lipford et al., 1997a,b) or 웁-galactosidase (웁-Gal) (Sparwasser et al., 1998) plus CpG[S]ODN (as adjuvant) developed strong peptide-speciﬁc CTLs in the draining LNs. Furthermore, speciﬁc antibody responses were augmented, and CpG-ODN switched the isotype pattern to a Th1-type proﬁle, i.e., antigen-speciﬁc IgG2a became dominant (Lipford et al., 1997a). Similar data were obtained in both BALB/ c (Th2 biased) and B10 D2 (Th1 biased) mice immunized with hen egg lysozyme (HEL). Use of CpG[S]ODN as adjuvant switched the immune response to a Th1-dominated cytokine pattern associated with strong antiHEL IgG2a antibody responses (Chu et al., 1997). Finally, insect DNA and CpG[S]ODN surpassed the activity of complete Freund’s adjuvant (CFA) as measured by three different parameters, i.e., T proliferative responses, IFN-웂 synthesis, and production of IgG2a antibodies (Sun et al., 1998). Subunit or inactivated virus vaccines typically induce Th2 responses with high titers of neutralizing antibodies, but fail to induce signiﬁcant cellmediated immunity (Manickan et al., 1995; Raz et al., 1994, 1996). Supplementation of proteinaceous antigens with noncoding plasmid DNA rich in CpG motifs ampliﬁed humoral as well as T cell-mediated immune responses, as did CpG[S]ODN (Roman et al., 1997). CpG-ODN suppressed IgE synthesis but promoted IgG2a responses and IFN-웂 production by antigen-reactive CD4 T cells. CpG DNA transfected into IFN웂-primed (Sweet et al., 1998) murine or human macrophages initiated

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

349

production of IFN-움/웁, IFN-웂, IL-12, and IL-18 (Roman et al., 1997). Induction of Th1 responses by DNA immunization was also shown to be due to immunostimulatory effects of the DNA backbone (Leclerc et al., 1997). Interestingly, mice immunized with recombinant hepatitis B virus surface antigen (HBsAg) together with CpG[S]ODN (as adjuvant) developed high titers of anti-HBs antibodies, which were further augmented by the additional use of the standard adjuvant, aluminum hydroxide (alum). Even when admixed with alum, CpG[S]ODN triggered strong Th1 responses, i.e., predominantly IgG2a antibodies, and primed for CTLs (Davis et al., 1998). Taken together, these studies recommend plasmid CpG DNA sequences and synthetic CpG[S]ODN as powerful Th1 adjuvants for immune responses to natural or recombinant proteins. In one study (Sun et al., 1998), CpG[S]ODN even surpassed the ‘‘golden standard’’ of CFA. XV. CpG DNA Mediates Harmful Effects in Vivo

Similar to LPS, challenge of mice with CpG[S]ODN sequence-speciﬁcally triggered a lethal cytokine syndrome due to release of too high amounts of TNF-움 from macrophages and DCs (Sparwasser et al., 1997b). Interestingly, bacterial DNA as well as unmethylated CpG[S]ODN caused inﬂammation similar to inﬂammatory lung disease when intratracheally instilled in mice. Lavage of intratracheally challenged mice revealed a massive increase in cellularity, in particular of neutrophils. In addition, TNF-움, IL-6, and MIP-2 concentrations were high. Bacterial DNA extracted from sputum of cystis ﬁbrosis (CF) patients mimicked these effects. A possible relevance of these data to human lung diseases was suggested (Schwartz et al., 1997). XVI. CpG DNA Acts as Adjuvant for Antitumor Responses

Conceptionally, attempts to augment antitumor responses may beneﬁt from the ability of bacterial DNA and CpG[S]ODN to activate B cells (Messina et al., 1991; Krieg et al., 1995), NK cells (Yamamoto et al., 1992a; Ballas et al., 1996), and macrophages and DCs (Sparwasser et al., 1997b, 1998; Jakob et al., 1998), and the adjuvanticity of CpG[S]ODN for CTL induction toward proteinaceous antigens (Lipford et al., 1997a). Furthermore, CpG-ODN acted as adjuvant for Th1 polarization of cellular responses (Lipford et al., 1997a; Roman et al., 1997; Davis et al., 1998; Sun et al., 1998). Although studies using CpG-ODN to augment antitumor responses are still in their infancy, initial reports appear promising.

350

HERMANN WAGNER

Passive therapy with monoclonal antibodies (mAbs) can have antitumor effects in patients (Maloney et al., 1997). NK cells and monocytes/macrophages participated in antibody-dependent cellular cytotoxicity (ADCC), which played a role in the response to antibody therapy (Kaminski et al., 1986). CpG[S]ODNs were assayed for enhancing antitumor effects of mAbs in a murine lymphoma model system. In vivo, there was a synergy between CpG[S]ODNs and the protective effects of antitumor mAbs. In vitro, CpG[S]ODN increased ADCC (Wooldridge et al., 1997). Using the secreted antibody from 38C13 murine lymphoma to immunize against a protein ‘‘tumor’’antigen, CpG[S]ODNs were assessed for adjuvanticity (Weiner et al., 1997). Adjuvanticity of CpG[S]ODN was as effective as that of CFA in promoting protection against tumor challenge. Of note, higher titers of IgG2a antitumor antibodies were observed when CpG-ODNs were used. The ability of CpG[S]ODNs to promote as Th1 responses has been extended to model tumor systems. Growth of murine lymphoma cells stably transfected with ovalbumin was inhibited in mice vaccinated with ovalbumin plus CpG[S]ODN (R. M. Vabulas, personal communication). In this system, tumor protection was mediated, at least in part, by an (ovalbumin) SIINFEKL peptide-speciﬁc CTL expressing the CD8⫹CD4⫺ phenotype. CpG[S]ODN also acted as adjuvant for CD4⫹ T cell response induction toward tumor-speciﬁc peptide antigens (Sun et al., 1998). XVII. CpG DNA Reverts Th2-Oriented Pathology

Epidemiological studies indicate an inverse relationship between Th1promoting infections and the propensity of individuals to develop atopic disorders (Shirakawa et al., 1997; von Mutius et al., 1992). In mice, mycobacteria have been shown to prevent the pathogenesis of asthmatic inﬂammation (Erb et al., 1998). This raised the question as to whether bacterial DNA and CpG[S]ODNs could mimic the Th1-promoting effect of bacterial infections. Coadministration of CpG[S]ODN (used as adjuvant) with Schistosoma mansoni eggs (used as antigen) prevented airway eosinophilia, Th2 cytokine induction, IgE production, and bronchial hyperreactivity in this murine model of asthma (Kline et al., 1998). The Th1-promoting adjuvanticity of CpG[S]ODN may still be operative in IL-12⫺/⫺ and IFN웂⫺/⫺ knockout mice, although reduced (Kline et al., 1998). Because, on balance, IFN-움/웁 favors Th1 responses (Brinkmann et al., 1993; Vieillard et al., 1997; Cousens et al., 1997), a need to analyze the effect of CpGODN-induced IFN-움/웁 was suggested. Infection of BALB/c mice with Leishmania major is another model for Th2-driven disease (Locksley and Louis, 1992). Nitric oxide (NO) gener-

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

351

ated in APCs by the inducible NO synthase (iNOS) was crucial for the control of L. major (Stenger et al., 1996). CpG[S]ODNs promoted Th1 development in LNs draining the site of L. major challenge and conveyed NO-dependent protective immunity to disease-prone BALB/c mice. Conversion from the L. major-susceptible Th2 phenotype to a Th1-resistant phenotype was associated with IL-12 production, and sustained expression of the IL-12 receptor 웁 chain. Strikingly, CpG[S]ODNs could cure when administered as late as 15–20 days after infection (Zimmermann et al., 1998; Lipford et al., 1999b). XVIII. CpG DNA Acts as Adjuvant for Mucosal Immunity

Following intranasal immunization with formalin-inactivated inﬂuenza virus, bacterial DNA as well as CpG[S]ODNs promoted the increased production of inﬂuenza-speciﬁc antibodies in serum, saliva, and the genital tract (Moldoveanu et al., 1998). In another study the mucosal adjuvanticity of CpG[S]DNA was compared to that of cholera toxin (CT). Intranasal immunization with HBsAg together with CpG[S]ODN triggered stronger mucosal immune responses as compared to CT (McCluskie and Davis, 1998). Because mucous membrane surfaces of the respiratory, digestive, and genito-urinary systems represent the prime transmission site for many pathogens, mucosal adjuvanticity of CpG-ODN is promising. To date, no mucosal adjuvant has been licensed for human use. XIX. CpG DNA Causes Extramedullary Hematopoiesis

CpG[S]ODNs cause extramedullary hematopoiesis (Sparwasser et al., 1999a). Challenge of mice with CpG[S]ODNs led sequence speciﬁcally to a transient splenomegaly with up to threefold increase in spleen weight at day 6. Cell surface phenotyping revealed a relative increase in cellularity of B220⫺/CD3⫺ double-negative cells. In this cell compartment, the number of GM-CFUs, of BFU-E, and of CFU-S was dramatically increased. As a consequence, CpG-ODNs, could be used to mediate protective effects in sublethally irradiated mice. For example, CpG[S]ODNs compensated radiation-induced damage of the lympho-hematopoietic system by accelerating the regeneration of the immune system, as shown in two experimental systems: the induction of CTLs to ovalbumin and resistance to the intracellular pathogen Listeria monocytogenes (Sparwasser et al., 1999a). Enlarged spleens (day 6) of CpG[S]ODN-treated mice represented a convenient source of progenitors for DCs. On culture in GM-CSF conditioned medium, large numbers of DCs could be obtained (T. Sparwasser, personal communication).

352

HERMANN WAGNER

XX. CpG DNA Activates Human Immune Cells

The MY-1 fraction from BCG and Escherichia coli DNA were shown to augment lytic NK cell activity in human PBMCs (Yamamoto et al., 1994a; Ballas et al., 1996). HIV rev antisense [S]ODNs induced B cell proliferation in human PBMCs, but the results were not linked to CpG motifs (Branda et al., 1996b). Palindromic CpG[S]ODNs derived from ampR of plasmid vectors induced the release of IFN, and possibly of IL-12 and IL-18, from human PBMCs. However, cationic lipids were used to transfect the CpG-ODN directly into cells (Roman et al., 1997). An additional uncertainty clouds these types of studies because human CD14positive PBMCs were found to be extremely sensitive to LPS. For example, concentrations as low 1–5 pg/ml of LPS triggered IL-6 and TNF-움 production in CD14-positive PBMCs as assessed by intracellular cytokine staining (S. Bendigs, unpublished data). Such low LPS concentrations may not be detectable in conventional Limulus amebocyte lysate (LAL) assays. Another problem is highlighted by recent evidence that optimal ﬂanking regions of CpG motifs may be species speciﬁc, and thus distinct in human and mouse DNA (A. M. Krieg, personal communication). CpG[S]ODN 1668 [known for efﬁcient stimulation of murine B cells (Krieg et al., 1995) and DCs (Sparwasser et al., 1998)], the HIV anti-rev ODN DSP30 [shown to be mitogenic for human B cells (Branda et al., 1996a; Liang et al., 1996)], and CpG[S]ODN 2006 (active toward human cells; A. M. Krieg, personal communication) were compared for their ability to stimulate cells within human PMBCs. Sequence speciﬁcally, CpG[S]ODN 2006 and DSP30 were mitogenic toward B cells, they activated monocytes to release IL-6 and IL-12, and induced up-regulation of MHC class I and II, as well as CD40 and CD86 on macrophages and B cells. However, TNF-움 cytokine induction was poor (Bauer et al., 1999). Interestingly, the ‘‘murine’’ CpG[S]ODN 1668 was almost inactive toward human cells whereas the ‘‘human’’ CpG-ODN DSP30 and 2006 was poor stimulators of murine cells. Overall these data suggest that a CpG motif as such is not sufﬁcient to activate human immune cells and that ﬂanking regions inﬂuence human immune cell activation. Of note, the immunostimulating effect of DSP30 toward human B cells (Liang et al., 1996) was chloroquine sensitive, implying necessity of endosomal maturation (Bauer et al., 1999). XXI. CpG DNA Mediates Signaling: Stimulation versus Costimulation

By virtue of their capacity for surveillance and phagocytotic/endocytotic properties, DCs are likely to engage pathogens through PRRs at an early stage of infectious challenge (Banchereau and Steinman, 1998). Signals

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

353

induced by pattern recognition may be grouped into three categories: (1) proinﬂammatory responses, including the secretion of TNF-움, IL-6, IFN-움웁, and chemokines, (2) costimulation of T cell activation, including CD80, CD86, and CD40, and (3) induction of effector cytokines, including IL-10, IL-12, TGF-움, and IFN-웂. Bacterial DNA and CpG[S]ODNs directly stimulated DCs and macrophages to deliver signals from all of these three categories. On the other hand, CpG-ODN-mediated costimulation of T cells (Lipford et al., 1999a) and NK cells (Chace et al., 1997) was not direct but required preceding activation signals. B cells, however, responded both to CpG-ODNs as the only stimulus and when presented as a costimulus, because cross-linking B cell receptors (BCRs) greatly augmented responsiveness (Krieg et al., 1995). Direct CpG-ODN responsiveness of B cells was sensitive to lysosomatropic compounds known to block endosomal maturation (Yi and Krieg, 1998b). CpG-ODN-mediated costimulation of T cells was insensitive to lysosomotropic compounds (Lipford et al., 1999a). Sensitivity to these compounds of NK cells and of B cells triggered via BCR cross-linking has not been evaluated as yet. It has been suggested that costimulation of T cells, B cells, and NK cells becomes triggered via signaling competent cell surface receptors (Lipford et al., 1998b) (Fig. 2). XXII. CpG DNA Allows MHC Class I-Restricted CTL Responses to Exogeneous Antigens

Two separate processing pathways for presentation of exogenous and endogenous antigens are known to exist. MHC class II-restricted CD4⫹ helper T cells respond to extracellular antigens, whereas MHC class I-restricted CD8⫹ CTL responses are directed against endogeneous cytosolic antigens (Germain and Margulies, 1993; Bevan, 1995). This strict dichotomy has been challenged by observations indicating that peptides generated from exogenous proteins can gain access to the cytosol and therefore be presented by MHC class I molecules ( Jondal et al., 1996). Professional APCs such as DCs play an important role in cross-priming during induction of primary CD8 T cell-mediated responses to soluble antigen (Paglia et al., 1996). It is still controversial as to which pathways exogenous antigens are presented on MHC class I molecules. In the case of soluble protein antigens such as ovalbumin, sampling of antigen via micropinocytosis allowed access to the cytosol of DCs and processing via the conventional MHC class I presentation pathway [known to be ‘‘transporter associated with antigen presentation’’ (TAP) dependent and brefeldin A sensitive] (Brossart and Bevan, 1997; Lehner and Cresswell, 1996). Exogenous antigens such as bacteria, aggregated proteins, or anti-

354

HERMANN WAGNER

FIG. 2. CpG DNA-mediated signaling: direct stimulation versus costimulation. Bacterial CpG DNA directly stimulates B cells and APCs such as dendritic cells and macrophages. Stimulation is sensitive to lysosomotropic compounds, implying that endosomal maturation is a critical step. Costimulation of T cells is resistant to lysosomotropic compounds. It is hypothesized that costimulation occurs via signaling-competent cell surface receptors. The mechanism of B cell and NK cell costimulation is unclear.

gens incorporated into liposomes are internalized by phagocytosis and may leak into the cytosol to be processed via the conventional MHC class I presentation pathway (Rock, 1996). Inactivated or viruslike particles and glycopeptides, however, appear to access the MHC class I presentation pathway via a chloroquine-sensitive but TAP-independent and brefeldin A-resistant endosomal pathway, similar to the endosomal class II MHC presentation pathway ( Jondal et al., 1996). Finally, synthetic peptides that have the correct MHC class I binding motif do not require processing and, in the presence of adjuvants, prime for CTL responses in vivo (Aichele et al., 1990). One may conclude that APCs have several pathways at their disposal to allow exogenous antigens to access the MHC class I presentation pathway. However, to activate primary antigen-reactive T cells, DCs need to mature into professional APCs and this maturation is driven efﬁciently by bacterial DNA or CpG[S]ODNs (Sparwasser et al., 1998). Use of CpG[S]ODNs as adjuvant permitted CD8⫹ CTL responses not only to proteins such as ovalbumin and 웁-Gal entrapped in liposomes (Lipford et

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

355

al., 1997a; Sparwasser et al., 1998), but also toward the soluble proteins as well as the immunodominant peptides from these proteins (G. B. Lipford and R. M. Vabulas, unpublished data). Strong CTL responses were induced in LNs, provided the LNs were draining the site of both antigen and CpGODN (adjuvant) challenge. CpG-ODNs caused translocation into T cell areas of DCs staining for intracytoplasmic IL-12 (Sparwasser et al., 1999b). CpG-ODN-mediated direct activation of DCs was independent of CD40– CD40L interaction. This was considered key to explain the strong CTL responses induced. To date, it is unclear whether CpG-ODNs facilitate MHC class I presentation via the endogeneous or exogeneous antigenprocessing pathway. However, use of CpG DNA as adjuvant and thus CD40–CD40L-independent but direct activation of APCs allowed bypassing the CD4 T helper cell requirement for the induction of CTL responses toward soluble proteins. This may explain why class I MHC binding peptides can act in the presence of the adjuvant CpG-ODN as strong antigen in CD4 T cell knockout mice (T. Sparwasser, personal communication). XXIII. Concluding Remarks

Although the immunobiology of bacterial DNA and mimicking CpG DNA motifs has been discovered only relatively recently, a remarkable level of knowledge has been reached, and many possible therapeutic applications are being tested or have been proposed. Central to its physiological relevance is the question whether ‘‘sensing’’ of CpG DNA reﬂects an evolutionary vestige of the innate system to signal infectious danger. If pathogen-derived CpG DNA is sensed and evaluated via pattern recognition receptors, the brisk cell activation observed may be classiﬁed as an ancestral stress response to potentially genotoxic foreign DNA. However, determining the relative role of bacterial CpG DNA during infection will be a formidable task. Another burning issue relates to the question of whether bacterial CpG DNA mimics Th1-promoting infections. If so, CpG DNA may inﬂuence the propensity of individuals to develop atopic disorders. Whatever the outcome, it has become clear that CpG-DNA sequences speciﬁcally alert cells of the innate immune system to interact with cells of the adaptive immune system by providing costimulatory molecules and effector cytokines and by causing proinﬂammatory responses. By conditioning for Th1 responses, bacterial CpG DNA may interfere with peripheral tolerance and allow productive T cell responses. Although DNA vaccines and antisense therapy expose the host to CpG DNA, deciphering CpGODN sequences, which effectively stimulate human immune cells, represents a major and immediate task. Undoubtedly, the study of the immunobiology of bacterial CpG DNA has elucidated rules of the interplay of innate

356

HERMANN WAGNER

and adaptive immunity. To exploit the unique and potent capacity of CpG DNA for therapeutic application in humans is one of the many challenges of modern science. ACKNOWLEDGMENTS This work was in part supported by the Deutsche Forschungsgemeinschaft, SFB 391 and SFB 464, a grant from the Bundesministerium fu¨r Bildung, Wissenschaft und Technologie and CpG-Immunopharmaceuticals GmbH and the Wilhelm-Sander-Foundation. I thank my collegues G. Lipford, K. Heeg, G. Hacker, H. Hacker, T. Sparwasser, and T. Miethke for continuing valuable discussions and help in preparing this manuscript, and A. M. Krieg for sharing unpublished data. Thanks to C. Ba¨uerle for excellent secretarial assistance.

REFERENCES Abbas, A. K., Murphy, K. M., and Sher, A. (1996). Functional diversity of helper T lymphocytes. Nature (London) 383, 787–793. Aichele, P., Hengartner, H., Zinkernagel, R. M., and Schulz, M. (1990). Antiviral cytotoxic T cell response induced by in vivo priming with a free synthetic peptide. J. Exp. Med. 171, 1815–1820. Albert, M. L., Pearce, S. F. A., Francisco, L. M., Sauter, B., Roy, P., Silverstein, R. L., and Bhardwaj, N. (1998). Immature dendritic cells phagocytose apoptotic cells via alphavbeta5 and CD36, and cross-present antigens to cytotoxic T lymphocytes J. Exp. Med. 188, 1359– 1368. Anitescu, M., Chace, J. H., Tuetken, R., Yi, A. K., Berg, D. J., Krieg, A. M., and Cowdery, J. S. (1997). Interleukin-10 functions in vitro and in vivo to inhibit bacterial DNA-induced secretion of interleukin-12. J. Interferon Cytokine. Res. 17, 781–788. Ballas, Z. K., Rasmussen, W. L., and Krieg, A. M. (1996). Induction of NK activity in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial DNA. J. Immunol. 157, 1840–1845. Banchereau, J., and Steinman, R. M. (1998). Dendritic cells and the control of immunity. Nature (London) 392, 245–252. Barker, R. H., Jr., Metelev, V., Rapaport, E., and Zamecnik, P. (1996). Inhibition of Plasmodium falciparum malaria using antisense oligodeoxynucleotides. Proc. Natl. Acad. Sci. U.S.A. 93, 514–518. Bast, R. C., and Bast, B. S. (1976). Critical review of previously reported animal studies of tumor immunotherapy with nonspeciﬁc immunostimulants. Ann. N.Y. Acad. Sci. 277, 60–93. Bauer, M., Heeg, K., Wagner, H., and Lipford, G. B. (1999). DNA activates human immune cells through a CpG sequence dependent manner. Immunology In press. Beckman, E. M. Porcelli, S. A., Morita, C. T., Behar, S. M., Furlong, S. T., and Brenner, M. B. (1994) Recognition of a lipid antigen by CD1-restricted alpha beta⫹ T cells [see comments]. Nature (London) 372, 691–694. Bendelac, A., and Fearon, D. T. (1997). Innate pathways that control acquired immunity [editorial]. Curr. Opin. Immunol. 9, 1–3. Bendelac, A., Rivera, M. N., Park, S. H., and Roark, J. H. (1997) Mouse CD1-speciﬁc NK1 T cells: Development, speciﬁcity, and function. Annu. Rev. Immunol. 15, 535–562. Bendigs, S., Salzer, U., Lipford, G. B., Wagner, H., and Heeg, K. (1999). CpG-oligodeoxynucleotides costimulate primary T cells in the absence of antigen presenting cells. Eur. J. Immunol. In press.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

357

Benimetskaya, L., Berton, M., Kolbanovsky, A., Benimetsky, S., and Stein, C. A. (1997a). Formation of a G-tetrad and higher order structures correlates with biological activity of the RelA (NF-kappaB p65) ‘antisense’ oligodeoxynucleotide. Nucleic Acids Res. 25, 2648– 2656. Benimetskaya, L., Loike, J. D., Khaled, Z., Loike, G., Silverstein, S. C., Cao, L., el Khoury, J., Cai, T. Q., and Stein, C. A. (1997b). Mac-1 (CD11b/CD18) is an oligodeoxynucleotidebinding protein [see comments]. Nature Med. 3, 414–420. Bennett, R. M., Gabor, G. T., and Merritt, M. M. (1985). DNA binding to human leukocytes. Evidence for a receptor-mediated association, internalization, and degradation of DNA. J. Clin. Invest. 76, 2182–2190. Bevan, M. J. (1995). Antigen presentation to cytotoxic T lymphocytes in vivo. J. Exp. Med. 182, 639–641. Bird, A. P. (1980). DNA methylation and the frequency of CpG in animal DNA. Nucleic Acids Res. 8, 1499–1504. Bird, A. P., (1993). Functions for DNA methylation in vertebrates. Cold Spring Harb. Symp. Quant. Biol. 58, 281–285. Bird, A. P., Taggart, M. H., Nicholls, R. D., and Higgs, D. R. (1987). Non-methylated CpG-rich islands at the human alpha-globin locus: Implications for evolution of the alphaglobin pseudogene. EMBO J. 6, 999–1004. Branch, A. D. (1998). A good antisense molecule is hard to ﬁnd. Trends Biochem. Sci. 23, 45–50. Branda, R. F., Moore, A. L., Mathews, L., McCormack, J. J., and Zon, G. (1993). Immune stimulation by an antisense oligomer complementary to the rev gene of HIV-1. Biochem. Pharmacol. 45, 2037–2043. Branda, R. F., Moore, A. L., Hong, R., McCormack, J. J., Zon, G., and Cunningham-Rundles, C. (1996a). B-cell proliferation and differentiation in common variable immunodeﬁciency patients produced by an antisense oligomer to the rev gene of HIV-1. Clin. Immunol. Immunopathol. 79, 115–121. Branda, R. F., Moore, A. L., Lafayette, A. R., Mathews, L., Hong, R., Zon, G., Brown, T., and McCormack, J. J. (1996b). Ampliﬁcation of antibody production by phosphorothioate oligodeoxynucleotides. J. Lab. Clin. Med. 128, 329–338. Bretscher, P., and Cohn, M. (1970). A theory of self–nonself discrimination. Science 169, 1042–1049. Brinkmann, V., Geiger, T., Alkan, S., and Heusser, C. H. (1993). Interferon alpha increases the frequency of interferon gamma-producing human CD4⫹ T cells. J. Exp. Med. 178, 1655–1663. Brossart, P., and Bevan, M. J. (1997). Presentation of exogenous protein antigens on major histocompatibility complex class I molecules by dendritic cells: Pathway of presentation and regulation by cytokines. Blood 90, 1594–1599. Burge, C., Campbell, A.M., and Karlin, S. (1992). Over- and under-representation of short oligonucleotides in DNA sequences. Proc. Natl. Acad. Sci. U.S.A. 89, 1358–1362. Castier, Y., Chemla, E., Nierat, J., Heudes, D., Vasseur, M. A., Rajnoch, C., Bruneval, P., Carpentier, A., and Fabiani, J.N. (1998). The activity of c-myb antisense oligonucleotide to prevent intimal hyperplasia is nonspeciﬁc. J. Cardiovasc. Surg. (Torino) 39, 1–7. Cella, M., Engering, A., Pinet, V., Pieters, J., and Lanzavecchia, A. (1997). Inﬂammatory stimuli induce accumulation of MHC class II complexes on dendritic cells [see comments]. Nature (London) 388, 782–787. Chace, J. H., Hooker, N. A., Mildenstein, K. L., Krieg, A. M., and Cowdery, J. S. (1997). Bacterial DNA-induced NK cell IFN-gamma production is dependent on macrophage secretion of IL-12. Clin. Immunol. Immunopathol. 84, 185–193.

358

HERMANN WAGNER

Chu, R. S., Targoni, O. S., Krieg, A. M., Lehmann, P. V., and Harding, C. V. (1997). CpG oligodeoxynucleotides act as adjuvants that switch on T helper 1 (Th 1) immunity. J. Exp. Med. 186, 1623–1631. Clipstone, N. A., and Crabtree, G. R. (1992). Identiﬁcation of calcineurin as a key signalling enzyme in T-lymphocyte activation. Nature (London) 357, 695–697. Condon, C., Watkins, S. C., Celluzzi, C. M., Thompson, K., and Falo, L. D., Jr. (1996). DNAbased immunization by in vivo transfection of dendritic cells. Nature, Med. 2, 1122–1128. Coulondre, C., Miller, J. H., Farabaugh, P. J., and Gilbert, W. (1978). Molecular basis of base substitution hotspots in Escherichia coli. Nature (London) 274, 775–780. Cousens, L. P., Orange, J. S., Su, H. C., and Biron, C. A. (1997). Interferon-alpha/beta inhibition of interleukin 12 and interferon-gamma production in vitro and endogenously during viral infection. Proc. Natl. Acad. Sci. U.S.A. 94, 634–639. Cowdery, J. S., Chace, J. H., Yi, A. K., and Krieg, A. M. (1996). Bacterial DNA induces NK cells to produce IFN-gamma in vivo and increases the toxicity of lipopolysaccharides. J. Immunol. 156, 4570–4575. Crooke, R. M., Crooke, S. T., Graham, M. J., and Cooke, M. E. (1996). Effect of antisense oligonucleotides on cytokine release from human keratinocytes in an in vitro model of skin. Toxicol. Appl. Pharmacol. 140, 85–93. Davis, H. L., Weeranta, R., Waldschmidt, T. J., Tygrett, L., Schorr, J., and Krieg, A. M. (1998). CpG DNA is a potent enhancer of speciﬁc immunity in mice immunized with recombinant hepatitis B surface antigen. J. Immunol. 160, 870–876. Delgado, S., Gomez, M., Bird, A., and Antequera, F. (1998). Initiation of DNA replication at CpG islands in mammalian chromosomes. EMBO J. 17, 2426–2435. Dempsey, P. W., Allison, M. E., Akkaraju, S., Goodnow, C. C., and Fearon, D. T. (1996). C3d of complement as a molecular adjuvant: Bridging innate and acquired immunity. Science 271, 348–350. De Smedt, T., Pajak, B., Muraille, E., Lespagnard, L., Heinen, E., De Baetselier, P., Urbain, J., Leo, O., and Moser, M. (1996). Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J. Exp. Med. 184, 1413–1424. Doerﬂer, W. (1991). Patterns of DNA methylation—evolutionary vestiges of foreign DNA inactivation as a host defense mechanism. A proposal. Biol. Chem. Hoppe Seyler 372, 557–564. Donnelly, J. J., Ulmer, J. B., Shiver, J. W., and Liu, M. A. (1997). DNA vaccines. Annu. Rev. Immunol. 15, 617–648. Emlen, W., Pisetsky, D. S., and Taylor, R. P. (1986). Antibodies to DNA. A perspective. Arthritis Rheum. 29, 1417–1426. Erb, K. J., Holloway, J. W., Sobeck, A., Moll, H., and Le Gros, G. (1998). Infection of mice with Mycobacterium bovis-Bacillus Calmette–Guerin (BCG) suppresses allergeninduced airway eosinophilia. J. Exp. Med. 187, 561–569. Fearon, D. T. (1997). Seeking wisdom in innate immunity [news; comment]. Nature (London) 388, 323–324. Fearon, D. T., and Locksley, R. M. (1996). The instructive role of innate immunity in the acquired immune response. Science 272, 50–53. Fisher, T. L., Terhorst, T., Cao, X., and Wagner, R. W. (1993). Intracellular disposition and metabolism of ﬂuorescently-labeled unmodiﬁed and modiﬁed oligonucleotides microinjected into mammalian cells. Nucleic Acids Res. 21, 3857–3865. Fornace, A. J. Jr., Alamo, I. Jr., and Hollander, M. C. (1988). DNA damage-inducible transcripts in mammalian cells. Proc. Natl. Acad. Sci. U.S.A. 85, 8800–8804. Fox, R. I. (1993). Mechanism of action of hydroxychloroquine as an antirheumatic drug. Semin. Arthritis Rheum. 23, 82–91.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

359

Freimark, B. D., Blezinger, H. P., Florack, V. J., Nordstrom, J. L., Long, S. D., Deshpande, D. S., Nochumson, S., and Petrak, K. L. (1998). Cationic lipids enhance cytokine and cell inﬂux levels in the lung following administration of plasmid: Cationic lipid complexes. J. Immunol. 160, 4580–4586. Fynan, E. F., Webster, R. G., Fuller, D. H., Haynes, J. R., Santoro, J. C., and Robinson, H. L. (1993). DNA vaccines: Protective immunizations by parenteral, mucosal, and genegun inoculations. Proc. Natl. Acad. Sci. U.S.A. 90, 11478–11482. Gellert, M. (1997). Recent advances in understanding V(D)J recombination Adv. Immunol. 64, 39–64. Germain, R. N., and Margulies, D. H. (1993). The biochemistry and cell biology of antigen processing and presentation. Annu. Rev. Immunol. 11, 403–450. Gluck, R., Mischler, R., Brantschen, S., Just, M., Althaus, B., and Cryz, S. J. Jr. (1992). Immunopotentiating reconstituted inﬂuenza virus virosome vaccine delivery system for immunization against hepatitis A. J. Clin. Invest. 90, 2491–2495. Green, D. R., and Scott, D. W. (1994). Activation-induced apoptosis in lymphocytes. Curr. Opin. Immunol. 6, 476–487. Guvakova, M. A., Yakubov, L. A., Vlodavsky, I., Tonkinson, J. L., and Stein, C. A. (1995). Phosphorothioate oligodeoxynucleotides bind to basic ﬁbroblast growth factor, inhibit its binding to cell surface receptors, and remove it from low afﬁnity binding sites on extracellular matrix. J. Biol. Chem. 270, 2620–2627. Hacker, H., Mischak, H., Miethke, T., Liptay, S., Schmid, R., Sparwasser, T., Heeg, K., Lipford, G. B., and Wagner, H. (1998). CpG-DNA-speciﬁc activation of antigen-presenting cells requires stress kinase activity and is preceded by non-speciﬁc endocytosis and endosomal maturation. EMBO J. 17, 6230–6240. Halpern, M. D., and Pisetsky, D. S. (1995). In vitro inhibition of murine IFN gamma production by phosphorothioate deoxyguanosine oligomers. Immunopharmacology 29, 47–52. Halpern, M. D., Kurlander, R. J., and Pisetsky, D. S. (1996). Bacterial DNA induces murine interferon-gamma production by stimulation of interleukin-12 and tumor necrosis factoralpha. Cell Immunol. 167, 72–78. Han, J., Lee, J. D., Bibbs, L., and Ulevitch, R. J. (1994). A MAP kinase targeted by endotoxin and hyperosmolarity in mammalian cells. Science 265, 808–811. Hartmann, G., Krug, A., Waller-Fontaine, K., and Endres, S. (1996). Oligodeoxynucleotides enhance lipopolysaccharide-stimulated synthesis of tumor necrosis factor: Dependence on phosphorothioate modiﬁcation and reversal by heparin. Mol. Med. 2, 429–438. Hashimoto, W., Takeda, K., Anzai, R., Ogasawara, K., Sakihara, H., Sugiura, K., Seki, S., and Kumagai, K. (1995). Cytotoxic NK 1.1 Ag⫹ alpha beta T cells with intermediate TCR induced in the liver of mice by IL-12. J. Immunol. 154, 4333–4340. Hertl, M., Neckers, L. M., and Katz, S. I. (1995). Inhibition of interferon-gamma-induced intercellular adhesion molecule-1 expression on human keratinocytes by phosphorothioate antisense oligodeoxynucleotides is the consequence of antisense-speciﬁc and antisensenon-speciﬁc effects. J. Invest. Dermatol. 104, 813–818. Hibi, M., Lin, A., Smeal, T., Minden, A., and Karin, M. (1993). Identiﬁcation of an oncoprotein- and UV-responsive protein kinase that binds and potentiates the c-Jun activation domain. Genes Dev. 7, 2135–2148. Jakob, T., Walker, P. S., Krieg, A. M., Udey, M. C., and Vogel, J. C. (1998). Activation of cutaneous dendritic cells by CpG-containing oligodeoxynucleotides: A role for dendritic cells in the augmentation of Th1 responses by immunostimulatory DNA. J. Immunol. 161, 3042–3049.

360

HERMANN WAGNER

Janeway, C. A. Jr. (1989). Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harb. Symp. Quant. Biol. 54, 1–13. Jondal, M., Schirmbeck, R., and Reimann, J. (1996). MHC class I-restricted CTL responses to exogenous antigens. Immunity 5, 295–302. Jones, P. A., Rideout, W. M., 3d, Shen, J. C., Spruck, C. H., and Tsai, Y. C. (1992). Methylation, mutation and cancer. Bioessays 14, 33–36. Kaminski, M. S., Kitamura, K., Maloney, D. G., Campbell, M. J., and Levy, R. (1986). Importance of antibody isotype in monoclonal anti-idiotype therapy of a murine B cell lymphoma. A study of hybridoma class switch variants. J. Immunol. 136, 1123–1130. Karin, M. (1995). The regulation of AP-1 activity by mitogen-activated protein kinases. J. Biol. Chem. 270, 16483–16486. Karlin, S., Doerﬂer, W., and Cardon, L. R. (1994). Why is CpG suppressed in the genomes of virtually all small eukaryotic viruses but not in those of large eukaryotic viruses? J. Virol. 68, 2889–2897. Karounous, D. G., Grudier, J. P., and Pisetsky, D. S. (1968). Spontaneous expression of antobodies to DNA of various species origin in sera of normal subjects and patients with systemic lupus erythematosus. J. Immunol. 140, 451–459. Kimura, Y., Sonehara, K., Kuramoto, E., Makino, T., Yamamoto, S., Yamamoto, T., Kataoka, T., and Tokunaga, T. (1994). Binding of oligoguanylate to scavenger receptors is required for oligonucleotides to augment NK cell activity and induce IFN. J. Biochem. (Tokyo) 116, 991–994. Kline, J. N., Waldschmidt, T. J., Businga, T. R., Lemish, J. E., Weinstock, J. V., Thorne, P. S., and Krieg, A. M. (1998). Modulation of airway inﬂammation by CpG oligodeoxynucleotides in a murine model of asthma. J. Immunol. 160, 2555–2559. Klinman, D. M., Yi, A. K., Beaucage, S. L., Conover, J., and Krieg, A. M. (1996). CpG motifs present in bacteria DNA rapidly induce lymphocytes to secrete interleukin 6, interleukin 12, and interferon gamma. Proc. Natl. Acad. Sci. U.S.A. 93, 2879–2883. Klinman, D. M., Yamshchikov, G., and Ishigatsubo, Y. (1997). Contribution of CpG motifs to the immunogenicity of DNA vaccines. J. Immunol. 158, 3635–3639. Krieg, A. M. (1996). An innate immune defense mechanism based on the recognition of CpG motifs in microbial DNA.J. Lab. Clin. Med. 128, 128–133. Krieg, A. M., Gause, W. C., Gourley, M. F., and Steinberg, A. D. (1989). A role for endogenous retroviral sequences in the regulation of lymphocyte activation. J. Immunol. 143, 2448–2451. Krieg, A. M., Yi, A. K., Matson, S., Waldschmidt, T. J., Bishop, G. A., Teasdale, R., Koretzky, G. A., and Klinman, D. M. (1995). CpG motifs in bacterial DNA trigger direct B-C cell activation. Nature (London) 374, 546–549. Krieg, A. M., Matson, S., and Fisher, E. (1996). Oligodeoxynucleotide modiﬁcations determine the magnitude of B cell stimulation by CpG motifs. Antisense Nucleic Acid Drug Dev. 6, 133–139. Krieg, A. M., Love-Homan, L., Yi, A. K., and Harty, J. T. (1998a). CpG DNA induces sustained IL-12 expression in vivo and resistance to Listeria monocytogenes challenge. J. Immunol. 161, 2428–2434. Krieg, A. M., Wu, T., Weeratna, R., Eﬂer, S. M., Love-Homan, L., Yang, L., Yi, A. K., Short, D., and Davis, H. L. (1998b). Sequence motifs in adenoviral DNA block immune activation by stimulatory CpG motifs. Proc. Natl. Acad. Sci. U.S.A. 95, 12631–12636. Krieg, A. M., Yi, A. K., Schorr, J., and Davis, H. L. (1998c). The role of CpG dinucleotides in DNA vaccines. Trends Microbiol. 6, 23–27. Krieger, M. (1992). Molecular ﬂypaper and atherosclerosis: Structure of the macrophage scavenger receptor. Trends. Biochem. Sci. 17, 141–146.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

361

Kyriakis, J. M., and Avruch, J. (1996). Sounding the alarm: Protein kinase cascades activated by stress and inﬂammation. J. Biol. Chem. 271, 24313–24316. Kyriakis, J. M., Banerjee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F., Avruch, J., and Woodgett, J. R. (1994). The stress-activated protein kinase subfamily of c-Jun kinases. Nature (London) 369, 156–160. Lafferty, K. J., and Woolnough, J. (1977). The origin and mechanism of the allograft reaction. Immunol. Rev. 35, 231–262. Lang, R., Hu¨ltner, L., Lipford, G. B., Wagner, H., and Heeg, K. (1999). G-Rich oligodeoxynucleotides induce proliferation of bone marrow cells. Submitted. Lapatschek, M. S., Gilbert, R., Wagner, H., and Miethke, T. (1998). Activation of macrophages and B-lymphocytes by an oligodeoxynucleotide derived from an acutely pathogenic simian immunodeﬁciency virus. Antisense Nucleic Acid Drug Dev. 8, 357–370. Leclerc, C., Deriaud, E., Rojas, M., and Whalen, R. G. (1997). The preferential induction of a Th1 immune response by DNA-based immunization is mediated by the immunostimulatory effect of plasmid DNA. Cell Immunol. 179, 97–106. Lee, F. S., Hagler, J., Chen, Z. J., and Maniatis, T. (1997). Activation of the IkappaB alpha kinase complex by MEKK1, a kinase of the JNK pathway. Cell 88, 213–222. Lehner, P. J., and Cresswell, P. (1996). Processing and delivery of peptides presented by MHC class I molecules. Curr. Opin. Immunol. 8, 59–67. Lenschow, D. J., Walunas, T. L., and Bluestone, J. A. (1996). CD28/B7 system of T cell costimulation. Annu. Rev. Immunol. 14, 233–258. Li, E., Beard, C., and Jaenisch, R. (1993). Role for DNA methylation in genomic imprinting [see comments]. Nature 366, 362–365. Liang, H., Nishioka, Y., Reich, C. F., Pisetsky, D. S., and Lipsky, P. E. (1996). Activation of human B cells by phosphorothioate oligodeoxynucleotides. J. Clin. Invest. 98, 1119–1129. Lipford, G. B., Bauer, M., Blank, C., Reiter, R., Wagner, H., and Heeg, K. (1997a). CpGcontaining synthetic oligonucleotides promote B and cytotoxic T cell responses to protein antigen: A new class of vaccine adjuvants. Eur. J. Immunol. 27, 2340–2344. Lipford, G. B., Sparwasser, T., Bauer, M., Zimmerman, S., Koch, E. S., Heeg, K., and Wagner, H. (1997b). Immunostimulatory DNA: Sequence-dependent production of potentially harmful or useful cytokines. Eur. J. Immunol. 27, 3420–3426. Lipford, G. B., Heeg, K., and Wagner, H. (1998). Bacterial DNA as immune cell activator. Trends Microbiol. 6, 496–500. Lipford, G. B., Bendigs, S., Wagner, H., and Heeg, K. (1999a). Bacterial CpG-DNA mediated T cell co-stimulation: Sequence dependence but independence of APCs. Submitted. Lipford, G. B., Sparwasser, T., Zimmermann, S., Heeg, K., and Wagner, H. (1999b). Bacterial CpG-DNA primes draining lymphnodes for protective Th1 cell responsiveness. Submitted. Locksley, R. M., and Louis, J. A. (1992). Immunology of leishmaniasis. Curr. Opin. Immunol. 4, 413–418. Loke, S. L., Stein, C. A., Zhang, X. H., Mori, K., Nakanishi, M., Subasinghe, C., Cohen, J. S., and Neckers, L. M. (1989). Characterization of oligonucleotide transport into living cells. Proc. Natl. Acad. Sci. U.S.A. 86, 3474–3478. Macfarlane, D. E., Manzel, L., and Krieg, A. M. (1997). Unmethylated CpG-containing oligodeoxynucleotides inhibit apoptosis in WEHI 231 B lymphocytes induced by several agents: Evidence for blockade of apoptosis at a distal signalling step. Immunology 91, 586–593. Macfarlane, D. E., and Manzel, L. (1998). Antagonism of immunostimulatory CpGoligodeoxynucleotides by quinacrine, chloroquine, and structurally related compounds. J. Immunol. 160, 1122–1131.

362

HERMANN WAGNER

Maloney, D. G., Grillo-Lopez, A. J., Bodkin, D. J., White, C. A., Liles, T. M., Royston, I., Varns, C., Rosenberg, J., and Levy, R. (1997). IDEC-C2B8: Results of a phase I multipledose trial in patients with relapsed non-Hodgkin’s lymphoma. J. Clin. Oncol. 15, 3266– 3274. Manickan, E., Rouse, R. J., Yu, Z., Wire, W. S., and Rouse, B. T. (1995). Genetic immunization against herpes simplex virus. Protection is mediated by CD4⫹ T lymphocytes. J. Immunol. 155, 259–265. Manickan, E., Karem, K. L., and Rouse, B. T. (1997). DNA vaccines—A modern gimmick or a boon to vaccinology? Crit. Rev. Immunol. 17, 139–154. Matzinger, P. (1994). Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12, 991–1045. McCluskie, M. J., and Davis, H. L. (1998). CpG DNA is a potent enhancer of systemic and mucosal immune response against hepatitis B surface antigen with intranasal administration to mice. J. Immunol. 161, 4463–4466. McIntyre, K. W., Lombard-Gillooly, K., Perez, J. R., Kunsch, C., Sarmiento, U. M., Larigan, J. D., Landreth, K. T., and Narayanan, R. (1993). A sense phosphorothioate oligonucleotide directed to the initiation codon of transcription factor NF-kappa B p65 causes sequencespeciﬁc immune stimulation. Antisense Res. Dev. 3, 309–322. Medzhitov, R., and Janeway, C. A. Jr. (1996). On the semantics of immune recognition. Res. Immunol. 147, 208–214. Medzhitov, R., and Janeway, C. A. Jr. (1997). Innate immunity: Impact on the adaptive immune response. Curr. Opin. Immunol. 9, 4–9. Meehan, R. R., Lewis, J. D., McKay, S., Kleiner, E. L., and Bird, A. P. (1989). Identiﬁcation of a mammalian protein that binds speciﬁcally to DNA containing methylated CpGs. Cell 58, 499–507. Messina, J. P., Gilkeson, G. S., and Pisetsky, D. S. (1991). Stimulation of in vitro murine lymphocyte proliferation by bacterial DNA. J. Immunol. 147, 1759–1764. Messina, J. P., Gilkeson, G. S., and Pisetsky, D. S. (1993). The inﬂuence of DNA structure on the in vitro stimulation of murine lymphocytes by natural and synthetic polynucleotide antigens. Cell Immunol. 147, 148–157. Milligan, J. F., Matteucci, M. D., and Martin, J. C. (1993). Current concepts in antisense drug design. J. Med. Chem. 36, 1923–37. Mojcik, C. F., Gourley, M. F., Klinman, D. M., Krieg, A. M., Gmelig-Meyling, F., and Steinberg, A. D. (1993). Administration of a phosphorothioate oligonucleotide antisense to murine endogenous retroviral MCF env causes immune effects in vivo in a sequencespeciﬁc manner. Clin. Immunol. Immunopathol. 67, 130–136. Moldoveanu, Z., Love-Homan, L., Huang, W. Q., and Krieg, A. M. (1998). CpG DNA, a novel immune enhancer for systemic and mucosal immunization with inﬂuenza virus. Vaccine 16, 1216–1224. Monteith, D. K., Henry, S. P., Howard, R. B., Flournoy, S., Levin, A., Bennett, C. F., and Crooke, S. T. (1997). Immune stimulation—A class effect of phosphorothioate oligodeoxynucleotides in rodents. Anticancer Drug Design 12, 421–432. Mosmann, T. R., and Coffman, R. L. (1989). TH1 and TH2 cells: Different patterns of lymphokine secretion lead to different functional propeties. Annu. Rev. Immunol. 7, 145–173. Nan, X., Campoy, F. J., and Bird, A. (1997). MeCP2 is a transcriptional repressor with abundant binding sites in genomic chromatin. Cell 88, 471–481. Paglia, P., Chiodoni, C., Rodolfo, M., and Colombo, M. P. (1996). Murine dendritic cells loaded in vitro with soluble protein prime cytotoxic T lymphocytes against tumor antigen in vivo [see comments]. J. Exp. Med. 183, 317–322.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

363

Parillo, J. E. (1993). Pathogenetic mechanisms of septic shock. N. Engl. J. Med. 328, 1471– 1477. Pearson, A. M., Rich, A., and Krieger, M. (1993). Polynucleotide binding to macrophage scavenger receptors depends on the formation of base-quartet-stabilized four-stranded helices. J. Biol. Chem. 268, 3546–3554. Perez, J. R., Li, Y., Stein, C. A., Majumder, S., van Oorschot, A., and Narayanan, R. (1994). Sequence-independent induction of Sp1 transcription factor activity by phosphorothioate oligodeoxynucleotides. Proc. Natl. Acad. Sci. U. S. A. 91, 5957–5961. Pisetsky, D. S. (1996). Immune activation by bacterial DNA: A new genetic code. Immunity 5, 303–310. Pisetsky, D. S., and Reich, C. F. (1994). Stimulation of murine lymphocyte proliferation by a phosphorothioate oligonucleotide with antisense activity for herpes simplex virus. Life Sci. 54, 101–107. Pisetsky, D. S., Grudier, J. P., and Gilkeson, G. S. (1990). A role for immunogenic DNA in the pathogenesis of systemic lupus erythematosus. Arthritis Rheum. 33, 153–159. Plescia, O. J., Palczuk, N. C., Cora-Figueroa, E., Mukherjee, A., and Braun, W. (1965). Production of antibodies to soluble RNA (sRNA). Proc. Natl. Acad. Sci. U.S.A. 54, 1281– 1285. Porter, K. R., Kochel, T. J., Wu, S. J., Raviprakash, K., Phillips, I., and Hayes, C. G. (1998). Protective efﬁcacy of a dengue 2 DNA vaccine in mice and the effect of CpG immunostimulatory motifs on antibody responses. Arch. Virol. 143, 997–1003. Prestridge, D. S., and Burks, C. (1993). The density of transcriptional elements in promoter and non-promoter sequences. Human Mol. Genet. 2, 1449–1453. Ramanathan, M., Lantz, M., MacGregor, R. D., Garovoy, M. R., and Hunt, C. A. (1994). Characterization of the oligodeoxynucleotide-mediated inhibition of interferon-gammainduced major histocompatibility complex class I and intercellular adhesion molecule-1. J. Biol. Chem. 269, 24564–24574. Ramasamy, R., Kanagaratnam, R., Misiura, K., Rebowski, G., Amerakoon, R., and Stec, W. J. (1996). Anti-sense oligodeoxynucleoside phosphorothioates nonspeciﬁcally inhibit invasion of red blood cells by malaria parasites. Biochem. Biophys. Res. Commun. 218, 930–933. Ray, A., Sassone-Corsi, P., and Sehgal, P. B. (1989). A multiple cytokine- and second messenger-responsive element in the enhancer of the human interleukin-6 gene: Similarities with c-fos gene regulation. Mol. Cell Biol. 9, 5537–5547. Raz, E., Carson, D. A., Parker, S. E., Parr, T. B., Abai, A. M., Aichinger, G., Gromkowski, S. H., Singh, M., Lew, D., and Yankauckas, M. A. (1994). Intradermal gene immunization: The possible role of DNA uptake in the induction of cellular immunity to viruses. Proc. Natl. Acad. Sci. U.S.A. 91, 9519–9523. Raz, E., Tighe, H., Sato, Y., Corr, M., Dudler, J. A., Roman, M., Swain, S. L., Spiegelberg, H. L., and Carson, D. A. (1996). Preferential induction of a Th1 immune response and inhibition of speciﬁc IgE antibody formation by plasmid DNA immunization. Proc. Natl. Acad. Sci. U.S.A. 93, 5141–5145. Redford, T. W., Yi, A. K., Ward, C. T., and Krieg, A. M. (1998). Cyclosporin A enhances IL-12 production by CpG motifs in bacterial DNA and synthetic oligodeoxynucleotides. J. Immunol. 161, 3930–3935. Rietschel, E. T., and Wagner, H., eds (1996). ‘‘Pathology of Septic Shock.’’ Springer-Verlag, Berlin and New York. Rock, K. L. (1996). A new foreign policy: MHC class I molecules monitor the outside world. Immunol. Today 17, 131–137.

364

HERMANN WAGNER

Roman, M., Martin-Orozco, E., Goodman, J. S., Nguyen, M. D., Sato, Y., Ronaghy, A., Kornbluth, R. S., Richman, D. D., Carson, D. A., and Raz, E. (1997). Immunostimulatory DNA sequences function as T helper-1-promoting adjuvants. Nature Med. 3, 849–854. Saijo, Y., Uchiyama, B., Abe, T., Satoh, K., and Nukiwa, T. (1997). Contiguous four-guanosine sequence in c-myc antisense phosphorothioate oligonucleotides inhibits cell growth on human lung cancer cells: Possible involvement of cell adhesion inhibition. Jpn. J. Cancer Res. 88, 26–33. Sasaki, T., Horiuchi, S., Yamazaki, M., and Yui, S. (1996). Stimulation of macrophage DNA synthesis by polyanionic substances through binding to the macrophage scavenger receptor. Biol. Pharm. Bull. 19, 449–455. Sato, Y., Roman, M., Tighe, H., Lee, D., Corr, M., Nguyen, M. D., Silverman, G. J., Lotz, M., Carson, D. A., and Raz, E. (1996). Immunostimulatory DNA sequences necessary for effective intradermal gene immunization. Science 273, 352–354. Schubbert, R., Lettmann, C., and Doerﬂer, W. (1994). Ingested foreign (phage M13) DNA survives transiently in the gastrointestinal tract and enters the bloodstream of mice. Mol. Gen. Genet. 242, 495–504. Schubbert, R., Renz, D., Schmitz, B., and Doerﬂer, W. (1997). Foreign (M13) DNA ingested by mice reaches peripheral leukocytes, spleen, and liver via the intestinal wall mucosa and can be covalently linked to mouse DNA. Proc. Natl. Acad. Sci. U.S.A. 94, 961–966. Schuler, G., Thurner, B., and Romani, N. (1997). Dendritic cells: From ignored cells to major players in T-cell-mediated immunity. Int. Arch. Allergy Immunol. 112, 317–322. Schwartz, D. A., Quinn, T. J., Thorne, P. S., Sayeed, S., Yi, A. K., and Krieg, A. M. (1997). CpG motifs in bacterial DNA cause inﬂammation in the lower respiratory tract. J. Clin. Invest. 100, 68–73. Segal, B. M., Klinman, D. M., and Shevach, E. M., (1997). Microbial products induce autoimmune disease by an IL-12-dependent pathway. J. Immunol. 158, 5087–5090. Sen, G. C., and Lengyel, P. (1992). The interferon system. A bird’s eye view of its biochemistry. J. Biol. Chem. 267, 5017–5020. Shimada, S., Yano, O., Inoue, H., Kuramoto, E., Fukuda, T., Yamamoto, H., Kataoka, T., and Tokunaga, T. (1985). Antitumor activity of the DNA fraction from Mycobacterium bovis BCG. II. Effects on various syngeneic mouse tumors. J. Natl. Cancer Inst. 74, 681–688. Shimada, S., Yano, O., and Tokunaga, T. (1986). In vivo augmentation of natural killer cell activity with a deoxyribonucleic acid fraction of BCG. Jpn. J. Cancer Res. 77, 808–816. Shirakwa, T., Enomoto, T., Shimazu, S., and Hopkin, J. M. (1997). The inverse association between tuberculin responses and atopic disorder [see comments]. Science 275, 77–79. Sonehara, K., Saito, H., Kuramoto, E., Yamamoto, S., Yamamoto, T., and Tokunaga, T. (1996). Hexamer palindromic oligonucleotides with 5⬘-CG-3⬘ motif(s) induce production of interferon. J. Interferon Cytokine Res. 16, 799–803. Sousa, C. R., Hieny, S., Scharton-Kersten, T., Jankovic, D., Charest, H., Germain, R. N., and Sher, A. (1997). In vivo microbial stimulation induces rapid CD40 ligand–independent production of interleukin 12 by dendritic cells and their redistribution to T cell areas [see comments]. J. Exp. Med. 186, 1819–1829. Sparwasser, T., Miethke, T., Lipford, G., Borschert, K., Hacker, H., Heeg, K., and Wagner, H. (1997a). Bacterial DNA causes septic shock. Nature (London) 386, 336–337. Sparwasser, T., Miethke, T., Lipford, G., Erdmann, A., Hacker, H., Heeg, K., and Wagner, H. (1997b). Macrophages sense pathogens via DNA motifs: Induction of tumor necrosis factor-alpha-mediated shock. Eur. J. Immunol. 27, 1671–1679. Sparwasser, T., Koch, E. S., Vabulas, R. M., Heeg, K., Lipford, G. B., Ellwart, J. W., and Wagner, H. (1998). Bacterial DNA and immunostimulatory CpG oligonucleotides trigger maturation and activation of murine dendritic cells. Eur. J. Immunol. 28, 2045-2054.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

365

Sparwasser, T., Hu¨ltner, L., Koch, E. S., Luz, A., Lipford, G. B., and Wagner, H. (1999a). Immunostimulatory CpG-oligonucleotides cause extramedullary murine hemotopoiesis. J. Immunol. 162, 2368–2374. Sparwasser, T., Vabulas, M., Lipford, G., and Wagner, H. (1999b). Bacterial CpG-DNA driven transient lymphadenopathy is associated with T and B cell inﬂux, chemokine production and sustained propensity for Th1 immune reactivity. Submitted. Staber, F. G., and Metcalf, D. (1980). Cellular and molecular basis of the increased splenic hemopoiesis in mice treated with bacterial cell wall components. Proc. Natl. Acad. Sci. U. S. A. 77, 4322–4325. Stacey, K. J., Sweet, M. J., and Hume, D. A. (1996). Macrophages ingest and are activated by bacterial DNA. J. Immunol. 157, 2116–2122. Starnes, C. O. (1992). Coly’s toxins in perspective. Nature (London) 357, 11–12. Stein, C. A. (1995). Does antisense exist? Nature Med. 1, 1119–1121. Stein, C. A. (1996). Phosphorothioate antisense oligodeoxynucleotides: Questions of speciﬁcity. Trends Biotechnol. 14, 147–149. Stein, C. A., and Cheng, Y. C. (1993). Antisense oligonucleotides as therapeutic agents—is the bullet really magical? Science 261, 1004–1012. Stein, C. A., and Cohen, J. S. (1988). Oligodeoxynucleotides as inhibitors of gene expression: A review. Cancer Res. 48, 2659–2668. Stein, C. A., Tonkinson, J. L., Zhang, L. M., Yakubov, L., Gervasoni, J., Taub, R., and Rotenberg, S. A. (1993) Dynamics of the internalization of phosphodiester oligodeoxynucleotides in HL60 cells. Biochemistry 32, 4855–4861. Steinman, R. M., Pack, M., and Inaba, K. (1997). Dendritic cells in the T-cell areas of lymphoid organs. Immunol. Rev. 156, 25–37. Stenger, S., Donhauser, N., Thuring, H., Rollinghoff, M., and Bogdan, C. (1996). Reactivation of latent leishmaniasis by inhibition of inducible nitric oxide synthase. J. Exp. Med. 183, 1501–1514. Sun, S., Cai, Z., Langlade-Demoyen, P., Kosaka, H., Brunmark, A., Jackson, M. R., Peterson, P. A., and Sprent, J. (1996). Dual function of Drosophila cells as APCs for naive CD8⫹ T cells: Implications for tumor immunotherapy. Immunity 4, 555–564. Sun, S., Beard, C., Jaenisch, R., Jones, P., and Sprent, J. (1997). Mitogenicity of DNA from different organisms for murine B cells. J. Immunol. 159, 3119–3125. Sun, S., Kishimoto, H., and Sprent, J. (1998). DNA as an adjuvant: Capacity of insect DNA and synthetic oligodeoxynucleotides to augment T cell responses to speciﬁc antigen. J. Exp. Med. 187, 1145–1150. Sweet, M. J., Stacey, K. J., Kakuda, D. K., Markovich, D., and Hume, D. A. (1998). IFNgamma primes macrophage responses to bacterial DNA. J. Interferon Cytokine Res. 18, 263-271. Takekawa, M., and Saito, H. (1998). A family of stress-inducible GADD45-like proteins mediate activation of the stress-responsive MTK1/MEKK4 MAPKKK. Cell 95, 521–530. Tanaka, T., Chu, C. C., and Paul, W. E. (1992). An antisense oligonucleotide complementary to a sequence in I gamma 2b increases gamma 2b germline transcripts, stimulates B cell DNA synthesis, and inhibits immunoglobulin secretion. J. Exp. Med. 175, 597–607. Tang, D. C., DeVit, M., and Johnston, S. A. (1992). Genetic immunization is a simple method for eliciting an immune response. Nature (London) 356, 152–154. Terada, K., Okuhara, E., and Kawarada, Y. (1991). Antigen DNA isolated from immune complexes in plasma of patients with systemic lupus erythematosus hybridizes with the Escherichia coli lac Z gene. Clin. Exp. Immunol. 85, 66–69. Terry, W. D., and Yamamura, Y. (1979). ‘‘Immunobiology and Immunotherapy of Cancer: Developments in Immunology.’’ Elsevier/North-Holland, Amsterdam.

366

HERMANN WAGNER

Thompson, C. B. (1995). New insights into V(D)J recombination and its role in the evolution of the immune system. Immunity 3, 531–539. Tokunaga, T., Yamamoto, H., Shimada, S., Abe, H., Fukuda, T., Fujisawa, Y., Furutani, Y., Yano, O., Kataoka, T., and Sudo, T. (1984). Antitumor activity of deoxyribonucleic acid fraction from Mycobacterium bovis BCG. I. Isolation, physicochemical characterization, and antitumor activity. J. Natl. Cancer Inst. 72, 955–962. Tokunaga, T., Yano, O., Kuramoto, E., Kimura, Y., Yamamoto, T., Kataoka, T., and Yamamoto, S. (1992) Synthetic oligonucleotides with particular base sequences from the cDNA encoding proteins of Mycobacterium bovis BCG induce interferons and activate natural killer cells. Microbiol. Immunol. 36, 55–66. Tonkinson, J. L., and Stein, C. A. (1994). Patterns of intracellular compartmentalization, trafﬁcking and acidiﬁcation of 5⬘-ﬂuorescein labeled phosphodiester and phosphorothioate oligodeoxynucleotides in HL60 cells. Nucleic Acids Res. 22, 4268–4275. Tough, D. F., Borrow, P., and Sprent, J. (1996). Induction of bystander T cell proliferation by viruses and type I interferon in vivo [see comments]. Science 272, 1947–1950. Tough, D. F., Sun, S., and Sprent, J. (1997). T cell stimulation in vivo by lipopolysaccharide (LPS). J. Exp. Med. 185, 2089–2094. Trinchieri, G. (1998). Interleukin-12; A cytokine at the interface of inﬂammation and immunity. Adv. Immunol. 70, 83–243. Tsubata, T., Wu, J., and Honjo, T. (1993). B-cell apoptosis induced by antigen receptor crosslinking is blocked by a T-cell signal through CD40. Nature (London) 364, 645–648. Ulevitch, R. J., and Tobias, P. S., (1995). Receptor-dependent mechanisms of cell stimulation by bacterial endotoxin. Annu. Rev. Immunol. 13, 437–457. Unanue, E. R. (1997). Inter-relationship among macrophages, natural killer cells and neutrophils in early stages of Listeria resistance. Curr. Opin. Immunol. 9, 35–43. Vaerman, J. L., Moureau, P., Deldime, F., Lewalle, P., Lammineur, C., Morschhauser, F., and Martiat, P. (1997). Antisense oligodeoxyribonucleotides suppress hematologic cell growth through stepwise release of deoxyribonucleotides. Blood 90, 331–339. Vieillard, V., Cremer, I., Lauret, E., Rozenbaum, W., Debre, P., Autran, B., and De Maeyer, E. (1997). Interferon beta transduction of peripheral blood lymphocytes from HIVinfected donors increases Th1-type cytokine production and improves the proliferative response to recall antigens. Proc. Natl. Acad. Sci. U.S.A. 94, 11595–11600. von Mutius, E., Fritzsch, C., Weiland, S. K., Roll, G., and Magnussen, H. (1992). Prevalence of asthma and allergic disorders among children in united Germany: A descriptive comparison. BMJ 305, 1395–1399. Wagner, R. (1994). Gene inhibition using antisense oligodeoxynucleotides. Nature (London) 372, 333–335. Weeratna, R., Brazolot Millan, C. L., Krieg, A. M., and Davis, H. L. (1998). Reduction of antigen expression from DNA vaccines by coadministered oligodeoxynucleotides. Antisense Nucleic Acid Drug Dev. 8, 351–356. Weiner, G. J., Liu, H. M., Wooldridge, J. E., Dahle, C. E., and Kreig, A. M. (1997). Immunostimulatory oligodeoxynucleotides containing the CpG motif are effective as immune adjuvants in tumor antigen immunization. Proc. Natl. Acad. Sci. U.S.A. 94, 10833– 10837. White, J. R., Gordon-Smith, E. C., and Rutherford, T. R. (1996). Phosphorothioate-capped antisense oligonucleotides to Ras GAP inhibit cell proliferation and trigger apoptosis but fail to downregulate GAP gene expression. Biochem. Biophys. Res. Commun. 227, 118–124. Williamson, J. R., Raghuraman, M. K., and Cech, T. R. (1989) Monovalent cation-induced structure of telomeric DNA: The G-quarter model. Cell 59, 871–880.

IMMUNE CELL ACTIVATION BY BACTERIAL CpG DNA

367

Wooldridge, J. E., Ballas, Z., Krieg, A. M., and Weiner, G. J. (1997). Immunostimulatory oligodeoxynucleotides containing CpG motifs enhance the efﬁcacy of monoclonal antibody therapy of lymphoma. Blood 89, 2994–2998. Yamamoto, S., Yamamoto, T., Kataoka, T., Kuramoto, E., Yano, O., and Tokunaga, T. (1992a). Unique palindromic sequences in synthetic oligonucleotides are required to induce IFN and augment IFN-mediated natural killer activity. J. Immunol. 148, 4072– 4076. Yamamoto, S., Yamamoto, T., Shimada, S., Kuramoto, E., Yano, O., Kataoka, T., and Tokunaga, T. (1992b). DNA from bacteria, but not from vertebrates, induces interferons, activates natural killer cells and inhibits tumor growth. Microbiol. Immunol. 36, 983– 997. Yamamoto, T., Yamamoto, S., Kataoka, T., Komuro, K., Kohase, M., and Tokunaga, T. (1994a). Synthetic oligonucleotides with certain palindromes stimulate interferon production of human peripheral blood lymphocytes in vitro. Jpn. J. Cancer Res. 85, 775– 779. Yamamoto, T., Yamamoto, S., Kataoka, T., and Tokunaga, T. (1994b). Lipofection of synthetic oligodeoxyribonucleotide having a palindromic sequence of AACGTT to murine splenocytes enhances interferon production and natural killer activity. Microbiol. Immunol. 38, 831–836. Yang, R. B., Mark, M. R., Gray, A., Huang, A., Xie, M. H., Zhang, M., Goddard, A., Wood, W. I., Gurney, A. L., and Godowski, P. J. (1998). Toll-like receptor-2 mediates lipopolysaccharide-induced cellular signalling [see comments]. Nature (London) 395, 284–288. Yi, A. K., and Krieg, A. M. (1998a). CpG DNA rescue from anti-IgM-induced WEHI-231 B lymphoma apoptosis via modulation of I kappa B alpha and I kappa B beta and sustained activation of nuclear factor-kappa B/c-Rel. J. Immunol. 160, 1240–1245. Yi, A. K., and Krieg, A. M. (1998b). Rapid induction of mitogen-activated protein kinases by immune stimulatory CpG DNA. J. Immunol. 161, 4493-4497. Yi, A. K., Chace, J. H., Cowdery, J. S., and Krieg, A. M. (1996a). IFN-gamma promotes IL-6 and IgM secretion in response to CpG motifs in bacterial DNA and oligodeoxynucleotides. J. Immunol. 156, 558–564. Yi, A. K., Hornbeck, P., Lafrenz, D. E., and Krieg, A. M. (1996b). CpG DNA rescue of murine B lymphoma cells from anti-IgM-induced growth arrest and programmed cell death is associated with increased expression of c-myc and bcI-xL. J. Immunol. 157, 4918– 4925. Yi, A. K., Klinman, D. M., Martin, T. L., Matson, S., and Krieg, A. M. (1996c). Rapid immune activation by CpG motifs in bacterial DNA. Systemic induction of IL-6 transcription through an antioxidant-sensitive pathway. J. Immunol. 157, 5394–5402. Yi, A. K., Chang, M., Peckham, D. W., Krieg, A. M., and Ashman, R. F. (1998a). CpG oligodeoxyribonucleotides rescue mature spleen B cells from spontaneous apoptosis and promote cell cycle entry. J. Immunol. 160, 5898–5906. Yi, A. K., Tuetken, R., Redford, T., Waldschmidt, M., Kirsch, J., and Krieg, A. M. (1998b) CpG motifs in bacterial DNA activate leukocytes through the pH-dependent generation of reactive oxygen species. J. Immunol. 160, 4755–4761. Yoshimori, T., Yamamoto, A., Moriyama, Y., Futai, M., and Tashiro, Y. (1991). Baﬁlomycin A1, a speciﬁc inhibitor of vacuolar-type H(⫹)-ATPase, inhibits acidiﬁcation and protein degradation in Iysosomes of cultured cells. J. Biol. Chem. 266, 17707–17712. Zbar, B., Bernstein, I. D., and Rapp, H. J. (1971). Suppression of tumor growth at the site of infection with living Bacillus Calmette–Guerin. J. Natl. Cancer Inst. 46, 831–839.

368

HERMANN WAGNER

Zbar, B., Ribi, E., Meyer, T., Azuma, I., and Rapp, H. J. (1974). Immunotherapy of cancer: Regression of established intradermal tumors after intralesional injection of mycobacterial cell walls attached to oil droplets. J. Natl. Cancer Inst. 52, 1571–1577. Zhang, X., Sun, S., Hwang, I., Tough, D. F., and Sprent, J. (1998). Potent and selective stimulation of memory-phenotype CD8⫹ T cells in vivo by IL-15. Immunity 8, 591– 599. Zimmermann, S., Egeter, O., Hausmann, S., Lipford, G. B., Rocken, M., Wagner, H., and Heeg, K. (1998). CpG oligodeoxynucleotides trigger protective and curative Th 1 responses in lethal murine leishmaniasis. J. Immunol. 160, 3627–3630.

ADVANCES IN IMMUNOLOGY, VOL. 73

Neutrophil-Derived Proteins: Selling Cytokines by the Pound MARCO ANTONIO CASSATELLA Department of Pathology, Section of General Pathology, Faculty of Medicine, University of Verona, 37134 Verona, Italy

I. Introduction

Neutrophils act as the ﬁrst line of defense against invading bacteria and other microorganisms. Traditionally, mature neutrophils have been considered to be terminally differentiated cells lacking the ability to synthesize proteins. In recent years, however, it has become increasingly clear that this view is outdated. Studies conducted in a large number of laboratories have indeed shown that neutrophils not only synthesize numerous proteins that directly participate in their effector functions, but that they can also produce a variety of pro- and antiinﬂammatory polypeptides. For instance, there now exists compelling evidence that neutrophils can release a number of cytokines, chemokines, growth factors, and interferons, both in vitro and in vivo. Furthermore, these studies have demonstrated that the interaction of neutrophils with a given agonist not only results in a characteristic pattern of cytokine release, but also that this response can be modulated by immunoregulatory cytokines such as interleukin-10 (IL10) and interferon-웂 (IFN-웂). In view of the broad spectrum of biological activities exerted by cytokines, it can be reasonably inferred that neutrophils not only play an important role in eliciting and sustaining inﬂammation, but may also signiﬁcantly contribute to the regulation of immune reactions. These considerations alone make it clear that the current conception of the roles played by neutrophils in pathophysiological processes is in dire need of being redeﬁned. It is my purpose to summarize here the current knowledge on the production of cytokines by neutrophils in vitro and in vivo, particularly molecular regulation and other biological and pathophysiological aspects. II. General Features of Cytokine Production by Neutrophils

Polymorphonuclear neutrophil granulocytes (PMNs) are the predominant inﬁltrating cell type present in the cellular phase of the acute inﬂammatory response, and therefore act as a ﬁrst line of defense against invading bacteria and other microorganisms (1). Although monocytes and lymphocytes are thought to represent a major source of cytokines among blood cells (2), it is now evident that cytokines are generated by many other cell 369

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

370

MARCO ANTONIO CASSATELLA

types as well, including neutrophils (3, 4). Most of the studies addressing neutrophil cytokine release have been conducted in vitro, but the validity of their ﬁndings has nonetheless often been conﬁrmed in vivo, principally through the use of molecular biology techniques or other sensitive approaches such as immunohistochemistry (IH) and in situ hybridization (ISH) (5). Thus, the ability of neutrophils to generate cytokines points to new mechanisms by which PMNs, might inﬂuence multiple aspects of the inﬂammatory and immune responses. In view of the fact that neutrophils usually represent the ﬁrst cell type to migrate toward inﬂamed tissue, and given that most neutrophil responses are stimulus speciﬁc, these cells might play a key role in determining (or at least inﬂuencing) the evolution of subsequent host responses. Table I lists the cytokines that, to date, have been shown to be released by PMNs in vitro, either constitutively or following appropriate stimulation. At ﬁrst glance, the ability of PMNs to produce such a variety of cytokines TABLE I CYTOKINES EXPRESSED BY NEUTROPHILS in Vitro Proinﬂammatory cytokines TNF-움 IL-1움 IL-1웁 IL-12 IFN-움 IFN-웂a IL-6a Antiinﬂammatory cytokines IL-1RA TGF-웁 C-X-C chemokines IL-8 GRO-움 GRO-웁b CINC-1 CINC-2움 IP-10 MIG I-TACb C-C chemokines MIP-1움 MIP-1웁 MCP-1a a b

Requires deﬁnitive conﬁrmation. mRNA only.

