The Biology of Extracellular Molecular Chaperones (Novartis Foundation Symposium 291)

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Ch...

Author: Novartis Foundation

26 downloads 765 Views 9MB 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

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

THE BIOLOGY OF EXTRACELLULAR MOLECULAR CHAPERONES

The Novartis Foundation is an international scientific and educational charity (UK Registered Charity No. 313574). Known until September 1997 as the Ciba Foundation, it was established in 1947 by the CIBA company of Basle, which merged with Sandoz in 1996, to form Novartis. The Foundation operates independently in London under English trust law. It was formally opened on 22 June 1949. The Foundation promotes the study and general knowledge of science and in particular encourages international co-operation in scientific research. To this end, it organizes internationally acclaimed meetings (typically eight symposia and allied open meetings and 15–20 discussion meetings each year) and publishes eight books per year featuring the presented papers and discussions from the symposia. Although primarily an operational rather than a grant-making foundation, it awards bursaries to young scientists to attend the symposia and afterwards work with one of the other participants. The Foundation’s headquarters at 41 Portland Place, London W1B 1BN, provide library facilities, open to graduates in science and allied disciplines. Media relations are fostered by regular press conferences and by articles prepared by the Foundation’s Science Writer in Residence. The Foundation offers accommodation and meeting facilities to visiting scientists and their societies. Information on all Foundation activities can be found at http://www.novartisfound.org.uk

Novartis Foundation Symposium 291

THE BIOLOGY OF EXTRACELLULAR MOLECULAR CHAPERONES

Copyright © Novartis Foundation 2008 Published in 2008 by John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester PO19 8SQ, UK National 01243 779777 International (+44) 1243 779777 e-mail (for orders and customer service enquiries): [email protected] Visit our Home Page on http://www.wileyeurope.com or http://www.wiley.com All Rights Reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road, London W1T 4LP, UK, without the permission in writing of the Publisher. Requests to the Publisher should be addressed to the Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or emailed to [email protected], or faxed to (+44) 1243 770620. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. Other Wiley Editorial Offices John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany John Wiley & Sons Australia Ltd, 33 Park Road, Milton, Queensland 4064, Australia John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 6045 Freemont Blvd, Mississauga, Ontario, Canada L5R 4J3 Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Novartis Foundation Symposium 291 x + 224 pages, 26 figures, 15 tables

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 978-0470-72333-3 Typeset in 10½ on 12½ pt Garamond by SNP Best-set Typesetter Ltd., Hong Kong Printed and bound in Great Britain by T. J. International Ltd, Padstow, Cornwall.

Contents

Symposium on The biolog y of extracellular molecular chaperones, held at the Novartis Foundation, London, 5–7 June 2007 Editors: Derek J. Chadwick (Organizer) and Jamie Goode This symposium is based on a proposal made by Brian Henderson, R. John Ellis and A. Graham Pockley Péter Csermely Chair’s introduction Jodie Haak and Kevin C. Kregel Discussion 15

1

1962–2007: a cell stress odyssey 3

Peter A. Lund and R. John Ellis The chaperone function: meanings and myths 23 Discussion 36 Péter Csermely, Tamás Korcsmáros, István A. Kovács, Máté S. Szalay and Csaba Soti Systems biology of molecular ´´ chaperone networks 45 Discussion 54 Radhey S. Gupta , Nallur B. Ramachandra, Timothy Bowes and Bhag Singh Unusual cellular disposition of the mitochondrial molecular chaperones Hsp60, Hsp70 and Hsp10 59 Discussion 69 Martha Triantafi lou, Daniel Sawyer, Abdiaziz Nor, Emmanouil Vakakis and Kathy Triantafi lou Cell surface molecular chaperones as endogenous modulators of the innate immune response 74 Discussion 79 A. Graham Pockley and Gabriele Multhoff extracellular fluids: friend or foe? 86 Discussion 96 v

Cell stress proteins in

vi

CONTENTS

Francisco J. Quintana and Irun R. Cohen HSP60 speaks to the immune system in many voices 101 Discussion 111 Stuart K. Calderwood, Jianlin Gong, Jimmy R. Theriault, Salamatu S. Mambula and Philip J. Gray Jnr Cell stress proteins: novel immunotherapeutics 115 Discussion 131 General discussion

137

Brian Henderson Cell stress proteins as modulators of bacteria–host interactions 141 Discussion 154 Anthony R. M. Coates, Ana Cehovin and Yanmin Hu macrophage activation 160 Discussion 169 Alexzander Asea Discussion 179

Chaperonin 60 and

Hsp70: a chaperokine 173

Hajime Nakamura Extracellular functions of thioredoxin 184 Discussion 192 Carol L. Miller-Graziano, Asit De, Krzysztof Laudanski, Tara Herrmann and Sanjukta Bandyopadhyay HSP27: an anti-inflammatory and immunomodulatory stress protein acting to dampen immune function 196 Discussion 208 Gabriel S. Panayi and Valerie M. Corrigall BiP, an anti-inflammatory ER protein, is a potential new therapy for the treatment of rheumatoid arthritis 212 Discussion 216 Final discussion

221

Index of contributors 225 Subject index

227

Participants

Alexzander Asea Division of Investigative Pathology, Scott & White Memorial Hospital and Clinic and The Texas A & M University System Health Science Center, College of Medicine, 1901 South First Street, Building 205, Room 1R22, Temple, TX 76508, USA Stuart K. Calderwood Division of Molecular and Cellular Biology, Department of Radiation Oncology, Beth Israel Deaconess Medical Center and Harvard Medical School, 21–27 Burlington Avenue, Room 553B, Boston MA 02215, USA Claudia Campanella (Novartis Foundation Bursar) Human Anatomy Section, Department of Experimental Medicine, University of Palermo, via del Vespro 129, 90127 Palermo, Italy Michael Cheetham UCL Institute of Ophthalmology, 11–43 Bath St, London EC1V 9EL, UK Anthony R. M. Coates Medical Microbiology, Centre for Infection, Department of Cellular and Molecular Medicine, St George’s, University of London, Cranmer Terrace, London SW17 0RE, UK Everly Conway de Macario Center of Marine Biotechnology (COMB), University of Maryland Biotechnology Institute (UMBI), Columbus Center, 701 E. Pratt Street, Baltimore, MD 21202, USA Valerie M. Corrigall Department of Academic Rheumatology, Division of Immunology, Infection and Inflammatory Disease, King’s College London School of Medicine at Guy’s, King’s College and St Thomas’ Hospitals, King’s College London, 5th Floor Thomas Guy House, London SE1 9RT, UK Péter Csermely (Chair) Department of Medical Chemistry, Semmelweis University, Puskin Street 9, H-1088, Budapest, Hungary R. John Ellis Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, UK vii

viii

PARTICIPANTS

Martin E. Feder Department of Organismal Biology and Anatomy, The University of Chicago, 1027 East 57th Street, Anatomy Building (Room 309), Chicago, IL 60637-1508, USA J. S. Hill Gaston University of Cambridge School of Clinical Medicine, Box 157, Level 5, Department of Medicine, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, UK Radhey S. Gupta Department of Biochemistry HSC-4H2, McMaster University, 1200 Main St West, Hamilton, Ontario L8N 3Z5, Canada Christiane Habich Deutsches Diabetes-Zentrum DDZ, Leibniz-Zentrum für Diabetes-Forschung an der Heinrich-Heine-Universität Düsseldorf, Auf’m Hennekamp 65, 40225 Düsseldorf, Germany Brian Henderson Division of Microbial Diseases, UCL Eastman Dental Institute, University College London, 256 Gray’s Inn Road, London WC1X 8LD, UK Lawrence E. Hightower University of Connecticut, Molecular and Cell Biology, 91 North Eagleville Road, Unit 3125, Storrs, CT 06269-3125, USA Kevin C. Kregel, Departments of Integrative Physiology and Radiation Oncology, 532 FH, The University of Iowa, Iowa City, IA 52242, USA Peter A. Lund School of Biological Sciences, University of Birmingham, Birmingham B15 2TT, UK Alberto J. L. Macario Center of Marine Biotechnology (COMB), University of Maryland Biotechnology Institute (UMBI), Columbus Center, 701 E. Pratt Street, Baltimore, MD 21202, USA Carol L. Miller-Graziano Immunobiology and Stress Response Laboratories, Department of Surgery, University of Rochester School of Medicine and Dentistry, 601 Elmwood Avenue—SURG, Rochester, New York, NY 14642, USA Gabriele Multhoff Department of Radiotherapy and Radiooncology, Technische Universität München, Hospital rechts der Isar, Building 557, Room 1.33, Ismaningerstr. 22, 81675 München, Germany Hajime Nakamura Thioredoxin Project, Department of Experimental Therapeutics, Translational Research Center, Kyoto University, 54 ShogoinKawaharacho, Sakyo, Kyoto 606-8507, Japan