Growth factors G-CSF M-CSFa GM-CSFa IL-3a SCFa,b Angiogenic or ﬁbrogenic factors VEGF TGF-움 HGF LDGF CEMF Others Fas ligand CD30 ligand Oncostatin M GDF NGFb BDNFb NT-4b

NEUTROPHIL-DERIVED PROTEINS

371

readily suggests that these cells might regulate not only the inﬂammatory and immune responses, but also other processes, such as hematopoiesis, wound healing, angiogenesis, and antiviral defense. Several laboratories have generated convincing molecular evidence that PMNs can express proinﬂammatory cytokines such as tumor necrosis factor-움 (TNF-움) (6–8), IL-1움, IL-1웁 (9–11), and IL-12 (12, 13); antiinﬂammatory cytokines such as the IL-1 receptor antagonist (IL-1RA) (14, 15) and transforming growth factor-웁 (TGF-웁)(16–18); chemokines such as IL-8 (19, 20), growthrelated gene product-움 (GRO-움) (21, 22), macrophage inﬂammatory protein-1움 (MIP-1움) and MIP-1웁 (23–25), and cytokine-induced neutrophil chemoattractants (CINCs) (26, 27); and other cytokines such as IFN움 (28, 29), granulocyte colony-stimulating factor (G-CSF) (30, 31), Fas ligand (FasL) (32, 33), CD30 ligand (CD30L) (34, 35), vascular endothelial growth factor (VEGF) (36–38), and hepatocyte growth factor (HGF) (39, 40). In addition, PMNs have been reported, albeit in single instances, to express and/or release macrophage CSF (M-CSF), IL-3, GRO-웁, interferon-웂 inducible protein-10 (IP-10), monokine induced by IFN-웂 (MIG), TGF-움, oncostatin (OSM), neurotrophins, and other molecules; these studies therefore await further conﬁrmation (4, 5). Finally, conﬂicting data exist in the literature concerning the issue of whether IL-6, monocyte chemotactic protein-1 (MCP-1), granulocyte–macrophage CSF (GMCSF), stem cell factor (SCF), and IFN-웂 expression can be induced in human neutrophils (4, 5). In any instance, the very fact that neutrophils can synthesize, store, and release such a wide range of cytokines stresses the necessity to reconsider the general role of PMNs in physiopathology. Though the production of cytokines by activated neutrophils is striking in its diversity, the extent to which neutrophils generate cytokines is generally limited (at least in vitro), especially when compared to the scale of this response in peripheral mononuclear cells (MNCs) (4, 41). In this respect, it must be emphasized that despite their generally lower synthetic capacity, granulocytes nevertheless constitute the majority of inﬁltrating cells in inﬂamed tissues, and may therefore represent an important source of cytokines under such conditions. Depending on the cytokine, neutrophils possess 10–20 times less RNA per cell than do monocytes or lymphocytes, and synthesize 10- to 300-fold less cytokine than do monocytes on an individual basis (4, 5). It follows that if one is to investigate whether neutrophils produce a given cytokine (or other protein), it is absolutely mandatory to work with highly puriﬁed PMN populations (⬎ 99.5%). It is also highly recommended to exclude the possibility of a prestimulation of PMNs during their isolation procedures (42, 43), which may be driven, for instance, by contamination of reagents, solutions, or labware with trace levels of endotoxin (44), or by the use of inappropriate methods of erythro-

372

MARCO ANTONIO CASSATELLA

cyte lysis. In our experiments, as well as in those of others (45), more than 98% of freshly puriﬁed PMNs are usually positive for CD62L, a membranebound antigen that is rapidly released on neutrophil stimulation (46). That indicates that the isolation procedures and hypotonic lysis do not stimulate our neutrophil preparations to any signiﬁcant extent. In addition, the use of endotoxin-free serum and culture media, the veriﬁcation of endotoxin levels in all solutions by the Limulus amebocyte lysate (LAL) assay, and checking whether polymyxin B sulfate inhibits the effects of the stimuli under investigation (47) all represent some of the common measures taken to keep LPS contamination under control. Although sometimes tedious, these routine procedures are very helpful to avoid artifactual effects of some stimuli, especially because the latter have sometimes been reported in the literature, and are probably often at the root of controversial observations. The production of individual cytokines by neutrophils is inﬂuenced to a great extent by the stimulatory conditions. The most widely used stimuli are bacterial LPS and cytokines, but known inducers of cytokine generation in neutrophils belong also to other classes of molecules, including chemotactic factors [formyl-methionyl-leucyl-phenylalanine (fMLP), leukotriene B4 (LTB4), platelet-activating factor (PAF), and the complement component C5a], neuropeptides (substance P), phagocytic particles, and microorganisms such as fungi, viruses, and bacteria (5). In general, a given stimulus not only affects the magnitude and kinetics of cytokine release, but also inﬂuences the pattern of production as well. We have focused most of our investigations on the effects of LPS, fMLP, TNF-움 and Saccharomyces cerevisiae opsonized with IgG (Y-IgG) (8, 19, 21). This led to the realization that Y-IgG appears to be a potent trigger for the release of only some proinﬂammatory cytokines (TNF-움, IL-8, and GRO-움), whereas LPS and TNF-움 induce both proinﬂammatory and antiinﬂammatory (for example IL-1RA) cytokines (4, 5, 15). In contrast, fMLP (and chemoattractants in general) seem to trigger only a small, transient release of IL-8 and GRO움, but apparently not that of any other cytokine (4, 5). Similarly, IL-4 and IL-13 induce only IL-1RA (48, 49). Moreover, the production of cytokines by neutrophils can be a ﬁnely tuned process, as illustrated by the synthesis and release of IL-12, IP-10, or MIG, which occur only if neutrophils are costimulated with at least two agonists. One such combination is the IFN웂 plus LPS tandem (12). The induction of a given cytokine production in PMNs is usually preceded by an increased accumulation of the related mRNA transcripts. This can be detected by performing classical Northern blot analyses, or, in ribonuclease protection assays (RPAs), by reverse transcriptase– polymerase chain reaction (RT–PCR) or by in situ hybridization. Northern

NEUTROPHIL-DERIVED PROTEINS

373

blotting or RPAs have the advantage of making differences in cytokine mRNA expression readily visible, and when coupled with densitometric analysis or phosphorimaging, these approaches can be semiquantitative. In addition, the use of Northern blotting and of related techniques can yield important insights into the molecular mechanisms regulating cytokine gene expression in PMNs. Although quantitative RT–PCR is also possible, the exteme sensitivity of this technique entails the risk of amplifying cytokine mRNA from very few (⬍ 0.5%) contaminating monocytes, lymphocytes, or eosinophils (50). From a general point of view, however, Northern blotting, RPAs, and especially RT–PCR must be performed with adequate controls to avoid false positive results deriving from contaminating cells. To illustrate the necessity to carry out such controls, sufﬁce it to recall that monocytes and lymphocytes possess 10–20 times more RNA per cell than do neutrophils; thus, contamination of a neutrophil preparation with only 1% monocytes can easily translate into a total RNA preparation containing 20–25% monocyte-derived RNA. Based on reports from our group (8) as well as from other laboratories (51, 52), the presence of contaminating monocyte mRNA in Northern analyses can be assessed by using IL-6 cDNA as a probe. The combined absence of detectable IL-6 mRNA therefore indicates, in my opinion, that a given PMN population must be reasonably free from contaminating monocytes. Similarly, RANTES cDNA can be used to monitor eosinophil contamination (our unpublished observations). Finally, the production of cytokines by neutrophils can be modulated by immunomodulatory cytokines such as IFN-웂, IL-10, IL-4, and IL-13, raising the possibility that T helper type 1 (Th1) and Th2 lymphocytes (53) may selectively inﬂuence the production of cytokines by PMNs. Interestingly, depending on the stimulus used, the effects of these regulatory cytokines are different. For example, IFN-웂 potentiates the late release of IL-8 or GRO-움 by LPS-stimulated PMNs, but inhibits that induced by Y-IgG or fMLP (4). Similarly, IL-10 enhances the production of IL-1RA by LPS-stimulated PMNs, but does not inﬂuence that induced by Y-IgG or fMLP. Finally, though IL-10 inhibits the production of GRO-움 in LPSstimulated PMNs, it enhances it in response to TNF-움 (21). A condensed description on the experimental conditions used to induce the production of the various cytokines by neutrophils in vitro follows below. Not included in this review is eosinophil- and basophil-derived cytokine production. III. Production of Speciﬁc Cytokines by Neutrophils in Vitro

A. CHEMOKINES Chemokines represent a group of chemotactic cytokines whose importance in inﬂammatory processes is best illustrated by their ability to speciﬁ-

374

MARCO ANTONIO CASSATELLA

cally recruit discrete leukocyte populations. These polypeptides are usually small, with molecular masses in the range of 8 to 12 kDa, and they share 20 to 70% homology in amino acid sequences (54–56). Chemokines have been classiﬁed into four closely related subfamilies on the basis of the relative positions of the ﬁrst two cysteine residues, but only two of these subfamilies have been extensively characterized: the ‘‘C-X-C’’ and the ‘‘C-C’’ chemokines (54–56). The C-X-C subfamily includes IL-8 (the prototype), GRO-움, GRO-웁, GRO-웂, epithelial cell-derived and neutrophil-activating 78-amino acid peptide (ENA-78), MIP-2, CINC, IFN-웂-inducible protein of 10 kDa (IP-10), MIG, and others; these chemokines predominantly exert stimulatory and chemotactic activities toward neutrophils (54–57). The C-C subfamily includes monocyte chemotactic proteins, MIP-1움, MIP-1웁, MIP-1웂, and RANTES; these chemokines are predominantly chemotactic for monocytes, eosinophils, basophils, and certain T lymphocyte subsets (57). Two additional molecules, which are likely members of novel groups of chemokines, the C and CX3C subfamilies, have also been identiﬁed (54–57). To the former belongs lymphotactin, a potent attractant for lymphocytes, whereas to the latter belongs fractalkine or neurotactin, which is an integral membrane protein with a chemokine domain at its N terminus (54–56). It has become quite obvious that chemokines and their receptors are expressed by a wide variety of nonhematopoietic cells, and that chemokine functions extend well beyond leukocyte physiology. For instance, the connections among chemokines, their receptors, and HIV infection broaden the previously narrow focus on chemokines as mere chemoattractants (58). Among the cell types that generate chemokines, neutrophils have been convincingly and repeatedly shown to secrete both C-X-C and C-C chemokines. 1. C-X-C Chemokines a. Interleukin-8. Interleukin-8 is a key mediator for the recruitment of circulating neutrophils. This chemokine is expressed in response to inﬂammatory stimuli, and is secreted by a variety of cell types, including lymphocytes, epithelial cells, keratinocytes, ﬁbroblasts, endothelial cells (ECs), smooth muscle cells (57), and neutrophils. In the latter instance, IL-8 is one of the most abundantly secreted (and most extensively studied) cytokines produced by neutrophils. Conversely, neutrophils represent the primary cellular target for IL-8, to which they respond by chemotaxis, release of granule enzymes, respiratory burst, up-regulation of CR1 and CD11/CD18 surface expression, increased adherence to unstimulated endothelial cells, and transmigration across the endothelium (57). Other biological properties of IL-8 include its chemotactic activity toward T

NEUTROPHIL-DERIVED PROTEINS

375

lymphocytes, albeit less potently than in the case of neutrophils (57), the stimulation of histamine release from basophils, and the ability to promote angiogenesis (59). Subsequent to the early demonstrations that PMNs can release relevant amounts of biologically active IL-8 into culture supernatants after phagocytosis of Y-IgG (19), or after stimulation with LPS (19, 20), a large number of studies followed and extended these seminal ﬁndings. Strieter and colleagues (20) showed that PMNs (⬎99% pure, plated at 5 ⫻ 106 ml) expressed signiﬁcant steady-state levels of IL-8 mRNA after 4 hr of stimulation with LPS (100 ng/ml), and released about 100 pg/ml of IL-8 into culture supernatants after 24 hr. We obtained very similar results with LPS (1 애g/ml), and further showed that phagocytosis of Y-IgG represented a much more potent stimulus for the extracellular production of IL-8 release (from 10 to 50 ng/ml/18 hr 107 PMN) (19). In addition, we provided clear molecular and immunological evidence that the IL-8 recovered in the PMN cultures was not due to contaminating monocytes, and that it was biologically active (19). It soon emerged that fMLP (60–64), as well as other neutrophil chemotactic factors such as C5a (60, 65), PAF (60), and LTB4 (66), triggered the release of substantial amounts of IL-8 in PMNs. Chemoattractant-elicited IL-8 production was shown to be preceded by an enhanced expression of IL-8 mRNA, and to be dependent on de novo protein synthesis (60). Optimal release of IL-8 was observed using 10 nM of fMLP (a concentration that fails to stimulate a respiratory burst under the experimental conditions used), and was potentiated from two- to sixfold if PMNs were brieﬂy pretreated with cytochalasin B (60). Furthermore, secretion of IL-8 in response to fMLP was found to be transient, because maximal production was observed by 2–3 hr and returned to basal levels thereafter (20, 64), contrary to the sustained release of IL-8 induced by Y-IgG or LPS (19). It is therefore possible that IL-8, once produced in response to fMLP, is taken up by cell surface receptors or is degraded by proteolytic enzymes simultaneously released by PMNs (67). However, a more likely explanation might stem from the fact that fMLP induces the release of neither IL-1웁 nor TNF-움, two cytokines that have been shown to amplify the release of IL-8 from LPS-treated neutrophils in an autocrine/paracrine manner (4). In contrast to the above ﬁndings, other investigators failed to observe any signiﬁcant extracellular antigenic IL-8 (relative to unstimulated controls) in neutrophils stimulated with graded concentrations of C5a, fMLP, or LTB4 (68). Similar results were also reported when fMLP and PAF were used at doses higher than 10–100 nM (69). In the latter two studies, however, it must be pointed out that extracellular IL-8 was measured in supernatants harvested 24 hr poststimulation; as explained above, the

376

MARCO ANTONIO CASSATELLA

24-hr time point is far from optimal to detect IL-8 in supernatants from fMLP-stimulated PMNs (60). Nevertheless, Strieter et al. (68) reported that if neutrophils were exposed to chemotactic agonists in the presence of LPS, the production of IL-8 was synergistically elevated. The fact that chemotactic factors for neutrophils can induce the production of another chemoattractant (i.e., IL-8) is intriguing. IL-8 production induced by chemotactic factors could represent a feedback mechanism whereby greater numbers of PMNs are attracted to an inﬂammatory site, ensuring a persistent inﬂux of PMNs that perpetuate the inﬂammatory reaction. The release of IL-8 induced by LPS can be inhibited by FUT-175 (a protease inhibitor) (70), and by 3-morpholinosydnonimine (SIN-1) (a combined nitric oxide–superoxide donor) (71), but is additively enhanced either by 1, 2, 3, 4-oxatriazolium 5-amino chloride (GEA-3162) (a nitric oxide donor) (71), or by a combination of SNP and NAC. Similarly, IL-8 secretion can be enhanced by S-nitrosoglutathione, but not by dibutyrylcGMP (72). The latter results have important implications for patients with acute lung injury, who are commonly treated with inhaled nitric oxide, because enhanced production of neutrophil-derived IL-8 may represent a novel way by which nitric oxide could regulate chemotactic responses. The production of IL-8 induced by LPS is also potentiated by PMN binding to solid-phase ﬁbrinogen (73), but not to plastic surfaces precoated either with ﬁbronectin (Fn), a ligand of the integrin receptor 움5웁1, or with laminin (Ln), a ligand of the integrin receptor 움6웁1 (74). Interestingly, under the latter conditions, untreated PMNs released high levels of IL-8 compared to cells adhered to plastic (74). It is noteworthy that although most of the studies described above were performed using LPS derived from Escherichia coli, other LPS serotypes have been shown to stimulate PMNs to release IL-8. For instance, endotoxins from pathogenic bacteria of periodontitis lesions (such as Actinobacillus actinomycetemcomitans, Fusobacterium nucleatum, and, to lesser extent, Porphyromonas gingivalis and Capnocytophaga ochracea) have been reported to promote IL-8 secretion (75). The effect of P. gingivalis was preceded by a stimulation of NF␬B- and AP-1-binding activities (76), in keeping with the ability of this bacterium to concurrently induce the production of IL-1 and TNF-움 in neutrophils (75). LPS, Y-IgG phagocytosis, and chemotactic factors are not the only stimuli that induce the production of IL-8 by neutrophils. TNF-움 is another agent that stimulates IL-8 mRNA and secretion in a time- and dose-dependent manner (25, 68, 77, 78), at much higher levels than IL-1웁 (79, 80). However, IL-1웁 dramatically synergizes with TNF-움, if used at the same concentrations produced by neutrophils in response to LPS (79). Other studies identiﬁed GM-CSF as a very potent inducer of IL-8 by human PMNs (25,

NEUTROPHIL-DERIVED PROTEINS

377

81–83), and this activity of GM-CSF is potentiated by the immunosuppressive drugs, cyclosporin A (CsA) and rapamycin, but not by FK506 (84). Some investigators have reported that IL-2 (83, 85, 86), IL-13 (87), and TGF-웁 (88) induce IL-8 mRNA expression and secretion by PMNs, but others found these cytokines to be ineffective (see 89, 90, for IL-2; 50, for IL-13; 51, for TGF-웁). In contrast, there is no doubt that IL-15 induces IL-8 mRNA expression and secretion by neutrophils (86, 90), as it also does in monocytes (91). IL-15 is a recently discovered cytokine produced by a wide range of different cell types including ﬁbroblasts, keratinocytes, endothelial cells, and macrophages in response to lipopolysaccharide or microbial infection (92). IL-15 has been shown to act on various cells of the immune system, including T lymphocytes, B lymphocytes, NK cells, and, more recently, peripheral blood neutrophils (89). In PMNs IL-15 was observed to induce cytoskeletal rearrangements, to enhance phagocytosis, to increase the synthesis of several cellular proteins, and to delay apoptosis (89). Under identical conditions, however, IL-2 failed to affect any of these responses in neutrophils (89). The range of stimuli shown to induce the production of IL-8 by PMNs covers many more biological agents and continues to expand (Table II). For example, the inﬂammatory microcrystals monosodium urate (MSU) and calcium pyrophosphate dihydrate (CPPD), which are major mediators of gout and pseudogout, respectively, both increase the secretion of IL-8 by neutrophils, yet have no effect on that of MIP-1움 (25). Interestingly, the presence of MSU and CPPD synergistically enhances the production of IL-8 induced by TNF-움 and GM-CSF, but completely inhibits the secretion of MIP-1움 induced by TNF-움 (25). PMNs stimulated with phorbol myristate acetate (PMA) (61, 78, 93), staurosporine (a nonspeciﬁc inhibitor protein kinase C and of other kinases) (93), thapsigargin (an inhibitor of the endoplasmic Ca2⫹-sequestering ATPase) (94), and A23187 (61, 63, 69) and ionomycin (94) (two calcium ionophores) express high levels of IL-8 mRNA and release IL-8 in a time- and dose-dependent manner. IL-8 production induced by A23187 is partially inhibited by a selective PAF antagonist (69), in keeping with the known ability of A23187 to stimulate the synthesis of PAF in neutrophils (95). That PAF receptor antagonists can inhibit IL-8 release from neutrophils has been conﬁrmed in another recent study (96). Substance P, one of the main mediators of neurogenic inﬂammation, not only directly stimulates the release of IL-8 from human neutrophils, but also enhances the effect of TNF-움 and fMLP (97). Human or bovine lactoferrin (LF), iron-glycoproteins having fundamental antimicrobial activities, and lactoferricin (LFcin), a peptide derived from the N-terminal region of LF, all have been shown to stimulate the release IL-8 from

378

MARCO ANTONIO CASSATELLA

TABLE II AGENTS ABLE TO TRIGGER THE PRODUCTION OF IL-8 Cytokines and growth factors TNF-움 IL-1웁 IL-15 GM-CSF Thrombopoietin IL-2a IL-13a TGF-웁a Chemoattractants fMLP C5a PAF Leukotriene B4 Substance P Calcium ionophores Ionomycin A23187 Surface molecules Anti-CD30L antibodies Anti-ICAM3 antibodies Anti-PSGL-1 antibodies Enzymes Lactoferrin Lactoferricin Particulate agents Calcium pyrophosphate dihydrate microcrystals Monosodium urate microcrystals

a b

Requires deﬁnitive conﬁrmation. mRNA only.

BY

HUMAN NEUTROPHILS

Bacteria and related products LPS from Escherichia coli Listeria monocytogenes Yersinia enterocolitica Salmonalla typhimurium CPC from Bacillus fragilis Staphylococcus aureus, S. epidermidis Panton–Valentine leukocidin from S. aureus Heat-killed streptococci Erythrogenic toxin A from Streptococcus pyogenes Streptococcus pneumoniae Alveolysin from Bacillus alvei Pseudomonas aeruginosa Borrelia burgdorferi outer surface protein A Actinobacillus actinomycetemcomitans Fusobacterium nucleatum Porphytomonas gingivalis Capnocytophaga ochracea Mycobacterium tuberculosis and lipoarabinomannan (LAM) Protein-puriﬁed derivative Fungi and related products Cryptococcus neoformans and glucuronoxylomannan Candida albicians and derivatives Saccharomyces cerevisiae IgG-opsonized S. cerevisiae Zymosan Protozoa Plasmodium falciparum-infected erythrocytes Viruses Respiratory syncitial virus Epstein–Barr virus Other agents or conditions Protamine Sodium ﬂuoride PMA Thapsigargin Sulfatidesb Matrix proteins (ﬁbronectin, laminin)

NEUTROPHIL-DERIVED PROTEINS

379

human neutrophils (98). Interestingly, protamine, a basic peptide, exerted the same effect as that of LF and LFcin, suggesting the importance of the basic nature of LF and LFcin in mediating this stimulatory activity (98). Thrombopoietin, which regulates early and late stages of platelet formation as well as platelet activation, is yet another molecule that is able stimulate a sustained release of IL-8 from human neutrophils (64). Whereas sulfatides, which are established ligands for L-selectin, up-regulate the accumulation of IL-8 transcripts in neutrophils (99), ligation of P-selectin glycoprotein ligand-1 (PSGL-1) (100), or ICAM-3 (101), by immobilized antibodies to the plastic, triggers the release of IL-8 from neutrophils. Even reverse signaling via CD30L was demonstrated to induce the production of IL-8 by freshly isolated neutrophils (35). This phenomenon was obtained by cross-linking CD30L with speciﬁc monoclonal antibodies or by CD30–Fc fusion protein (35). Indirect effects through CD30 were clearly ruled out, because neutrophils do not express this surface molecule (34, 35). I have already mentioned that Y-IgG represents one of the most potent stimuli of IL-8 secretion by neutrophils (19). Similarly, coincubation of PMNs with zymosan (a yeast cell wall extract) (69), heat-killed Candida albicans (83, 86), or MP-F2, another candidial product (102), also results in the appearance of signiﬁcant quantities of immunoreactive IL-8 in culture supernatants of neutrophils. Although opsonization is not a prerequisite for the release of IL-8 under these conditions (45, 69, 103, 104), it does enhance the efﬁciency of stimulation (61, 103). Evidence that the release of IL-8 by neutrophils stimulated by zymosan is dependent on a CD11b/CD18 signaling pathway has also been provided (69). Other observations have revealed that two structurally distinct PAF antagonists suppressed zymosan-induced neutrophil IL-8 generation by approximately 70% (69). Similarly, cAMP-elevating agents proved to be effective negative modulators of IL-8 generation in zymosan-activated human neutrophils, mainly because they inhibited zymosan ingestion (104). Because IL-8 generation induced by LPS was shown to be unaffected by exogenous PGE2 (105), the data suggest that, depending on the activating stimuli, cAMPelevating agents differentially modulate the ability of neutrophils to produce IL-8. Neutrophils secrete copious amounts of IL-8 after phagocytosis of Listeria monocytogenes and Yersinia enterocolitica, and this is further potentiated by GM-CSF (106). Interestingly, a noninvasive L. monocytogenes strain was also found to be phagocytosed, and to induce high levels of IL8 release by PMNs, whereas the most important virulence factor of L. monocytogenes, listeriolysin O, was ineffective (106). Secretion of IL-8 by PMNs occurs in response to stimulation with Cryptococcus neoformans yeast cells, or with glucuronoxylomannan (GXM), the major C. neoformans

380

MARCO ANTONIO CASSATELLA

capsular polysaccharide (107), Streptococcus pyogenes (108), and Plasmodium falciparum-infected erythrocytes (109). Production of IL-8 also occurs after treatment of PMNs with various bacterial toxins: the Panton– Valentine leukocidin (Luk-PV, from Staphylococcus aureus V8) and alveolysin (Alv, from Bacillus alvei), two pore-forming toxins, as well as the erythrogenic toxin A (ETA, from Streptococcus pyogenes) (110). However, IL-8 release occurred only when the toxins were used at low concentrations (0.5–5 ng/107 cells) (110). Different pathways mediate the generation and release of IL-8 induced by these toxins, because inhibitors of protein tyrosine kinases reduced the release and the mRNA expression of IL-8 in PMNs challenged with Luk-PV and Alv, but not with ETA (110). Borrelia burgdorferi outer surface protein A (OspA), but not the unlipidated recombinant OspA, has been shown to represent a potent stimulant of neutrophil activities, including the secretion of IL-8 (62) (the spirochete B. burgdorferi is the etiologic agent of Lyme disease). Furthermore, a puriﬁed capsular polysaccharide complex (CPC) from Bacteroides fragilis, an anaerobic pathogen that accounts for the majority of infections that occur within the peritoneal cavity, yielded high levels of IL-8 production by PMNs, with a response more robust than, and time-dependently different from, that observed in autologous PBMCs (111). Finally, a very lowmolecular-weight (1 kDa) product of Pseudomonas aeruginosa was shown to induce the expression of IL-8 mRNA in neutrophils (112). In this regard, clinical P. aeruginosa isolates, the mucoid P. aeruginosa strain (CF3M) and its nonmucoid revertant (CF3), and puriﬁed P. aeruginosa mucoid exopolysaccharide (alginate) were all shown to produce a signiﬁcant increase in IL-8 release by human PMNs, in a dose- and time-dependent manner (113, 114). Stimulation of neutrophils with formalin-killed strain P. aeruginosa 5276 induced levels of IL-8 production that were more than 10 times higher than those induced by LPS or IL-1웁, after the same stimulation time (24 hr) (115). A moderate inhibitory effect of erythromycin (115) or erythromycin derivatives (116) on the production of IL-8 by formalin-killed Pseudomonas-stimulated neutrophils was also reported. In this respect, it is important to note that treatment with erythromycin of patients with chronic airway disease (CAD) and persistent P. aeruginosa infection causes signiﬁcant reductions of the inﬂammatory parameters in bronchoalveolar lavage (BAL) ﬂuids from these patients (115). Taken together, these observations have signiﬁcant implications in the context of cystic ﬁbrosis (CF), because they not only attribute an important role for PMN-derived IL-8 in maintaining neutrophil inﬂux, but provide a better understanding of the mechanisms by which P. aeruginosa bacteria determine the clinical outcome of CF and other diseases.

NEUTROPHIL-DERIVED PROTEINS

381

Other microorganisms promote the release of IL-8 by human PMNs. For instance, exposure of human neutrophils to Salmonella typhimurium and S. aureus strongly induced IL-8 (and MIP-1움) secretion, whereas Streptococcus pneumoniae and Staphylococcus epidermidis were less potent (114). Coincubation of neutrophils with TNF-움 and all of the above microorganisms led to a signiﬁcant enhancement of the induction of IL8. PMNs stimulated with Mycobacterium tuberculosis, or with a major mycobacterial cell wall component, lipoarabinomannan (LAM), release both IL-8 and GRO-움 (117). Release of IL-8 critically depends on the PMN: M. tuberculosis ratio, a 1 : 1 ratio being the optimal ratio (117). These ﬁndings have been conﬁrmed and extended with the demonstration that, in response to heat-killed M. tuberculosis and protein-puriﬁed derivatives (PPDs), neutrophils produce MIP-1움, but not MCP-1 (80), and that under those conditions, the release of IL-8 (but not that of MIP-1움) was potentiated by TNF-움 (80). Some viruses have been found to induce the production of IL-8 by PMNs. One example is the respiratory syncytial virus (RSV), which provokes in neutrophils an enhancement of IL-8 mRNA steady-state levels, accompanied by the secretion of IL-8, in a time- and dose-dependent manner (118). IL-8 synthesis depended on the adherence of viral particles and the subsequent phagocytic event, but not on the infection process. Indeed, stimulation of human PMNs with viable, heat-inactivated or UVinactivated RSV induced IL-8 production (protein plus mRNA) to a similar degree (61, 118). Stimulation of human PMNs with puriﬁed RSV Gprotein, a major capsid protein, resulted in an increased IL-8 release from human PMNs, but to a signiﬁcantly lesser degree compared with intact RSV (118). When RSV particles were opsonized with mAbs directed to the RSV fusion protein (F protein), the release of IL-8 from PMNs increased in comparison with RSV alone, and this increase occurred without a concomitant enhancement of IL-8 mRNA levels (61, 118). These results are highly reminiscent of those obtained by using yeast particles (Y) and Y-IgG as stimuli (103). In spite of an equivalent increase of IL-8 mRNA levels induced by both Y-IgG and Y, Y-IgG was more potent than Y in inducing the release of IL-8 (103). Engagement of the Fc웂-receptors might play a role in enhancing the synthesis and/or rate of the de novo-synthesized cytoplasmic IL-8 pool. Epstein–Barr Virus (EBV), the etiologic agent of infectious mononucleosis and other diseases, also leads to an increased accumulation of mRNA encoding IL-8 and MIP-1움 in neutrophils, and to a time-dependent production of the chemokines, with a maximum after 24 hr (119). This response was dependent on the interaction between the glycoprotein gp350 of the viral envelope and the neutrophil surface and not due to viral replication

382

MARCO ANTONIO CASSATELLA

or the synthesis of one or more intermediate proteins (119). The IL-8 produced under these conditions was biologically active, as determined in a calcium mobilization assay. Pretreatment of neutrophils with GM-CSF resulted in a signiﬁcantly greater proportion of the newly synthesized IL8 being secreted (approximately 41% versus less than 10%) (120). Because both IL-8 and MIP-1움 are chemoattractants for T cells and for T and B cells, respectively, the ability of EBV to induce their production by neutrophils may enhance the ability of this virus to infect B and T lymphocytes via increased recruitment to sites of infection. In addition, the release of these chemokines combined with the ability of EBV to alter the ratio of IL-1RA/IL-1 in favor of immunosuppression (121) may be related to the infectivity of the virus in vivo. Finally, the presence of IL-8 mRNA in most samples of freshly isolated neutrophils has been repeatedly observed (60, 61, 68, 78, 83, 113). Usually, these constitutive IL-8 transcripts are almost completely phased out within a few hours of culture in the absence of stimulation (60). Although it was reported that adherence to plastic induces neutrophils and monocytes to secrete IL-8 (61, 122), we (60) and others (82) did not ﬁnd substantial differences in culturing neutrophils under adherent (polystyrene) or nonadherent (polypropylene) conditions. Despite the presence of speciﬁc mRNA, the secretion of IL-8 by unstimulated cultured human neutrophils is always very low (below 100 pg/ml), unless the cells have been inadvertently preactivated by the isolation procedures. This constitutive IL-8 mRNA may either result from mechanical stress during cell preparation or represent a constitutive RNA pool that facilitates the rapid appearance of the mature protein on PMN stimulation. Evidence for the latter possibility is that the constitutive presence of TNF-움 mRNA or IL-1웁 mRNA (60) in freshly isolated PMNs is much more variable compared to that of IL-8 mRNA. Nevertheless, it cannot be excluded that cell separation procedures might weakly induce IL-8 but not TNF-움 or IL-1웁 mRNA expression (45). b. Growth-Related Gene Product-움, GRO-웁, and Related Chemokines. IL-8 is not the only C-X-C chemokine with neutrophil-activating properties that is secreted by human PMNs. Indeed, our studies revealed that activated PMNs can also release GRO-움 (21) in amounts and with kinetics that vary depending on the agonist. Previous work showed that PMNs that adhered for 45 min to ﬁbronectin-coated plastic expressed selectively GRO-움 mRNA, whereas monocytes from the same individuals expressed all three GRO isoforms (GRO-움, GRO-웁, and GRO-웂) (123). However, only RT–PCR analysis was utilized in that study (123). In our hands, phagocytosis of Y-IgG caused only a moderate GRO-움 mRNA accumulation (as compared to LPS, TNF-움, or fMLP), but it was approxi-

NEUTROPHIL-DERIVED PROTEINS

383

mately two- to threefold more potent than LPS, and much more than TNF-움, in inducing the production of GRO-움 (21). Very small quantities of GRO-움 protein were detected in culture supernatants of PMNs stimulated with fMLP, in spite of a very high accumulation of GRO-움 mRNA transcripts (comparable to that induced by LPS) (21). It is therefore possible that, as was observed in the case of Y-IgG-stimulated IL-8 release (124), GRO-움 production in Y-IgG-treated PMNs is controlled at the translational or posttranslational level. By contrast, in monocytes, LPS was more potent than Y-IgG and TNF-움, indicating that the regulation of GRO-움 production is governed by distinct mechanisms depending on the cell type (21). Other studies have conﬁrmed the ability of neutrophils to produce GRO움. In one study, PMNs obtained from peripheral blood (PB) or from reumathoid arthritis synovial ﬂuids (RA SF) were compared for their ability to generate GRO-움. After culture for 24hr, both RA SF PMNs and PB PMNs constitutively produced very high levels of GRO움, and this production was slightly increased on stimulation with LPS (22). In another work, neutrophils were stimulated for up to 12 hr with TNF-움, GM-CSF, MSU, and CPPD, used either individually or in combination, and the levels of GRO-움 in the culture supernatants were then assessed (25). Low levels of GRO-움 were detected in supernatants from both control and stimulated cells, regardless of the triggering conditions (25). Although the number of PMNs used in these experiments were not speciﬁed, the quantitative results shown with TNF-움 were in the order of picograms per milliliter, as in our case (25). Similarly, neutrophils treated with thapsigargin were found to release higher levels of GRO-움 compared to untreated cells, but this increase was not statistically signiﬁcant (94). More recently, evidence that PMNs activated with M. tuberculosis or LAM release GRO-움 has been clearly provided (117). LAM produced a concentration-dependent increase of GRO-움 release by PMNs, whereas M. tuberculosis stimulated a 12-fold increase of GRO-움 levels into neutrophil supernatants (117). Furthermore, though polymixin B did not inﬂuence GRO-움 or IL-8 production, treatment of neutrophils with an LTB4 inhibitor prior to stimulation with M. tuberculosis or LAM partially blocked GRO움 and IL-8 secretion (117). This suggests that the 5-lipooxygenase pathway is involved in the signal transduction cascade induced by M. tuberculosis and LAM to induce IL-8 and GRO-움. Based on their ﬁndings, the authors concluded that PMNs, through the generation of GRO-움 and IL-8, may contribute to the early host response against M. tuberculosis, increasing local inﬂammation and recruiting more phagocytes to the site of injury. The ability of PMNs to produce GRO-움 is relevant for several reasons. For instance, among its various biological activities, GRO-움 has mitogenic effects on normal and transformed human melanocyte cell lines (125), and

384

MARCO ANTONIO CASSATELLA

this raises the possibility that neutrophil-derived GRO-움 may play a role in melanocyte transformation. Like IL-8, GRO-움 also acts as a mediator of inﬂammation, in that it exerts powerful chemotactic and activatory properties toward PMNs, including degranulation, increased expression of adhesion molecules, and in vivo recruitment of neutrophils to sites of injection (54, 57). Thus, the generation of GRO-움 by neutrophils, in addition to that of IL-8, may stimulate the recruitment of neutrophils to inﬂammatory sites and their subsequent activation. In this respect, one study (126) has shown that the concentrations of GRO-움 and IL-8 were markedly elevated in BAL ﬂuids of three acute pathologic states: bacterial pneumonia (BPN), adult respiratory distress syndrome (ARDS), and Pneumocystis carinii pneumonia (PCP). The levels of these two chemokines were basically undetectable in 16 subjects of control group, but much higher in the ARDS and BPN groups than in PCP group (126). Interestingly, the levels of GRO-움 were biologically active and consistently higher than the levels of IL-8. GRO-움 or IL-8 levels both correlated with the absolute neutrophil number per milliliter when all groups were studied, again suggesting that GRO-움 could be as important as IL-8 in acute lung inﬂammatory process (126). Although this study did not establish the precise cellular source of GRO-움 and IL-8 in the alveoli, the results emphasize that in addition to IL8, other neutrophil-derived chemokines such as GRO-움 can be produced at high concentrations and are likely to act in concert with IL-8 in the lung. Human PMNs also express basal levels of GRO-웁 mRNA that increase up to 10-fold on stimulation with LPS for 24 hr (127). In monocytes, the same treatment results in a more than 100-fold increase of GRO-웁 transcripts (127). Because of the lack of available speciﬁc immunological assays to measure this chemokine, there exists only little information on GRO-웁 secretion by leukocytes. In contrast, the neutrophil expression of the various GRO proteins has been investigated in other species. For instance, rat peritoneal neutrophils harvested 16 hr after casein injection and incubated with staurosporine and PMA were shown to accumulate, in a time-dependent manner, high levels of CINC-3/MIP-2 mRNA and release signiﬁcant levels of CINC-1, CINC-2움, and CINC-3, but not CINC-2웁 (27). It should be pointed out here that on the basis of the similarities of amino acid sequences, CINCs are the rat counterparts of the human GRO proteins (128). CINC-3/MIP-2 was actually the chemokine most abundantly produced and was biologically active (27). mRNA accumulation and production of staurosporine- and PMA-induced CINCs were down-regulated by the protein kinase C inhibitors, H-7, calphostin C, and Ro 31-8425, and by the tyrosine kinase inhibitor, genistein (27). These data corroborate our previous ﬁndings on the ability of staurosporine and PMA to directly induce IL-8 production by human neutrophils (93).

NEUTROPHIL-DERIVED PROTEINS

385

The ability of rat neutrophils and macrophages to produce CINCs and MIP-1움 after phagocytosis of heat-killed yeast cells in vitro was assessed in another study (129). Neutrophils produced considerable amounts of CINC-2 and MIP-1움, but very low quantities of CINC-3 and no CINC1 (129). In contrast, macrophages produced high amounts of CINC-3 and MIP-1움, and considerable levels CINC-1 and CINC-2. Accordingly, supernatants collected from phagocytosing rat neutrophils and macrophages possessed chemotactic activities toward neutrophils, which were markedly inhibited by anti-CINCs or anti-MIP-1움 antibodies (129). Furthermore, dexamethasone suppressed the production of CINC-2 and MIP1움, due to its ability to inhibit the phagocytic process (129). Finally, the capacity of PMNs and macrophages to express CINC and other chemokines after adherence in vitro and exposure to endotoxin was assessed by Wu and colleagues (26). PMNs were obtained from the peritoneal cavity of rats previously injected with glycogen. PMNs expressed mRNA for CINC, MIP-2, MIP-1움, MIP-1웁, and MCP-1, either after adherence or with endotoxin stimulation at levels higher than macrophages (26). c. Interferon-웂-Inducible Protein-10, Monokine Induced by IFN-웂, and Interferon-Inducible T Cell 움 Chemoattractant. The C-X-C chemokines can be further subdivided into two classes, depending on the presence of the glutamate-leucine-arginine (ELR) motif preceding the ﬁrst two cysteines (54–56). IL-8, GRO, and other members express this motif and predominantly exert stimulatory and chemotactic activities toward neutrophils (54–56). In contrast, platelet factor 4 (PF-4), stromal-derived factor (SDF), IP-10, MIG, and the recently cloned interferon-inducible T cell 움 chemoattractant (I-TAC) (130) lack the ELR sequence (130, 131). Based on their structure and function, MIG, IP-10, and I-TAC consist of three very closely related chemokines that act mainly on T lympocytes and NK cells (130–132). MIG, I-TAC, and IP-10 are usually strongly induced by IFN-웂 in a range of cell types, including monocytes, keratinocytes, endothelial cells, and astrocytes (130–132). Though the receptors for the C-X-C chemokines containing the ELR motif are expressed on different types of leukocytes, there is thus far only a single receptor, CXCR3, that is known to bind MIG, I-TAC, and IP-10 (130, 131, 133). Importantly, CXCR3 (and CCR5) seem to be preferentially expressed on activated T lymphocytes of the T helper type 1 phenotype (134–137). Studies in vitro have shown that MIG and IP-10 are also active as inhibitors of colony formation by hematopoietic cells (138, 139). In vivo, MIG and IP-10 inhibit neovascularization and exert antitumor effects in murine models (140–145). We have observed that both IP-10 and MIG are produced and released by human neutrophils (146, 147). We also found that I-TAC mRNA can

386

MARCO ANTONIO CASSATELLA

be expressed in neutrophils under speciﬁc stimulatory conditions (147). To our surprise, IFN-웂 alone exerts only a modest effect on IP-10, MIG, and I-TAC mRNA accumulation in neutrophils (146, 147). However, stimulation of PMNs with IFN-웂 in combination with either TNF-움 or LPS (but not with Y-IgG, fMLP, IL-8, PMA, or G-CSF) resulted in a considrable induction of IP-10, MIG, and I-TAC mRNA transcripts, as well as in the extracellular release of the IP-10 and MIG proteins. Though IP-10 and MIG production in IFN-웂 treated neutrophils was dependent on the concentration of LPS or TNF-움, kinetics of chemokine release in culture supernatants were different, in the sense that IP-10 was detected after 4 hr, and MIG was measured only after 18–21 hr (146, 147). By contrast, IFN-웂 proved to be a very potent stimulus for IP-10, MIG, and I-TAC mRNA experession and IP-10 and MIG production by autologous PBMCs, suggesting that the production of these chemokines is controlled in different ways, depending on the cell type. Another striking difference between PBMCs and PMNs, was that costimulation of IFN-웂-treated PBMCs with LPS led to a diminished production of IP-10 and MIG with respect to PBMCs treated with IFN-웂 alone (146, 147). IL-10, and, less efﬁciently, IL-4, suppressed extracellular production of IP-10 and MIG in neutrophils stimulated with IFN-웂 plus LPS or TNF-움 (146, 147). Finally, in agreement with previous ﬁndings on monocytes (148), IFN-움 alone induced only IP10 mRNA in neutrophils, whereas in PBMCs it induced both IP-10 and I-TAC mRNA (147). Immunological assays on IP-10 protein release revealed that IFN-움 in combination with IFN-웂, LPS, TNF-움, or IL-1웁 determined a considerable release of antigenic IP-10 into the neutrophil supernatants (147). The ability of PMNs to produce MIG, IP-10, and likely I-TAC might be signiﬁcant considering the various biological functions that these three chemokines exert. Generation of MIG, IP-10, and I-TAC by PMNs may, for instance, contribute to recruit Th1 lymphocytes to sites of inﬂammation (134–137). The selective activity of MIG, I-TAC, and IP10 on activated T cells and probably NK cells is consistent with a role in regulating the trafﬁcking and/or function of effector cells during an immune response. Because IFN-움, in addition to IL-12, IFN-웂, TGF-웁 and hormones, has been reported to regulate Th1 development (149), production of IP-10 by IFN-움-stimulated neutrophils may represent one of the mechanisms that contributes to Th1 responses essential for clearance of pathogens such as viruses. Alternatively, neutrophil-derived IP-10, MIG, or I-TAC might negatively regulate hematopoiesis (138, 139), or, in light of to the established angiostatic properties of IP-10 and MIG (140–145), may represent one of the mechanisms whereby these cells exert antitumor effects (150). Neutrophils can in fact mediate tumor cell killing through direct or bystander effects, and can participate in the cross-talk with CD8 T cells,

NEUTROPHIL-DERIVED PROTEINS

387

which is instrumental in the rejection of speciﬁc cytokine-transduced tumors (151–153). 2. C-C Chemokines a. Macrophage Inﬂammatory Protein-1움 and MIP-1웁. MIP-1움 and MIP-1웁 are members of the C-C subfamily of chemokines, which also includes MCP-2, MCP-3, I-309/TCA3, and RANTES (57). Data from a number of laboratories suggest that both MIP-1움 and MIP-1웁 are expressed primarily in T cells, B cells, monocytes, neutrophils, and Langerhans cells stimulated with antigens, speciﬁc agonists, or mitogens (154). MIP-1움 acts as a potent chemotactic/activating factor for monocytes and subpopulations of T and B lymphocytes and eosinophils, and also activates several effector functions of macrophages and neutrophils, such as the generation of hydrogen peroxide (57) and the secretion of TNF-움, IL-1움, and IL-6 (154). MIP-1움 may be pyrogenic and might also be implicated in negative regulation of myelopoiesis, because it has suppressive activity on hematopoietic stem cells and early subsets of hematopoietic progenitor cells, both in vitro and in vivo (57, 154). A preferential migration of activated CD4⫹ and CD8⫹ T cells in response to MIP-1움 and MIP-1웁 has been reported (155). The ability of stimulated neutrophils to secrete MIP-1움 and MIP-1웁 has been fully characterized by Kasama and co-workers (23, 24). It was initially observed that cell-free supernatants from human PMNs stimulated with increasing concentrations of LPS possessed a chemotactic activity for human monocytes; this activity was signiﬁcantly attenuated (by approximately 60%) in the presence of neutralizing antihuman MIP-1움 antibodies. By Northern blot analysis, enzyme-linked immunosorbent assay (ELISA), and immunocytochemistry, the group succeeded in demonstrating that LPS induces mRNA expression and extracellular production of MIP-1움 (23,) and of MIP-1웁 as well (24). Though GM-CSF, G-CSF, and IL-3 failed by themselves to induce the expression of MIP-1움, stimulation of PMNs with LPS in the presence of GM-CSF resulted in a synergistic expression of both MIP-1움 mRNA and protein, compared with LPS alone (23). Other stimuli were subsequently shown to possess the ability to induce neutrophils to secrete MIP-1움 (25). For instance, MSU, CPPD, and GMCSF were found to be ineffective, whereas TNF-움 signiﬁcantly up-regulaed MIP-1움 mRNA expression and secretion (25). Secretion of MIP-1움 induced by TNF-움 was completely inhibited by the presence of either MSU or CPPD, yet the production of IL-8 under the same conditions was synergistically enhanced (25). These results are of great relevance because they imply that the failure of inﬂammatory microcrystals to induce MIP-

388

MARCO ANTONIO CASSATELLA

1움 production directly, or to inhibit the release of MIP-1움 in response to TNF-움, can prevent the generation of neutrophil-derived chemotactic signals that could attract mononuclear cells into the synovial enviornment. Exposure of human neutrophils to microorganisms represents another condition that triggers MIP-1움 mRNA and secretion in human neutrophils. This is indeed the case for P. aeruginosa, S. aureus, S. typhimurium, S. pneumoniae, and to a lesser extent, S. epidermidis, and C. albicans (114). In contrast, Saccharomyces cerevisiae and zymosan failed to induce MIP1움 release (114). Production of MIP-1움 induced by TNF-움 was inhibited by the simultaneous treatment with zymosan or S. cerevisiae, whereas it was potentiated by S. typhimurium, P. aeruginosa, or S. pneumoniae (114). These results indicate that whereas neutrophils exposed to some micro organisms in the presence of TNF-움 will produce both IL-8 and MIP-1움, thus resulting in generation of signals for the recruitment of neutrophils and mononuclear leukocytes, respectively, certain types of microorganisms can skew this response toward the synthesis of only IL-8. MIP-1움 (as well as IL-8 and GRO-움) is also produced by neutrophils in response to heat-killed M. tuberculosis and PPD (80), although at lower levels than after stimulation with LPS (80). The yields of antigenic MIP1움 detected in neutrophils stimulated with M. tuberculosis or PPD in combination with TNF-움 (or IL-1웁) did not differ from those measured after stimulation with mycobacterial derivatives alone (80), despite a higher expression of MIP-1움 mRNA (80). The reason for the discrepancy between MIP-1움 mRNA expression and MIP-1움 antigen release from neutrophils stimulated with M. tuberculosis or PPD in the presence of TNF-움 was attributed to a reduction of neutrophil viability induced by TNF-움 (80). As mentioned previously, exposure of neutrophils to EBV leads to an increased accumulation of MIP-1움 and IL-8 mRNA, followed by a timedependent production of the related antigenic molecule (119). EBV increases the synthesis of MIP-1움 in neutrophils by about 176-fold, approximately 87% of which is secreted (119). Under similar conditions, only 27% of the total IL-8 is released (119). Treatment of neutrophils with GMCSF before EBV activation enhances the production of both MIP-1움 and IL-8, without modifying the levels of their mRNA compared with those obtained in response to either GM-CSF or EBV alone (120). The ability of GM-CSF to potentiate the production of MIP-1움 by EBV-stimulated neutrophils may enhance the ability of this virus to infect B and T lymphocytes via increased recruitment to sites of infection. Also, supernatants from rat neutrophils (and macrophages) collected after phagocytosis of heat-killed yeast cells were shown to display toward neutrophils chemotactic activities that were markedly inhibited by dexamethasone and anti-MIP-1움 antibodies; these supernatants contained ele-

NEUTROPHIL-DERIVED PROTEINS

389

vated levels of MIP-1움 (129). In neutrophils, the peak of MIP-1움 release was observed after 5 hr of incubation, whereas MIP-1움 mRNA levels were still elevated at 24 hr (129). The lack of correspondence between MIP1움 mRNA expression and MIP-1움 extracellular detection was ascribed to the degradation of antigenic MIP-1움 by serine proteases released from neutrophils during phagocytosis. In this regard, the addition of several protease inhibitors (including phenylmethylsulfonyl ﬂuoride, leupeptin, chymostatin, or antipain) to the 8-hr supernatants of neutrophils cultured with yeast cells and incubated for another 20 hr signiﬁcantly inhibited the degradation of MIP-1움 occurring under these conditions (129). Finally, increase in MIP-1움 and MIP-1웁 mRNA expression in rat PMNs after adherence in vitro and exposure to endotoxin to levels higher than in macrophages was also observed. The ability of PMNs to secrete MIP-1움 and MIP-1웁 is of considerable importance. The fact that neutrophils produce MIP-1움 and MIP-1웁 (in addition to IL-8, GRO-움, IP-10, MIG, and I-TAC) suggests that once PMNs arrive at an inﬂammatory site, they not only promote the further recruitment of neutrophils, but also the subsequent accumulation and activation of monocytes/macrophages, eosinophils, and lymphocytes. It is reasonable to speculate that human PMNs, through the production of IL8, GRO-움, CINCs, MIP-1움, MIP-1웁, IP-10, MIG, and I-TAC, could play a pivotal role in regulating the switch of the type of leukocyte inﬁltration typically observed during the evolution of the inﬂammatory response from acute to chronic stages. Such data reinforce the concept that neutrophils have the potential to regulate the migration of various leukocytic cellular types into inﬂammatory sites. b. Monocyte Chemotactic Proteins. Monocyte chemotactic protein-1 is a glycoprotein with a molecular mass of 9–15 kDa, it belongs to the CC family of chemokines (54, 57) and is also known as monocyte chemoattractant and activating factor (MCAF). Almost all cells or tissues examined make MCP-1 on stimulation with a variety of agents, but the targets are limited to monocytes and basophils (54, 57). MCP-1/JE is known to strongly stimulate chemotaxis of peripheral blood monocytes, and to modulate several monocyte responses, including the respiratory burst, calcium inﬂux, adhesion molecule expression, release of lysosomal enzymes, and monocyte-mediated inhibition of tumor cell growth (54, 57). MCP-1 also exhibits chemotactic activity toward basophils, eosinophils, and lymphocytes, but not neutrophils (54, 57). In addition, MCP-1 directly up-regulates monocyte cytokine production, and cytostatic activity (54, 154). Interestingly, monocytes are the primary source of this cytokine, and respond to TNF-움 with increased MCP-1 production in an autocrine manner (54,

390

MARCO ANTONIO CASSATELLA

154). Reports have revealed the existence of two related proteins, MCP2 and MCP-3, which, respectively, share 62 and 73% amino acid homology with MCP-1 (156). MCP-2 and MCP-3 also share the chemoattractant speciﬁcity of MCP-1 for monocytes in vivo (156). Attempts to determine whether MCP-1 mRNA gene expression can be induced in human neutrophils, either after adhesion or after stimulation with LPS, produced negative results (20). Subsequently, Van Damme and colleagues developed sensitive RIAs for MCP-1 and MCP-2 (157). In agreement with the previous study (20), they could not detect any extracellular release of either MCP-1 or MCP-2 by PMNs, even though neutrophils were stimulated for up to 48 hr with optimal doses of many agonists, including IL-1웁, IFN-웁, IFN-웂, GM-CSF, PMA, and LPS (157). In contrast, Burn et al. (158), by using Northern analysis, RT–PCR, and IH, reported that MCP-1 transcripts and antigenic protein were absent in freshly isolated neutrophils, but were clearly detectable in neutrophils incubated for 20 hr in tissue culture medium. Under similar experimental conditions, however, we were unable to detect any MCP-1 transcripts by Northern blot (M. P. Russo, unpublished observations). Similarly, Hachicha et al. (25) reported that mRNA for MCP-1, MCP-2, MCP-3, RANTES, or I-309 was undetectable in PMNs stimulated for 3 hr with TNF-움, GM-CSF, MSU, and CPPD, either alone or in combination. Additionally, neither MCP-1 mRNA nor protein was detected in neutrophils treated for up to 24 hr with graded doses (0.001–100 애g/ml) LPS, heat-killed M. tuberculosis, or PPD (80), or in response to immobilized anti-ICAM-3 antibodies (101). Contradictory ﬁndings on the in vivo expression of MCP1 in neutrophils also exist, and will be discussed below. These various observations emphasize the need to conduct further studies aimed at clarifying whether, or under which speciﬁc conditions, human neutrophils express MCP-1 mRNA and protein. B. TUMOR NECROSIS FACTOR-움 Tumor necrosis factor-움, a paracrine and endocrine mediator with potent immunomodulatory and proinﬂammatory properties, plays a major role in host defense (159). Biologically active TNF-움 is a homotrimer consisting of 17-kDa subunits that is produced mainly by activated monocytes and macrophages (159). TNF-움 belongs to a superfamily of membraneanchored and soluble cytokines, including lymphotoxin-움 (LT-움) and LT웁, Fas ligand, CD30 and CD40L, as well as their respective receptors, which collectively constitute a class of molecules that are centrally involved in T cell immunity (160). TNF-움 exerts a variety of strong proinﬂammatory effects on different cell types, such as the induction of catabolic states in adipocytes, of endothelial cell adhesion molecules, of plasminogen activator

NEUTROPHIL-DERIVED PROTEINS

391

and its inhibitor, as well as that of other cytokines, and so forth (159). As a result, TNF-움 signiﬁcantly contributes to vasodilatation, thrombosis, leukocyte recruitment, bone resorption, matrix degradation in cartilage, changes in liver metabolism, pannus formation, and cachexia, consistent with the important role attributed to TNF-움 in many disease states (159). In addition, TNF-움 has been shown to be a primary mediator of the pathology seen in endotoxic shock, and to be selectively cytotoxic for many transformed cells, especially in combination with IFN-웂. Many of the biological actions of TNF-움 occur in combination with other cytokines, as part of a cytokine network. The cellular targets of TNF-움 include monocytes, macrophages, lymphocytes, eosinophils, and neutrophils (159). In the latter cell type, TNF-움 increases PMN adherence to vascular endothelium, enhances neutrophil phagocytosis and antibody-dependent cell cytotoxicity, and triggers neutrophil degranulation, as well as the release of reactive oxygen metabolites (161). Moreover, under certain conditions, TNF-움 is also a very potent priming agent for neutrophil functions (161). The ﬁrst evidence on TNF-움 mRNA expression in PMNs was provided by Lindemann et al. (162), who analyzed neutrophils stimulated with GM-CSF. Under these conditions, however, PMNs did not release the TNF-움 protein—an observation that was repeatedly conﬁrmed in subsequent studies (50, 163, 164). Indirect evidence for the involvement of released TNF-움 in the granulocyte-mediated killing of two extremely TNF움-sensitive cell lines, namely, WEHI sarcoma 164 and L929 cells, was published almost at the same time by Mandy and colleagues (165). Similarly, indications for a constitutive release of a TNF-움-like molecule by inﬂammatory PMNs, harvested from the peritoneal cavity of mice injected with casein were also reported but not substantiated at the molecular level (166). Molecular evidence of the ability of PMNs to express TNF-움 mRNA and secrete the related protein was ﬁrst obtained using LPS as a stimulus (6, 7). Indeed, PMNs exhibited increased levels of TNF-움 mRNA, and released approximately 160–190 pg/ml/5 ⫻ 105 cells of TNF-움 into culture supernatants in response to 5 애g/ml LPS (6). Djeu et al. (7) conﬁrmed that LPS (1 to 1000 ng/ml) was a good inducer of TNF-움, but additionally reported that incubation of neutrophils with the opportunistic fungus, C. albicans, leads to a substantial extracellular release of TNF-움. Mandy et al. (167) similarly found that human neutrophils can be stimulated by C. albicans to produce TNF-움 in vitro. Because PMNs are crucial effector cells responsible for the elimination of C. albicans, and because TNF-움 is a potent activator of this PMN function (168), the ability of PMNs to produce TNF-움 suggests the possible existence of an autocrine feedback loop for self-activation to muster a host defense against microbes. In this

392

MARCO ANTONIO CASSATELLA

regard, it has been shown that supernatants from PMNs incubated with C. albicans for 18 hr signiﬁcantly decrease the number of fresh PMNs demonstrating features of apoptosis, while increasing the percentage of viable PMNs during in vitro culture (169). Neutralization of TNF-움 biologic activity with speciﬁc monoclonal antibodies partially abrogated the supernatant-mediated prolongation of PMN survival (169). Mandy’s group reported that, in addition to C. albicans, LPS, PMA, and other microorganisms (e.g., E. coli, S. aureus, and Klebsiella pneumoniae) can also stimulate neutrophils to produce (TNF-움) (167). More recently, endogenously generated TNF-움 induced by S. aureus was found to mediate the up-regulation of ICAM-1 observed in S. aureus-treated neutrophils (163). Furthermore, release of S. aureus-induced TNF-움 was inhibited by the phosphodiesterase inhibitors pentoxifylline and rolipram, suggesting that elevation of cAMP may represent a mechanism by which the production of TNF-움 is inhibited in neutrophils (163). In this regard, exogenous dibutyryl cyclic AMP (DBcAMP) signiﬁcantly inhibited the production of TNF-움 in LPSstimulated PMNs (170). In comparison to LPS, we showed that phagocytosis of Y-IgG by neutrophils induces an extracellular release of TNF-움 at much higher levels (8). In contrast to neutrophils, LPS proved to be a more potent activator of TNF-움 release than phagocytosis in monocytes or PBMCs (8, 171). Interestingly, maximal yields of TNF-움 in neutrophil supernatants in response to Y-IgG or LPS are detected after 5–10 hr of stimulation, and then decline over time (6–8, 171–173). In fact, it can be difﬁcult to detect any TNF-움 at all in culture supernatants from PMNs stimulated for 24 hr with substimulatory doses of LPS (45, 174). However, if very high doses of LPS are used, then a more sustained TNF-움 release over time can be observed (107). Though it is not clear yet why the yields of released TNF움 decay with time, this also happens to be the case with MIP-1움, which is released by phagocytosing neutrophils (101). In this regard, phagocytosis carried out in the presence of protease inhibitors such as 움1-antitrypsin increased the recovery of TNF-움 (103), suggesting that the stability of the TNF-움 protein in the culture medium is inﬂuenced by proteolytic enzymes. This is keeping with the ability of elastase and cathepsin G to degrade TNF-움 speciﬁcally (175, 176), and with the fact that phagocytosis is a potent stimulus for neutrophil degranulation (177). On what perhaps constitutes a related observation, Takeichi et al. (52) reported that neither LPS, concanavalin A (Con A), nor zymosan was an effective stimulus for TNF-움 release by neutrophils, even though they noted that TNF-움 mRNA was induced by those stimuli. Similarly, in another study (178), the same group reported that highly puriﬁed PMNs (⬎99.5%) isolated from periapical exudates (PEs) expressed signiﬁcant levels of mRNA for TNF-움 (and IL-

NEUTROPHIL-DERIVED PROTEINS

393

1움/웁), but that antigenic TNF-움 was not detected in those exudates. The authors attributed these conﬂicting results to the fact that their TNF-움 assay was not sensitive enough (52), or that TNF-움 produced by PMNs was adsorbed by active TNF receptor-expressing cells (178). In view of the convincing evidence generated by several other groups, that LPSstimulated PMNs do release TNF-움 (50, 51, 124, 179) and even express the TNF-움 membrane-bound form (170), another potential explanation for the negative ﬁndings of Takeichi’s group might be that under their experimental conditions the TNF-움 protein was proteolyzed. An important role for surface CD14 in mediating the ability of LPS to induce TNF-움 secretion in neutrophils emerged from a recent study by Haziot et al. (180). In that study, stimulation of neutrophils under serumfree conditions with complexes of puriﬁed LPS-binding protein (LBP) and low concentrations of LPS (0.1 to 5 ng/ml) resulted in a release of TNF움 that was severely inhibited by anti-CD14 antibodies (180). Similarly, concentrations as low as 0.1 ng/ml of Salmonella abortus equi endotoxin (which are in the same range as those observed in patients with sepsis) (181) were shown to stimulate signiﬁcantly the production of TNF-움 by PMNs (173). This response was inhibited in a concentration-dependent fashion by several adenosine receptor agonists, the A2 receptor agonist [5⬘-(N-cyclopropyl)carboxamidoadenosine, CPCA] being 1000 times more effective than the 301 agonist (2-chloro-N6cyclopentyladenosine) (173). Noteworthy is that CPCA inhibited the TNF-움 production even when administered 2 hr after LPS stimulation (173). Binding of adenosine to A2 receptors is linked to the activation of the adenylate cyclase and subsequent increase of intracellular cAMP, and there is abundant evidence that cAMP is a potent negative regulator of the synthesis of TNF-움 (182)—for instance, the aforementioned effects of pentoxyﬁlline and rolipram (163). Table III lists the stimuli that are known to induce TNF-움 expression in human neutrophils. Brieﬂy, these include sulfatides (99), thapsigargin (94), MP-F2 (102), L. monocytogenes and Y. enterocolitica (183), C. neoformans and its derivatives (107), LPS from various periodontopatic bacteria (75), and P. falciparum-infected erythrocytes (109). In addition to the above stimuli, Balazovich and colleagues (184) reported that in PMNs plated onto ﬁbrinogen, fMLP stimulates a detectable release of TNF-움 within 45 min, with maximal production by 90 min (184). Under these conditions, PMNs also released H2O2 and lactoferrin with kinetics similar to those of TNF-움. Because TNF-움 is a potent agonist of adherent PMNs (185), the authors additionally investigated the effects of neutralizing antiTNF-움 antibodies in their experimental model. The latter inhibited both H2O2 and lactoferrin release, whereas rabbit IgG, anti-HLA-A, anti-HLAB, anti-HLA-C, anti-CD14, and anti-IL-8 antibodies were all ineffective

394

MARCO ANTONIO CASSATELLA

TABLE III AGENTS ABLE TO TRIGGER THE PRODUCTION OF TNF-움

BY

HUMAN NEUTROPHILS

Cytokines and growth factors Bacteria and related products TNF-움 LPS from E coli IL-1웁 LPS from A. actinomycetemcomitans LPS from F. nucleatum IL-2a GM-CSF LPS from P. gingivalis Other agents or conditions LPS from C. ochracea PMA Klebsiella pneumoniae Concanavalin A Staphylococcus aureus Escherichia coli Sulfatidesb Matrix proteins (ﬁbronectin, Yersinia enterocolitica laminin) Listeria monocytogenes Fungi and related products Cryptococcus neoformans and glucuronoxylomannan Candida albicans and derivatives Saccharomyces cerevisiae IgG-opsonized S. cerevisiae Zymosan Protozoa Plasmodium falciparum-infected erythrocytes a b

Requires deﬁnitive conﬁrmation. mRNA only.

(184). Furthermore, treatment of PMNs with either actynomycin D (Act D) or cyclohexymide (CHX) resulted in a partial (33%) inhibition of H2O2 and lactoferrin release, suggesting that protein synthesis was required for this fMLP-mediated activation of adherent PMNs. The addition of TNF움 to either CHX or of Act D-treated PMNs overcame their inhibition, indicating that the effect was speciﬁc for TNF-움. Finally, a combination of antibodies against both TNF-움 receptors resulted in a signiﬁcant inhibition of fMLP-mediated activation of H2O2 and lactoferrin release. Taken together, these ﬁndings support the hypothesis that TNF-움 release and ligation of TNF-움 receptors are central for fMLP-stimulated oxidant release from PMNs adherent to ﬁbrinogen. In contrast, Derevianko and colleagues were not able to detect any TNF-움 released by PMNs previously adhered to plastic surfaces, or to surfaces precoated either with ﬁbronectin (Fn) or laminin (Ln), and then stimulated for up to 24 hr with fMLP (100 nM ) (74). Under these conditions, however, IL-8 was clearly released by fMLP-stimulated cells (74). In agreement with Derevianko et al. (74), we (5) and others (186) have been unable to demonstrate any effect of fMLP on TNF gene expression or release in PMNs. Even though we cannot formally exclude the possibility that the experimental conditions that we routinely use might not be optimal (5), it was recently reported

NEUTROPHIL-DERIVED PROTEINS

395

that fMLP, over a wide concentration range (1 nM–1 애M ), failed to induce the production of TNF-움 by PMNs (170). Moreover, the latter authors found that fMLP (1 애M ) actually inhibited the release of TNF-움 by neutrophils treated with LPS (170); similarly, fMLP was shown to inhibit TNF-움 production in PMNs costimulated with IFN-웂 and LPS. The mechanisms underlying this inhibitory action of fMLP were also explored, and a number of interesting observations were made (170). First, the pattern of CD14 and LPS binding sites remained unaltered in PMNs after fMLP treatment, whereas the CD11/CD18 complex, which has also been described as an LPS receptor (187), was up-regulated (170). Second, it was excluded that the lower levels of TNF-움 found in supernatants of fMLPtreated PMNs could be due to the adsorption of TNF by soluble TNF-움 receptors released from PMNs (170). Third, fMLP did not diminish the levels of TNF-움 mRNA induced by LPS. Finally, as evaluated by ﬂow cytometry, fMLP decreased to basal levels the membrane-bound form of TNF-움 induced by LPS (170). Together, these ﬁndings strongly support the view that fMLP does not stimulate TNF-움 gene expression or release in neutrophils. From a more general point of view, these studies also indicate that fMLP, which is usually viewed as a proinﬂammatory agent, can assume an anti-inﬂammatory role under certain speciﬁc circumstances. Another agonist that has been proposed to induce TNF-움 gene expression and secretion in PMNs is IL-2 (188). Release of TNF-움 protein by IL-2-treated PMNs was reported to be dose and time dependent (188). Though CHX did not affect the accumulation of TNF-움 mRNA in IL-2treated cells (1000 U/ml), it did block that induced by GM-CSF (188). In addition, CHX superinduced TNF-움 gene expression in cells stimulated with IL-8 or heat-killed C. albicans. Thus, it appeared from that study that various induction pathways exist in PMNs for the TNF-움 gene (188). Interestingly, the IL-2-induced TNF-움 release was inhibited by monoclonal antibodies raised against the 웁 chain of the IL-2 receptor (IL-2R웁) (188). This is consistent with previous studies by the same group, which demonstrated that neutrophils express IL-2R웁 (but not IL-2R움) a molecule also found to be involved in the killing of C. albicans by IL-2-stimulated PMNs (189). On a related note, the cloning of the IL-2 receptor 웂 chain (IL-2R웂) in Sugamura’s laboratory (190, 191) led to the observation that this molecule is constitutively expressed in granulocytes (192); this was subsequently conﬁrmed by several other groups (193–196), including our own (90). The relevance of the latter ﬁnding is illustrated by the fact that IL-2R웂 is an indispensable component of high- and intermediate-afﬁnity IL-2 receptor complexes, because its cytoplasmic domain plays a critical role in mediating intracellular signal transduction (190, 191). On the basis of the above considerations, it would appear that neutrophils express inter-

396

MARCO ANTONIO CASSATELLA

mediate-afﬁnity IL-2 receptors consisting of the 웁 and 웂 chains. However, Sugamura’s group did not limit their investigations to IL-2R웂; in particular, they found no evidence that neutrophils express either IL-2R움 or IL-2R웁 (192); the lack of 웁 chain expression in PMNs was also observed by Philips et al. (193). These discrepancies with the data of Djeu’s group (189) were attributed to nonspeciﬁc antibody binding in the latter case, because PMNs were not pretreated with serum or human IgG (189), a step that is essential for the speciﬁc staining of the 웁 chain on the cell surface (192). This being said, another study did provide evidence for IL-2R웁 surface expression in PMNs, using adequate staining procedures (194). Though the latter study somewhat fueled the controversy surrounding the issue of IL-2R웁 expression in PMNs, it also provided a highly interesting piece of information. Indeed, it showed that even though up to 70% of neutrophils expressed IL-2R웁 on their surface, less than 20% of the cells were able to bind ﬂuorescent-labeled IL-2 (194). Thus, it appears that neutrophil populations bind IL-2 very poorly, regardless of whether the cells express IL-2R웁. In contrast to the IL-2R웁 controversy, numerous groups have observed that neutrophils do not express IL-2R움 (86, 189), a subunit that is required for high-afﬁnity binding of IL-2. This indicates that in addition to binding poorly to neutrophils, IL-2 does not even bind with high afﬁnity. If true, the latter conclusion raises serious concerns over the reported ability of IL-2 to stimulate neutrophil function. In this regard, we repeatedly failed to observe any effect of IL-2 toward TNF-움 or IL-8 mRNA expression and release by human neutrophils. Our negative results could not be attributed to a poor biological activity of the three different commercial preparations of IL-2 that we used (at doses ranging from 1 to 10,000 U/ ml), because they were all effective in inducing lymphocyte proliferation and cytokine production from autologous mononuclear cells. In keeping with our data, Girard and colleagues found that IL-2 neither affected de novo RNA synthesis, nor did it induce the synthesis of any individual protein species in neutrophils (194). Interestingly, the latter study also revealed that in PMNs stimulated with IL-2 in combination with a ﬁxed concentration of GM-CSF (but not of TNF-움 or fMLP), a dose-dependent effect of IL-2 toward the induction of both RNA and protein synthesis was observed (194). This therefore suggests that IL-2 may actually have the ability to stimulate neutrophils, yet only under very speciﬁc conditions—such as when the cells are ‘‘primed’’ by GM-CSF (or perhaps by other agonists). The ability of PMNs to release TNF-움 in response to many different stimuli suggests that the role of granulocytes within the context of host defense goes well beyond the killing of invading microorganisms in septic infections. Secretion of TNF-움 by neutrophils may contribute to the nu-

NEUTROPHIL-DERIVED PROTEINS

397

merous metabolic effects that TNF-움 elicits in the host, but may also represent a mechanism whereby neutrophils can activate themselves in an autocrine/paracrine fashion (197). C. FAS LIGAND Accumulating evidence indicates the Fas (APO-1/CD95)/Fas ligand (FasL) system represents an important cellular pathway responsible for the induction of apoptosis in various tissues (198). In addition to its role in the induction of apoptosis, the Fas system has also been suggested to function as an autocrine pathway to T cell suicide (199). Fas/CD95, a member of the TNF/nerve growth factor family, is a widely expressed 45-kDa type I membrane protein which mediates apoptosis following its ligation with agonistic anti-Fas IgM Ab or FasL (200). FasL, a 40kDa type II protein, is also a member of the TNF/NGF superfamily; in contrast to Fas, however, the tissue expression of FasL is relatively restricted, but has been reported to be inducible (200). Importantly, soluble forms of FasL capable of inducing apoptosis have been found to be secreted by activated cells (201). Among the leukocytes, mature human neutrophils have the shortest lifespan and die rapidly via apoptosis, both in vivo and in vitro (202). It must be stressed, however, that neutrophil survival can be signiﬁcantly extended by certain growth factors (G-CSF, GM-CSF), proinﬂammatory cytokines (IFN-웂, IL-1, TNF, and IL-6), bacterial products (LPS), and even glucocorticoids (203). Within the context of the ongoing search for the mechanisms ultimately responsible for the regulation of apoptosis in neutrophils, the ﬁndings that normal PMNs are highly susceptible to Fas-induced death raised the possibility that the Fas system could play a fundamental role in the regulation of spontaneous neutrophil apoptosis (204). Further support for this hypothesis was the demonstration of the presence of FasL on the surface of mature human neutrophils (32). Experiments performed to examine the effects of antagonistic anti-Fas IgG1 and Fas-Ig (deployed as a soluble receptor for FasL) signiﬁcantly reduced the spontaneous neutrophil death in vitro by approximately 50% during a 72-hr time course (32). In addition, neutrophils were shown to release a biologically active, soluble form of FasL capable of speciﬁcally reducing the viability of target cells in a coculture system (32). Taken together, these ﬁndings not only suggest that the Fas system could represent a key element controlling the rapid spontaneous turnover of neutrophils, but also implicate a possible role for FasL in neutrophil-mediated cytotoxicity (205). The coexpression of FasL and Fas in PMNs, as well as their involvement in inducing spontaneous PMN apoptosis, have been fully conﬁrmed (33, 206). In one study, the effects of EBV on the expression of the Fas/FasL

398

MARCO ANTONIO CASSATELLA

antigens were evaluated, in view of the fact that EBV can induce neutrophil apoptosis (206). Exposure of neutrophils to EBV was found to increase the expression of Fas and FasL signiﬁcantly on the neutrophil outer surface; moreover, the kinetics of this increase paralleled those of EBV-induced neutrophil death (206). However, even though EBV exposure concurrently induced the release of soluble FasL, treatment of freshly isolated neutrophils with supernatants from EBV-treated neutrophils failed to modify signiﬁcantly the rate of spontaneous apoptosis of the naive neutrophils (206). Thus, whereas this study did not clearly establish the involvement of the Fas/FasL system in the EBV-induced apoptotic process, the possibility that the Fas/FasL system might nevertheless play an important role, perhaps in synergy with virally encoded proteins, remains. Despite the uncertainties that remain about the intricacies of the Fas/FasL system in neutrophils, the number of studies aiming to elucidate the signal transduction pathway linking Fas engagement to neutrophil apoptosis is currently growing at a rapid rate. In time, such studies are likely to provide important insights into the mechanisms whereby neutrophil apoptosis is induced through the Fas system, and how neutrophil apoptosis/survival are modulated by inﬂammatory agents. D. CD30 LIGAND CD30, a member of the NGF/TNF receptor superfamily, is preferentially expressed by Th2-type CD4⫹ T cell clones as well as by CD8⫹ T cell clones showing a Th2 proﬁle of cytokine secretion (207). The natural ligand for CD30, CD30L, has been shown to be a type II transmembrane glycoprotein of the TNF ligand superfamily (207). CD30L is reportedly expressed by B lymphocytes, as well as by a subset of activated macrophages and T cells (207). We were among the ﬁrst to show that CD30 engagement induces NF-␬B activation in human T cells (208), a ﬁnding that has potential implications for the development of Th2 responses and the control of HIV replication (209). Gruss et al. (34) initially reported that PMNs constitutively express CD30L mRNA (but not CD30 mRNA), and that after 24 hr of stimulation with LPS, IFN-웂, or GM-CSF, the steady-state levels of CD30L mRNA were slightly increased. They also found that in reactive lymph nodes and tonsils, CD30L was expressed by a small subset of lymphoid cells, histiocytes, and neutrophils, as determined by immunostaining and ﬂow cytometry (34). Higher levels of CD30L expression were noted in Hodgkin’s disease (HD) lesions among bystander cells, which included T cells and PMNs (210). Remarkably, native CD30L displayed at the cell surface was functionally active, as shown by the ability of ﬁxed PMNs to interact with CD30-positive cell lines and to induce their proliferation (210). Our

NEUTROPHIL-DERIVED PROTEINS

399

unpublished experiments conﬁrm that neutrophils express high levels of CD30L mRNA, especially after stimulation with LPS or fMLP. In addition, we observed a variable CD30L protein expression pattern in neutrophils in immunoﬂuorescence analyses, depending on the donor. The biological implications of CD30L expression on neutrophils require further investigations. For instance, it is known that the interaction between CD30L and CD30 triggers the replication of HIV (209, 211). In addition, Ho et al. (212) showed that human neutrophils can potentiate HIV replication in infected mononuclear leukocytes, but did not identify the mechanisms or the molecules implicated in these effects. In the light of the potential ability of neutrophils to express CD30L, one obvious candidate for this interesting effect could be the CD30L. It has been shown that crosslinking of CD30L by a mAb or by a CD30–Fc fusion protein induces the production of IL-8 by freshly isolated neutrophils (35). These results indicate that cross-linked CD30L can transduce a signal to the ligandbearing cell. This ‘‘reverse signaling’’ via CD30L strongly suggest that TNF family members and their cognate receptors signal bidirectionally, blurring the distinction between ligand and receptor. E. INTERLEUKIN-1 Interleukin-1 is an important mediator of the host defense response to injury and infection, and has both protective and proinﬂammatory effects (213). The production and action of IL-1 are so important that they are regulated by multiple control pathways, some of which are unique to this cytokine, in a so-called IL-1 system (214). This IL-1 system consists of two agonists, IL-1움 and IL-1웁, a speciﬁc activation enzyme (IL-1-converting enzyme, ICE), a receptor antagonist (IL-1RA) produced in different isoforms, and two surface-binding molecules, the type I and type II IL-1 receptors (IL-1RI, IL-1RII) (213, 214). A wide variety of cells, including monocytes, macrophages, dendritic cells, lymphocytes, NK cells, endothelium, ﬁbroblast, glial cells, and keratinocytes, are able to secrete IL-1 (213). A vast phenomenological literature on the activities of IL-1 exists, and it is almost impossible to list them. IL-1 affects the hematopoietic system at various levels, from immature precursors to mature myelomonocytic and lymphoid elements. In particular, IL-1 costimulates T cell proliferation in the classic costimulator assay, induces cytokine production in monocytes, prolongs the in vitro survival of PMNs, activates endothelial cells in a proinﬂammatory, prothrombotic sense, and stimulates the release of corticotropin-releasing hormone (CRH) by the hypothalamus, ultimately causing a release of corticosteroids in the bloodstream by the adrenals (213) Furthermore, IL-1 is a key mediator of the acute-phase response, and is the main endogenous pyrogen, an activity shared with other cytokines,

400

MARCO ANTONIO CASSATELLA

including TNF-움 and IL-6. In vivo, IL-1 induces hypotension, fever, weight loss, neutrophilia, and acute-phase response. IL-1 has a number of local effects that have been termed catabolic, and plays a role in destructive joint and bone diseases (215), in addition to having an important toxic activity against insulin-producing 웁-cells in the Langerhans islets. In addition, a number of observations support a role of IL-1 in the pathogenesis of many diseases (215). Although the issue of whether or not PMNs synthesize and secrete IL1 has been a matter of debate for several years (216–218), molecular studies have clariﬁed this issue. There is no longer any doubt that neutrophils can be induced to produce small amounts of IL-1, not only in humans, but also in mice (219), rats (220), rabbits (221–223), and cattle (224). Reasons for past failures to detect IL-1 may include a poor sensitivity of the assays available at the time (217, 218), and/or inappropriate stimulatory conditions. For instance, it has been reported that muramyl dipeptides (MDPs) or their derivatives do not induce IL-1 activity in neutrophil supernatants or lysates (225). In that study, however, neutrophils were cultured in serum-free medium (225), a procedure that adversely affects neutrophil viability (226). In other cases, contaminating monocytes present in cell preparations have been proposed to account for the IL-1 production attributed to PMNs (217, 218). That human PMNs, produce an IL-1-like activity after stimulation with particulate and soluble agents, such as zymosan and PMA, respectively, was clearly documented by Tiku and colleagues (227). This IL-1-like activity was detected in the supernatants of PMNs only if cells were stimulated for at least 4–5 hr, and was also induced by A23187 plus cytochalasin B, as well as by LPS, albeit at much lower levels (227). Interestingly, though no IL-1-like activity was present in a preformed state or stored in PMN granules (228), an indirect indication that cell lysates of resting PMNs contained an endogenous inhibitor of IL-1 was also provided (228). Subsequently, neutrophils stimulated with GM-CSF were found to express the mRNA and to release both IL-1움 and IL-1웁 (9, 229, 230). Although PMNs produced both IL-1움 and IL-1웁 in response to GM-CSF, the number of cells producing IL-1움 was greater and they were more strongly stained than those secreting IL-1웁, as revealed by an ELISASPOT (229). What is more, exposure to GM-CSF of PMNs isolated from saliva did not result in a further increase of IL-1웁 production (231), implying that these cells were already maximally activated in situ, presumably in response to the oral ﬂora. Indeed, oral PMNs accumulated IL-1웁 mRNA and produced IL-1웁 in amounts severalfold greater compared to those released by peripheral PMNs activated in vitro with GM-CSF (231). Partial elucidation of the mechanisms whereby GM-CSF regulates IL-1웁 gene

NEUTROPHIL-DERIVED PROTEINS

401

expression in PMNs has been reported by Fernandez et al. (232), who demonstrated effects of GM-CSF at both the transcriptional and the posttranscriptional level. Human PMNs have also been shown to produce both forms of IL-1 in response to LPS, zymosan, and Con A, IL-1웁 being produced and released before IL-1움, over a 72-hr culture period (52, 233). Lord et al. (11) not only provided evidence that PMNs transcribe and translate the IL-1움 and IL-1웁 genes after stimulation with LPS or IL-1움, but clearly established that IL-1움 and IL-1웁 synthesis attributed to PMNs was not accounted for by the low level of contaminating PBMCs. Although virtually all PMNs showed diffuse staining with both IL-1움- and IL-1웁-speciﬁc antisera, synthesis of IL-1웁 exceeded that of IL-1움, and little or no IL-1움 was released by PMNs over an 18- to 24-hr time course (11). Another very interesting observation made by this group was that although increases in IL-1 mRNA after stimulation of PMNs and PBMCs with LPS were similar, PMNs were less efﬁcient than PBMCs in translating IL-1 mRNA (11). Because IL-1움 and IL-1웁 mRNA from either PMNs or PBMCs were translated with equal efﬁciency in rabbit reticulocyte lysates, it was speculated that synthesis of IL-1 in PMNs is subject to some form of translational control (11). Other researchers have once again questioned whether LPS-treated neutrophils produce IL-1 following endotoxin treatment (229, 234). In an ex vivo model of localized cytokine production that closely mimics in vivo conditions, that is, stimulation of human whole blood with LPS, Hsi and Remick (234) found that positivity for IL-1웁 in PMNs by immunohistochemistry was very low and did not increase over time. In the same model, monocytes displayed a marked IL-1웁 signal that increased by 4 hr of LPS stimulation (234). In contrast, several demonstrations by other groups not only conﬁrm that both LPS (51, 103, 124, 172, 235, 236) and GM-CSF (237) up-regulate the expression of IL-1웁 in human PMNs, but provided evidence that endogenous IL-1웁 exerts an essential role in prolonging neutrophil survival (238, 239). Increased expression of IL-1웁 mRNA and release determined by LPS and GM-CSF were also shown to be reduced by herbimycin A, a tyrosine kinase inhibitor, and pyrrolidine dithiocarbamate, an NF-␬B inhibitor (239), but not by diﬂuoromethylornithine (DFMO), a selective inhibitor of ornithine decarboxylase (240, 241). IL-1웁 mRNA accumulation is induced in a dose-dependent manner also by TNF-움 and/or IL-1웁 (10), but not by IL-2, IL-3, IL-4, IL-5, IL-6, or IFN-웂 (230). Though the relative levels of the cell-associated antigenic IL-1웁 induced by TNF-움 and/or IL-1웁 strongly correlated with the induction of IL-1웁 mRNA, no IL-1웁 antigen can be detected in the supernatants of PMNs incubated with or without cytokines for less than 4 hr (10). In agreement with these observations, we (124) and others (242) also found

402

MARCO ANTONIO CASSATELLA

that neutrophils produce and release IL-1웁 not before 5–6 hr following stimulation with TNF-움 (124). Interestingly, neonatal PMNs were shown to accumulate cell-associated antigenic IL-1웁 when stimulated by TNF움 and LPS, and this expression was consistently higher than that of adult PMNs (186). The elevated IL-1웁 expression by neonatal PMNs may be signiﬁcant for immune and inﬂammatory regulation in neonates. Table IV lists the stimuli able to trigger the production of IL-1웁 by PMNs. Included are superoxide anion (243), rheumatoid synovial ﬂuid (244), C. neoformans and glucuronoxylomannan (107), L. monocytogenes and Y. enterocolitica (183), C. albicans or MP-F2 (245), OK-432, an immunopotentiator prepared from streptococci (246), the nonoral organism Fusobacterium mortiferum (247), LPS from F. nucleatum and A. actinomycetemcomitans, P. gingivalis and C. ochracea (75), and EBV (121). Another condition inducing the mRNA expression and production of IL-1웁 by neutrophils is the phagocytosis of yeast particles (103). Though kinetics of IL-1웁, TNF-움 and IL-8 secretions in response to opsonized (Y-IgG) and unopsonized (Y) yeast particles were similar, the absolute amounts of IL1웁 detected in the supernatants of PMNs stimulated with yeasts alone were usually higher than those obtained with Y-IgG-stimulated PMNs, whereas the amounts of TNF-움 and IL-8 were lower (103). Similarly, serum-opsonized zymosan was more potent than LPS in triggering the TABLE IV AGENTS ABLE TO TRIGGER THE PRODUCTION OF IL-1 Cytokines and growth factors TNF-움 IL-1움 IL-1웁 GM-CSF Surface molecules Anti-CD32 (Fc웂RII) antibodies Particulate agents Calcium pyrophosphate dihydrate microcrystals Monosodium urate microcrystals Other agents or conditions PMA Concanavalin A Anticytoplasmic antibodies Synovial ﬂuid from rheumatoid arthritis Cellular density Matrix proteins (ﬁbronectin, laminin)

BY

HUMAN NEUTROPHILS

Bacteria and related products LPS from E. coli LPS from F. mortiferum LPS from A. actinomycetemcomitans LPS from F. nucleatum LPS from P. gingivalis LPS from C. ochracea Streptococcal OK-432 Staphylococcus aureus Yersinia enterocolitica Listeria monocytogenes Fungi and related products Cryptococcus neoformans and glucuronoxylomannan Candida albicans and derivatives Saccharomyces cerevisiae IgG-opsonized S. cerevisiae Zymosan Viruses Epstein–Barr virus

NEUTROPHIL-DERIVED PROTEINS

403

extracellular production of IL-1웁 (171). Even fMLP (100 nM ) has been shown to induce the release of detectable amounts of IL-1웁 by PMNs (74), and its effect, as well as that of LPS, was enhanced by the precoating of tissue culture plates with the matrix proteins Fn and Ln (74). In contrast to these ﬁndings, and similar to the case of fMLP-stimulated TNF-움 release, we (4, 248) and others (186) have never been able to measure any intra- or extracellular IL-1웁 in response to fMLP (10–100 nM ). However, we have repeatedly detected a transient induction of IL-1웁 mRNA accumulation in fMLP-treated PMNs (unpublished experiments). Brooks and co-workers (242) demonstrated that neutrophils can be stimulated to express IL-1웁 mRNA and protein by antineutrophil cytoplasmic antibodies (ANCAs), which are present in patients with systemic vasculitis. Both human ANCAs and mAbs to a variety of autoantigens recognized by ANCAs, including proteinase 3, myeloperoxidase (MPO), bactericidal/ permeability-increasing protein, and elastase, were effective (242). This response could be inhibited by Act D and CHX, but not by FK506 or cyclosporin A (242). Besides, human anti-MPO IgG F(ab)2 were completely without effect on IL-1웁 production (and on superoxide production), suggesting the requirement for Fc receptor involvement. Indeed, the mAb Fab IV.3 (anti-Fc웂RII) induced IL-1웁 production from neutrophils treated with normal serum (242). Because neutrophils accumulate in the acute blood vessel lesions of patients with autoimmune systemic vasculitis, the data suggest that, in these individuals, ANCAs may stimulate PMNs to release IL-1웁, other than activating neutrophils to produce reactive oxygen radicals and to degranulate (249). The possibility that neutrophil-derived IL-1웁 may contribute to augment the local inﬂammatory response by the activation of vascular endothelial cells and inﬁltrating leukocytes must thus be considered for therapeutic intervention. The capacity of neutrophils to synthesize and release IL-1움/웁 species after phagocytosis of MSU and CPPD, alone, or in combination with TNF움 and GM-CSF, was tested in other studies (250, 251). Phagocytosis of MSU and CPPD crystals by joint neutrophils is in fact associated with acute gout and pseudogout (252). PMNs are the major, if not the only, cell type within the synovial ﬂuid in the initial phase of such crystal-induced attacks. It was demonstrated that MSU crystals are more potent inducers of both IL-1움 and IL-1웁 generation compared to CPPD or unopsonized zymosan (250). With all three stimuli, the synthesis of IL-1웁 was 5- to 14fold greater than that of IL-1움, and up to 10 times more IL-1웁 than IL1움 was usually secreted (250). Colchicine, a speciﬁc and effective drug in the clinical treatment of acute gout, partly inhibited the secretion of IL1 by neutrophils during phagocytosis of solid particles (250). The results suggested that neutrophil-derived IL-1 contributes to the pathogenesis of

404

MARCO ANTONIO CASSATELLA

crystal-induced arthritis. Subsequently (251), it was observed that treatment of PMNs with GM-CSF and TNF-움 before incubation with suboptimal concentrations of the crystals enhanced the total synthesis of microcrystal-induced IL-1 by ﬁve- to ninefold, over a 12-hr incubation period. Under the same conditions, the levels of total IL-1RA produced by neutrophils were approximately 35 to 43% lower than the expected amounts produced by cytokine-treated cells (251). As a result of this shift in IL-1 and IL1RA production by GM-CSF- and TNF-움-activated neutrophils, the net biologic activity of IL-1 secreted in response to the microcrystals was enhanced. Treatment of neutrophils with colchicine prior to incubation with GM-CSF or TNF-움 inhibited the crystal-induced IL-1 production by 50 to 55%, but failed to affect that of IL-1RA. In this situation, therefore, the IL-1RA to IL-1 ratio increased signiﬁcantly by 185–220% (250). These results not only demonstrated that IL-1 and IL-1RA production by human neutrophils is differentially regulated, but that the combined presence of GM-CSF or TNF-움 and microcrystals favors the production of biologically active IL-1 over that of IL-1RA, thereby potentially amplifying the inﬂammatory response associated with crystal-induced diseases. In addition, because colchicine selectively inhibited IL-1 without affecting IL-1RA production, a novel putative site of action of this drug was uncovered. Finally, neutrophils from healthy subjects were analyzed under culture conditions of low (5 ⫻ 106/ml) and high (100 ⫻ 106/ml) cell density for 18 hr, to investigate the effect of cellular density on IL-1웁, 5-lipoxygenase (5-LO), and 5-lipoxygenase-activating protein (FLAP) gene expression (253). Only high-density conditions lead to the induction of IL-1웁 gene at the RNA and protein levels (within 2.5 hr), as measured by RT–PCR and by immunoprecipitation. In contrast, 5-LO mRNA accumulation decreased in cells cultured at high density, whereas FLAP mRNA was not affected (253). Interestingly, experiments with synovial ﬂuid neutrophils obtained from rheumatoid arthritis patient demonstrated that the pattern of IL-1웁, 5-LO, and FLAP production was similar to that observed in peripheral blood neutrophils cultured at high density (253). These results suggest that cellular density may play a role in gene modulation when neutrophils are accumulating at an inﬂammatory site at high density. In so many situations neutrophils can produce IL-1움 and IL-1웁, with profound effects during inﬂammation and, as a consequence, pathophysiological implications. Release of IL-1 in concert with TNF-움 may trigger a number of effects, such as recruitment and regulation of leukocye function, priming of neutrophils for oxidative activity, phagocytosis, degranulation and stimulation of the production of arachidonic acid metabolites, regulation of ﬁbroblast, osteoclast, and synovial cell activation, induction of fever, and so forth (1, 213). These actions may fulﬁll immunomodulatory functions

NEUTROPHIL-DERIVED PROTEINS

405

that may be beneﬁcial for the host, but, if dysregulated, may contribute to the pathogenesis of the diseases in which IL-1 is involved. F. INTERLEUKIN-1 RECEPTOR ANTAGONIST The IL-1 receptor antagonist has been characterized as a 23- to 25-kDa glycosylated protein made by the same cells that secrete IL-1 (254). IL1RA is an anticytokine that speciﬁcally inhibits the proinﬂammatory actions of IL-1 by binding to IL-1RI and IL-1RII located on a target cells, without initiating any signal transduction or biological activity (255). Different structural variants of IL-1RA have been described: the secreted form(s) produced by mononuclear cells, and two intracellular isoforms (icIL-1RA) found in keratinocytes and other epithelial and myelomonocytic cells (254). The biological function played by the cell-associated fraction of IL-1RA remains largely obscure. The ﬁrst reports suggesting that neutrophils might secrete products featuring IL-1 inhibitory activity were those by Tiku and colleagues (227, 228), but the molecular demonstration that this IL-1 inhibitory activity corresponded to IL-1RA was provided later by McColl’s group (14). While studying the neutrophil production of an IL-1-like factor (227), Tiku’s group observed that the addition of PMNs to monocytes cultured in the presence of zymosan led to a decreased biological IL-1 activity of the resulting supernatants, relative to those harvested from monocytes stimulated in the absence of PMNs (228). A possible explanation was that PMNs released an inhibitor of IL-1. A better characterization of this PMN-derived IL-1 inhibitory activity revealed that (1), it was constitutively present either in lysates of freshly isolated PMNs or in cell-free supernatants obtained from PMNs stimulated or not for 18 hr, (2) it could be generated in the absence of serum and it was not produced as a result of the activity of neutrophil-proteases, and (3) it was constitutively contained in greater amounts than IL-1 (228). The latter observations might provide another explanation for the difﬁculties reported by various investigators in detecting the production of IL-1 by PMNs, especially when measured in biological assays (see above). McColl’s group not only conﬁrmed the data that neutrophils constitutively produce IL-1RA (14), but also showed that following activation with TNF-움 and GM-CSF, PMNs express the mRNA and secrete increased amounts of IL-1RA protein over a 24-hr period (approximately 0.5 ng/ml/106 cells) (14). The same authors concluded that IL-1RA constitutes one of the most abundant de novo-synthesized proteins of activated neutrophils. They also calculated that IL-1RA is produced by neutrophils in excess of IL-1 by a factor of at least 100; such a ratio ﬁts perfectly with the reported amounts of IL-1RA needed to inhibit the proinﬂammatory effects of IL-1 (213, 254, 255). In addition, they found that none of a large

406

MARCO ANTONIO CASSATELLA

array of classical neutrophil stimuli (fMLP, G-CSF, IL-1웁, IL-4, TGF-웁, or IFN-웂), including LPS, inﬂuenced the extracellular production of IL1RA by PMNs. All of them were used over a 24-hr period, and at concentrations ranging from sub- to supraoptimal doses (14). The reasons that McColl and co-workers were not able to identify LPS as an effective stimulus for the induction of IL-1RA synthesis (14) or de novo RNA synthesis (256) by neutrophils remain intriguing. Possible factors that could have hampered the detection of any modiﬁcation of neutrophil gene expression in response to LPS are the conditions used to isolate and culture PMNs—for instance, inappropriate use of serum-free culture medium. It is well known that in order to stimulate phagocytes at low doses, LPS necessitates binding different serum proteins, which in turn bind to CD14 (257). However, it has also been demonstrated that high doses of LPS (⬎100 ng/ml) can bypass the requirement for serum proteins, and alone can efﬁciently stimulate the target cells (257). Finally, it is also possible that the particular batch or strain of LPS used by McColl’s group was not properly working. That PMNs express increased levels of IL-1RA mRNA and protein after incubation with LPS was originally demonstrated by Ulich and co-workers (15). Importantly, extent of IL-1RA mRNA expression found in LPStreated PMNs was in magnitude similar to, or even greater than, that found in PBMCs (15). LPS proved to be a very efﬁcient stimulus for the production of IL-1RA in PMNs also in our hands, being much more potent than Y-IgG (248, 258). Consistent with a poor ability of Y-IgG to induce IL-1RA in neutrophils, Malyak et al. (235) also reported that adherent IgG did not trigger the production of IL-1RA from PMNs. It is noteworthy that the stimulation of human monocytes with LPS or IgG, these latter either in a soluble form (259) or attached to plastic surfaces (260), led instead to approximately the same amount of IL-1RA production. These results suggest that the ability of PMNs and monocytes to secrete IL1RA in response to speciﬁc agonists, acting presumably through the same receptors, is regulated by distinct mechanisms. Whether this difference between monocytes and neutrophils is due to a greater ability of the monocyte Fc receptors to generate intracellular signals for expression, translation, and secretion of IL-1RA remains to be investigated. In this respect, data by Chang et al. (261) indicated that adherent anti-Fc웂RIII (CD16) mAbs, but not anti-Fc웂RI or anti-Fc웂RII mAbs, induce the production of IL-1RA from PMNs. Elegant studies performed by Colotta and collaborators have shown that in addition to LPS and GM-CSF (48), also IL-4 (48), IL-13 (49), and TGF-웁1 (262) are efﬁcient inducers of IL-1RA mRNA expression and secretion in PMNs. These studies have also revealed that in PMNs, either IL-13 or TGF-웁1 induces the mRNA expression of an intracellular form

NEUTROPHIL-DERIVED PROTEINS

407

of IL-1RA (icIL-1RA) (49, 262). The observation that TGF-웁 (262, 263) is able to trigger the production of IL-1RA might constitute one of the mechanism(s) whereby IL-1RA is constitutively produced by PMNs (14, 228, 235, 258). TGF-웁 activities seem in fact also to be constitutively released by PMNs (16, 17), and could therefore stimulate the same cells to produce IL-1RA through an autocrine loop. Colotta’s group cloned and characterized a new intracellular isoform of IL-1RA expressed in PMNs and other cells, which was termed icIL-1RA type II (icIL-1RAII) (264). icIL-1RAII was found to be potentially biologically active, and to differ from the previously known icIL-1RA (icIL-1RAI) only by an additional stretch of 21 amino acids located within the NH2-terminal portion of the molecule (264). Table V summarizes the list of the agents able to induce IL-1RA release by PMNs in vitro. These include A23187 and zymosan (228), thapsigargin (94), MSU, CPPD (251), 10 nM fMLP (248), and P. falciparum-infected erythrocytes (109). But there are other experimental conditions in which IL-1RA is expressed in neutrophils. For instance, functional studies have shown that stimulation of PMNs with a number of periodontopathic bacteria results in the production of an IL-1 inhibitor. These bacteria include P. gingivalis, Bacteroides forsythus, A. actinomycetemcomitans, F. nucleatum, and the nonoral F. mortiferum (247). More recently, these data have been TABLE V AGENTS ABLE TO TRIGGER THE PRODUCTION OF IL-1RA Cytokines and growth factors TNF-움 IL-1웁 IL-4 IL-13 GM-CSF TGF-웁1 IL-10(?) Chemoattractants fMLP Calcium ionophores Ionomycin, A23187 Surface molecules Anti-CD16 (Fc웂RIII) antibodies Particulate agents Calcium pyrophosphate dihydrate microcrystals Monosodium urate microcrystals

BY

HUMAN NEUTROPHILS

Bacteria and related products LPS from E. coli LPS from A. actinomycetemcomitans LPS from F. nucleatum LPS from P. gingivalis LPS from C. ochracea Bacteroides forsythus Fusobacterium mortiferum Fungi and related products Candida albicans Saccharomyces cerevisiae IgG-opsonized S. cerevisiae Zymosan Protozoa Plasmodium falciparum-infected erythrocytes Viruses Epstein–Barr virus Other agents or conditions PMA Thapsigargin

408

MARCO ANTONIO CASSATELLA

conﬁrmed, in that substantially great amounts of IL-1RA were found to be released from PMNs stimulated with LPS deriving from P. gingivalis, A. actinomycetemcomitans, C. ochracea, and F. nucleatum (75). In addition, the inhibitory effects of IL-1RA on the biological activity of IL-1 in PMN supernatants were examined by the thymocyte comitogen proliferation assay. PMN supernatants stimulated with these lipopolysaccharides, and especially these from P. gingivalis or C. ochracea, showed less biological IL-1 activity as compared with the same doses of recombinant human IL1웁 detected by ELISA (75). These ﬁndings demonstrated that LPS from periodontopathic bacteria are able of stimulating PMNs to release not only proinﬂammatory cytokines such as IL-1, TNF-움, and IL-8 (75), but also antiinﬂammatory substances, such as IL-1RA. Therefore, different secretion levels of these cytokines and their biological activities induced by the various lipopolysaccharides might be important in the onset and progression of periodontal diseases (265). EBV is another agent that alters the gene expression and protein synthesis of IL-1 and IL-1RA in human peripheral blood neutrophils (121, 266). EBV induces a rapid accumulation of IL-1 and IL-1RA mRNA in neutrophils, followed by a later appearance of considerable synthesis of IL-1움, IL-1웁, and IL-1RA proteins. In response to EBV, neutrophils secreted approximately 3200 and 610 times more IL-1RA than IL-1움 or IL-1웁, respectively (121). Pretreatment of cells with CHX or phosphonoacetic acid (known to inhibit viral DNA polymerase activity) did not abrogate the effect of EBV, suggesting that EBV does not penetrate the cell, but that a virion’s structural molecule is required to induce such an effect. In this respect, neutralization of the viral particles with a monoclonal antibody reacting with glycoprotein gp350 of the viral envelope inhibited the production of IL-1 and IL-1RA. GM-CSF differentially modulates EBV-induced IL-1 and IL-1RA production by neutrophils, in that it synergistically enhances the production of IL-1움 and IL-1웁, while it poorly affects IL-1RA synthesis (267). However, more than 94% of the IL-1움 and IL-1웁 produced remain cell associated (267). The elevated levels of IL-1RA produced by EBVstimulated PMNs might be useful for the virus to induce immunosuppression and to avoid rejection at the early stage of its infectious process. Conversely, the effect of GM-CSF may enhance host response by altering the ratio of IL-1RA to IL-1 in favor of IL-1. In conclusion, the ﬁndings that PMNs, in addition to exerting a series of proinﬂammatory activities, may also produce IL-1 receptor agonist suggest that these cells can also mediate antiinﬂammatory actions. Because IL-1RA blocks the activities of IL-1 both in vitro and in vivo, the production of relatively large amounts of IL-1RA by neutrophils could be of major

NEUTROPHIL-DERIVED PROTEINS

409

biological signiﬁcance in that it could contribute to inhibition of established inﬂammatory or immune responses mediated by IL-1 (213, 254). G. INTERLEUKIN-6 IL-6 is a multifunctional cytokine that exerts multiple effects on many different types of target cells (268). It is a terminal differentiation factor for B cell maturation, a growth factor for myeloma, plasmacytoma, and hybridoma cells, T cells, mesangial cells, and megakaryocytes, and (depending on the context) a pro- or antiinﬂammatory factor (268). IL-6 shares several activities with IL-1 and TNF-움, insofar as it potently induces the acute-phase response in the liver, has pyrogenic properties, and can stimulate the proliferation of hematopoietic progenitor cells, from which osteoclastic cells are derived. IL-6 is synthesized by a wide variety of cells, including ﬁbroblasts, T and B cells, macrophages, endothelial cells, astrocytes, and keratinocytes; when puriﬁed from natural sources, IL-6 consists of a single chain of 21–28 kDa (268). There is no general consensus among researchers involved in neutrophil research about whether or not human PMNs express or release IL-6. This controversial issue can be said to have originated in two articles that were published in 1990. In one of them (269), it was shown that following stimulation of whole blood with LPS or Con A, the IL-6-expressing cells were exclusively monocytes. In the other article, Cicco et al. (270) reported that stimulation of PMNs with GM-CSF, and to a lesser extent with TNF움, leads to a rapid accumulation of IL-6 mRNA and subsequent secretion of the protein, as measured in a bioassay. In addition to LPS, PMA and CHX were also described as inducers of IL-6 transcripts in PMNs (270). In disagreement with the latter study, we provided clear evidence for the total lack of IL-6 gene expression by neutrophil preparations that are free of contaminating monocytes, whether or not these PMN preparations are cultured for up to 18 hr in the presence of LPS or Y-IgG (8). Though it could be argued that our failure to detect IL-6 in neutrophils (as opposed to the autologous PBMCs) might reﬂect limits in the sensitivity of Northern blot analysis, it should be pointed out that IL-6 mRNA transcripts were equally undetectable following RT–PCR analysis of neutrophil preparations, be it in unstimulated cells (90) or following PMN treatment with LPS, GM-CSF, or CHX (M. P. Russo et al., unpublished observations). In keeping with our observations, and with the initial report by Kato and colleagues (269), several other studies have emphasized that the detection of IL-6 transcripts in human PMNs might be accounted for by contaminating monocytes. In particular, it was demonstrated that if monocyte contamination of neutrophil preparations is kept below 0.8%, then IL-6 release from PMNs is undetectable, even following LPS stimulation (51). Con-

410

MARCO ANTONIO CASSATELLA

versely, LPS stimulation was shown to be a potent inducer of IL-6 gene expression (51), as shown previously by us (8) and many other groups before us. In another study, in which peripheral blood PMNs monocytes, and lymphocytes were stimulated with zymosan, Con A, or LPS (52), both monocytes and lymphocytes were observed to secrete high levels of IL-6, whereas no immunoreactive IL-6 was detectable in PMNs. In that study, the authors’ use of a very sensitive radioactive RT–PCR analysis further revealed that though higher mRNA levels of IL-1움, IL-1웁, and TNF-움 were expressed in zymosan-stimulated PMNs than in cells stimulated with LPS or Con A, IL-6 mRNA remained undetectable under any of these conditions (52). Again, IL-6 transcripts were highly expressed in activated autologous monocytes and lymphocytes (52). Identical results were observed in PMNs isolated from the gingival crevicular ﬂuid from patients affected by periodontitis (52), or from alveolar bone-derived PMNs (178). On a ﬁnal note, other investigators also reported on the inability of neutrophils to secrete IL-6 in response to less common stimuli. For instance, PBMCs were found to secrete large amounts of IL-6 in response to P. falciparum-infected erythrocytes, whereas autologous granulocytes did not (109). Similarly, stimulation of highly pure PMN preparations with thapsigargin led to the release of IL-8, GRO-움, and IL-1RA, but not IL-6 (94). The studies reviewed thus far strongly suggest that PMNs do not express or secrete IL-6, and that any such effect is artifactual, but there nevertheless exist several published articles supporting the opposite conclusion. For instance, Mianji and colleagues reported that LPS induces IL-6 mRNA expression and release in PMNs (271). Curiously, in this latter work, neither PMA nor GM-CSF had an effect on IL-6 mRNA abundance (271), in contrast with other ﬁndings (237, 270). Similarly, Palma et al. (172) reported a time-dependent release of IL-6 in response to LPS, which reached a maximum after 24 hr; substantial levels of IL-6 transcripts were also detected in resting neutrophils (172). Though other investigators reported that no IL-6 mRNA signal is detectable in freshly isolated neutrophils (as determined by RT–PCR) (272), stimulation with LPS or zymosan for 1 or 3 hr nevertheless resulted in the accumulation of IL-6 mRNA (45). In the latter study, however, no IL-6 protein could be detected in the corresponding culture medium (45). In contrast, zymosan was reported to stimulate the IL-6 secretion in PMNs isolated from normal subjects, but not from breast cancer patients (273). Similarly, PMNs were reported to release large amounts of IL-6 following phagocytosis of yeasts (C. neoformans) or of bacteria such as L. monocytogenes, Y. enterocolitica, and E. coli, or even following exposure to RSV particles (61, 107, 183). Furthermore, release of IL-6 was also found with neutrophils stimulated for 24 hr with E. coli LPS serotype B5 or B12 and, to a much lesser extent,

NEUTROPHIL-DERIVED PROTEINS

411

with bacterial extracts from A. actinomycetemcomitans or P. gingivalis (274). Taken together, these studies indicate that LPS and phagocytic stimuli can promote IL-6 mRNA accumulation in PMNs; with the exception of one study (45), it also appears that the IL-6 protein is released in response to the same stimuli. Finally, IL-6 mRNA was detected by RT– PCR in PMNs after Fc웂RI or Fc웂RII cross-linking, and after in vitro or in vivo treatment with G-CSF (275). The IL-6 protein was synthesized and secreted at very low levels, as determined by ﬂow cytometric analysis of intracellular proteins and ELISA, respectively (275). A synergistic increase in IL-6 secretion was also observed when PMNs isolated from patients treated with G-CSF were stimulated in vitro by anti-Fc웂RI or anti-Fc웂RII cross-linking (275). It is, however, well known that antigenic Fc웂RI is nearly undetectable on freshly isolated neutrophils of normal donors (276, 277). On the basis of the aforementioned studies, it therefore appears that in human PMNs, certain stimulatory conditions do result in IL-6 gene expression and release. Among the proponents of this idea, however, a far less obvious conclusion is the issue of whether unstimulated neutrophils express IL-6 mRNA, and similarly, whether the IL-6 protein is constitutively released. Indeed, some of these investigators report that IL-6 mRNA and protein are undetectable in resting neutrophils (13, 45), yet others found high constitutive levels of IL-6 transcripts (172), or high IL-6 release in resting PMNs from both normal donors and breast cancer patients (273). In this regard, Melani et al. (272) reported that freshly isolated circulating human granulocytes (98% neutrophils, 1.2% eosinophils, 0.8% mononucleated cells) constitutively express IL-6 mRNA (as determined by RT–PCR or ISH). Interestingly, the constitutive expression of the IL-6 gene was very weak when neutrophils were isolated from buffy coats instead of fresh, heparinized peripheral blood. Furthermore, IL-6 expression became undetectable if PMN suspensions were incubated as brieﬂy as 60 min on ice, or at 37⬚C in Dulbecco’s modiﬁed Eagle’s medium. Yet IL-6 expression was still inducible by the addition of GM-CSF (50 ng/ ml) for another hour (272), as observed by others (237). Thus it appears that PMN isolation procedure, or the different anticoagulants used, can substantially affect the constitutive expression of IL-6 in PMNs (272). Though this study has the distinguished merit of shedding some light on the confusion surrounding the issue of IL-6 gene expression in unstimulated neutrophils, its ﬁndings regarding the inducibility of the IL-6 gene by GMCSF are in complete disagreement with those of Mianji et al. (271), who reported that neither GM-CSF nor PMA had any effect toward IL-6 mRNA levels in human neutrophils. In summary, the issue of whether IL-6 expression and release occurs in human neutrophils remains an open question, based on the available

412

MARCO ANTONIO CASSATELLA

literature. This being said, there exists compelling evidence emphasizing the need to exclude false positives due to contaminating monocytes (8, 51). Similarly, the potential role of contaminating eosinophils should also be taken into account; ﬁrst, because the proportion of eosinophils in various neutrophil preparations typically ranges from 2 to 5%, and second, because eosinophils have been reported to express IL-6 mRNA actively and to release the related protein in considerable amounts (272, 278, 279). Finally, it was reported that though PMNs from normal donors exhibit little IL-6 mRNA (RT–PCR) and protein even after LPS stimulation, the production of IL-6 is three to six times more elevated in PMNs from HIV⫹ patients at all stages of infection, both under basal conditions and after stimulation with LPS (13, 102). Thus, the discrepancies observed in studies addressing the production of IL-6 in human neutrophils may well be related to the presence of contaminating eosinophils and/or monocytes in different neutrophil preparations, or perhaps (to a lesser extent) also related to whether HIV⫹ individuals are unknowingly used as blood donors. Alternatively, it is also conceivable that particular stimuli (or pathological conditions) might speciﬁcally induce IL-6 in human PMNs. Whatever the case may be, further studies are clearly required before any deﬁnitive conclusions can be drawn about the ability of human PMNs to express or release IL-6. In contrast to their human counterparts, the ability of murine neutrophils to produce IL-6, both in vitro and in vivo, appears to be undisputed (see below). H. ONCOSTATIN M Oncostatin M (OSM), a member of the IL-6 family of cytokines that is synthesized by activated human T lymphocytes and monocytes, was originally identiﬁed as a growth regulator for certain tumor and non-tumorderived cells (280). Although the physiological role of this cytokine remains largely undeﬁned, OSM has been shown to promote the stimulation of acute-phase protein synthesis, as well as the induction of IL-6 expression, in several cell types (280). Another study showed also that OSM stimulates several PMN functions, including their transmigration across conﬂuent monolayers of primary endothelial cells (281). Conversely, the potential ability of neutrophils to serve as a source of OSM was recently investigated (281). Highly puriﬁed PMNs were found to contain preformed OSM in intracellular stores that were rapidly mobilized by GM-CSF or PMA. In addition, PMNs were also observed to produce OSM after a few hours of stimulation with various agonists. The most potent effect was observed following costimulation with LPS and GM-CSF, a combination that upregulated both mRNA and protein levels of OSM. Interestingly, OSM production was down-regulated by dexamethasone, but not by IL-10. The

NEUTROPHIL-DERIVED PROTEINS

413

OSM produced by PMNs was also found to be biologically active, as demonstrated by its ability to induce 움1-acid glycoprotein synthesis by HepG2 cells (282). I. INTERLEUKIN-12 IL-12, a heterodimeric factor of 70 kDa (p70), is formed by two covalently linked glycosylated chains of approximately 40 and 35 kDa, respectively, whose coexpression is necessary for biologic activity (283). IL-12 is produced mainly by mononuclear phagocytes, B cells, neutrophils, and dendritic cells (283). The biological activities of IL-12 are mainly directed toward NK cells and T lymphocytes, in which it induces the production of lymphokines (particularly IFN-웂); it enhances cell-mediated cytotoxicicity in addition to exerting mitogenic effects. IL-12 is also required for the optimal differentiation of cytotoxic T lymphocytes and appears to have an obligatory role as inducer of differentiation of T helper lymphocytes toward the Th1 type, as well as suppression of IgE synthesis (283, 284). The ﬁrst demonstrations that mature human neutrophils produce and release both the IL-12 p40 free chain and the IL-12 p70 heterodimer originated from my laboratory (12). Our experiments showed that among a wide range of stimuli tested, only LPS in combination with IFN-웂 efﬁciently induced a signiﬁcant production of biologically active IL-12, whereas in monocytes LPS was effective by itself (12). This was attributed, on the one hand, to the fact that LPS induced a 100-fold increase in IL-12 p40 mRNA without having an effect on IL-12 p35 mRNA accumulation, and on the other hand, to the fact that IFN-웂 directly induced a severalfold increase in the accumulation of IL-12 p35 mRNA and enhanced the LPSinduced accumulation of IL-12 p40 mRNA. Therefore, the combined effect of LPS and IFN-웂 induced sufﬁcient expression of both IL-12 p40 and IL-12 p35 mRNA to achieve production of the biologically active IL-12 p70 heterodimer, at physiologically relevant concentrations (12). A study by Cassone et al. (13) has fully conﬁrmed and extended our initial ﬁndings, by showing that neither LPS nor MP-F2, used alone, was able to induce biologically active IL-12 production by PMNs, whereas these two microbial-derived imunomodulators induced monocytes to release appreciable levels of IL-12 (13). However, when used in combination with IFN웂, both MP-F2 and LPS were capable of inducing PMNs to produce IL-12 (13). In agreement with these observations, several studies have established that the genes encoding the two IL-12 subunits are independently regulated (285–287). As previously shown in monocytes (288), or for other proinﬂammatory cytokines produced by neutrophils (289), we also found that IL-10 suppressed the LPS-induced IL-12 p40 mRNA and the secretion of IL-12 in PMNs costimulated with LPS and IFN-웂 (12).

414

MARCO ANTONIO CASSATELLA

In a subsequent study, we additionally observed that the release of IL-12 induced by LPS plus IFN-웂 did not signiﬁcantly differ for PMNs from HIV⫹ patients and HIV⫺ individuals (290). In autologous PBMCs, however, the production of IL-12 occurring in response to LPS, or LPS plus IFN웂, was signiﬁcantly higher in cells isolated from HIV⫹ patients (290). Expression and production of IL-12 have also been reported in murine neutrophils stimulated in vitro and in vivo, using different experimental approaches, including immunoﬂuorescence staining and microscopic analyses (291, 292). Thioglycollate-elicited neutrophils cultured for 24 hr in the presence of IFN-웂 plus LPS or with C. albicans (either the agerminative live vaccine strain PCA-2, which provokes healing infections, or the highly virulent CA-6 strain, which induces nonhealing infections), were tested for cytokine production (291, 292). Low levels of bioactive IL-12 and IL10 were found in PMN cultures stimulated with IFN-웂 plus LPS. IL-12, but not IL-10, was produced in response to the agerminative strain PCA2, whereas the opposite pattern occurred in response to CA-6 (291). A similar cytokine production proﬁle was observed in stimulated neutrophils isolated from peripheral blood (291). As revealed by RT–PCR, IL-12 p40 and IL-10 messages, together with those of TNF-움 and IL-6, were present in elicited neutrophils cultured with LPS plus IFN-웂 for 2 hr (292). In contrast, under the same stimulatory conditions, the IL-6 and TNF-움 messages, but not those encoding IL-10 and IL-12 p40, were present in peritoneal macrophages (291). The release of IL-12 by murine neutrophils stimulated by LPS plus IFN-웂 was greatly enhanced if the cells were preincubated with IL-4 (293). Although we did not observe a similar effect of IL-4 in human neutrophils (our unpublished observations), similar ﬁndings were reported in the case of human monocytes (294). PMNs also produce IL-12 in other situations. For instance, peritoneal neutrophils and splenic macrophages harvested from iron-overloaded mice and stimulated in vitro with IFN-웂 and LPS were tested for candidacidal activity and expression of inducible nitric oxide synthase (iNOS) and IL12 p40 genes in another study (295). Iron overload greatly increases murine susceptibility to disseminated infection with low-virulence C. albicans cells (295). The results showed that iron treatment reduced the candidacidal activity from both uninfected and infected mice. In contrast, modulation of expression of iNOS and IL-12 p40 genes was dependent on the types of cells and on the infectious state. Iron loading caused the disappearance of both types of messages in neutrophils from uninfected or infected mice, as well as in macrophages from uninfected mice (295). Likewise, the in vitro production of NO and IL-12 was impaired in neutrophils (295). However, treatment of iron-overloaded mice with desferoxamine (an iron chelator) cured the infection, and restored both the candidacidal potential

NEUTROPHIL-DERIVED PROTEINS

415

and the appearance of iNOS and IL-12 p40 messages in neutrophils (295): consequently, a Th1 protective antifungal response took place (295). Neutrophils isolated from TNF/LT-움 double-deﬁcient mice were stimulated with IFN-웂 plus LPS in vitro, showing no impairment and releasing IL12 whether the cells were obtained from uninfected animals or from mice infected with C. albicans (296). Enriched populations of dendritic cells, macrophages, and neutrophils (80% pure), isolated from the peritoneal cavity of mice following thioglycollate administration, all responded to UV-inactivated or live herpes simplex virus (HSV), as well as to LPS, by up-regulating the expression of IL-12 p40 mRNA (297). These experiments were performed to clarify which types of cells are responsible of the IL-12 expression and release in the cornea and draining lymph node on ocular infection with HSV (297). The data indicated that among the inﬂammatory cells that inﬁltrate the cornea in response to HSV-1 infection, the neutrophil is the prodominant cell type, and might therefore represent an important source of IL-12. In light of the important immunoregulatory functions of IL-12, in particular the induction of IFN-웂 production and facilitation of Th1-type responses, the ability of neutrophils to produce IL-12 suggests that these cells may play an active role in the regulatory interactions between innate resistance and adaptive immunity. Production of IL-12, TNF-움, IL-1웁, and other molecules by PMNs, and of IFN-웂 by NK cells, suggests that the activation of innate resistance early during infection represents not only a ﬁrst line of defense against invading microorganisms, but also that it could prime the immune system and determine the characteristics of the ensuing antigen-speciﬁc adaptive immune response. It is even possible to imagine a dynamic situation in which IL-12 produced by PMNs (and other phagocytic cells) in response to bacterial or parasitic infections induces the production of IFN-웂 by T and NK cells, which in turn activates the phagocytic cells (298) and at the same time favors a Th1-type immune response. If so, IL-12 could be involved in an effective positive feedback mechanism that, through IFN-웂, could induce the activation of PMNs and monocytes. As such, IL-12 might represent a bridge between innate resistance and adaptive immunity. Evidence of the ability of neutrophils to mount Th1/Th2 responses in vivo is described below. J. INTERFERON-움 The interferons are a family of closely related inducible secreted proteins; they are important not only in defense against a wide range of viruses, but also in the regulation of the immune response and in hematopoietic cell development (299). Though there is only one IFN-웂 and one IFN-웁 gene, to date there are 13 IFN-움 functional genes reported, encoding as many

416

MARCO ANTONIO CASSATELLA

IFN-움 protein species (300). IFN-움 is produced by most cells (but mainly by leukocytes) in early response to viral infections, following stimulation with natural or synthetic double-stranded RNA such as poly(I:C), as well as in response to bacterial, mycoplasma, and protozoan infections, and to certain cytokines (299). In addition to inducing a potent antiviral state in uninfected cells, IFN-움, at higher concentrations, has antiproliferative activity against both normal and tumor cells and is important in regulating cell-mediated immunity as well as some aspects of humoral immunity, because can enhance, for example, the expression of class I MHC gene products, natural killer cell activity, and macrophage activation (299). The ﬁrst exhaustive molecular demonstration of the ability of neutrophils to express and release the IFN-움 protein was reported by Shirafuji et al. (28). Prior to that, however, another study had showed that PMN-enriched fractions possessed very low levels of IFN-움 mRNA transcripts, as measured by S1-nuclease mapping, but contained no data at the protein level (301). By Northern blot analysis, Shirafuji et al. (28) were able to demonstrate that G-CSF induced the mRNA for IFN-움 type 1 in neutrophils, but not in PBMCs. Moreover, a radioimmunoassay speciﬁc for IFN-움 revealed that the levels of IFN-움 in the culture media of G-CSF-treated neutrophils rose, in a time-dependent manner, up to 100 U/ml per 107 cells after 12 hr, whereas those of unstimulated cells remained at about 20 U/ml. Under the same experimental conditions, neither LPS (10 ng/ ml) nor fMLP (1 nM ) effectively stimulated the expression of IFN-움 mRNA or protein in neutrophils. It is worth recalling that LPS and fMLP, unlike G-CSF (82), induce the expression of IL-8 in PMNs (60). In view of the recently demonstrated ability of IL-8 to inhibit IFN-움 activity and to attenuate the IFN-mediated inhibition of viral replication (302), the selective production of IFN-움 by G-CSF-treated neutrophils might represent a speciﬁc effector response against viral infection. Other studies have conﬁrmed that PMNs produce IFN-움. In one of these, Brandt et al. (29), using RT–PCR, were able to show that PMNs accumulate IFN-움 mRNA either in a constitutive manner or on infection with Sendai virus, being IFN-움1, IFN-움2, and IFN-움4, the species predominantly expressed in these cells (29). A biological assay revealed that the levels of antiviral activity of supernatants recovered from PMNs stimulated with Sendai virus were very similar to those detected in MNCs, and, surprisingly, much more abundant than those measured from puriﬁed T and B cells (29). In earlier studies (303), bovine neutrophils were also shown to produce a factor with antiviral, IFN-like properties. This IFNlike factor appeared in detectable amounts 12 to 18 hr after exposure of PMNs to fragmented herpesvirus-infected cells (303), and could not be neutralized by antibodies to bovine IFN-움, -웁, and -웂 (304). Neutrophils

NEUTROPHIL-DERIVED PROTEINS

417

infected with BHV-2, herpes simplex virus type 2, equine herpesvirus 1, bovine respiratory syncytial virus, bovine viral diarrhea virus, or parainﬂuenza virus 3 were unable to produce this IFN-like factor (304). Although only limited attention has been paid to the role of neutrophils in viral infection, the fact that neutrophils can be signiﬁcant sources of IFN움 or IFN-like molecules emphasizes the potentially important role of PMNs in host defenses against viral infection, or in other biological processes in which interferons are implicated. Furthermore, because IFN-움 inhibits neutrophil colony formation in vitro (300), release of IFN-움 by PMNs may be also viewed as a feedback regulatory mechanism for neutropoiesis. K. INTERFERON-웂 IFN-웂, also known as type II interferon or immune interferon, is a glycoprotein mainly secreted by T cells and NK cells having weak antiviral and antiproliferative activity. Through its ability to activate macrophages to become tumoricidal and to kill intracellular parasites, IFN-웂 is responsible for inducing nonspeciﬁc cell-mediated mechanisms of host defense (305). IFN-웂 induces the expression of many key molecules, including MHC class I and II antigens, cell surface ICAM-1, nitric oxide synthase, and chemokines (305). Other than being a macrophage-activating factor, evidence has shown that IFN-웂 is also a neutrophil-activating factor (306). IFN-웂 induces the surface expression of Fc웂RI, enhances neutrophil antibody-dependent cell cytotoxicity (ADCC), potentiates the generation of reactive oxygen intermediates (ROI) and granule release in response to different stimuli, and modulates neutrophil microbicidal activity (306). Clinically, IFN-웂 has been shown to alleviate the symptoms of chronic granulomatous disease and to exhibit potential beneﬁts in the treatment of infectious disease and neoplasia (307). Evidence that PMNs might synthesize IFN-웂 has been provided by Yeaman et al. (308). Highly puriﬁed human PMNs were cultured for 18 hr with or without G-CSF, in the presence or absence of IL-12, and then examined by confocal microscopy. An accumulation of speciﬁc IFN-웂 immunoreactivity in the PMN population cultured either in the presence or in the absence of G-CSF was found, although the staining was more intense in cells treated with IL-12 (308). IFN-웂 positivity was present only in one-third of the total PMN population, and was dependent on brefeldin A treatment (308). To determine whether IFN-웂 was secreted, cell-free supernatants were assayed after incubation of neutrophils with LPS, IL12, TNF-움, and IL-12 plus TNF-움 for 24 hr. ELISA testing revealed increased levels of IFN-웂 in supernatants collected from LPS-, TNF-움-, and especially IL-12 plus TNF-움-treated neutrophils (308). These ﬁndings were supported by other observations on the production of IFN-웂 in

418

MARCO ANTONIO CASSATELLA

human endometria, by using confocal microscopic analysis (308). Extracellular IFN-웂 was mostly associated with matrix components, and was located immediately beneath the luminal epithelium and along the glandular epithelium proximal to the lumen. As evidenced by intracellular staining, IFN-웂 was present in both stromal cells and intraepithelial lymphocytes through all stages of the menstrual cycle (308). Interestingly, the stromal cell containing intracellular IFN-웂 was identiﬁed as a polymorphonuclear neutrophil on the basis of its nuclear morphology (308). As a result, it was suggested that polymorphonuclear neutrophils play an important role in determining immune responsiveness within the female reproductive tract. Though this may turn out to be true, other investigators failed to detect IFN-웂 in supernatants from neutrophils stimulated for 24 hr with LPS (309). Thus, additional studies are clearly required before any further speculation can be made on the potential biological signiﬁcance of IFN웂 production by PMNs. L. TRANSFORMING GROWTH FACTOR-움 TGF-움 is a polypeptide belonging to the family of EGF-related proteins (310). It is secreted by several cell types, including cells of hematopoietic origin, such as eosinophils (311) and stimulated macrophages (312). TGF움 elicits different effects on target cells, such as mitogenic signaling and promotion of neovascularization, and it is also involved in wound healing and in tumor development (310). Calafat and colleagues (313), using immunoelectron microscopy, have clearly demonstrated that human neutrophils and monocytes store TGF움 in cytoplasmic vesicles. Interestingly, a previous report (314) also showed that granulocytes stimulated with fMLP plus cytochalasin B released immunoreactive TGF-움, but even though neutrophils were not excluded as possible source, it was concluded that eosinophils were the major manufacturers of TGF-움. In neutrophils, TGF-움 was localized in electron-dense vesicles, whereas in monocytes TGF-움 was found in the peroxidasenegative granules (313). No colocalization of TGF-움 with components of azurophilic, speciﬁc, or gelatinase granules, or secretory vesicles, was observed in neutrophils. This implied that TGF-움 is stored in a novel, not previously described compartment. Other experiments indicated that the TGF-움 gene was expressed early in myeloid development, and the protein was stored in granules. Monocytes and neutrophils were also pelleted and lysed to quantify TGF-움 using ELISA. Monocytes were found to contain approximately 257 pg of TGF-움/106 cells, but neutrophils contained much less, 2.5 pg of TGF-움/106 cells. The ﬁndings that neutrophils contain TGF-움 suggest that these cells might contribute not only to the beneﬁcial effects caused by chronic in-

NEUTROPHIL-DERIVED PROTEINS

419

ﬂammatory diseases, such as wound healing, but also to the development of complications, such as ﬁbrosis. M. TRANSFORMING GROWTH FACTOR-웁 The TGF-웁 family of cytokines is composed of several ubiquitous molecules that are multifunctional and essential to survival (315). Several TGF웁 isoforms (TGF-웁1, TGF-웁2, and TGF-웁3) exist, and these are secreted as latent precursors whose activation is achieved through the action of proteases. TGF-웁 has multiple cell surface receptors, of which at least two mediate signal transduction (315). Although TGF-웁 was originally discovered as a secreted factor that induced malignant transformation in vitro, it was subsequently shown to play important roles in inﬂammation, growth, and development, in would repair, and in host immunity (315). As a general but by no means exclusive rule, TGF-웁 serves as a conversion factor, converting an active inﬂammatory site into one dominated by resolution and repair (316). Grotendorst and colleagues were the ﬁrst to report that cultured human neutrophils constitutively express TGF-웁1 mRNA and secrete high levels of the protein (16). Secreted TGF-웁1 was found to be structurally similar to that which is released by human platelets, and appeared to be in a fully active form, because acid treatment did not increase the amount of TGF웁 activity (16). After incubation of neutrophils with LPS, fMLP, or immune complexes for 24 hr, no difference in the levels of TGF-웁 transcripts or protein were detected relative to untreated cells (approximately 50 ng/ml/ 106 cells) (16). Though stimulation of monocytes resulted in a strong increase of TGF-웁 secretion, unstimulated PMNs secreted approximately ﬁve times more TGF-웁-like activity than an equal number of unstimulated monocytes (16). Although the authors correctly hypothesized that the constitutive TGF-웁 secretion by neutrophils might have been due to neutrophil activation during the isolation procedure, these results bring about additional considerations. First, the production of TGF-웁 in neutrophils and monocytes seems to be differentially regulated, and second, TGF-웁 production in monocytes might be controlled by posttranscriptional mechanisms. Our unpublished observations conﬁrm that neutrophils constitutively express high levels of TGF-웁1 mRNA that do not change in LPS-, IL-10-, or fMLP-activated cells. Similarly, TGF-웁1 transcripts were found to be constitutively expressed in synovial ﬂuid neutrophils and to be unaffected by cell exposure to GM-CSF (237). That PMNs represent an important potential source of TGF-웁 in inﬂammatory inﬁltrates was conﬁrmed by Fava et al. (17). They reported that a preexisting TGF-웁 bioactivity could be released from freshly isolated peripheral PMNs, if incubated with PMA at 37⬚C for 30 min. The identity

420

MARCO ANTONIO CASSATELLA

of this TGF-웁 activity as TGF-웁1 was conﬁrmed by the use of a speciﬁc ELISA. However, in contrast with the ﬁndings of Grotendorst et al. (16), Fava et al. (17) found a little detectable TGF-웁1 activity released from unstimulated PMNs. Such discrepancies might be explained by the fact that neutrophils were cultured for only 30 min (17), as opposed to 24 hr (16). Worthy of note is that to date, no study has examined whether PMNs speciﬁcally or differentially express TGF-웁2 or TGF-웁3 mRNA and proteins. This important aspect obviously requires urgent investigation. The capacity of neutrophils to produce bioactive TGF-웁 suggests that PMNs may play an important role in situations characterized by inﬂammation and ﬁbrosis, such as in chronic immunodriven inﬂammations and immune responses, in wound repair, or in the pathogenesis of ﬁbrotic disease. Furthermore, in view of its ability to exert chemotactic activities toward neutrophils and monocytes, the release of TGF-웁 by PMNs could promote strong local proinﬂammatory effects (317–319). N. VASCULAR ENDOTHELIAL GROWTH FACTOR VEGF/vascular permeability factor (VPF), originally identiﬁed as a tumor cell-derived substance, is a multifunctional cytokine that exerts a key role in physiological and pathological neoangiogenesis, by stimulating endothelial cell proliferation and vessel hyperpermeability (320, 321). VEGF has also a major role in the pathogenesis of many diseases, including rheumatoid arthritis, cutaneous diseases, proliferative retinopathies, and hypervascularized tumors (320, 321). VEGF is a glycosylated heparinbinding homodimer that exists as one of four different isoforms, VEGF 121, VEGF 165, VEGF 189, and VEGF 206, and is produced by several cell types, including monocytes, macrophages, T lymphocytes, keratinocytes, and ﬁbroblasts (320–322). Several studies have demonstrated that human neutrophils represent a source of VEGF (36–38). In the ﬁrst one (36), neutrophils isolated from blood have been found to release VEGF into the cell-free supernatants in response to ionomycin, phorbol dibutyrate, and serum-opsonized zymosan (SOZ) particles. The magnitude of the response varied with the stimulant applied and was dose and time dependent in a period of examination of 2 hr. By immunohistochemistry, it was also demonstrated that neutrophils inﬁltrating inﬂammatory lesions, including synovial biopsies from rheumatoid arthritis, periapical dental granulomas, chronically inﬂamed gingival tissues, and mucocele and nonspeciﬁc oral ulcerations, contain VEGF (36). In another contemporary article (37), not only PMA, but also fMLP and TNF-움, were shown to trigger a time-dependent secretion of VEGF by human neutrophils. PMA-treated neutrophils released signiﬁcant amounts of VEGF after 15 min, and these remained elevated over a period of 2 to

NEUTROPHIL-DERIVED PROTEINS

421

24 hr. CHX had no effect on the PMA-induced secretion of VEGF, suggesting that the PMA-induced secretion of VEGF acted on a preformed pool of this molecule. The existence of an intracellular pool of VEGF was conﬁrmed by measuring the intracellular content of VEGF in resting neutrophils (37). Whereas basically 100% of the intracellular VEGF content was released in response to PMA, approximately 28 and 31% were released by adherent neutrophils in response to fMLP and TNF-움, respectively (37). Kinetics of fMLP- and of TNF-움-induced secretion of lactoferrin and VEGF were similar, suggesting the intracellular localization of VEGF in the speciﬁc granule fraction (37). Accordingly, subcellular fractionation of human neutrophils and Western blot analysis showed a granule-speciﬁc distribution of the intracellular pool of VEGF in resting neutrophils (37). More recently, RT–PCR studies showed that PMNs express mRNA for the two most common VEGF splice variants, VEGF 121 and VEGF 165 (38). However, it was not determined which variant is the predominant VEGF form secreted by PMNs. Taken together, these observations suggest that neutrophil-derived VEGF may be instrumental in the development and progression of angiogenic and other vascular phenomena in inﬂammatory processes and tumors. Through the production of VEGF and other angiogenic factors such as TGF-움, IL-8, and TNF-움, neutrophils may signiﬁcantly inﬂuence the evolution of inﬂammatory pathologies involving modulation of vascular permeability as seen in ischemia–reperfusion injury, rheumatoid arthritis, and PMN-mediated glomerulonephritis. Furthermore, in view of the established chemoattractant properties of VEGF for monocytes (323), neutrophils may reinforce the recruitment of monocytes to sites of acute inﬂammation following the initial inﬂux of PMNs. O. HEPATOCYTE GROWTH FACTOR HGF, also known as scatter factor, is a heparin-binding heterodimer related to plasminogen; it was ﬁrst recognized as a molecule that stimulates hepatocyte proliferation (324). It is now known to be a cytokine with numerous functions, including stimulatory effects on hepatocytes, melanocytes, and epithelial cells, induction of endothelial cells and hematopoietic cell growth, chemotactic and chemokinetic activities, activation of increased motility and scattering of cell colonies, and modulation of organ development (324). HGF has also a role in wound repair, liver and kidney regeneration, and tumor propagation (324). It is produced as a single-chain precursor protein and is proteolitically cleaved to generate the mature molecule, and its gene expression and synthesis is regulated by several factors. HGF binds with high afﬁnity to its receptor, c-met, whose intracellular domain has tyrosine kinase activity and transduces all the effects of the cytokine

422

MARCO ANTONIO CASSATELLA

(325). HGF can be produced by various cells, including ﬁbroblasts, epithelial and endothelial cells, malignant cells, and neutrophils (324). PMN capacity to express HGF was discovered in studies aimed to deﬁne and compare the cellular localization of hepatocyte growth factor in human and rat tissues (39). For this purpose, parafﬁn-embedded sections and frozen sections were examined by immunohistochemistry using a polyclonal antiserum to hepatocyte growth factor (39). The distribution of hepatocyte growth factor was almost identical in humans and rats, and, among a variety of cell types, strong cytoplasmic immunoreactivity for hepatocyte growth factor was found to be present in granulocytes (39). Subsequently, by electron microscopy studies performed in tissue samples obtained from livers of control subjects and patients affected by acute and chronic hepatitis and cirrhosis, it was conﬁrmed that PMNs represent a source of circulating HGF (40). P. GRANULOCYTE– COLONY-STIMULATING FACTOR AND OTHER GROWTH FACTORS Lindemann et al. (30) reported that stimulation of PMNs with GM-CSF induces the mRNA expression and release of M-CSF and G-CSF. Two other groups have reported that neutrophils cultured with LPS for 24 hr release signiﬁcant amounts of G-CSF (31, 309). Remarkably, in one study (31), the endogenous release of G-CSF was correlated with the maintenance effect of LPS on neutrophil superoxide generation and survival. GCSF is produced by bone marrow stromal cells, endothelial cells, and ﬁbroblasts in response to speciﬁc stimuli, and acts on both granulocyte precursors and mature cells (326). G-CSF not only enhances the proliferation of granulocyte precursors in bone marrow, but also primes mature neutrophils for superoxide anion production, phagocytosis, and antibodymediated cytotoxicity against targets such as tumor cells, and delays their apoptosis (202, 326). Various observations in the literature indicate that neutrophils can express and release other cytokines, but the conclusions reached are not always concordant. For example, in one of these cases, Ramenghi et al. (327) evaluated by RT–PCR the cell types in peripheral blood that express the mRNA for stem cell factor (SCF), the ligand of the c-kit protooncogene. Only granulocytes, but not puriﬁed lymphocytes, monocytes, or the total cell population obtained by Ficoll puriﬁcation, appeared to express transcripts encoding both the soluble and transmembrane forms of SCF (327). In contrast, in identical types of experiments, we could demonstrate neither constitutive nor induced SCF mRNA, nor secretion of the related protein (our unpublished observations). Kita and co-workers (328), on the basis of effects detected in biological assays, indirectly assumed that neutrophils

NEUTROPHIL-DERIVED PROTEINS

423

stimulated with PMA in combination with ionomycin (a calcium ionophore) produce GM-CSF and IL-3. They were, however, unable to measure IL3 and GM-CSF directly in neutrophil culture supernatants, by speciﬁc immunological techniques (328). Similarly, other researchers (186, 309) failed to detect antigenic GM-CSF in cell lysates and supernatants from PMNs stimulated for up to 24 hr with TNF-움, LPS, fMLP, or IL-1웁. Conversely, other studies have reported the detection of measurable quantities of GM-CSF in supernatants harvested from PMNs stimulated with MSU microcrystals (121), or cultured with plasma from patients with severe burns (329). There also exist several other reports on the ability of neutrophils to produce additional factors whose identity has been only partially characterized to date. For instance, a novel cytokine mitogenic for connective tissue cells, termed ‘‘leukocyte-derived growth factor’’ (LDGF), was identiﬁed in supernatants of LPS-activated neutrophils and monocytes, and PHAtreated T cells (330). The LDGF cDNA hybridized to a 1.3-kb transcript present in total RNA isolated from activated human neutrophils, macrophages, and T lymphocytes, but not from ﬁbroblasts, bovine smooth muscle cells, or umbilical vein endothelial cells. The peptide produced by T cells exhibited electrophoretic mobility identical to the peptide produced by macrophages, but the protein produced by neutrophils migrated slightly slower (330). Sequence analysis of the LDGF peptide revealed that this polypeptide is a precursor of other known peptides, including platelet basic protein (PBP), connective tissue activating peptide III (CTAP-III), and neutrophil-activating peptide 2 (NAP-2). However, none of these shorter peptides is active as a mitogen for ﬁbroblasts. This suggests that the LDGF gene produces multiple peptides possesing divergent biological activities that could regulate different phases of the repair process. The fact that PMNs may produce LDGF in addition to other biological mediators that orchestrate the cellular events during wound healing (TGF-웁, VEGF) reinforces the concept that neutrophils are much more important than originally believed in the recruitment and activation of connective tissue cells to wound sites. Human PMNs, but not PBMCs, have been shown to secrete constitutively a mediator designated ‘‘granulocyte-derived factor’’ (GDF) (331); the GDF levels did not change after treatment of PMNs with activating agents such as phorbol ester, calcium ionophore, poly(I:C), and IFN-웂, or metabolic inhibitors such as actinomycin D and cycloheximide. This factor was shown to enhance the uptake of [3H]thymidine by the Molt-3, CTV-1, and K562 leukemic cell lines in a dose-dependent manner, and to amplify the level of thymidine kinase activity in the sensitive cells (331). However, cell number and the rate of DNA synthesis in GDF-responsive

424

MARCO ANTONIO CASSATELLA

cultures remained unchanged (331). GDF had an apparent molecular mass of about 40 k Da, was heat and acid labile, and was shown to be distinct from a panel of different lymphokines and monokines (IL-1, IL-2, IL-4, IL6, TNF-움, GM-CSF, IFN-움) in antigenicity, biochemical, and functional characteristics (331). Although GDF is possibly a novel cytokine that can alter the pattern of DNA synthesis and growth characteristics of certain hematopoietic cells, its biologic and physiologic signiﬁcance remains to be clariﬁed. Neutrophils from other species also produce speciﬁc growth factors, A series of studies on neutrophils from the inﬂamed peritoneal cavity of rabbits demonstrated that these cells produce a trypsin-sensitive polypeptide factor able to affect corneal endothelial cells. This factor, ‘‘corneal endothelium modulation factor’’ (CEMF) (332), is a polypepiyde that affects cell shape and collagen gene expression in corneal endothelial cells. The modulation consists of phenotypic switches from a polygonal cell shape to a ﬁbroblastic morphology and from basement membrane collagen (type IV-rich) synthesis to ﬁbrillar collagen (type I-rich) synthesis. Basic ﬁbroblast growth factor (bFGF) supplemented with heparin was able to induce effects identical to those of CEMF (333), thus a subsequent study examined whether bFGF was the direct mediator for corneal endothelium modulation and whether CEMF played a role in inducing bFGF production (334). Neither exogenous bFGF nor CEMF alone caused induction of bFGF mRNA in corneal endothelial cells, but simultaneous treatment with bFGF and CEMF selectively enhanced the 4.9-kb transcripts encoding bFGF (334). When protein synthesis was inhibited by cycloheximide, bFGF synthesis was blocked in the presence of CEMF, and as a consequence corneal endothelium modulation was inhibited. Furthermore, bFGF antisense oligonucleotides blocked by 50% the enhanced growth potential mediated by bFGF induced with CEMF. Taken together, these ﬁndings suggested that de novo synthesis of bFGF induced by CEMF is required for corneal endothelium modulation (334). Culture medium or homogenates of casein-induced rat peritoneal PMNs were shown to induce markedly the production of collagenase and PGE2 by normal synovial cells. Collagenase activity and PGE2 induction were abolished by IL-1움-neutralizing antibodies and PMN-derived culture medium contained relatively small amounts of IL-1움 (335). However, the effects induced in synovial cells by recombinant rat IL-1움 were less than those induced by rat PMN culture medium. A combination of endogenous IL-1움-deleted PMN-derived supernatant and recombinant IL-1움 restored the synovial cell collagenase and PGE2 production to the levels of the original PMN-derived culture medium (335), thus proving that PMN culture medium contained not only IL-1움, but also an additional factor syner-

NEUTROPHIL-DERIVED PROTEINS

425

gizing with IL-1움 (335). This new factor, which was subsequently shown to be a negatively charged protein of about 80 kDa (ruling out that is a member of the IL-6 or epidermal growth factor families) (335), may have important roles in connective tissue destruction and chronic inﬂammation in diseases such as rheumatoid arthritis. Inﬂammatory neutrophils collected from rat peritoneal exudates were shown to secrete a factor with hepatocyte-stimulating activity (HSF) (336); this factor induced 움2-macroglobulin (움2M) synthesis in primary rat hepatocytes (337). Unfortunately, it was not assessed whether HSF corresponds to IL-6. Various other studies showed that inﬂammatory PMNs collected from the pleural cavity of rats released factors that enhanced the phytohemagglutinin (PHA)-induced proliferative response of normal lymph node cells or thymocytes (338). These lymphocyte-stimulating activities were found in PMN supernatants and in their lysates, and were signiﬁcantly upregulated following pretreatment of neutrophils with Con A (338, 339). However, the nature of the latter activities (338, 339) and of other similar factors (340–344) was not further elucidated. Q. NEUTROPHINS Laurenzi et al. (345) examined the expression of neurotrophin and neurotrophin receptor mRNA in human granulocytes and bone marrow cells. By RPA and RT–PCR, granulocytes were found to express constitutively mRNA encoding nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and neurotrophin-4 (NT-4), but not neurotrophin-3 (NT3). Expression of NT-4 (but not NGF or BDGF) mRNA was speciﬁcally up-regulated by LTB4 (345). Whether the reported accumulation of neurotrophin mRNA was accompanied by the synthesis of the corresponding proteins was not investigated. Such information is urgently required. R. CYTOKINES THAT ARE NOT PRODUCED BY NEUTROPHILS Human neutrophils do not produce RANTES, MCP-2, MCP-3, or I309 (25), IL-2 (233), or IL-10 or IL-13 (50, 309, 346). A very elegant demonstration was actually provided in the case of IL-10 and IL-13 (50). Human PMNs, puriﬁed by immunomagnetic depletion of class II-positive cells, were found unable to synthesize, store, or release IL-10 and IL-13 in numerous activatory conditions. Various stimuli were utilized: LPS and/ or TNF-움 with or without IFN-웂 or GM-CSF, fMLP, PAF, LTB4, PMA, opsonized or unopsonized zymosan, IL-1웁, IL-6, IL-8, IL-10, IL-4, IL13, TGF-웁, OSM, leukemia inhibitory factor, and keratinocyte growth factor, all of which were used for up to 72 hr. Lack of IL-10 and IL-13 production was conﬁrmed by the absence of IL-10 and IL-13 mRNA expression by RT–PCR. The addition of protease inhibitors, which was

426

MARCO ANTONIO CASSATELLA

used to detect IL-10 in supernatants of stimulated murine neutrophils (291, 292), did not allow the detection of IL-10 in human PMN cultures (50). This was consistent with clinical studies on patients with acute distress respiratory syndrome (ARDS), in which alveolar PMN culture supernatants contained no IL-10 (347). Very importantly, however, studies performed with PMN suspensions isolated by means of classical methods (dextran sedimentation followed by Ficoll centrifugation) and containing 1 to 2% of monocytes and lymphocytes showed detectable production of IL-10 and IL-13 (50). The latter experiments might explain some reports of (348) the presence of IL-10 in supernatants from LPS-stimulated PMNs, whereby IL-10 plays a negative role in the regulation of prostanoid production induced by LPS. IV. Production of Cytokines by Neutrophils Isolated from Individuals Affected by Different Pathologies

Investigations on the capacity of neutrophils to produce cytokines in human pathological conditions are not as numerous as those regarding mononuclear cells. Because neutrophils are abundant in the circulation and are readily accessible to experimental investigations, more information will likely appear in the future. A concise description of what has been reported to date is presented below. A. PRODUCTION OF IL-8 IN VARIOUS DISEASES Some researchers have examined the capacity of PMNs to produce IL8 in the context of various diseases. Kuhns et al. (349) observed that during the evolution of the inﬂammatory response associated with skin lesions raised by suction, IL-8 reached levels up to 175 ng/ml in the media bathing the lesions. Accumulation of IL-8 strictly correlated with the accumulation of the exudative neutrophils at this inﬂammatory site (63). These neutrophils exhibited 100-fold greater levels of cell-associated IL-8, and spontaneously released up to 50-fold more IL-8 than freshly isolated peripheral blood neutrophils from the same donors (63). These data indicate that neutrophils migrating to an inﬂammatory focus have up-regulated their production of IL-8, suggesting that, by releasing IL-8, PMNs may play a relevant role in the autocrine regulation of the inﬂammatory response. Neutrophils isolated from the sputum of patients with cystic ﬁbrosis (CF) were found to release elevated amounts of IL-8 and nitrite, PMNs from the sputum of subjects with active disease were more active than corresponding cells from stable subjects (350). Accordingly, signiﬁcant levels of IL-8 and nitrite were present in the soluble phase of sputum of CF subjects, once again at much higher levels in subjects with disease

NEUTROPHIL-DERIVED PROTEINS

427

exacerbation (350). IL-8 and nitric oxide produced by neutrophils and other cells could therefore represent novel therapeutic targets to block lung damage in CF. Accordingly, long-term erythromycin treatment of patients affected by chronic airway disease (CAD) have proved to be effective, due to a moderate inhibitory effect of the drug on the production of IL-8 by Pseudomonas-stimulated neutrophils (115). Involvement of PMNs as mediators of lung injury in the pathogenesis of pulmonary complications associated with cocaine abuse has been supported by Baldwin et al. (85), who investigated whether controlled in vivo administration of cocaine altered the function of neutrophils. It was found that exposure to cocaine in vivo enhanced the antibacterial and antitumor activities of PMNs, but also potentiated their ability to release IL-8 (85). The role of IL-8 in inducing neutrophil accumulation in the nasal discharge of patients with chronic sinusitis was examined by Suzuki and colleagues (351). Chronic sinusitis is a common inﬂammatory paranasal sinus disease, clinically characterized by edematous hypertrophy of the paranasal mucosa and persistent purulent nasal discharge and paranasal sinus effusion with selective and vigorous recruitment of neutrophils. By immunohistochemistry (IH) and in situ hybridization, the researchers examined specimens obtained from two groups of patients, those with chronic sinusitis and those with allergic rhinitis. The IL-8 levels in nasal discharge were signiﬁcantly higher in the chronic sinusitis group compared to the allergic rhinitis group (351). Immunoreactivity to IL-8 was observed in emigrated PMNs from nasal smears, in nasal gland duct cells, and in epithelial cells of the chronic sinusitis group, whereas those of the allergic rhinitis group mostly showed little or no reaction. Similar patterns of localization were shown by ISH for IL-8 mRNA. The results of Suzuki et al. (351) suggest that chemotactic factors in sinus effusion, including IL8 derived from nasal gland duct cells and epithelial cells, ﬁrst attract neutrophils out of the mucosa, and then neutrophils that have emigrated into the sinus effusion secrete IL-8. This might induce a further neutrophil accumulation in the sinus of chronic sinusitis patients. Though another article conﬁrms that neutrophils from asthmatic or atopic dermatitis patients do not express signiﬁcantly more IL-8 than control neutrophils (352), a different study showed that PMNs isolated from the sputum of subjects with bronchiectasis produced elevated amounts of IL-8, IL-1웁, and, to a lesser extent, TNF-움 (353). In the latter situation, an autocrine loop involving IL-1웁 was responsible for the maintenance of IL-8 production by PMNs within the bronchus lumen (353). Lin and Huang evaluated the IL-8 gene expression and release by PMNs and peritoneal macrophages (PMs) during peritonitis caused by S. aureus, in uremic patients on continuous ambulatory peritoneal dialysis (CAPD)

428

MARCO ANTONIO CASSATELLA

(354). Their previous study indicated that IL-8 was detectable in drain dialysate of these patients during the early acute stage of peritonitis, at variable levels depending on the microorganism type (355). These authors revealed that both the IL-8 levels and the amount of IL-8 mRNA expression were highly correlated with the PMN count found in the drain dialysate, because both were high at the onset of peritonitis and then decreased progressively (354). However, PMs expressed more IL-8 mRNA than did PMNs. These data suggest that through the release of IL-8, PMNs may also be considered potential contributors to the pathogenesis of peritoneal injury. Different groups of investigators (356, 357) examined the expression of neutrophil-speciﬁc chemoattractants in psoriatic lesions. Dense focal accumulation of neutrophils in the upper epidermis is a hallmark of psoriasis (358), but the signals for neutrophil diapedesis and migration in the disease are not fully understood. The studies provided evidence for a differential expression of three neutrophil-attracting chemokines, namely IL-8, GRO-움, and ENA-78. ISH and IH of serial sections of psoriatic lesions were employed to identify and localize the cells producing these chemokines. A series of observations were made. Along with keratinocytes, neutrophils residing in microabscesses in the stratum corneum expressed IL-8 messages at high levels (356). As opposed to IL-8, GRO-움 was highly expressed by single cells in the papillary dermis (vessel-associated cells and inﬁltrating cells) (356, 357). GRO-움 mRNA expression was highly variable and, in the upper epidermidis, GRO-움 and IL-8 mRNA were typically coexpressed by clusters of keratinocytes (356, 357). Though mRNA expression of the highly homologous chemokine ENA-78 was absent (357), focal expression of GRO-움 and IL-8 in the epidermis was associated with a focal inﬁltration of neutrophils (356, 357). However, GRO-움 hybridization patterns did not correspond to the distribution patterns of neutrophils, indicating that neutrophils were not a major source for GRO-움 (356, 357). Taken together, the data indicated that both IL-8 and GRO-움 are important chemoattractants for neutrophil diapedesis in vivo, and that migration of neutrophils and formation of micropustules appear to be inﬂuenced by the cooperative action of both IL-8 and GRO-움. B. INFLAMMATORY BOWEL DISEASE Neutrophils are important cellular mediators in inﬂammatory bowel diseases (IBDs) such as ulcerative colitis (UC) and Crohn’s disease (CD) (359). To visualize the distribution of IL-8 in the intestinal mucosa and to understand its possible role in the induction and promotion of IBD, Mazzucchelli and colleagues (360) analyzed the expression of the IL-8 gene in resected bowel segments of 14 patients with active CD or UC. In

NEUTROPHIL-DERIVED PROTEINS

429

situ hybridization revealed, in the affected mucosa, strong and speciﬁc IL8 signals that correlated with the histological grade of active inﬂammation. IL-8-expressing cells were mainly macrophages, neutrophils, and epithelial cells, and in patients with UC were diffusely distributed over the entire affected mucosa, whereas in patients with CD, IL-8-expressing cells showed a focal distribution pattern (360). Neutrophils very likely represented the main IL-8-expressing cell population at the site of active IBD, at the base of ulcers, in inﬂammatory exudates on the mucosal surfaces, in crypt abscesses, and at the borders of ﬁstulae (360). These ﬁndings were in agreement with previous studies showing an association between increased IL-8 and myeloperoxidase levels in corresponding regions of the colonic IBD mucosa (361). IL-8 gene expression in uninﬂamed intestinal tissue resected for colon carcinoma and in inﬂamed colonic tissue resected from IBD was also studied by Grimm and colleagues (362). IL-8 mRNA was detected by ISH in macrophages and neutrophils adjacent to ulceration in inﬂamed bowel, as well as in many neutrophils in lamina propria, ulcer slough, and crypt abscesses, but was not detected in uninﬂamed mucosa displaying intact epithelium or from carcinoma resections (362). IL-8 protein, as detected by IH, was present in the same distribution as IL-8 mRNA. Recently recruited macrophages were also responsible for some of this IL-8 expression, whereas epithelial cells in normal and inﬂamed tissue showed neither IL-8 message nor IL-8 protein. These data, further supported by other more recent ﬁndings (363, 364), suggest that neutrophils and recently recruited macrophages are responsible for production of IL-8 in IBD, and thus may trigger a mechanism for a continuing cycle of neutrophil attraction and activation. TNF-움 and IL-1웁 are also implicated in the initiation and perpetuation of intestinal inﬂammation in IBD (365). Beil and colleagues (366) used an ultrastructural immunogold morphologic and morphometric analysis to identify the cellular and subcellular sites of TNF-움 in colonic biopsies obtained from patients with CD. They found that neutrophils, in addition to eosinophils, macrophages, mast cells, ﬁbroblasts, and epithelial and Paneth cells, expressed TNF-움. The cytokine was not present in cytoplasmic granules of neutrophils, but it was associated with the membranes of Golgi structures and cytoplasmic vesicles, or in lipid bodies (366). More recently (346), PMNs isolated from blood of patients with IBD were shown to release increased levels of TNF-움 IL-1웁, and IL-1RA in response to LPS compared to normal control neutrophils. The enhanced secretion of TNF-움 and IL-1웁 did not appear to be speciﬁc for IBD because PMNs from patients with infectious colitis also showed increased release of proinﬂammatory cytokines (346). It was concluded that PMNs are an important source of proinﬂammatory cytokines in patients with intestinal inﬂamma-

430

MARCO ANTONIO CASSATELLA

tion, and therefore deserve more attention in the immunopathogenic mechanisms of these diseases (346). C. NEUTROPHIL-DERIVED CYTOKINES IN RHEUMATOID ARTHRITIS AND SYSTEMIC LUPUS ERYTHEMATOSUS Rheumatoid arthritis, a systemic autoimmune disease characterized by chronic inﬂammation of the synovium, often leads to the destruction of articular cartilage and juxtaarticular bone. Although the etiology of RA is unknown, considerable evidence suggests that cytokines play a critical role in its pathogenesis (367). In view of the fact that synovial ﬂuid from patients with active RA are heavily inﬁltrated with neutrophils (85–95% of the total cell population), several groups have investigated whether neutrophils can be a potential source of cytokines. PMNs possess the greatest capacity to induce cartilage degradation directly and therefore to act as mediators in the pathogenesis of tissue damage observed in RA (161). Malyak et al. (368), found a very strong correlation between SF IL-1RA levels and the number of neutrophils present in these ﬂuids. Synovial ﬂuid PMNs contained preexisting IL-1웁 and IL-1RA proteins in the absence of detectable mRNA, and both LPS and GM-CSF induced modest increases in IL-1웁 and IL-1RA mRNA and protein in cultured SF PMNs, as well as in normal cultured PMNs (368). Thus, with regard to the IL1웁 and IL-1RA proteins, PMNs isolated from inﬂammatory SF were found to be qualitatively similar to PMNs from normal peripheral blood, and, therefore not to be activated in vivo. Interestingly, SF samples from patients with noninﬂammatory arthropathies contained undetectable levels of IL-1RA, and therefore it was inferred that normal SF does not contain IL-1RA (368). The authors therefore concluded that neutrophils might signiﬁcantly contribute to the total IL-1RA levels in SF in patients with active RA. The ability of RA blood-derived and SF-derived PMNs to produce IL-1움 and IL-1웁, in the absence or presence of LPS, was also assessed by Dularay et al. (229). In contrast with the previous results (368), no production of IL-1웁 (or IL-1움) by SF PMNs was found, whereas blood PMNs from three of eight RA patients produced IL-1. Opposite results were reported also by Edwards and colleagues (237, 369). This group initially reported that in RA patients SF PMNs, but not blood PMNs, contained levels of IL-1웁 mRNA ranging from 0.5 to 3% of the maximal levels that could be induced by GM-CSF treatment of blood neutrophils (237). Subsequently, they detected IL-1웁 mRNA transcripts in blood PMNs of patients with RA, at much greater levels than those detected in paired SF PMNs (369). In contrast, in two patients with seronegative arthritis and two patients with ankylosing spondylitis, the hybridization signals for IL-1웁 mRNA in the blood neutrophils were undetectable.