PARTICIPANTS

ix

Brett Nixon Life Sciences Building, Room LS4-22, School of Environmental and Life Sciences, University of Newcastle, Callaghan, 2308 NSW, Australia Gabriel S. Panayi Guy’s, King’s and St Thomas’ School of Medicine, King’s College London, Department of Academic Rheumatology, Guy’s Hospital, St Thomas’ Street, London SE1 9RT, UK A. Graham Pockley Immunobiology Research Unit, University of Sheffield, Room LU115, L Floor, Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF, UK Francisco J. Quintana Brigham & Women’s Hospital, Harvard Medical School, 177 Ave. L. Pasteur HIM 720, Boston, MA 02115, USA Kathy Triantafi lou University of Sussex, School of Life Sciences, Falmer, Brighton BN1 9QG, UK

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Chair’s introduction Péter Csermely Department of Medical Chemistry, Semmelweis University, Puskin Street 9, H-1088 Budapest, Hungary

This meeting has two major parts. The fi rst deals with the mechanisms of extracellular chaperones. We may pose the following key questions here: • What are the extracellular chaperones doing in general, and how do they get out of the cell? • Is this a nicely regulated process or are we just witnessing the agony of some cells and the product of this? These are all open questions. The second part of the meeting deals with the effects of extracellular molecular chaperones. This is an area where we probably know much more. However, this field is much wider than the previous one. Our meeting will give us a great opportunity to survey all the possible applications of extracellular chaperones as well as the various therapeutic approaches. This will be an exciting exercise for all of us. I would like to raise some additional questions, which we might want to discuss over the next few days. • Plasma membrane dynamics, membrane rafts and microheterogeneities. To begin with, I am curious as to whether we have enough knowledge about the dynamics of the plasma membrane system in our experiments. Let’s remember the paradigm change that occurred in nuclear transport: fi rst we thought this was a specific transport process, then we believed that it is a relatively free process, where only the retention of proteins in the nucleus is defining which of them is getting out and which of them stays in, and now we have returned again to the specific transport hypothesis. We have to be aware of the importance of rafts and microheterogeneities in the plasma membrane, and their possible roles in the secretion of extracellular chaperones. • Personal stress history. There will be lots of discussion on immunomodulation, and the pro- and anti-inflammatory role of the chaperones outside the cells. We are close to the point where we might start to talk about a personal stress history of patients reflected by molecular chaperones and the antigens 1

2

CSERMELY

against them, including the prenatal and postnatal period. It is a highly complex picture that is emerging. • Personal symbiotic status. It also important that we should take into account the personal symbiotic status: the plethora of those bacteria that inhabit us, and whether there is some conflict between the two which may lead to serious health problems. I also have some questions, which may have a lesser importance, but which might turn out to be quite exciting later on. • Role of extracellular ATP in the complex formation of extracellular chaperones. We know a lot about the role of ATP in the function of intracellular chaperones. What about extracellular chaperones? I know that generally ATP is not stable outside the cell, but what about its local role? How does ATP affect the local complex formation of extracellular chaperones? It is an interesting open question. • Chaperones of the extracellular matrix, oligosaccharide-chaperones. What about the extracellular matrix? Are there special chaperones for certain parts of the extracellular matrix, and how general and necessary is that? Is it just necessary from time to time? There are a lot of oligosaccharides outside the cell: do they need or have chaperones at all? These are questions that could be important but which are seldom addressed.

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

1962–2007: a cell stress odyssey Jodie Haak and Kevin C. Kregel l Department of Integrative Physiolog y, The University of Iowa, 532 FH, Iowa City, IA 52242, USA

Abstract. The induction of a cellular stress response was fi rst observed in 1962 in a set of serendipitous experiments in Drosophila melanogaster larvae, which led to the discovery of a family of intracellular polypeptides known as heat shock proteins (HSPs). These highly conserved proteins are present in both prokaryotic and eukaryotic species, suggesting that they play important roles in fundamental cellular processes. Moreover, these proteins are induced in response to a range of stimuli, implicating HSPs as important modifying factors in an organism’s response to a variety of physiological conditions. HSPs were initially regarded as intracellular molecules mediating cytoprotective, regulatory and chaperoning functions. However, the past two decades have seen an explosion of information related to the cell stress response, with a primary focus on molecular chaperones, which are a class of multifunctional intracellular proteins that assist in folding and assembly of other proteins. Stress proteins have also been identified on cell surfaces and in extracellular fluids, and are now viewed as potential immunomodulators, pro-inflammatory signalling molecules, and anti-inflammatory proteins in disease states. This chapter serves as an overview of the rapidly expanding world of cell stress proteins and aims to provide the reader with a foundation for more detailed presentations in subsequent sections of this book. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 3–22

In the face of stress, cells utilize a variety of intricate coping mechanisms. One key system involves a broad array of highly conserved stress proteins that have a range of functions aimed at maintaining cellular homeostasis. These molecules, which include a high-profi le collection of proteins termed molecular chaperones, play vital roles in mediating polypeptide folding, signalling, chaperoning and cytoprotection inside cells. However, several provocative lines of investigation now implicate these proteins as important players in other realms, including the extracellular milieu, with involvement in critical stress-related processes related to inflammation and immunity. Much of the current work in the field is now aimed 1

This paper was presented at the symposium by Kevin Kregel, to whom correspondence should be addressed. 3

4

HAAK & KREGEL

at delineating the molecular and cellular mechanisms by which these different protein families function in maintaining homeostasis and modulating stress responses. Historical overview The first evidence for a cellular stress response is generally associated with a set of experiments in fl ies in 1962 (Ritossa 1962). Larvae of Drosophila melanogaster were accidentally kept overnight at an elevated temperature and upon inspection the next day, an unusual puffing pattern was noted on salivary gland chromosomes, indicating that the ‘heat shock’ episode had caused marked changes in the gene expression pattern of the larvae. However, the term ‘heat shock protein’ (HSP) was not coined until 1974, when specific protein products associated with these genes were identified in fl ies (Tissieres et al 1974). Over the next three decades, the cellular stress response, as the heat shock response came to be known, has been extensively studied, particularly with regard to stress proteins and their everexpanding array of functions (Benjamin & McMillan 1998, Hightower 1991, Kregel 2002, Lindquist & Craig 1988, Morimoto et al 1994, Welch 1992). Relevant to this symposium, the term ‘molecular chaperone’ appeared in 1987 as a general descriptor of a newly identified group of intracellular proteins (Ellis 1987). These chaperones were made up of several diverse and unrelated families of proteins that were proposed to assist in the correct folding of proteins and function in protein assembly and disassembly. As this field has ‘unfolded’ since the early 1990s, two observations deserve mention. First, it has been a challenge for investigators to understand and apply the correct terminology as it relates to HSPs and molecular chaperones (Ellis 2005). For the purposes of this chapter and the necessity of presenting a broad, general perspective on the cellular stress response in mammalian systems, it should be appreciated that many chaperones are inducible by stress and assist other proteins in achieving proper folding. Thus, these molecular chaperones are included in the family of heat shock proteins (or cell stress proteins). However, it should also be noted that some molecular chaperones are not stress proteins. Second, it has become very apparent over the past two decades that molecular chaperones have additional functions besides their proteinfolding actions. For instance, paradoxical immune functions have been identified for some molecular chaperones, as they are now known to play a role in both autoimmunity and as immunogens in association with challenges such as infection (Henderson & Shamaei-Tousi 2005). Another interesting twist in the stress response story involved the observation that some molecular chaperones are localized to the surface of cells (Ferrarini et al 1992, Multhoff 2002). These reports were followed by studies suggesting that molecular chaperones have actions that mimic pro-inflammatory cytokines as well as anti-inflammatory and immunosup-