NEUTROPHIL-DERIVED PROTEINS

431

The results of their second study (369) implied that activation of IL-1웁 expression by PMNs in RA occurs in the circulation before the cells enter into diseased joints. As a consequence, only a small subpopulation of SF PMNs expressing IL-1웁 can be found at any moment, namely those that have been most recently attracted into the inﬂamed joint. More recently (244), the same group demonstrated that cell-free SF induced the synthesis of IL-1웁 in neutrophils isolated from fresh buffy coats. Both soluble and insoluble immune complexes, which are found in RA ﬂuid, induced the synthesis of IL-1웁, but very little of this cytokine was secreted (244). That neutrophils isolated from the SF of patients with RA constitutively synthesize IL-1웁 has been observed by Jobin and Gauthier (253). IL-1웁 expression was considered to occur as the result of a neutrophil–neutrophil interaction caused by the high cellular density observed in the SF of patients with RA (ranging from 2 ⫻ 106/ml to 100 ⫻ 106/ml). Contributing to the description of neutrophil-derived cytokines in RA, Edwards’s group showed that both RA SF and blood PMNs express high levels of TGF-웁1 but no TNF-움 mRNA (237), whereas Taichman et al. (36) identiﬁed VEGF in neutrophils from RA synovial biopsies, by immunohistochemical staining. Furthermore, Beaulieu and McColl (370) found RA SF PMNs to be signiﬁcantly less efﬁcient in producing IL-1RA, compared with matched blood PMNs. The spontaneous or GM-CSF- or TNF-움-induced production of IL-1RA by SF neutrophils was signiﬁcantly decreased when compared with blood neutrophils isolated from the same individuals. Under the same experimental conditions, production of IL1웁 and IL-8 was up-regulated, suggesting that there was not a general down-regulation of cell function in SF PMNs. These results were also paralleled by a comparable modulation at the level of cytokine mRNA expression (370). Beaulieu and McColl therefore concluded that neutrophils are likely to be an important source of IL-8 and IL-1웁 in the RA joint, but that SF neutrophils appear incapable of mounting a response as high as that of blood neutrophils in terms of IL-1RA production (370). The potential role of neutrophil-derived chemokines to attract PMNs to the SF was further investigated by Koch et al. (22), who found signiﬁcantly greater levels of antigenic GRO-움 in SF from patients with RA as compared with osteoarthritis (OA) or other noninﬂammatory arthritides. This GRO움 accounted for 28% of the whole chemotactic activity for PMNs found in RA SF (22), suggesting that also GRO-움 plays an important role in the migration of PMNs into the inﬂamed RA joints. Both RA SF PMNs and normal blood PMNs produced signiﬁcant amounts of GRO-움, either constitutively or after stimulation with LPS (22). Therefore, production of IL8 (370), GRO-움 (22), and likely more chemokines by neutrophils and other

432

MARCO ANTONIO CASSATELLA

cells may lead to the recruitment of more leukocytes to the joint, and therefore perpetuate RA. PMNs from patients with systemic lupus erythematosus (SLE) are known to exhibit several functional abnormalities that cause an increased susceptibility to infections (371), which is one of the hallmarks of SLE (372). Hsieh et al. (373) found that the spontaneous and LPS-stimulated production of IL-8 by the peripheral blood PMNs of active SLE patients was impaired as compared to inactive SLE or healthy individuals. This impaired IL-8 production by SLE PMNs was not linked to the administration of steroids, because incubation of normal PMNs or inactive SLE PMNs with prednisolone for 24 hr did not signiﬁcantly affect IL-8 production (373). However, the possibility that a long-term immunosuppressive treatment may have led to defective IL-8 production in active SLE patients was not excluded. The same group also reported that the spontaneous and LPS-stimulated production of IL-1RA by PMNs, but not by PBMCs, in patients with active SLE was signiﬁcantly lower than that in patients with inactive SLE or in normal groups (374). This impaired IL-1RA production by SLE PMNs seemed, again, not linked to the administration of corticosteroids, because prednisolone did not affect the IL-1RA production of normal neutrophils in vitro, either spontaneously or after LPS stimulation (374). Moreover, IL-1RA production by active SLE PMNs increased concomitantly with clinical and laboratory measures of improvement after effective clinical therapy, but not in the patients who did not respond to the treatment (374). Taken together, the results suggested that decreased IL-8 and IL1RA production may constitute speciﬁc functional defects of PMNs in patients with active SLE, and may predispose to infections and may also be regarded as novel indicators of disease activity in patients with active SLE. However, a more recent study in which SLE patients were not taking any antirheumatic or immunosuppressive medications (375) partially contradicts the latter ﬁndings (374), because only in monocytes of SLE patients was IL-1RA production somewhat lower than in normal individuals. Furthermore, no signiﬁcant difference between SLE patients and controls was found in the production of IL-1웁 by neutrophils or monocytes after LPS stimulation (375). The ratio of IL-1RA/IL-1 production from neutrophils and monocytes was also not different for lupus patients and normal donors (375). D. NEUTROPHIL-DERIVED CYTOKINES IN PATIENTS WITH SEPSIS IL-1웁 is one of the genes that appear to play an essential role in the pathogenesis of sepsis syndrome (376). McCall’s group reported that circulating PMNs of patients with sepsis syndrome (sepsis-PMNs) were, in vitro, tolerant to endotoxin-induced expression of the IL-1웁 gene (377).

NEUTROPHIL-DERIVED PROTEINS

433

This tolerance consisted of a combined reduction in LPS-stimulated levels of IL-1웁 mRNA and a decreased synthesis of the immunoreactive IL-1웁 protein. It was excluded that the mechanisms responsible for the tolerance of sepsis-PMNs were due to the loss of the CD14 surface protein, the receptor required for endotoxin-mediated gene induction in PMNs (180), or that they were the result of a global reduction in the functional responses of PMNs (377). The down-regulation of the IL-1웁 gene in sepsis-PMNs occurred concomitantly with an up-regulation of the constitutive expression of the type II IL-1 receptor (IL-1RII) (378), and did not persist in PMNs of patients recovering from the sepsis syndrome. Tolerance involved speciﬁc signal transduction pathways triggered by LPS, because sepsis-PMNs normally synthesized IL-1웁 in response to S. aureus, and secreted elastase (377). Interestingly, tolerance was not limited to infection by gram-negative bacteria, but was also observed when sepsis was induced by gram-positive bacteria, Rickettsia, Candida species, or staphylococcal exotoxins, and not in patients seriously ill without detectable infection. In agreement with these ﬁndings, the release of IL-8 by PMNs isolated from patients with sepsis was also signiﬁcantly reduced whether activated by LPS or by heat-killed streptococci (108). Similar results were obtained by studying neutrophils of patients who had undergone cardiac surgery with cardiopulmonary bypass (CPB), a condition that provokes a sort of inﬂammatory stress (108). In vitro experiments demonstrated that the observed diminished neutrophil ability to produce IL-8 did not reﬂect an endotoxin tolerance effect stricto sensu, because LPS was unable to desensitize the cells to a second challenge by LPS, as reported for monocytes (379). However, pretreatment of PMNs with IL-10 rendered the cells less reactive to a subsequent stimulation with LPS (108), suggesting that IL-10 may have been responsible for the observed down-regulated capacity to produce IL-8 in those patients. In addition to circulating IL-8, high levels of cell-associated IL-8 were also detected in blood samples from patients with sepsis (380). Other than erythrocytes, already known as a‘‘sink’’ for IL-8 (381) PMNs and PBMCs were found to contribute substantially to the detection of cell-associated IL-8 (380). These ﬁndings were not surprising, because we (124) and others (63) had previously shown that cell-associated IL-8 is indeed measurable in neutrophils either in vitro (124) or in vivo (63). Strikingly, in sepsis patients the amount of IL-8 associated with PMNs on a per cell basis was twice that found with PBMCs, and 3000 times that found in erythrocytes (380). Taking into account the relative number of circulating cells, 78.5–92% of the detectable cell-associated IL-8 in blood was thus linked to PMNs. The analysis of cell-associated cytokines was extended to another component of the inﬂammatory response, IL-1RA, which also can exist as an

434

MARCO ANTONIO CASSATELLA

intracellular form (254). Cell-associated IL-1RA was detected in septic patients (380). However, in contrast to IL-8, the levels of cell-associated IL-1RA were not always higher than those of circulating IL-1RA. What is more, IL-8 and IL-1RA mRNA were detected in PBMCs and PMNs from some septic patients and from some healthy controls, thus illustrating that circulating cells can contribute to the production of both IL-8 and IL1RA (380). The authors concluded that measurement of cell-associated proinﬂammatory and antiinﬂammatory cytokines in sepsis patients is a more reliable reﬂection of their production than is the simple measurement in plasma. E. NEUTROPHIL-DERIVED CYTOKINES IN PATIENTS WITH DENTAL INFECTIONS Adult periodontitis is a chronic infectious disease associated with active tissue damage. PMNs usually constitute the great majority (⬎95%) of cells in gingival crevicular ﬂuid (GCF) obtained from gingival inﬂammation. Furthermore, in GCF, it is possible to reveal, with considerable frequency, signiﬁcant levels of both IL-1 and TNF-움 (382). Whether PMNs derived from adult periodontitis could express cytokine genes in vivo was investigated by Takeichi and colleagues (52). Signiﬁcant levels of biologically active IL-1움 and IL-1웁, but not TNF-움 or IL-6, were found in GCF of adult periodontitis patients (52). In addition, by using RT–PCR associated with a slot-blot analysis, they provided elegant evidence that highly puriﬁed PMNs (⬎99.5%) collected from GCF express IL-1움, IL-1웁, and TNF-움 messenger RNA, but not IL-6 transcripts (52). Takahashi et al. (383) examined by ISH the IL-1웁 mRNA-expressing cells in GCF harvested at 15 diseased sites from ﬁve patients with adult periodontitis and 8 clinically periodontal healthy sites from three volunteers. ISH showed IL-1웁 transcripts in both PMNs and MNCs, but not in epithelial cells, in all GCF samples from diseased and healthy sites. The latter was not surprising, because a small inﬂammatory reaction to bacterial plaque is constantly present even in clinically healthy sites. PMNs were the predominant leukocytes in diseased and healthy sites, and the percentages or IL-1웁 mRNApositive PMNs in GCF samples from diseased and healthy sites were 92.3 and 80.9%, respectively. Besides, the mean amounts of IL-1웁 mRNA expression in PMNs were higher compared to MNCs in all samples and there was heterogeneity within the populations of PMNs and MNCs in their ability to express the IL-1웁 gene (383). These ﬁndings indicated that IL-1웁 is predominantly produced by PMNs in the gingival crevice of patients with adult periodontitis and in healthy controls. In line with these observations, Hendley et al. (231) showed that freshly obtained oral PMNs accumulated IL-1웁 mRNA and released the soluble cytokine after 3 hr of

NEUTROPHIL-DERIVED PROTEINS

435

culture, but failed to respond to further stimulation with GM-CSF. In contrast, peripheral PMNs cultured in the absence of stimulus did produce this cytokine, but responded well to GM-CSF (231). Furthermore, the amount of IL-1웁 produced by oral PMNs was strikingly greater than that produced by circulating PMNs in vitro (231). The same group conﬁrmed the observations on IL-1웁 in a larger number of cases (40 patients with adult periodontitis and 40 orally healthy matched controls) and extended the study to include the production of TNF-움 (164). They found that oral PMNs released considerable amounts of IL1웁 and TNF-움, but there was no difference between patients and controls when the cytokine levels were corrected for cell number. However, when the effect of disease activity was examined, cytokine release by oral PMNs was found to be greater in patients with advanced disease (164). No accumulation of mRNA for either IL-1웁 or TNF-움 was demonstrable in oral PMNs that had been cultured for 3 hr (164), whereas freshly puriﬁed neutrophils expressed both messages (164, 231). Interestingly, within the healthy control group, IL-1웁 production by oral PMNs was signiﬁcantly higher in males, but no effect of race on cytokine production could be discerned in patients or controls (164). Examination of IL-1웁 production by peripheral blood PMNs exposed to GM-CSF revealed no difference between the patient and control groups. In contrast, IL-1웁 production by peripheral blood PMNs was signiﬁcantly reduced in patients with advanced periodontitis (164). Finally, evidence that PMNs in inﬂammatory periradicular tissues may be a signiﬁcant source of IL-1움 and IL-1웁 has been further generated by Miller et al. (384). Cell suspensions from selected periapical granuloma specimens, as well as from puriﬁed peripheral blood PMNs and PBMCs, were subjected to IL-1 quantitation using an ELISA procedure. Such cell suspensions were found to produce signiﬁcant levels of IL-1 and could be stimulated to produce increased levels after coculture with LPS. They have also found a positive staining for IL-6 in a subpopulation (15–20%) of neutrophils resident in inﬂamed periradicular lesions (274). Because a marked variation in cytokine production among individuals was discovered (164, 383), Galbraith and colleagues started to investigate a possible association between TNF-움 genotype and the level of TNF-움 production by oral PMN’s obtaining promising preliminary results (385). Whether a particular TNF genotype is a risk factor for severity of disease in adult periodontitis is an interesting working hypothesis. The levels of expression of some chemokines have been evaluated by Tonetti et al. (386), who performed in situ hybridization of IL-8 and MCP1 genes in frozen tissue sections from patients affected by adult periodontal infections. Maximal IL-8 expression was found in the junctional epithelium

436

MARCO ANTONIO CASSATELLA

adjacent to the infecting microorganisms where PMN inﬁltration was more prominent, whereas MCP-1 was expressed in the chronic inﬂammatory inﬁltrate and along the basal layer of the oral epithelium where only cells of the monocyte/macrophage lineage were present (386). These topographically speciﬁc tissue locations of chemokine mRNA expression were consistent with a hypothetical establishment of a discrete chemotactic source for an effective leukocyte recruitment. The pattern of IL-8 production in gingiva observed by Tonetti et al. (386) was quite different from what was observed in patients affected by rapidly progressive periodontitis (RPP) (387). The latter disease is a rapid and destructive form of periodontitis, which typically emerges in the late second or third decade of life, and that is characterized by a massive inﬁltration of PMNs in inﬂammatory gingival sites (388). High IL-8 plasma levels as well as strong IL-8 mRNA expression in both epithelial and connective gingival cells from patients with RPP was found (387). Moreover, the gingival PMNs contained IL-8 mRNA, suggesting an autoampliﬁcation of PMN recruitment and activation in the gingiva. In addition, resting PMNs of RPP patients showed reduced Lselectin, Lewis x, and sialyl Lewis x antigen expression as well as increased H2O2 production (387). PMNs showed a lack of increased response (H2O2 production) to formyl peptides after ex vivo priming with IL-8, possibly caused by IL-8 desensitization. Medical treatment of gingival inﬂammation was accompanied by normalization of IL-8 values, PMN oxidative burst, and PMN priming by IL-8, suggesting a link among these features of the disease in which PMN-derived IL-8 is probably self-amplifying (387). These modiﬁcations of PMN adhesion molecule expression, together with their increased basal oxidative burst and excessive IL-8 production, may contribute to the noxious inﬂammatory reaction, which may in turn be autopotentiated by PMN production of IL-8. Chronic apical periodontitis is an infectious disease, characterized by granuloma formation and progressive bone resorption around the apex of the tooth. IL-1움, IL-1웁, and TNF-움 are known to induce osteoblastic bone resorption, inhibit bone formation in vitro, and decrease bone collagen synthesis. Based on previous evidence that PMNs derived from chronic inﬂamed tissues can produce proinﬂammatory cytokines (52), Takeichi et al (178) characterized the ability of alveolar bone-derived PMNs to produce proinﬂammatory cytokines. High concentrations of IL-1움, IL-1웁, and IL6, but not of TNF-움, were detected in periapical exudates (PEs), collected from periapical lesions with chronic periapical periodontitis through root canals (178). Because the PE contains predominantly PMNs with a small percentage of lymphocytes and/or macrophages, the latter were puriﬁed and analyzed for cytokine mRNA expression using a cytokine-speciﬁc RT– PCR. Highly puriﬁed PMNs isolated from PEs expressed signiﬁcant levels

NEUTROPHIL-DERIVED PROTEINS

437

of mRNA for IL-1움, IL-1웁, and TNF-움. IL-6 mRNA was not detected, despite a high concentration of IL-6 detected in supernatants of PEs (178), likely deriving from macrophages, T lymphocytes, osteoblasts, or ﬁbroblasts around periapical lesions. These data strongly suggested that human PMNs derived from alveolar bone spontaneously produce IL-1움, IL-1웁, and TNF움 at sites of inﬂammation, and probably initiate inﬂammation and regulate augmentation of bone resorption in vivo. F. NEUTROPHIL-DERIVED CYTOKINES IN PATIENTS WITH OTHER TYPES OF INFECTIONS To determine the cytokine-producing cells during the early and late stages of shigellosis, Raqib et al. (389) examined, by immunohistochemistry, cryopreserved tissues from Shigella-infected patients. Shigella infection is usually accompanied by an intestinal activation of epithelial cells, T cells, and macrophages within the inﬂamed colonic mucosa. Histopathologically, Shigella infection is characterized by the presence of chronic inﬂammatory cells with or without neutrophils and microulcers in the lamina propria, crypt distortion, and, less frequently, crypt abscess. Raqib and co-workers found that Shigella-infected patients had signiﬁcantly higher numbers of cytokine-producing cells for all the cytokines studied (IL-1움/웁, IL-1RA, IL-4, IL-6, IL-8, IL-10, IFN-웂, TNF-움, lymphotoxin-움, TGF-웁1, -웁2, and -웁3), compared to healthy controls. However, production of the various cytokines in rectal biopsies during acute and convalescent periods was not signiﬁcantly different, with the exceptions of TGF-웁 and IL-1RA (389). In the acute Shigella infection, the PMNs present in the crypt abscess were clearly seen to contain IL-1웁 (389). This observation is revelant in view of the evidence that IL-1 is a key player in the cascade mediating invasion and inﬂammation by Shigella of the intestinal mucosa (390). Meningitis is an acute inﬂammatory disease of the pia and arachnoid and the ﬂuid in the subarachnoid space. Meningitis is accompanied by a differential immigration of leukocytes into the subarachnoid space, but the mechanisms regulating leukocyte invasion during meningitis are still incompletely understood. Obviously, chemokines may represent the major chemoattractant stimuli for the differential recruitment of leukocytes into the subarachnoid space during meningitis. To better understand the role of chemokines, Sprenger and colleagues (391) analyzed 48 paired cerebrospinal ﬂuid (CSF) and serum samples from patients hospitalized for meningitic symptoms. Very high levels of IL-8, GRO-움, and MCP-1 were detected in the CSF during bacterial and nonbacterial meningitis, whereas MIP-1움 and RANTES were below detection limits. In contrast, the levels of chemokines in the blood serum samples were not elevated (391). The levels of IL-8 and GRO-움 in the CSF signiﬁcantly correlated with the

438

MARCO ANTONIO CASSATELLA

immigrated granulocytes in the CSF of patients with bacterial meningitis. A similar correlation between MCP-1 levels and the mononuclear cell count in nonbacterial meningitis was found (391). In another work, protein concentration and mRNA expression of TNF움 and TGF-웁1 in the CSF of 23 patients with bacterial or viral meningitis were investigated (392). High amounts of cytokines (especially of TNF움) at both protein and mRNA levels were detected in bacterial as well viral infections (392). A preponderance of TNF-움 compared to TGF-웁1 mRNA was visible in CSF cells of patients with bacterial meningitis, whereas a balance of TNF-움 and TGF-웁1 mRNA or a higher expression of TGF-웁1 mRNA was detected in viral meningitis. Surprisingly, in the acute-phase bacterial meningitis, but even in viral infections, neutrophils expressed more TNF-움 and TGF-웁1 mRNA than did lymphocytes or monocytes/macrophages, which instead dominated cytokine synthesis during the healing phase. TNF-움 and TGF-웁1 mRNA were expressed by neutrophils at the time of signiﬁcant clinical worsening (392). Finally, PMNs from injured patients with elevated Candida antigen titers were evaluated for their ability to activate the anticandidal functions and to produce TNF-움 and IL-8 (393). Though PMNs demonstrated impaired function against C. albicans growth when compared with PMNs from injury-matched controls, cytokine production and autocrine activation remained intact (393), demonstrating that the PMN dysfunction in these patients is not global. G. NEUTROPHIL-DERIVED CYTOKINES IN PATIENTS WITH VIRAL INFECTIONS Investigations by ultrastructural localization of HGF in tissue samples obtained from control subjects and patients affected by acute and chronic hepatitis and cirrhosis revealed that, in addition to biliary epithelial cells, PMNs represented a source of circulating HGF (40). However, the number of PMNs stained for HGF in the liver from patients with acute hepatitis was not greater than that from control subjects, probably because the specimens were obtained from recovering patients (40). Electron microscopy conﬁrmed that the immunostained cells with segmented nuclei were polymorphonuclear leukocytes and that the stained grains were on the membranes of rough endoplasmic reticulum, around speciﬁc or azurophilic granules. It was therefore proposed that PMNs could produce HGF during liver regeneration, and this cytokine, in turn, may affect hepatocyte and biliary epithelial cell proliferation (324). Bortolami and colleagues assessed the effect of different doses of IFN움 on the production of TNF-움 by resting and LPS-activated human neutrophils from normal and hepatitis C virus (HCV)-infected patients (394).

NEUTROPHIL-DERIVED PROTEINS

439

The results revealed that non of the IFN-움 concentrations (25–5000 U/ ml) alone induced TNF-움 from PMNs, and that TNF-움 production by PMNs after LPS stimulation was similar in normal and HCV-infected patients. However, IFN-움 associated with LPS induced a marked increase in TNF-움 secretion by PMNs from HCV-infected patients, but had minimal effects on PMNs from healthy controls (394). The data uncovered a new biological property of IFN-움, which may be taken into account during therapy for HCV-related chronic hepatitis. Cassone’s group reported that neutrophils from HIV-infected (HIV⫹ ) patients were able to synthesize amounts of IL-1웁 and IL-6 comparable to those made by neutrophils from healthy subjects, suggesting that cells from HIV⫹ patients were good responders to activating signals (245). More recently, their analysis of the ability of neutrophils from AIDS patients to produce cytokines was numerically extended (13, 102). IL-1웁, IL-6, IL8, and TNF-움 production by PMNs from 21 HIV⫹ patients, including 11 with full-blown AIDS, and 20 HIV-uninfected (HIV⫺) subjects (matched for age and sex to HIV⫹ subjects) was studied by RT–PCR and ELISA. In all subjects, cytokine gene expression was strongly stimulated by MPF2 or LPS, and inhibited by IL-10 (13, 102). In these studies, PMNs from HIV⫹ subjects showed increased IL-6 and TNF-움 gene expression and produced more IL-6 and TNF-움 than did PMNs from HIV⫺ controls (13, 102). Quantitatively similar expression of IL-1웁 transcripts, as well as production of IL-1웁 and IL-8 proteins, was observed in cells from HIV⫹ and HIV⫺ subjects (13, 102. We also investigated the ability of neutrophils isolated from HIV⫹ patients to produce proinﬂammatory cytokines (290). We determined the in vitro responsiveness of PMNs and PBMCs to LPS, used in the presence or absence of IFN-웂, in 47 HIV⫹ patients with advanced stages of virus infection. Release of TNF-움 and IL-8 from HIV⫹ PMNs was found to be higher than that from normal PMNs. Conversely, release of IL-1웁 and IL-1RA in response to LPS, or to LPS plus IFN-웂, was found to be lower in PMNs from HIV⫹ patients than that from controls. The release of IL12 induced by LPS, or by LPS plus IFN-웂, did not signiﬁcantly differ for PMNs from HIV⫹ patients and healthy donors. For all the cytokines, the capacity of IFN-웂 to modulate their production was not modiﬁed in HIV⫹ patients. However, the production of TNF-움 and IL-12 in response to LPS, or to LPS plus IFN-웂, was found to be signiﬁcantly higher in PBMCs isolated from HIV⫹ patients, whereas the release of IL-1웁 was signiﬁcantly lower. In the case of IL-8, no statistically signiﬁcant difference was found for PBMCs isolated from HIV⫹ patients and healthy donors (290). Collectively, Cassone’s group data and ours suggest that in HIV⫹ patients with advanced stages of disease, the ability of PMNs (and PBMCs) to produce

440

MARCO ANTONIO CASSATELLA

speciﬁc cytokines in response to LPS is signiﬁcantly altered, and this might condition the evolution of HIV disease. H. NEUTROPHIL-DERIVED CYTOKINES IN PATIENTS WITH NEOPLASTIC DISEASES Clinical and experimental studies from Jablonska et al. (273, 395, 396) have focused on the measurement of cytokines by neutrophils and PBMCs from breast cancer patients. The authors found signiﬁcant differences concerning the parameters examined for PMNs and PBMCs from cancer patients as compared with normal individuals (395). Neutrophils of cancer patients stimulated with zymosan and LPS were found to produce signiﬁcantly lower amounts of TNF-움 as compared to PMNs of healthy controls (396). Conversely, neutrophils isolated from patients with breast cancer displayed an‘‘increased’’ capacity to constitutively produce IL-6 as compared to controls, but did not respond to a further stimulation with LPS or zymosan (273). However, in this latter study, extremely high levels of IL-6 were released by resting neutrophils from normal or patient donors (273). In line with these results, another group has reported that neutrophils isolated from patients with early oral carcinoma produce IL-6 and IL-1웁 in response to OK-432, but, again, at levels so elevated that they were comparable to those found in autologous PBMCs (246). Finally, Ericson and colleagues (275) have reported that circulating neutrophils from patients with metastatic breast cancers undergoing G-CSF therapy express detectable levels of IL-6 mRNA by RT–PCR, as well as detectable intracellular IL-6 protein (15 pg/ml) (275). Cross-linking Fc웂RI or Fc웂RII on these G-CSF-treated PMNs resulted in a synergistic increase in the amount of secreted IL-6 (275). V. Modulation of Cytokine Production in Human Neutrophils

An important aspect that has emerged from many studies is that cytokine expression in neutrophils often changes as a consequence of the‘‘functional’’ state of these cells at the moment of receiving the stimulus. Various types of inﬂammatory or immunomodulating agents can positively or negatively affect the functional responses of neutrophils, including the production of cytokines (397). The mechanisms whereby cytokine production is potentiated or down-regulated are (in most cases) still a mystery, the elucidation of which is rendered even more difﬁcult by the fact that these immunomodulating agents, usually, but not always, do not directly trigger any response in neutrophils. Though IFN-웂 and IL-10 represent the best studied examples, other cytokines (or experimental conditions) have been described to modulate neutrophilic cytokine production, as outlined below.

441

NEUTROPHIL-DERIVED PROTEINS

A. EFFECTS OF IFN웂 As evidenced in Table VI, IFN-웂 mainly potentiates the production of cytokines by neutrophils. This up-regulation occurs regardless of the agonist used, and is especially evident after long incubation times (i.e., 16 hr or more). In particular, the LPS-elicited secretion of TNF-움 and IL-1웁, as well as the TNF-움-stimulated release of IL-1웁, are markedly enhanced by IFN-웂 (124). Under some circumstances however, IFN-웂 exerts an inhibitory effect toward neutrophil-derived cytokine production. For instance, preincubation of PMNs with IFN-웂 signiﬁcantly inhibits the release of IL-8 after 2 hr of stimulation with fMLP or Y-IgG (77). Similarly, IFN웂 inhibits the early production of IL-8 induced by LPS or TNF-움, but augments the release of IL-8 at later time points (124, 179). An identical biphasic action of IFN-웂 has been observed in the case of MIP-1움, MIP-1웁, and GRO-움 release in LPS-stimulated neutrophils (21, 179). Importantly, Northern blot analyses have revealed that these various effects of IFN-웂 are paralleled by corresponding changes at the level of cytokine (TNF-움 and IL-1웁) or chemokine (IL-8, MIP-1움, MIP-1웁, and GRO-움) mRNA expression. Other studies conﬁrm that treatment of human neutrophils with IFN웂 and LPS for 20 hr markedly increases the production of TNF-움 over that of neutrophils treated with LPS alone (50, 170, 398), and have further extended this priming effect on the release of TNF-움 to IFN-웂 used in combination with GM-CSF (50). Moreover, the addition of sodium nitroprusside (SNP), an exogenous source of nitric oxide, together with TABLE VI EFFECTS OF IFN-웂 ON THE RELEASE OF NEUTROPHIL-DERIVED CYTOKINES in Vitro Stimulus Useda Cytokine TNF-움 IL-1움/웁 IL-1RA IL-8 IL-12 GRO-움 MIP-1움/웁 IP-10, MIG

LPS 앖앖 앖앖 앖 앖 앖앖앖 앖 앖 앖앖앖

TNF-움

fMLP

Y-IgG

MP-F2 앖

앖

앖 ⫽

앖 앖 앖 ⫽

앖

⫽

⫽

앖

앖앖앖

IL-1웁

앖

앖앖앖

a Abbreviations: LPS, lipopolysaccharide; fMLP, formyl-methionyl-leucyl-phenylalanine; Y-IgG; S. cerevisiae opsonized with IgG; MP-F2; mannoprotein fraction of C. albicans. Effects: The number of arrows reﬂects the extent of the up or down regulatory effect of IFN-웂 in neutrophils stimulated for 20 hr; ⫽, no effect.

442

MARCO ANTONIO CASSATELLA

N-acetylcysteine (NAC), which increases the bioavailability of nitric oxide, was shown to further potentiate the IFN-웂 plus LPS-induced TNF-움 and IL-8 secretion over a 20-hr period of incubation (72, 398). It should be noted that exposure of LPS-treated PMNs to SNP or SNP plus NAC, but not to NAC alone, also increased the production of TNF-움 (398) and IL8 (72), as compared with exposure to LPS alone. Interestingly, though the up-modulation of IL-8 production was associated with an increased expression of LPS-induced IL-8 mRNA (72), this was not the case with TNF-움 (398). The potential effect of nitric oxide on LPS-induced TNF움 and IL-8 production by human neutrophils may represent a mechanism by which endothelial and vascular smooth muscle cells can augment the activation state of neutrophils as the latter migrate from the blood to sites of infection. It may also represent a novel way by which nitric oxide regulates chemotactic responses. Conversely, these responses could also contribute to the cytokine-mediated tissue injury and organ failure observed during systemic infections. The modulatory actions of IFN-웂 affect also the expression of IL-1RA. We have demonstrated that in LPS-stimulated neutrophils, IFN-웂 increases the production of IL-1RA by enhancing the release of de novosynthesized IL-1RA, without modifying IL-1RA mRNA accumulation (248). We also observed that in response to fMLP (10 nM ), neutrophils released small but signiﬁcant amounts of IL-1RA: following IFN-웂 treatment, this fMLP-elicited IL-1RA production was greatly potentiated and was accompanied by a transient up-regulation of IL-1RA mRNA accumulation (248). In contrast, IL-1웁 protein was undetectable in culture supernatants from neutrophils stimulated with fMLP (248), even though an increased accumulation of IL-1웁 mRNA was noticed under these conditions (5). Furthermore, prior treatment of granulocytes with IFN-웂 enhanced the IgG-opsonized yeast particle-induced IL-1RA formation, but only at late times of incubation. Under all conditions tested, IFN-웂 by itself failed to exert any detectable effect toward gene expression and release of IL1RA or IL-1웁 (248). It therefore appears that, depending on the type of stimulus used, IFN-웂 enhances IL-1RA production by acting through different mechanisms. The ability of IFN-웂 to modulate cytokine production by PMNs may have important implications in vivo. For instance, the ability of IFN-웂 to up-regulate cytokine and chemokine production by neutrophils exposed to provocative doses of LPS might be one of the key factors contributing to the pathogenesis of endotoxic shock (399). It might also represent one of the key mechanisms contributing to the improvement of host defense against microbes in CGD patients undergoing IFN-웂 therapy (307, 400).

443

NEUTROPHIL-DERIVED PROTEINS

B. EFFECTS OF IL-10 The effects of IL-10 on the production of cytokines by human neutrophils (289) and monocytes (401) have been reviewed, and herein I will brieﬂy deal with this information (see also Table VII). IL-10 was initially reported to inhibit the extracellular production of TNF-움, IL-1웁, and IL-8 triggered by LPS and Y-IgG (79). Subsequently, it was demonstrated that IL-10 suppresses also the production of IL-1움 by LPS-stimulated neutrophils, and that all its inhibitory effects are signiﬁcantly greater than the inhibitions observed with either IL-4 or TGF-웁 (51). Interestingly, Northern blot analyses revealed that IL-10 diminished the LPS-induced mRNA accumulation of proinﬂammatory cytokines (TNF-움, IL-1, and IL-8) only at time points later than 4-5 hr, but not at onset (1–2 hr) (51, 79). All of these observations have been repeatedly conﬁrmed by several groups (88, 108, 309, 346, 348, 402, 403) and further extended by the demonstration that IL-10-mediated inhibition of neutrophil-derived cytokines also occurs in neutrophils isolated from HIV-infected individuals and stimulated with MP-F2 from C. albicans (13, 102). Evidence that MIP-1움 and MIP-1웁 release by neutrophils stimulated with LPS is inhibited by IL-10 has been provided by Kunkel’s group (24). Consistent with our previous studies on IL-8 (79), the extracellular production of MIP-1움 and MIP-1웁, as well as the LPS-induced accumulation of mRNA encoding these chemokines, was signiﬁcantly suppressed by IL-10 only after 4–8 hr of culture, and not before (24). Subsequently, TABLE VII EFFECTS OF IL-10 ON THE RELEASE OF NEUTROPHIL-DERIVED CYTOKINES in Vitro Stimulus useda Cytokine

LPS

TNF-움 IL-1움/웁 IL-1RA IL-8 IL-12 GRO-움 MIP-1움/웁 IP-10, MIG

앗앗 앗앗 앖 앗앗 앗앗b 앗 앗 앗앗b

TNF-움

fMLP

Y-IgG

MP-F2

앖 ⫽

⫽ ⫽

앗 앗 ⫽ 앗

앗앗 앗앗

앖

⫽

앗

앗앗

IL-1웁

⫽

앗b

a Abbreviations: LPS, lipopolysaccharide; fMLP, formyl-methionyl-leucyl-phenlalanine; Y-IgG, S. cerevisiae opsonized with IgG; MP-F2, mannoprotein fraction of C. albicans. Effects: The number of arrows reﬂects the extent of the up or down regulatory effect of IL-10 in neutrophils stimulated for 20 hr; ⫽, no effect. b Stimulation in the presence of IFN-웂.

444

MARCO ANTONIO CASSATELLA

we showed that IL-10 also inﬂuences GRO-움 production induced by LPS with a pattern identical to those of IL-8 and MIP-1움/웁 (21). In addition, we demonstrated that IL-10 slightly inhibits the release of GRO-움 triggered by Y-IgG but potentiates the release of GRO-움 induced by TNF-움 (21), evidence that not all of the effects of IL-10 toward granulocytes are inhibitory. The studies of how IL-10 and IFN-웂 affect cytokine and chemokine production have revealed the existence of a PMN-centered cytokine network regulating the production of IL-8, MIP-1움, MIP-1웁, and probably also GRO-움, in LPS-stimulated PMNs (4). This consists of two distinct phases (4). The early phase accounts for a low level of chemokine release, directly induced by LPS. This initial wave of chemokine release is followed by a second delayed phase, in which endogenous TNF-움 and IL-1웁 synergize with LPS in inducing dramatically elevated levels of IL-8, MIP-1움, MIP-1웁, and probably also of GRO-움. This sequential production of chemokines by LPS-activated PMNs, regulated by TNF-움 and IL-1웁, might be ﬁnalized to amplify the recruitment and activation of neutrophils and other leukocytes during an inﬂammatory response to LPS. Supporting the potential existence of this network, the in vitro proﬁle of TNF-움, IL-1웁, and IL-8 secretion in response to C. neoformans was found analogous to that observed with LPS (107). In contrast, an autocrine loop involving only IL-1웁 was shown to maintain the secretion of IL-8 by PMNs isolated from the sputum of subjects with chronic bronchial sepsis (353). In the latter conditions, neutrophil-derived IL-8 was inhibited by IL-10 at levels signiﬁcantly lower than in peripheral blood neutrophils stimulated with LPS (353). It must, however, be speciﬁed that sputum neutrophils were presumably in an‘‘activated’’ state, because they constitutively secreted large amounts of IL-8, IL-1웁, and TNF-움 that were only slightly increased in the presence of LPS (353). IL-10 has been proved to be a potent inhibitor of the extracellular release of IP-10 and MIG induced by IFN-웂 plus either LPS or TNF-움 (146, 147), and of the extracellular production of the p40 chain of heterodimeric IL-12 in human PMNs treated with LPS (12). As a consequence of the latter suppressive action, no biologically active IL-12 could be measured in supernatants of neutrophils stimulated by LPS plus IFN-웂 in the presence of IL-10 (12). Once again, these inhibitory effects mediated by IL10 were paralleled by alterations at the level of cytokine mRNA accumulation. In view of the fundamental role of IL-12, IP-10, and MIG to initiate and perpetuate a Th1 response (134–137, 283), IL-10-mediated suppression of neutrophil-derived IL-12, IP-10, and MIG could represent a mechanism, acting at the level of innate immunity, whereby IL-10 prevents Th1 cell generation and attraction during infections (404).

NEUTROPHIL-DERIVED PROTEINS

445

In contrast to the inhibitory effect on the production of proinﬂammatory cytokines and chemokines, IL-10 markedly potentiates the mRNA and extracellular yields of IL-1RA (by two- to threefold after 18 hr) in LPSstimulated neutrophils, but not in Y-IgG-treated cells (258). Additionally, IL-10 has been shown to up-regulate the constitutive neutrophil secretion of IL-1RA (346). This being said, Jenkins et al. (405) reported that in neutrophils treated with LPS over a period of 22–24 hr, IL-10 signiﬁcantly reduced the overall IL-1웁 production, but did not inﬂuence that of IL1RA. However, they did not discriminate between the net effects of IL10 on cell-associated versus secreted IL-1웁 or IL-1RA. In addition, they showed that although IL-10 moderately decreased the LPS-induced IL1웁 mRNA at times later than 4 hr, in agreement with our previous results (79), IL-10 did not affect LPS-induced IL-1RA mRNA levels, in sharp contrast to our ﬁndings (258). We do not have an explanation for this discrepancy, although a possible hypothesis might be partially related to the different concentrations of LPS used in the two studies. Nevertheless, the net result of the effect of IL-10 on neutrophils was to increase the ratio of IL-1RA to IL-1웁 (405), consistent with our previous data (79, 258). In another study, in which the yields of IL-1RA detected in cellfree supernatants of agonist-stimulated neutrophils were much higher than the values found by other investigators (14, 258, 405), only IL-4 was reported to amplify the extracellular production of IL-1RA elicited by LPS (used at 100 ng/ml) (263). Conversely, IL-10 as well as IL-4 displayed an extremely high potency in up-regulating the production of IL-1RA induced by TNF-움 (263), suggesting that the effects of IL-10 on IL-1RA production triggered by LPS or TNF-움 act through different pathways. By regulating PMN-derived cytokine production, IL-10 may have an important regulatory role in limiting the duration and extent of acute inﬂammatory responses, for instance in lethal endotoxemia. In this regard, Marie et al. (108) have recently shown that, unlike monocytes (406), neutrophils exposed to LPS for 24 hr before a secondary stimulation with LPS were not ‘‘tolerant,’’ but instead displayed an enhanced IL-8 production. But if IL-10 and LPS were added simultaneously to PMN cultures during the pretreatment period, the ‘‘priming’’ effects of LPS were completely obliterated (108). These results suggest that the IL-10 produced during sepsis (407) may render neutrophils unresponsive to a further stimulation with LPS. On the other hand, the fact that the release of cytokines can be modulated not only by IFN-웂, but also by IL-10, IL-4, and IL-13 (see below), argues for the possibility that both Th1 and Th2 lymphocytes might play a role in inﬂuencing the production of cytokines by PMNs. Studies addressing the in vivo effects of IL-10 on neutrophil-derived cytokines are urgently required.

446

MARCO ANTONIO CASSATELLA

C. EFFECTS OF IL-4 AND IL-13 IL-4 and IL-13 represent additional molecules able to modulate moderately the release of some cytokines by neutrophils. For instance, treatment of PMNs with IL-4 results in a substantial reduction in the IL-8 steadystate mRNA levels and production stimulated by LPS (51, 88, 105). In contrast, IL-4 is a poor inhibitor of PMN-derived IL-8 in response to YIgG (5) and, similarly to IL-13, does not inhibit the release of IL-8 triggered by high doses of TNF-움 (20 ng/nl), whether or not TNF-움 is used in combination with LPS (88). In its inhibitory action on LPS-induced IL-8 release, IL-4 is more potent than equivalent concenterations of TGF-웁 or IL-13, but less potent than IL-10 (51, 88). Interestingly, though the inhibitory effects of IL-4 (as well as of IL-13) greatly vary depending on the donor, the effects of IL-10 do not (88). Despite this donor variability, IL-4 potentiates the negative effect of IL-10 on IL-8 and other proinﬂammatory cytokine release by neutrophils (our unpublished observations). In fact, other than IL-8, IL-4 has also been shown to markedly decrease (by up to 50%) the synthesis of IL-1웁 induced by LPS over a 22-hr period (235). This occurred without affecting the LPS-stimulated IL-1웁 mRNA levels (235, and our unpublished experiments), suggesting that IL-4 acts at posttranscriptional levels. Furthermore, IL-4 induces a substantial production of IL-1RA from neutrophils, which is more than additively increased by coincubation with LPS (235, 263). As mentioned, IL-4 or IL-10, but not IL-13 or TGF-웁, also ampliﬁes the production of IL-1RA induced by TNF-움 (263). Although there are no reports on the effects of IL-4 or IL13 on the production of TNF-움 by PMNs, our preliminary data suggest that TNF-움 mRNA expression and release induced by LPS are inhibited by IL-4, but less efﬁciently than by IL-10 (5). It remains to be investigated whether IL-4, as IL-10, down-regulates the release of IL-8 induced by LPS because of its inhibition on endogenous TNF-움 and IL-1웁 production (4). D. OTHER MODULATORY EFFECTS Under some circumstances GM-CSF may act as a priming agent for neutrophil-derived cytokines. For instance, GM-CSF may prime neutrophils for an enhanced TNF-움 and IL-8 production induced by LPS (50). Similarly, GM-CSF, but not G-CSF, was shown to potentiate IL-8 mRNA expression and release after internalization of L. monocytogenes and Y. enterocolitica (106). Other studies have demonstrated the GM-CSF primes the release of IL-8 induced by fMLP (113), which also modulates cytokine and chemokine production in EBV-stimulated neutrophils (120, 267). In the latter case, pretreatment of neutrophils with GM-CSF prior to EBV addition synergistically enhanced the total production of IL-1움, IL-1웁,

NEUTROPHIL-DERIVED PROTEINS

447

IL-8, and MIP-1움, whereas it only marginally affected IL-1RA synthesis. These actions occurred without modiﬁcations of the IL-1움, IL-1웁, IL-8, and MIP-1움 mRNA steady-state levels (120, 267), supporting the notion that GM-CSF acts at a translational level. Though the majority of IL-1움, IL-1웁, and IL-8 newly synthesized in response to either GM-CSF or EBV remained inside the cell, the combination of the two agonists resulted in signiﬁcantly more of the IL-8 being secreted (267), whereas IL-1움 and IL-1 remained cell associated (120). This was in contrast to what observed with MIP-1움, the majority of which was secreted irrespective of the treatment (120). Several other experimental conditions have been described as further modulating the production of cytokines by activated neutrophils. For example, a synergistic release of IL-8 by neutrophils has been shown to occur in response to IL-15 plus inactivated C. albicans (86), to LPS plus TNF움 (50, 108) or fMLP (68), and to some bacteria or mycobacterial derivatives plus TNF-움 (114, 120). On a related note, the release of IL-6 induced by cross-linking of Fc웂RI or Fc웂RII was potentiated by G-CSF (275). The effects of inﬂammatory microcrystals on neutrophil-derived cytokines have already been extensively summarized (5). VI. Molecular Regulation of Cytokine Production in Neutrophils

An increasing body of knowledge is now available on the molecular mechanisms that regulate cytokine gene expression in neutrophils. Studies addressing cytokine release at the molecular level have revealed that cytokine production is usually preceded by an increased accumulation of the related mRNA transcripts. As in monocytic cells (which have been more extensively studied in this regard), the control of cytokine gene expression in neutrophils can take place at the transcriptional, posttranscriptional, translational, and posttranslational levels. This evidence derives from either direct experimental demonstrations or indirect indications obtained through the use of selective metabolic drugs (5). The latter include actinomycin D (ACT D), which is a blocker of RNA synthesis that can indicate that the induction of a given gene is regulated at the transcriptional level, or cycloheximide, puromycin, and emetine, which are inhibitors of protein synthesis. Results obtained with ACT D have indirectly suggested that the induction of IL-1웁, IL-1RA, IL-8, and MIP-1움 transcripts in neutrophils stimulated with various agonists involves transcriptional events. As will be discussed below, this is already known to be the case for IL-1웁, IL-8, and MIP-1움. CHX has been shown to inhibit, superinduce, or have no effect on the induction of a given cytokine mRNA, depending on the particular mRNA under investigation and the stimulatory conditions. In any case,

448

MARCO ANTONIO CASSATELLA

the use of CHX has made it clear that the inﬂuence of de novo protein synthesis on the steady-state level of a given cytokine mRNA can vary considerably depending on the triggering stimulus. A. TRANSCRIPTIONAL REGULATION The transcriptional activity of a given gene can be directly determined by the nuclear run-on assay technique. However, very few investigators have performed such assays in PMNs probably due to the extreme difﬁculties encountered in attempting to detect the very low transcriptional activity of these cells (51, 408). This being said, the use of this approach has nevertheless revealed that neutrophils actively transcribe the IL-1웁 gene in response to IL-1, TNF-움, IL-1 plus TNF-움 (409), LPS (410), and GMCSF (232). Similarly, the IL-8 and MIP-1움 genes have been shown to be transcriptionally induced in response to LPS in neutrophils (410, 411). Very little is known concerning the regulation of gene transcription in neutrophils, but it is noteworthy that many cytokine genes depend on the activation of transcription factors, such as NF-␬B (412). For this reason, we have started to investigate the expression and distribution of NF-␬B/ Rel proteins and of their cytoplasmic inhibitor, I␬B-움, in human peripheral blood neutrophils, as well as their respective fate on cell activation (413). Among a wide range of neutrophil agonists, TNF-움, LPS, fMLP, and PMA were found to induce efﬁciently both the nuclear accumulation of NF␬B/Rel proteins and the concomitant degradation of cytoplasmic I␬B움, whereas IL-1웁, LTB4, and PAF proved to be weaker stimuli at the concentrations tested (413). Moreover, the onset of both processes was paralleled by the activation of nuclear NF-␬B DNA-binding activity. In contrast, GM-CSF, G-CSF, IFN-움, IFN-웂, IL-8, and IL-10 exerted no detectable effect on any of these responses (up to 120 min) (413). The stimuli that promote the nuclear accumulation of Rel family proteins and the concomitant activation of NF-␬B can be subdivided into two general categories: those whose action is rapid, and the slower acting ones. The effect of the former (TNF-움 and LPS, and on a much smaller scale, IL1웁) was evident by 10 min and reached a maximum by 30 min. In contrast, the latter agonists (fMLP and PMA, and to a lesser extent, PAF and LTB4) required at least 30 min to exert a similar action, a maximal effect usually being observed between 45 and 60 min. We ﬁnally demonstrated that in neutrophils, the degradation of I␬B-움 protein and the concomitant nuclear activation of NF-␬B were accompanied by a marked accumulation of I␬B움 mRNA transcripts, resulting in the reexpression of the protein, through de novo synthesis (413). More recently, we have also reported that NF-␬B is rapidly and transiently induced in neutrophils undergoing phagocytosis of Y-IgG particles (414) or stimulated by IL-15, but not by IL-2 (90). Other

NEUTROPHIL-DERIVED PROTEINS

449

groups have conﬁrmed our results (76, 239, 415), except in the case of GM-CSF, which, in contrast to our ﬁndings (413), was found to induce NF-␬B activity (239). Therefore, further studies are required to clarify whether GM-CSF activates NF-␬B in PMNs. B. REGULATION AT THE LEVEL OF MRNA STABILITY As opposed to transcriptional regulation, there are many demonstrations that cytokine gene expression of neutrophils can be regulated at the level of mRNA stability. For instance, this is one of the mechanisms explaining the up-regulation of IL-1웁 mRNA expression in neutrophils treated with IL-1, TNF-움, IL-1 plus TNF-움 (409), or GM-CSF (232). mRNA stabilization analyses have also demonstrated that MIP-1움 mRNA isolated from PMNs stimulated in the presence of GM-CSF plus LPS had a prolonged half-life, relative to LPS alone (23). Another cytokine mRNA, that encoding IL-1RA, was found to be mainly regulated at the posttranscriptional level in neutrophils. The augmented expression of IL-1RA mRNA in PMNs treated with IL-13 (49) and TGF-웁1 (262) was shown to depend on a marked increase in IL-1RA transcript stability. The possible contribution of transcriptional induction of the IL-1RA gene could not, however, be excluded, in that ACT D partially blocked the enhancing effect that IL13 (49) and TGF-웁1 (262) had on IL-1RA mRNA steady-state levels. Finally, the half-life of IL-1RA mRNA was prolonged in PMNs stimulated in the presence of IL-10 and LPS, as compared with cells stimulated with LPS alone (258). The mechanisms underlying some of the modulatory effects of IL-10 and IFN-웂 on cytokine mRNA accumulation in LPS-treated neutrophils have been extensively analyzed. The stimulatory effect of LPS on IL-1웁 gene transcription was found to be uninﬂuenced by either IFN-웂 or IL10 (410). Under identical experimental conditions, both IL-10 (258) and IFN-웂 (our unpublished observations) also failed to have a signiﬁcant effect on the stability of IL-1웁 mRNA isolated from LPS-treated PMNs. In contrast, the inhibitory effect of IL-10 on LPS-induced IL-8 mRNA accumulation was shown to correlate with an enhancement of IL-8 mRNA degradation (24, 118), and our experiments indicated that IL-10 and IFN웂 also inhibit the rate of LPS-stimulated IL-8 gene transcription in PMNs (410). It can therefore be envisaged that IL-10 inhibits LPS-induced IL8 mRNA accumulation both through inhibition of IL-8 gene transcription and through enhanced degradation of IL-8 mRNA. MIP-1움 represents another gene whose early down-modulation by IFN-웂 in LPS-treated PMNs (24, 179) is regulated at the level of transcription (411). As also observed for IL-8 (79, 179) and MIP-1웁 (179), IFN-웂 inhibits mRNA expression and production of MIP-1움 from LPS-stimulated PMNs at early

450

MARCO ANTONIO CASSATELLA

time points, but augments MIP-1움 expression later on (179). IFN-웂 did not affect MIP-1움 mRNA stability in PMNs treated with LPS for 2–4 hr (179, 411). However, after stimulation of neutrophils with LPS for 24 hr, the stability of MIP-1움 mRNA was shown to be signiﬁcantly prolonged by IFN-웂 (179). In contrast to IFN-웂, IL-10 inhibited LPS-induced MIP1움 mRNA accumulation through enhanced mRNA degradation (24). We found that IL-10 does not inhibit the rate of LPS-stimulated MIP-1움 gene transcription in PMNs (our unpublished observations), therefore conﬁrming that IL-10 primarily acts at the level of MIP-1움 mRNA stability. Although no studies on MIP-1웁 gene transcription have been reported to date, MIP-1웁 mRNA expression was inﬂuenced by IL-10 and IFN-웂 in the same manner as MIP-1움, that is, at the posttranscriptional level (24, 179). Stability of MIP-1웁 mRNA in LPS-treated neutrophils was signiﬁcantly prolonged by IFN-웂 at 24 hr (179), but not at the 4-hr time point, and was markedly diminished by IL-10 at 4 and 8 hr (24). C. TRANSLATIONAL REGULATION Another series of observations indicate that mechanisms other than gene transcription or mRNA stabilization are involved in the regulation of cytokine production in neutrophils. A clear demonstration of a form of translational control in PMNs has been shown for IL-1 (11). It was observed that although LPS induced the expression of IL-1움 and IL-1웁 mRNA in PMNs, these cells were much less efﬁcient in translating these transcripts when compared with PBMCs (11). In contrast, both PMN- and PBMCderived IL-1 mRNAs were translated with equal efﬁciency in an in vitro protein synthesis system (11). IL-4 markedly decreased the total IL-1웁 protein synthesis in LPS-induced PMNs, without reducing the LPSstimulated IL-1웁 mRNA levels, suggesting that the effects of IL-4 were mediated at the translational level (235). Control at the translational level has also been suggested for IL-1RA and MIP-1움 mRNA, under speciﬁc experimental conditions. Pretreatment of neutrophils with TNF-움 before stimulation with calcium pyrophosphate dihydrate microcrystals up-regulates the levels of IL-1RA mRNA as compared with induction by TNF-움 alone, whereas cell-associated and secreted protein levels of IL-1RA were inhibited in the presence of CPPD (251). Under the same conditions, CPPD crystals synergistically increased both IL-1웁 mRNA and protein levels in TNF-움-treated neutrophils. These data suggest that inhibition of IL-1RA synthesis by CCPD in TNF-움-treated neutrophils occurs at the translational level (251). In the case of MIP-1움, both CPPD and monosodium urate monohydrate inhibited the immunodetectable MIP-1움 induced by TNF-움, without affecting MIP-1움 steadystate mRNA levels (25). Because the crystals neither enhanced the degra-

NEUTROPHIL-DERIVED PROTEINS

451

dation of MIP-1움 protein nor interfered with the immunodetection of MIP-1움 by ELISA, the data suggest that the inhibitory effect of microcrystals on MIP-1움 protein production is primarily translational (25). Both zymosan and S. cerevisiae have been shown to inhibit the production of MIP-1움 induced by TNF-움 (114). As hypothesized for CPPD and MSU, a possible inhibitory mechanism at of phagocytic agonists was postulated to occur at the level of the MIP-1움 gene translation. This was because zymosan and S. cerevisiae, while inducing the accumulation of MIP-1움 mRNA, neither inhibited MIP-1움 secretion nor caused MIP-1움 extracellular degradation (114). Similarly, when neutrophils were stimulated by M. tuberculosis or PPD in the presence of TNF-움, expression of MIP-1움 mRNA was signiﬁcantly higher than that stimulated by M. tuberculosis or PPD alone, but production of antigenic MIP-1움 was diminished, not increased (80). In this case, a retained enhanced neutrophil apoptosis rate, induced by mycobacterium derivatives plus TNF-움, was responsible for the reduced production of MIP-1움 (80). Further evidence that cytokine or chemokine mRNA transcripts may be under translational control has been derived from studies addressing the effects of GM-CSF in EBV-treated neutrophils. Pretreatment of neutrophils with GM-CSF prior to EBV stimulation, synergistically enhanced the production of IL-1움 and IL-1웁, but only slightly affected the synthesis of IL-1RA (267). The fact that cytokine synthesis was increased by GMCSF in the absence of changes in mRNA levels suggests a more effective translation of IL-1 and IL-1RA mRNA (267). Additional studies from the same group showed that GM-CSF increased the production of EBVinduced neutrophil IL-8 and MIP-1움 by approximately two- to threefold (120). Pretreatment with GM-CSF failed to increase levels of the steadystate mRNA encoding IL-1 and IL-1RA, supporting the idea that GM-CSF predominantly primes EBV-treated neutrophils at a posttranscriptional or a translational level (120). Finally, it was reported that the induction by GM-CSF of TNF-움 mRNA accumulation in PMNs is not accompanied by TNF-움 synthesis or release (162). Although the reasons for this lack of TNF-움 synthesis were not further investigated, it nevertheless may be speculated that TNF-움 production might be subjected to translational regulation. If so, it is also conceivable that by increasing the steady-state level of TNF-움 mRNA, GM-CSF may act as a ‘‘priming’’ agent for a subsequent triggering stimulus, as in fact has been demonstrated to occur on LPS stimulation (50). In another study, nitric oxide-generating compounds increased neutrophilic LPSinduced TNF-움 production, yet without increasing TNF-움 mRNA levels (398). It must be stressed, however, that Northern blots were performed at a single time point (1 hr). If Northern blots performed at later time

452

MARCO ANTONIO CASSATELLA

points could conﬁrm the lack of effect of nitric oxide-generating compounds on TNF-움 mRNA steady-state levels, then the enhancement of TNF움 production by these compounds would likely reﬂect a modulation of translation or secretion. D. POSTTRANSLATIONAL REGULATION Several reports describe regulatory mechanisms in neutrophils at the level of cytokine secretion. A more efﬁeint secretion of IL-1RA as compared to that obtained with the single agonists was observed in PMNs stimulated with IL-4 plus LPS (235), or with IL-10 plus TNF-움 (263). Similarly, the enhancement of IL-1RA release by IFN-웂 (248) or GM-CSF (267) in neutrophils stimulated with LPS or EBV, respectively, appeared essentially to take place at the level of secretion. IL-8 is another cytokine for which there is numerous experimental evidence that production may be regulated at the level of secretion. In one of our studies we separately quantiﬁed the IL-8 immunoreactivity that remained cell associated, and that was released by the cells, in order to better examine whether the effects of IFN-웂 on agonist-induced IL-8 production might reﬂect changes in IL-8 synthesis or secretion (124). This study brought forward a number of interesting observations (124). First, the total production of IL-8 in LPS- and Y-IgG-treated PMNs (as well as in resting cells) continuously increased, up to 18 hr. Second, the amounts of IL-8 released by resting PMNs, as well as by LPS-, TNF-움-, and YIgG-activated cells, increased throughout the incubation period, but the amounts of secreted IL-8 were inferior to those that were cell-associated, except for Y-IgG. Third, although the accumulation of IL-8 mRNA paralleled the increase in total IL-8 protein levels in both LPS- and Y-IgGactivated PMNs, the latter secreted IL-8 more efﬁciently than did LPSstimulated cells. Fourth, in PMNs pretreated with IFN-웂 and then stimulated for up to 6 hr, the total synthesis of IL-8 was signiﬁcantly lower than when the cells were exposed to the stimuli in the absence of IFN-웂. However, total IL-8 production after 18 hr of incubation with various stimuli was not signiﬁcantly affected by the presence or absence of IFN웂, except for Y-IgG-treated cells. All these effects were paralleled by changes at the mRNA level. Moreover, the percentage of IL-8 secreted after stimulation with LPS and TNF-움 for up to 18 hr, or with Y-IgG for 2 hr, was signiﬁcantly higher in PMNs that were pretreated with IFN-웂 in comparison with untreated cells. Thus, even though IFN-웂-treated PMNs synthesized less IL-8 than untreated PMNs, they secreted IL-8 more efﬁciently after stimulation with LPS or TNF-움, at all time points examined. Altogether, these studies indicated that the up-regulatory effects of IFN-웂 on LPS- and TNF-움-induced secretion of IL-8, observed after 18 hr, were largely explained by the potentiating effect of IFN-웂 at the

NEUTROPHIL-DERIVED PROTEINS

453

level of IL-8 secretion. That IL-8 accumulation is under translational control was also revealed by the fact that CHX, but not ACT D, inhibited the accumulation of cell-associated IL-8 after culture at 37⬚C for 2 hr (63). In vitro assays of LPS-stimulated neutrophils from healthy or septic patients demonstrate equivalent levels of cell-associated IL-8 (108). Because PMNs from septic patients release smaller amounts of IL-8 compared to those of control subjects, it is evident that in septic neutrophils the proportion of IL-8 secreted is lower (108). Finally, the percentage of IL8 secreted by PMNs in response to LPS was shown to increase in correlation with the doses of LPS used (380), and the enhancement of IL-8 production by GM-CSF, in EBV-treated neutrophils, appeared to take place at the level of translation and secretion (120). Experiments performed in my laboratory revealed that the extracellular production of GRO-움 by PMNs does not always correlate with equivalent changes at the level GRO-움 mRNA expression. That did occur in YIgG-stimulated PMNs, and the extracellular production of GRO-움 was approximately two- to threefold more effective than with LPS or TNF-움 stimulation (21). However, the increase of GRO-움 transcripts was markedly less pronounced (21). These ﬁndings indicated that Y-IgG phagocytosis regulates GRO-움 production at the level of secretion, and this is in keeping our observations on Y-IgG-stimulated secretion of IL-8, as summarized above (124). In contrast, culture of PMNs with fMLP strongly promoted GRO-움 mRNA accumulation but contained amounts of GRO-움 that were not signiﬁcantly higher than those from untreated cells (21). A possible explanation is that GRO-움 is released but rapidly degraded by the proteolytic enzymes that are simultaneously secreted in response to fMLP. Conversely, it cannot be excluded that in spite of its ability to strongly induce GRO-움 mRNA, fMLP fails to provide the intracellular signals necessary to either translate or secrete GRO-움. If so, this would suggest that GRO움 production is controlled at the translational or posttranslational level, as previously observed for GRO-움 during malignant transformation of normal melanocytes (416). What is more, IL-10 did not affect the LPS-elicited accumulation of GRO-움 mRNA (21), suggesting that its inhibitory effect on LPS-induced GRO-움 production likely occurs at the translational or posttranslational level. A last consideration is that in most of the studies quantifying the proportion of cell-associated cytokines, the percentages of the cytokines linked to their receptors were not measured. This should be considered, however, because it may profoundly inﬂuence the results and the conclusions. VII. Cytokine Production by Neutrophils in Vivo

Studies evaluating the possibility that PMNs are a signiﬁcant source of cytokines in vivo are ongoing. In speciﬁc experimental animal models the

454

MARCO ANTONIO CASSATELLA

production of cytokines by neutrophils appears to be fundamental for the evolution and/or resolution of the induced pathological process. Table VIII summarizes the cytokines that have been described to be produced by neutrophils in vivo. A. EFFECTS OF LPS ADMINISTRATION ON NEUTROPHIL-DERIVED CYTOKINES in Vivo A classic way to study neutrophil cytokine expression in vivo is an LPSinduced acute inﬂammation in animals. An example is the intratracheal injection (IT) of endotoxin, which in rats causes a dramatic inﬂux of PMNs into the bronchoalvolar space. When kinetics of cytokine mRNA expression in the lung were investigated by Northern analysis, it was found that IL-1움 mRNA peaked at 2 to 6 hr, whereas IL-1웁/IL-1RA mRNA peaked at 6 hr, concurrent with the maximum inﬂux of neutrophils (15). It was hypothesized that the production of IL-1 by PMNs may play a role in the activation of lymphocytes during the transition between acute neutrophilic and chronic mononuclear inﬂammation, whereas the synthesis of IL-1RA might function as a negative feedback mechanism that down-regulates neutrophil inﬂux into inﬂammatory sites. Fractionation of alveolar macrophage (AM) -enriched and PMN-enriched subpopulations from bronchoalveolar lavage cells revealed that neutrophils were the predominant source for both IL-1움/웁 and IL-1RA mRNA (15). Using a similar model, Xing et al. (417) provided in vivo evidence that PMNs can represent a signiﬁcant source of TNF-움 at sites of acute inﬂammation in rats. By Northern blot analysis, they found that PMNs displayed several times more TNF-움 mRNA than did AMs at 6 and 12 hr after IT instillation of LPS. By in situ hybridization, most of the cells positive for TNF-움 mRNA were PMNs localized within the inﬂamed tissue near bronchioles or vessels. By IH, TNF-움 protein was localized mainly to AMs at early times after LPS challenge (1–3 hr), whereas thereafter (6–12 hr) PMNs were the predominant source of TNF-움 protein (417). The same group subsequently demonstrated that in lung PMNs and AMs, LPS triggers a distinct cytokine response, by selectively increasing mRNA transcripts encoding TNF-움, IL-1웁, IL-6, and MIP-2, but not RANTES or TGF-웁1 (418). At a time (1 hr) when only a minimal PMN inﬁltration was present, AMs appeared to be the predominant source of all cytokines examined, whereas at later times (6 and 12 hr), when PMN inﬁltration became maximal, PMNs were the prominent source of those cytokines. A low, basal, noninducible signal for TGF-웁1 (but not for RANTES) mRNA was detected in both AMs and PMNs (418). Interestingly, ISH of the lung tissue revealed that among the cells that stained for MIP-2 mRNA in response to LPS were, particularly, the PMNs located in the vicinity of bronchioles and vasculature, but

455

NEUTROPHIL-DERIVED PROTEINS

TABLE VIII CYTOKINES EXPRESSED BY NEUTROPHILS in Vivo Cytokine Produced/Experimental Model TNF-␣ Carrageenan pretreatment in mice injected with LPS Mice injected intraperitoneally with LPS Mice undergoing hyperoxia Listeria monocytogenes-infected mice Herpetic stromal keratitis in mice Mice injected with a colon adenocarcinoma releasing G-CSF Cutaneous injury in mice Rats instilled intratracheally with LPS Apical periodontitis in rats Brain injury in rats Ethanol intoxication in rats Shwartzman reaction in rabbit lung LPS infusion of rabbits Intravitreal injection with LPS Rabbits injected intraarticularly with MSU crystals IL-1␣ Mice injected intraperitoneally with LPS Listeria monocytogenes-infected mice Mice injected with a colon adenocarcinoma releasing G-CSF Cutaneous injury in mice Rats injected intratracheally with LPS Apical periodontitis in rats IL-1␤ Mice with endotoxemia and hemorrage Mice undergoing hyperoxia Mice injected with a colon adenocarcinoma releasing G-CSF Acute pancreatitis in mice Cutaneous injury in mice Spinal cord injury in mice Rats injected intratracheally with LPS Rats injected intravenously with LPS Intravitreal injection with LPS Transient retinal ischemia in rats Apical periodontitis in rats Brain injury in rats Shwartzman reaction in rabbit lung Rabbits injected intraarticularly with LPS Rabbits injected intraarticularly with IL-8 Rabbits injected intraperitoneally with casein Rabbits injected intraarticulary with MSU crystals IL-1RA Mice orogastrically infected with Yersenia enterocolitica Rats injected intratracheally with LPS Shwartzman reaction in rabbit lung Rabbits injected intraarticulary with IL-8

Ref. 429 430 454 459 460 151, 475 480 417, 418 476, 477 482 427 422 426 439 443 430 459 151 480 15 476, 477 424 453, 454 151 479 480 481 15, 418 423 439 449 476, 477 482 422 432, 435 434 442 443 457 15 422 434 (continues)

456

MARCO ANTONIO CASSATELLA

TABLE VIII CYTOKINES EXPRESSED BY NEUTROPHILS in Vivo (Continued) Cytokine Produced/Experimental Model Rabbits injected intraarticularly with LPS Intravitreal injection with LPS Rabbits injected intraarticulary with MSU crystals IL-6 Mice injected intraperitoneally with LPS Rats instilled intratracheally with LPS Brain injury in rats Reperfusion of ischemic myocardium in dogs IL-10 Candida albicans-infected mice Mice injected intraperitoneally with LPS IL-12 Candida Albicans-infected mice TGF-␤ Rats instilled intratracheally with LPS Brain injury in rats Developing endochondral bone in rats IL-8 Shwartzman reaction in rabbit lung Rabbits injected intraarticularly with LPS Rabbits injected intraperitoneally with casein Rabbits injected intraarticularly with MSU crystals Reperfusion of ischemic myocardium in rabbits Reperfusion of ischemic brain in rabbits Rabbits undergoing hyperoxia Dog trachea superfused with Pseudomonas supernatants MIP-2 Antiglomerular basement membrane nephritis in rats Instilled intratracheally with LPS Rats instilled intratracheally with LPS, or injected intraperitoneally with thioglycollate CINC Antiglomerular basement membrane nephritis in rats Rats injected intraperitoneally with LPS KC Rats instilled intratracheally with LPS, or injected intraperitoneally with thioglycollate MCP-1 Intratracheal instillation with bleomycin in rats Rabbits injected intraarticularly with LPS Rabbits injected intraarticularly with MSU crystals MIP-1␤ Rabbits injected intraperitoneally with casein M-CSF Brain injury in rats

Ref. 435 439 443 418 482 447 291, 292 430 291–293, 296 418 482 18 422 435 441 443, 445 445 448 452 112 26 418 419 26 420 419 450 439 439 441 482

NEUTROPHIL-DERIVED PROTEINS

457

not within the vasculature (418). Under similar experimental conditions (IT instillation of LPS), recruited PMNs could rapidly (30 min) and persistently (up to 16 hr) be induced to express MIP-2 and KC/GRO-움/MGSA (419). Although Northern analyses were performed on pooled BAL cell mRNA, these BAL cell populations changed from predominantly macrophages (⬎95%) to mostly PMNs (60% within 2 hr, 91% after 16 hr) after LPS instillation (419). Furthermore, expression of MIP-2 and KC mRNA was also observed within exudative neutrophils obtained after intraperitoneal injection of thioglycollate (419). Collectively, these observations support the notion that in these lung models of LPS-elicited inﬂammation, inﬁltrating PMNs represent a signiﬁcant source of proinﬂammatory cytokines. Lukaszewicz and colleagues (420) studied CINC mRNA induction in lungs of normal, neutropenic, and adrenalectomized rats after intraperitoneal injection of E. coli LPS. After a single dose of LPS, rapid induction of CINC mRNA coincided with neutrophil inﬁltration into lungs, a response that lasted approximately 12 to 24 hr. However, it was not formally proved to be a causal relationship between CINC expression and neutrophil lung inﬁltration (420). Interestingly, CINC mRNA induction in lungs was heightened 30% in adrenalectomized animals, consistent with the role of glucocorticosteroids as potent inhibitors of chemokine production (421), and 400% in neutropenic ones (420). The latter data suggest that neutrophils may act to inhibit expression of CINC, limiting in this way their own inﬂux into tissue via a negative feedback mechanism. Production of TNF-움, IL-1웁, IL-8, and IL-1RA was also analyzed in a model of local Shwartzman reaction (LSR) studied in rabbit lung (422). In this model, myeloperoxidase activity (representing neutrophil accumulation) peaked at 1–2 hr and was sustained for 48 hr after intravenous (IV) challenge with LPS. Kinetics of cytokine production revealed that TNF-움 was the ﬁrst to appear, peaking at 0.5 hr, whereas IL-1웁 and IL-8 increased later and peaked at 2 hr (422). IL-1RA was present even before the challenge, and its increased production showed a dual peak: at 0.5–2 and at 48 hr. The authors speculated that this endogenous IL-1RA might serve to suppress part of IL-1 activity and function through a negative feedback mechanism to prevent an excessive inﬂammatory response. IH showed that the cellular sources of these cytokines were inﬁltrating neutrophils and AMs (422), and it was inferred that production of cytokines was not a direct mediator for the initiation of LSR, but likely the consequence of events following leukocyte inﬁltration. Circulating PMNs can also constitute a prominent source of cytokines in LPS-mediated inﬂammations. Williams et al. (423) showed that following IV infusion of LPS for 2 hr in rats, PMNs rather than PBMCs in the

458

MARCO ANTONIO CASSATELLA

pulmonary vasculature were the major source of IL-1웁 transcripts. In contrast, no induction of IL-1웁 expression was observed in airway or circulating leukocytes. Under these experimental conditions, LPS increased pulmonary vascular sequestration of leukocytes, recruiting most prominently an activated pool of neutrophils that were more adherent, that were primed for increased reactive oxygen intermediate (ROI) production, and that expressed increased IL-1웁 messages. Thus, this early study suggested a more prominent role than previously appreciated for sequestered neutrophils in sepsis-induced lung inﬂammation. Another study has shown that neutrophils that trafﬁc to the lungs are a major source of IL-1웁 in the lungs after hemorrhage and endotoxemia (424). In particular, IL-1웁 was detected by IH in enriched pulmonary neutrophil populations isolated 1 hr after hemorrhage or endotoxemia, but not in lymphocytes (424). In neutropenic mice IL-1웁 expression was lower than in control animals, providing a possible explanation of previous observations that neutropenia decreases alveolar leak and acute lung injury in endotoxin-treated animals (425). In addition, neutrophils seen in the pulmonary vasculature were noted to express IL-1웁 (424), suggesting that endotoxin or hemorrhage induces IL-1웁 production by neutrophils before these cells enter the lung parenchyma. These data indicate that IL-1웁-producing neutrophils trafﬁc to the lungs rapidly in response to hemorrhage or endotoxemia and support the concept that proinﬂammatory cytokine production by lung neutrophils may contribute to the development of lung injury after blood loss and sepsis. In a different study, Cirelli et al. (426) conﬁrmed that an accumulation of intravascular mononuclear phagocytes and neutrophils in the pulmonary circulation takes place during a continuous infusion of endotoxin in sheep. These authors detected an increased cytoplasmic TNF-움 immunoreactivity in both mononuclear phagocytes and neutrophils sequestered in pulmonary arterioles, capillaries, and venules (426). Coincidentally, plasma levels of TNF-움 signiﬁcantly increased, suggesting that both neutrophils and mononuclear phagocytes contributed to the rise in the circulating levels of TNF-움, and the development of acute lung injury (426). Other studies demonstrated that ethanol intoxication inhibits iNOS but not TNF-움 mRNA expression in AMs or neutrophils recruited to the rat lung in response to LPS (427), and that the release of constitutive TNF-움 by inﬂammatory neutrophils isolated from pleural exudate of rats treated with prolactin is markedly enhanced (428). Furthermore, carrageenan (CAR) pretreatment was shown to prime mice for an enhanced LPSinduced TNF-움 production in sera and increased their mortality rate (429). It was then observed that CAR treatment enhances the LPS-induced TNF-움 activity in the supernatants of neutrophils, but not of mononuclear cells (429). Because both serum TNF-움 and mortality risk were signiﬁcantly

NEUTROPHIL-DERIVED PROTEINS

459

lower in neutropenic mice, the data suggest that neutrophils are the major TNF-움-producing cells, thus being responsible for mortality (429). Nill et al. (430) compared the temporal sequence of endotoxin-induced TNF-움, IL-1움, and IL-10 gene expression and cellular localization of cytokine proteins in pulmonary tissue of two strains of mice that have a genetically based differential sensitivity to endotoxin. Cytokine mRNA were studied by RT–PCR and ISH in lung tissue obtained from the endotoxin-sensitive C3H/HeN and endotoxin-resistant C3H/HeJ mice at different times after IP injection of LPS (430). Although levels of TNF-움 mRNA and protein in the two mouse strains were similar at 1–2 hr, IL-1움 gene and protein expression in pulmonary tissue isolated from endotoxinresistant mice was lower at any time point examined (430). IL-10 mRNA and protein levels were up-regulated and continued to increase over a 12-hr time period in C3H/HeN mice, whereas they were basically undetectable in CH3/HeJ endotoxin-resistant mice. In both types of mouse strains, TNF-움, IL-1움, and IL-10 immunoreactive proteins were localized primarily to the inﬁltrating neutrophils, as well as to AMs and type II pneumocytes (430). However, quantitation of neutrophil inﬁltration into pulmonary tissue demonstrated that there was a signiﬁcant decrease in the inﬂammatory inﬁltrate in pulmonary tissue isolated from CH3/HeJ mice following LPS administration, which correlated with decreased levels of immunoreactive cytokine proteins within pulmonary cells (430). These results unequivocally implicate that inﬁltrating neutrophils are important cellular mediators of pulmonary tissue damage induced by endotoxin in the CH3H/HeN endotoxin-sensitive mice. Terebuth et al. (431) performed IH to localize cells expressing IL-6 in selected organs of normal and endotoxin-challenged mice. In normal mice, a constitutive cytoplasmic IL-6 immunoreactivity was detected in blood monocytes and their precursors, in bone marrow and splenic stromal macrophages, and in granulocytes as well (431). Though signiﬁcant serum levels of IL-6 were absent, cell-associated IL-6 bioactivity was found in circulating PMNs but not in lymphocytes. However, after IP injection of LPS, there was a two- to threefold increase in PMN cell-associated IL-6 bioactivity from 1 to 3 hr, followed by an almost complete depletion at 6 hr. Concomitantly, serum levels of IL-6 peaked at 3 hr after LPS challenge and dropped signiﬁcantly by 6 hr. Interestingly, constitutive, increased intracellular IL-6 in circulating PMNs was detected in the absence of IL-6 mRNA, which was instead present in granulocytic/monocytic progenitors in the bone marrow. In the latter cells, IL-6 transcripts increased with a similar time course after LPS challenge (431). These data suggest a scenario in which circulating granulocytes bear IL-6 as a stored component, likely acquired during bone marrow maturation. During endotoxemia,

460

MARCO ANTONIO CASSATELLA

granulocytes release IL-6 as a result of appropriate signals received, for example, during margination or chemotaxis. To better clarify the pathogenesis of arthritis, Yoshinaga’s group used a model of rabbit arthritis induced by intraarticular injection of LPS. Initially, they observed that the IL-1웁 produced by neutrophils and macrophages in the synovial exudate was responsible for both leukocyte inﬁltration and loss of proteoglycans (432). Destruction of cartilage was the consequence of leukocyte-derived cartilage-degrading substances, such as elastase and superoxide anion (432). In leukopenic rabbits, injection of LPS induced neither production of IL-1웁 nor loss of PG. They also investigated the generation kinetics of IL-1RA, and its signiﬁcance in the pathogenesis of LPS arthritis. Production of IL-1RA (by leukocytes) was delayed compared to IL-1웁, was sustained for 1 week, and was in 180- to 200-fold molar excess of IL-1웁. LPS-induced leukocyte inﬁltration was inhibited by 70–75% by rabbit IL-1RA (432). Furthermore, the administration of anti-IL-1RA mAb with LPS into rabbit knee joints increased the IL-1 activity fourfold and the production of antigenic IL-1웁 by 30–50%. This treatment also enhanced the LPS-induced leukocyte inﬁltration and protein leakage by 20–40% (433), suggesting that endogenous IL-1RA may suppress a part of IL-1 activity in situ, but that its amount was too low for suppression of all the biologic effects of IL-1웁. Subsequently, the same group injected homologous IL-8 in rabbit knee joints and investigated the inﬂammatory response (434). In these experiments, IL-8 induced a massive accumulation of neutrophils (but no appreciable numbers of lymphocytes) and provoked the release of neutrophil elastase, which led to cartilage destruction. In addition, injection of IL-8 induced bioactive and immunoreactive IL-1웁 and IL-1RA, but not TNF-움 in the joint cavity. As determined by IH, IL-1웁- and IL-1RA-positive cells were, again, inﬁltrating leukocytes (434). Production kinetics of immunoreactive IL-1RA in SF overlapped that of IL-1웁, but the peak concentration of IL-1RA exceeded that of IL-1웁 by a 40- to 50-fold molar ratio. Strikingly, in neutrophil-depleted rabbits, IL-8 induced no cartilage destruction and far lesser concentrations of IL-1웁 and IL-1RA as compared with normal rabbits (434), proving that inﬁltrating neutrophils were the main producers of these cytokines and that they were responsible for cartilage destruction. What is more, IL-8 induced little macrophage/lymphocyte accumulation in neutrophildepleted rabbits, suggesting that early neutrophil accumulation may affect the later accumulation of macrophages or lymphocytes, likely through the production of speciﬁc chemoattractants. The authors thus concluded that IL-8 is a potent neutrophil activator in vivo and may have a crucial role in the biology of inﬂammation and the pathogenesis of inﬂammatory processes, including septic arthritis (434). In more recent papers, the same

NEUTROPHIL-DERIVED PROTEINS

461

group investigated the network and involvement of inﬂammatory cytokines in their rabbit model of LPS-induced arthritis (435). The study was based on the assumption that production of TNF-움 precedes that of IL-1웁 and IL-1RA and that IL-8 is detectable after IL-1. Surprisingly, maximum levels of TNF-움 and IL-8 were detected 2 hr after LPS injection, whereas IL-1웁 and IL-1RA were detected after 6 and 9 hr, respectively. By IH, synovial cells were positive for TNF-움 and IL-8, and inﬁltrating leukocytes were positive for IL-1웁, IL-1RA, and IL-8. In neutrophil-depleted rabbits, the levels of TNF-움 and IL-8 were similar to those in normal rabbits. In contrast, no IL-1웁 was detected in neutrophil-depleted rabbits and the levels of IL-1RA were much lower (435). The effects of neutralizing antibodies against TNF-움, IL-1웁, and IL-8 led to many important observations. First, TNF-움 and IL-8 were produced by synovial lining cells, were the ﬁrst cytokines appearing at the site of inﬂammation, and induced subsequent production of IL-1웁 and IL-1RA by neutrophils. Second, IL-1웁 induced further production of IL-1웁, and, endogenous IL-1RA down-regulated the production of IL-1웁 but not that of TNF-움 or IL-8 (435). Interestingly, the early phase of the leukocyte inﬂux was not blocked by inhibitors of each cytokine, indicating that this phase was dependent on other chemoattractants, such as C5a, PAF, or LTB4 (435). Furthermore, late accumulation of neutrophils was inhibited only by 40 to 60% by anti-TNF-움 mAbs, recombinant IL-1RA, or anti-IL-8 IgG, suggesting that factor(s) other than IL-8 (probably GRO-움, MIP-2, or ENA-78) were involved in the late phase of leukocyte inﬂux in LPS-induced arthritis (435). That IL-8 is partly involved in TNF-움- but not in IL-1웁-induced neutrophil recruitment (436) was subsequently conﬁrmed by Matsushima’s group (437, 438). Finally, MCP-1 was immunohistochemically detected in synovial lining cells and inﬁltrating neutrophils, but the amounts of MCP-1 detected in SF from neutrophil-depleted rabbits were similar to those in normal rabbits, suggesting that synovial lining cells were the main source of MCP-1 detected in SF (439). Administration of neutralizing anti-MCP-1 antibody inhibited LPS-induced monocyte inﬁltration by 58.4%, suggesting that synovial production of MCP-1 plays an important role in the recruitment of monocytes in these arthritis models (439). Yoshinaga and colleagues (440) investigated the involvement of TNF움, IL-1웁, and IL-1RA in LPS-induced uveitis. Intravitreal injection of LPS in rabbits induced a massive leukocyte inﬁltration and protein leakage into the aqueous humor, with the peak 24 hr postinjection. TNF-움, IL-1움, and IL-1RA were signiﬁcantly augmented, and leukocyte-depletion studies showed that inﬁltrating leukocytes were the major cellular sources of these cytokines. Further experiments revealed that TNF-움 and IL-1웁 were the

462

MARCO ANTONIO CASSATELLA

principal mediators of LPS-induced uveitis and that endogenous IL-1RA was able to down-regulate inﬂammatory reactions (440). B. ADDITIONAL in Vivo MODELS OF ACUTE INFLAMMATION INVOLVING NEUTROPHIL-DERIVED CYTOKINES Mori et al. (441) examined by Northern blot analysis the expression of 12 different genes in peritoneal exudate neutrophils, harvested at 5 and 24 hr after IP injection of casein in rabbits. Though mRNA from IL-1움, TNF-움, and MCP-1 mRNA was below detection levels during the entire inﬂammatory period of observation, the remaining nine gene products were classiﬁed into three categories (441). The ﬁrst group included 웂actin, MRP-8, and MRP-14, the latter two being calcium-binding proteins and components of a complex molecule with inhibitory activity against casein kinase I and II. These messages were constitutively expressed in blood neutrophils and were also rapidly induced after emigration into inﬂammatory sites. The second group of gene products included IL-1웁, IL-8, MIP-1웁, and the fMLP-R, which were induced rapidly after the onset of inﬂammation (2–5 hr), but returned to basal levels of expression by 24 hr. The functions of the second group of gene products relate especially to chemotaxis, one of the hallmarks of early inﬂammation. To the third group of expressed genes, only ferritin-related mRNAs (F and H chains) were ascribed, because they were induced slowly (4–7 hr) and increased with the progression of the inﬂammatory process. This study not only underlined that neutrophils contribute to the acute inﬂammatory reactions by synthesizing a variety of proteins for a fairly long period, but also evidenced that such response is regulated and subjected to a programmed sequence. Expression of IL-1웁 was in agreement with previous ﬁndings reported by the same group, who, by using IH staining at the single-cell level, unequivocally proved that polymorphonuclear leukocytes were the major cells synthesizing IL-1 following casein-induced inﬂammation in rabbits (222, 223, 442). Furthermore, an IL-1-like activity was produced by neutrophils obtained by bronchoalveolar lavage from experimentally inﬂammed rat lung. Activity was released spontaneously from neutrophils at high levels but it was enhanced by stimulation with endotoxin in vitro (220). Yoshinaga’s group analyzed the cytokine network involving TNF-움, IL1웁, IL-8, and IL-1RA in a rabbit experimental model of acute gout (443). Though the production of TNF-움 in synovial ﬂuids reached a peak at 2 hr after intraarticular injection of monosodium urate crystals, the production of IL-1웁 and IL-8 occurred in two phases, the ﬁrst at 2 hr and the second at 9 and 12 hr, respectively. By contrast, the production of endogenous IL-1RA reached a peak at 9 hr (443). Inﬁltrating leukocytes,

NEUTROPHIL-DERIVED PROTEINS

463

including neutrophils, were the source of IL-1웁 and IL-1RA and of the second-phase IL-8 (443). Studies with speciﬁc neutralizing antibodies or other inhibitors demonstrated that IL-8 and TNF-움 were independently produced in the early stage of MSU crystal-induced arthritis and were responsible for the production of IL-1웁 and IL-1RA, whereas IL-1웁 was responsible for the second phase of expression of IL-1웁 and IL-8. Thus, the production kinetics of IL-8 and IL-1웁 in MSU crystal-induced arthritis was completely different from that observed in LPS-induced arthritis (see above) (435). Intravenous injection of colchicine completely inhibited neutrophil inﬁltration and damage without affecting the production of TNF움 or the ﬁrst peak of IL-8, suggesting that colchicine inhibits MSU crystalinduced arthritis by directly inhibiting the migration of neutrophils (443). Further evidence on the pathogenic roles of locally produced IL-8 in rabbit MSU crystal-induced gouty arthritis has been provided by Nishimura and colleagues (444), who detected immunoreactive IL-8 protein in synovial lining cells at 12–24 hr after the MSU crystal injection, and also in inﬁltrated neutrophils in synovium (444). MCP-1 was also detected in synovival lining cells and inﬁltrating neutrophils by immunohistochemistry (439). Production of MCP-1 in crystal-induced arthritis was independent of TNF-움 or IL-1, but was shown to play an important role in the recruitment of monocytes in this arthritis model (439). In work aiming to identify the neutrophil chemoattractants generated in a model of myorcadial infarction in the rabbit, Ivey et al. (445) attributed important roles to the complement fragment C5a and to IL-8. Ischemia induces all the typical changes characteristic of an acute inﬂammatory response, among which an early neutrophil accumulation is a prominent feature. A determinant step in neutrophil accumulation is the local generation of chemical signals responsible for leukocyte recruitment. Neutrophil accumulation is markedly accelerated during reperfusion after ischemia, and early studies have implicated PMNs in the generation of tissue damage associated with reperfusion (446). In their study, Ivey et al. (445) demonstrated that immunoreactive C5a and IL-8 were present in myocardial tissue after ischemia and reperfusion, but the time course of their appearance was quite different. C5a was detected after 5 min of the initiation of reperfusion, whereas IL-8 concentrations rose slowly and were signiﬁcantly elevated at 1.5 hr and were highest at 4.5 hr, in close parallel with leukocyte inﬁltration. Further experiments revealed that neutrophil depletion virtually abolished IL-8 generation in the myocardium, but had no inﬂuence on C5a generation. Therefore, in this model, C5a was probably liberated from preformed substrates as early as a few minutes after the initiation of reperfusion and induced a ﬁrst phase of neutrophil inﬁltration. Once in the tissue, neutrophils became the source of IL-8 in the myocardium, and

464

MARCO ANTONIO CASSATELLA

this IL-8 generation was responsible for a subsequent wave of neutrophil accumulation (445). In studies attempting to clarify the mechanisms by which mononuclear cells are attracted to a previously ischemic and infarcted myocardium, Youker and colleagues, in a dog experimental model, provided preliminary evidence that even neutrophils play a role in myocardial healing (447). What is more, in this study PMNs were shown to express IL-6 mRNA and protein, but the signiﬁcance of this was not discussed (447). Other ﬁndings have involved locally produced IL-8 as a pivotal mediator of cerebral reperfusion (448). Reperfusion of rabbit brain after a transient focal ischemia induced a perivascular neutrophil inﬁltration and aggregation as well as tissue damage (448). Brain tissue levels of IL-8 increased signiﬁcantly at 6 hr after reperfusion, without a noticeable elevation of plasma IL-8 levels. IL-8 protein was detected by IH in the vascular wall and, to a lesser degree, in inﬁltrated neutrophils. In addition, a neutralizing anti-IL-8 antibody signiﬁcantly diminished neutrophil inﬁltration and reduced brain edema and infarct size in comparison to rabbits receiving a control antibody (448). As a whole, the results suggested not only a crucial role of neutrophil inﬁltration in this cerebral reperfusion injury model but also that IL-8 might be considered as a novel target for the intervention of this injury. In a rat model of transient retinal ischemia, a condition that leads to neuronal damage, Hangai et al. (449) studied the levels of IL-1 gene expression by semiquantiative RT–PCR, and also used ISH and IH. Little expression of IL-1움 and IL-1웁 genes was observed in normal retina, but this was highly up-regulated after ischemia and subsequent reperfusion, in a time-dependent manner (449). Time courses of IL-1움 and IL-1웁 mRNA expression were also different, in that induction of IL-1움 mRNA occurred before that of IL-1웁 mRNA. For IL-1웁, three types of cells, including neutrophils, were identiﬁed as the cellular origin of mRNA. The authors speculated that neutrophils recruited after ischemia are activated and consequently synthesize IL-1, which then promotes secretion of products that damage the microvasculature and retinal tissue (449). In a rat model of lung injury obtained by IT instillation of bleomycin, which subsequently leads to ﬁbrosis, Sakanashi et al. (450) investigated the kinetics and the molecular mechanisms underlying macrophage inﬁltration. Northern blot analysis revealed that the expression of MCP-1 mRNA in the lung was most prominent the ﬁrst day after instillation and declined thereafter, thus preceding the numerical change of the exudate monocytes. IH disclosed that the main sources of MCP-1 production were alveolar and interstitial macrophages, as well as PMNs (450). Based on these results, the authors speculated that MCP-1 produced by PMNs and by alveolar

NEUTROPHIL-DERIVED PROTEINS

465

and interstitial macrophages induced the inﬁltration of blood monocytes in the very early phase, and that the subsequent accumulation of macrophages was enhanced by the MCP-1 production by monocyte-derived exudate macrophages (450). Under similar experimental conditions, erythromycin was shown to exert a prophylactic effect on the acute lung injury induced by intratracheal administration of bleomycin, an effect probably linked to a down-regulation of neutrophil-derived elastase and other mediators (cytokines?) (451). An important component of the pathophysiologic response to hyperoxia is pulmonary inﬂammation, although the roles of speciﬁc inﬂammatory mediators during pulmonary hyperoxia toxicity are not completely known. In rabbits exposed to hyperoxia (452), a quantitative ISH of BAL cells showed that both IL-8 and MCP-1 were expressed in AMs, whereas only IL-8 was present in recruited PMNs. Interestingly, IL-8 mRNA production in PMNs was elevated throughout the time that PMNs were available for analysis, and although no data on IL-8 protein were provided, the presence of increased levels of IL-8 mRNA in PMNs entering the alveolus implies an autocrine role for this cytokine in PMN activation (452). Other studies have focused on cytokine regulation during hyperoxic lung injury in adult and neonatal mice (453, 454). In one study, lungs from adult mice assayed by Northern blot analysis displayed increased levels of IL-1웁 mRNA after 2 days of hyperoxia, whereas IL-1움 mRNA was barely detectable (453). In situ hybridization and immunohistochemical analyses revealed an accumulation of IL-1웁 transcripts and protein in pulmonary interstitial macrophages and in a subset of neutrophils (453). In another study, Johnston et al. (454) examined hyperoxic lung injury in neonatal and adult mice. Neonatal animals of several species are more tolerant to hyperoxic exposure than are adults. However, the mechanisms of increased neonatal tolerance are unknown, as are the cell types that contribute to oxygen resistance. Adults and neonatal mice were exposed to more than 95% oxygen for 78 hr and 10 days, respectively, and lung mRNA was assayed by ribonuclease protection assays (RPA). After 72 hr of exposure, the messages encoding TNF-움, IL-1웁, and IL-6 were increased twofold in adult lungs. However, at this time point these mice were dead nearly dead. No alterations in neonatal lung mRNA were detected until 7 days of oxygen exposure. At that time neonatal mice demonstrated increases in lung mRNA encoding TNF-움, IL-1웁, and IL-6 of three-, ﬁve-, and eightfold, respectively. Acute alveolitis and slight edema were detected, but lethality was not observed until 10 days of exposure. In situ hybridization in neonatal mice revealed an accumulation of TNF-움 and IL-1웁 transcripts in pulmonary interstitial macrophages and in a subset of neutrophils after 7 days of exposure. Messages encoding IL-1움, IL-2, IL-3, IL-4, IL-5, IL-10, IFN-웂, and

466

MARCO ANTONIO CASSATELLA

LT-움 were not altered in either adult or neonatal mice controls at any time point examined. In conclusion, adult mice demonstrated little change in cytokine mRNA until lethality was imminent, whereas newborn mice demonstrated an acute induction of TNF-움, IL-1웁, and IL-6 early in the development of hyperoxic injury. A rapid cytokine response early in the development of hyperoxic injury may play an important role in the adaptation of neonatal lungs to toxicity from prolonged oxygen exposure. Using the model of antiglomerular basement membrane (anti-GBM) nephritis in rats, Wu et al. (26) investigated the mechanisms underlying in situ chemokine expression and the in vivo function of these molecules during the acute phase of this inﬂammatory model. Previous studies had established that CXC chemokines are important for the acute inﬂux of PMNs in anti-GBM nephritis and the consequent damage to the glomerulus (manifested as proteinuria) (455). Other studies had also implicated CC chemokines in the pathophysiology of glomerulonephritis (456). Wu et al. (26) described that during the evolution of anti-GBM nephritis, CXC chemokine expression (MIP-2 and CINC) was monophasic and paralleled neutrophil inﬂux, whereas CC chemokine expression (MIP-1움, MIP-1웁, and MCP-1) was biphasic, with peaks coinciding with the inﬂux of PMNs and then macrophages. IL-1웁 and TNF-움 mRNA expression exhibited features intermediate between those of CXC and CC chemokines. The initial peak of chemokine expression was attenuated by decomplementation (which selectively attenuates PMN inﬂux), neutropenia (by cell depletion with a speciﬁc antiserum), and irradiation-induced leukopenia. The delayed peak was attenuated only by leukopenia, but was augmented in the accelerated form of this disease model, corresponding to an increase in macrophage inﬂux. Differential expression of chemokines by PMNs and macrophages was not an intrinsic property of these cells, because these leukocytes expressed similar proﬁles of chemokines in vitro, either after adherence or with endotoxin stimulation (26). In general, macrophage expression of chemokine mRNA was quantitatively more modest than in PMNs. IH for MIP-1움 in acute nephritis validated that expression during acute nephritis was accompanied by local protein production. Moreover, neutralizing Abs to MIP-1움 attenuated the acute-phase proteinuria, but not the accompanying inﬂux of PMNs or macrophages. In comparison, neutralizing mAbs to CINC inhibited both PMN inﬂux and proteinuria. A combination of the two antibodies was not signiﬁcantly more effective than either antibody alone (26). The study conclusively established that myeloid cells are necessary for glomerular chemokine expression during nephritis, and that the differential expression of CXC and CC chemokines not only relates speciﬁcally to the differential inﬂux of leukocyte subsets, but must involve additional factors. In addition, this study argued against the simpliﬁed scheme

NEUTROPHIL-DERIVED PROTEINS

467

that CXC chemokines are mediators of acute inﬂammation, and CC chemokines are mediators of chronic inﬂammation. C. NEUTROPHIL-DERIVED CYTOKINES IN in Vivo INFECTIONS In a very interesting study, Jordan et al. (457) determined the endogenous mediators involved in the induction of IL-1RA during oral infection of mice with the enteropathogenic Y. enterocolitica. These bacteria initially proliferate in the tissue of the terminal ileum, predominantly in the Peyer’s patches (PPs), where the immediate antibacterial host defense is characterized by an inﬁltration of granulocytes and monocytes. By ISH, Northern blot, and IH, the authors found expression of IL-1RA mRNA and synthesis of IL-1RA in PPs, as well as in noninfected organs such as spleen, but not in the liver (457). In contrast, the mRNA for IL-1웁 in PPs was expressed considerably earlier, because sessile macrophages were its primary source. No temporal differences were observed between IL-1움 and IL-1RA (457). Circulating and recruited neutrophils, but not PBMCs, were identiﬁed to be the primary source of IL-1RA in tissues, whereas approximately 20% of the positive IL-1RA-staining cells were accounted for by inﬂammatory macrophages. In addition, ISH of adjacent sections of PPs revealed a distinct hybridization pattern for each cytokine, suggesting that IL-1움, IL1웁, and IL-1RA were produced independently by different cell types, or alternatively by cells of the same phenotype located within different tissue areas. Strikingly, neutralization with an antiserum of IL-6, a cytokine that was also promptly induced by Y. enterocolitica infection, caused suppression of both IL-1RA mRNA in PPs and synthesis of IL-1RA in circulating neutrophils. In support of these in vivo ﬁndings, IL-6 induced IL-1RA expression in cultures of macrophages and PMNs in vitro, and anti-IL-6 antiserum blocked these effects of IL-6 (457). In this respect, previous studies in humans had demonstrated that IL-6 infused into cancer patients rapidly increased the levels of circulating IL-1RA (458). Altogether, the observations of Jordan et al. (457) uncovered important interrelationships among IL-1, IL-6, and IL-1RA in Y. enterocolitica infections. For example, after the production of IL-1 and IL-6 early after Yersinia infection, IL-6 in turn induces IL-1RA, which then may inhibit IL-1 activities through a negative feedback loop, thus facilitating the resolution of the inﬂammatory response locally and presumably at remote sites of infection. More insights into the mechanisms underlying neutrophil inﬂux in the airways during chronic bacterial infection were uncovered by Inoue et al. (112), who studied the in vivo effect in dog trachea of P. aeruginosa supernatants on the expression and localization of IL-8 mRNA in airways. Application of the supernatants stimulated IL-8 mRNA expression in epithelial and gland duct cells and in recruited neutrophils (112). The molecule

468

MARCO ANTONIO CASSATELLA

responsible for this IL-8 induction was a small-molecular-mass (1-kDa) product of P. aeruginosa (112). IL-8 expression in recruited neutrophils suggests a potential mechanism for amplifying the inﬂammatory response and for a positive feedback of a protective antibacterial response, for example, by rendering phagocytosis more effective. Expression of TNF-움 and IL-1움 transcripts, but not of IL-6, in neutrophils elicited from L. monocytogenes-infected mice was demonstrated by Dai et al. (459). These observations have been made in elicited peritoneal PMNs from IFN-웂 receptor-deﬁcient (IFN-웂R⫺/⫺) mice, used as a model to study the innate immune responses during infection with L. monocytogenes (459). In agreement with previous reports on the essential role of IFN-웂 to limit bacterial spread, these mutant mice were unable to limit bacterial growth and died of sepsis even with an infection dose of seventy Listeria microorganisms. The authors detected large inﬂammatory foci of infection in the spleen or liver of IFN-웂R⫺/⫺ mice; the foci were populated mainly by PMNs that, interestingly, were fully able to kill Listeria (459). Nevertheless, despite their fundamental contribution in innate immune responses against Listeria, neutrophils were unable to rescue mice from fatal listeriosis, arguing for their limited protective role in IFN-웂R⫺/⫺ mice. Thomas and colleagues investigated the role of neutrophils in herpetic stromal keratitis (HSK) (460). HSK is an immunopathologic response observed in immunocompetent mice after corneal infection with herpes simplex virus-1. The earliest sign of disease is a speciﬁc neutrophil inﬁltration, which lasts for 48 to 72 hr and then disappears. This rapid PMN response most likely contributes to curtailing viral replication, minimizing viral dissemination. In search of how PMNs exert antiviral effects, the researchers performed ISH experiments. Preliminary results revealed TNF-움 and iNOS signals in cells resembling PMNs in acute inﬂamed cornea (460). These molecules might thus play an antiviral activity against HSV. In addition, results from the same group showed that enriched PMNs from mice peritoneum produce IL-12 on exposure to HSV (297). Finally, Romani’s group (291, 292) has provided fundamental information on the role of neutrophils in the generation of murine T helper responses to C. albicans. It is well known that subsets of CD4⫹ T helper cells can be characterized on the basis of their pattern of cytokine production either in mice (461) or in human systems (53). Th1 cells predominantly produce IL-2, IFN-웂, and LT-움, and are effective inducers of delayedtype hypersensitivity (DTH), whereas Th2 cells mainly produce IL-4, IL5, and IL-10 and provide more effective help for B cells. Human Th1like cells preferentially develop during infections by intracellular bacteria, protozoa, and viruses, whereas Th2-like cells predominate during helminthic infestations and in response to common environmental allergens (53).

NEUTROPHIL-DERIVED PROTEINS

469

Strongly polarized human Th1-type and Th2-type responses play different roles in protection and can also promote different immunopathological reactions (53). Candida albicans is a commensal microorganism that, especially in immunonocompromised hosts, may represent an important cause of morbidity and mortality. Studies in mice have clearly established that multiple mechanisms may control the outcome of experimental infections (462). In immunized mice the outcome is greatly conditioned by the type of predominant T helper cell subset activated by the initial exposure to the yeast: Th1 cell activation leads to resistance and onset of durable protection, whereas Th2 cell responses are associated with susceptibility to progressive disease (462). Neutrophils have a major role in providing a ﬁrst line of defense in C. albicans infection (463, 464). Previous studies in Romani’s laboratory had indicated that numerous factors are involved in preferential induction of murine Th1 or Th2 cell reponses to Candida (462). Cytokines emerged, obviously, as key regulators in the development of CD4⫹ subsets from precursor Th cells (462), and for the Th1 responses several lines of evidence indicated production of IL-12 as the determinant (463, 465). Importantly, depletion of granulocytes in resistant mice led to the onset of Th2 rather than Th1 responses, indicating that the latter cells may participate in Candida-driven Th1 development (463). It is worth remembering here that neutropenia constitutes in humans one of the major factors responsible for fungal dissemination to visceral organs. Using a live vaccine strain or virulent challenge in mucosal or systemic infections of mice with C. albicans, Romani’s group (291) initially examined the effect of depletion of neutrophils on the course of primary and secondary challenge and on development of CD4⫹ cell-dependent immunity. They obtained evidence of deleterious effects of neutrophil depletion occurring at the time of infection under all conditions of testing, both in naive and in previously immunized mice (291). Neutrophil depletion concurrent with infection also resulted in the selective appearance of the IL-4 message in puriﬁed CD4⫹ splenocytes, an early indicator of Th2 development. In contrast, PMN depletion appeared to beneﬁt the hosts late in the course of an overwhelming systemic infection. In an attempt to correlate neutrophil function with the nature of the T cell response, the authors also tested the ability of neutrophils to produce cytokines associated with functionally distinct CD4⫹ Th cell responses to Candida. They found that neutrophils were endowed with the capacity to secrete IL-12 and IL-10 in vitro in response to different strains of C. albicans or to IFN-웂 plus LPS. In particular, IL-10 and bioactive IL-12, but not IL-4, were found in cultures stimulated with IFN-웂 plus LPS. IL-12 but not IL-10 was produced in response to the live vaccine strain PCA-2, which causes healing infection,

470

MARCO ANTONIO CASSATELLA

while the opposite pattern was observed in response to the highly virulent CA-6 strain, which causes nonhealing infection (291). The authors therefore concluded that neutrophil ablation early in the course of Th1associated self-limiting infection appeared to change the qualitative development of the T cell response, and rendered mice susceptible to infection (292). Subsequently, the ability of neutrophils to release IL-12 and IL-10 in vivo, during the course of C. albicans systemic infection, was investigated by injecting intravenously the live vaccine strain PCA-2, or the CA-6 strain (292). Under those conditions, neutrophils expressed secreted IL-12 and IL-10, correlating with the respective development of self-limiting (Th1associated) and progressive (Th2-associated) disease (292). Importantly, macrophages were characterized as cells with a poor ability to secrete IL-10 and, even less, IL-12 (293). Neutrophil depletion prevented the development of protective Th1 responses in healer mice, but exogenous IL-12 was effective in protecting neutropenic hosts susceptible to infection, consistent with a role for neutrophil-derived IL-12 in Th1 development (292). Neutrophil depletion, however, increased resistance later in infection of susceptible host, the latter ﬁnding being related to a decreased IL-10 production (292). Another very important observation in the studies of Romani and colleagues was that the balance between IL-10 and IL-12 production by neutrophils was modiﬁed by exogenous IL-12, in that PMN release of IL-10 increased after IL-12 treatment in both uninfected and infected mice (292). Although this IL-12-induced production of IL-10 by neutrophils might have been the result of indirect mediators stimulating PMNs, this mechanism could act as a regulatory response to challenge with IL-12. In addition, such an effect of IL-12 might account for an observation previously made by Romani’s group of a paradoxical effect of IL-12 in the resistant host. They in fact reported that administration of IL-12 not only fails to promote (enhance) protective anticandidal immunity in nongranulocytopenic mice, but actually promoted Th2 development in a healing infection with detectable levels of circulating IL-10/IL-4 (292). The increased production of IL-10 by neutrophils after IL-12 treatment might be the explanation for this, or could contribute to the failure of IL12 to exert protective effects in mice with candidiasis (292). An unbalanced overproduction of IFN-웂 mediated by IL-12 administration has been also proposed as a factor that may lead to an enhanced susceptibility to Candida infections (466). In summary, the results of Romani et al. (291, 292) are very important because they demonstrate that PMNs, through the release of IL-12 and IL-10, may signiﬁcantly contribute to the patterns of susceptibility and resistance in mice with candidiasis. Moreover, they indicate that neutrophils have not only an effector role in C. albicans infection, but also an immunomodulatory one, regulating Th1 and Th2 differentiation. More

NEUTROPHIL-DERIVED PROTEINS

471

strikingly, the work of Romani et al. (292, 295) has demonstrated for the ﬁrst time that neutrophils, via their ability to release cytokines, play an active role in determining the qualitative development of the T cell response (462). Further evidence for an in vivo role of IL-12-producing neutrophils as initiators of a Th-1 cell-mediated immunity has been observed in response to Toxoplasma gondii or in the IFN-웂R⫺/⫺ mice, as mentioned previously (467). All the latter concepts likely originated from the work of my group (12). Romani’s group has also uncovered a previously unappreciated role for endogenous IL-4 to induce a protective antifungal CD4⫹ Th1 response in mice with C. albicans infection (293). IL-4 deﬁcient mice, while having an impaired Th2 response, did not default to the Th1 pathway, thus becoming highly susceptible in the late stage of C. albicans infection (293). Defective IFN-웂 and IL-12 production, but not IL-12 responsiveness, was observed in IL-4-deﬁcient mice that failed to mount protective Th1-mediated acquired immunity in response to a live vaccine strain of the yeast or on mucosal immunization in vivo. However, late treatment with exogenous IL-4, while improving the outcome of infection, potentiated CD4⫹ Th1 responses even in the absence of neutrophils. In the same study, the authors found that IL-4 efﬁciently primed neutrophils for IL-12 production in response to the fungus, and the effect was associated with the induction of IL-4 receptor on these cells (293). Priming with IL-4 also resulted in the release of high levels of IL-6 by neutrophils in vitro. The ﬁndings indicate that endogenous IL-4 is required for the induction and maintenance of IL-12-dependent protective antifungal responses, possibly through combined activity on cells of the innate and adaptive immune systems. Other studies from the same group reported that TNF/LT-움 doubledeﬁcient mice are more susceptible to infection caused by virulent or low-virulence C. albicans cells, and this susceptibility correlates with an impaired development of protective Th1 response (296). In this model, neutrophils from controls and TNF/LT-움 double-deﬁcient mice produced comparable levels of bioactive IL-12 (296). This suggested that secretion of IL-12 by PMNs occurs independently of TNF signaling and that the impaired development of protective antifungal Th1 responses in TNF/LT움-deﬁcient mice occurs in spite of IL-12 production (296). In contrast, a negative effect of excess iron on the antifungal effector functions of neutrophils and macrophages, and on the expression of IL-12 mRNA, particularly in neutrophils from infected mice, was found in another work (293). Iron overload greatly increased susceptibility to disseminated infection with low-virulence C. albicans cells (295). Therefore, it appeared that one likely mechanism by which iron overload increases the susceptibility to infection is through inhibition of a directive cytokine, such as IL-12 (295).

472

MARCO ANTONIO CASSATELLA

D. A ROLE FOR NEUTROPHIL-DERIVED CYTOKINES IN OTHER in Vivo EXPERIMENTAL MODELS Other studies have indicated that the in vivo production of cytokines by PMNs may signiﬁcantly affect other processes, for instance, the antitumor and immune response. In a series of papers published by Sendo and co-workers (468–472), for example, cell-mediated immune responses and antibody production were analyzed in rats that were depleted of PMNs using a speciﬁc antineutrophil monoclonal antibody designated RP-3. Those experiments demonstrated that both the priming and the effector phases of delayed-type hypersensitivity to sheep red blood cells (SRBCs) (468) were partially inhibited in PMN-depleted rats, possibly through inhibition of MNC recruitment in DTH (469, 471), suggesting that neutrophils enhance DTH to SRBCs. The same group previously demonstrated that IL-8-induced CD4⫹ T lymphocyte recruitment into subcutaneous tissues of rats was inhibited by the RP-3 treatment (471). Furthermore, by assessing the direct or indirect splenic plaque-forming cell (PFC) response to SRBCs in rats depleted of PMNs 6–12 hr before immunization, the authors detected an increased number of anti-SRBC antibody-producing cells (470). This phenomenon was observed only when the antigen was administered intraperitoneally and not with IV immunization (471) and suggested that neutrophils could suppress antibody production in certain situations. More recently, they also showed that G-CSF administered at the elicitation phase enhances DTH response to SRBC and MNL recruitment (472). Using a similar experimental animal model, the same group also demonstrated that transplantation immunity against cancer and generation of CD8⫹ effector T cells in response to tumor-associated antigens were abrogated by selective depletion of neutrophils (473, 474). Although the precise mechanisms underlying all these phenomena were not elucidated, it can be envisaged that they are related to the lack of PMN-derived cytokines. These molecules would affect, for instance, antigen presentation (IL-1), lymphocyte proliferation and activation (IL-1, TNF-움), macrophage activation (TNF-움), and leukocyte recruitment (IL-8, IP-10, MIG, MIP-1움/MIP-1웁, CINC, and MIP-2). The potential ability of PMNs to mediate antitumor activity in vivo has been clearly elucidated by Stopacciaro et al. (151). They took advantage of the murine colon adenocarcinoma C-26 cell line engineered to release G-CSF (C-26/G-CSF), to study the mechanisms responsible for inhibition of tumor uptake in syngeneic animals, and of regression of an established tumor in sublethally irradiated mice injected with these cells. Using C-26/ G-CSF they identiﬁed the cell types that inﬁltrate the tumor and the cytokines expressed in situ. It was found that inhibition of tumor uptake and

NEUTROPHIL-DERIVED PROTEINS

473

regression of an established tumor in sublethally irradiated mice occurred through different mechanisms. In the former case, PMNs were the main cells responsible for inhibiting the uptake of C26/G-CSF. In the latter case PMNs, macrophages, and T cells, including CD8⫹ T cells, which are required for IFN-웂-mediated tumor regression, determined the rejection of a C26/G-CSF nodule initially grown in sublethally irradiated mice. Both depletion of CD8⫹ T cells and neutralization of IFN-웂 produced by CD8⫹ T cells resulted in a reduction of PMN number and TNF-움 expression, and therefore in tumor progression (151). Notably, as evidenced by IH and ISH, either newly recruited granulocytes surrounding the injected neoplastic cells, or, in sublethally irradiated mice, the PMNs inﬁltrating the C-26/G-CSF tumor during its initial growing phase, expressed transcripts for IL-1움, IL-1웁, and TNF-움 (151). In another study (475), analysis of the phenotypic changes resulting from cytokine activities during rejection of C26/G-CSF, and inhibition of such changes by anti cytokine antibodies, indicated that TNF-움 was instrumental in tumor regression. C-26/ G-CSF regresssing tumors were characterized by hemorrhagic necrosis dependent on the inﬁltrating leukocytes and the cytotoxic cytokines they produced (475). Complete tumor regression was the result of tumor cell hypoxia following damage of the tumor microvasculature, which was the target of the cytotoxic cytokines (TNF-움) and the PMNs (475). Locally produced IL-1 and TNF-움 induced VCAM-1 and E-selectin on tumor vessels, and thus indirectly attracted T lymphocytes (475). Treatment with monoclonal antibodies to IFN-웂 or TNF-움 blocked tumor regression by inhibiting VCAM-1 and E-selectin expression on tumor-associated endothelial cells, and this resulted in a reduced number of inﬁltrating leukocytes. Thus, whereas tumor inhibition was mediated mainly by PMNs, tumor regression occurred because of the cooperation of PMNs and T cells, as well as of a combination of cytokines, for which T cell-derived IFN-웂 and PMN-derived TNF-움 were necessary. Neutrophil production of cytokines in vivo is implicated in other processes. For instance, in studies aiming to elucidate the pathogenesis of pulpitis and apical periodontitis, Tani-Ishii et al. (476) identiﬁed cells that express IL-1움 and TNF움- in infected pulps and in developing rat periapical lesions after surgical pulp exposure. As detected by immunohistochemistry, IL-1움 and TNF-움 positive cells were present as early as 2 days after pulp exposure in both the pulp and the periapical region (476). In contrast, cells expressing IL-1웁 and LT-움 were not found in pulp or periapical lesions during this period. Cells expressing IL-1움 and TNF-움 were identiﬁed primarily as macrophages and ﬁbroblasts, with occasional staining of PMNs (476). Osteoblasts and osteoclasts were also positive, whereas lymphocytes were negative. In general, cytokine-expressing cells were lo-

474

MARCO ANTONIO CASSATELLA

cated proximal to abscesses and the root apex (476). The data demonstrated that resident connective tissue cells, as well as inﬁltrating leukocytes, express bone-resorptive cytokine in response to infection in these lesions. Expression of IL-1움 and TNF-움 in neutrophils inﬁltrating early periapical lesions of normal and immunodeﬁcient mice was subsequently conﬁrmed (477) and its importance further demonstrated (478). The importance of IL-1웁 in the pathogenesis of acute pancreatitis has been demonstrated by dramatic attenuation of pancreatic destruction and signiﬁcant increases in survival when its actions are inhibited. Hypothesizing that inﬁltrating leukocytes contribute substantially to the intrapancreatic production of IL-1웁, Fink and Norman (479) examined the speciﬁc role of those cells. Mice were assigned to one of four groups 48 hr prior to induction of pancreatitis: (1) PMN depletion via antimurine PMN antiserum (PMN-d), (2) macrophage (M␾) depletion via antimacrophage antiserum (M␾-d)(3) PMN and M␾ depletion [PMN ⫹ M␾-d], and (4) immunocompetent pancreatitis. Edematous pancreatitis was then induced in all experimental groups by caerulein, and intrapancreatic IL-1웁 production was determined by IH and RT–PCR (479). The experiments performed by these authors demonstrated that intrapancreatic IL-1웁 production was primarily attributable to the leukocytes inﬁltrating the gland during the progression of the disease. IH techniques suggested that the macrophage was the major contributor of IL-1 protein. On the other hand, there was greater attenuation of IL-1 mRNA levels in animals that were devoid of neutrophils (479). Elimination of either macrophages or neutrophils (and their inﬂammatory products, including IL-1웁) could thus have beneﬁcial effects and signiﬁcantly decrease the severity of pancreatic destruction. To obtain greater insights into the pattern of cytokine expression in wound tissues, and in their regulation during the repair process, Hubner and colleagues (480) have been able to show a strong and early induction of IL-1움, IL-1웁, and TNF-움 expression after cutaneous injury in normal mice. The highest levels of these cytokines were seen as early as 12– 24 hr after wounding, and after completion of the proliferative phase of wound healing, mRNA levels of these cytokines returned to the basal level. Remarkably, during the early phase of wound repair, proinﬂammatory cytokines were predominantly expressed in PMNs. At later stages of the repair process, expression of IL-1움, IL-1웁, and of TNF-움 was also seen in macrophages. Induction of these cytokines after injury was signiﬁcantly reduced during wound repair in healing-impaired glucocorticoid-treated mice (480). These ﬁndings demonstrate that wound healing defects are associated with impaired IL-1움, IL-1웁, and TNF-움 expression and suggest that early induction of these genes is important for normal repair. More importantly, the data provide evidence for a novel function of PMNs as

NEUTROPHIL-DERIVED PROTEINS

475

regulators of inﬂammatory process and as initiators of proliferative processes. This hypothesis was supported by the colocalization of cytokineexpressing PMNs and, for example, KGF-expressing ﬁbroblasts at the wound edge. Thus, the early expression of proinﬂammatory cytokines by PMNs that invade the wound seems to be of major signiﬁcance for induction of growth factor expression and thus for the initiation of wound repair. In another model of local wounding response secondary to spinal cord injury in mice, Bartholdi et al. (481) investigated the expression pattern of proinﬂammatory and chemoattractant cytokines. They could show by ISH that transcripts for TNF-움 and IL-1 as well as MIP-1움 and MIP-1웁 were up-regulated within the ﬁrst hour following injury. In this early phase, expression of the proinﬂammatory cytokines was restricted to cells, probably resident CNS cells, in the area surrounding the lesion. Though TNF-움 was expressed in a very early time window, IL-1 could be detected in a subset of PMNs that immigrated into the spinal cord around 6 hr. Messages for the chemokines MIP-1움/웁 were expressed in a generalized way in the grey matter of the entire spinal cord around 24 hr and again were restricted to the cellular inﬁltrate at the lesion site at 4 days following injury. The data suggest that resident CNS cells, most probably microglial cells, and not peripheral inﬂammatory cells, are the main source for cytokine and chemokine mRNA (481). The temporal mRNA expression patterns for TNF-움, IL-1웁, IL-6, M-CSF, and TGF-웁1 in two rat injury models with very different cellular inﬂammatory reactions have been assessws; in particular, contussion of the spinal cord and axotomy of the facial nerve were investigated (482). Comparative analyses using semiquantitative RT–PCR show an early and robust but transient up-regulation of IL-1웁, TNF-움, IL-6, and M-CSF mRNA in spinal cord after contusion injury. In contrast, expression of IL-1웁 and TNF-움 mRNA in the axotomized facial nucleus was minimal and delayed, and levels of M-CSF mRNA remained unaltered. Similar to injured spinal cord, the axotomized nucleus showed a dramatic and early up-regulation of IL-6 mRNA, but unlike spinal cord, IL-6 mRNA levels subsided only gradually. Both injury types showed gradually increasing levels of TGF-웁1 mRNA that were maximal at 7 days postinjury. RT–PCR analyses were also performed on isolated blood-borne mononuclear cells and neutrophils and showed that these cells contained high levels of IL-1웁 and M-CSF mRNA, moderate levels of TGF-웁1 and TNF-움 mRNA, and minimal levels of IL-6 mRNA. However, RT–PCR analyses together with histological observations indicated that expression of the proinﬂammatory cytokines IL-1웁, TNF-움, and IL-6 is short-lived and self-limited after contusion injury, and that it occurs primarily within endogenous glial cells.

476

MARCO ANTONIO CASSATELLA

Last, the time course of appearance of the TGF-웁 and its localization in developing endochondral bone was examined by Carrington and colleagues (18). These authors used the demineralized matrix-induced bone-forming system in rats. For the ﬁrst time, TGF-웁 was detected in developing endochondral bone in vivo. Intracellular immunohistochemical localization of TGF-웁 revealed that the cell types in which TGF-웁 could be detected varied with the time after implantation of the demineralized matrix: ﬁrst were inﬂammatory cells, and then cells in late hypertrophying and calcifying cartilage, i.e., the osteoblasts and, interestingly, also bone marrow granulocytes (18). Therefore, production of TGF-웁 by granulocytes may contribute to the regulation of ossiﬁcation during endochondral bone development. VIII. Concluding Remarks

The classic role attributed to neutrophils is still based on the obsolete view that PMNs are terminally differentiated, short-lived cells, with minimal (if any) transcriptional or translational activity. However, the studies summarized in this review clearly demonstrate the ability of neutrophils to synthesize and release various cytokines. Moreover, the fact that neutrophils clearly predominate over other cell types under various in vivo conditions suggests that, at least under some circumstances, the contribution of PMN-derived cytokines can be of foremost importance. In this respect, there already exists evidence suggesting that under some circumstances, the contribution of PMN-derived cytokines can be of foremost importance to the evolution of certain pathologies. Within the ﬁeld of neutrophil-derived cytokines, one of the facets that, in my opinion, urgently needs to be further elucidated is the identiﬁcation of all the stimuli that are able to induce cytokine synthesis in neutrophils. It has become clear that the interaction of PMNs with a given agonist produces a characteristic response, thus such studies might prove to be especially helpful in understanding the pathogenesis of diseases in which neutrophils represent (or are presumed to be) the ﬁrst cell type encountering, and interacting with, the etiologic agent. For example, the inﬂux of the different leukocyte populations to inﬂammatory lesions might very well reﬂect the individual chemokines being produced by neutrophils: IL-8 and GRO-움 predominantly recruit neutrophils, whereas MIP-1움/웁, IP-10, or MIG essentially recruit monocytes and lymphocyte subtypes. Thus, depending on the nature of the primary insult, and its effect toward the production of chemokines by neutrophils, the evolution of a given type of inﬂammatory reaction may be anticipated. Another very important aspect that, to date, has been mostly ignored by investigators working in the ﬁeld of cytokine production by neutrophils

NEUTROPHIL-DERIVED PROTEINS

477

is the potential involvement of transcription factors in the regulation of cytokine gene transcription. This partially reﬂects the fact that very little is known about the transcriptional events that control cytokine and chemokine gene expression in PMNs. However, the inhibitory effects of actinomycin D toward the inducible accumulation of cytokine mRNA in PMNs, and a limited amount of direct evidence (see Section VI), suggest that transcriptional events might play a central role in this process. Many of the regulatory elements located in the promoter regions of most cytokines, as well as the families of transcription factors that bind to them and control their transcription, have been well identiﬁed and characterized in various cell types. However, there can exist substantial differences in the exact pattern of transcription factor binding, depending on the cell type. Thus, performing such studies in PMNs could represent a step forward in our understanding of the cell-speciﬁc regulation of cytokine gene expression. Among other things, these studies could lead to the identiﬁcation of novel transcription factors, and eventually neutrophil-speciﬁc factors (483). Moreover, they could potentially elucidate the molecular bases of the many qualitative and quantitative differences observed between neutrophils and monocytes in terms of their ability to produce individual cytokines, such as IL-6, IL-12, IP-10, and so forth. Finally, the expression of neutrophil transcription factors could be the result of a regulated myeloid differentiation program, which in certain hematopoietic diseases or malignancies may be altered. In a broader context, studies addressing the mechanisms that regulate the intracellular distribution and release of cytokines are very scarce, both in neutrophils and other cell types. The need for such studies is best illustrated considering that they would advantageously complement the considerable knowledge already accumulated on cytokine gene and protein expression. In this regard, we recently found that neither STAT1 nor STAT3 is tyrosine phosphorylated in response to stimulation with IL10 in neutrophils (484), whereas in autologous PBMCs, we conﬁrmed that IL-10 rapidly triggers the phosphorylation of both proteins on tyrosine residues, as previously reported (485–488). Under appropriate stimulatory conditions, however, both STAT1 and STAT3 can undergo tyrosine phosphorylation in neutrophils (484, 489). The reasons for which IL-10 does not trigger tyrosine phosphorylation of STAT1 and STAT3 in neutrophils remain unknown, but our data make it likely that, in neutrophils, the activation of STAT1 and STAT3 tyrosine phosphorylation is not required for the modulatory effects of IL-10 toward cytokine production (4, 289). More importantly, our results also raise the possibility that the regulation of cytokine generation by IL-10 in other cell types might also occur independently of STAT protein activation.

478

MARCO ANTONIO CASSATELLA

An increasing body of evidence, mainly derived from in vitro studies, indicates that PMN survival can be greatly extended following exposure to microenvironmental signals, such as LPS, inactivated streptococci, IL1웁, TNF-움, IL-6, IFN-웂, G-CSF, and GM-CSF (202, 203). These observations raise the possibility that PMN viability in vivo may be considerably greater than what has been heretofore believed. If so, the ability of neutrophils to synthesize immunomodulatory cytokines could prove to be a phenomenon of considerable pathophysiological importance. In any case, it has become clear that PMNs should be considered not only as active and central elements of the inﬂammatory response, but also as cells that, through cytokine secretion, may signiﬁcantly inﬂuence the direction and evolution of the inﬂammatory and immune processes. In such a scenario, PMNs would play a pivotal role in regulatory interactions between innate resistance (mediated by phagocytic cells and NK cells) and adaptive immunity (mediated by T and B cells). In addition, recent studies have indicated that PMNs can synthesize and express MHC class II molecules on their surface (490, 491; our unpublished observations), and there is now even evidence showing that PMNs are capable of supporting T cell activation by bacterial superantigens (492). Furthermore, highly puriﬁed lactoferrinpositive precursors of end-stage PMNs cultured with the cytokine combination GM-CSF plus IL-4 and TNF-움 were shown to develop dendritic cell morphology, and to acquire features characteristic of dendritic cells, including potent T cell-stimulating activity in allogeneic, as well as autologous, mixed lymphocyte reactions (MLRs) (493). Surprisingly, these neutrophil-derived dendritic cells were found to be at least 10,000 times more efﬁcient in presenting soluble antigen to autologous T cells, relative to freshly isolated monocytes (493). This therefore raises the possibility that PMNs might have the ability to initiate a cellular immune response. This putative function of neutrophils would again imply an important role of these cells in many pathological conditions. Although relatively novel, research addressing cytokine production by neutrophils has brought forward new and exciting discoveries. In view of the variety of cytokines and chemokines secreted by neutrophils, it can be envisaged that PMNs can orchestrate the inﬁltration of leukocytes into sites of injury, and therefore determine the evolution of the host response. This being said, it is still premature to assess the true biological signiﬁcance of cytokine production by neutrophils. Even though our understanding of cytokine production by PMNs is far from complete, particularly in humans in vivo, its full appreciation is likely to yield valuable clues as to potential therapeutic approaches designed to control various disorders known to be inﬂuenced by PMNs. In vivo studies will ultimately be essential for critically

NEUTROPHIL-DERIVED PROTEINS

479

testing speciﬁc hypotheses about the biological impact of neutrophil cytokine production in health and disease. ACKNOWLEDGMENTS The author thanks Jose´ Lapinet Vera, Patrizia Scapini, and Elena Caveggion for critical reading of the manuscript. The author particularly thanks P. P. McDonald for many helpful suggestions and critical editing of most of this review. This work was supported by grants from MURST (40% and 60% funds, and coﬁnanziamento MURST-Universita´), AIRC, and Progetto Sanita`, Fondazione Cassa di Risparmio VR-VI-BL-AN.

REFERENCES 1. Edwards, S. (1994). ‘‘Biochemistry and Physiology of the Neutrophil.’’ Press Syndicate of the University of Cambridge, Cambridge, UK. 2. ‘‘Cytokines in Health and Disease,’’ 2nd Ed. (1993). Dekker, Inc., New York. 3. Lloyd, A. R., and Oppenheim, J. J. (1992). Immunol Today 13, 69–72. 4. Cassatella, M. A. (1995). The production of cytokines by polymorphonuclear neutrophils Immunol. Today 16, 21–26. 5. Cassatella, M. A. (1996). ‘‘Cytokines Produced by Polymorphonuclear Neutrophils: Molecular and Biological Aspects.’’ Springer, Landes Co., Berlin, Heidelberg, and New York. 6. Dubravec, D. B., Spriggs, D. R., Mannick, J. A., and Rodrick, M. L. (1990). Circulating human peripheral blood granulocytes synthesize and secrete tumor necrosis factor 움. Proc. Natl. Acad. Sci. U.S.A. 87, 6758–6761. 7. Djeu, J. Y., Serbousek, D., and Blanchard, D. K. (1990). Release of tumor necrosis factor by human polymorphonuclear leukocytes. Blood 76, 1405–1409. 8. Bazzoni, F., Cassatella, M. A., Laudanna C., and Rossi, F. (1991). Phagocytosis of opsonized yeast induces TNF움 mRNA accumulation and protein release by human polymorphonuclear leukocytes. J. Leukoc. Biol. 50, 223–228. 9. Lindemann, A., Riedel, D., Oster, W., Meuer, S. C., Blohm, D., Mertelsmann, R. H., and Herrmann, F. (1988). Granulocyte/macrophage colony-stimulating factor induces interleukin 1 production by human polymorphonuclear neutrophils. J. Immunol. 140, 837–883. 10. Marucha, P. T., Zeff, R. A., and Kreutzer, D. L. (1990). Cytokine regulation of IL-1 웁 gene expression in the human polymorphonuclear leukocyte. J. Immunol. 145, 2932– 2937. 11. Lord, P. C., Wilmoth, L. M., Mizel, S. B., and McCall, C. E. (1991). Expression of interleukin-1 움 and 웁 genes by human blood polymorphonuclear leukocytes. J. Clin. Invest. 87, 1312–1321. 12. Cassatella, M. A., Meda, L., Gasperini, S., D’Andrea, A. Ma, X., and Trinchieri, G. (1995). Interleukin-12 production by human polymorphonuclear leukocytes. Eur. J. Immunol. 25, 1–5. 13. Cassone, A., Chiani, P., Quinti, I., and Torosantucci, A. (1997). Possible participation of polymorphonuclear cells stimulated by microbial immunomodulators in the dysregulated cytokine patterns of AIDS patients. J. Leukoc. Biol. 62, 60–66. 14. McColl, S. R., Paquin, R., Menard, C., and Beaulieu, A. D. (1992). Human neutrophils produce high levels of the interleukin 1 receptor antagonist in response to granulocyte/ macrophage colony-stimulating factor and tumor necrosis factor 움. J. Exp. Med. 176, 593–598.

480

MARCO ANTONIO CASSATELLA

15. Ulich, T. R., Guo, K., Yin, S., del Castillo, J., Yi, E. S., Thompson, R. C., and Eisenberg, S. P. (1992). Endotoxin-induced cytokine gene expression in vivo. IV. Expression of interleukin-1움/웁 and interleukin-1 receptor antagonist mRNA during endotoxemia and during endotoxin-initiated local acute inﬂammation. Am. J. Pathol. 141, 61–68. 16. Grotendorst, G. R., Smale, G., and Pencev, D. (1989). Production of transforming growth factor 웁 by human peripheral blood monocytes and neutrophils. J. Cell. Phys. 140, 396–402. 17. Fava, R. A., Olsen, N. J., Postlethwaite, A. E., Broadley, K. N., Davidson, J. M., Nanney, L. B., Lucas, C., and Townes, A. S. (1991). Transforming growth factor 웁1 (TGF웁1) induced neutrophil recruitment to synovial tissues: Implications for TGF웁driven synovial inﬂammation and hyperplasia. J. Exp. Med. 173, 1121–1132. 18. Carrington, J. L., Roberts, A. B., Flanders, K. C., Roche, N. S., and Reddi, A. H. (1988). Accumulation, localization, and compartmentation of transforming growth factor 웁 during endochondral bone development. J. Cell Biol. 107, 1969–1975. 19. Bazzoni, F., Cassatella, M. A., Rossi, F., Ceska, M., Dewald, B., and Baggiolini, M. (1993). Phagocytosing neutrophils produce and release high amounts of the neutrophil activating peptide-1/interleukin 8. J. Exp. Med. 173, 771–774. 20. Strieter, R. M., Kasahara, K., Allen, R., Showell, H. J., Standiford, T. J., and Kunkel, S. L. (1990). Human neutrophils exhibit disparate chemotaotic gene expression. Biochem. Biophys. Res. Commun. 173, 725–730. 21. Gasperini, S., Calzetti, F., Russo, M. P., De Gironcoli, M., and Cassatella, M. A. (1995). Regulation of GRO움 production in human granulocytes. J. Inﬂamm. 45, 143–151. 22. Koch, A. E, Kunkel, S. L, Shah, M. R., Hosaka, S., Halloran, M. M., Haines, G. K., Burdick, M. D., Pope, R. M., and Strieter, R. M. (1995). Growth-related gene product 움. A chemotatcic cytokine for neutrophils in rheumatoid arthritis. J. Clin. Invest. 155, 3660–3666. 23. Kasama, T., Strieter, R. M., Standiford, T. J., Burdick, M. D., and Kunkel, S. L. (1993). Expression and regulation of human neutrophil-derived macrophage inﬂammatory protein 1-움. J. Exp. Med. 178, 63–72. 24. Kasama, T., Strieter, R. M., Lukacs, N. W., Burdick, M. D., and Kunkel, S. L. (1994). Regulation of neutrophil-derived chemokine expression by IL-10. J. Immunol. 152, 3559–3569. 25. Hachicha, M., Naccache, P. H., and McColl, S. R. (1995). Inﬂammatory mycrocrystal differentially regulate the secretion of macrophage inﬂammatory protein 1 and interleukin 8 by human neutrophils: A possible mechanism of neutrophil recruitment to sites of inﬂammation in synovitis. J. Exp. Med. 182, 2019–2025. 26. Wu, X., Dolecki, G. J., Sherry, B., Zagorski, J., and Lefkowith, J. B. (1997). Chemokines are expressed in a myeloid cell-dependent fashion and mediate distinct functions in immune complex glomerulonephritis in rat. J. Immunol. 158, 3917–3924. 27. Edamatsu, T., Xiao, Y. Q., Tanabe, J., Mue, S., and Ohuchi, K. (1997). Induction of neutrophil chemotactic factor production by staurosporine in rat peritoneal neutrophils. Br. J. Pharmacol. 121, 1651–1658. 28. Shirafuji, N., Matsuda, S., Ogura, H., Tani, K., Kodo, H., Ozawa, K., Nagata, S., Asano, S., and Takaku, F. (1990). Granulocyte colony-stimulating factor stimulates human mature neutrophilic granulocytes to produce interferon-움. Blood 75, 17–19. 29. Brandt, E. R., Linnane, A. W., and Devenish, R. J. (1994). Expression of IFN움 genes in subpopulations of peripheral blood cells. Br. J. Haematol. 86, 717–725. 30. Lindemann, A., Riedel, D., Oster, W., Ziegler-Heitbrock, H. W., Mertelsmann, R., and Herrmann, F. (1989). Granulocyte-macrophage colony-stimulating factor induces

NEUTROPHIL-DERIVED PROTEINS

31.

32.

33.

34.

35. 36. 37.

38.

39.

40.

41. 42.

43.

44.

45. 46.

481

cytokine secretion by human polymorphonuclear leukocytes. J. Clin. Invest. 83, 1308– 1312. Ichinose, Y., Hara, N., Ohta, M., Aso, H., Chikama, H., Kawasaki, M., Kubota, I., Shimizu, T., and Yagawa, K. (1990). Recombinant granulocyte colony-stimulating factor and lipopolysaccharide maintain the phenotype of and superoxide anion generation by neutrophils. Infect. Immun. 58, 1647–1652. Liles, W. C., Kiener, P. A, Ledbetter, J. A, Aruffo, A., and Klebanoff, S. J. (1996). Differential expression of Fas (CD95) and Fas ligand on normal human phagocytes: Implications for the regulation of apoptosis in neutrophils. J. Exp. Med. 184, 429–440. Hsieh, S. C., Huang, M. H., Tsai, C. Y., Tsai, Y. Y., Tsai, S. T., Sun, K. H., Yu, H. S., Han, S. H., and Yu, C. L. (1997). The expression of genes modulating programmed cell death in normal human polymorphonuclear neutrophils. Biochem. Biophys. Res. Commun. 233, 700–706. Gruss, H. J., DaSilva, N., Hu, Z. B., Uphoff, C. C., Goodwin, R. G., and Drexler, H. G. (1994). Expression and regulation of CD30 ligand and CD30 in human leukemia– lymphoma cell lines. Leukemia 8, 2083–2094. Wiley, S. R., Goodwin, R. G., and Smith, C. A. (1996). Reverse signaling via CD30 ligand. J. Immunol. 157, 3635–3639. Taichman, N. S., Young, S., Cruchley, A. T., Taylor, P., and Paleolog, E. (1997). Human neutrophils secrete vascular endothelial growth factor. J. Leukoc. Biol. 62, 397–400. Gaudry, M., Bregerie, O., Andrieu, V., El Benna, J., Pocidalo, M. A., and Hakim, J. (1997). Intracellular pool of vascular endothelial growth factor in human neutrophils. Blood 90, 4153–4161. Webb, N. J., Myers, C. R., Watson, C. J., Bottomley, M. J., and Brenchley, P. E. (1998). Activated human neutrophils express vascular endothelial growth factor. Cytokine 10, 254–257. Wolf, H. K., Zarnegar, R., and Michalopoulos, G. K. (1991). Localization of hepatocyte growth factor in human and rat tissues: An immunohistochemical study. Hepatology 14, 488–494. Sakaguchi, H., Seki, S., Tsubouchi, H., Daikuhara, Y., Niitani, Y., and Kobayashi, K. (1994). Ultrastructural location of human hepatocyte growth factor in human liver. Hepatology 19, 1157–1163. Cassatella, M. A., Gasperini, S., and Russo, M. P. (1997). Cytokine expression and release by neutrophils. Ann. N.Y. Acad. Sci. 832, 233–242. Kuijpers, T. W., Tool, A. T., van der Schoot, C. E., Ginsel, L. A., Onderwater, J. J., Roos, D., and Verhoeven, A. J. (1991). Membrane surface antigen expression on neutrophils: A reappraisal of the use of surface markers for neutrophil activation. Blood 78, 1105–1111. Venaille, T. J., Misso, N. L., Phillips, M. J., Robinson, B. W., and Thompson, P. J. (1994). Effects of different density gradient separation techniques on neutrophil function. Scand. J. Clin. Lab. Invest. 54, 385–391. Haslett, C., Guthrie, L. A., Kopaniak, M. M., Johnston, R. B., Jr., and Henson, P. M. (1985). Modulation of multiple neutrophil functions by preparative methods or trace concentrations of bacterial lipopolysaccharide. Am. J. Pathol. 119, 101–110. Altstaedt, J., Kirchner, H., and Rink, L. (1996). Cytokine production of neutrophils is limited to interleukin-8. Immunology 89, 563–568. Stibenz, D., and Bu¨hrer, C. (1994). Down-regulation of L-selectin surface expression by various leukocyte isolation procedures (corrected and republished in with original paging). Scand. J. Immunol. 39, 59–63.

482

MARCO ANTONIO CASSATELLA

47. Morrison, D. C., and Jacobs, D. M. (1976). Binding of polymyxin B to the lipid A portion of bacterial lipopolysaccharides. Immunochemistry 13, 813–818. 48. Re, F., Mengozzi, M., Muzio, M., Dinarello, C. A., Mantovani, A., and Colotta, F. (1993). Expression of interleukin-1 receptor antagonist (IL-lra) by human circulating polymorphonuclear cells. Eur. J. Immunol. 23, 570–573. 49. Muzio, M., Re, F., Sironi, M., Polentarutti, N., Minty, A., Caput, D., Ferrara, P., Mantovani, A., and Colotta, F. (1994). Interleukin-13 induces the production of interleukin-1 receptor antagonist (IL-lra) and the expression of the mRNA for the intracellular (keratinocyte) form of IL-lra in human myelomonocytic cells. Blood 83, 1738–1743. 50. Reglier, H., Arce-Vicioso, M., Fay, M., Gougerot-Pocidalo, M. A., and Chollet-Martin, S. (1998). Lack of IL-10 and IL-13 production by human polymorphonuclear neutrophils. Cytokine 10, 192–198. 51. Wang, P., Wu, P., Anthes, J. C., Siegel, M. I., Egan, R. W., and Billah, M. M. (1994). Interleukin-10 inhibits Interleukin-8 production in human neutrophils. Blood 83, 2678–2683. 52. Takeichi, O., Saito, I., Tsurumachi, T., Saito, T., and Moro, I. (1994). Human polymorphonuclear leukocytes derived from chronically inﬂamed tissue express inﬂammatory cytokines in vivo. Cell Immunol. 156, 296–309. 53. Romagnani, S. (1994). Lymphokine production by human T cells in disease state. Annu. Rev. Immunol. 12, 227–257. 54. Rollins, B. J. (1997). Chemokines. Blood 90, 909–928. 55. Luster, A. D. (1998). Chemokines: chemotactic cytokines that mediate inﬂammation. N. Engl. J. Med. 338, 436–445. 56. Baggiolini, M. (1998). Chemokines and leukocyte trafﬁc. Nature (London) 392, 565–568. 57. Oppenheim, J. J., Zacharie, C. O. C., Mukaida, N., and Matsushima, K. (1991). Properties of the novel proinﬂammatory supergene ‘‘intercrine’’ cytokine family. Annu. Rev. Immunol. 9, 617–648. 58. Horuk, R. (1998). Chemokines beyond inﬂammation. Nature (London) 393, 524–525. 59. Koch, A. E., Polverini, P. J., Kunkel, S. L., Harlow, L. A., DiPietro, L. A., Elner, V. M., Elner, S. G., and Strieter, R. M. (1992). Interleukin-8 as a macrophage-derived mediator of angiogenesis. Science 258, 1798–1801. 60. Cassatella, M. A., Bazzoni, F., Ceska, M., Ferro, I., Baggiolini, M., and Berton, G. (1992). Interleukin 8 production by human polymorphonuclear leukocytes. The chemoattractant formyl-methionyl-leucyl-phenylalanine induces the gene expression and release of interleukin 8 through a pertussis toxin sensitive pathway. J. Immunol. 148, 3216–3220. 61. Arnold, R., Werner, F., Humbert, B., Werchau, H., and Konig, W. (1994). Effect of respiratory syncytial virus–antibody complexes on cytokine (IL-8, IL-6, TNF움) release and respiratory burst in human granulocytes. Immunology 82, 184–191. 62. Morrison, T. B., Weis, J. H., and Weis, J. J. (1997). Borrelia burgdorferi outer surface protein A (OspA) activates and primes human neutrophils. J. Immunol. 158, 4838– 4845. 63. Kuhns, D., and Gallin, J. I. (1995). Increased cell-associated IL-8 in human exudative and A23187-treated peripheral blood neutrophils. J. Immunol. 154, 6556–6562. 64. Brizzi, M. F., Battaglia, E., Rosso, A., Strippoli, P., Montrucchio, G., Camussi, G., and Pegoraro, L. (1997). Regulation of polymorphonuclear cell activation by thrombopoietin. J. Clin. Invest. 99, 1576–1584. 65. Ember, J. A., Sanderson, S. D., Hugli, T. E., and Morgan, E. L. (1994). Induction of IL8 synthesis from monocytes by human C5a anaphylotoxin. Am. J. Pathol. 144, 393–403.

NEUTROPHIL-DERIVED PROTEINS

483

66. McCain, R. W., Holden, E. P., Blackwell, T. R., and Christman, J. W. (1994). Leukotriene B4 stimulates human polymorphonuclear leukocytes to synthesize and release interleukin-8 in vitro. Am. Respir. Cell. Mol. Biol. 10, 651–657. 67. Bentwood, B. J., and Henson, P. M. (1980). The sequential release of granule constitutents from human neutrophils. J. Immunol. 124, 855–862. 68. Strieter, R. M., Kasahara, K., Allen, R. M., Standiford, T. J., Rolfe, M. W., DeForge, L. E., Becker, F. S., Chensue, S. W., and Kunkel, S. L. (1992). Cytokine-induced neutrophil-derived interleukin-8. Am. J. Pathol. 141, 397–407. 69. Au, B., Williams, T. J., and Collins, P. D. (1994). Zymosam-induced IL-8 release from human neutrophils involves activation via the CD11b/CD18 receptor and endogenous platelet-activating factor as an autocrine modulator. J. Immunol. 152, 5411–5419. 70. Kikuchi, M., Endo, S., Inada, K., Yamashita, H., Takakuwa, T., Nakae, H., Kasai, T., Baba, N., and Yamada, Y. (1995). Inhibitory effect of FUT-175 on the production of interleukin 8 and polymorphonuclear-leukocyte elastase. Res. Commun. Mol. Pathol. Pharmacol. 87, 269–274. 71. Galley, H. F., Dubbels, A. M., and Webster, N. R. (1998). The effect of midazolam and propofol on interleukin-8 from human polymorphonuclear leukocytes. Anesth. Analg. 86, 1289–1293. 72. Corriveau, C. C., Madara, P. J., Van Dervort, A. L., Tropea, M. M., Wesley, R. A., and Danner, R. L. (1998). Effects of nitric oxide on chemotaxis and endotoxin-induced interleukin-8 production in human neutrophils. J. Infect. Dis. 177, 116–126. 73. Pakianathan, D. R. (1995). Extracellular matrix proteins and leukocyte. J. Leukoc. Biol. 57, 699–702. 74. Derevianko, A., D’Amico, R., and Simms, H. (1996). Polymorphonuclear leucocyte (PMN)-derived inﬂammatory cytokines-regulation by oxygen tension and extracellular matrix. Clin. Exp. Immunol. 106, 560–567. 75. Yoshimura, A., Hara, Y., Kaneko, T., and Kato, I. (1997). Secretion of IL-1웁, TNF움, IL-8 and IL-lra by human polymorphonuclear leukocytes in response to lipopolysaccharides from periodontopathic bacteria. J. Periodont. Res. 32, 279–286. 76. Sugita, N., Kimura, A., Matsuki, Y., Yamamoto, T., Yoshie, H., and Hara, K. (1998). Activation of transcription factors and IL-8 expression in neutrophils stimulated with lipopolysaccharide from Porphyromonas gingivalis. Inﬂammation 22, 253–267. 77. Cassatella, M. A., Guasparri, I., Ceska, M., Bazzoni, F., and Rossi, F. (1993). Interferon웂 inhibits interleukin-8 production by human polymorphonuclear leukocytes. Immunology 78, 177–184. 78. Fujishima, S., Hoffman, A. R., Vu, T., Kim, J., Zheng, H., Daniel, D., Kim, Y., Wallace, E. F., Larrick, J. W., and Rafﬁn, T. A. (1993). Regulation of neutrophil interleukin 8 gene expression and protein secretion by LPS, TNF움 and IL-1웁. J. Cell. Physiol. 154, 478–485. 79. Cassatella, M. A., Meda, L., Bonora, S., Ceska, M., and Constantin, G. (1993). Interleukin 10 inhibits the release of proinﬂammatory cytokines from human polymorphonuclear leukocytes. Evidence for an autocrine role of TNF움 and IL-1웁 in mediating the production of IL-8 triggered by lipopolysaccharide. J. Exp. Med. 178, 2207–2211. 80. Kasahara, K., Sato, I., Ogura, K., Takeuchi, H., Kobayashi, K., and Adachi, M. (1998). Expression of chemokines and induction of rapid cell death in human blood neutrophils by Mycobacterium tuberculosis. J. Infect. Dis. 178, 127–137. 81. McCain, R. W., Dessypris, E. N., and Christman, J. W. (1993). GM-CSF stimulates human polymorphonuclear leukocytes to produce interleukin-8 in vitro. Am. J. Respir. Cell. Mol. Biol. 8, 28–34.

484

MARCO ANTONIO CASSATELLA

82. Takahashi, G. W., Andrews, D. F., Lilly, M. B., Singer, J. W., and Alderson, M. R. (1993). Effect of GM-CSF and IL-3 on IL-8 production by human neutrophils and monocytes. Blood 81, 357–364. 83. Wei, S., Liu, J. H., Blanchard, D. K., and Djeu, J. Y. (1994). Induction of IL-8 gene expression in human polymorphonuclear neutrophils by recombinant IL-2. J. Immunol. 152, 3630–3636. 84. Hilger, R. A., Koller, M., and Konig, W. (1995). Immunosuppressive agents enhance the cytokine-induced priming of inﬂammatory cells. Int. Arch. Allergy Immunol. 107, 259–260. 85. Baldwin, G. C., Buckley, D. M., Roth, M. D., Kleerup, E. C., and Tashkin, D. P. (1997). Acute activation of circulating polymorphonuclear neutrophils following in vivo administration of cocaine. A potential etiology for pulmonary injury. Chest 111, 698–705. 86. Musso, T., Calosso, L., Zucca, M., Millesimo, M., Puliti, M., Bulfone-Paus, S., Merlino, C., Savoia, D., Cavallo, R., Ponzi, A. N., and Badolato, R. (1998). Interleukin-15 activates proinﬂammatory and antimicrobial functions in polymorphonuclear cells. Infect. Immun. 66, 2640–2647. 87. Girard, D., Paquin, R., Naccache, P. H., and Beaulieu, A. D. (1996). Effects of interleukin-13 on human neutrophil functions. J. Leukoc. Biol. 59, 412–419. 88. Marie, C., Pitton, C., Fitting, C., and Cavaillon, J. M. (1996). Regulation by antiinﬂammatory cytokines (IL-4, IL-10, IL-13 and TGF웁) of IL-8 production by LPS- and or TNF움-activated human polymorphonuclear cells. Mediators Inﬂamm. 5, 334–340. 89. Girard, D., Paquet, M. E., Paquin, R., and Beaulieu, A. D. (1996). Differential effects of interleukin-15 (IL-15) and IL-2 on human neutrophils: Modulation of phagocytosis, cytoskeleton rearrangement, gene expression, and apoptosis by IL-15. Blood 88, 3176– 3184. 90. McDonald, P. P., Russo, M. P., Ferrini, S., and Cassatella, M. A. (1998). Interleukin15 activates. Blood 92, 4828–4835. 91. Badolato, R., Ponzi, A. N., Millesimo, M., Notarangelo, L. D., and Musso, T. (1997). Interleukin-15 (IL-15) induces IL-8 and monocyte chemotactic protein 1 production in human monocytes. Blood 90, 2804–2809. 92. Waldmann, T., Tagaya, Y., and Bamford, R. (1998). Interleukin-2, interleukin-15, and their receptors. Int. Rev. Immunol. 16, 205–226. 93. Cassatella, M. A., Aste, M., Calzetti, F., Constantin, G., Guasparri, I., Ceska, M., and Rossi, F. (1993). Studies on the regulatory mechanisms of interleukin-8 gene expression in resting and IFN웂-treated neutrophils. Evidence on the capability of staurosporine of inducing the production of IL-8 by human neutrophils. Biochem. Biophys. Res. Commun. 190, 660–667 (published erratum in 1993, 192 324). 94. Kuhns, D., Young, H. A., Gallin, E. K., and Gallin, J. I. (1998). J. Immunol. 161, 4332– 4339. 95. Sisson, J. H., Prescott, S. M., McIntyre, T. M., and Zimmerman, G. A. (1987). Production of platelet-activating factor by stimulated human polymorphonuclear leukocytes. Correlation of synthesis with release, functional events, and leukotriene B4 metabolism. J. Immunol. 138, 3918-3926. 96. Hilger, R. A., Koller, M., and Konig, W. (1996). Inhibition of leukotriene formation and IL-8 release by the PAF-receptor antagonist SM-12502. Inﬂammation 20, 57–70. 97. Serra, M. C., Calzetti, F., Ceska, M., and Cassatella, M. A. (1994). Effect of substance P on superoxide anion and IL-8 production by human PMN. Immunology 82, 63–69. 98. Shinoda, I., Takase, M., Fukuwatari, Y., Shimamura, S., Koller, M., and Konig, W. (1996). Effects of lactoferrin and lactoferricin on the release of interleukin 8 from human polymorphonuclear leukocytes. Biosci. Biotechnol. Biochem. 60, 521–523.

NEUTROPHIL-DERIVED PROTEINS

485

99. Laudanna, C., Constantin, G., Baron, P. L., Scarpini, E., Scarlato, G., Cabrini, G., Dechecchi, C., Rossi, F., Cassatella, M. A., and Berton, G. (1994). Sulfatides trigger increase of cytosolic free calcium and enhanced expression of tumor necrosis factor움 and interleukin-8 mRNA in human neutrophils. Evidence for a role of L selectin as a signaling molecule. J. Biol. Chem. 269, 4021–4026. 100. Hidari, K. I., Weyrich, A. S., Zimmerman, G. A., and McEver, R. P. (1997). Engagement of PSGL-1 enhances tyrosine phosphorylation and activates MAP kinases in human neutrophils. J. Biol. Chem. 272, 28750–28756. 101. Kessel, J. M., Hayﬂick, J., Weyrich, A. S., Hoffman, P. A., Gallatin, M., McIntyre, T. M., Prescott, S. M., and Zimmerman, G. A. (1998). Coengagement of ICAM-3 and Fc receptors induces chemokine secretion and spreading by myeloid leukocytes. J. Immunol. 160, 5579–5587. 102. Torosantucci, A., Chiani, P., Quinti, I., Ausiello, C. M., Mezzaroma, I., and Cassone, A. (1997). Responsiveness of human polymorphonuclear cells (PMNL) to stimulation by a mannoprotein fraction (MP-F2) of Candida albicans; enhanced production of IL-6 and tumour necrosis factor 움 (TNF-움) by MP-F2-stimulated PMNL from HIVinfected subjects. Clin. Exp. Immunol. 107, 451–457. 103. Cassatella, M. A., Bazzoni, F., D’Andrea, A., Tronchin, M., Ceska, M., De Gironcoli, M., and Rossi, F. (1993). Studies on the production of proinﬂammatory cytokines and on the modulation of gene expression for some NADPH oxidase components by phagocytosing human neutrophils. Fund. Clin. Immunol. 1, 99–106. 104. Au, B. T., Teixeira, M. M., Collins, P. D., and Williams, T. J. (1998). Effect of PDE4inhibitors on zymosaninduced IL-8 release from human neutrophils. Synergism with prostanoids and salbutamol. Br. J. Pharmacol. 123, 1260–1266. 105. Wertheim, W. A., Kunkel, S. L., Standiford, T. J., Burdick, M. D., Becker, F. S., Wilke, C. A., Gilbert, A. R., and Strieter, R. M. (1993). Regulation of neutrophilderived IL-8: the role of prostaglandin E2, dexamethasone, and IL-4. J. Immunol. 151, 2166–2175. 106. Arnold, R., and Konig, W. (1998). Interleukin-8 release from human neutrophils after phagocytosis of Listeria monocytogenes and Yersinia enterocolitica. J. Med. Microbiol. 47, 55–62. 107. Retini, C., Vecchiarelli, A., Monari, C., Tascini, C., Bistoni, F., and Kozel, T. R. (1996). Capsular polysaccharide of Cryptococcus neoformans induces proinﬂammatory cytokine release by human neutrophils. Infect. Immun. 64, 2897–2903. 108. Marie, C., Muret, J., Fitting, C., Losser, M. R., Payen, D., and Cavaillon, J. M. (1998). Reduced ex vivo interleukin-8 production by neutrophils in septic and nonseptic systemic inﬂammatory response syndrome. Blood 91, 3439–3446. 109. Wahlgren, M., Abrams, J. S., Fernandez, V., Bejarano, M. T., Azuma, M., Torii, M., Aikawa, M., and Howard, R. J. (1995). Adhesion of Plasmodium falciparum-infectederythrocytes to human cells and secretion of cytokines (IL-1-웁, IL-1RA, IL-6, IL-8, IL-10, TGF웁, TNF움, G-CSF, GM-CSF). Scand. J. Immunol. 42, 626–636. 110. Konig, B., Koller, M., Prevost, G., Piemont, Y., Alouf, J. E., Schreiner, A., and Konig, W. (1994). Activation of human effector cells by different bacterial toxins (leukocidin, alveolysin, and erythrogenic toxin A). Generation of interleukin & Infect. Immun. 62, 4831–4837. 111. Gibson, F. C., Tzianabos, A. O., and Onderdonk, A. B. (1996). The capsular polysaccharide complex of Bacteroides fragilis induces cytokine production from human and murine phagocytic cells. Infect. Immun. 64, 1065–1069. 112. Inoue, H., Massion, P., Ueki, I. F., Grattan, K. M., Hara, M., Dohrman Chan, B., Lausier, J. A., Golden, J. A., and Nadel, J. A. (1994). Pseudomonas stimulates interleu-

486

113. 114. 115.