CELL STRESS ODYSSEY

5

pressive properties. In addition, there is now evidence that some molecular chaperones are present in extracellular fluids and function as cell-to-cell signalling molecules (Frostegard & Pockley 2005, Pockley 2003). Taken together, a modern view of the cell stress response must include the notion that various stress-related proteins, including HSPs and molecular chaperones, are multifunctional and have a range of activities both inside and outside the cell. The classical view of the cell stress response Stress proteins are present in both prokaryotic and eukaryotic cells and have a high level of conservation, suggesting that they play an important role in fundamental cellular processes. A variety of intracellular and extracellular stimuli can elicit a stress response, with a general profi le that includes the activation of stress genes and the inhibition of housekeeping genes. Importantly, the physiological impact of this response can be observed at molecular, cellular and systemic levels. Thus, it has become evident that the regulation and function of stress proteins are key components of the stress response in a physiological setting. HSP families The HSPs are a subset of the more broadly defi ned category of cell stress proteins. As their name indicates, HSPs were identified in cells as the product of genes induced by heat shock. While the initial view of HSPs was that their induction was associated with a heat shock episode, a more modern usage of the HSP term now encompasses a larger class of proteins that have a range of molecular chaperone functions. A large number of proteins have been discovered within the HSP family, and they are generally divided into groups based on both their size and function (Table 1). These proteins are present in the cytosol, mitochondria, endoplasmic reticulum, peroxisomes and nucleus inside cells, as well as the cellular membrane and extracellular fluids. Specific locations vary depending on the particular protein. One of the most studied families of HSPs is the collectively conserved ATP-dependent Hsp70 family. Expressed in virtually every cell type, up-regulation of Hsp70 can be triggered by exposure to a wide range of environmental conditions, including heat, metabolic stress, hypoxia, oxidative stress, infection, and inflammation. The classical paradigm The classical paradigm associated with the heat shock response for a mammalian system is depicted in Fig. 1. Application of a cellular stress such as heating can cause the denaturing of native proteins and aggregation and/or misfolding of

6

HAAK & KREGEL

TABLE 1 Nomenclature and examples of intracellular locations and proposed functions of mammalian heat shock protein families Major protein families and their members

Cellular location

Proposed cellular functions

Small HSPs α -crystallin

Cytoplasm

Chaperone activity; structural control of the cytoskeleton Microfi lament stabilization; anti-inflammation Haem catabolism; antioxidant

Hsp27 Hsp32 (haem oxygenase) Hsp40 Hsp40 Hsp47 Hsp60 (chaperonins) Hsp60 TCP1 Hsp70 Hsc70

Cytoplasm, nucleus, cell surface, extracellular Cytoplasm

Cytoplasm, nucleus Endoplasmic reticulum Mitochondria, cell surface, extracellular Cytoplasm

Cytoplasm, peroxisomes

Hsp70

Cytoplasm, nucleus, cell surface, extracelllular

mtHsp70 (Grp75)

Mitochondria

Grp78(BiP)

Endoplasmic reticulum, cell surface

Hsp90 Hsp90

Grp94/Gp96

Cytoplasm, cell surface, extracellular Endoplasmic reticulum, extracellular

Hsp110 Hsp110

Cytoplasm, nucleolus

Ubiquitin

Cytoplasm

Thioredoxin (TRX) TRX1 TRX2

Cytoplasm, nucleus, extracellular Mitochondria

Protein folding; regulates Hsp70 activity Processing of pro-collagen Folding of newly synthesized and denatured proteins; protein import; inflammation Protein folding; maintenance of actin cytoskeleton Constitutively expressed; molecular chaperone; assists with protein translocation Stress-inducible; protein folding and transport; disassembly of oligomers; inflammation; antigen presentation; natural killer cell activation Cytoprotection; molecular chaperone; protein import Cytoprotection; molecular chaperone; anti-inflammation Constitutively expressed; molecular chaperone; assists with protein translocation; pro-inflammation; antigen presentation Protein folding and transport; disassembly of oligomers; proinflammation; antigen presentation Protein folding and assembly; cytoprotection Protein degradation Inhibit apoptosis; antioxidant Inhibit apoptosis; antioxidant

CELL STRESS ODYSSEY

7

nascent polypeptides in the cell. These misfolded and aggregated proteins can either lose or alter their function, in addition to becoming unable to reach their target cellular location. To protect against extensive protein damage caused by heat, the cell invokes a cell stress response. In the simplified scheme depicted in Fig. 1, a stressor activates a heat shock transcription factor, which then translocates to the nucleus and binds to a heat shock element in the promoter region of the HSP gene. This results in rapid induction of HSPs, which function in protein folding

stress

1

• •••••••• • • •• • • • • • • • • • •• • •••• • • • •• •••• •• • •••• • •• •• •• •• • •• • •••• • •• • • • • •• • • • •• • • • • • • •• ••••• • • • • ••

native protein

HSF

HSF HSF

Nucleus

transports to endoplasmic reticulum

Hsp70 mRNA

unfolds

2

translation

5 transports to mitochondria

refolds

HSE

Hsp70

to

binds

misfolded protein

3 binds to

lds

refo

transports

native protein Stress-denatured or nascent chain peptides

4

Prevents aggregation

FIG. 1. A schematic representation of the classic cell stress response. This example includes the activation of the inducible form of the 72 kDa heat shock protein (Hsp70), and some of the intracellular roles played by this stress protein are highlighted. (1) A diverse range of physiological signals, including stressors such as hyperthermia, hypoxia, energy depletion, and reactive oxygen species, can activate the heat shock transcription factor (HSF). The activated HSF enters the nucleus and binds to heat shock elements (HSE) in the promoter region of the Hsp70 gene. Hsp70 mRNA is then transcribed and leaves the nucleus for the cytosol, where new Hsp70 is synthesized. (2) In response to a stress, Hsp70 can bind to misfolded proteins in the cytoplasm and assist in the unfolding and subsequent refolding of these proteins to their native form. Hsp70 can also bind to stress-denatured or nascent chain peptides, resulting in (3) refolding of the protein to its native form, or (4) prevention of aggregation of these proteins. (5) Hsp70 can also be bound to stress-denatured or nascent chain peptides, and this complex can then be transported to various organelles, including mitochondria and the endoplasmic reticulum.

8

HAAK & KREGEL

and repair, protein transport to various intracellular organelles, cytoprotection and cell signalling. It is important to note that stress proteins also function in nonstressful conditions, assisting in the folding of nascent polypeptides and acting as monitors of the cellular milieu. Thus, these proteins contribute in vital and varied ways to both cellular homeostasis and survival, and also have physiological effects that impact at tissue, organ and systemic levels. Development of thermotolerance In addition to their protein-repair roles subsequent to an acute stress, numerous studies have demonstrated that HSP induction can protect against ensuing stressors. In the 1970s and 1980s, it was demonstrated that stimulation of Hsp70 production by a mild stress can confer transient protection against subsequent, more intense stress (Landry et al 1982). This phenomenon was initially termed ‘thermotolerance’ because it was demonstrated in experiments involving heat stress. In the classic case of induced thermotolerance, exposure to a mild heat shock can induce stress proteins and other cellular changes to make cells more resistant to a subsequent, more severe (i.e. lethal) heat shock. In general, the degree of conferred thermotolerance is considered directly proportional to the level of HSP induction. The development of thermotolerance consists of a specific series of responses to a moderate level of stress. For instance, within hours of exposure to a preconditioning agent, a cell exhibits increased synthesis of Hsp70, which confers a transitory resistance to protein damage and cell death that lasts for hours to a few days. While initial studies regarding the protective effect of Hsp70 preconditioning were performed under in vitro conditions and focused on heat as a stressor, a succession of studies have shown that resistance to a variety of insults can be developed upon up-regulation of Hsp70 (Kregel 2002). This tolerance phenomenon is also very important in vivo, and involves resistance to stressors (e.g. hypoxia, ischaemia/reperfusion, acidosis, ethanol, radiation) applied at the whole-organism level. Indeed, as will be discussed in subsequent sections, additional reports have linked HSPs to a multitude of physiological functions in animals, thus implying that pre-conditioning could be a potential therapy for a vast range of perturbations. A modern view of the cell stress response While the classic view of the stress response primarily focused on cellular housekeeping functions such as protein synthesis and transport across biological membranes (Fig. 1), a more modern perspective has developed over the past two decades that includes a primary focus on stress proteins that can be categorized

CELL STRESS ODYSSEY

9

as molecular chaperones (Fig. 2). These molecular chaperones, which have also been identified within cells and particular organelles, on cell surfaces, and in extracellular fluids, have now been integrated into physiologically relevant settings and viewed as potential immunomodulators, pro-inflammatory and antiinflammatory signalling molecules, and regulators of cell survival.

Inhibition of apoptotic pathways One important cellular stress response is the initiation of an apoptotic cascade that can ultimately lead to the death of a cell. Hsp70, along with other molecular

Hsp90 Hsp27

3

MHC-1

NK cell

Hsp70 Gp96

4

Hsp60

Hsp60

Gp96

Tcell

Hsp90

5

Hsp27

Hsp27

cytc release

Anti-inflammation ( ↑ IL-10, ↓ TNFa)

Hsp60

Hsp70

Hsp70 Hsp27 Hsp90

2

Hsp60

•••••••••• • • • • ••• •••• • •••••••••••••••••••••••••••• •••••••••••• •••••• • • • • • ••• • • • ••••••••••• • ••••

Apoptotic caspase cascade

6

( ↑ IL-10, ↓ TNFa)

1

Gp96

Hsp27

Gp96

↑ Pro-inflammatory cytokines (TNFa, IL-1, IL-6, IL-12, IL-15)

Hsp90

Hsp27

Hsp60

FIG. 2. A modern view of molecular chaperones and the cell stress response. In this schematic, cellular stress leads to the activation of several different molecular chaperones (e.g. Hsp27, Hsp60, Hsp70, Hsp90, and Gp96), which in turn can stimulate many major signalling pathways. In this scenario, responses can involve the separate activation of selected signalling pathways or a co-ordinated response among many pathways. Responses include: (1) stimulation of pro-inflammatory cytokines (e.g. TNFα , IL1 and IL6); (2) modulation of anti-inflammatory effectors (e.g. stimulating IL10 and inhibiting TNFα ); (3) extracellular release of molecular chaperones; (4) surface expression of molecular chaperones and antigen presentation; (5) immune cell activation and migration associated with surface expression of molecular chaperones; and (6) inhibition of apoptotic pathways via blockade of cytochrome c (cyt c) induced stimulation of apoptotic caspase cascades. Molecular chaperones can also move from selected organelles (e.g. Hsp27 in the nucleus) to other locations within and outside the cell.

10

HAAK & KREGEL

chaperones, is thought to interfere with the apoptotic signalling pathways (Fig. 2), thus leading to cell survival (Garrido et al 2006, Nadeau & Landry 2007). Hsp70 induction also reduces activation of stress kinase c-Jun N-terminal kinase (JNK), a critical activator of the apoptotic death cascade. In turn, reduction of JNK activity suppresses the apoptotic cytochrome c/caspase cascade that can ultimately lead to cell death. These discoveries serve to highlight the fact that chaperone functions initially attributed to Hsp70 were merely the tip of the iceberg, and a multitude of other critical functions involving numerous molecular chaperones are still being unravelled. Paradoxical inflammatory actions of stress proteins In addition to their regulatory roles in apoptotic/cell survival pathways, the majority of molecular chaperones are also involved in signalling processes associated with inflammation pathways. Independent of their protein-binding capacities, these chaperones may also have powerful effects on inflammation via direct cytokine-like activity and these pro-inflammatory properties support the concept of their behaviour as ‘chaperokines’ (Asea et al 2000). However, molecular chaperones also present an apparently paradoxical picture with regard to inflammatory processes, as they can exhibit both pro-inflammatory and anti-inflammatory properties. Though molecular chaperones play a primary role in both protecting and repairing damaged polypeptides, they are also involved in the induction of an inflammatory response in injured cells. Specifically, the Hsp60, Hsp70, Hsp90 and Gp96 families have potent pro-inflammatory effects that are independent of their ability to bind to proteins (Tsan & Gao 2004). These HSPs have been shown to induce production of several pro-inflammatory cytokines, including tumour necrosis factor (TNF) α , interleukin (IL)1, IL6, IL12 and IL15 (Fig. 2). Additionally, Hsp60 and Hsp70 have been linked to increasing T-lymphocyte production of inflammatory cytokines and macrophage production of TNFα . Finally, elevations in circulating levels of HSPs have been found in patients with some types of cancers, cardiovascular disease, renal disease, and acute conditions such as infections (De et al 2000, Johnson & Fleshner 2006, Panayi et al 2004, Wright et al 2000). These extracellular HSPs may be involved in sending a ‘danger signal’ to the rest of the organism, stimulating and sustaining a systemic inflammatory response (Multhoff 2002). Ironically, molecular chaperones have been found to be involved not only in the initiation of inflammation, but also in the suppression of some inflammatory processes (Fig. 2). Both Hsp27 and the HSP known as BiP have been found to induce production of the anti-inflammatory cytokine IL10 (De et al 2000, Panayi et al 2004). Additionally, in sublethal stress conditions, Hsp60 and Hsp70 family

CELL STRESS ODYSSEY

11

members have been implicated in reducing inflammation by suppressing both TNFα production and promoting IL10 induction (Johnson & Fleshner 2006, Moseley 2000). These responses may protect the cell from further damage due to the inflammatory response and potentially contribute to the aforementioned thermotolerance phenomenon. Immune responses While the resolution of these seemingly contradictory reports regarding stress proteins and their impact on inflammation at both cellular and systemic levels will certainly take more investigation, it appears that the extent and mode of chaperone production likely determines the net effect of their influence on the innate inflammatory response. Not only have they been linked to inflammation, but many molecular chaperones have been associated with the adaptive immune response as well. Several molecular chaperones, including Hsp70, Hsp90 and Gp96, have been linked to roles in antigen presentation (Fig. 2), and these proteins are suspected of carrying immunogenic peptides for recognition and activation of the major histocompatability complex class I (MHCI) molecules (Johnson & Fleshner 2006, Pockley 2003, Tsan & Gao 2004). HSPs have also been linked to antigen presentation for cytotoxic T cell activation. Additionally, Hsp70 and others have been found to bind specifically to dendritic cells, which are among the main antigen presenting cells (Asea 2005). Adding to the complexity of the integrated picture of molecular chaperone function, proteins such as Hsp90 and Hsp70 have been identified on the surface of cancer and virally infected cells, in both in vitro and in vivo conditions. It has been postulated that these cell surface locations allow chaperones to stimulate the activation and migration of natural killer cells. Additionally, molecular chaperones have been shown to exhibit strong binding to B cells, which are responsible for antibody production (Asea 2005). These data implicate molecular chaperones as potential key players in stimulating immune responses. Indeed, vaccination of mice with Hsp70, Hsp90 or Gp96 has been shown to stimulate immune responses that suppress tumour progression and promote tumour rejection (Tsan & Gao 2004). As evidence accumulates, it is becoming clear that molecular chaperones have functions well beyond the classical understanding of molecular chaperone protein folding. Potential ramifications of stress protein dysfunction Ironically, while their primary job is to protect the integrity of various proteins in the cell, stress proteins can also fall victim to the same variety of insults that cause their target proteins to malfunction. Alterations in the structural integrity of