116.

117. 118. 119. 120. 121. 122. 123. 124. 125.

126.

127.

MARCO ANTONIO CASSATELLA

kin-8 mRNA expression selectively in airway epithelium, in gland ducts, and in recruited neutrophils. Am. J. Respir. Cell. Mol. Biol. 11, 651–663. Konig, B., Ceska, M., and Konig, W. (1995). Effect of Pseudomonas aeruginosa on interlerukin-8 release from human phagocytes. Int. Arch. Allergy. Immunol. 106, 357–365. Hachicha, M., Rathanaswami, P., Naccache, P. H., and McColl, S. R. (1998). Regulation of chemokine gene expression in human peripheral blood neutrophils phagocytosing microbial pathogens. J. Immunol. 160, 449–454. Oishi, K., Sonoda, F., Kobayashi, S., Iwagaki, A., Nagatake, T., Matsushima, K., and Matsumoto, K. (1994). Role of interleukin-8 (IL-8) and an inhibitory effect of erythromycin on IL-8 release in the airways of patients with chronic airway diseases. Infect. Immun. 62, 4145–4152. Yoshimine, H., Oishi, K., Nagatake, T., Matsushima, K., and Omura, T. (1998). Inhibition of neutrophil-derived IL-8 production caused by EM (erythromycin) 201 derivative: Modiﬁcation of the activities by the 5-position desosamine side chain. Jpn. J. Antibiot. 51 (Suppl.), 134–136. Riedel, D. D., and Kaufmann, S. H. (1997). Chemokine secretion by human polymorphonuclear granulocytes after stimulation with Mycobacterium tuberculosis and lipoarabinomannan. Infect. Immun. 65, 4620–4623. Konig, B., Krusat, T., Streckert, H. J., and Konig, W. (1996). IL-8 release from human neutrophils by the respiratory syncytial virus is independent of viral replication. J. Leukoc. Biol. 60, 253–260. McColl, S. R., Roberge, C. J., Larochelle, B., and Gosselin, J. (1997). EBV induces the production and release of IL-8 and macrophage inﬂammatory protein-1움 in human neutrophils. J. Immunol. 159, 6164–6168. Roberge, C. J., McColl, S. R., Larochelle, B., and Gosselin, J. (1998). Granulocytemacrophage colony-stimulating factor enhances EBV-induced synthesis of chemotaotic factors in human neutrophils. J. Immunol. 160, 2442–2448. Roberge, C. J., Poubelle, P. E., Beaulieu, A. D., Heitz, D., and Gosselin, J. (1996). The IL-1 and IL-1 receptor antagonist (IL-1Ra) response of human neutrophils to EBV stimulation. Preponderance of IL-Ra detection. J. Immunol. 156, 4884–4891. Kasahara, K., Strieter, R. M., Chensue, S. W., Standiford, T. J., and Kunkel, S. L. (1991). Mononuclear cell adherence induces neutrophil chemotactic factor/interleukin8 gene expression. J. Leukoc. Biol. 50, 287–295. Haskill, S., Peace, A., Morris, J., Sporn, S. A., Anisowicz, A., Lee, S. W., Smith, T., Martin, G., Ralph, P., and Sager, R. (1990). Identiﬁcation of three related human GRO genes encoding cytokine functions. Proc. Natl. Acad. Sci. U.S.A. 87, 7732–7736. Meda, L., Gasperini, S., Ceska, M., and Cassatella, M. A. (1994). Modulation of proinﬂammatory cytokine release from human polymorphonuclear leukocytes by 웂 interferon. Cell. Immunol. 57, 448–461. Luan, J., Shattuck-Brandt, R., Haghnegahdar, H., Owen, J. D., Strieter, R., Burdick, M., Nirodi, C., Beauchamp, D., Johnson, K. N., and Richmond, A. (1997). Mechanism and biological signiﬁcance of constitutive expression of MGSA/GRO chemokines in malignant melanoma tumor progression. J. Leukoc. Biol. 62, 588–597. Villard, J., Dayer-Pastore, F., Hamacher, J., Aubert, J. D., Schlegel-Haueter, S., and Nicod, L. P. (1995). GRO움 and interleukin-8 in Pneumocystis carinii or bacterial pneumonia and adult respiratory distress syndrome. Am. J. Respir. Crit. Care. Med. 152, 1549–1554. Iida, N., and Grotendorst, G. R. (1990). Cloning and sequencing of a new gro transcript from activated human monocytes: Expression in leukocytes and wound tissue. Mol. Cell. Biol. 10, 5596–5599.

NEUTROPHIL-DERIVED PROTEINS

487

128. Nakagawa, H., Komorita, N., Shibata, F., Ikesue, A., Konishi, K., Fujioka, M., and Kato, H. (1994). Identiﬁcation of cytokine-induced neutrophil chemoattractants (CINC), rat GRO/CINC-2움 and CINC-2웁, produced by granulation tissue in culture: Puriﬁcation, complete amino acid sequences and characterization. Biochem. J. 301, 545–550. 129. al-Mokdad, M., Shibata, F., Takano, K., and Nakagawa, H. (1998). Differential production of chemokines by phagocytosing rat neutrophils and macrophages. Inﬂammation 22, 145–159. 130. Cole, K. E., Strick, C. A., Paradis, T. J., Ogborne, K. T., Loetscher, M., Gladue, R. P., Lin, W., Boyd, J. G., Moser, B., Wood, D. E., Sahagan, B. G., and Neote, K. (1998). Interferon-inducible T cell 움 chemoattractant (I-TAC): A novel non-ELR CXC chemokine with potent activity on activated T cells through selective high afﬁnity binding to CXCR3. J. Exp. Med. 187, 2009–2021. 131. Farber, J. M. (1997). Mig and IP-10: CXC chemokines that target lymphocytes. J. Leukoc. Biol. 61, 246–257. 132. Taub, D. D., Sayers, T. J., Carter, C. R., and Ortaldo, J. R. (1995). 움 and 웁 chemokines induce NK cell migration and enhance NK-mediated cytolysis. J. Immunol. 155, 3877– 3888. 133. Loetscher, M., Gerber, B., Loetscher, P., Jones, S. A., Piali, L., Clark-Lewis, I., Baggiolini, M., and Moser, B. (1996). Chemokine receptor speciﬁc for IP10 and mig: Structure, function, and expression in activated T-lymphocytes. J. Exp. Med. 184, 963–1154. 134. Qin, S., Rottman, J. B., Myers, P., Kassam, N., Weinblatt, M., Loetscher, M., Koch, A. E., Moser, B., and Mackay, C. R. (1998). The chemokine receptors CXCR3 and CCR5 mark subsets of T cells associated with certain inﬂammatory reactions. J. Clin. Invest. 101, 746–754. 135. Bonecchi, R., Bianchi, G., Bordignon, P. P., D’Ambrosio, D., Lang, R., Borsatti, A., Sozzani, S., Allavena, P., Gray, P. A., Mantovani, A., and Sinigaglia, F. (1998). Differential expression of chemokine receptors and chemotactic responsiveness of type 1 T helper cells (Th1s) and Th2s. J. Exp. Med. 187, 129–134. 136. Sallusto, F., Lenig D., Mackay, C. R., and Lanzavecchia, A. (1998). Flexible programs of chemokine receptor expression on human polarized T helper 1 and 2 lymphocytes. J. Exp. Med. 187, 875–883. 137. Loetscher, P., Uguccioni, M., Bordoli, L., Baggiolini, M., Moser, B., Chizzolini, C., and Dayer, J. M. (1998). CCR5 is characteristic of Th1 lymphocytes. Nature (London) 391, 344–345. 138. Sarris, A. H., Broxmeyer, H. E., Wirthmueller, U., Karasavvas, N., Cooper, S., Lu, L., Krueger, J., and Ravetch, J. V. (1993). Human interferon-inducible protein 10: Expression and puriﬁcation of recombinant protein demonstrate inhibition of early human hematopoietic progenitors. J. Exp. Med. 178, 1127–1132. 139. Schwartz, G. N., Liao, F., Gress, R. E., and Farber, J. M. (1997). Suppressive effects of recombinant human monokine induced by IFN-웂 (rHuMig) chemokine on the number of committed and primitive hemopoietic progenitors in liquid cultures of CD34⫹ human bone marrow cells. J. Immunol. 159, 895–904. 140. Arenberg, D. A., Kunkel, S. L., Polverini, P. J., Morris, S. B., Burdick, M. D., Glass, M. C., Taub, D. T., Iannettoni, M. D., Whyte, R. I., and Strieter, R. M. (1996). Interferon-웂-inducible protein 10 (IP-10) is an angiostatic factor that inhibits human non-small cell lung cancer (NSCLC) tumorigenesis and spontaneous metastases. J. Exp. Med. 184, 981–992. 141. Sgadari, C., Farber, J. M., Angiolillo, A. L., Liao, F., Teruya-Feldstein, J., Burd, P. R., Yao, L., Gupta, G., Kanegane, C., and Tosato, G. (1997). Mig, the monokine induced by interferon-웂, promotes tumor necrosis in vivo. Blood 89, 2635–2643.

488

MARCO ANTONIO CASSATELLA

142. Angiolillo, A. L., Sgadari, C., Sheikh, N., Reaman, G. H., and Tosato, G. (1995). Regression of experimental human leukemias and solid tumors induced by Epstein– Barr virus immortalized B cells. Leuk. Lymphoma 19, 267–276. 143. Luster, A. D., and Leder, P. (1993). IP-10, a -C-X-C- chemokine, elicits a potent thymus-dependent antitumor response in vivo. J. Exp. Med. 178, 1057–1065. 144. Tannenbaum, C. S., Tubbs, R., Armstrong, D., Finke, J. H., Bukowski, R. M., and Hamilton, T. A. (1998). The CXC chemokines IP-10 and Mig are necessary for IL12-mediated regression of the mouse RENCA tumor. J. Immunol. 162, 927–932. 145. Coughlin, C. M., Salhany, K. E., Gee, M. S., LaTemple, D. C., Kotenko, S., Ma, X. J., Gri, G., Wysocka, M., Kim, J. E., Liu, L., Liao, F., Farber, J. M., Pestka, S., Trinchieri, G., and Lee, W. M. F. (1998). Tumor cell responses to IFN affect tumorigenicity and response to IL-12 therapy and antiangiogenesis. Immunity 9, 25–34. 146. Cassatella, M. A., Gasperini, S., Calzetti, F., Bertagnin, A., Luster, A. D., and McDonald, P. P. (1997). Regulated production of the interferon-웂-inducible protein-10 (IP-10) chemokine by human neutrophils. Eur. J. Immunol, 27. 111–115. 147. Gasperini, S., Marchi, M., Calzetti, F., Vicentini, L., Olsen, H., Murphy, M., Liao, F., Farber, J., and Cassatella, M. A. (1999). Gene expression and production of the MIG (monokine induced by IFN웂), I-TAC (IFN-inducible T cell 움 chemoattractant) and IP-10 (IFN웂 inducible protein-10) chemokines by human neutrophils. J. Immunol. 162, in press. 148. Farber, J. M. (1993). HuMig: A new human member of the chemokine family of cytokines. Biochem. Biophys. Res. Commun. 192, 223–230. 149. Maggi, E. (1998). The TH1/TH2 paradigm in allergy. Immunotechnology 3, 233–244. 150. Colombo, M. P., Modesti, A., Parmiani, G., and Forni, G. (1992). Local cytokine availability elicits tumor rejection and systemic immunity through granulocyte–Tlymphocyte cross-talk. Cancer Res. 52, 4853–4857. 151. Stoppacciaro, A., Melani, C., Parenza, M., Mastracchio, A., Bassi, C., Baroni, C., Parmiani, G., and Colombo, M. P. (1993). Regression of an established tumor genetically modiﬁed to release granulocyte-colony stimulating factor requires granulocyte–T cell cooperation and T cell-produced interferon 웂. J. Exp. Med. 178, 151–161. 152. Cavallo, F., Signorelli, P., Giovarelli, M., Musiani, P., Modesti, A., Brunda, M. J., Colombo, M. P., and Forni, G. (1997). Antitumor-efﬁcacy of adenocarcinoma cells engineered to produce interleukin 12 (IL-12) or other cytokines compared with exogenous IL-12. J. Natl. Cancer Inst. 89, 1049–1058. 153. Musiani, P., Allione, A., Modica, A., Lollini, P. L., Giovarelli, M., Cavallo, F., Belardelli, F., Forni, G., and Modesti, A. (1996). Role of neutrophils and lymphocytes in inhibition of a mouse mammary adenocarcinoma engineered to release IL-2, IL-4, IL-7, IL-10, IFN-움, IFN-애, and TNF-움 Lab. Invest. 74, 146–157. 154. Mantovani, A. (1997). The interplay between primary and secondary cytokines Cytokines. involved in the regulation of monocyte recruitment. Drugs 54 (Suppl. 1), 15–23. 155. Taub, D. D., Conlon, K., Lloyd, A. R., Oppenheim, J. J., and Kelvin, D. J. (1993). Preferential migration of activated CD4⫹ and CD8⫹ T cells in response to MIP-1움 and MIP-1웁. Science 260, 355–358. 156. Van Damme, J., Proost, P., Lenaerts, J. P., and Opdenakker, G. (1992). Structural and functional identiﬁcation of two human, tumor-derived monocyte chemotactic proteins (MCP-2 and MCP-3) belonging to the chemokine family. J. Exp. Med. 176, 59–65. 157. Van Damme, J., Proost, P., Put, W., Arens, S., Lenaerts, J. P., Conings, R., Opdenakker, G., Heremans, H., and Billiau, A. (1994). Induction of monocyte chemotactic proteins MCP-1 and MCP-2 in human ﬁbroblasts and leukocytes by cytokines and cytokine

NEUTROPHIL-DERIVED PROTEINS

158. 159. 160. 161. 162.

163. 164. 165. 166. 167. 168. 169. 170.

171. 172.

173. 174.

489

inducers. Chemicals synthesis of MCP-2 and development of a speciﬁc RIA. J. Immunol. 152, 5495–5502. Burn, T. C., Petrovick, M. S., Hohaus, S., Rollins, B. J., and Tenen, D. G. (1994). Monocyte chemoattractant protein-1 gene is expressed in activated neutrophils and retinoic acid-induced human myeloid cell lines. Blood 84, 2776–2783. Tracey, K. J., and Cerami, A. (1994). Tumor necrosis factor: A pleiotropic cytokine and therapeutic target. Annu. Rev. Med. 45, 491–503. Bazzoni, F., and Beutler, B. (1996). The tumor necrosis factor ligand and receptor families. N. Engl. J. Med. 334, 1717–1725. Edwards, S. W., and Hallett, M. B. (1997). Seeing the wood for the trees: The forgotten role of neutrophils in rheumatoid arthritis. Immunol. Today 18, 320–324. Lindemann, A., Riedel, D., Oster, W., Loems Ziegler-Heitbrock, H. W., Mertelsmann, R., and Hermann, F. (1989). Granulocyte–macrophage colony-stimulating factor induces cytokine secretion by human polymorphonuclear leukocytes. J. Clin Invest. 83, 1308–1312. Mandi, Y., Nagy, Z., Ocsovski, I., and Farkas, G. (1997). Effects of tumor necrosis factor and pentoxifylline on ICAM-1 expression on human polymorphonuclear granulocytes. Int. Arch. Allergy Immunol. 114, 329–335. Galbraith, G. M., Hagan, C., Steed, R. B., Sanders, J. J., and Javed, T. (1997). Cytokine production by oral and peripheral blood neutrophils in adult periodontitis. J. Periodontol. 68, 832–838. Mandi, Y., Degre, M., and Beladi, I. (1989). Involvement of tumor necrosis factor in human granulocyte-mediated killing of WEHI-164 cells. Int. Arch. Allergy Appl. Immunol. 90, 411–413. Yamazaki, M., Ikenami, M., and Sugiyama, T. (1989). Cytotoxin from polymorphonuclear leukocytes and inﬂammatory ascitic ﬂuids. Br. J. Cancer 59, 353–355. Mandi, Y., Endresz, V., Krenacs, L., Regely, K., Degre, M., and Beladi, I. (1991). Tumor necrosis factor production by human granulocytes. Int. Arch. Allergy Appl. Immunol. 96, 102–106. Djeu, J. Y. (1992). Cytokines and anti-fungal immunity. Adv. Exp. Med. Biol. 319, 217–223. Sweeney, J. F., Nguyen, P. K., Omann, G. M., and Hinshaw, D. B. (1998). Autocrine/ paracrinemodulation of polymorphonuclear leukocyte su rival after exposure to Candida albicans. Shock 9, 146–152. Vulcano, M., Alves Rosa, M. F., Minnucci, F. S., Chernavsky, A. C., and Isturiz, M. A. (1998). N-Formyl-methionyl-leucyl-phenylalanine (fMLP) inhibits tumour necrosis factor-움 (TNF-움) production on lipopolysaccharide (LPS)-stimulated human neutrophils. Clin. Exp. Immunol. 113, 39–47. Ohlsson, K., Linder, C., Lundberg, E., and Axelsson, L. (1996). Release of cytokines and proteases from human peripheral blood mononuclear and polymorphonuclear cells following phagocytosis and LPS stimulation. Scand. J. Clin. Lab. Invest. 56, 461–470. Palma, C., Cassone, A., Serbousek, D., Pearson, C. A., and Djeu, J. Y. (1992). Lactoferrin release and interleukin-1, interleukin 6, and tumor necrosis factor production by human polymorphonuclear cells stimulated by various lipopolysaccharides: relationship to growth inhibition of Candida albicans. Infect. Immun. 60, 4604–4611. Thiel, M., and Chouker, A. (1995). Acting via A2 receptors, adenosine inhibits the production of tumor necrosis factor-움 of endotoxin-stimulated human polymorphonuclear leukocytes. J. Lab. Clin. Med. 126, 275–282. Terashima, T., Soejima, K., Waki, Y., Nakamura, H., Fujishima, S., Suzuki, Y., Ishizaka, A., and Kanazawa, M. (1995). Neutrophils activated by G-CSF suppress TNF움 release from monocytes stimulated by endotoxin. Am. J. Resp. Cell. Mol. Biol. 13, 69–73.

490

MARCO ANTONIO CASSATELLA

175. Scuderi, P., Nez, P. A., Duerr, M. L., Wong, B. J., and Vadez, J. M. (1991). Cathepsin G and leukocyte elastase inactivate human tumor necrosis factor and lymphotoxin. Cell. Immunol. 135, 299–313. 176. Van kessel, K. P., Van Strip, J. A., and Verhoef, J. (1991). Inactivation of recombinant human tumor necrosis factor-움 by proteolytic enzymes released from stimulated human neutrophils. J. Immunol. 147, 3862–3868. 177. Steadman, R., Petersen, M. M., Topley, N., Matthews, N., Spur, B., and Williams, J. D. (1990). Differential augmentation by recombinant human tumor necrosis factor움 of neutrophil responses to particulate zymosan and glucan. J. Immunol. 144, 2712– 2718. 178. Takeichi, O., Saito, I., Tsurumachi, T., Moro, I., and Saito, T. (1996). Expression of inﬂammatory cytokine genes in vivo by human alveolar bone-derived polymorphonuclear leukocytes isolated from chronically inﬂamed sites of bone resorption. Calcif. Tissue Int. 58, 244–248. 179. Kasama, T., Strieter, R. M., Lukacs, N. W., Lincoln, P. M., Burdick, M. D., and Kunkel, S. L. (1995). Interferon 웂 modulates the expression of neutrophil-derived chemokines. J. Invest. Med. 43, 58–67. 180. Haziot, A., Tsuberi, B. Z., and Goyert, S. M. (1993). Neutrophil CD14: Biochemical properties and role in the secretion of tumor necrosis factor-움 response to lipopolysaccharide. J. Immunol. 150, 5556–5565. 181. Brandtzaeg, P., Mollnes, T. E., and Kierulf, P. (1989). Complement activation and endotoxin levels in systemic meningococcal disease. J. Infect. Dis. 160, 58–65. 182. Strieter, R. M., Remick, D. G., Ward, P. A., Spengler, R. N., Lynch, J. P., Larrick, J., and Kunkel, S. L. (1988). Cellular and molecular regulation of tumor necrosis factor움 production by pentoxifylling. Biochem. Biophys. Res. Commun. 155, 1230–1236. 183. Arnold, R., Scheffer, J., Konig, B., and Konig, W. (1993). Effects of Listeria monocytogenes and Yersinia enterocolitica on cytokine gene expression and release from human polymorphonuclear granuocytes and epithelial (HEp-2) cells. Infect. Immun. 61, 2545– 2552. 184. Balazovich, K. J., Suchard, S. J., Remick, D. G., and Boxer, L. A. (1996). Tumor necrosis factor-움 and FMLP receptors are functionally linked during FMLP stimulated activation of adherent human neutrophils. Blood 88, 690–696. 185. Nathan, C. F. (1987). Neutrophil activation on biological surfaces. Massive secretion of hydrogen peroxide in response to products of macrophages and lymphocytes. J. Clin. Invest. 80, 1550–1560. 186. Contrino, J., Krause, P. J., Slover, N., and Kreutzer, D. (1993). Elevated interleukin1 expression in human neonatal neutrophils. Pediatr. Res. Sep. 34, 249–252. 187. Wright, S. D., and Jong, M. T. (1986). Adhesion-promoting receptors on human macrophages recognize Escherichia coli by binding to lipopolysaccharide. J. Exp. Med. 164, 1876–1888. 188. Wei, S., Blanchard, D. K., Liu, J. H., Leonard, W. J., and Djeu, J. Y. (1993). Activation of tumor nacrosis factor-움 production from human neutrophils by IL-2 via IL-R웁. J. Immunol. 150, 1979–1987. 189. Djeu, J. Y., Liu, J. H., Wei, S., Rui, H., Pearson, C. A., Leonard, W. J., and Blanchard, D. K. (1993). Function associated with IL-2 receptor 웁 expression on human neutrophils . Mechanism of activation of antifungal activity against Candida albicans by IL 2. J. Immunol. 150, 960–970. 190. Takeshita, T., Asao, H., Ohtani, K., Ishii, N., Kumaki, S., Tanaka, N., Munakata, H., Nakamura, M., and Sugamura, K. (1992). Cloning of the 애 chain of the human IL2 receptor. Science 257, 379–382.

NEUTROPHIL-DERIVED PROTEINS

491

191. Asao, H., Takeshita, T., Ishii, N., Kurnaki, S., Nakamura, M., and Sugamura, K. (1993). Reconstitution of functional IL-2 receptor complexes on ﬁbroblastoid cells: Involvement of the cytoplasmic domain of the 애 chain in two distinct signaling pathways. Proc. Natl. Acad. Sci. U.S.A. 90, 4127–4131. 192. Ishii, N., Takeshita, T., Kimura, Y., Tada, K., Kondo, M., Nakamura, M., and Sugamura, K. (1994). Expression of the IL-2 receptor 애 chain on various populations in human peripheral blood. Int. Immunol. 6, 1273–1277. 193. Philips, J. H., Takeshita, T., Sugamura, K., and Lanier, L. (1989). Activation of natural killer cells via the p75 interleukin-2 receptor. J. Exp. Med. 170, 291–296. 194. Girard, D., Gosselin, J., Heitz, D., Paquin, R., and Beaulieu, A. D. (1995). Effects of IL-2 on gene expression in human neutrophils. Blood 86, 1170–1176. 195. Liu, J. H., Wei, S., Ussery, D. W., Epling-Burnette, P. K., Leonard, W. J., and Djeu, J. Y. (1994). Expression of the IL-2 receptor 웂 chain on human neutrophils. Blood 84, 3780–3875. 196. Schumann, R. R., Nakarai, T., Gruss, H. J., Brach, M. A., von Arnim, U., Kirschning, C., Karawajew, L., Ludwig, W. D., Renauld, J. C., Ritz, J., and Herrmann, F. (1996). Transcript synthesis and surface expression of the interleukin-2 receptor (움웁 and 웂chain) by normal and malignant myeloid cells. Blood 87, 2419–2427. 197. Kowanko, I. C., Ferrante, A., Clemente, G., Youssef, P. P., and Smith, M. (1996). Tumor necrosis factor priming of peripheral blood neutrophils from rheumatoid arthritis patients. J. Clin. Immunol. 16, 216–221. 198. Chinnaiyan, A. M., and Dixit, V. M. (1997). Portrait of an executioner: The molecular mechanism of FAS/APO-1-induced apoptosis. Semin. Immunol. 9, 69–76. 199. Dhein, J., Walczak, H., Baumler, C., Debatin, K., and Krammer, M. (1995). Autocrine T-cell suicide mediated by APO-1/CD95. Nature (London) 373, 438–441. 200. Nagata, S. (1998). Fas induced apoptosis Intern. Med. 37, 179–181. 201. 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. 202. Savill, J. (1997). Apoptosis in resolution of inﬂammation. J. Leukoc. Biol. 61, 375–380. 203. Haslett, C. (1997). Granulocyte apoptosis and inﬂammatory disease. Br. Med Bull. 53, 669–683. 204. Iwai, K., Miyawaki, T., Takizawa, T., Konno, A., Ohta, K., Yachie, A., Seki, H., and Taniguchi, N. (1994). Differential expression of bcl-2 and susceptibility to anti-Fasmediated cell death in peripheral blood lymphocytes, monocytes, and neutrophils. Blood 84, 1201–1208. 205. Liles, W. C., and Klebanoff, S. J. (1995). Regulation of apoptosis in neutrophils—Fas track to death? J. Immunol. 155, 3289–3291. 206. Larochelle, B., Flamand, L., Gourde, P., Beauchamp, D., and Gosselin, J. (1998). Epstein–Barr virus infects and induces apoptosis in human neutrophils. Blood 92, 291–299. 207. Gruss, H. J., and Dower, S. K. (1995). The TNF ligand superfamily and its relevance for human diseases. Cytokines Mol. Ther. 1, 75–105. 208. McDonald, P. P., Cassatella, M. A., Bald, A., Maggi, E., Romagnani, S., Gruss, H. J., and Pizzolo, G. (1995). CD30 ligation induces nuclear factor kappa B activation in human T cell lines. Eur. J. Immunol. 25, 2870–2876. 209. Biswas, P., Smith, C. A., Goletti, D., Hardy, E. C., Jackson, R. W., and Fauci, A. S. (1995). Cross-linking of CD30 induces HIV expression in chronically infected T cells. Immunity 2, 587–596. 210. Gruss, H. J., Pinto, A., Gloghini, A., Wehnes, E., Wright, B., Boiani, N., Aldinucci, D., Gattei, V., Zagonel, V., Smith, C. A., Kadin, M. E., von Schilling, C., Goodwin,

492

211.

212.

213. 214. 215. 216.

217.

218.

219.

220. 221.

222.

223.

224. 225.

226.

227.

MARCO ANTONIO CASSATELLA

R. G., Herrmann, F., and Carbone, A. (1996). CD30 ligand expression in nonmalignant and Hodgkin’s disease-involved lymphoid tissues. Am. J. Pathol. 149, 469–481. Maggi, E., Annunziatio, F., Manetti, R., Biagiotti, R., Giudizi, M. G., Ravina, A., Almerigogna, F., Boiani, N., Alderson, M., Romagnani, S. (1995). Activation of HIV expression by CD30 triggering in CD4⫹ T cells from HIV-infected individuals. Immunity 3, 251–255. Ho, J. L., He, S., Hu, A., Geng, J., Basile, F. G., Almeida, M. G., Saito, A. Y., Laurence, J., and Johnson, W. D. (1995). Neutrophils from human immunodeﬁciency virus (HIV)seronegative donors induce HIV replication from HIV-infected patients’ mononuclear cells and cell lines: An in vitro model of HIV transmission facilitated by Chlamydia trachomatis. J. Exp. Med. 181, 1493–1505. Dinarello, C. A. (1997). Interleukin-1. Cytokine Growth Factor Rev. 8, 253–265. Colotta, F., Dower, S. K., Sims, J. E., and Mantovani, A. (1994). The type II ‘docoy’ receptor: A novel regulatory pathway for interleukin-1. Immunol. Today 15, 562–566. Dinarello, C. A. (1996). Biologic basis for interleukin-1 in disease. Blood 87, 2095–2147. Bennet, I. L., and Beeson, P. B. (1953). Studies on the pathogenesis of fever-producing substance from the polymorphonuclear leukocytes and from the ﬂuid of sterile exudates. J. Exp. Med. 98, 493–498. Hanson, D. F., Murphy, P. A., and Windle, B. E. (1980). Failure of rabbit neutrophils to secrete endogenous pyrogen when stimulated with staphylococci. J. Exp. Med. 151, 1360–1371. Windle, B. E., Murphy, P. A., and Cooperman, S. (1983). Rabbit polymorphonuclear leukocytes do not secrete endogenous pyrogens or interleukin-1 when stimulated by endotoxin polyinosine:polycytosine or muramyl dipeptide. Infect. Immun. 39, 1142– 1146. Goto, F., Nakamura, S., Goto, K., and Yoshinaga, M. (1984). Production of a lymphocyte proliferation potentiating factor by puriﬁed polymorphonuclear leucocytes from mice and rabbits. Immunology 53, 683–692. Kusaka, Y., and Donaldson, K. (1990). Production of interleukin-1-like activity by neutrophils derived from rat lung. Thorax 45, 281–282. Kitamura, M. Goto, F., Ohkawara, S., and Yoshinaga, M. (1986). Production of pyrogen by polymorphonuclear leukocytes during the course of casein induced peritonitis in rabbits. Acta Pathol. Jpn. 36, 791–803. Ohkawara, S., Goto, K., Mori, S., Goto, F., Saita, N., Sagara, T., and Yoshinaga, M. (1989). Interleukin-1 production by polymorphonuclear leukocytes during the course of acute inﬂammation in rabbits. Dermatologica 179 (Suppl. 1), 84–90. Ohkawara, S., Matsukawa, A., and Yoshinaga, M. (1992). Neutrophils as the major producer of interleukin-1 and its inhibitor at the inﬂmmatory site in rabbits. Arch. Immunol. Ther. Exp. (Warsz.) 40, 11–16. Canning, P. C., and Neill, J. D. (1989). Isolation and characterization of interleukin1 from bovine polymorphonuclear leukocytes. J. Leukoc. Biol. 45, 21–28. Jupin, C., Parant, M., Chedid, L., and Damais, C. (1987). Enhanced oxidative burst without interleukin-1 production by normal human polymorphonuclear leukocytes primed with muramyl dipeptides. Inﬂammation 11, 153–161. Payne, C. M., Glasser, L., Tischler, M. E., Wyckoff, D., Cromey, D., Fiederlein, R., and Bohnert, O. (1994). Programmed cell death of the normal human neutrohil: An in vitro model of senescence. Microsc. Res. Tech. 28, 327–344. Tiku, K., Tiku, M. L., and Skosey, J. L. (1986). Interleukin-1 production by human polymorphonuelear neutrophils. J. Immunol. 136, 3677–3685.

NEUTROPHIL-DERIVED PROTEINS

493

228. Tiku, K., Tiku, M. L., Liu, S., and Skosey, J. L. (1986). Normal human neutrophils are a source of a speciﬁc interleukin-1 inhibitor. J. Immunol. 136, 3686–3692. 229. Dularay, B., Westacott, C. I., and Elson, C. J. (1992). IL1 secreting cell assay and its application to cells from patients with rheumatoid arthritis. Br. J. Rheumatol. 31, 19–24. 230. Marucha, P. T., Zeff, R. A., and Kreutzer, D. L. (1991). Regulation of IL1 웁 gene expression in human peripheral blood PMN. J. Periodont. Res. 26, 264–267. 231. Hendley, T. M., Steed, R. B., and Galbraith, G. M. (1995). Interleukin-1 웁 gene expression in human oral polymorphonuclear leukocytes. J. Periodontol. 66, 761–765. 232. Fernandez, M. C., Walters, J., and Marucha, P. (1996). Transcriptional and posttranscriptional regulation of GM-CSF induceed IL1 웁 gene expression in PMN. J. Leukoc. Biol. 59, 598–603. 233. Goh, K., Furusawa, S., Kawa, Y., Negishi-Okitsu, S., and Mizoguchi, M. (1989). Production of interleukin-1 움 and 웁 by human peripheral polymorphonuclear neutrophils. Int. Arch. Allergy Appl. Immunol. 88, 297–303. 234. His, E. D., and Remick, D. G. (1995). Monocytes are the major producers of interleukin-1 웁 in an ex vivo model of local cytokine production. J. Interferon Cytokine Res. 15, 89–94. 235. Malyak, M., Smith, M. F., Abel, A. A., and Arend, W. P. (1994). Peripheral blood neutrophil production of interleukin-1 receptor antagonist and interleukin-1웁. J. Clin. Immunol. 14, 20–30. 236. Miller, G. A., DeMayo, T., and Hutter, J. W. (1996). Production of interleukin-1 by polymorphonuclear leukocytes resident in periradicular tissues. J. Endod. 22, 346–351. 237. Quayle, J. A., Adams, S., Bucknall, R. C., and Edwards, S. W. (1994). Cytokine expression by inﬂammatory neutrophils. FEMS Immunol. Med. Microbiol. 8, 233–239. 238. Colotta, F., Re, F., Muzio, M., Bertini, R., Polentarutti, N., Sironi, M., Giri, J. G., Dower, S. K., Sims, J. E., and Mantovani, A. (1993). Interleukin-1 type II receptor: A decoy target for IL-1 that is regulated by IL-4. Science 261, 472–475. 239. William, R., Watson, G., Rotstein, O. D., Parodo, J., Bitar, R., and Marshall, J. C. (1998). The IL-1웁-converting enzyme (caspase-1) inhibits apoptosis of inﬂammatory neutrophils through activation of IL-1웁. J. Immunol. 161, 957–962. 240. Walters, J. D., Cario, A. C., Leblebicioglu, B., Fernandez, M. C., and Marucha, P. T. (1995). An inhibitor of polyamine biosynthesis impairs human polymorphonuclear leukocyte priming by tumor necrosis factor-움. J. Leukoc. Biol. 57, 282–286. 241. Walters, J. D., Cario, A. C., Danne, M. M., and Marucha, P. T. (1998). An inhibitor of ornithine decarboxylase antagonizes superoxide generation by primed human polymorphonuclear leukocytes. J. Inﬂamm. 48, 40–46. 242. Brooks, C. J., King, W. J., Radford, D. J., Adu, D., McGrath, M., and Savage, C. O. (1996). IL-1웁 production by human polymorphonuclear leucocytes stimulated by antineutrophil cytoplasmic autoantibodies: Relevance to systemic vasculitis. Clin. Exp. Immunol. 106, 273–279. 243. Kasama, T., Kobayashi, K., Fukushima, T., Tabata, M., Ohno, I., Negishi, M., Ide, H., Takahashi, T., and Niwa, Y. (1989). Production of interleukin-1-like factor from human peripheral blood monocytes and polymorphonuclear leukocytes by superoxide anion: The role of interleukin-1 and reactive oxygen species in inﬂamed sites. Clin. Immunol. Immunopathol. 53, 439–448. 244. White, F. S., Quayle, J. A., and Edwards, S. W. (1996). Gene expression by inﬂammatory neutrophils. Stimulation of interleukin-1 웁 production by rheumatoid synovial ﬂuid. Biochem. Soc. Trans. 24, 493S. 245. Cassone, A., Palma, C., Djeu, J. Y., Aiuti, F., and Quinti, I. (1993). Anticandidal activity and interleukin-1 웁 and interleukin-6 production by polymorphonuclear leukocytes are preserved in subjects with AIDS. J. Clin. Microbiol. 31, 1354–1357.

494

MARCO ANTONIO CASSATELLA

246. Ueta, E., Umazume, M., Yamamoto, T., and Osaki, T. (1994). Enhancement of polymorphonuclear leukocyte (PMN) function by OK 432. Int. J. Immunopharmacol. 16, 7–17. 247. Yamazaki, K., Polak, B., Bird, P. S., Gemmell, E., Hara, K., and Seymour, G. J. (1989). Effects of periodontopathic bacteria on IL-1 and IL-1 inhibitor production by human polymorphonuclear neutrophils. Oral Microbiol. Immunol. 4, 193–198. 248. McDonald, P. P., Gasperini, S., Calzetti, F., and Cassatella, M. A. (1998). Modulation by interferon-gamma of the production and gene expression of IL-1 receptor antagonist in human neutrophils. Cell Immunol. 184, 45–50. 249. Falk, R. J., Terrel, R. S., Charles, L. A., and Jennette, J. C. (1990). Anti-neutrophilcytoplasmic autoantibodies induce neutrophils to degranulate and produce oxygen radicals in vitro. Proc. Natl. Acad. Sci. U.S.A. 87, 4115–4119. 250. Roberge, C. J., Grassi, J., De Medicis, R., Frobert, Y., Lussier, A., Naccache, P. H., and Poubelle, P. E. (1991). Crystal–neutrophil interactions lead to interleukin-1 synthesis. Agents Actions 34, 38–41. 251. Roberge, C. J., de Medicis, R., Dayer, J. M., Rola-Pleszczynski, M., Naccache, P. H., and Poubelle, P. E. (1994). Crystal-induced neutrophil activation. V. Differential production of biologically active IL-1 and IL-1 receptor antagonist. J. Immunol. 152, 5485–5494. 252. Genth, E. (1993). Crystal-induced arthropathies. Gerontology 26, 24–33. 253. Jobin, C., and Gauthier, J. (1997). Differential effects of cell density on 5-lipoxygenase (5-LO), ﬁve-lipoxygenase-activating protein (FLAP) and interleukin-1웁 (IL-1웁) expression in human neutrophils. Inﬂammation 21, 235–250. 254. Arend, W. P., Malyak, M., Guthridge, C. J., and Gabay, C. (1998). Interleukin-1 receptor antagonist: Role in biology. Annu. Rev. Immunol. 16, 27–55. 255. Dripps, D. J., Brandhuber, B. J., Thompson, R. C., and Eisenberg, S. P. (1991). Interleukin-1 (IL-1) receptor antagonist binds to the 80-kDa IL-1 receptor but does not initiate IL-1 signal transduction. J. Biol. Chem. 266, 10331–10336. 256. Beaulieu, A. D., Paquin, R., Rathanaswami, P., and McColl, S. R. (1992). Nuclear signaling in human neutrophils. Stimulation of RNA synthesis is a response to a limited number of proinﬂammatory agonists. J. Biol. Chem. 267, 426–432. 257. Ulevitch, R. J., and Tobias, P. S. (1995). Receptor-dependent mechanisms of cell stimulation by bacterial endotoxin. Annu. Rev. Immunol. 13, 437–457. 258. Cassatella, M. A., Meda, L., Gasperini, S., Calzetti, F., and Bonora, S. (1994). Interleukin 10 (IL-10) upregulates IL-1 receptor antagonist production from lipopolysaccharide-stimulated human polymorphonuclear leukocytes by delaying mRNA degradation. J. Exp. Med. 179, 1695–1699. 259. Poutsiaka, D. D., Clark, B. D., Vannier, E., and Dinarello, C. A. (1991). Production of interleukin-1 receptorantagonist and interleukin-1웁 by peripheral blood mononuclear cells is differentially, regulated. Blood 78, 1275–1281. 260. Arend, W. P., Smith, M. F., Janson, R. W., and Joslin, F. G. (1991). IL-1 receptor antagonist and IL-1 웁 production in human monocytes are regulated differently. J. Immunol. 147, 1530–1536. 261. Chang, D. M. (1995). Cellular signals for the induction of human interleukin-1 receptor antagonist. Clin. Immunol. Immunopathol. 74, 23–30. 262. Muzio, M., Sironi, M., Polentarutti, N., Mantovani, A., and Colotta, F. (1994). Induction by transforming growth factor 웁1 of the interleukin1 receptor antagonist and of its intracellular form in humen polymorphonuclear cells. Eur. J. Immunol. 24, 3194–3198. 263. Marie, C., Pitton, C., Fitting, C., and Cavaillon, J. M. (1996). IL-10 and IL-4 synergize with TNF-움 to induce IL-1ra production by human neutrophils. Cytokine 8, 147–151.

NEUTROPHIL-DERIVED PROTEINS

495

264. Muzio, M., Polentarutti, N., Sironi, M., Poli, G., De Gioia, L., Introna, M., Mantovani, A., and Colotta, F. (1995). Cloning and characterization of a new isoform of the interleukin-1 receptor antagonist. J. Exp. Med. 182, 623–628. 265. Seymour, G. J., Gemmell, E., Reinhardt, R. A., Eastcott, J., and Taubman, M. A. (1993). Immunopathogenesis of chronic inﬂammatory periodontal disease: Cellular and molecular mechanisms. J. Periodont. Res. 28, 478–486. 266. Beaulieu, A. D., Paquin, R., and Gosselin, J. (1995). Epstein–Barr virus modulates de novo protein synthesis in human neutrophils Blood 86, 2789–2798. 267. Roberge, C. J., Larochelle, B., Rola-Pleszczynski, M., and Gosselin, J. (1997). Epstein– Barr virus induces GM-CSF synthesis by monocytes: Effect on EBV-induced IL-1 and IL-1 receptor antagonist production in neutrophils. Virology 238, 344–352. 268. Hirano, T. (1998). Interleukin 6 and its receptor: Ten years later. Int. Rev. Immunol. 16, 249–284. 269. Kato, K., Yokoi, T., Takano, N., Kanegane, H., Yachie, A., Miyawaki, T., and Taniguchi, N. (1990). Detection by in situ hybridization and phenotypic characterization of cell expressing IL-6 mRNA in human stimulated blood. J. Immunol. 144, 1317–1322. 270. Cicco, N. A., Lindemann, A., Content, J., Vandenbussche, P., Lubbert, M., Gauss, J., Mertelsmann, R., and Hermann, F. (1990). Inducible production of interleukin-6 by human neutrophils: Role of granulocyte-macrophage colony-stimulating factor and tumor necrosis factor 움. Blood 75, 2049–2052. 271. Mianji, S., Hamasaki, Y., Yamamoto, S., and Miyazaki, S. (1996). Inhibition by dexamethasone of the lipopolysaccharide-induced increase in IL-6 mRNA abundance and IL-6 production in human polymorphonuclear leukocytes. Int. J. Immunopharmacol. 18, 339–346. 272. Melani, C., Mattia, G. F., Silvani, A., Care´, A., Rivoltini, L., Parmiani, G., and Colombo, M. P. (1993). Interleukin-6 expression in human neutrophil and eosinophil peripheral blood granulocytes. Blood 81, 2744–2749. 273. Jablonska, E., and Pietruska, Z. (1998). Changes in soluble IL-6 receptor and IL-6 production by polymorphonuclear cells and whole blood cells of breast cancer patients. Arch. Immunol. Ther. Exp. (Warsz.) 46, 25–29. 274. Euler, G. J., Miller, G. A., Hutter, J. W., and D’Alesandro, M. M. (1998). Interleukin6 in neutrophils from peripheral blood and inﬂammatory periradicular tissues. J. Endod. 24, 480–484. 275. Ericson, S. G., Zhao, Y., Gao, H., Miller, K. L. Gibson, L. F., Lynch, J. P., and Landreth, K. S. (1998). Interleukin-6 production by human neutrophils after Fcreceptor cross-linking or exposure to granulocyte colony-stimulating factor. Blood 91, 2099–2107. 276. Perussia, B., Dayton, E. T., Fanning, V., Thiagarajan, P., Hoxie, J., and Trinchieri, G. (1983). Immune interferon and leukocyte-conditioned medium induce normal and leukemic myeloid cells to differentiate along the monocytic pathway. J. Exp. Med. 158, 2058–2080. 277. Hoffmeyer, F., Witte, K., and Schmidt, R. E. (1997). The high-afﬁnity Fc웂RI on PMN: Regulation of expression and signal transduction. Immunology 92, 544–552. 278. Hamid, Q., Barkans, J., Meng, Q., Ying, S., Abrams, J. S., Kay, A. B., and Moqbel, R. (1992). Human eosinophils synthesize and secrete interleukin-6, in vitro. Blood 80, 1496–1501. 279. Gleich, G. J., and Adolphson, C. R. (1986). The eosinophilic leukocyte. Structure and function. Adv. Immunol. 39, 177–253. 280. Taga, T., and Kishimoto, T. (1997). Gp130 and the interleukin-6 family of cytokines. Annu. Rev. Immunol. 15, 797–819.

496

MARCO ANTONIO CASSATELLA

281. Modur, V., Feldhaus, M. J., Weyrich, A. S., Jicha, D. L., Prescott, S. M., Zimmerman, G. A., and McIntyre, T. M. (1997). Oncostatin M is a proinﬂammatory mediator. In vivo effects correlate with endothelial cell expression of inﬂammatory cytokines and adhesion molecules. J. Clin. Invest. 100, 158–168. 282. Grenier, A., Dehoux, M., Boutten, A., Arce-Vicioso, M., Durand, G., GougerotPocidalo, M. A., and Chollet-Martin, S. (1998). Oncostatin M production and regulation by human polymorphonuclear neutrophils. Blood 93, 1413–1421. 283. Trinchieri, G. (1998). Interleukin-I2: A citokine at the interface of inﬂammation and immunity. Adv. Immunol. 70, 83–243. 284. Gately, M. K., Renzetti, L. M., Magram, J., Stern, A. S., Adorini, L., Gubler, U., and Presky, D. H. (1998). The interleukin-12/interleukin-12 receptor system: role in normal and pathologic immune responses. Annu. Rev. Immunol. 16, 495–521. 285. Hayes, M. P., Wang, J., and Norcross, M. A. (1995). Regulation of interleukin-12 expression in human monocytes: Selective priming by interferon-애 of lipopolysaccharide-inducible p35 and p40 genes. Blood 86, 646–650. 286. Snijders, A., Hilkens, C.M., van der Pouw Kraan, T. C., Engel, M., Aarden, L.A., and Kapsenberg, M. L. (1996). Regulation of bioactive IL-12 production in lipopolysaccharide-stimulated human monocytes is determined by the expression of the p35 subunit. J. Immunol. 156, 1207–1212. 287. Grewe, M., Czech, W., Morita, A., Werfel, T., Klammer, M., Kapp, A., Ruzicka, T., Schopf, E., and Krutmann, J. (1998). Human eosinophils produce biologically active IL-12: Implications for control of T cell responses. J. Immunol. 161, 415–420. 288. D’Andrea, A., Aste-Amezaga, M., Valiante, N. M., Ma, X., Kubin, and Trinchieri, G. (1993). Interleukin10 (IL-10) inhibits human lymphocyte interferon 웂-production by suppressing natural killer cell stimulatory factor/IL-12 synthesis in accessory cells. J. Exp. Med. 178, 1041–1048. 289. Cassatella, M. A. (1998). The neutrophil: One of the cellular targets of interleukin10. Int. J. Clin. Lab. Res. 28, 148–161. 290. Gasperini, S., Zambello, R., Agostini, C., Trentin, L., Tassinari, C., Cadrobbi, P., Semenzato, G., and Cassatella, M. A. (1998). Impaired cytokine production by neutrophils isolated from patients with AIDS. AIDS 12, 373–379. 291. Romani, L., Mencacci, A., Cenci, E., Del Sero, G., Bistoni, F., and Puccetti, P. (1997). An immunoregulatory role for neutrophils in CD4⫹ T helper subset selection in mice with candidiasis. J. Immunol. 158, 2356–2362. 292. Romani, L., Mencacci, A., Cenci, E., Spaccapelo, R., Del Sero, G., Nicoletti, I., Trinchieri, G., Bistoni, F., and Puccetti, P. (1997). Neutrophil production of IL-12 and IL-10 in candidiasis and efﬁcacy of IL-12 therapy in neutropenic mice. J. Immunol. 158, 5349–5356. 293. Mencacci, A., Del Sero, G., Cenci, E., d’Ostiani, C. F., Bacci, A., Montagnoli, C., Kopf, M., and Romani, L. (1998). Endogenous interleukin 4 is required for development of protective CD4⫹ T helper type 1 cell responses to Candida albicans. J. Exp. Med. 187, 307–317. 294. D’Andrea, A., Ma, X., Aste-Amezaga, M., Paganin, C., and Trinchieri, G. (1995). Stimulatory and inhibitory effects of interleukin (IL)-4 and IL-13 on the production of cytokines by human peripheral blood mononuclear cells: priming for IL-12 and tumor necrosis factor 움 production. J. Exp. Med. 181, 537–546. 295. Mencacci, A., Cenci, E., Boelaert, J. R., Bucci, P., Mosci, P., Fe., d’Ostiani, C., Bistoni, F., and Romani, L. (1997). Iron overload alters innate and T helper cell responses to Candida albicans in mice. J. Infect. Dis. 175, 1467–1476.

NEUTROPHIL-DERIVED PROTEINS

497

296. Mencacci, A., Cenci, E., Del Sero, G., Fe d’Ostiani, C., Mosci, P., Montagnoli, C., Bacci, A., Bistoni, F., Quesniaux, V. F., Ryffel, B., and Romani, L. (1998). Defective co-stimulation and impaired Th1 development in tumor necrosis factor/lymphotoxin움 double-deﬁcient mice infected with Candida albicans. Int. Immunol. 10, 37–48. 297. Kanangat, S., Thomas, J., Gangappa, S., Babu, J. S., and Rouse, B. T. (1996). Herpes simplex virus type 1-mediated up-regulation of IL-12 (p40) mRNA expression. Implications in immunopathogenesis and protection. J. Immunol. 156, 1110–1116. 298. Trinchieri, G., Kubin, M., Bellone, G., and Cassatella, M. A. (1993). Cytokine crosstalk between phagocytic cells and lymphocytes: Relevance for differentiation/activation of phagocytic cells and regulation of adaptive immunity. J. Cell Biochem. 53, 301–308. 299. Pfeffer, L. M., Dinarello, C. A., Herberman, R. B., Williams, B. R., Borden, E. C., Bordens, R., Walter, M. R., Nagabhushan, T. L., Trotta, P. P., and Pestka, S. (1998). Biological properties of recombinant 움-interferons: 40th anniversary of the discovery of interferons. Cancer Res. 58, 2489–2499. 300. Baron, S. et al., eds. (1992). ‘‘Interferon. Principles and Application.’’ The University of Texas Medical Branch Department of Microbiology, Galveston, Texas. 301. Hiscott, J., Ryals, J., Dierks, P., Hofmann, V., and Weissmann, C. (1984). The expression of human interferon 움 genes. Philos. Trans. R. Soc. Lond., Ser. B., Biol. Sci. 307, 217–226. 302. Khabar, K. S., Al-Zoghaibi, F., Al-Ahdal, M. N., Murayama, T., Dhalla, M., Mukaida, N., Taha, M., Al-Sedairy, S. T., Siddiqui Y, Kessie G, and Matsushima K. (1997). The 움 chemokine, interleukin-8, inhibits the antiviral action of interferon 움. J. Exp. Med. 186, 1077–1085. 303. Rouse, B. T., Babiuk, L. A., and Henson, P. M. (1980). Neutrophils in antiviral immunity: Inhibition of virus replication by a mediator produced bybovine neutrophils. J. Infect. Dis. 141, 223–232. 304. Ohmann, H. B., Campos, M., Fitzpatrick, D. R., Rapin, N., and Babiuk, L. A. (1989). A neutrophil-derived antiviral protein: Induction requirements and biological properties. J. Virol. 63, 1916–1923. 305. Farrar, M. A., and Schreiber, R. D. (1993). The molecular cell biology of interferon웂 and its receptor. Annu. Rev. Immunol. 11, 571–611. 306. Cassatella, M. A., and Berton, G. (1992). Modulation of neutrophil-functions by IFN웂. In ‘‘Interferon. Principles and Application’’ (S. Baron et al., eds.), pp. 387–398. The University of Texas Medical Branch Department of Microbiology, Galveston, Texas. 307. Gallin, J. I. (1991). Interferon-웂 in the management of chronic granulomatous disease. Rev. Infect. Dis. 13, 973–978. 308. Yeaman, G. R., Collins, J. E., Currie, J. K., Guyre, P. M., Wira, C. R., and Fanger, M. W. (1998). IFN-웂 is produced by polymorphonuclear neutrophils in human uterine endometrium and by cultured peripheral blood polymorphonuclear neutrophils. J. Immunol. 160, 5145–5453. 309. Keel, M., Ungethum, U., Steckholzer, U., Niederer, E., Hartung, T., Trentz, O., and Ertel, W. (1997). Interleukin-10 counterregulates proinﬂammatory cytokine induced inhibition of neutrophil apoptosis during severe sepsis. Blood 90, 3356–3363. 310. Kumar, V., Bustin, S. A., and McKay, I. A. (1995). Transforming-growth factor 움. Cell. Biol. Int. 19, 373–388. 311. Wong, D. T., Weller, P. F., Galli, S. J., Elovic, A., Rand, T. H., Gallagher, G. T., Chiang, T., Chou, M. Y., Matossian, K., McBride, J., et al. (1990). Human eosinophils express transforming growth factor 움. J. Exp. Med. 172, 673–681. 312. Madtes, D. K., Raines, E. W., Sakariassen, K. S., Assoian, R. K., Sporn, M. B., Bell, G. I., and Ross, R. (1988). Induction of transforming growth 움 in activated human alveolar macrophages. Cell 53, 285–293.

498

MARCO ANTONIO CASSATELLA

313. Calafat, J., Janssen, H., Stahle-Backdahl, M., Zuurbier, A. E., Knol, E. F., and Egesten, A. (1997). Human monocytes and neutrophils store transforming growth factor 움 in a subpopulation of cytoplasmic granules. Blood 90, 1255–1266. 314. Bry, K., Hallman, M., and Lappalainen, U. (1994). Cytokines released by granulocytes and mononuclear cells stimulate amnion cell prostaglandin E2 production. Prostaglandins 48, 389–399. 315. Clark, D. A., and Coker, R. (1998). Transforming growth factor-웁 (TGF-웁). Int. J. Biochem. Cell. Biol. 30, 293–298. 316. Wahl, S. M. (1994). Transforming growth factor beta: The good, the bad, and the ugly. J. Exp. Med. 180, 1587–1590. 317. Wiseman, D. M., Polverini, P. J., Kamp, D. W., and Leibovich, S. J. (1988). Transforming growth factor-beta (TGF beta) is chemotactic for human monocytes and induces their expression of angiogenic activity. Biochem. Biophys. Res. Commun. 157, 793–800. 318. Balazovich, K. J., Fernandez, R., Hinkovska-Galcheva, V., Suchard, S. J., and Boxer, L. A. (1996). Transforming growth factor-웁1 stimulates degranulation and oxidant release by adherent human neutrophils. J. Leukoc. Biol. 60, 772–777. 319. Brandes, M. E., Mai, U. E., Ohura, K., and Wahl, S. M. (1991). Type I transforming growth factor-웁 receptors on neutrophils mediate chemotaxis to transforming growth factor-웁. J. Immunol. 147, 1600–1606. 320. Amoroso, A., Del Porto F., Di Monaco, C., Manfredini, P., and Afeltra, A. (1997). Vascular endothelial growth factor: A key mediator of neoangiogenesis. Eur. Rev. Med. Pharmacol. Sci. 1, 17–25. 321. Ferrara, N., and Davis-Smyth, T. (1997). The biology of vascular endothelial growth factor. Endocrinol. Rev. 18, 4–25. 322. Brown, L. F., Detmar, M., Claffey, K., Nagy, J. A., Feng, D., Dvorak, A. M., and Dvorak, H. F. (1997). Vascular permeability factor/vascular endothelial growth factor: A multifunctional angiogenic cytokine. EXS 79, 233–269. 323. Barleon, B., Sozzani, S., Zhou, D., Weich, H. A., Mantovani, A., and Marme, D. (1996). Migration of human monocytes in response to vascular endothelial growth factor (VEGF) is mediated via the VEGF receptor ﬂt 1. Blood 87, 3336–3343. 324. Boros, P., and Miller, C. M. (1995). Hepatocyte growth factor: A multifunctional cytokine. Lancet 345, 293–295. 325. To, C. T. T., and Tsao, M. S. (1998). The roles of hepatocyte growth factor/scatter factor and met receptor in human cancers. Oncol. Rep. 5, 1013–1024. 326. Demetri, G. D., and Grifﬁn, J. D. (1991). Granulocyte colony-stimulating factor and its receptor. Blood 78, 2791–2808. 327. Ramenghi, U., Ruggieri, L., Dianzani, I., Rosso, C., Brizzi, M. F., Camaschella, C., Pietsch, T., and Saglio, G. (1994). Human peripheral blood granulocytes and myeloid leukemic cell lines express both transcripts encoding for stem cell factor. Stem Cells (Dayton) 12, 521–526. 328. Kita, H., Ohnishi, T., Okubo, Y., Weiler, D., Abrams, J. S., and Gleich, G. J. (1991). Granulocyte/macrophage colony-stimulating factor and interleukin 3 release from human peripheral blood eosinophils and neutrophils. J. Exp. Med. 174, 745–748. 329. Chitnis, D., Dickerson, C., Munster, A. M., and Winchurch, R. A. (1996). Inhibition of apoptosis in polymorphonuclear neutrophils from burn patients. J. Leukoc. Biol. 59, 835–839. 330. Iida, N., Haisa, M., Igarashi, A., Pencev, D., and Grotendorst, G. R. (1996). Leukocyte derived growth factor links the PDGF and CXC chemokine families of peptides. FASEB J. 10, 1336–1345.

NEUTROPHIL-DERIVED PROTEINS

499

331. Ho, C. K., Ou, B. R., Hsu, M. L., Su, S. N., Yung, C. H., and Wang, S. Y. (1990). Induction of thymidine kinase activity and clonal growth of certain leukemic cell lines by a granulocyte-derived factor. Blood 75, 2438–2444. 332. Kay, E. P., Rivela, L., and He, Y. G. (1990). Corneal endothelium modulation factor released by polymorphonuclear leukocytes. Partial puriﬁcation and initial characterization. Invest. Ophthalmol. Vis. Sci. 31, 313–322. 333. Kay, E. P., Gu, X., Ninomiya, Y., and Smith, R. E. (1993). Corneal endothelial modulation: A factor released by leukocytes induces basic ﬁbroblast growth factor that modulates cell shape and collagen. Invest. Ophthalmol. Vis. Sci. 34, 663–672. 334. Kay, E. P., Gu, X., and Smith, R. E. (1994). Corneal endothelial modulation: bFGF as direct mediator and corneal endothelium modulation factor as induces. Invest. Ophthalmol. Vis. Sci. 35, 2427–2435. 335. Hashida, R., Kuwada, M., Chiba, K. I., Horizoe, T., Shirota, H., and Nagai, Y. (1996). A factor derived from polymorphonuclear leukocytes enhances interleukin-1-induced synovial cell collagenase and prostaglandin E2 production in rate. Eur. J. Biochem. 236, 517–522. 336. Ishibashi, H., Kurokawa, S., Tsuchiya, Y., Hayashida, K., Hirata, Y., and Okubo, H. (1989). Induction of polymorphonuclear leukocyte-derived hepatocyte stimulating factor and effects on 움 2-macroglobulin synthesis. J. Clin. Lab. Immunol. 29, 147–150. 337. Hirata, Y., Kurokawa, S., Ishibashi, H., Hayashida, K., Kimura, H., Nagano, M., and Okubo, H. (1986). Polymorphonuclear leukocytes and the induction of 움2 macroglobulin synthesis. J. Clin. Lab. Immunol. 21, 125–130. 338. Sheng, Y. J., Bird, J., Pompidou, A., Florentin, I., and Giroud, J. P. (1984). Rat and human polymorphonuclear leukocyte derived lymphocyte stimulatory factors. Biomed. Pharmacother. 38, 304–308. 339. Sheng, Y. J., Bird, J., and Giroud, J. P. (1984). Effects of supernatants and lysates of polymorphonuclear leukocytes—Lymphocyte mitogenic factors. Int. J. Immunopharmacol. 6, 557–562. 340. Bird, J., Sheng, Y. J., and Giroud, J. P. (1984). Effects of supernatants and lysates of polymorphonuclear leucocytes: Macrophage stimulatory factors. Br. J. Exp. Pathol. 65, 243–250. 341. Yoshinaga, M., Nishime, K., Nakamura, S., and Goto, F. (1980). A PMN derived factor that enhances DNA-synthesis in PHA or antigen stimulated lymphocytes. J. Immunol. 124, 94–99. 342. Rodrick, M. L., Lamster, I. B., Sonis, S. T., Pender, S. G., Kolodkin, A. B., Fitzgerald, J. E., and Wilson, R. E. (1982). Effects of supernatants of polymorphonuclear neutrophils recruited by different inﬂammatory substances on mitogen responses of lymphocytes. Inﬂammation 6, 1–11. 343. Harris, P. J. (1982). Neutrophil product with lymphocyte activating factor activity. Clin. Exp. Immunol. 50, 474–478. 344. Fitzgerald, J., West, K., Sonis, S., and Wilson, R. (1983). Polymorphonuclear leukocyte regulation of lymphocyte proliferation and differentiation. Immunobiology 165, 421–431. 345. Laurenzi, M. A., Beccari, T., Stenke, L., Sjolinder, M., Stinchi, S., and Lindgren, J. A. (1998). Expression of mRNA encoding neurotrophins and neurotrophin receptors in human granulocytes and bone marrow cells—Enhanced neurotrophin-4 expression induced by LTB4. J. Leukoc. Biol. 64, 228–234. 346. Nikolaus, S., Bauditz, J., Gionchetti, P., Witt, C., Lochs, H., and Schreiber, S. (1998). Increased secretion of pro-inﬂammatory cytokines by circulating polymorphonuclear

500

347.

348.

349.

350. 351.

352.

353.

354.

355.

356.

357.

358. 359.

360.

361.

362.

MARCO ANTONIO CASSATELLA

neutrophils and regulation by interleukin 10 during intestinal inﬂammation. Gut 42, 470–476. Chollet-Martin, S., Gatecel, C., Kermarrec, N., Gougerot-Pocidalo, M. A., Payen, D. M. (1996). Alveolar neutrophil functions and cytokine levels in patients with the adult respiratory distress syndrome during nitric oxide inhalation. Am. J. Respir. Crit. Care Med. 153, 985–990. Niiro, H., Otsuka, T., Izuhara, K., Yamaoka, K., Ohshima, K., Tanabe, T., Hara, S., Nemoto, Y., Tanaka, Y., Nakashima, H., and Niho, Y. (1997). Regulation by interleukin10 and interleukin 4 of cyclooxygenase-2 expression in human neutrophils. Blood 89, 1621–1628. Kuhns, D. B., DeCarlo, E., Hawk, D. M., and Gallin, J. (1992). Dynamics of the cellular and humoral components of the inﬂammatory response elicited in skin blisters in humans. J. Clin. Invest. 89, 1734–1740. Francoeur, C., and Denis, M. (1995). Nitric oxide and interleukin 8 as inﬂammatory components of cystic ﬁbrosis. Inﬂammation 19, 587–598. Suzuki, H., Takahashi, Y., Wataya, H., Ikeda, K., Nakabayashi, S., Shimomura, A., and Takasaka, T. (1996). Mechanism of neutrophil recruitment induced by IL-8 in chronic sinusitis. J. Allergy Clin. Immunol. 98, 659–670. Simon, H. U., Youseﬁ, S., Weber, M., Simon, D., Holzer, C., Hartung, K., and Blaser, K. (1995). Human peripheral blood eosinophils express and release interleukin 8. Int. Arch. Allergy Immunol. 107, 124–126. Pang, G., Ortega, M., Zighang, R., Reeves, G., and Clancy, R. (1997). Autocrine modulation of IL-8 production by sputum neutrophils in chronic bronchial sepsis. Am. J. Respir. Crit. Care Med. 155, 726–731. Lin, C. Y., and Huang, T. P. (1994). Gene expression and release of interleukin-8 by peritoneal macrophages and polymorphonuclear leukocytes during peritonitis in uremic patients on continuous ambulatory peritoneal dialysis. Nephron 68, 437–441. Lin, C. Y., Lin, C. C., and Huang T. P. (1993). Serial changes of interleukin-6 and interleukin-8 levels in drain dialysate of uremic patients with continuous ambulatory peritoneal dialysis during peritonitis. Nephron 63, 404–408. Gillitzer, R., Ritter, U., Spandau, U., Goebeler, M., and Brocker, E. B. (1996). Differential expression of GRO-움 and IL-8 mRNA in psoriasis: A model for neutrophil migration and accumulation. in vivo. J. Invest. Dermatol. 107, 778–782. Kulke, R., Todt-Pingel, I., Rademacher, D., Rowert, J., Schroder, J. M., and Christophers, E. (1996). Co-localized overexpression of GRO-움 and IL-8 mRNA is restricted to the suprapapillary layers of psoriatic lesions. J. Invest. Dermatol. 106, 526–530. Schroder, J. M. (1995). Cytokine networks in the skin. J. Invest. Dermatol. 105 (Suppl. 1), 20S–24S. Gebbers, J. O., and Otto, H. F. (1977). Immunohistochemical and electromicroscopic observations on the local immune response in ulcerative colitis. Virchows Arch. A Pathol., Pathol. Anat. 374, 271–273. Mazzucchelli, L., Hauser, C., Zgraggen, K., Wagner, H., Hess, M., Laissue J. A., and Mueller, C. (1994). Expression of interleukin-8 gene in inﬂammatory bowel disease is related to the histological grade of active inﬂammation. Am. J. Pathol. 144, 997–1007. Raab, Y., Gerdin, B., Ahlstedt, S., and Hallgren, R. (1993). Neutrophil mucosal involvement is accompanied by enhanced local production of interleukin-8 in ulcerative colitis. Gut 34, 1203–1206. Grimm, M. C., Elsbury, S. K., Pavli, P., and Doe, W. F. (1996). Interleukin 8: Cells of origin in inﬂammatory bowel disease. Gut 38, 90–98.

NEUTROPHIL-DERIVED PROTEINS

501

363. Daig, R, Andus, T., Aschenbrenner, E., Falk, W., Scholmerich, J., and Gross, V. (1996). Increased interleukin 8 expression in the colon mucosa of patients with inﬂammatory bowel disease. Gut 38, 216–222. 364. Nielsen, O. H., Rudiger, N., Gaustadnes, M., and Horn, T. (1997). Intestinal interleukin-8 concentration and gene expression in inﬂammatory bowel disease. Scand. J. Gastroenterol. 32, 1028–1034. 365. Derkx, B., Taminiau, J., Radema, S., Stronkhorst, A., Wortel, C., Tytgat, G., and van Deventer, S. (1994). Tumour-necrosis-factor antibody treatment in Crohn’s disease. Lancet 342, 173–174. 366. Beil, W. J., Weller, P. F., Peppercorn, M. A., Galli, S. J., and Dvorak, A. M. (1995). Ultrastructural, immunogold localization of subcellular sites of TNF-움 in colonic Crohn’s disease. J. Leukoc. Biol. 58, 284–298. 367. Arend, W. P., and Dayer, J. M. (1990). Cytokines and cytokine inhibitors or antagonists in rheumatoid arthritis. Arthritis Rheum. 33, 305–315. 368. Malyak, M., Swaney, R. E., and Arend, W. P. (1993). Levels of synovial ﬂuid interleukin1 receptor antagonist in rheumatoid arthrits and other arthropathies. Potential contribution from synovial ﬂuid neutrophils. Arthrits Rheum. 36, 781–789. 369. Quayle, J. A., Adams, S., Bucknall, R. C., and Edwards, S. W. (1995). Interleukin-1 expression by neutrophils in rheumatoid arthritis. Ann. Rheum. Dis. 54, 930–933. 370. Beaulieu, A. D., and McColl, S. (1994). Differential expression of two major cytokines produced by neutrophils, interleukin-8 and the interleukin-1 receptor antagonist, in neutrophils isolated from the synovial ﬂuid and peripheral blood of patients with rheumatoid arthritis. Arthritis Rheum. 37, 855–859. 371. Landry, M. (1977). Phagocyte function and cell-mediated immunity in systemic lupus erythematosus. Arch. Dermatol. 113, 147–154. 372. Staples, P. J., Gerding, D. N., Decker, J. L., and Gordon, R. S., Jr. (1974). Incidence of infection in systemic lupus erythematosus. Arthritis Rheum. 17, 1–10. 373. Hsieh, S. C., Tsai, C. Y., Sun, K. H., Yu, H. S., Tsai, S. T., Wang, J. C., Tsai, Y. Y., Han, S. H., and Yu, C. L. (1994). Decreased spontaneous and lipopolysaccharide stimulated production of interleukin-8 by polymorphonuclear neutrophils of patients with active systemic lupus erythematosus. Clin. Exp. Rheum. 12, 627–633. 374. Hsieh, S. C., Tsai, C. Y., Sun, K. H., Tsai, Y. Y., Tsai, S. T., Han, S. H., Yu, H. S., and Yu, C. L. (1995). Defective spontaneous and bacterial lipopolysaccharide-stimulated production of interleukin-1 receptor antagonist by polymorphonuclear neutrophils of patients with active systemic lupus erythematosus. Br. J. Rheumatol. 34, 107–112. 375. Chang, D. M. (1997). Interleukin-1 and interleukin-1 receptor antagonist in systemic lupus erythematosus. Immunol. Invest. 26 , 649–659. 376. Dinarello, C. A. (1997). Proinﬂammatory and anti inﬂammatory cytokines as mediators in the pathogenesis of septic shock. Chest 112 (Suppl. 6), 321S–329S. 377. McCall, C. E., Grosso-Wilmoth, L. M., LaRue, K., Guzman, R. N., and Cousart, S. L. (1993). Tolerance to endotoxin-induced expression of the interleukin-1웁 gene in blood neutrophils of humans with the sepsis syndrome. J. Clin. Invest. 91, 853–861. 378. Fasano, M. B., Cousart, S., Neal, S., and McCall, C. E. (1991). Increased expression of the interleukin-1 receptor on blood neutrophils of humans with the sepsis syndrome. J. Clin. Invest. 88, 1452–1459. 379. Takasuka, N., Tokunaga, T., and Akagawa, K. S. (1991). Preexposure of macrophages to low doses of lipopolysaccharide inhibits the expression of tumor necrosis factor-움 mRNA but not of IL-1웁 mRNA. J. Immunol. 146, 3824–3830. 380. Marie, C., Fitting, C., Cheval, C., Losser, M. R., Carlet, J., Payen, D., Foster, K., and Cavaillon, J. M. (1997). Presence of high levels of leukocyte-associated interleukin-8 upon cell activation and in patients with sepsis syndrome. Infect. Immun. 65, 865–871.

502

MARCO ANTONIO CASSATELLA

381. Darbonne, W. C., Rice, G. C., Mohler, M. A., Apple, T., Hebert, C. A., Valente, A. J., and Baker, J. B. (1991). Red blood cells are a sink for interleukin 8, a leukocyte chemotaxin. J. Clin. Invest. 88, 1362–1369. 382. Kabashima, H., Maeda, K., Iwamoto, Y., Hirofuji, T., Yoneda, M., Yamashita, K., and Aono, M. (1990). Partial characterization of an interleukin-1-like factor in human gingival crevicular ﬂuid from patients with chronic inﬂammatory periodontal disease. Infect. Immun. 58, 2621–2627. 383. Takahashi, K., Poole, I., and Kinane, D. F. (1995). Detection of interleukin-1웁 mRNAexpressing cells in human gingival crevicular ﬂuid by in situ hybridization. Arch. Oral Biol. 40, 941–947. 384. Miller, G. A., DeMayo, T., and Hutter, J. W. (1996). Production of interleukin-1 by polymorphonuclear leukocytes resident in periradicular tissue. J. Endod. 22, 346–351. 385. Galbraith, G. M., Steed, R. B., Sanders, J. J., and Pandey, J. P. (1998). Tumor necrosis factor 움 production by oral leukocytes: Inﬂuence of tumor necrosis factor genotype. J. Periodontol. 69, 428–433. 386. Tonetti, M. S., Imboden, M. A., Gerber, L., Lang, N. P., Laissue, J., and Mueller, C. (1994). Localized expression of mRNA for phagocyte speciﬁc chemotactic cytokines in human periodontal infections. Infect. Immun. 62, 4005–4014. 387. Gainet, J., Chollet-Martin, S., Brion, M., Hakim, J., Gougerot-Pocidalo, M. A., and Elbim, C. (1998). Interleukin-8 production by polymorphonuclear neutrophils in patients with rapidly progressive periodontitis: An amplifying loop of polymorphonuclear neutrophil activation. Lab. Invest. 78, 755–762. 388. Garg, A. K., Ortea, A., and Machado, C. L. (1996). Recognition and treatment of radipdly progressive periodontitis. Gen. Dent. 44, 136–139. 389. Raqib, R., Lindberg, A. A., Wretlind, B., Bardhan, P. K., Andersson, U., and Andersson, J. (1995). Persistence of local cytokine production in shigellosis in acute and convalescent stages. Infect. Immun. 63, 289–296. 390. Sansonetti, P. J., Arondel, J., Cavaillon, J. M., and Huerre, M. (1995). Role of interleukin-1 in the pathogenesis of experimental shigellosis. J. Clin. Invest. 96, 884–892. 391. Sprenger, H., Rosler, A., Tonn, P., Braune, H. J., Huffmann, G., and Gemsa, D. (1996). Chemokines in the cerebrospinal ﬂuid of patients with meningitis. Clin. Immunol. Immunopathol. 80, 155–161. 392. Ossege, L. M., Sindern, E., Voss, B., and Malin, J. P. (1996). Expression of tumor necrosis factor-움 and transforming growth factor-웁1 in cerebrospinal ﬂuid cells in meningitis. J. Neurol. Sci. 144, 1–13. 393. Sweeney, J. F., Rosemurgy, A. S., Wei, S., and Djeu, J. Y. (1998). Intact autocrine activation and cytokine production by PMNs from injured adults with elevated Candida antigen titres. Injury 29, 35–40. 394. Bortolami, M., Carlotto, C., Fregona, I., Quaino, V., Venturi, C., and Naccarato, R. (1995). Effects of interferon-움 on the production by polimorphonuclear cells. Fund. Clin. Immunol. 3, 153–156. 395. Jablonska, E., Kiluk, M., Piotrowski, L., Grabowska, Z., Markiewicz, W., and Jablonski, J. (1998). Tumornecrosis factor-움 and soluble tumor necrosis factor receptors in the culture supernatants of polymorphonuclear cells and peripheral blood mononuclear cells from cancer patients. Eur. Cytokine Netw. 9, 155–159. 396. Jablonska, E. (1998). Serum levels of tumor necrosis factor 움 and production of this cytokine by polymorphonuclear cells in breast cancer patients. Arch. Immunol. Ther. Exp. (Warsz.) 46, 93–96. 397. ‘‘Granulocyte Responses to Cytokines. Basic and Clinical Aspects’’ (1992). Dekker, Inc., New York.