12

HAAK & KREGEL

molecular chaperones can disrupt their function and potentially lead to pathological conditions that have been termed ‘chaperonopathies’ (Macario & Conway de Macario 2005). These chaperonopathies, which are implicated in a wide range of disease states, are postulated to develop through mutations or post-translational defects (Fig. 3). The modification of molecular chaperones at post-translational levels can occur through a variety of mechanisms and result in the loss of chaperone function. In a condition that has been termed ‘acquired chaperonopathy,’ alteration of these molecular chaperones can lead to an accumulation of protein damage and protein aggregation. This type of pathology has been linked to a number of neurological disorders that are associated with aggregates of various proteins (e.g. Parkinson’s, Huntington’s and Alzheimer’s disease) (Meriin & Sherman 2005). However, not all chaperonopathies involve structural modification of the chaperones themselves. Many pathological states that have been connected with these proteins are more specifically targeted to gene regulation of chaperone expression. In this scenario, the proteins themselves are structurally normal, but they are improperly expressed, resulting in either elevations or reductions in their levels at intracellular and extracellular locations. The mechanisms responsible for these alterations are unclear, but are postulated to be

Chaperone Pathologies

Examples Alzheimer’s Parkinson’s Huntington’s Retinopathies Cataracts

Protein aggregation

Examples

Post-translational modification Loss of function

Impaired immune response

Altered gene regulation Abnormal protein expression

Decreased ability to cope with stressors

Altered cell signalling

Cancer Aging Cardiopathy Neuropathy

Damage accumulation

FIG. 3. Examples of chaperone pathologies. Different mechanisms at the cellular level can contribute to modifications in molecular chaperones, resulting in a variety of pathophysiological conditions and functional alterations that are manifested at cellular and systemic levels.

CELL STRESS ODYSSEY

13

related to possible mutations in heat shock transcription factors. Perhaps most relevant, abnormal expression of HSPs has been associated with several conditions involving physiological decline, including ageing, neurodegenerative disorders and cardiopathies. Ageing is one of the primary biological processes that has been associated with decreased stress protein expression. In addition, the ability of an organism to upregulate stress proteins in response to stress is altered with advancing age (Arslan et al 2006, Kregel 2002). For instance, our own laboratory has shown an altered HSP expression in response to stress when comparing aged and young animals (Hall et al 2000, Kregel 2002). This blunted cell stress response can have broad physiological ramifications, leading to an accumulation of protein damage, reductions in repair processes, functional declines at cellular and systemic levels, and susceptibility to further insult (Zhang et al 2003). Loss of chaperone function can also lead to impairment of the immune response (Pockley 2003). As was noted in earlier sections of this chapter, molecular chaperones play various roles in regulating the immune response; therefore, alteration of their function, whether through damage to the protein itself or changes in protein levels due to gene control, could have dire implications for the immune response to a stressor. Atypical levels of several molecular chaperones have also been found in a broad range of tumour cells and cancerous tissues. In fact, tumour cells appear to be dependent on increased HSP expression for survival and growth. It is thought that because these stress proteins play a role in the control of cell growth, and more specifically, the inhibition of apoptosis, they may aid in tumour formation and proliferation (Jolly & Morimoto 2000, Nollen & Morimoto 2002, Soti et al 2005). Overexpression of several molecular chaperones has been found to result in tumour formation, and these provocative observations further implicate altered HSP expression in oncogenesis (Calderwood et al 2006). Conclusion Due to the wide range of functions ascribed to molecular chaperones and the myriad conditions associated with stress protein dysfunction (Fig. 3), it is no surprise that these proteins are emerging as potential therapeutic targets in a broad range of fields. Current and future research will examine whether manipulation of molecular chaperones can successfully treat or prevent conditions ranging from ageing to cancer. As the molecular chaperone field moves forward, many opportunities will present themselves to investigators. The integration of disciplines such as genomics, molecular biology, systems biology, and translational medicine will provide scientists with an array of tools to aid in the delineation of diagnoses and treatments of numerous diseases and pathological conditions.

14

HAAK & KREGEL

Acknowledgements The Kregel Laboratory is supported by funding from the National Institutes of Health (AG12350). The authors acknowledge the expert secretarial help of Ms. Joan Seye.

References Arslan MA, Csermely P, Soti C 2006 Protein homeostasis and molecular chaperones in aging. Biogerontology 7:383–389 Asea A 2005 Stress proteins and initiation of immune response: chaperokine activity of hsp72. Exerc Immunol Rev 11:34–45 Asea A, Kraeft SK, Kurt-Jones EA et al 2000 HSP70 stimulates cytokine production through a CD14-dependant pathway, demonstrating its dual role as a chaperone and cytokine. Nat Med 6:435–442 Benjamin IJ, McMillan DR 1998 Stress (heat shock) proteins: molecular chaperones in cardiovascular biology and disease. Circ Res 83:117–132 Calderwood SK, Khaleque MA, Sawyer DB, Ciocca DR 2006 Heat shock proteins in cancer: chaperones of tumorigenesis. Trends Biochem Sci 31:164–172 De AK, Kodys KM, Yeh BS, Miller-Graziano C 2000 Exaggerated human monocyte IL-10 concomitant to minimal TNF-alpha induction by heat-shock protein 27 (Hsp27) suggests Hsp27 is primarily an antiinflammatory stimulus. J Immunol 165:3951–3958 Ellis RJ 1987 Proteins as molecular chaperones. Nature 328:378–379 Ellis RJ 2005 Chaperone function: the orthodox view. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, New York, p 3–21 Ferrarini M, Heltai S, Zocchi MR, Rugarli C 1992 Unusual expression and localization of heatshock proteins in human tumor cells. Int J Cancer 51:613–619 Frostegard J, Pockley AG 2005 Heat shock protein release and naturally occurring exogenous heat shock proteins. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, New York, p 195–219 Garrido C, Brunet M, Didelot C, Zermati Y, Schmitt E, Kroemer G 2006 Heat shock proteins 27 and 70: anti-apoptotic proteins with tumorigenic properties. Cell Cycle 5:2592–2601 Hall DM, Xu L, Drake VJ, Oberley LW, Oberley TD, Moseley PL, Kregel KC 2000 Aging reduces adaptive capacity and stress protein expression in the liver after heat stress. J Appl Physiol 89:749–759 Henderson B, Shamaei-Tousi A 2005 Molecular chaperones: the unorthodox view. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, New York, p 78–96 Hightower LE 1991 Heat shock, stress proteins, chaperones, and proteotoxicity. Cell 66:191– 197 Johnson JD, Fleshner M 2006 Releasing signals, secretory pathways, and immune function of endogenous extracellular heat shock protein 72. J Leukoc Biol 79:425–434 Jolly C, Morimoto RI 2000 Role of the heat shock response and molecular chaperones in oncogenesis and cell death. J Natl Cancer Inst 92:1564–1572 Kregel KC 2002 Heat shock proteins: modifying factors in physiological stress responses and acquired thermotolerance. J Appl Physiol 92:2177–2186 Landry J, Bernier D, Chretien P, Nicole LM, Tanguay RM, Marceau N 1982 Synthesis and degradation of heat shock proteins during development and decay of thermotolerance. Cancer Res 42:2457–2461 Lindquist S, Craig EA 1988 The heat-shock proteins. Annu Rev Genet 22:631–677 Macario AJ, Conway de Macario E 2005 Sick chaperones, cellular stress, and disease. N Engl J Med 353:1489–1501

CELL STRESS ODYSSEY

15

Meriin AB, Sherman MY 2005 Role of molecular chaperones in neurodegenerative disorders. Int J Hyperthermia 21:403–419 Morimoto RI, Tissieres A, Georgopoulos C 1994 Progress and perspectives on the biology of heat shock proteins and molecular chaperones. In: Morimoto RI, Tissieres A, Georgopoulos C (eds) The biology of heat shock proteins and molecular chaperones. Cold Spring Harbor Press, Cold Spring Harbor, NY, p 1–30 Moseley P 2000 Stress proteins and the immune response. Immunopharmacology 48:299– 302 Multhoff G 2002 Activation of natural killer cells by heat shock protein 70. Int J Hyperthermia 18:576–585 Nadeau SI, Landry J 2007 Mechanisms of activation and regulation of the heat shock-sensitive signaling pathways. Adv Exp Med Biol 594:100–113 Nollen EA, Morimoto RI 2002 Chaperoning signaling pathways: molecular chaperones as stress-sensing ‘heat shock’ proteins. J Cell Sci 115:2809–2816 Panayi GS, Corrigall VM, Henderson B 2004 Stress cytokines: pivotal proteins in immune regulatory networks; Opinion. Curr Opin Immunol 16:531–534 Pockley AG 2003 Heat shock proteins as regulators of the immune response. Lancet 362:469–476 Ritossa F 1962 A new puffi ng pattern induced by temperature shock and DNP in Drosophila. Experientia 18:571–573 Soti C, Nagy E, Giricz Z, Vigh L, Csermely P, Ferdinandy P 2005 Heat shock proteins as emerging therapeutic targets. Br J Pharmacol 146:769–780 Tissieres A, Mitchell HK, Tracy UM 1974 Protein synthesis in salivary glands of Drosophila melanogaster: relation to chromosome puffs. J Mol Biol 84:389–398 Tsan MF, Gao B 2004 Cytokine function of heat shock proteins. Am J Physiol Cell Physiol 286:C739–744 Welch WJ 1992 Mammalian stress response: cell physiology, structure/function of stress proteins, and implications for medicine and disease. Physiol Rev 72:1063–1081 Wright BH, Corton JM, El-Nahas AM, Wood RF, Pockley AG 2000 Elevated levels of circulating heat shock protein 70 (Hsp70) in peripheral and renal vascular disease. Heart Vessels 15:18–22 Zhang HJ, Xu L, Drake VJ, Xie L, Oberley LW, Kregel KC 2003 Heat-induced liver injury in old rats is associated with exaggerated oxidative stress and altered transcription factor activation. FASEB J 17:2293–2295

DISCUSSION Macario: One important thing we have learned from the study of chaperones is that some of them are highly conserved in evolution. Hsp70, which is supposed to be one of the most conserved proteins of all, was not present in some organisms of the domain Archaea (Macario et al 1999, Macario & Conway de Macario 2001). The parallel development in the study of stress proteins and in the study of phylogeny, applying molecular techniques and leading to the discovery of three major evolutionary domains, Archaea, Bacteria and Eukarya (eukaryotes), revealed that in one of these domains, the Archaea, some organisms do not have the Hsp70 system. We also learned that when Hsp70(DnaK) occurs in a genome, the other two components of the chaperone machine, Hsp40(DnaJ) and GrpE, are also always present. If one is missing, so are the other two (Macario & Conway de

16

HAAK & KREGEL

Macario 1999). Interestingly, and rather puzzlingly, within the domain Archaea, only the Crenarcheota do not have Hsp70. Within the other Archaea, belonging to other sub-Domains (e.g. the Euryarchaeota), those that grow at 70 °C or higher do not have the genes (Macario et al 2004). The others that grow at lower temperatures have the genes. In addition, there are some genera that don’t have the gene, regardless of the temperature of their environment. In contrast, all bacteria have this system (Macario et al 2006). The point I am making is this: why is there this sharply defined absence, and what replaces this important system in those organisms that don’t have it, but otherwise look more-or-less the same? Ellis: The suggestion made by the Hartl laboratory is that the role of Hsp70 in those Archaea that lack this chaperone is taken by the protein prefoldin. This suggestion fits the scenario that says that Archaea and Bacteria originated independently, which is a view proposed by Martin & Russell (2003). The occurrence of Hsp70 in some Archaea can be explained by saying this was acquired later from Bacteria by horizontal gene transfer. Prefoldin occurs also in the eukaryotic cytosol and its crystal structure is known: it is quite unlike any other chaperone, but it does seem to do the same sort of job as Hsp70 (Siegert et al 2000). Csermely: Let me broaden the question a little. In bacteria, not only is Hsp70 sometimes missing, but some species also lack the GroE system (Wong & Houry 2004). Ellis: The GroEL system seems to be specific for a particular class of proteins that are especially prone to aggregation. The αβ domain proteins are a major component of these. Those bacteria which lack GroE have small genomes, so it might be that they don’t make this type of protein that suffers from the aggregation problem. Has anyone done a survey of the 85 proteins in Escherichia coli that absolutely require GroEL in those organisms? Lund: Of the organisms that are missing GroEL, three are mycoplasmas. The proteins that are folded by GroEL in E. coli are certainly present in these mycoplasmas: things like EF-Tu, which you can’t imagine an organism not having. This doesn’t mean that they can’t form independently of GroEL in those organisms. There is a research effort in Munich to take some of these proteins from the mycoplasmas, put them in E. coli and see whether they are still GroEL independent. Ellis: It might come down to the individual propensity to aggregate; single residue changes can affect this propensity. Csermely: Do any higher organisms lack some of these chaperone classes, or is it specific to bacteria? Ellis: There is the famous case of the Hsp104 family, which according to genome sequence comparisons, is apparently missing from all metazoa, but is present in all other organisms. This is the only known chaperone family that can dissolve insoluble protein aggregates. One speculation is that nervous systems have a particular problem with protein aggregation. An interesting experiment

CELL STRESS ODYSSEY

17

would be to put Hsp104 into a neuronal cell line in which you could induce some aggregation. Cheetham: I have heard the argument that Hsp104 is present in organisms that can’t walk away from stress, as an extreme recovery mechanism. Ellis: There is a complication here. In yeast, Hsp104 is required to initiate amyloid fibril formation so if the Hsp104 gene is deleted, amyloid fibril formation is prevented. However, if you overexpress Hsp104, you don’t get amyloid fibril formation either. So it appears that a critical level of Hsp104 is needed to allow protein aggregation. Here you have an example where a chaperone potentiates a disease condition. The suggestion is that for the amyloid fibrils to spread, you need continual seeds to be produced, and these seeds are produced by splitting of the fibrils—a process that Hsp104 stimulates (Tanaka et al 2006). So Hsp104 might be a bad chaperone to have in the brain because it makes the aggregation problem worse than it already is. Hightower: Work done in the early 1980s by Michael Tytell showed that Hsp70 was one of a number of proteins he dubbed as glial axon transfer proteins. In his model of the squid giant axon the glial cells were capable of transferring these proteins to neuronal cells. This was one of the first indications that stress proteins can get out of one cell and into another. In one of Kevin Kregel’s very nice time lines, he mentions that stress proteins were shown on cell surfaces in the 1990s. Since we will be talking a lot about the extracellular appearance of chaperones one might as well broach the issue of how this was greeted. It was greeted with broad disbelief in the 1990s, but not by everyone. There were people who weren’t surprised. One reason is that there was work in the late 1970s done by Susan Lindquist (Velazquez et al 1980, Velazquez & Lindquist 1984) showing that in cultured Drosophila cells Hsp70 was at or near the plasma membrane. It was a piece of her study that was missed for a long time. Then in the early 1980s Michael Tytell showed that some of the stress proteins could be transferred from cell to cell (Tytell & Lasek 1984, Tytell et al 1986), and in the late 1980s my student Peter Guidon and I showed that these proteins could be released from cultured newborn rat cells. In fact, they were easily released: all we had to do was change the culture medium and these were released. If cells were stressed, then some of the heat shock proteins such as stress-inducible Hsp70 joined other proteins such as the cognate protein Hsc70 at being released. Actin was also released. These were released by non-endoplasmic reticulum (ER)/Golgi-dependent mechanisms (Hightower & Guidon 1989). To get this paper published, satisfying the reviewers with controls took longer than the original work! Finally, they couldn’t come up with sufficient objections not to publish. These historical works didn’t physically show the association on the outside of the cell, but they showed that these proteins could get out of the cells by a mechanism other than cell necrosis.