NEUTROPHIL-DERIVED PROTEINS

503

398. Van Dervort, A. L., Yan, L., Madara, P. J., Cobb, J. P., Wesley, R. A., Corriveau, C. C., Tropea, M. M., and Danner, R. L. (1994). Nitric oxide regulates endotoxininduced TNF-움 production by human neutrophils. J. Immunol. 152, 4102–4109. 399. Galanos, C., and Freudenberg, M. A. (1993). Mechanisms of endotoxin shock and endotoxin hypersensitivity. Immunobiology 187, 346–356. 400. Ahlin, A., Elinder, G., and Palmblad, J. (1997). Dose-dependent enhancements by interferon-웂 on functional responses of neutrophils from chronic granulomatous disease patients. Blood 89, 3396–3401. 401. Moore, K. W., O’Garra, A., de Waal Malefyt, R., Vieira, P., and Mosmann, T. R. (1993). Interleukin-10. Annu. Rev. Immunol. 11, 165–190. 402. Maloney, C. G., Kutchera, W. A., Albertine, K. H., McIntyre, T. M., Prescott, S. M., and Zimmerman, G. A. (1998). Inﬂammatory agonists induce cytooxygenase type 2 expression by human neutrophils. J. Immunol. 160, 1402–1410. 403. Bussolati, B., Mariano, F., Montrucchio, G., Piccoli, G., and Camussi, G. (1997). Modulatory effect of interleukin-10 on the production of platelet-activating factor and superoxide anions by human leucocytes. Immunology 90, 440–447. 404. Morel, P. A., and Oriss, T. B. (1998). Crossregulation between Th1 and Th2 cells. Crit. Rev. Immunol. 18, 275–303. 405. Jenkins, J. K., Malyak, M., and Arend, W. P. (1994). The effects of interleukin-10 on interleukin-1 receptor antagonist and interleukin-1웁 production in human monocytes and neutrophils. Lymphokine Cytokine Res. 13, 47–54. 406. Ziegler-Heitbrock, H. W., Frankenberger, M., and Wedel, A. (1995). Tolerance to lipopolysaccharide in human blood monocytes. Immunobiology 193, 217–223. 407. Friedman, G., Jankowski, S., Marchant, A., Goldman, M., Kahn, R. J., and Vincent, J. L. (1997). Blood interleukin-10 levels parallel the severity of septic shock. J. Crit. Care 12, 183–187. 408. Re, F., Muzio, M., De Rossi, M., Polentarutti, N., Giri, J. G., Mantovani, A., and Colotta, F. (1994). The type II ‘‘receptor’’ as a decoy target for interleukin-1 in polymorphonuclear leukocytes: Characterization of induction by dexamethasone and ligand binding properties of the released decoy receptor. J. Exp. Med. 179, 739–743. 409. Marucha, P. T., Zeff, R. A., and Kreutzer, D. L. (1991). Cytokine-induced IL-1웁 gene expression in the human polymorphonuclear leukocyte: Transcriptional and posttranscriptional regulation by tumor necrosis factor and IL-1. J. Immunol. 147, 2603– 2608. 410. Cassatella, M. A., Gasperini, S., Calzetti, F., McDonald, P. P., and Trinchieri, G. (1995). Lipopolysaccharide-induced interleukin-8 gene expression in human granulocytes: Transcriptional inhibition by interferon-웂. Biochem. J. 310, 751–755. 411. Cassatella, M. A. (1996). Interferon 웂 inhibits the lipopolysaccharide induced macrophage inﬂammatory protein-1움 gene transcription in human neutrophils. Immunol. Lett. 49, 79–83. 412. Ghosh, S., May, M. J., and Kopp, E. B. (1998). NF-␬B and Rel proteins: Evolutionarily conserved mediators of immune responses. Annu. Rev. Immunol. 16, 225–260. 413. McDonald, P. P., Bald, A., and Cassatella, M. A. (1997). Activation of the NF-␬B pathway by inﬂammatory stimuli in human neutrophils. Blood 89, 3421–3433. 414. McDonald, P. P., and Cassatella, M. A. (1997). Activation of transcription factor NF␬B by phagocytic stimuli in human neutrophils. FEBS Lett. 412, 583–586. 415. Vollebregt, M., McPherson, K. G., Hampton, M. B., and Winterbourn, C. C. (1998). Activation of NF-␬B in human neutrphils during phagocytosis of bacteria independently of oxidant generation. FEBS Lett. 432, 40–44.

504

MARCO ANTONIO CASSATELLA

416. Bordoni, R., Fine, R., Murray, D., and Richmond, A. (1990). Characterization of the role of melanoma growth factor stimulatory activity (MGSA) in the growth of normal melanocytes, nevocytes, and malignant melanocytes. J. Cell Biochem. 44, 207–219. 417. Xing, Z., Kirpalani, H., Torry, D., Jordana, M., and Gauldie, J. (1993). Polymorphonuclear leukocytes as a signiﬁcant source of tumor necrosis factor-움 in endotoxinchallenged lung tissue. Am. J. Pathol. 143, 1009–1015. 418. Xing, Z., Jordana, M., Braciak, T., Ohtoshi, T., and Gauldie, J. (1993). Lipopolysaccharide induces expression of granulocyte/macrophage colony-stimulating factor, interleukin-8, and interleukin-6 in, human nasal, but not lung, ﬁbroblasts: Evidence for heterogeneity within the respiratory tract. Am. J. Respir. Cell. Mol. Biol. 9, 255–263. 419. Huang, S., Paulauskis, J. D., Goldleski, J. J., and Kobzik, L. (1992). Expression of macrophage inﬂammatory protein-2 and KC mRNA in pulmonary inﬂammation. Am. J. Pathol. 41, 981–988. 420. Lukaszewicz, G. C., Souba, W. W., and Abcouwer, S. F. (1996). Induction of cytokine induced neutrophil chemoattractant (CINC) mRNA in the lungs of septic rats. J. Trauma 41, 222–228. 421. Cato, A. C., and Wade, E. (1996). Molecular mechanisms of antiinﬂammatory action of glucocorticoids. Bioessays 18, 371–378. 422. Imamura, S., Matsukawa, A., Ohkawara, S., Kagayama, M., and Yoshinaga, M. (1997). Involvement of tumor necrosis factor-움 interleukin-1웁 interleukin-8, and interleukin1 receptor antagonist in acute lung injury caused by local Shwartzman reaction. Pathol. Int. 47, 16–24. 423. Williams, J. H., Jr., Patel, S. K., Hatakeyama, D., Arian, R., Guo, K., Hickey, T. J., Liao, S. Y., and Ulich, T. R. (1993). Activated pulmonary vascular neutrophils as early mediators of endotoxin induced lung inﬂammation. Am. J. Respir. Cell Mol. Biol. 8, 134–144. 424. Parsey, M. V., Tuder, R. M., and Abraham, E. (1998). Neutrophils are major contributors to intraparenchymal lung IL-1웁 expression after hemorrhage and endotoxemia. J. Immunol. 160, 1007–1013. 425. Chang, S. W., Feddersen, C. O., Henson, P. M., and Voelkel, N. F. (1987). Platelet activating factor mediates hemodynamic changes and lung injury in endotoxin-treated rats. J. Clin. Invest. 79, 1498–1509. 426. Cirelli, R. A., Carey, L. A., Fisher, J. K., Rosolia, D. L., Elsasser, T. H., Caperna, T. J., Gee, M. H., and Albertine, K. H. (1995). Endotoxin infusion in anesthetized sheep is associated with intrapulmonary sequestration of leukocytes that immunohistochemically express tumor necrosis factor-움. J. Leukoc. Biol. 57, 820–826. 427. Kolls, J. K., Xie, J., Lei, D., Greenberg, S., Summer, W. R., and Nelson, S. (1995). Differential effects of in vivo ethanol on LPS-induced TNF and nitric oxide production in the lung. Am. J. Physiol. 268, L991–L998. 428. Meli, R., Raso, G. M., Gualillo, O., Pacilio, M., and Di Carlo, R. (1997). Prolactin modulation of nitric oxide and TNF-움 production by peripheral neutrophils in rats. Life Sci. 61, 1395–1403. 429. Matsumoto, T., Yoshida, S., Shiga, Y., Kikuchi, M., Sata, T., and Shigematsu, A. (1997). Contribution of neutrophils to lipopolysaccharide-induced tumor necrosis factor production and mortality in a carrageenan-pretreated mouse model. FEMS Immunol. Med. Microbiol. 17, 171–178. 430. Nill, M. R., Oberyszyn, T. M., Ross, M. S., Oberyszyn, A. S., and Robertson, F. M. (1995). Temporal sequence of pulmonary cytokine gene expression in response to endotoxin in C3H/HeN endotoxin-sensitive and C3H/HeJ endotoxin resistant mice. J. Leukoc. Biol. 58, 563–574.

NEUTROPHIL-DERIVED PROTEINS

505

431. Terebuth, P. D., Otterness, I. G., Strieter, R. M., Lincoln, P. M., Danforth, J. M., Kunkel, S. L., and Chensue, S. W. (1992). Biologic and immunohistochemical analysis of interleukin-6 expression in vivo. Constitutive and induced expression in murine polymorphonuclear and mononuclear phagocytes Am. J. Pathol. 140, 649–657. 432. Matsukawa, A., Ohkawara, S., Maeda, T., Takagi, K., and Yoshinaga, M. (1993). Production of IL-1 and IL-1 receptor antagonist and the pathological signiﬁcance in lipopolysaccharide-induced arthritis in rabbits. Clin. Exp. Immunol. 93, 206–211. 433. Fukumoto, T., Matsukawa, A., Ohkawara, S., Takagi, K., and Yoshinaga, M. (1996). Administration of neutralizing antibody against rabbit IL-1 receptor antagonist exacerbates lipopolysaccharide-induced arthritis in rabbits. Inﬂamm. Res. 45, 479–485. 434. Matsukawa, A., Yoshimura, T., Maeda, T., Ohkawara, S., Takagi, K., and Yoshinaga, M. (1995). Neutrophil accumulation and activation by homologous IL-8 in rabbits. IL8 induces destruction of cartilage and production of IL-1 and IL-1 receptor antagonist in vivo. J. Immunol. 154, 5418–5425. 435. Matsukawa, A., Yoshimura, T., Miyamoto, K., Ohkawara, S., and Yoshinaga, M. (1997). Analysis of the inﬂammatory cytokine network among TNF 움, IL-1웁, IL-1 receptor antagonist, and IL-8 in LPS-induced rabbit arthritis. Lab. Invest. 76, 629–638. 436. Miyamoto, K., Matsukawa, A., Ohkawara, S., Takagi, K., and Yoshinaga, M. (1997). IL-8 is involved in homologous TNF 움-, but not in IL-1웁-induced neutrophil inﬁltration in rabbits. Inﬂamm. Res. 46, 472–477. 437. Endo, H., Akahoshi, T., Takagishi, K., Kashiwazaki, S., and Matsushima, K. (1991). Elevation of interleukin-8 (IL-8) levels in joint ﬂuids of patients with rheumatoid arthritis and the induction by IL-8 of leukocyte inﬁltration and synovitis in rabbit joints. Lymphokine Cytokine Res. 10, 245–252. 438. Akahoshi, T., Endo, H., Kondo, H., Kashiwazaki, S., Kasahara, T., Mukaida, N., Harada, A., and Matsushima, K. (1994). Essential involvement of interleukin-8 in neutrophil recruitment in rabbits with acute experimental arthritis induced by lipopolysaccharide and interleukin-1. Lymphokine Cytokine Res. 13, 113–116. 439. Matsukawa, A., Miyazaki, S., Maeda, T., Tanase, S., Feng, L., Ohkawara, S., Yoshinaga, M., and Yoshimura, T. (1998). Production and regulation of monocyte chemoattractant protein-1 in lipopolysaccharide- or monosodium urate crystal-induced arthritis in rabbits: Roles of tumor necrosis factor alpha, interleukin-1, and interleukin-8 Lab. Invest. 78, 973–985. 440. Mo, J. S., Matsukawa, A., Ohkawara, S., and Yoshinaga, M. (1998). Involvement of TNF 움, IL-1웁 and IL-1 receptor antagonist in LPS-induced rabbit uveitis. Exp. Eye Res. 66, 547–557. 441. Mori, S., Goto, K., Goto, F., Murakami, K., Ohkawara, S., and Yoshinaga, M. (1994). Dynamic changes in mRNA expression of neutrophils during the course of acute inﬂammation in rabbits. Int. Immunol. 6, 149–156. 442. Goto, F., Goto, K., Mori, S., Ohkawara, S., and Yoshinaga, M. (1989). Biosynthesis of interleukin-1웁 at inﬂammatory site in rabbits: Kinetics and producing cells. Br. J. Exp. Pathol. 70, 597–606. 443. Matsukawa, A., Yoshimura, T., Maeda, T., Takahashi, T., Ohkawara, S., and Yoshinaga, M. (1998). Analysis of the cytokine network among tumor necrosis factor 움, interleukin1웁, interleukin-8, and interleukin-1 receptor antagonist in monosodium urate crystalinduced rabbit arthritis. Lab. Invest. 78, 559–569. 444. Nishimura, A., Akahoshi, T., Takahashi, M., Takagishi, K., Itoman, M., Kondo, H., Takahashi, Y., Yokoi, K., Mukaida, N., and Matsushima, K. (1997). Attenuation of monosodium urate crystal induced arthritis in rabbits by a neutralizing antibody against interleukin-8. J. Leukoc. Biol. 62, 444–449.

506

MARCO ANTONIO CASSATELLA

445. Ivey, C. L., Williams, F. M., Collins, P. D., Jose, P. J., and Williams, T. J. (1995). Neutrophil chemoattractants generated in two phases during reperfusion of ischemic myocardium in the rabbit. Evidence for a role for C5a and interleukin-8. J. Clin. Invest. 95, 2720–2728. 446. Williams, F. M., Kus, M., Tanda, K., and Williams, T. J. (1994). Effect of duration of ischaemia on reduction of myocardial infarct size by inhibition of neutrophil accumulation using an anti-CD18 monoclonal antibody. Br. J. Pharmacol. 111, 1123–1128. 447. Youker, K. A., Birdsall, H. H., Frangogiannis, N. G., Kumar, A. G., Lindsey, M. L., Ballantyne, C. M., Smith, C. W., Rossen, R. D., and Entman, M. L. (1997). Phagocytes in ischemia injury. Ann. N.Y. Acad. Sci. 832, 243–265. 448. Matsumoto, T., Ikeda, K., Mukaida, N., Harada, A., Matsumoto, Y., Yamashita, J., and Matsushima, K. (1997). Prevention of cerebral edema and infarct in cerebral reperfusion injury by an antibody to interleukin-8. Lab. Invest. 77, 119–125. 449. Hangai, M., Yoshimura, N., Yoshida, M., Yabuuchi, K., and Honda, Y. (1995). Interleukin-1 gene expression in transient retinal ischemia in the rat. Invest. Ophthalmol. Vis. Sci. 36, 571–578. 450. Sakanashi, Y., Takeya, M., Yoshimura, T., Feng, L., Morioka, T., and Takahashi, K. (1994). Kinetics of macrophage subpopulations and expression of monocyte chemoattractant protein-1 (MCP-1) in bleomycin-induced lung injury of rats studied by a novel monoclonal antibody agianst rat MGR-1. J. Leukoc. Biol. 56, 741–750. 451. Azuma, A., Furuta, T., Enomoto, T., Hashimoto, Y., Uematsu, K., Nukariya, N., Murata, A., and Kudoh, S. (1998). Preventive effect of erythromycin on experimental bleomycin induced acute lung injury in rats. Thorax 53, 186–189. 452. D’Angio, C. T., Sinkin, R. A., LoMonaco, M. B., and Finkelstein, J. N. (1995). Interleukin-8 and monocyte chemoattractant protein-1 mRNAs in oxygen-injured rabbit lung. Am. J. Physiol. 268, L826–L831. 453. Piedboeuf, B., Horowitz, S., Johnston, C. J., Gamache, M., Belanger, S., Poubelle, P. E., Welty, S. E., and Watkins, R. H. (1998). Interleukin-1 expression during hyperoxic lung injury in the mouse. Free Radic. Biol. Med. 24, 1446–1454. 454. Johnston, C. J., Wright, T. W., Reed, C. K., and Finkelstein, J. N. (1997). Comparison of adult and newborn pulmonary cytokine mRNA expression after hyperoxia. Exp. Lung Res. 23, 537–552. 455. Wada, T., Tomosugi, N., Naito, T., Yokoyama, H., Kobayashi, K., Harada, A., Mukaida, N., and Matsushima, K. (1994). Prevention of proteinuria by the administration of anti-interleukin 8 antibody in experimental acute immune complex-induced glomerulonephritis. J. Exp. Med. 180, 1135–1140. 456. Feng, L., Xia, Y., Yoshimura, T., and Wilson, C. B. (1995). Modulation of neutrophil inﬂux in glomerulonephritis in the rat with anti-macrophage inﬂammatory protein-2 (MIP-2) antibody. J. Clin. Invest. 95, 1009–1017. 457. Jordan, M., Otterness, I. G., Ng, R., Gessner, A., Rollinghoff, M., and Beuscher, H. U. (1995). Neutralization of endogenous IL-6 suppresses induction of IL-1 receptor antagonist. J. Immunol. 154, 4081–4090. 458. Tilg, H., Trehu, E., Atkins, M. B., Dinarello, C. A., and Mier, and J. W. (1994). Interleukin-6 (IL-6) as an anti-inﬂammatory cytokine: Induction of circulating IL-1 receptor antagonist and soluble tumor necrosis factor receptor p55. Blood 83, 113–118. 459. Dai, W. J., Bartens, W., Kohler, G., Hufnagel, M., Kopf, M., and Brombacher, F. (1997). Impaired macrophage listericidal and cytokine activities are responsible for the rapid death of Listeria monocytogenes-infected IFN-웂 receptor-deﬁcient mice. J. Immunol. 158, 5297–5306.

NEUTROPHIL-DERIVED PROTEINS

507

460. Thomas, J., Gangappa, S., Kanangat, S., and Rouse, B. T. (1997). On the essential involvement of neutrophils in the immunopathologic disease: Herpetic stromal keratitis. J. Immunol. 158, 1383–1391. 461. Mosmann, T. R., and Coffman, R. L. (1989). Th1 and Th2 cells: Different patterns of lymphokine secretion lead to different functional properties. Annu. Rev. Immunol. 7, 145–173. 462. Romani, L. (1997). The T cell response against fungal infections. Curr. Opin. Immunol. 9, 484–490. 463. Romani, L., Puccetti, P., and Bistoni, F. (1996). Initiation of T-helper cell immunity to Candida albicans by IL-12. The role of neutrophils Chem. Immunol. 63, 115–137. 464. Romani, L., Mencacci, A., Cenci, E., Puccetti, P., and Bistoni, F. (1996). Neutrophils and the adaptive immune response to Candida albicans. Res. Immunol. 147, 512–518. 465. Romani, L., Puccetti, P., and Bistoni, F. (1997). Interleukin-12 in infectious diseases. Clin. Microbiol. Rev. 10, 611–636. 466. Lavigne, L. M., Schopf, L. R., Chung, C. L., Maylor, R., and Sypek, J. P. (1998). The role of recombinant murine IL-12 and IFN-웂 in the pathogenesis of a murine systemic Candida albicans infection. J. Immunol. 160, 284–292. 467. Fresno, M., Kopf, M., and Rivas, L. (1997). Cytokines and infectious diseases. Immunol. Today 18, 56–58. 468. Kudo, C., Yamashita, T., Araki, A., Terashita, M., Watanabe, T., Atsumi, M., Tamura, M., and Sendo, F. (1993). Modulation of in vivo immune response by selective depletion of neutrophils using a monoclonal antibody, RP-3, I. Inhibition by RP-3 treatment of the priming and effector phases of delayed-type hypersensitivity to sheep red blood cells in rats. J. Immunol. 150, 3728–3738. 469. Kudo, C., Yamashita, T., Terashita, M., and Sendo, F. (1993). Modulation of in vivo immune response by selective depletion of neutrophils using a monoclonal antibody. RP-3 II Inhibition by RP-3 treatment of mononuclear leukocyte recruitment in delayedtype hypersensitivity to sheep red blood cells in rats. J. Immunol. 150, 3739–3746. 470. Kudo, C., Araki, A., Matsushima, K., and Sendo, F. (1991). Inhibition of IL-8-induced W3/25⫹ (CD4⫹ ) T lymphocyte recruitment into subcutaneous tissues of rats by selective depletion of in vivo neutrophils with a monoclonal antibody. J. Immunol. 147, 2196– 2201. 471. Tamura, M., Sekiya, S., Terashita, M., and Sendo, F. (1994). Modulation of the in vivo immune response by selective depletion of neutrophils using a monoclonal antibody, RP-3. III. Enhancement by RP-3 treatment of the anti-sheep red blood cell plaque-forming cell response in rats. J. Immunol. 153, 1301–1308. 472. Terashita, M., Kudo, C., Yamashita, T., Greita, T., Gresser, I., and Sendo F. (1996). Enhancement of delayed-type hypersensitivity to sheep red blood cells in mice by granulocyte colony stimulating factor administration at the elicitation phase. J. Immunol. 156, 4638–4643. 473. Midorikawa, Y., Yamashita, T., and Sendo F. (1990). Modulation of the immune response to transplanted tumors in rats by selective depletion of neutrophils in vivo using a monoclonal antibody: Abrogation of speciﬁc transplantation resistance to chemical carcinogen-induced syngeneic tumors by selective depletion of neutrophils in vivo. Cancer Res. 50, 6243–6247. 474. Tanaka, E., and Sendo F. (1993). Abrogation of tumor-inhibitory MRC-OX8⫹ (CD8⫹ ) effector T cell generation in rats by selective depletion of neutrophils in vivo using a monoclonal antibody. Int. J. Cancer 54, 131–136. 475. Colombo, M. P., Lombardi, L., Melani, C., Parenza, M., Baroni, C., Ruco, L., and Stoppacciaro, A. (1996). Hypoxic tumor cell death and modulation of endothelial

508

476. 477. 478. 479. 480. 481. 482.

483. 484. 485. 486.

487.

488. 489.

490. 491.

MARCO ANTONIO CASSATELLA

adhesion molecules in the regression of G-CSF transduced tumors. Am. J. Pathol. 148, 473–483. Tani-Ishii, N., Wang, C. Y., and Stashenko, P. (1995). Immunolocalization of bone resorptive cytokines in rat pulp and periapical lesions following surgical pulp exposure. Oral Microbiol. Immunol. 10, 213–219. Fouad, A. F. (1997). IL-1움 and TNF-움 expression in early periapical lesions of normal and immunodeﬁcient mice. Dent. Res. 76, 1548–1554. Assuma, R., Oates, T., Cochran, D., Amar, S., and Graves, D. T. (1998). IL-1 and TNF antagonists inhibit the inﬂammatory response and bone loss in experimental periodontitis. J. Immunol. 160, 403–409. Fink, G. W., and Norman, J. G. (1996). Intrapancreatic interleukin-1웁 gene expression by speciﬁc leukocyte populations during acute pancreatitis. J. Surg. Res. 63, 369–373. Hubner, G., Brauchle, M., Smola, H., Madlener, M., Fassler, R., and Werner, S. (1996). Differential regulation of pro-inﬂammatory cytokines during wound healing in normal and glucocorticoid-treated mice. Cytokine 8, 548–556. Bartholdi, D., and Schwab, M. E. (1997). Expression of pro-inﬂammatory cytokine and chemokine mRNA upon experimental spinal cord injury in mouse: An in situ hybridization study Eur. J. Neurosci. 9, 1422–1438. Streit, W. J., Semple-Rowland, S. L., Hurley, S. D., Miller, R. C., Popovich, P. G., and Stokes, B. T. (1998). Cytokine mRNA proﬁles in confused spinal cord and axotomized facial nucleus suggest a beneﬁcial role for inﬂammation and gliosis. Exp. Neurol. 152, 74–87. Itoh, K., Okubo, K., Utiyama, H., Hirano. T., Yoshii, J., and Matsubara, K. (1998). Expression proﬁle of active genes in granulocy Blood 9, 1432–1441. Gasperini, S., Bovolenta, C., Vollebregt, M., and Cassatella, M. A. (1998). IL-10 fails to induce the tyrosine phosphorylation of STAT1 and STAT3 in human neutrophils. Submitted. Finbloom, D. S., and Winestock, K. D. (1995). IL-10 induces the tyrosine phosphorylation of Tyk2 ad Jak1 and the differnetial assembly of STAT1움 and STAT3 complexes in human T cells and monocytes. J. Immunol. 155, 1079–1090. 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. Feldman, G. M., Chuang, E. J., and Finbloom, D. S. (1995). IgG immune complexes inhibit IFN-웂-induced transcription of the Fc웂RI gene in human monocytes by preventing the tyrosine phosphorylation of the p91 (Stat 1) transcription factor. J. Immunol. 154, 318–325. Ho, A. S., Wei, S. H., Mui, A. L., Miyajima, A., Moore, K. W. (1995). Functionalregions of the mouse interleukin-10 receptor cytoplasmic domain. Mol. Cell. Biol. 15, 5043–5053. Bovolenta, C., Gasperini, S., and Cassatella, M. A. (1996). Granulocyte colony-stimulating factor induces the binding of STAT1 and STAT3 to the IFN웂 response region within the promoter of-the Fc웂RI/CD64 gene in human neutrophils. FEBS Lett. 386, 239–242. Gosselin, E. J., Wardwell, K., Rigby, W. F., and Guyre, P. M. (1993). Induction of MHC class II on human polymorphonuclear neutrophils by granulocyte/macrophage colony-stimulating factor, IFN웂 and IL-3. J. Immunol. 151, 1482–1490. Smith, W. B., Guida, L., Sun, Q., Korpelainen, E. I., van den Heuvel, C., Gillis, D., Hawrylowicz, C. M., Vadas, M. A., and Lopez, A. F. (1995). Neutrophils activated by

NEUTROPHIL-DERIVED PROTEINS

509

granulocyte-macrophage colony-stimulating factor express receptors for interleukin-3 which mediate class II expression. Blood 86, 3938–3944. 492. Fanger, N. A., Liu, C., Guyre, P. M., Wardwell, K., O’Neil, J., Guo, T. L., Christian, T. P., Mudzinski, S. P., and Gosselin, E. J. (1997). Activation of human T cells by major histocompatability complex class II expressing neutrophils: Proliferation in the presence of superantigen, but not tetanus toxoid. Blood 89, 4128–4135. 493. Oehler, L., Majdic, O., Pickl, W. F., Stockl, J., Riedl, E., Drach, J., Rappersberger, K., Geissler, K., and Knapp, W. (1998). Neutrophil granulocyte-committed cells can be driven to acquire dendritic cell, characteristics. J. Exp. Med. 187, 1019–1028.

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 73

Murine Models of Thymic Lymphomas: Premalignant Scenarios Amenable to Prophylactic Therapy EITAN YEFENOF The Lautenberg Center for General and Tumor Immunology, The Hebrew University—Hadassah Medical Center, Jerusalem 91120, Israel

I. Introduction

Retroviruses that induce thymic lymphomas became important in the ﬁeld of immunology owing to their instrumental role in the discovery of T lymphocytes and T cell immunological function (Miller, 1961 a,b). Concomitantly, they became a powerful tool for the study of the multistep nature of tumor development and the genomic modiﬁcations involved in lymphoma progression (Corcoran et al., 1984; Cuypers et al., 1984; Li et al., 1984; O’Donnel et al., 1985; Selten et al., 1984; Steffen, 1984). The landmark discovery by Ludwig Gross (1951) that spontaneous leukemias of thymic origin in susceptible AKR mice are induced by an endogenous retrovirus introduced a new era of molecular cancer research, which eventually led to the discovery of viral and cellular oncogenes (reviewed in Cofﬁn, 1990). On another front, the work by Gross motivated the attempt to isolate retroviruses from many other types of cancers and to ascribe a viral etiology in a whole range of malignancies (Gross, 1978, 1980). Although this search was futile in most instances (Weinberg, 1996; McCann, 1998), virally induced tumors in general, and retrovirally induced thymic lymphomas in particular, provided experimental systems for the study of genetic, biochemical, cellular, and immunological aspects of tumorigenesis (reviewed in Tsichlis and Lazo, 1991). Oncogenic retroviruses isolated from a variety of animal species have been classiﬁed into two major categories (Varmus, 1984; Nusse, 1986): the acute viruses, which transform susceptible cells in vitro and induce tumors in vivo shortly after their inoculation, and the chronic viruses, which do not transform cells in vitro but induce tumors in vivo after a prolonged latency period. Chronic retroviruses are heterogeneous with regard to their structure, genome, susceptible host, tissue tropism, and the type of tumor induced. However, they share an essential common feature: they are nondefective, replication-competent viruses whose oncogenic activity depends on repeated cycles of infection and integration (Teich, 1982). Yet, because the genome of such viruses does not contain an oncogene, proviral integration into cellular DNA does not lead to immediate neoplastic transformation. Rather, it initiates a complex sequence of 511

Copyright 䉷 1999 by Academic Press. All rights of reproduction in any form reserved. 0065-2776/99 $30.00

512

EITAN YEFENOF

cellular and molecular events that gradually progress toward the development of a frank tumor (Varmus, 1988). A multitude of factors underlying this progressive malignant transformation have been characterized (Tsichlis and Lazo, 1991). Some of them (e.g., insertional mutagenesis) occur at the end stage of the latent period, whereas others (e.g., transactivation of cellular genes or receptor-mediated growth stimulation) begin to operate shortly after the initial infection. Early isolates of chronic retroviruses, such as Gross/AKR and Moloney leukemia virus, predominantly induced lymphomas and leukemias of thymic origin (reviewed in Teich et al., 1983). Because identical tumors could be induced by X-rays, gamma irradiation, and alkalizing carcinogens (e.g., methylnitrosourea), it was proposed that these mutagens activate an endogenous retrovirus that is the common etiologic agent of thymic lymphomas (Fischinger et al., 1981, 1982; Frei, 1980; Janowski et al., 1986). Indeed, in sporadic experiments, an oncogenic retrovirus was isolated from lymphomas arising in mice exposed to X-ray irradiation (Haran-Ghera, 1966; Kaplan, 1974). This virus induced primary thymic lymphomas when inoculated into susceptible mouse strains and accordingly has been termed radiation leukemia virus (RadLV) (Haran-Ghera, 1966; Kaplan, 1967; Haran-Ghera, 1971; Decleve et al., 1974; Lieberman et al., 1978). Subsequent experiments, however, have indicated that induction of lymphomas by radiation does not involve a viral etiology, the resemblance between X-ray- and RadLV-induced lymphomas notwithstanding (Ihle et al., 1976a,b, Yefenof, 1980; Yefenof et al., 1980a,b, Janowski et al., 1990). RadLV is, thus, yet another isolate of an oncogenic endogenous retrovirus that is reactivated in radiation lymphomas on repeated transfers in vivo or propagation in vivo, and is not the primary cause of X-ray-induced lymphomagenesis. Even though murine leukemia retroviruses can infect a range of cell types and tissues in vitro, the outcome of their in vivo inoculation is, by and large, a lymphoma of thymic origin (Cofﬁn, 1990; Tsichlis and Lazo, 1991). Moreover, in the case of AKR lymphomas, the genetically transmitted ecotropic virus is not leukeogenic by itself (Rowe, 1972; Chattopadhyay et al., 1980; Yanagihara et al., 1982). Paradoxically, however, its recombination with an endogenous xenotropic virus, which increases the spectrum of susceptible target cells to viral infection on the one hand, generates an oncogenic virus that induces exclusively thymic lymphomas on the other hand (O’Donnell et al., 1981; Herr and Gilbert, 1983). Likewise, irradiation or methylnitrosourea is mutagenic in a whole range of cells in vitro, yet in vivo such treatments produce malignancies of a predominantly thymic source (Lieberman and Kaplan, 1959; Frei, 1980; Frei and Lawley, 1980; Newcomb et al., 1988). A general conclusion that can be drawn from

MURINE MODELS OF THYMIC LYMPHOMAS

513

these observations is that the mutagenic potential of several, independent carcinogens is realized most effectively within the thymic microenvironment. What makes the lymphocytes of the thymus a preferential target for leukemogenesis? Research on thymocyte development during the past decade has highlighted the unique function of the thymus as a primary lymphoid gland that sustains the differentiation of T lymphocytes and shapes the repertoire of immunocompetent T cells (reviewed in Fowlkes and Pardoll, 1989; Anderson et al., 1996). These features, summarized in the following section, account for the high susceptibility of thymic lymphocytes to transformation by the aforementioned experimental modalities. II. Immunobiology of the Thymus in Relation to Lymphomagenesis

The thymic microenvironment is unique in its ability to uphold the complex process required for generation of immunocompetent T lymphocytes. Similar to all other types of blood cells, T cells are descended from hematopoietic stem cells (HSCs) residing in the bone marrow (Ikuta et al., 1992). An intermediate step in the differentiation of HSCs along the T lymphoid lineage is the formation of prothymocyte precursor cells in the bone marrow, which are recruited to populate the thymus by chemoattractants produced in thymic stromal cells (Champion et al., 1986; Imhof et al., 1988; Bauvois et al., 1989; Deugnier et al., 1989; Carr et al., 1994). On entering the thymus, the prothymocytes engage in a complex sequence of interactions with nonlymphoid stromal cells, leading to their further differentiation and maturation (Boyd et al., 1993; Scollay and Godfrey, 1995). During their intrathymic residence, which lasts 4–5 days, the prothymocytes travel from the corticomedullary junction through the thymic parenchyma and toward the subcapsular cortex (Boyd and Hugo, 1991; Van-Ewijk, 1991; Van-Ewijk et al., 1994). From there the differentiating thymocytes migrate down the cortex and enter the thymic medulla as mature T lymphocytes. While migrating through the intrathymic compartments, the thymocytes proliferate, generating a sufﬁciently large number of TCR-움웁⫹ and TCR-웂␦⫹ cells for repertoire selection (Von-Boehmer, 1990; Jameson et al., 1995). Concomitantly, they acquire receptors for growth factors and adhesion molecules, and costimulatory receptors such as CD4, CD8, and CD28 (Scollay, 1991; Anderson et al., 1996; Anderson and Jenkinson, 1997). The temporal order of these events ensures the shaping of the T cell repertoire such that T cells acquiring receptors with a high afﬁnity for self-antigens/self-MHC are eliminated via apoptosis, whereas those expressing receptors with an afﬁnity to foreign peptides/ self-MHC mature and survive (Huesmann et al., 1991; Ignatowicz et al.,

514

EITAN YEFENOF

1996, Janeway et al., 1998). Eventually, the positively selected T cells leave the thymus and populate secondary lymph nodes, where they provide immune protection against potentially hazardous foreign antigens. Hence, the thymus is a dynamic gland in which lymphocytes undergo extensive proliferation, differentiation, and maturation processes predicated on genetic activation, repression, and recombinatorial events operating in individual cells within a short window of time. These modiﬁcations are mandatory for the maintenance of thymic physiology and function, but at the same time they create a state of genetic instability that renders the population of thymic lymphocytes susceptible to viral, chemical, and radiation carcinogenesis. These events also contribute to the activation of endogeneous retroviral sequences and their recombinations, leading to the generation of a carcinogenic retrovirus, as in the highly leukemic AKR mouse strain. III. Thymic Lymphomas of AKR Mice

The AKR inbred mouse strain has been instrumental in studying the steps underlying retroviral transformation of thymic lymphomas (Rowe, 1978). At least two proviral sequences are carried by the AKR germ line at two distant chromosomal loci designated AKV-1 and AKV-2 (Rowe, 1972; Chattopadhyay et al., 1980). These endogenous viruses are expressed early in the course of embryogenesis, causing viremia from birth on (O’Donnel et al., 1981; Famulari, 1983), but none of them is leukemogenic per se (Yanagihara et al., 1982; Fredrickson et al., 1984). However in mice 5–6 months old, the viruses undergo recombinational events with endogenous, xenotropic viral sequences, resulting in the de novo production of a dual tropic virus (DTV), also known as mink cell focus-forming (MCF) virus, due to its ability to infect mink cells in vitro (Fischinger et al., 1975; Hartley et al., 1977). The DTV is expressed predominantly in the aged thymus and is the etiologic agent of thymic lymphomas that begin to arise 2–3 months later (Pedersen et al., 1981; Hays et al., 1982; Herr and Gilbert, 1983). Injection of DTV into thymuses of young AKR mice accelerates signiﬁcantly the onset of disease and reduces the latent period to 60–90 days following viral administration (Cloyd, 1983; O’Donnell et al., 1984; Staal and Hartley, 1988; Hays et al., 1989a). This is the length of time required for selection of a clonal lymphoma with a proviral integration upstream from the c-myc locus or with a duplication of chromosome 15 that carries an activated c-myc (Corcoran et al., 1984; Li et al., 1984; Steffen, 1984; Wirschubsky et al., 1984a,b; O’Donnell et al., 1985). Hence, the thymic compartment in which premature T lymphocytes proliferate and differentiate provides a microenvironment that favors the sequential occurrence of two stochastic molecular events mandatory for lymphoma

MURINE MODELS OF THYMIC LYMPHOMAS

515

progression: (1) recombination between ecotropic and xenotropic endogenous viral sequences, resulting in the formation of the DTV, and (2) speciﬁc proviral insertion or chromosomal duplication. Whereas the former step involves multiple cycles of virus infection and propagation in an oligoclonal population of thymic lymphocytes, the latter stage entails the selection of a single cell that becomes the sole progenitor of a clonal lymphoma (O’Donnell et al., 1981; Takeuchi et al., 1984; O’Donnell et al., 1985; Cuypers et al., 1986; Hays et al., 1989b, 1990). An inquiry into the nature of lymphocyte populations that provide a cellular source for the evolving lymphoma has led to the characterization of potential lymphoma cells in AKR and other mouse strains susceptible to the induction of primary thymic lymphomas. IV. Prelymphoma Cells in AKR Mice

The term ‘‘prelymphoma cells’’ (PLCs) was introduced by Haran-Ghera (1980a) to depict the presence of potential lymphoma cells in young AKR mice long before the development of overt thymic lymphoma. The concept of PLCs evolved through studies demonstrating that thymectomy of AKR mice at the age of 1–3 months prevented development of spontaneous lymphomas (McEndy et al., 1944; Peled and Haran-Ghera, 1985). However, grafting of a thymus to thymectomized mice enabled progression of the disease, culminating in the emergence of thymic lymphomas arising from the host rather than from the thymic donor cells (Hays, 1982; Takeuchi et al., 1984; Haran-Ghera et al., 1987; Haran-Ghera, 1994). Reciprocal experiments demonstrated that transfer of AKR bone marrow cells to irradiated F1 recipients results in thymic lymphomas of donor origin. These results were interpreted as indicating that cells with a malignant potential are present in the bone marrow of young AKR mice, and later seed in the thymus, where they acquire a fully malignant phenotype. Another line of evidence advocating the existence of PLCs in the bone marrow was the 20–30% incidence of B cell lymphomas developing in thymectomized, aged AKR mice (Peled and Haran-Ghera, 1985). These data, collected in the 1980s, indicated that the DTV replicating in the thymus was not the sole factor triggering lymphoma progression. Another prerequisite is the generation in the bone marrow of PLCs, which, following migration to the thymus, become susceptible targets for the DTV. This concept, however, should be revisited in light of population dynamic studies demonstrating that thymic lymphocytes do not persist in the thymus for more than 4–5 days, the time required to accomplish a full course of positive and negative selection of maturing T cells (Huesmann et al., 1991; Fowlkes and Pardoll, 1989). At the end of this selection interval T lymphocytes must emigrate from the thymus and populate a secondary lymph node or die via apoptosis

516

EITAN YEFENOF

(Fowlkes and Schweighhoffer, 1995; Jameson et al., 1995). It is, thus, obvious that thymic lymphomas arise from lymphocytes that transiently reside in the thymus for a few days. The source of such lymphocytes is bone marrow stem cells that differentiate along the lymphoid lineage into prothymocytes, which then seed in the thymus, where they mature into a functional repertoire of mature T cell clones in a relatively short period of time (Ikuta et al., 1992). Once a DTV has appeared in the thymus following recombination between ecotropic and xenotropic endogenous viral sequences, it will continue to propagate by repeated cycles of infections, affecting lymphocytes trafﬁcking through the thymus and thymic stromal cells (Hays et al., 1984). Such cells are not transformed by the DTV infection and continue to differentiate to maturity or die. However, when a stochastic integration of the provirus upstream from c-myc, duplication of chromosome 15, or another as yet unidentiﬁed genetic event has occurred in a single infected T cell, it will become the progenitor of a clonal lymphoma. The introduction of the PLC concept through the study of AKR lymphomagenesis constituted an important contribution to the ﬁeld of experimental oncology, in that it motivated the search for and the identiﬁcation of PLCs in other murine lymphoma models. These then provided the tools for the study of prophylactic intervention in multistage oncogenesis by targeting and attacking potentially malignant cells, as will be discussed in Section VIII. V. Carcinogen-Induced Lymphomas

AKR mice are susceptible to the induction of thymic lymphomas by the chemical carcinogen N-methyl-N-nitrosourea (NMU). A single injection of NMU into AKR mice at 6 weeks of age results in the accelerated development of thymic lymphomas 2–4 months later, prior to the onset of the spontaneous disease (Frei, 1980; Richie et al., 1985). It was, therefore, assumed that NMU induces genetic recombination between endogenous retroviruses, thereby expediting the generation of a lymphomagenic DTV (Frei, 1980). However, molecular analysis revealed that carcinogeninduced AKR lymphomas contain proviral sequences of ecotropic origin only, which are distinct from the lymphomagenic DTV prevalent in the genome of spontaneous AKR lymphomas (Richie et al., 1985, 1988). Somatic integration of ecotropic proviral sequences in NMU-induced AKR lymphomas does not contribute to lymphomagenesis, based on the ﬁnding that AKR/FV-1B congenic mice were equally sensitive to NMU carcinogenesis (Richie et al., 1991). The B allele of the FV-1 locus restricts the integration and replication of endogenous N-tropic viruses, thereby

MURINE MODELS OF THYMIC LYMPHOMAS

517

protecting AKR (FV-1N ) mice against the development of spontaneous lymphomas (Lilly and Pincus, 1973; Jolicoeur and Rassart, 1980). On the other hand, such mice are equally sensitive to NMU-induced lymphomagenesis, and the resulting lymphomas do not contain somatically acquired proviral integrations (Richie et al., 1991). In addition, treatment of AKR mice with antiviral envelope antibodies, which drastically reduces the incidence of spontaneous lymphomas, had no effect on NMU-induced lymphomagenesis (Haran-Ghera, 1994). Finally, NMU induces thymic lymphomas in mouse strains (e.g., C57L/6J) that do not carry an endogenous ecotropic viral genome, albeit with a longer latency and lower incidence as compared with the AKR strain (Newcomb et al., 1988, 1990; Brathwaite et al., 1992; Kubota et al., 1995). It may, therefore, be surmised that induction of thymic lymphomas by NMU does not involve somatic integration of the proviral genome, which is crucial for the development of spontaneous AKR lymphomas. This conclusion is corroborated by the ﬁnding that a high proportion of the NMU-induced lymphomas in AKR (and other mouse strains) had an activated K-ras, displaying point mutations at codons 12, 61, or 146 of the gene (Geurrero and Pellicer, 1987; Corominas et al., 1991). By contrast, activation of K-ras has not been detected in spontaneous or virally induced lymphomas of these mouse strains (Warren et al., 1987). Nevertheless, the higher incidence and shorter latency of NMU-induced lymphomas in AKR mice, as compared to other inbred strains, suggest cooperation between the chemical carcinogen and endogenous murine leukemia viruses (MuLVs) in lymphoma development. Indeed, NMU treatment stimulated high-level expression of infectious ecotropic MuLV, suggesting that viral products enhance lymphoma progression (Richie et al., 1988). Likewise, the presence of the AKV-1 locus in congenic NFS-N mice, which are otherwise relatively resistant to NMU carcinogenesis, strongly enhanced the formation of thymic lymphomas following NMU treatment (Becker, 1990). This experiment indicates that the AKV-1 locus confers susceptibility to lymphoma induction by increasing the number of target cells responding to NMU. This may occur through multiple infection with the AKV-1-encoded ecotropic virus or by the action of another gene linked to the AKV-1 locus. It seems clear, accordingly, that the etiologies of viral- and NMU-induced lymphomas are distinct. Yet, it is striking that these independent modalities converge to develop a common, tissue-speciﬁc tumor. This outcome emphasizes the uniqueness of the thymic microenvironment as an optimal target for lymphomagenesis. The continuous progression of prothymocytes migrating from the bone marrow to the thymus, followed by extensive proliferation and differentiation before emigration to secondary lymph nodes or death, sustains the high frequency of recombinational events

518

EITAN YEFENOF

between endogenous viruses on the one hand, and the acquisition of speciﬁc point mutations that are induced by the chemical carcinogen on the other hand. VI. Thymic Lymphomas Induced by Fractionated Irradiation

Another modality by which murine thymic lymphomas are induced is whole-body exposure to irradiation (Kaplan, 1967). Most of the studies on radiation-induced lymphomagenesis have been performed in the C57BL mouse strain, which has an extremely low incidence of spontaneous lymphoma. Kaplan and Brown (1952) demonstrated that exposure of C57BL/ Ka mice to four weekly doses of 1.7 Gy induced the development of thymic lymphomas in the majority of the mice, after 4–7 months. Because thymic lymphoma is the only type of tumor emerging in split-irradiated mice, it was initially assumed that the thymus is a direct target organ for cell transformation by fractionated irradiation (Kaplan, 1967, 1974). However, later experiments indicated that X-ray lymphomagenesis proceeds through an indirect mechanism of induction. Thus, irradiation of only the thymus was insufﬁcient to produce an overt disease (Kaplan, 1967). Furthermore, full or partial shielding of the bone marrow during irradiation, as well as infusion of bone marrow cells from unirradiated syngeneic mice, protected the irradiated mouse against lymphoma development (Wallis et al., 1966; Ilbery, 1967; Peled and Haran-Ghera, 1969). Finally, whereas thymectomy rescued irradiated mice from subsequent development of lymphomas, progression of the disease could be restored by subscutaneous grafting of a nonirradiated thymus (Kaplan and Brown, 1954). These observations pointed to involvement of an endogeneous infectious virus in the etiology of the disease, one that is activated in the irradiated bone marrow but infects, and subsequently transforms, lymphocytes of the thymus. A search for a radiation-induced leukemogenic virus led to the isolation of the radiation leukemia virus (RadLV) from the cells of an X-ray-induced thymic lymphoma (Lieberman and Kaplan, 1959; ManteuilBrutlag et al., 1980). This virus induced thymic lymphomas when injected intrathymically (i.t.) into adult C57BL mice (Kaplan, 1961; Decleve et al., 1978; Haran-Ghera et al., 1966). It was thus inferred that RadLV is an endogenous virus activated by fractionated irradiation and that lymphocytes proliferating in the postirradiated thymus are targets for infection and transformation by the reactivated virus. Later experiments, however, revealed that the vast majority of X-ray-induced lymphomas do not produce a leukemogenic virus and that exposure of mice to fractionated irradiation does not activate an endogeneous virus (Lieberman et al., 1976; Decleve et al., 1977). This led to two alternative hypotheses: Haran-Ghera (1976)

MURINE MODELS OF THYMIC LYMPHOMAS

519

suggested that fractionated irradiation eliminates the bulk of the thymic lymphocytes and at the same time causes a genetic lesion in premature bone marrow cells, rendering them potentially lymphomagenic (PLCs). Migration of the PLCs to the thymus and their proliferation in order to regenerate the intrathymic lymphoid pool result in further progression to full malignancy. Lieberman et al. (1987), on the other hand, postulated that irradiation activates in the bone marrow a leukemogenic factor that is transmitted to target cells in the thymus. In both cases, the thymus provides the appropriate microenvironment in which initial premalignant events that take place in bone marrow cells can eventually be altered into a mature lymphoma. To assess the validity of these hypotheses, Boniver et al. (1981) transplanted bone marrow or thymic cells from mice previously exposed to fractionated irradiation into mice receiving a single, nonleukemogenic irradiation treatment. Because the donors were C57BL/Thy1.2 and the recipients C57BL/Thy1.1 congenic mice, the origin of the ensuing thymic lymphomas could be traced by Thy1 phenotyping. The ﬁndings indicated the presence of PLCs in the thymus 30–60 days after irradiation, which progressed to full lymphoma on transfer to syngeneic recipients. No PLCs were detected among bone marrow cells in this study. Haran-Ghera (1980b) performed similar experiments in which thymic or bone marrow cells from irradiated C57BL/6 mice were transferred to (BALB/CXC57BL/6)F1 mice, thus allowing for a distinction between donor- and recipient-type cells by H-2 phenotyping. In this study the recipient mice developed thymic lymphomas of donor origin when transfer was performed with bone marrow, but not with thymic lymphocytes, indicating the presence of PLCs in the bone marrow of the mice following their exposure to fractionated irradiation. In another study (Lieberman et al., 1987), thymectomized, irradiated C57BL/Ka/Thy1.2 mice were grafted with a thymus from a Thy1.1 congenic, neonatal mouse before or after exposure to four fractions of irradiation. The genetic origin of thymic lymphomas developing in the treated mice was analyzed by Thy1 phenotyping. The results indicated that when the mice were grafted before irradiation, most of the lymphomas were of host origin. In contrast, lymphomas in mice receiving a thymic graft after irradiation were predominantly of donor origin. These results were interpreted by the investigators as suggesting that split-dose irradiation induces production and secretion from cells in the bone marrow of a transmittable leukemogenic factor that acts on lymphocytes proliferating in the thymus, bringing about their transformation into lymphoma cells. The leukemogenic factor was not an endogenous RadLV, because neither the bone marrow nor the lymphoma cells produced a detectable amount of virus.

MURINE MODELS OF THYMIC LYMPHOMAS

521

variant (A-RadLV) that induced high-incidence thymic lymphomas when injected i.t. into adult, nonirradiated mice (Haran-Ghera, 1971). Establishment of cell lines from RadLV-induced lymphomas enabled molecular analysis of the virus. Ben-David et al. (1987a) cloned a highly leukemogenic thymotropic virus from the A-RadLV-induced lymphoma cell line 136.5 (Haas, 1974). The virus displayed a unique genomic RNA containing ecotropic and xenotropic endogenous sequences. Rassart et al. (1986) isolated a highly leukemogenic RadLV produced by the BL/VL3 cell line (Lieberman et al., 1979), whose RNA contained xenotropic sequences in the long-terminal repeat (LTR) and envelope(env) regions. It also had 43-base-pair tandem repeats in the LTR-U3 region, as well as additional point mutations (Merregaert et al., 1985; Rassart et al., 1986; Gorska-Flipot et al., 1992). It is conceivable that the unique LTR sequence of RadLV restricts the tissue tropism of the virus and that the tandem repeats function as transcriptional enhancers. Indeed, Gorska-Flipot and Joulicoeur (1990) identiﬁed a factor, termed Rad-1, that was present in T cells but not in ﬁbroblasts. The Rad-1 protein binds to a unique RNA sequence located immediately downstream of the core consensus region, which has a sequence motif in its minus-DNA strand. Rad-1 may also interact with other factors bound to the LTR core sequence. The LTR-U3 of RadLV produced by the BL/VL3 cell line induces in ﬁbroblasts synthesis of a suppressive factor that blocks the replication of RadLV and other MuLVs (Rassart et al., 1988). This ﬁnding indicates that the U3 restricts the tropism of RadLV by inducing in nonlymphoid cells a state of resistance, which interferes with virus replication following infection (Gorska-Flipot et al., 1992). The env protein plays a signiﬁcant role in the pathogenesis of certain MuLV-induced tumors. To assess the function of the RadLV env, Poliquin (1992) constructed a series of recombinant viruses, using RadLV and a nonleukemogenic endogenous virus derived from a BALB/c mouse. Infectivity assays with these variant viruses indicated that the thymotropism of RadLV is conferred by the env region of its genome. However, the BALB/c endogenous virus could be rendered thymotropic by replacing its env or LTR with those of RadLV. It was concluded, accordingly, that the thymotropism of RadLV is determined by a complementarity between the env and the LTR of the virus. Thus, the restricted infection and transformation of thymic lymphocytes by RadLV might be dependent on the formation of a particular env–LTR–target cell combination within the thymic microenvironment. Activation of cellular protooncogenes via proviral integration has been implicated in the induction of malignancies by a number of chronic retroviruses (Peters, 1990). Thus, c-myc, lck, pim-1, gin-1, and mlvi have been

522

EITAN YEFENOF

identiﬁed as common integration sites for MuLVs, and genomic rearrangement of these regions was found to correlate with development of T cell lymphomas and leukemias (Tsichlis et al., 1983, 1984; Corcoran et al., 1984; Cuypers et al., 1984; Lemay and Jolicoeur, 1984; Li et al., 1984; Selten et al., 1984; Steffen, 1984; Graham et al., 1985; O’Donnell et al., 1985; Villeneuve et al., 1986). However, integration of RadLV in the vicinity of these genetic loci in thymic lymphomas has not been demonstrated. On the other hand, Tremblay et al. (1992) detected a novel gene, designated vin-1, which underwent rearrangement by insertion of proviral RadLV in some RadLV-induced lymphomas. The vin-1 gene was later identiﬁed by Hanna et al. (1993) as coding for the cycline-D2 protein, which regulates cell cycle transition from G1 to the S phase (Hunter and Pines, 1991). These investigators suggested that constitutive activation of cycline-D2 following insertional mutagenesis may contribute to oncogenesis by driving the cells to continuous proliferation. It is unlikely, however, that cyclineD2 expression per se could account for the malignant transformation of T cells by RadLV. Proviral integration at the vin-1/cycline-D2 locus occurred in only 5% of the RadLV-induced lymphomas tested, indicating that cycline-D2 activation is neither a mandatory nor a common genetic alteration occurring during RadLV-induced leukemogenesis. The process apparently requires additional genetic events (Tremblay et al., 1992). The involvement of c-myc and mlvi in RadLV leukemogenesis was indicated by the high frequency (⬎60%) of RadLV-induced lymphomas displaying chromosome 15 trisomy (Wiener et al., 1978a). The essential segment involved in chromosome 15 duplication was localized to the distal region in which c-myc and mlvi are located (Wiener et al., 1978b). Acquisition of chromosome 15 trisomy is a late event in the genesis of T cell lymphomas and occurs independently of insertional mutagenesis or c-myc rearrangement (Wirschubsky et al., 1984c). The genetic overdose of c-myc due to chromosome 15 duplication apparently has a decisive inﬂuence on the expression of the malignant phenotype (Wirschubsky et al., 1986). The multiplicity of mechanisms presumably underlying RadLV-induced leukemogenesis suggests the existence of alternative pathways related to cellular and molecular events that are activated by primary infection with RadLV, eventually leading to the development of a common malignant phenotype. This conclusion is reinforced by data demonstrating that regardless of the eventual transforming event, RadLV-induced lymphomas stem from a population of potentially malignant cells that appear shortly after virus inoculation and persist in the thymus during the entire premalignant latency. Using a transplantation assay that distinguishes between donorand recipient-type lymphomas, Haran-Ghera (1980b) demonstrated that

MURINE MODELS OF THYMIC LYMPHOMAS

523

the thymus of a C57BL mouse inoculated with A-RadLV contains PLCs, which are detectable as early as 10 days after virus inoculation. The virally induced PLCs were characterized as premature, continuously dividing CD4⫹8⫹ thymic lymphocytes (Gokhman et al., 1990). These cells constitute the major thymic population of lymphocytes subject to positive and negative selection. We estimated the number of PLCs in the thymus of RadLV-inoculated mice by transplanting graded numbers of thymocytes from mice inoculated with RadLV into a number of recipients (Yefenof et al., 1991). The minimum number of thymocytes required to convey lymphomagenesis was 103/ mouse. Because the average number of lymphocytes in the thymus is 8 ⫻ 107, we assumed that 3 weeks after virus inoculation, the thymus contains some 앑8 ⫻ 104 ‘‘leukemogenic units,’’ each capable of initiating independent lymphoma progression when transferred to a susceptible thymus. Ben-David et al. (1987b) identiﬁed virus-producing cells in the premalignant thymus, by staining them with a monoclonal antibody directed against the RadLV envelope glycoprotein (gp70). Virus-producing cells were ﬁrst detected 10–15 hr following inoculation of the virus, their frequency steadily increasing and reaching one-third of the total thymic cell population 1–4 days later. Thereafter, the percentage of virus-positive cells declined, plummeting to 2–3% in the third week following virus inoculation. This low level of virus-positive cells remained constant during the remainder of the premalignant latency and until the outbreak of overt lymphoma, when the thymus was repopulated by lymphoma cells, all of which were infected by the virus. The ﬂuctuation in the relative number of viruspositive cells in the premalignant thymus indicates that shortly after its inoculation, RadLV infects a large population of thymic lymphocytes, the majority of which are subsequently eliminated due to intrathymic lymphocyte turnover. However, several millions of virus-infected cells are retained in the thymus over an extended period of time, constituting a pool of PLCs from which a mature thymic lymphoma eventually arises. The relatively large number of PLCs in the thymus of RadLV-inoculated mice led to the assumption that the cells constitute a pleioclonal population of T lymphocytes. The monoclonal origin of mature RadLV-induced lymphomas has been proved by detection of unique rearrangements in the T웁 gene of the T cell receptor (Ben-David et al., 1987b; Yefenof et al., 1991). To determine the clonal nature of PLCs we developed a split-transfer assay in which thymocytes from a single RadLV-inoculated mouse were injected i.t. into several syngeneic recipients (Avni et al., 1995). Donortype lymphomas were then analyzed by Southern hybridization using a T웁-speciﬁc probe. The clonal makeup of the PLCs could be inferred, as

524

EITAN YEFENOF

different T웁 rearrangements within the array of generated lymphomas reﬂected the clonal heterogeneity of the PLC progenitors at the time of transfer. These experiments revealed that when recipient mice are injected with limited numbers (103 –104) of thymic cells, explanted 3–6 weeks after virus inoculation, each of the ensuing donor-type lymphomas displays a unique and distinct T웁 rearrangement (Yefenof et al., 1991). This pattern indicates that the PLCs constitute a pleioclonal population of thymic lymphocytes, with a single cell becoming the sole progenitor of a malignant lymphoma. When the PLCs remain within the thymus of the virusinoculated mouse, a clonal lymphoma eventually develops due to the selection of a particular PLC that has progressively acquired a fully malignant phenotype. However, split injection of PLCs from a single thymus into several mice results in the independent progression of leukemogenesis in each recipient, yielding clonal lymphomas derived from different PL precursors. The long-term persistence of PLCs in the thymus during the premalignant latency of the disease required particular attention in view of the short-term residency of maturing T cells in the thymus, which does not exceed 4–5 days (Fowlkes and Pardoll, 1989; Huesmann et al., 1991). Indeed, most of the RadLV-infected cells are eliminated from the thymus within the ﬁrst week following virus infection (Ben-David et al., 1987b). However, 1–3 million virus-infected PLCs remain in the thymus, where they survive for several months. It was, therefore, postulated that infection by RadLV confers on such cells the ability to survive in the thymus for at least 12–14 weeks, the period required for the development of a fully mature lymphoma. Because RadLV-induced PLCs have been characterized as activated, continuously dividing T lymphocytes (Gokhman et al., 1990), it seemed likely that their maintenance in the thymus is facilitated by the self-renewal of a PLC pool. Haas et al. (1984) reported that early developing T cell lymphomas induced by RadLV are dependent on a continuous response to an undeﬁned growth factor. In a subsequent study we found that some, but not all, RadLV-induced lymphomas secrete interleukin-4 (IL-4), even though their growth is not dependent on the factor (Yefenof et al., 1992b). RadLV does not infect or transform T lymphocytes in vitro. However, incubation of thymic cells with RadLV resulted in IL-4 secretion that was inhibited in the presence of anti-RadLV antibodies (Yefenof et al., 1992b). Whereas untreated normal thymic cells did not secrete IL-4, thymocytes from RadLV-inoculated prelymphoma mice constitutively produced the factor, and production was enhanced by RadLV. As opposed to mature lymphomas, in vitro growth of PLCs was dependent on IL-4 (Yefenof et al., 1991). It is, therefore, conceivable that IL-4 secretion in the former

MURINE MODELS OF THYMIC LYMPHOMAS

525

is a vestige of the PLC progenitors, whose long-term survival is dependent on an IL-4-driven autocrine stimulus, because progression to the malignant state involves the selected outgrowth of a PLC that is emancipated from IL-4 dependency. Such a cell will become the progenitor of a clonal lymphoma that is no longer dependent on the factor for continuous growth. How does RadLV induce the secretion of IL-4 in a manner that is inhibited by anti-RadLV antibodies? O’Neil et al. (1987) found that RadLV cross-ligates CD3/TCR and CD4 expressed on the membrane of TH cells. These molecules are directly involved in relaying activation signals to T helper (Th) cells, recognizing an antigen presented by MHC class II molecules. On presentation to Th2 cells, an IL-4-dependent autocrine growth stimulation loop begins to operate, bringing about the clonal expansion of cells bearing receptors speciﬁc for the stimulating antigens (Fernandez-Botran et al., 1986). This ﬁnding suggests that through its binding to CD3 and CD4, RadLV mimics the action of antigen on Th2 cells and induces autocrine growth mediated by IL-4. Two basic features, however, distinguish antigen from RadLV-driven stimulation. (1) Whereas antigen activates an autocrine growth stimulation loop for clonal expansion of antigen-speciﬁc T cells, RadLV acts on a pleioclonal population of T lymphocytes. (2) In the course of an immune response to an antigen, IL-4 secretion is limited and subsides as soon as the antigen is eliminated. Autocrine activation by RadLV, on the other hand, is continuous, because the virus integrates and replicates within the infected cells. A persistent autocrine response to IL-4 enables the cells to proliferate in the thymus and maintain a population of pleioclonal PLCs, one of which is eventually transformed. This model may be construed as a modiﬁcation of the receptor-mediated lymphomagenesis model proposed by Weissman and McGrath (1982), which implies that RadLV transforms T cell clones expressing a TCR speciﬁc for the virus envelope glycoprotein. Binding of the virus or its env products to a clonotypic TCR could induce antigenic stimulation resulting in continuous proliferation. This hypothesis, however, cannot be reconciled with the need for an antigen to be processed and presented by MHC molecules of antigen-presenting cells before it can be recognized by a speciﬁc TCR (Germain, 1993). Our model proposes the interaction of the transforming retrovirus with the antigen receptor as a growth-promoting event during the premalignant phase of the disease, but excludes the elements of speciﬁc antigenic recognition of viral antigens by clones expressing a unique TCR for the viral envelope glycoprotein. VIII. Preventive Therapy of Prelymphoma Mice

The AKR inbred strain became the ﬁrst murine model for investigating the long-term progression of thymic lymphomas (Hays et al., 1989a,b,

526

EITAN YEFENOF

1990). Apart from allowing identiﬁcation of the DTV as the etiologic agent of the disease, it provided an experimental tool for studying the prospects of prophylactic therapy in protecting the mice against later development of overt lymphoma. The drastic reduction in lymphoma incidence among AKR mice that have been thymectomized at 1–2 months of age (McEndy et al., 1944), encouraged the search for additional nonsurgical means that could be applied during the latent period to rescue the mice from early death. Because DTV is the cause of thymic lymphomas in AKR mice, treatments targeting viral production or infectivity are useful in preventing lymphoma progression. Indeed, injection of a nonleukemogenic virus (24666) into young AKR mice (1–2 months old) drastically reduced the incidence of spontaneous thymic lymphomas in the adult animals (Peled and Haran-Ghera, 1991). The antilymphomagenic activity of the 24-666 virus depended on its ability to abolish the DTV expression in the thymus, which retained its intact structure and cellular composition over time. Another effective modality that protects AKR mice against lymphoma development is passive immunotherapy with antiviral antibodies (HaranGhera et al., 1995). Administration of monoclonal antibodies (18-5 on Hy-72) directed against the envelope glycoprotein of DTV, to young AKR mice (Portis et al., 1982), reduced the incidence of spontaneous lymphomas from 87 to 7% and abolished the changes in thymic cell populations characteristic of the adult animals (Haran-Ghera and Peled, 1991). Active, nonspeciﬁc immune stimulation also effectively reduced the incidence of thymic lymphomas. Thus, injection of heat-killed Lactobacillus casei, which acts as a nonspeciﬁc immunopotentiator, into 2- to 4-month-old AKR mice protected the animals against later development of lymphomas (Watanbe, 1996). The reduction in lymphoma incidence correlated with the development of antiviral immunity and a reduction in the level of DTV produced in the thymus. Another modality that provides prophylactic interference is low doses of total-body irradiation. Exposure of 2-month-old AKR mice to fractions of 15 cGy twice a week for 40 weeks reduced the incidence of lymphomas by 40% (Ishii et al., 1996). Ishii and co-workers suggested that the protective effect of total-body irradiation against lymphoma development results from potentiated immunity, previously described in chronically irradiated mice (Anderson et al., 1988). Another possibility is the direct suppressive inﬂuence of irradiation on intrathymic production of DTV, which is mandatory for the development of spontaneous lymphoma. All of the modalities that effectively reduce the incidence of spontaneous lymphomas directly or indirectly affect the level of DTV produced in the thymus or the infectivity of thymic lymphocytes by that virus. The ability to induce a prelymphoma state by viral or nonviral carcinogenesis has facilitated the study of prophylactic intervention using treatments that

MURINE MODELS OF THYMIC LYMPHOMAS

527

interfere with the persistence of PLCs or their progression to full lymphoma. As discussed in Section VI, the induction of thymic lymphomas by fractionated X-ray irradiation is preceded by the emergence of PLCs in the bone marrow. The protective effect of partial bone shielding during irradiation (Wallis et al., 1966) indicated that nonirradiated marrow cells interfere in the progression of PLCs to overt thymic lymphoma. Indeed, administration of 107 bone marrow cells from a nonirradiated syngeneic donor 2 hr after termination of X-ray irradiation was sufﬁcient to protect the mice against the development of thymic lymphomas (Boniver et al., 1981; Van Bekkum et al., 1984). When the bone marrow cells are grafted 1 month later, this maneuver fails to inhibit the emergence of the tumors (Humblet et al., 1997). Gorelick et al. (1984) suggested that NK cells residing in the bone marrow recognize and eliminate X-ray-induced PLCs. Such effector lymphocytes are radiation sensitive and are therefore destroyed by fractionated irradiation, but infusion of nonirradiated bone marrow restores this inhibitory activity. However, later studies by Lieberman et al. (1992) demonstrated that bone marrow of severe combined immune deﬁciency (SCID) mice is not protective during radiation leukemogenesis, even though the activity of natural killer (NK) cells in the SCID marrow is intact. The inability of SCID bone marrow to prevent X-ray lymphomagenesis may thus be attributed to the SCID mutation, which does not allow stem cells in the thymus to mature into immunocompetent T lymphocytes (Bosma and Carroll, 1991). Because radiation leukemogenesis requires gradual intrathymic accumulation of PLCs that originate in radiation-damaged stem cells in the bone marrow, reconstitution with nonirradiated bone marrow may interfere with this process by competing with irradiated stem cells for thymic repopulation. However, such interference can occur only with bone marrow cells capable of maturing into functional T lymphocytes, whereas stem cells defective in differentiation competence along the T lineage lack this ability. Boniver et al. (1989, 1992) found that repeated injection of TNF-움 or interferon-웂 following fractionated irradiation prevented the onset of thymic lymphomas. Administration of 2.5 ⫻ 104 units of TNF-움 or 4 ⫻ 104 units of interferon-웂, three times a week during the ﬁrst 6 weeks after irradiation, reduced the incidence of lymphomas by more than 50%. Both treatments eliminated the presence of PLCs and restored the size and cellular composition of the irradiated thymus. In addition to the direct cytotoxic effect of TNF-움 against PLCs, both TNF-움 and interferon-웂 may have acted indirectly by restoring thymic lymphopoiesis, thereby interfering in the seeding of PLCs in the thymus (Humblet et al., 1994). The protective activity of nonirradiated bone marrow infused

528

EITAN YEFENOF

immediately after split-dose irradiation may also be mediated, in part, by the TNF-움 and interferon-웂 produced by the grafted cells, because antibodies directed against these cytokines partially restored lymphoma induction (Humblet et al., 1996). Potworowski et al. (1996) showed that mice exposed to fractionated irradiation have a reduced incidence of thymic lymphomas if injected i.t. with 105 dendritic cells every ﬁfth week after termination of the radiation treatment. Transfer of thymic cells from treated to untreated mice indicated a drastic reduction in the number of PLCs following the administration of dendritic cells. The antilymphoma activity has been attributed to the ability of dendritic cells to augment speciﬁc immunity against PLCs by restoring the capacity of antigen presentation in the irradiated thymus, and to interact directly with PLCs, thereby leading to their death via apoptosis. These studies demonstrate that the development of a full-ﬂedged lymphoma may be prevented by prophylactic intervention that aborts the progression of PLCs to overt disease. The most compelling evidence that this approach is indeed appropriate comes from studies of mice inoculated with RadLV. Both the initiation and the promotion of a RadLV-induced lymphoma occurred in the thymus, the site where PLCs are generated, propagated, and progress to primary lymphoma. Hence, treatments designed to target biological properties enabling the persistence of PLCs in the thymus are presumed to be prophylactically effective. Because survival of PLCs in the thymus is dependent on IL-4-driven autocrine growth stimulation, treatments antagonizing IL-4 activity should presumably prevent lymphoma development. A suitable drug for this purpose is cyclosporin A (CSA), an immunosuppressive cyclic peptide that inhibits cytokine production by activated T cells (Schreiber and Crabtree, 1992). Because CSA suppresses T cell immunity, it is of no beneﬁt for cancer therapy, which seeks to improve, rather than to suppress, the immunocompetence of the host. However, as the intrathymic persistence of PLCs induced by RadLV is enabled by IL-4, suppressed production of the cytokine could be effective in retarding lymphoma progression. In vitro studies (Yefenof et al., 1992a) demonstrated that CSA markedly reduces the secretion of IL-4 by RadLV-induced lymphomas and PLCs. These results provided the basis for applying CSA in prophylactic intervention during the period of premalignant latency. Administration of CSA 3–6 weeks after virus inoculation signiﬁcantly delayed the onset of lymphoma, the most effective regimen being intraperitoneal (i.p.) injections of 50 mg/kg CSA twice a week for 2 weeks (Yefenof et al., 1992a). Although optimal CSA treatment prolonged the latency of the disease by 5–7 weeks, it did not prevent the onset of lymphoma.

MURINE MODELS OF THYMIC LYMPHOMAS

529

A similar effect was obtained by speciﬁc targeting of PLCs with an immunotoxin (IT) consisting of a deglycosylated ricin A chain and a monoclonal antibody directed against the RadLV envelope glycoprotein (gp70), which is expressed on the surface of virus-infected cells (Yefenof et al., 1991). Biweekly intravenous (i.v.) administration of 1 mg/kg IT during the fourth and ﬁfth week after virus inoculation extended the survival of the mice by 40–45 days (Yefenof et al., 1992b) but did not prevent lymphoma development. When, however, the two drugs were given together, they synergistically protected more than 80% of the mice against lymphoma development for up to 1 year after virus inoculation. It is, therefore, assumed that the virus-speciﬁc IT effectively eradicates a great majority, but not all, of the virus-infected PLCs, thereby delaying, but not preventing the development of lymphoma. When CSA is concomitantly administered, it arrests the growth of the few residual PLCs escaping IT treatment, whose survival is dependent on IL-4 secretion. These complementary activities prevent the progression of the disease and provide full protection to the mice. Another antagonist of IL-4 is a monoclonal antibody (11B11) that speciﬁcally neutralizes the activity of IL-4 (O’Hara and Paul, 1985). By inhibiting IL-4 binding to its receptors, 11B11 effectively and speciﬁcally curbed the growth of IL-4-dependent cells in vitro (Yefenof et al., 1991). It is also an effective immunomodulatory drug in vivo owing to its speciﬁc IL-4 antagonistic activity (Vicari and Papiernik, 1993; Cheever et al., 1994). Based on these data, we postulated that administration of 11B11 to RadLVinoculated mice could interfere with lymphoma progression. Indeed, when the premalignant mice were twice injected with 2 mg/kg 11B11 in the early latent period, a signiﬁcant delay in the onset of the disease was recorded (Yefenof et al., 1992a). Moreover, if the dose of 11B11 was increased to 20 mg/kg, two injections during the third week of latency were sufﬁcient to afford full protection, the animals remaining diseasefree for 1 year after inoculation of the virus (Epszteyn et al., 1997). The multiple procedures and relative ease with which the development of primary thymic lymphomas may be delayed and prevented allow the prediction that prophylactic therapy may be adequate and effective in various premalignant conditions, whether of a lymphoid or a nonlymphoid nature. Once the existence and unique biological properties of PLCs are identiﬁed, it should be possible to intervene in their progression to fullﬂedged disease by designing modalities that decrease the size, growth rate, and survival of the PLC population. The laboratory experience with murine thymic lymphomas suggests that this approach would be efﬁcient, reliable, and curative.

530

EITAN YEFENOF

IX. Concluding Remarks

The signiﬁcance of the thymus for the ontogeny of the immune system became apparent through investigations into its involvement in retrovirally induced oncogenesis. This discovery led to a sequence of studies that unraveled the role of the thymus as a primary lymphoid organ where bone marrow-derived progenitors proliferate, differentiate, and undergo complex selection procedures, yielding a functional population of T lymphocytes. Alongside with these developments in the immunological arena, the thymus continued to attract the attention of experimental oncologists as a preferential site for malignant transformation induced by viral, chemical, or physical carcinogens. The present article highlights several distinct pathways of murine lymphomagenesis, all leading to the development of thymic lymphomas, and outlines the unique features of the thymus, which provides an optimal microenvironment sustaining the stepwise progression of the oncogenic process, initiated by various independent agents. The long interval of latency that precedes the development of a fullblown lymphoma, and its conﬁnement to the thymus, offered an experimental tool for investigating premalignant steps in the development of the disease. Analysis of these steps led to the detection of PLCs, whose survival and progression to overt lymphoma depend on their interaction with cells and factors in the thymic microenvironment. Their characterization enabled the design and testing of prophylactic modalities targeted at various biological characteristics of PLC populations and aiming to restrict their proliferation and progression toward full malignancy. These experiments demonstrated that the ultimate outcome of lymphomagenesis may be delayed, and sometimes prevented, by applying regimens that perturb unique aspects of PLC physiology. As such, the study of experimental thymic lymphomas has generated knowledge extending beyond its relevance to a speciﬁc experimental tumor. It argues in favor of the concept that premalignant cells emerging in the early stages of tumor progression could become potential targets for prophylactic therapy at a time when it could be categorically effective. Identiﬁcation and characterization of potential malignant cells in other experimental and clinical settings may expand the scope of opportunities for chemoimmuno-prevention of malignant diseases, creating a better alternative to conventional therapy of an already existing malignancy. ACKNOWLEDGMENTS I would like to thank Drs. E. Klein, D. Weiss, and A. Mahler for their critical review of the manuscript, Ms. G. Rubinstein for library services, and Ms. D. Ben-Dov and Ms. S. Saunders for secretarial assistance. Studies performed in the author’s laboratory were sup-

MURINE MODELS OF THYMIC LYMPHOMAS

531

ported by the U.S.–Israel Binational Science Foundation, the Israel Science Foundation, and Concern Foundation, Los Angeles.

REFERENCES Anderson, G., and Jenkinson, E. (1997). Piecing together the thymic puzzle. Immunol. Today 18, 363–364. Anderson, R. E., Williams, W. L., and Tokuda, S. (1988). Effect of low-dose irradiation upon T cell subsets involved in the response of primed A/J mice to Sal cells. Int. J. Radiat. Biol. Relat. Stud. Phys. Chem. Med. 53, 103–118. Anderson, G., Moore, N. C., Owen, J. J. T., and Jenkinson, E. J. (1996). Cellular interactions in thymocyte development. Annu. Rev. Immunol. 14, 73–99. Avni, O., Yefenof, E., and Kotler, M. (1995). Clonal dominance of RadLV-induced lymphomas. Leuk. Res. 19, 971–976. Bauvois, B., Ezine, S., Imhof, B., Denoyelle, M., and Thiery, J. P. (1989). A role for the thymic epithelium in the selection of pre-T cells from murine bone marrow. J. Immunol 143, 1077–1086. Becker, F. F. (1990). Role of the AKR gene locus Akv-1 in susceptibility to chemical induction of thymic lymphomas. Cancer Res. 50, 6235–6237. Ben-David, Y., Kotler, M., and Yefenof, E. (1987a). A highly leukemogenic radiation leukemia virus isolate is a thymotropic, immunosuppressive retrovirus with a unique RNA structure. Int. J. Cancer 39, 494–497. Ben-David, Y., Yefenof, E., and Kotler, M. (1987b). Clonal analysis of radiation leukemia virus-induced leukemic and preleukemic murine cells. Cancer Res. 47, 6590–6594. Boniver, J., Decleve, A., Lieberman, M., Honsik, C., Travis, M., and Kaplan, H. S. (1981). Marrow thymus interactions during radiation leukemogenesis in C57BL/Ka mice. Cancer Res. 41, 390–392. Boniver, J., Humblet, C., and Defresne, M. P. (1989). Tumor necrosis factor and interferon 웂 inhibit the development of radiation-induced thymic lymphomas in C57BL/Ka mice. Leukemia 3, 611–613. Boniver, J., Humblet, C., Delvenne, P., Deman, J., Rongy, A. M., Greimers, R., and Defresne, M. P. (1992). TNF-움 and radiation-induced thymic lymphomas. Leukemia 3, 835–845. Bosma, M. J., and Carroll, A. M. (1991). The SCID mouse mutant: Deﬁnition, characterization and potential uses. Annu. Rev. Immunol. 9, 323–350. Boyd, R. L., and Hugo, P. (1991). Towards an integrated view of thymopoiesis. Immunol. Today 12, 71–79. Boyd, R. L., Tucek, C. L., Godfrey, D. I., Izon, D. J., Wilson, T. J., Davidson, N. J., Bean, A. G., Ladyman, H. M., Ritter, M. A., and Hugo, P. (1993). The thymic microenvironment. Immunol. Today 14, 445–459. Brathwaite, O., Bayona, W., and Newcomb, E. W. (1992). p53 mutations in C57BL/6J murine thymic lymphomas induced by 웂-irradiation and N-methylnitrosourea. Cancer Res. 52, 3791–3795. Carr, M. W., Roth, S. J., Luther, E., Rose, S. S., and Springer, T. A. (1994). Monocyte chemoattractant protein 1 acts as a T-lymphocyte chemoattractant. Proc. Natl. Acad. Sci. U.S.A. 91, 3652–3656. Champion, S., Imhof, B. A., Savagner, P., and Thiery, J. P. (1986). The embryonic thymus produces chemotactic peptides involved in the homing of hemopoietic precursors. Cell 44, 781–790. Chattopadhyay, S. K., Lander, M. R., Rands, E., and Lowy, D. R. (1980). Structure of endogenous murine leukemia virus DNA in mouse genomes. Proc. Natl. Acad. Sci. U.S.A. 77, 5774–5778.