18

HAAK & KREGEL

Cheetham: The Mike Tytell giant squid axon work is interesting (Tytell et al 1986). You have to think about the physical size of the neuron and how it is very difficult to transcribe mRNA in the nucleus, make protein in the cell body, and transport the protein down the axon. It is much quicker to use the cells nearby to make the protein and then transfer it. Recently, people have followed this up. Boris Margulis and Irena Guzhova have shown that the glia can release Hsp70 and neurones can take up Hsp70, and the proteins have protective effects against a variety of stresses (Guzhova et al 2001, Novoselova et al 2005). Ellis: I take your point. But I was talking to Roger Morris recently who was making the point that exactly where a protein is made in the strengthening of a synapse is vital. It has to end up in the right synapse, so where the protein is made could be crucial. He was suggesting that such a protein isn’t made in the cell body but is actually made close to the synapse, using mRNA that is resident at that synapse. Cheetham: I agree there are well documented mechanisms for transport, but there is a difference between synaptic plasticity with remodelling of neuronal connections, and what a cell has to do in a crisis. In a crisis you don’t want the fire brigade to be coming from the next city. Henderson: Brett Nixon, do you want to say anything about the finding of Hsp60 in the sperm, in the process of capacitation? Nixon: We have found that during the activation of mouse spermatozoa, Hsp60 and Hsp90b1 become surface expressed. This event is apparently species specific as we do not find similar surface expression of these proteins on human spermatozoa. We are trying to elucidate the mechanisms underlying their delivery to the mouse sperm surface and their function in these cells. Our hypothesis is that they are involved in the assembly of a receptor complex to enable the sperm to bind to the oocyte. Consistent with this hypothesis, the surface expressed chaperones are present in lipid raft microdomains, and these structures undergo a lateral migration to the site of sperm interaction with the egg at the time of fertilization. But we are puzzled why this is true in the mouse but not the human, particularly as the mouse and human forms of Hsp60 share 97% amino acid sequence identity. Ellis: Have you tried adding the proteins from the outside? Nixon: No. Ellis: Do you think they come from the inside to the outside, or are they coming across the membrane, released and then reabsorbed? Nixon: Our evidence suggests that they are moving from the inside to the outside of the spermatozoa. Ellis: Has anyone tried sampling seminal fluid minus sperm to see what is in it? Nixon: Interestingly, as the mouse spermatozoa mature within the epididymis, we find they come in close apposition to dense bodies in the epididymal fluid that

CELL STRESS ODYSSEY

19

contain Hsp60 and Hsp90b1. We believe that these structures may facilitate the simultaneous transfer of the chaperones and possibly client proteins to the sperm during the transit through this organ. Incidentally, this interaction occurs in the precise region of the epididymis in which mouse sperm develop the potential to bind to and fertilize an oocyte. However, the chaperone proteins are not detected on the surface of the sperm at this stage of their development. Multhoff: Clifford Lingwood found Hsp70 on sperm surfaces, and found that this is working like an adhesion molecule for making contact with the egg (Boulanger et al 1995). Nixon: We have added the Hsp60 antibodies to mouse spermatozoa to try to block their interaction with oocytes, but this approach does not inhibit the interaction. On the basis of these data we believe the chaperones may play an indirect role in sperm–oocyte interaction by mediating the productive assembly of other receptors into a functional complex. Ellis: My problem with that would be how Hsp60 would work on the outer leaflet of the membrane. Nixon: What we are finding is that there are a number of sperm proteins implicated in the sperm–oocyte interaction. The existing paradigm of a simple lock and key mechanism fails to account for the fact that the majority of knockout mice generated for putative sperm–oocyte receptors actually retain their fertility. We now think there are multiple recognition molecules that are somehow involved in the interaction. We think that the chaperones are somehow piecing these together to enable the sequential interaction between different sperm receptors and oocyte ligands to take place. Ellis: There is some interesting work showing that there are intrinsic membrane proteins that have chaperone activity, in the sense that they prevent the aggregation of other multiple-spanning proteins that are being inserted into the membrane laterally from the insert site (Kota & Ljungdahl 2005). But this chaperone activity is taking place within the lipid environment, whereas I presume your proteins are on the outside of the leaflet. Nixon: At present we do not know anything about the orientation of the chaperones within the sperm membrane but I do not know of any evidence to suggest that chaperones do become incorporated through the membrane. However, there is evidence to suggest that chaperones such as Hsp90 are present on the surface of certain cancer cells and can promote tumour invasiveness (Eustace & Jay 2004). Given that Hsp90 requires ATP for its activity, I wonder how that could be the case. Ellis: We know that chaperones have particular functions in the cell, but we have to be careful not to assume that once they are on the cell surface the same chaperone functions are active. Proteins are large compared to their active sites: it could be that there are several different active sites on a given chaperone but

20

HAAK & KREGEL

that we know about only some of them. So it could be that what is important here is not the chaperone function that operates inside the cell, but some other function. Lund: I’d like to ask Kevin about the point you made about the decline in the ability of cells to respond to heat shock with age. You raised a point that it is probably going to be changes in regulation that will be more significant as far as pathology is concerned. Is this a developmental programme, or is it due to the accumulations of mutations in the regulators or binding regions? And do you see the same thing in simpler organisms in which there are good models for ageing? Kregel: My take on this is that the data in ageing systems have been mostly descriptive. These are observational studies. The mechanistic questions surrounding these changes in gene expression have not been teased out. In terms of the model systems, people working on Drosophila and Caenorhabditis elegans seem to have moved to a variety of different stress proteins. I’ve not seen a lot with the ageing: people are more focused on things like oxidative stress as a mechanism for ageing. We have been doing a lot of studies related to this. There is a black box surrounding some of the changes that occur with the ageing processes, where there is macromolecular damage that may be caused by reactive oxygen species (ROS) generation, and how this may stimulate or blunt HSP responses. It is a pretty integrated picture, but people working on ageing have rather glossed over the stress protein responses at this point. Lund: I guess there is the broader question: you alluded to the fact that the cell stress responses are global. What is the nature of the damage that they are responding to? The paradigm for HSPs is that they recognize unfolded proteins, but the way oxidative stress damages proteins is different from the way that heat shock damages proteins. Are there transduction routes to gene expression that we are not looking at at the moment? Kregel: I think this is the case. Some of these ROS species are cell signalling molecules. Perhaps they come into play in this scenario in terms of stimulating some of these cell stress responses. This integrated system of cell signalling when there is a cell stress will probably translate to a point where people can decipher the stimulation of some of these chaperones. If the free radical theory of ageing suggests that there are more ROS, how does this stimulate a cell signalling response that may manifest differently in terms of the chaperone expression? Pockley: The effect of ageing on the stress response is very interesting, particularly when we are looking at circulating proteins. We and others have done some work on this (Rea et al 2001, Jin et al 2004, Terry et al 2004). The limitation is that these are cross-sectional and not longitudinal studies, looking at circulating levels of these proteins with ageing. If we look in the elderly group, and ours was 90 years plus (Rea et al 2001), these otherwise healthy people have much lower