532

EITAN YEFENOF

Cheever, A. W., Williams, M. E., Wynn, T. A., Finkelman, F. D., Seder, R. A., Cox, T. M., Hieny, S., Caspar, P., and Sher, A. (1994). Anti-IL-4 treatment of Schistosoma mansoni-infected mice inhibits development of T cells and non-B, non-T cells expressing Th2 cytokines while decreasing egg-induced hepatic ﬁbrosis. J. Immunol. 153, 753–759. Cloyd, M. W. (1983). Characterization of target cells for MCF in AKR mice. Cell 32, 217–225. Cofﬁn, J. M. (1990). Retroviridae and their replication. In ‘‘Virology’’ (B. N. Fields and D. N. Knipe, eds.), pp. 1437–1500, Raven Press, New York. Corcoran, L. M., Adams, J. M., Dunn, A. R., and Cory, S. (1984). Murine T lymphomas in which the cellular myc oncogene has been activated by retroviral insertion. Cell 37, 113–122. Corominas, M., Perucho, M., Newcomb, E. W., and Pellicer, A. (1991). Differential expression of the normal and mutated K-ras alleles in chemically induced thymic lymphomas. Cancer Res. 51, 5129–5133. Cuypers, H. T., Selten, G., Quint, W., Zijlstra, M., Maandag, E. R., Boelens, W., VanWezenheek, P., Melief, C., and Berns, A. (1984). Murine leukemia virus-induced Tcell lymphomagenesis: Integration of proviruses in a distinct chromosomal region. Cell 37, 141–150. Cuypers, H. T., Selten, G., Zijilstra, M., DeGoede, R. E., Melief, C. J., and Berns, A. J. (1986). Tumor progression in murine leukemia virus-induced T-cell lymphomas: Monitoring clonal selections with viral and cellular probes. J. Virol. 60, 230–241. Decleve, A., Sato, C., Lieberman, M., and Kaplan, H. S. (1974). Selective thymic localization of murine leukemia virus-related antigens in C57BLKa mice after inoculation with radiation virus. Proc. Natl. Acad. Sci. U.S.A. 71, 3124–3128. Decleve, A., Lieberman, M., Ihle, J. N., and Kaplan, H. S. (1977). Biological and serological characterization of the C-type RNA viruses isolated from the C57BL/Ka strain of mice. III. Characterization of the isolates and their interaction in vitro and in vivo. In ‘‘Radiationinduced Leukemogenesis and Related Viruses’’ ( J. F. Duplan, ed.), pp. 247–260. Elsevier/ North-Holland Biomedical Press, Amsterdam. Decleve, A., Lieberman, M., Ihle, J. N., Rosenthal, P. N., Lung, M. L., and Kaplan, H. S. (1978). Physiochemical, biological and serological properties of a leukemogenic virus isolated from cultured RadLV-induced lymphomas of C57BL/Ka mice. Virology 90, 23–35. Deugnier, M. A., Imhof, B. A., Bauvois, B., Dunon, D., Denoyell, M., and Thiery, J. P. (1989). Characterization of rat T cell precursors sorted by chemotactic migration toward thymotaxin. Cell 56, 1073–1083. Epzsteyn, S., Vitetta, E. S., and Yefenof, E. (1997). Prophylactic intervention in RadLVinduced lymphogenesis with an anti-IL-4 monoclonal antibody. J. Immunother. 20, 287–291. Famulari, N. G. (1983). Murine leukemia viruses with recombinant env genes: A discussion of their role in leukemogenesis. Curr. Top. Microbiol. Immunol. 103, 75–108. Fernandez-Botran, R., Sanders, V. M., Oliver, K. G., Chen, Y. W., Krammer, P. H., Uhr, J. W., and Vitetta, E. S. (1986). Interleukin 4 mediates autocrine growth of helper T cells after antigenic stimulation. Proc. Natl. Acad. Sci. U. S. A. 83, 9689–9693. Fischinger, P. J., Nomura, S., and Bolognesi, D. P. (1975). A novel murine oncornavirus with dual eco- and xenotropic properties. Proc. Natl. Acad. Sci. U.S.A. 72, 5150–5155. Fischinger, P. J., Thiel, H. J., Ihle, J. N., Lee, J. C, and Elder, J. H. (1981). Detection of a recombinant murine leukemia virus-related glycoprotein on virus negative thymoma cells. Proc. Natl. Acad. Sci. U.S.A. 78, 1920–1924.

MURINE MODELS OF THYMIC LYMPHOMAS

533

Fischinger, P. J., Thiel, H. J., Lieberman, M., Kaplan, H. S., Dunlop, N. M., and Robey, W. G. (1982). Presence of a novel recombinant murine leukemia virus-like glycoprotein on the surface of virus-negative C57BL lymphoma cells. Cancer Res. 42, 4650–4657. Fowlkes, B. J., and Pardoll, D. M. (1989). Molecular and cellular events of T cell development. Adv. Immunol. 44, 207–264. Fowlkes, B. J., and Schweighoffer, E. (1995). Positive selection of T cells. Curr. Opin. Immunol. 7, 188–195. Fredrickson, T. N., Morse, H. C., III, and Rose, W. P. (1984). Spontaneous tumors of NFS mice congenic for ecotropic murine leukemia virus induction loci. J. Natl. Cancer Inst. 73, 521–524. Frei, J. V. (1980). Methylnitrosourea induction of thymomas in AKR mice requires one or two ‘‘hits’’only. Carcinogenesis 1, 721–723. Frei, J. V., and Lawley, P. D. (1980). Thymomas induced by simple alkylating agents in C57BL/Cbi mice: Kinetics of the dose response. J. Natl. Cancer Inst. 64, 845–856. Germain, R. N. (1993). Antigen processing and presentation. In ‘‘Fundamental Immunology’’ (W. E. Paul, ed.), 3rd ed., pp. 629–676. Raven Press, New York. Gokhman, I., Peled, A., and Haran-Ghera, N. (1990). Characteristics of potential lymphomainducing cells in mice sensitive or resistant to lymphomagenesis by Radiation Leukemia Virus variants. Cancer Res. 50, 2554–2561. Gorelik, E., Rosen, B., Copeland, D., Weatherly, B., and Herberman, R. B. (1984). Evaluation of the role of natural killer cells in radiation-induced leukemogenesis in mice. J. Natl. Cancer Inst. 72, 1397–1403. Gorska-Flipot, I., and Jolicoeur, P. (1990). DNA-binding proteins that interact with the long terminal repeat of radiation leukemia virus. J. Virol. 64, 1566–1572. Gorska-Flipot, I., Huang, M., Cantin, M., Rassart, E., Masse, G., and Jolicoeur, P. (1992). U3 long terminal repeat-mediated induction of intracellular immunity by a murine retrovirus: A novel model of latency for retrovirusesJ. Virol. 66, 7201–7210. Graham, M., Adams, J. M., and Cory, S. (1985). Murine T lymphomas with retroviral inserts in the chromosomal 15 locus for plasmacytoma variant translocations. Nature (London) 314, 740–743. Gross, L. (1951). Spontaneous leukemia developing in C3H mice following inoculation in infancy with AK leukemic extracts or AK embryos. Proc. Soc. Exp. Biol. Med. 76, 27–32. Gross, L. (1978). Viral etiology of cancer and leukemia: A look into the past, present and future. Cancer Res. 3, 485–493. Gross, L. (1980). The search for viruses as etiological agents in leukemia and malignant lymphomas: The role of the happy accident and the prepared mind. Cancer Res. 40, 3405– 3407. Guerrero, I., and Pellicer, A. (1987). Mutational activation of oncogenes in animal model systems of carcinogenesis. Mutat. Res. 185, 293–308. Haas, M. (1974). Continuous production of radiation leukemia virus in C57BL thymoma tissue culture lines: Puriﬁcation of the leukemogenic virus. Cell 1, 79–83. Haas, M., Altman, A., Rothenberg, E., Bogart, M. H., and Jones, O. W. (1984). Mechanism of T-cell lymphomagenesis: Transformation of growth-factor-dependent T-lymphoblastoma cells to growth-factor-independent T-lymphoma cells. Proc. Natl. Acad. Sci. U.S.A. 81, 1742–1746. Hanna, Z., Jankowski, M., Tremblay, P., Jiang, X., Milatovich, A., Francke, J., and Jolicoeur, P. (1993). The vin-1 gene, identiﬁed by provirus insertional mutagenesis, is the cyclin D2. Oncogene 8, 1661–1666. Haran-Ghera, N. (1966). Leukemogenic activity of centriguates from irradiated mouse thymus and bone marrow. Int. J. Cancer 1, 81–87.

534

EITAN YEFENOF

Haran-Ghera, N. (1971). Inﬂuence of host factors on leukemogenesis by the radiation leukemia virus. Isr. J. Med. Sci. 7, 17–25. Haran-Ghera, N. (1976). Pathways in murine leukemogenesis. In ‘‘Biology of Radiation Carcinogenesis’’ ( J. M. Yuhas, R. W. Tennant, and J. D. Regan, eds.), pp. 245–252. Raven Press, New York. Haran-Ghera, N. (1980a). Potential leukemic cells among bone marrow cells of young AKR/ J mice. Proc. Natl. Acad. Sci. U.S.A. 77, 2923–2926. Haran-Ghera, N. (1980b). Pathogenesis of murine leukemia. In ‘‘Viral Oncology’’ (G. Klein, ed.), pp. 161–185. Raven Press, New York. Haran-Ghera, N. (1994). Lymphomagenesis in AKR mice: B cell lymphomas as a model of tumor dormancy. Adv. Cancer Res. 63, 245–293. Haran-Ghera, N., and Peled, A. (1991). Prevention of spontaneous AKR T cell lymphomagenesis by elimination of potential lymphoma cells with antibody to speciﬁc gp 71 determinants. Virology 181, 536–540. Haran-Ghera, N., Lieberman, M., and Kaplan, H. S. (1966). Direct action of a leukemogenic virus on the thymus. Cancer Res. 26, 438–442. Haran-Ghera, N., Peled, A., Leef, F., Hoffman, A. D., and Levy, A. J. (1987). Enhanced AKR leukemogenesis by the dual tropic viruses. I. The time and site of origin of potential leukemic cells. Leukemia 1, 442–449. Haran-Ghera, N., Peled, A., Wu, L., Shortman, K., Brightman, B. K., and Fan, H. (1995). The effects of passive antiviral immunotherapy in AKR mice: I. The susceptibility of AKR mice to spontaneous and induced T cell lymphomagenesis. Leukemia 9, 1199–1206. Hartley, J. W., Wolford, M. K., Old, L. J., and Rowe, W. P. (1977). A new class of murine leukemia virus associated with the development of spontaneous lymphomas. Proc. Natl. Acad. Sci. U.S.A. 74, 789–792. Hays, E. F. (1982). Bone marrow progenitor cells of AKR mice give rise to thymic lymphoma cells. Leuk. Res. 6, 429–432. Hays, E. F., Margaretten, N., and Swanson, S. K. (1982). Spontaneous leukemia viruses: Lymphomagenic ecotropic viruses of AKR mice. J. Natl. Cancer Inst. 69, 1077–1082. Hays, E. F., Bristol, G. C., Lugo, J. P., and Wang, X. F. (1989a). Progression to development of lymphoma in the thymus of AKR mice treated neonatally with SL 3-3 virus. Exp. Hematol. 17, 1116–1121. Hays, E. F., Bristol, G. C., McDougall, S., Klotz, S. L., and Kronenberg, M. (1989b). Development of lymphoma in the thymus of AKR mice treated with the lymphomagenic virus SL3-3. Cancer Res. 49, 4225–4230. Hays, E. F., Bristol, G., and McDougall, S. (1990). Mechanisms of thymic lymphomagenesis by the retrovirus SL3-3. Cancer Res. 50 5631s–5635s. Hays, E. F., Swanson, S. K., Hale, L., and Margaretten, N. (1984). Thymic stroma in AR mice: Its function and virus production. Leuk. Res. 8, 637–645. Herr, W., and Gilbert, W. (1983). Somatically acquired recombinant murine leukemia proviruses in thymic leukemias of AKR/J mice. J. Virol. 46, 70–82. Huesmann, M., Scott, B., Kisielow, P., and Von-Boehmer, H. (1991). Kinetics and efﬁcacy of positive selection in the thymus of normal and T cell receptor transgenic mice. Cell 66, 533–540. Humblet, C., Deman, J., Rongy, A. M., Greimers, R., Boniver, J., and Defresne, M. P. (1994). TNF-움 is involved in the mechanism of murine thymic lymphoma prevention by bone marrow grafting. In ‘‘in Vivo Immunology’’ (E. Heiner et al., eds.), pp. 195–199, Plenum, New York. Humblet, C., Greimers, R., Delvenne, P., Deman, J., Boniver, J., and Defresne, M. P. (1996). Prevention of murine radiogenic thymic lymphomas by tumor necrosis factor or by marrow grafting. J. Natl. Cancer Inst. 12, 824–831.

MURINE MODELS OF THYMIC LYMPHOMAS

535

Humblet, C., Greimers, R., Boniver, J., and Defresne, M. P. (1997). Stages in the development of radiation-induced thymic lymphomas in C57BL/Ka mice: Preleukemic cells become progressively resistant to the tumor preventing effects of a bone marrow graft. Exp. Hematol. 25, 109–113. Hunter, T., and Pines, J. (1991). Cyclins and cancer. Cell 66, 1071–1074. Ignatowicz, L., Kappler, J., and Marrack, P. (1996). The repertoire of T cells shaped by a single MHC/peptide ligand. Cell 84, 521–529. Ihle, J. N., Jozeph, D. R., and Pazmino, N. H. (1976a). Radiation leukemia in C57BL/6 mice. II. Lack of ecotropic virus expression in the majority of lymphomas. J. Exp. Med. 144, 1406–1423. Ihle, J. N., McEwan, R., and Bengali, K. (1976b). Radiation leukemia in C57BL/6 mice. I. Lack of serological evidence for the role of endogenous ecotropic viruses in pathogenesis. J. Exp. Med. 144, 1391–1405. Ikuta, K., Uchida, N., Friedman, J., and Weissman, I. L. (1992). Lymphocyte development from stem cells. Annu. Rev. Immunol. 10, 759–783. Ilbery, P. L. T. (1967). Prevention of radioleukemia by lymph node shielding. Nature (London) 215, 656–657. Imhof, B. A., Deugnier, M. A., Girault, J. M., Champion, S., Damais, C., Itoh, T., and Thiery, J. P. (1988). Thymotaxin: A thymic epithelial peptide chemotactic for T-cell precursors. Proc. Natl. Acad. Sci. USA 85, 7699–7703. Ishii, K., Hosoi, Y., Yamada, S., Ono, T., and Sakamoto, K. (1996). Decreased incidence of thymic lymphoma in AKR mice as a result of chronic, fractionated low-dose total-body X irradiation. Radiat Res. 146, 582–585. Jameson, S. C., Hogquist, K. A., and Bevan, M. (1995). Positive selection of thymocytes. Annu. Rev. Immunol. 13, 93–126. Janeway, Jr., C. A., Kupfer, A., Viret, C., Boursalian, T., Goverman, J., Bottomly, K., and Sant’Angelo, D. (1998). T-cell development, survival and signaling: A new concept of the role of self-peptide: Self-MHC complexes. Immunologist 6, 5–12. Janowski, M., Merregaert, J., and Reddy, P. (1986). Retroviruses in radiation-induced lymphomas. Leuk. Res. 10, 833–842. Janowski, M., Cox, R., and Strauss, P. G. (1990). The molecular biology of radiation-induced carcinogenesis: Thymic lymphoma, myeloid leukemia and osteosarcoma. Int. J. Radiat. Biol. 57, 677–691. Jolicoeur, P., and Rassart, E. (1980). Effect of Fv-1 gene product on synthesis of linear and supercoiled viral DNA in cells infected with murine leukemia virus. J. Virol. 33, 183–195. Kaplan, H. S. (1961). The role of cell differentiation as a determinant of susceptibility to virus carcinogenesis. Cancer Res. 21, 981–988. Kaplan, H. S. (1967). On the natural history of the murine leukemias. Cancer Res. 27, 1325– 1340. Kaplan, H. S. (1974). Leukemia and lymphoma in experimental and domestic animals. Ser. Haematol. 7, 94–163. Kaplan, H. S., and Brown, M. B. (1952). A quantitative dose–response study of lymphoid tumor development in irradiated mice. J. Natl. Cancer Inst. 13, 185–190. Kaplan, H. S., and Brown, M. B. (1954). Development of lymphoid tumors in non-irradiated thymic grafts in thymectomized irradiated mice. Science 119, 439–441. Kotler, M., Ruchlemer, R., Avni, O., and Yefenof, E. (1994). Radiation leukemogenesis: Quantitative relationship between pre-leukemic cells in the thymus and lymphoma induction. Int. J. Cancer 56, 761–765. Kubota, Y., Murakami-Murofushi, K., Shimada, Y., Ogiu, T., and Oikawa, T. (1995). Reduced ﬁdelity of DNA synthesis in cell extracts from chemically induced primary thymic lymphomas of mice. Cancer Res. 55, 3777–3780.

536

EITAN YEFENOF

Lemay, G., and Jolicoeur, P. (1984). Rearrangement of a DNA sequence homologous to a cell-virus junction fragment in several Moloney murine leukemia virus-induced rat thymomas. Proc. Natl. Acad. Sci. U.S.A. 81, 38–42. Li, Y., Holland, C. A., Hartley, J. W., and Hopkins, N. (1984). Viral integration near cmyc in 10–20% of MCF-247-induced AKR lymphomas. Proc. Natl. Acad. Sci. U.S.A. 81, 6808–6811. Lieberman, M., and Kaplan, H. S. (1959). Leukemogenic activity of ﬁltrates from radiationinduced lymphoid tumors of mice. Science 130, 387–388. Lieberman, M., Kaplan, H. S., and Decleve, A. (1976). Anomalous virus expression in radiogenic lymphomas of C57BL/Ka mice. In ‘‘Biology of Radiation Carcinogenesis’’ ( J. M. Yuhas, R. W. Tennant, and J. D. Regan, eds), Raven Press, New York, pp. 237–258. Lieberman, M., Decleve, A., and Kaplan, H. S. (1978). Rapid in vitro assay for thymotropic, leukemogenic murine C-type RNA viruses. Virology 90, 274–278. Lieberman, M., Decleve, A., Ricciardi-Castagnoli, P., Boniver, J., Finn, O. J., and Kaplan, H. S. (1979). Establishment, characterization and virus expression of cell lines derived from radiation- and virus-induced lymphomas of C57BL/Ka mice. Int. J. Cancer 24, 168–177. Lieberman, M., Hansteen, G. A., McCune, J. M., Scott, M. L., White, J. H., and Weissman, I. L. (1987). Indirect induction of radiation lymphomas in mice: Evidence of a novel, transmissible leukemogen. J. Exp. Med. 166, 1883–1993. Lieberman, M., Hansteen, G. A., Waller, E. K., Weissman, I. L., and Sen-Majumdar, A. (1992). Unexpected effects of the severe combined immunodeﬁciency mutation on murine lymphomagenesis. J. Exp. Med. 176, 399–405. Lilly, F., and Pincus, T. (1973). Genetic control of murine viral leukemogenesis. Adv. Cancer Res. 17, 231–277. Manteuil-Brutlag, S., Liu, S. L., and Kaplan, H. S. (1980), Radiation leukemia virus contains two distinct viral RNAs. Cell 19, 643–652. McCann, J. (1998), Infections and cancer: Viruses are still prime suspects [news]. J. Natl. Cancer Inst. 90, 418–420. McEndy, D. P., Boom, M. P., and Furth, J. (1944). On the role of thymus, spleen and gonads in the development of leukemia in high leukemic strains of mice. Cancer Res. 4, 377–383. Merregaert, J., Nuyten, J. M., and Janowski, M. (1985). Nucleotide sequence of the envelope gene of radiation leukemia virus. Virology 144, 457–467. Miller, J. F. A. P. (1961a). Analysis of the thymus inﬂuence in leukemogenesis. Nature (London) 191, 248–249. Miller, J. F. A. P. (1961b). Immunological function of the thymus. Lancet 2, 748–749. Newcomb, E. W., Steinberg., J. J., and Pellicer, A. (1988). Ras oncogenes and phenotypic staging in N-methylnitrosourea and gamma-irradiation-induced thymic lymphomas in C57BL/6J mice. Cancer Res. 48, 5514–5521. Newcomb, E. W., Pellicer, A., and Cordon, C. (1990). Comparative analysis and anatomic distribution of Ras p21, IL-2R, and MEL-14 in malignant and hyperplastic murine thymus. Am. J. Pathol. 136, 307–317. Nusse, R. (1986). The activation of cellular oncogenes by retroviral insertion. Trends Genet. 3, 244–247. O’Donnell, P. V., Stockert, E., Obata, Y., and Old, L. J. (1981). Leukemogenic properties of AKR dual tropic (MCF) viruses: Ampliﬁcation of murine leukemia virus-related antigens on thymocytes and acceleration of leukemia development in AKR mice. Virology 112, 548–563. O’Donnell, P. V., Woller, R, and Chu, A. (1984). Stages in development of mink cell focusinducing (MCF) virus-accelerated leukemia in AKR mice. J. Exp. Med. 160, 914–934.

MURINE MODELS OF THYMIC LYMPHOMAS

537

O’Donnell, P. V., Fleissner, E., Lonial, H., Koehne, C. F., and Reicin, A. (1985). Early clonality and high-frequency proviral integration into the c-myc locus in AKR leukemias. J. Virol. 55, 500–503. O’Hara, J., and Paul, W. E. (1985), Production of a monoclonal antibody to and molecular characterization of B-cell stimulatory factor-1. Nature (London) 315, 333–336. O’Neill, H. C., Allison, J. P., McGrath, M. S., and Weissman, I. L. (1987). A subset of T cell receptors associated with L3T4 molecules mediates C6VL leukemia cell binding of its cognate retrovirus. Cell 49, 143–151. Pedersen, F. S., Crowther, R. L., Tenney, D. Y., Reimold, A. M., and Haseltine, W. A. (1981). Novel leukemogenic retroviruses isolated from a cell line derived from a spontaneous AKR tumor. Nature (London) 292, 167–170. Peled, A., and Haran-Ghera, N. (1969). Prevention of radiation leukemia by peripheral blood lymphocytes. Isr. J. Med. Sci. 5, 12–13. Peled, A., and Haran-Ghera, N. (1985). High incidence of B cell lymphomas derived from thymectomized AKR mice expressing TL. 4 antigen. J. Exp. Med. 162, 1081–1086. Peled, A., and Haran-Ghera, N. (1991). Prevention of spontaneous AKR T cell lymphomagenesis by 24-666, a virus isolated from an AKR B cell lymphoma. Virology 181, 528–535. Peters, G. (1990). Oncogenes at viral integration sites. Cell Growth Differ. 1, 503–510. Poliquin, L., Bergeron, D., Fortier, J. L., Paquette, Y., Bergeron, R., and Rassart, E. (1992). Determinants of thymotropism in Kaplan radiation leukemia virus and nucleotide sequence of its envelope region. J. Virol. 66, 5141–5146. Portis, J. L., McAtee, F. J., and Cloyd, M. W. (1982). Monoclonal antibodies to xenotropic and MCF murine leukemia viruses derived during the graft-versus-host reaction. Virology 118, 181–190. Potworowski, E. F., Gagnon, F., Beauchemin, C., and St. Pierre, Y. (1996). Dendritic cells prevent radiation-induced thymic lymphoma. Leukemia 10, 1639–1647. Rassart, E., Shang, M., Boie, Y., and Jolicoeur, P. (1986). Studies of emerging radiation leukemia virus variants in C57BL/Ka mice. J. Virol. 58, 96–106. Rassart, E., Paquette, Y., and Jolicoeur, P. (1988). Inability of Kaplan radiation leukemia virus to replicate on mouse ﬁbroblasts is conferred by its long terminal repeat. J. Virol. 62, 3840–3848. Richie, E. R., Nairn, R. S., and Becker, F. F. (1985). Proviral structure and differentiation antigen phenotype of spontaneous and chemically induced AKR lymphomas. Cancer Res. 45, 2802–2806. Richie, E. R., Angel, J. M., and McEntrie, B. (1988). N-Methyl-N-nitrosourea-induced Tlymphomas of AKR/J mice contain somatically acquired ecotropic-like murine leukemia proviruses. Virology 164, 211–219. Richie, E. R., Angel, J. M., and Cloyd, M. W. (1991). Inﬂuence of murine leukemia proviral integrations on development of N-methyl-N-nitrosourea-induced thymic lymphomas in AKR mice. J. Virol. 65, 5751–5756. Rowe, W. P. (1972). Studies of genetic transmission of murine leukemia virus by AKR mice. J. Exp. Med. 136, 1272–1285. Rowe, W. P. (1978). Leukemia virus genomes in the chromosomal DNA of the mouse. Harvey Lect. 71, 173–192. Schreiber, S. L., and Crabtree, G. R. (1992). The mechanism of action of cyclosporin A and FK506. Immunol. Today 13, 136–142. Scollay, R. (1991). T-cell subset relationships in thymocyte development. Curr. Opin. Immunol. 3, 204–209. Scollay, R., and Godfrey, D. I. (1995). Thymic emigration: Conveyor belt or lucky dips. Immunol. Today. 16, 268–273.

538

EITAN YEFENOF

Selten, G., Cuypers, H. T., Zijlstra, M., Melief, C., and Berns, A. (1984). Involvement of c-myc in MuLV-induced T cell lymphomas in mice: Frequency and mechanisms of activation. EMBO J. 3, 3215–3222. Staal, S. P., and Hartley, J. W. (1988). Thymic lymphoma induction by the AKT8 murine retrovirus. J. Exp. Med. 167, 1259–1264. Steffen, D. (1984). Proviruses are adjacent to c-myc in some murine leukemia virus-induced lymphomas. Proc. Natl. Acad. Sci. U.S.A. 81, 2097–2101. Takeuchi, H., Kato, A., and Hays, E. F. (1984). Presence of prelymphoma cells in the bone marrow of the lymphomagenic virus-treated AKR mouse. Cancer Res. 44, 1008–1011. Teich, N. (1982). Taxonomy of retroviruses. In ‘‘RNA Tumor Viruses’’ (R. Weiss, N. Teich, H. Varmus, and J. Cofﬁn, eds.), pp. 25–207. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Teich, N., Wyke, J., Mak T., Bernstein, A., and Hardy, W. (1983). Pathogenesis of retrovirus. In ‘‘Molecular Biology of Tumor Viruses’’ (N. Teich, R. Weiss, H. Varmus, and J. Cofﬁn, eds.), pp. 785–998. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Tremblay, P. J., Kozak, C. A., and Jolicoeur, P. (1992). Identiﬁcation of a novel gene, Vin1, in murine leukemia virus-induced T-cell leukemias by provirus insertional mutagenesis. J. Virol. 66, 5176–5182. Tsichlis, P. N., and Lazo, P. A. (1991). Virus–host interactions and the pathogenesis of murine and human oncogenic retroviruses. Curr. Top. Microbiol. Immunol. 171, 95–171. Tsichlis, P. N., Strauss, P. G., and Hu, L. F. (1983). A common region for proviral DNA integration in MoMuLV-induced rat thymic lymphomas. Nature (London) 302, 445–449. Tsichlis, P. N., Strauss, P. G., and Kozak, C. A. (1984). Cellular DNA region involved in induction of thymic lymphomas (Mlvi-2) maps to mouse chromosome 15. Mol. Cell. Biol. 4, 997–1000. Van-Bekkum, D. W., Boersma, W. J., Eliason, J. F., and Knaan, S. (1984). The role of prothymocytes in radiation-induced leukemogenesis in C57BL/Rij mice. Leuk. Res. 8, 461–471. Van-Ewijk, W. (1991). T-Cell differentiation is inﬂuenced by thymic microenvironments. Annu. Rev. Immunol. 9, 591–615. Van-Ewijk, W., Shores, E. W., and Singer, A. (1994). Crosstalk in the mouse thymus. Immunol. Today 15, 214–217. Varmus, H. E. (1984). The molecular genetics of cellular oncogenes. Annu. Rev. Genet. 18, 553–612. Varmus, H. E. (1988). Retroviruses. Science 240, 1427–1435. Vicari, A., and Papiernik, M. (1993), Multiple effects caused by anti-IL-4 mAb inoculation in the thymus and spleen of adult mice. Eur. Cytokine Netw. 4, 111-119. Villeneuve, L., Rassart, E., Jolicoeur, P., Graham, M., and Adams, J. M. (1986). Proviral integration site Mis-1 in rat thymomas corresponds to the pvt-1 translocation breakpoint in murine plasmacytomas. Mol. Cell. Biol. 6, 1834–1837. Von-Boehmer, H. (1990). Developmental biology of T cells in T cell-receptor transgenic mice. Annu. Rev. Immunol. 8, 531–556. Wallis, V., Davies, A. J., and Koller, P. C. (1966). Inhibition of radiation-induced leukaemia by the injection of haematopoietic tissue: A study of chimerism. Nature (London) 210, 500–504. Warren, W., Lawley, P. D., Gardner, E., Harris, G., Ball, J. K., and Cooper, C. W. (1987). Induction of thymomas by N-methyl-N-nitrosourea in AKR mice: Interaction between the chemical carcinogen and endogenous murine leukaemia viruses. Carcinogenesis (London) 8, 163–172.

MURINE MODELS OF THYMIC LYMPHOMAS

539

Watanabe, T. (1996). Suppressive effects of Lactobacillus casei cells, a bacterial immunostimulant, on the incidence of spontaneous thymic lymphoma in AKR mice. Cancer Immunol. Immunother. 42, 285–290. Weinberg, R. A. (1996). Very special cancer viruses. In ‘‘Racing to the Beginning of the Road: The Search for the Origin of Cancer,’’ pp. 66–84. Harmony Books, New York. Weissman, I. L., and McGrath, M. S. (1982). Retrovirus lymphoma-genesis: Relationship of normal immune receptors to malignant cell proliferation. Curr. Top. Microbiol. Immunol. 98, 103–112. Wiener, F., Ohno, S., Spria, J., Haran-Ghera, N., and Klein, G. (1978a). Chromosome changes (trisomies #15 and 17) associated with tumor progression in leukemias induced by radiation leukemia virus. J. Natl. Cancer Inst. 61, 227–237. Wiener, F., Ohno, S., Spira, J., Haran-Ghera, N., and Klein, G. (1978b). Cytogenetic mapping of the trisomic segment of chromosome 15 in murine T-cell leukaemia. Nature (London) 275, 658–660. Wirschubsky, Z., Wiener, F., Bregula, U., and Klein, G. (1984a). Further studies on the asymmetry of chromosome 15 duplication in trisomic leukemias of heterozygous origin: Preferential status of the AKR chromosome. Int. J. Cancer 34, 249–254. Wirschubsky, Z., Wiener, F., Spira, J., Sumegi, J., and Klein, G. (1984b). Triplication of one chromosome No. 15 with an altered c-myc containing EcoRI fragment and elimination of the normal homologue in a T-cell lymphoma line of AKR origin. Int. J. Cancer 33, 477–481. Wirschubsky, Z., Wiener, F., Spira, J., Sumegi, J., Perlmann, C., and Klein, G. (1984c). The role of speciﬁc chromosomal translocations and trisomies in the genesis of certain plasmacytomas and lymphomas in mice, rats and humans. Cancer Cells 2, 253–260. Wirschubsky, Z., Tsichlis, P., Klein, G., and Sumegi, J. (1986). Rearrangement of c-myc, pim-1 and Mlvi-1 and trisomy of chromosome 15 in MCF- and Moloney–MuLV-induced murine T-cell leukemias. Int. J. Cancer 38, 739–745. Yanagihara, K., Hamada, K., Seyama, T., Imamura, N., and Yokoro, K. (1982). In vitro studies of the mechanism of leukemogenesis. II. Characterization of endogenous murine leukemia viruses isolated from AKR thymic epithelial reticulum cell lines. J. Virol. 41, 360–366. Yefenof, E. (1980). In vitro generation of cytotoxic lymphocytes against radiation and radiation leukemia virus (RadLV) induced tumors. II. A radiation induced thymoma generates cytotoxic response in syngeneic but not in allogeneic lymphocytes. Immunology 41, 99–106. Yefenof, E., Meidav, A., and Kedar, E. (1980a). In vitro generation of cytotoxic lymphocytes against radiation and radiation leukemia virus (RadLV) induced tumors. III. Suppression of anti-tumor immunity in vitro by lymphocytes of mice undergoing radiation leukemia virus induced leukemogenesis. J. Exp. Med. 152, 1473–1483. Yefenof, E., Tchakirov, R., and Kedar, E. (1980b). In vitro generation of cytotoxic lymphocytes against radiation and radiation leukemia virus (RadLV) induced tumors. I. Role of viral antigenicity. Cancer Immunol. Immunother. 8, 171–178. Yefenof, E., Abboud, G., Epszteyn, S., and Vitetta, E. S. (1992a). Treatment of premalignancy: Prevention of lymphoma in radiation leukemia virus-inoculated mice by cyclosporin A and immunotoxin. Proc. Natl. Acad. Sci. U.S.A. 89, 728–732. Yefenof, E., Ela, C., Kotler, M., and Vitetta, E. S. (1992b). Induction of IL-4 by the Radiation Leukemia Virus (RadLV): Role in autocrine growth stimulation of RadLV infected preleukemic cells. Int. J. Cancer 50, 481–485. Yefenof, E., Epsztein, S., and Kotler, M. (1991). Quantitation, in vitro propagation, and characterization of preleukemic cells induced by radiation leukemia virus. Cancer Res. 51, 2179–2184.

520

EITAN YEFENOF

Lieberman et al. (1987) also maintained that the variance between their study and the study of Haran-Ghera (1980b), regarding involvement of the bone marrow in lymphoma progression, was attributable to the experimental systems used, namely, transfer of parental cells into F1 hybrid recipients versus Thy1.1 cells into Thy1.2 congenic mice. Kotler et al. (1994) transferred graded numbers of thymic lymphocytes from an irradiated mouse into syngeneic recipients at various intervals following termination of the leukemogenic treatment (four doses of 1.7 Gy). By recording the frequency of the thymic lymphomas developing in the recipient mice they estimated that PLCs ﬁrst appear in the thymus 6 weeks after irradiation and 9–15 weeks prior to the development of overt lymphoma. The initial proportion of PLCs in the thymus is ⱖ10⫺5; their frequency continously increases with time, reaching ⱖ10⫺3 at 10 weeks after termination of irradiation. These results indicate that the transition of the leukemogenic process from the bone marrow to the thymus is lengthy and gradual, whether manifested by the migration of PLCs or by the transfer of a transmittable leukemogen. A common conclusion that can be drawn from these studies is that the initial effect of fractionated irradiation is the emergence of PLCs, which appear in the bone marrow several months before the thymus is populated with lymphoma cells. The existence of PLCs for an extended period of time during the latent phase of the disease provided an experimental system in which it was possible to examine the effect of prophylactic intervention on lymphoma progression, by targeting the cells producing a leukemogenic factor or by interference in the transfer of PLCs from the bone marrow to the thymus. These studies are discussed in Section VIII. VII. RadLV-Induced Lymphomagenesis

The existence of thymic PLCs in the early stages of lymphomagenesis has been most persuasively demonstrated in mice inoculated with RadLV. RadLV was isolated from an X-ray-induced thymic lymphoma of a C57BL/ Ka mouse (Lieberman and Kaplan, 1959). The ﬁrst isolate was weakly leukemogenic, but serial passage of the virus in newborn mice resulted in the selection of a highly leukemogenic variant that induced thymic lymphomas when inoculated i.t. into adult mice (Kaplan, 1967). Another RadLV variant was isolated from bone marrow cells of a C57BL/6 mouse that had been exposed to fractionated irradiation (Haran-Ghera, 1966). This virus induced a high incidence of thymic lymphomas in mice receiving a single, nonleukemogenic dose of 4 Gy irradiation. Accordingly, it was designated radiation-dependent RadLV (D-RadLV). Repeated passage of D-RadLV in C57BL/6 mice yielded a highly leukemogenic, autonomous

This Page Intentionally Left Blank

INDEX

A

aggregated proteins, 49–50 artiﬁcial particles, 32–34 bacteria, 26–32 cell debris, 34–36 cellular sonicates, 50–51 complex quantitating, 17–19 denatured proteins, 49–50 endogenous antigens in vivo, 56–63 immunogenicity, 24–26 inactivated viruses, 23–24 ISCOM, 42–44 liposome, 42–44 mechanism, 19–23 native proteins, 44–45 pathways, 26–32 processing, 19–20 quantum theory, 51–52 receptor-mediated uptake, 46–49 srivastava saga, 52–56 viral, 32–38 virosome, 42–44 virus-like particles, 38–39 Ki67, expression dividing T cell humans, healthy, 316–319 humans, HIV-infected, 319–320 oral biologic ﬁltration, 188–189 dosage, 196–197 enzymatic digestion, 188–189 feeding regimes, 196–197 immunological effects, 153–155 induction factors, 183–187 inﬂammatory processes, 194–195 intestinal absorption, 189–194 Peyer’s patches, 197–200 structure, 187–188 proteinaceous, Th1 response, 348

Actinomycin D, 447 Acute distress respiratory syndrome, 426 Adenoviruses vectors, advantages, 62 Adjuvant arthritis, 213–214 Adult respiratory distress syndrome, 384 Age, oral tolerance, 178–181 AIDS, FasL role, 282–283 AKV-1 gene, 517 Alcohol-induced hepatitis, 286 Allergies animal models, 218, 221, 223 human models, 227–228, 231 Allografts, maternal donor, 229 Anergy, oral tolerance, 172–174 Antigen presenting cells anergy, 173–174 CpG effects, 336–338 development, 1–2 mucosal immune response, 192–193 oral tolerance induction, 161 pAPCS ablating in vivo, 23–24 generation, 14–15 types DCs, 15–16 Møs, 15 virosome presentation, 43 Antigens, see also speciﬁc antigens CTB, 182–183 endogenous MHC generation, 4–9 presentation in vivo, 56–63 exogenous internalization modes, 12–14 presentation ablating pAPCs in vivo, 23–24 541

542

INDEX

Antineutrophil cytoplasmic antibodies, 403 Antiphospholipid syndrome, 221 APC, see Antigen presenting cells Apoptosis, see Cell death Arthritis, see also Rheumatoid arthritis collagen-induced, 272 rheumatoid, see Rheumatoid arthritis septic, 460–461 Asthma CpG DNA role, 350–351 IL-8 production, 427 Autoimmune diseases, see also speciﬁc diseases DR3 production, 286–287 oral tolerance animal models, 218–223 humans, 227–228 worsening, 223–224 TNF-움 implications, 275–277 TRAIL production, 286–287 AXB cells, 59

B 웁2m, 40–41 Bacteria, see also speciﬁc bacterium artiﬁcial particles, 32–34 DNA, CpG motifs as antitumor adjuvant, 349–350 cellular uptake, 333 CTL response, 353–354 evolution, 346–348 extramedullary hematopoiesis-induced, 351 immune cell activation, 352 immune response, 331–332 mediative effects in vivo, 349 as mucosal adjuvant, 351 sequence effects APCs, 336–338 B cells, 334–335 immunosuppressive, 342 independent, 333–334 macrophage receptors, 341 NK cells, 339 T cells, 338–339 TNF-움 production, 342 signal mediation cell response, 352–353

cellular uptake, 343–344 chloroquine effects, 346 lysosomotropic compounds, 344–345 stress kinase pathway, 345–346 T cell effects as Th1 adjuvant, 348–349 Th2 pathology, 350–351 infections in vivo, 467–468 protein presentation, 26–32 Baﬁlomycin A, 346 BALT, see Bronchial-associated lymphoid tissue B cell linker protein–Grb2 complex, 96, 98 B cell receptors signal transduction CD95, 132–134 CD19 regulation, 118–119 CD22 regulation CD22-deﬁcient mice in vivo, 108–109 description, 107–108 MAP kinases, 115–119 SHP-1-deﬁcient mice in vivo, 110–112 CD40 regulation apoptosis, 124–127 MAP kinase role, 128–129 overview, 122–124 P13K pathway, 129 PTK role, 127–128 TRNR factors, 129–132 CDw150 regulation, 121–122 B cells antigen receptor complex, 89–91 cell death pathways mIgM model, 105–106 WEHI-231 model, 104–106 CpG effects, 334–335 death regulation, 84–87 DNA activation, 347 signal transduction Grb2-related molecules, 97–100 MAP kinases, 101–104 overview, 79–80 P13K, 92–93 PLC-웂, 92–93 pleckstrin homology domains, 101–102 PTKs, 91–92 Ras pathway, 95–97 Vav guanine-nucleotide exchange factor, 100

INDEX

size, 85–86 subpopulations, 88–89 Bcl-2 Bad, 95 CD40 regulation, 125 BFA, see Brefeldin A Biologic ﬁltration, 188–189 BLNK, see B cell linker protein Bone marrow function, 306–307 BRA, see Brefeldin A BrdU, see Bromodeoxyuridine Breast cancer IL-6 production, 440 TNF-움 production, 440 Brefeldin A function, 21 Bromodeoxyuridine, 315–316 Bronchial-associated lymphoid tissue, 203

C CAD, see Chronic airway disease Calcium pyrophosphate dihydrate IL-1 association, 403 IL-8 association, 378 Candida albicans, 469–471 CAPD, see Continuous ambulatory peritoneal dialysis CAR, see Carrageenan CARDIAK, 271 Cardiovascular disease, 284 Carrageenan, 458 Caspases caspase-1 IL-18 and, 274 IL-1웁 and, 271–273 INF-웁 and, 275 caspase-3 IL-16 and, 274 caspase-4 IL-1웁 processing, 273 caspase-8 CrmA, 278 DR3 ligand and, 286 FasL, 277–280 inﬂammatory process, 276–277 caspase-10 CrmA, 278 caspase-11, 273

543

classiﬁcation, 266 –cytokine interaction DR3 ligand, 286 FasL, 277–280 IL-1, 270–273 IL-6, 273 IL-16, 274 IL-18, 274–275 INF-웁, 274–275 overview, 267–270 TNF-움, 275–277 function, 265–266 recruitment domains, 266 sepsis and, 285 CD19, 118–119 CD22 B cell transduction CD22-deﬁcient mice in vivo, 108–109 description, 107–108 Lyn-deﬁcient mice in vivo, 109–110 MAP kinases regulation, 115–119 PTK pathway, 112–113, 115 PTPase pathway, 112–113, 115 SHP-1-deﬁcient mice in vivo, 110–112 CD40 BCR regulation apoptosis, 124–127 overview, 122–124 CD95, induced death, 132–134 CD30 ligand, 398–399 Cell death anti-Ig-induced, 124–127 B cells, regulation, 84–87 BCR pathways mIGD model, 105–106 mIgM model, 105–106 WEHI-231 model, 104–105 CD95-induced, 132–134 FasL-induced, 394–397 induction, 79–80 inhibition, 83–84 Cell debris, 34–36 CEMF, see Corneal endothelium modulation factor CF, see Cystic ﬁbrosis Chaperones function, 6, 8 immunization, 52–56 Chase–Sulzberger phenomenon, 156–157

544

INDEX

Chemokines characterization, 373, 375 types GROs, 382–385 IL-8, 375–382 MCPs, 389–390 MIPs, 387–389 Chloroquine, 346 Cholera toxin coupled-antigens, 182–183 tolerance induction, 186–187 Chromosomal damage, T cells, 312–313 Chronic airway disease PMN role, 427 treatment, 380 c-jun amino-terminal kinases, 101–104 Clathrin endocytosis, 12–13 nocytosis, 12–13 Colchicine, 403–404 Colitis, 221 Collagen induced arthritis caspase-1 effect, 272 oral tolerance model animal, 211–213 human, 230 nasal administration, 205 oral, systemic sclerosis, 231 Collagenase, 424–425 Colloral, 230 Continuous ambulatory peritoneal dialysis IL-8 production, 427–428 Corneal endothelium modulation factor, 424 Corticotropin-releasing hormone, 399–400 Cowpox virus protein FAS ligand and, 278 T cell selection, 280 TNF-움 and, 276 CpG dinucleotides as antitumor adjuvant, 349–350 CTL response, 353–354 description, 329–330 evolution, 346–348 extramedullary hematopoiesis-induced, 351 immune cell activation, 352 mediative effects in vivo, 349 as mucosal adjuvant, 351 sequence effects DNA vaccines, 340–341

immunosuppressive, 342 macrophage receptors, 341 TNF-움 production, 342 sequence-speciﬁc effects APCs, 336–338 B cells, 334–335 NK cells, 339 T cells, 338–339 signal mediation cell response, 352–353 cellular uptake, 343–344 chloroquine effects, 346 lysosomotropic compounds, 344–345 stress kinase pathway, 345–346 T cell effects as Th1 adjuvant, 348–349 Th2 pathology, 350–351 CPPD, see Calcium pyrophosphate dihydrate CRH, see Corticotropin-releasing hormone CrmA, see Cowpox virus protein Crohn’s disease, 231 Cross-priming, development, 2–3 CSA, see Cyclosporin Cycloheximide, 21 Cyclophilin, 336–337 Cyclosporin, 528 Cyclosporin A, 336–337 Cystic ﬁbrosis, 380, 426–427 Cytochalasin, 23 Cytokines, see also speciﬁc cytokines –caspase interaction DR3 ligand, 286 FAS ligand, 270–273 overview, 267–270 TRAIL, 286–287 CpG DNA, 352 neutrophil-derived acute inﬂammation in vivo, 462–467 antitumor response in vivo, 472–473 associated infections, 437 bone formation in vivo, 476 general features, 369–373, 375 HIV-associated, 439–440 immune response in vivo, 472–473 infections in vivo, 467–471 LPS effects in vivo, 454, 457–462 molecular regulation evidence, 447–448 mRNA stability, 449–450 postranslational, 452–453

545

INDEX

transcriptional, 448–449 translational, 450–452 pancreatitis in vivo, 474 periodontitis in vivo, 473–474 pulpitis in vivo, 473–474 speciﬁc substance, 373 wound healing in vivo, 474–475 neutrophils expression, see also speciﬁc cytokines nonexpression, 425–426 NK cells-produced, 415 production CpG effects, 335 neutrophil role, 369–373 T cell production, 338–339

D Death effector domains, 266 DED, see Death effector domains Delayed-type hypersensitivity inhibition, 177 measurement, 155 OVA reaction, 185–186 suppression, 174 Dendritic cells characterization, 15–16 control mechanism, 337 CpG effects, 340 inﬂuenza virus, 34 in vivo role, 16–17 OVA presentation, 44–45 Dental infections, 434–437 Dermatitis IL-8 production, 427 nickel sulﬁte, 229 Diabetes mellitus caspase-1 role, 281 FAS ligand role, 280–281 OVA-induced, 224 tolerance models, 215–216 Digestion, enzymatic, 188–189 DNA CpG motifs as antitumor adjuvant, 349–350 cellular uptake, 333 CTL response, 353–354 description, 329–330 evolution, 346–348

extramedullary hematopoiesis-induced, 351 immune cell activation, 352 mediative effects in vivo, 349 as mucosal adjuvant, 351 sequence-independent effects, 333–334 sequence-speciﬁc effects APCs, 336–338 B cells, 334–335 immunosuppressive, 342 macrophage receptors, 341 NK cells, 339 T cells, 338–339 TNF-움 production, 342 signal mediation cell response, 352–353 cellular uptake, 343–344 chloroquine effects, 346 lysosomotropic compounds, 344–345 stress kinase pathway, 345–346 T cell effects as Th1 adjuvant, 348–349 Th2 pathology, 350–351 plasmids, 56–60 vaccines, 340–341 DR3 ligand, 286 DTH, see Delayed-type hypersensitivity DTV, see Dual tropic virus Dual tropic virus characterization, 514–515 lymphomagenic, 516 thymic function, 515, 526

E EAE, see Experimential autoimmune encephalomyelitis EAMG, see Experimental autoimmune myasthenia gravis EAN, see Experimental allergic neuritis EBV, see Epstein-Barr virus EL4 cells, 51 Endocytosis clathrin-mediated, 12–13 inhibition, 23 non-clathrin mediated, 13–14 Endogenous antigens MHC generation, 4–9 presentation in vivo, 56–63

546

INDEX

Endoplasmic reticulum BFA inhibition, 21 extra, MHC loading, 9–12 transport, 6–9 Endosomal proteases, 23 Enzymatic digestion, 188–189 Epstein-Barr virus FasL production, 398 GM-CSF effects, 451 IL-8 production, 381–382 IL-1RI/IL-1RII production, 408 SAP link, 122 ERK, see Extracellular signal-regulated kinases Exogenous antigens internalization modes, 12–14 presentation ablating pAPCs in vivo, 23–24 aggregated proteins, 49–50 bacteria artiﬁcial particles, 32–34 pathways, 26–32 cell debris, 34–36 cellular sonicates, 50–51 complex quantitating, 17–19 denatured proteins, 49–50 endogenous antigens in vivo, 56–63 immunogenicity, 23–24 inactivated viruses, 23–24 ISCOM, 42–44 liposome, 42–44 mechanism, 19–23 native proteins, 44–45 processing, 19–20 quantum theory, 51–52 receptor-mediated uptake, 46–49 srivastava saga, 52–56 viral, 32–38 virosome, 42–44 virus-like particles, 38–39 Exosomes, characterization, 11 Experimental allergic neuritis, 221 Experimental autoimmune myasthenia gravis, 305 Experimential autoimmune encephalomyelitis MBP administration active suppression, 164–165, 206–208 glatiramer acetate, 211 Lewis rat, 208, 210 murine, 210–211

nasal aerosol, 211 neonatal, 223–224 oral tolerance models, 164–169 Extracellular signal-regulated kinases activation, BCR-induced, 96 B cell transduction CD22 regulation, 115–119 description, 101–104

F Fas ligand AIDS, 282–283 B cell regulation, 121–122, 132–134 cardiovascular disease, 284 caspase-8 and, 277–280 characterization, 277–278 GVHD, 283–294 Hashimoto’s thyroiditis, 281–282 hepatic diseases, 284–286 IBD, 286 IDDM and, 280–281 multiple sclerosis, 281 PMN expression, 397–398 TRAIL and, 287 FDC, see Follicular dendritic cells Felodipine, 231–232 Fetal thymus organ culture, 307 Fibrosis, 420 Filtration, biologic, 188–189 FLAP, see 5-Lipoxygenase-activating protein Follicular dendritic cells, 123 Food allergies, 223 FTOC, see Fetal thymus organ culture FV-1 gene, 516

G GALT, see Gut-associated lymphoid tissue GCs, see Germinal centers G-CSF, see Granulocyte-colony stimulating factor GEF, see Guanine nucleotide exchange factors Gelonin, 22 Germinal centers, 87–88 Gingival crevicular ﬂuid, 434–437 gp96, 53–54

547

INDEX

Graft-versus-host disease, 283–294 Granulocyte-colony stimulating factor, 422 Granulocyte-derived factor, 423–424 Granulocytes macrophage-stimulating factor neutrophil-derived EBV effects, 446–447 modulation, 446–447 Granzyme B, 271 Grapefruit juice, 231–232 Grb2 B cell pathway, 97–100 –BLNK complex, 96 GRO, see Growth-related gene products Growth-related gene products neutrophil-derived associated diseases, 428 characterization, 382–385 Guanine nucleotide exchange factors Grb2 interaction, 97–98 Ras activation, 95–96 Gut-associated lymphoid tissues age factors, 179 antigen production, 159 Th2 generation, 160 Gut mucosa, 153 GVHD, see Graft-versus-host disease

Herpetic stromal keratitis, 468 HGF, see Hepatocyte growth factor Highly active antiretroviral therapy description, 311–312 T cell depletion, 320–322 HSV, see Herpes simplex virus Human immunodeﬁciency virus, see also AIDs CD30 production, 399 cytokine production, 439–440 type 1 T cell depletion causes, 320–322 mechanism, 301–303 T cell renewal bone marrow function, 306–307 development, 307 impact, 305–306 thymic function, 307–308 T cells Ki67 expression, 319–320 turnover, 309–310 Hypersensitivity DTH, see Delayed-type hypersensitivity

I H HAART, see Highly active antiretroviral therapy Hashimoto’s thyroiditis, 281–282 HCV, see Hepatitis C virus Heat shock proteins HSP70, immunization, 52–53 HSP90, immunization, 52–53 HeLa cells, 12 HEL cells, 119 Hematopoiesis, extramedullary, 351 Hemocyanin, see Keyhole limpet hemocyanin Hepatitis alcohol-induced, 286 B/C, 285 HGF production, 438–439 Hepatitis B virus, 38–41 Hepatitis C virus, 438–439 Hepatocyte growth factor, 421–422 Hepatocyte-stimulating activity, 425 Herpes simplex virus, 415

IAP, see Inhibitors of apoptosis proteins IBD, see Inﬂammatory bowel disease IFD, Inﬂammatory bowel disease IFN-웂, see Interferon-웂 Immune complex disease, 221 Immunoglobulins IgA, 153–155, 175–176 IgM, 115 Immunoreceptor tyrosine-based inhibitory motifs, 107 Immunostimulating complex, 123–124 Infections cytokine role in vivo, 467–471 dental, 434–437 Inﬂammation cytokine effects in vivo, 454, 457–467 cytokine production, 378–379 IL-1RI/IL-1RII production, 408–409 lesions, VEGF production, 420–421 MOS role, 15 oral antigens, 194–195

548 pulmonary CpG DNA role Th1, 349 Th2, 350–351 IL-8 production, 380–381 TGF-웁 production, 420 TGR-움 production, 418–419 Inﬂammatory bowel disease animal models, 194–195 cytokine production, 428–430 FasL production, 286 IL-1 implications, 272–273 Inﬂammatory diseases, see also speciﬁc disease DR3 ligand role, 286 IL-1웁 implications, 270–271 INF-웁 production, 274–275 TNF-움 implications, 275–277 TRAIL production, 286–287 Inﬂuenza virus CpG motifs DNA role, 351 DC processing, 34 processing pathway, 36 Inhibitors of apoptosis proteins, 83 Inhibitor-user responsibility code, 19 Interferon-움/웁, 339 Interferon-웂 caspase-1 and, 275 characterization, 274 IL-12 expression, LPS-induced, 414–415 induced oral tolerance, 175 neuterophil expression, 415–417 neutrophil modulation, 441–442 production, CpG effects, 335 Interleukin-1 converting enzyme, 80 Interleukin-1웁-converting enzyme, 266 Interleukins IL-1 characterization, 270–270 neutrophil-derived characterization, 399–405 GDF, 434–437 modulation, 447 IL-1웁 caspase-1 and, 271–273 caspase-4 and, 273 MS, 281 neutrophil-derived IFD, 428–430 RA, 430–432

INDEX

sepsis, 432–434 SLE, 432 proteolytic processing, 271–273 IL-4 arthritis, 214 neutralization, 529 neutrophil-derived, 446 RadLV role, 524–525 IL-6 cancer, 409 caspase processing, 273 characterization, 409 CpG effects APCs, 336–338 B cells, 335 neutrophil-derived, 409–412 IL-8 neutrophil-derived associated diseases, 426–429 characterization, 375–382 sepsis, 432–434 IL-12 CpG effects APCs, 336–338 B cells, 335 neutrophil-derived, 413–415 IL-18 caspase-1 and, 274 characterization, 274 IL-10, neutrophil-derived, 443–445 IL-13, neutrophil-derived, 446 IL-16, neutrophil-derived, 274 IL-2R, TNF-움 release, 395–396 IL-1RI/IL-1RII, 405–409 TNF-움 neutrophil-derived, 434–437 Intestinal ﬂora, 183 Irradiation, induced thymic lymphoma, 518–520 ISCOM, see Immunostimulating complex ITIMs, see Immunoreceptor tyrosine-based inhibitory motifs

J JNK, see c-jun amino-terminal kinases

K Keratitis herpetic stromal, 468 Keyhole limpet hemocyanin, 228

549

INDEX

Ki67 antigen, 316–320 Kidney regeneration, 421 KLH, see Keyhole limpet hemocyanin

L Lactacystin, 20–21 Lactoferricin, 378–379 Lactoferrin, 378–379 Lamina propria cells, 193–194 LAT, see Linker for activation of T cells LCMV, see Lymphocytic choriomeningitis virus LDGF, see Leukocyte-derived growth factor Leishmania monocytogenes, 30–31 Leukemia virus Mø presentation, 34 radiation description, 518 induced lymphomagenesis, 520–525 Leukocyte-derived growth factor, 423 Linker for activation of T cells, 96–97 Lipopolysaccharide bacteria, 183 cytokine effect in vivo, 454, 457–562 FasL expression, 285 IL-12 expression, 414 PRR, 329–330 TNF-움 expression, 392–393 5-Lipoxygenase, 404 5-Lipoxygenase-activating protein, 404 Liver diseases FasL production, 284–286 oral tolerance, 200–201 Liver regeneration, 421 LLO, 48–49 Long-term nonprogressors, 307 LPCs, see Lamina propria cells LPS, see Lipopolysaccharide LTNPs, see Long-term nonprogressors Lung inﬂammation CpG DNA role Th1, 349 Th2, 350–351 cytokine role in vivo acute effects, 464–465 LPS effects, 454, 457–458 IL-8 production, 380–381 Lymphocytic choriomeningitis virus B. pertussis inactivation, 48 chaperone immunization, 53

diabetes induction, 216 Lymphomas Raji Burkitt’s, 122 thymic carcinogen-induced, 516–518 generation AKR mice, 514–515 immunobiology, 513–514 PLCs, 515 preventive therapy, 525–529 RadLV-induced, 520–525 radiation-induced, 518–520 Lyn, B cell regulation PTK pathway, 112–113, 115 PTPase pathway, 112–113, 115 role, 109–110

M Macrophage inﬂammatory proteins, 387–389 Macrophage monocytes bacterial exposure, 28–29 characterization, 15 in vivo role, 16–17 leukemia virus, 34 macropinocytosis, 44 Macropinocytosis, 14, 44 Major histocompatability complex class I antigen generation, 4–9 characterization, 1 CTL responses, 353–354 DC expression, 16 endosomal trafﬁcking, 11 evolution, 1 extra-ER loading, 9–12 gene products, 6–9 restriction, 2–3 SHP-1 regulation, 115 class II CTL responses, 353–354 DC expression, 16 SHP-1 regulation, 113 MAPK, see MAP kinases MAP kinases, BCR transduction CD22 regulation, 115–119 CD40 regulation, 128–129 description, 101–104 Maternal donor allografts, 229 MBP, see Myelin basic protein

550

INDEX

M cells, 189–190 MCP, see Monocyte chemotactic proteins Meningitis, 437–438 Methotrexate, 214 MHC, see Major histocompatability complex Microcrystals monosodium urate IL-1 association, 403 IL-8 association, 378 Monocyte chemotactic proteins, 389–390 mRNA stability, 449–450 MSU, see Microcrystals monosodium urate Mucosa, see Gut mucosa Mucosal-associated lymphoid tissue, 181 Multiple autoimmune diseases, 221–222 Multiple sclerosis FasL role, 281 oral tolerance, 229 Myelin basic protein bystander suppression, 168–169 EAE models active suppression, 164–168, 206–208 glatiramer acetate, 211 Lewis rat, 208, 210 murine, 210–211 nasal aerosol, 211 neonatal administration, 223–224

N Nasal tolerance, 203–206 Natural killer cells CpG effects, 339 cytokine production, 415 DNA activation, 347 radiation function, 527 Nerve growth factor characterization, 80, 83–84 Neutrophils CD30 ligand expression, 398–399 CEMF expression, 424 CINC production, 384–385 cytokine expression antitumor response in vivo, 472–473 bone formation in vivo, 476 general features, 369–373, 375 HIV-associated, 439–440 immune response in vivo, 472–473 infections in vivo, 467–471 LPS effects in vivo, 454, 457–462

meningitis-associated, 437–438 modulation, 440–447 molecular regulation evidence, 447–448 mRNA stability, 449–450 postranslational, 452–453 transcriptional, 448–449 translational, 450–452 pancreatitis in vivo, 474 shigellosis-associated, 437 speciﬁc substance, 373 wound healing in vivo, 474–475 FAS ligand production, 397–398 function, 369 G-CSF expression, 422 GDF expression, 423–424 GRO production, 382–385 HGF expression, 421–422 HSF expression, 425 IL-1 expression, 434–437 production, 399–405 IL-6 expression cancer, 440 characterization, 409–412 IL-8 expression associated diseases, 426–429 production, 375–382 sepsis, 432–434 IL-12 expression, 413–415 IL-1웁 expression, 428–430 IL-1웁 expression RA, 430–432 SLE, 432 IL-1웁 expression, 432–434 IL-1RI/IL-1RII production, 405–409 INF-움 expression, 415–417 LDGF expression, 423 MCP production, 387–389 MIP production, 387–389 noncytokine, 425–426 OSM expression, 412–413 SCF expression, 422–425 TGR-움 expression, 418–419 TGR-웁 expression, 419–420 TNF-움 expression cancer expression, 440 characterization, 390–397 GDF-associated, 434–437 IFD-associated, 428–430

INDEX

NG, see Nerve growth factor Nickel sensitization animal model, 222 human model, 229 N-methyl-N-nitrosourea, 516–517 NMU, see N-methyl-N-nitrosourea Nocytosis, clathrin-mediated, 12–13 Nonsteroidal antiinﬂammatory drugs, 230 NSAIDs, see Nonsteroidal antiinﬂammatory drugs

O ODNs, see Oligonucleotides Oligonucleotides, synthetic CpG as antitumor adjuvant, 349–350 cellular uptake, 333 CTL response, 353–354 evolution, 346–348 extramedullary hematopoiesis-induced, 351 immune cell activation, 352 immune responses, 331–332 mediative effects in vivo, 349 as mucosal adjuvant, 351 sequence effects APCs, 336–338 B cells, 334–335 immunosuppressive, 342 macrophage receptors, 341 NK cells, 339 T cells, 338–339 TNF-움 production, 342 sequence-independent effects, 333–334 signal mediation cell response, 352–353 cellular uptake, 343–344 chloroquine effects, 346 lysosomotropic compounds, 344–345 stress kinase pathway, 345–346 T cell effects as Th1 adjuvant, 348–349 Th2 pathology, 350–351 Oncostatin M, 412–413 Oral antigens biologic ﬁltration, 188–189 enzymatic digestion, 188–189, 196–197 feeding regimes, 196–197 immunological effects, 153–155 induction factors, 183–187 inﬂammatory processes, 194–195

intestinal absorption, 189–194 Peyer’s patches, 197–200 structure, 187–188 Oral cancer IL-6 production, 440 TNF-움 production, 440 Oral tolerance allergy models animal, 218, 221 human, 227–228, 231 arthritis models adjuvant, 213–214 antigen-induced, 214 collagen-induced, 211–213 human, 230 pristane-induced, 214 autoimmune models animal, 164–165, 206–208 worsening, 223–224 cellular, 174–176 Crohn’s disease, 231 deﬁnition, 157–158 diabetes models, 215–216 EAE models, MBP glatiramer acetate, 211 Lewis rat, 208, 210 murine, 210–211 nasal aerosol, 211 effector phase active suppression anergy, 172–174 bystander, 167–170 characterization, 163–167 clonal deletion, 172 웂␦ T cells, 170–171 idiotypic cells, 171–172 regulatory cells, 171–172 overview, 161–163 history, 155–157 humoral, 174–176 as immunologic event, 158–159 induction factors animal models age, 178 genetic background, 181 immunological status, 181–183 intestinal ﬂora, 183 types, 178 antigens biologic ﬁltration, 188–189

551

552 dosage, 196–197 enzymatic digestion, 188–189 feeding regimes, 196–197 inﬂammatory processes, 194–195 intestinal absorption, 189–194 liver’s role, 200–201 Peyer’s patches, 197–200 spleen’s role, 201 structure, 187–188 types, 183–187 inductive phase, 159–161 kinetics, 176–178 maintenance factors animal models age, 178 genetic background, 181 immunological status, 181–183 intestinal ﬂora, 183 types, 178 antigens biologic ﬁltration, 188–189 dosage, 196–197 enzymatic digestion, 188–189 feeding regimes, 196–197 inﬂammatory processes, 194–195 intestinal absorption, 189–194 liver’s role, 200–201 Peyer’s patches, 197–200 spleen’s role, 200–201 structure, 187–188 types, 183–187 maternal donor allografts, 229 modulation, 201–202 multiple sclerosis, 229 nasal, 203–206 nickel allergy, 229 senescence, 180 thyroid, 231 uveitis, 231 uveitis models, 216–217 Osteoarthritis, 431–432 OT-1 cells, 56–57 Ovalbumin liposomal association, 42–43 oral tolerance anergy, 173 biologic ﬁltration, 188–189 bystander suppression, 170 DTH, 185–186 enzymatic digestion, 188–189

INDEX

feed regimens, 196–197 high doses, 163 humoral responses, 177 induced-diabetes, 224 presentation native proteins, 44–45 pathways, 28–30 Oxidative stress, 285–286

P p38, 345 Pancreatitis, 474 Paraformaldehyde, 20 Paranasal sinus disease, 427 Pattern recognition receptors description, 329 LPS, 329–330 Periarteriolar lymphoid sheath, 86–87 Peyer’s patches oral tolerance, 197–200 tolerance induction, 186 PHA, see Phytohemagglutinin Phagocytosis characterization, 14 inhibition, 23 Phorbol myristic acetate, 44 Phosphatidylinositol 3-kinase, 92–93, 129 Phospholipase C-웂, 92–93 Phytohemagglutinin, 425 P13K, see Phosphatidylinositol 3-kinase Plasminogen activator inhibition, 84 PLC-웂, see Phospholipase C-웂 Pleckstrin homology domains, 101–102 PMA, see Phorbol myristic acetate Prelymphoma cells characterization, 515 emergence, 520 persistence, 524 TNF-움 against, 527 Pristane, 214 Prostaglandin2, 424–425 Proteases, endosomal, 23 Proteasome function, 6 inhibition, 20–21 Protein synthesis inhibitors, 21–22

INDEX

Protein tyrosine kinases BCR transduction biochemical pathways, 112–113, 115 CD40 regulation, 127–128 relation, 91–92 death capase cascade, 133 PRRs, see Pattern recognition receptors Psoriasis, 428

R Rad-1 factor, 521 RAG1/2 regulation, 89–90 Raji Burkitt’s lymphoma, 122 Respiratory syncytial virus, 381 Rheumatoid arthritis adjuvant, 213–214 antigen-induced, 214 collagen-induced animal model, 211–213 human model, 230 cytokine production, 430–432 pristane-induced, 214 synovial ﬂuids, 383 T cell depletion, 314–315 Rhinitis, 231 Ricin B chain, 22 RMA cells, 34

S SAPKS, see Stress-activated protein kinases Scatter factor, see Hepatocyte growth factor SCF, see Stem cell factor SCID, see Severe combined immune deﬁciency Sclerosis, systemic, 231 SCW, see Streptococcal cell wall SDS, see Sodium dodecyl sulfate Sendai virus processing pathway, 36–38 TCD8⫹ activation, 49 Senescence, 180 Sensitization, nickel, 222 Sepsis IL-1웁 production, 432–434 IL-8 production, 432–434 Serpins, 83–84

553

Severe combined immune deﬁciency, 527 Shigella, 271 Shigellosis, 437 SHIP Grb2 interaction, 97–98 PtdIns regulation, 95 SHP-1 B cell regulation PTK pathway, 112–113, 115 PTPase pathway, 112–113, 115 role, 110–112 CD22 binding, 107–108 SHP-2 B cell regulation CD22 binding, 107–108 Fc웂RIIB, 120–121 Shwartzman reaction, 457 Signal transduction B cells overview, 79–80 P13K, 92–93 PLC-웂, 92–93 pleckstrin homology domains, 101–102 PTKs, 91–92 Ras pathway, 95–97 Vav guanine-nucleotide exchange factor, 100 BCR CD19, 118–119 CD22 regulation, 115 CD22-deﬁcient mice in vivo, 108–109 description, 107–108 Lyn-deﬁcient mice in vivo, 109–110 MAP kinases, 115–119 PTK pathway in vivo, 112–113, 115 SHP-1-deﬁcient mice in vivo, 110–112 CD40 regulation apoptosis, 124–127 MAP kinase role, 128–129 overview, 122–124 P13K pathway, 129 PTK role, 127–128 TRNR factors, 129–132 CD95 regulation, 132–134 CDw150 regulation, 121–122 Fc웂RIIB, 120–121 Grb2-related molecules, 97–100 CpG-ODN cell response, 352–353 cellular uptake, 343–344

554

INDEX

chloroquine effects, 346 CTL response, 353–354 lysosomotropic compounds, 344–345 stress kinase pathway, 345–346 Simian immunodeﬁciency virus, 315–316 SIV, see Simian immunodeﬁciency virus SLE, see Systemic lupus erythematosus Sodium dodecyl sulfate, 48 S particles, 38–41 Spleen, 200–201 Stem cell factor, 422–425 Streptococcal cell wall, 186 Stress, oxidative, 285–286 Stress-activated protein kinases, 101–104 Substance P, 378 SV, see Sendai virus Systemic lupus erythematosus cytokine production, 432 immune complexes, 331

T TAP1 deﬁcient cells, 22 expression, 12 function, 6–9 TAP2 deﬁcient cells, 22 expression, 12 function, 6–9 T cells active suppression role, 170–171 CpG DNA role, 353–354 depletion HIV-1-induced, 320–322 recovery rates, 313–315 thymus relation, 322 division Ki67 mAb expression humans, healthy, 316–319 humans, HIV-infected, 319–320 endosome access, 11 HIV-1 depletion, 301–303 in vivo role, 16–17 LAT role, 96–97

memory responses, 24–25 pAPC triggering, 15 priming, 42–44 product, mathematical model, 323–324 production BrdU labeling, 315–316 division rates, 312–313 total body number, 310–312 turnover, 309–310 progenitor renewal HIV-1 infection bone marrow function, 306–307 development, 307 impact, 305–306 thymic function, 307–308 human, 304–305 mice, 303–304 normal, 303–304 response, 1–2, 17 SV, 49 Th1 response, 348–349 Th2 response, 350–351 triggering, 1–3 viral response, 3–4 Thymus development, 322 function decline, age-related, 305 T cell renewal, 307–308 lymphomas AKR mice, 514–515 generation immunobiology, 513–514 PLCs, 515 preventive therapy, 525–529 RadLV-induced, 520–525 organ culture, fetal, 307 Thyroid hormone replacement, 231 Thyroiditis, 222 Tolerance, see Oral tolerance Trachael eosinophilia, 222 TRAIL characterization, 286–287 FasL and, 287 Transforming growth factors TGR-움, 418–419 TGR-웁, 419–420 VEGF expression, 420–421 Transplantation, 222–223

555

INDEX

Tumor growth factor-웁, 165–167 Tumor necrosis factors receptor families CD40 signaling, 129–132 description, 80, 83–84 TNF-움 characterization, 390–391 CpG effects, 336–338 inﬂammatory disease implications, 275–277 LPS expression, 392–393 ODN effects, 342 PMN expression, 391–392, 394–397 prelymphoma cells, 527 Tumors cells, SCF effects, 422 CpG DNA effects, 349–350 cytokine role in vivo, 472–473 propagation, HGF, 421 VEGF effects, 420–421

V Vaccines DNA, 340–341 TCD8⫹ eliciting, 25 Vascular endothelial growth factor, 420–421 Vav guanine-nucleotide exchange factor, 101–102 VEGF, see Vascular endothelial growth factor Virosomes, 42–44 Viruses, inactivated, 23–24 Virus-like particles, 38–42 VLP, see Virus-like particles

W WEHI-231 cells, 125–126 Wilson’s disease, 284–285 Wound healing cytokine role in vivo, 474–475 HGF production, 421 LDGF production, 423 TGR-움 production, 418–419

U X Uveitis oral tolerance, 231 tolerance models, 216–217

X-linked lymphoproliferative disease, 122 XLP, see X-linked lymphoproliferative disease

This Page Intentionally Left Blank

CONTENTS OF RECENT VOLUMES

Volume 69 Molecular and Cellular Events in Early Thymocyte Development HANS-REIMER RODEWALD AND HANS JO¨ RG FEHLING Regulation of Immunoglobulin Light Chain Isotype Expression FREDERICK W. ALT AND JAMES R. GORMAN Role of Immunoreceptor Tyrosine-Based Activation Motif in Signal Transduction from Antigen and Fc Receptors NOAH ISAKOV Atypical Serine Proteases of the Complement System GE´ RARD J. ARLAUD, JOHN E. VOLANAKIS, NICOLE M. THIELENS, STHANAM V. L. NARAYANA, VE´ RONIQUE ROSSI, AND YUANYUAN XU

Olfactory Receptor Gene Regulation ANDREW CHESS INDEX

Volume 70 Biology of the Interleukin-2 Receptor BRAD H. NELSON AND DENNIS M. WILLERFORD Interleukin-12: A Cytokine at the Interface of Inﬂammation and Immunity GIORGIO TRINCHIERI Recent Progress on the Regulation of Apoptosis by Bcl-2 Family Members ANDY J. MINN, RACHEL E. SWAIN, AVERIL MA, AND CRAIG B. THOMPSON Interleukin-18: A Novel Cytokine That Augments Both Innate and Acquired Immunity HARUKI OKAMURA, HIROKO TSUTSUI, SHIN-ICHIRO KASHIWAMURA, TOMOHIRO YOSHIMOTO, AND KENJI NAKANISHI

Accessibility Control of V(D)J Recombination: Lessons from Gene Targeting WILLIAM M. HEMPEL, ISABELLE LEDUC, NOELLE MATHIEU, RAJ KAMAL TRIPATHI, AND PIERRE FERRIER Interactions between the Immune System and Gene Therapy Vectors: Bidirectional Regulation of Response and Expression JONATHAN S. BROMBERG, LISA DEBRUYNE, AND LIHUI QIN Major Histocompatibility Complex Genes Inﬂuence Individual Odors and Mating Preferences DUSTIN PENN AND WAYNE POTTS

CD4⫹ T-Cell Induction and Effector Functions: A Comparison of Immunity against Soluble Antigens and Viral Infections ANNETTE OXENIUS, ROLF M. ZINKERNAGEL, AND HANS HENGARTNER Current Views in Intracellular Transport: Insights from Studies in Immunology VICTOR W. HSU AND PETER J. PETERS 557

558

CONTENTS OF RECENT VOLUMES

Phylogenetic Emergence and Molecular Evolution of the Immunoglobulin Family JOHN J. MARCHALONIS, SAMUEL F. SCHLUTER, RALPH M. BERNSTEIN, SHANXIANG SHEN, AND ALLEN B. EDMUNDSON Current Insights into the ‘‘Antiphospholipid’’ Syndrome: Clinical, Immunological, and Molecular Aspects DAVID A. KANDIAH, ANDREJ SALI, YONGHUA SHENG, EDWARD J. VICTORIA, DAVID M. MARQUIS, STEPHEN M. COUTTS, AND STEVEN A. KRILIS INDEX

Volume 71 움웁/웂␦ Lineage Commitment in the Thymus of Normal and Genetically Manipulated Mice HANS JO¨ RG FEHLING, SUSAN GILFILLAN, AND RHODRI CEREDIG Immunoregulatory Functions of 웂␦ T Cells WILLI BORN, CAROL CADY, JESSICA JONESCARSON, AKIKO MUKASA, MICHAEL LAHN, AND REBECCA O’BRIEN STATs as Mediators of Cytokine-Induced Responses TIMOTHY HOEY AND MICHAEL J. GRUSBY CD95(APO-1/Fas)-Mediated Apoptosis: Live and Let Die PETER H. KRAMMER A CXC Chemokine SDF-1/PBSF: A Ligand for a HIV Coreceptor, CXCR4 TAKASHI NAGASAWA, KAZUNOBU TACHIBANA, AND KENJI KAWABATA T Lymphocyte Tolerance: From Thymic Deletion to Peripheral Control Mechanisms BRIGITTA STOCKINGER Confrontation between Intracellular Bacteria and the Immune System

ULRICH E. SCHAIBLE, HELEN L. COLLINS, AND STEFAN H. E. KAUFMANN INDEX

Volume 72 The Function of Small GTPases in Signaling by Immune Recognition and Other Leukocyte Receptors AMNON ALTMAN AND MARCEL DECKERT Function of the CD3 Subunits of the Pre-TCR and TCR Complexes during T Cell Development BERNARD MALISSEN, LAURENCE ARDOUIN, SHIH-YAO LIN, ANNE GILLET, AND MARIE MALISSEN Inhibitory Pathways Triggered by ITIMContaining Receptors SILVIA BOLLAND AND JEFFREY V. RAVETCH ATM in Lymphoid Development and Tumorigenesis YANG XU Comparison of Intact Antibody Structures and the Implications for Effector Function LISA J. HARRIS, STEVEN B. LARSON, AND ALEXANDER MCPHERSON Lymphocyte Trafﬁcking and Regional Immunity EUGENE C. BUTCHER, MARNA WILLIAMS, KENNETH YOUNGMAN, LUSIJAH ROTT, AND MICHAEL BRISKIN Dendritic Cells DIANA BELL, JAMES W. YOUNG, AND JACQUES BANCHEREAU Integrins in the Immune System YOJI SHIMIZU, DAVID M. ROSE, AND MARK H. GINSBERG INDEX