CELL STRESS ODYSSEY

21

levels of these proteins. If we combine these findings with those from Brian Henderson’s lab, which has looked at levels of Hsp60 in a longitudinal fashion in healthy individuals and found that levels are stable within an individual over time, then there has to be some fundamental change in the response of the elderly to what would appear to be a continual exposure to various stresses. We don’t know the physiological consequence of this. In some instances circulating levels of these proteins are associated with ongoing inflammatory responses such as cardiovascular disease, although counterintuitively, one study from our lab shows that high levels of circulating Hsp70 appear to protect against the subsequent development of atherosclerosis in individuals with established hypertension (Pockley et al 2003). The picture is confusing: we need a handle on what is driving the maintenance of this level of stress proteins. There is an incredible range of expression of Hsp60, for example, from zero to micrograms per millilitre, however within an individual this is stable over a period of time. Kregel: One challenge in this kind of scenario involves the evaluation of stress protein responses in healthy aged individuals. If we take a ‘snapshot’ of stress proteins in this age group, and if the system is not perturbed, it is very difficult to tease out what is occurring at either the cellular or systemic level. The situation in which these older individuals are stressed, and where some of these molecular chaperones need to turn on, could be the best circumstance to determine if differences are manifested with ageing. An obvious limitation with these types of ageing studies is that most are cross-sectional. Thus, if it is necessary to perturb a physiological system in order to assess chaperone function, one challenge is to apply similar levels of stress in the different age groups in order to make valid comparisons. Henderson: I’d like to add some confusion to this whole picture. We have looked at a cohort of 250 teenagers. Normally we find that half of the adult populations we measure have measurable Hsp60. In these teenagers it is about 25%. There’s loss with age, and loss with youth as well! References Boulanger J, Faulds D, Eddy EM, Lingwood CA 1995 Members of the 70 kDa heat shock protein family specifically recognize sulfoglycolipids: role in gamete recognition and mycoplasma-related infertility. J Cell Physiol 165:7–17 Eustace BK, Jay DG 2004 Extracellular roles for the molecular chaperone, hsp90. Cell Cycle. 3:1098–1100 Guzhova I, Kislyakova K, Moskaliova O et al 2001 In vitro studies show that Hsp70 can be released by glia and that exogenous Hsp70 can enhance neuronal stress tolerance. Brain Res 914:66–73 Hightower LE, Guidon PT Jr 1989 Selective release from cultured mammalian cells of heatshock (stress) proteins that resemble glia-axon transfer proteins. J Cell Physiol 138:257– 266

22

HAAK & KREGEL

Jin X, Wang R, Xiao C et al 2004 Serum and lymphocyte levels of heat shock protein 70 in aging: a study in the normal Chinese population. Cell Stress Chaperones 9:69–75 Kota J, Ljungdahl PO 2005 Specialized membrane-localized chaperones prevent aggregation of polytopic proteins in the ER. J Cell Biol 168:79–88 Macario AJL, Conway de Macario E 1999 The archaeal molecular chaperone machine: peculiarities and paradoxes. Genetics 152:1277–1283 Macario AJL, Conway de Macario E 2001 The molecular chaperone system and other antistress mechanisms in archaea. Front Biosci 6:d262–283 Macario AJL, Lange M, Ahring BK, Conway de Macario E 1999 Stress genes and proteins in the Archaea. Microbiol Mol Biol Rev 63:923–967 Macario AJL, Malz M, Conway de Macario E 2004 Evolution of assisted protein folding: the distribution of the main chaperoning systems within the phylogenetic domain Archaea. Front Biosci 9:1318–1332 Macario AJL, Brocchieri L, Shenoy A, Conway de Macario E 2006 Evolution of a proteinfolding machine: genomic and evolutionary analyses reveal three lineages of the archaeal hsp70(dnaK) gene. J Mol Evol 63:74–86 Martin W, Russell MJ 2003 On the origins of cells: a hypothesis for the evolutionary transitions from abiotic geochemistry to chemoautotrophic prokaryotes and from prokaryotes to nucleated cells. Philos Trans R Soc Lond B Biol Sci 358:59–85 Novoselova TV, Margulis BA, Novoselov SS et al 2005 Treatment with extracellular HSP70/ HSC70 protein can reduce polyglutamine toxicity and aggregation. J Neurochem 94:597–606 Pockley AG, Georgiades A, Thulin T, de Faire U, Frostegård J 2003 Serum heat shock protein 70 levels predict the development of atherosclerosis in subjects with established hypertension. Hypertension 42:235–238 Rea IM, McNerlan S, Pockley AG 2001 Serum heat shock protein and anti-heat shock protein antibody levels in aging. Exp Gerontol 36:341–352 Siegert R, Leroux MR, Scheufler C, Hartl FU, Moarefi I 2000 Structure of the molecular chaperone protein prefoldin: unique interaction of multiple coiled coil tentacles with unfolded proteins. Cell 103:621–632 Tanaka M, Collins SR, Toyama BH, Weissman JS 2006 The physical basis of how prion conformations determine strain phenotypes. Nature 442:585–589 Terry DF, McCormick M, Andersen S et al 2004 Cardiovascular disease delay in centenarian offspring: role of heat shock proteins. Ann N Y Acad Sci 1019:502–505 Tytell M, Lasek RJ 1984 Glial polypeptides transferred into the squid giant axon. Brain Res 324:223–232 Tytell M, Greenberg SJ, Lasek RJ 1986 Heat shock-like protein is transferred from glia to axon. Brain Res 363:161–164 Velazquez JM, Lindquist S 1984 Hsp70 nuclear concentration during environmental stress and cytoplasmic storage during recovery. Cell 36:655–662 Velazquez JM, DiDomenico BJ, Lindquist S 1980 Intracellular location of heat shock proteins in Drosophila. Cell 20:679–689 Wong P, Houry WA 2004 Chaperone networks in bacteria: analysis of protein homeostasis in minimal cells. J Struct Biol 146:79–89

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

The chaperone function: meanings and myths Peter A. Lund and R. John Ellis*1 School of Biological Sciences, University of Birmingham, Birmingham B15 2TT and *Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, UK

Abstract. What are molecular chaperones and how should we think about them? We propose that it is better to think in terms of a chaperone function rather than in terms of chaperone molecules. We define the chaperone function as the prevention or reversal of incorrect interactions that may occur when reactive macromolecular surfaces are transiently exposed to the intracellular environment. We suggest that this function is a distinct and essential cellular function, mediated by many different proteins. Chaperones have evolved to reduce, by a variety of mechanisms, the aggregation of proteins into non-functional, and sometimes cytotoxic, structures. Chaperones may also have evolved to have additional roles. A cellular or extracellular event mediated by a chaperone protein is not necessarily a consequence of that protein’s chaperone function. The aim of this article is to provide a brief summary of the origin and concepts used in the intracellular chaperone field, to provide a backdrop for discussion of their possible roles outside the cell. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 23–44

Protein self-assembly—a changing paradigm Proteins evolved to function inside cells, and (later) outside them, but their properties are commonly studied after purification and exposure to simpler, and rather different, conditions. Anfinsen received the Nobel Prize in 1972 for discovering that some purified denatured proteins will refold when the concentration of denaturant is lowered (Anfinsen 1973). We know now that most denatured proteins are capable of refolding in dilute media, needing neither other macromolecules nor an added energy source. This is also true of the assembly of some large multi-subunit structures such as ribosomes and certain viruses. These observations have been codified in the statement that proteins are capable of spontaneous self-assembly, the term ‘assembly’ being used to cover the folding of individual polypeptide 1

This paper was presented at the symposium by John Ellis, to whom correspondence should be addressed. 23

24

LUND & ELLIS

chains and any subsequent association into oligomers. The principle of protein self-assembly states that all the information needed for a protein chain to reach its correct, biologically active conformation is encoded in its aminoacyl sequence. Following Anfinsen, it was assumed that protein folding inside the cell would also not require any additional cofactors or energy, but studies initiated in the late 1970s challenged this idea. Some proteins were found to require ‘helper’ proteins for folding or assembly, and some of these helpers required ATP hydrolysis to function. The term ‘molecular chaperone’ (often abbreviated to ‘chaperone’) was proposed to describe these proteins (Laskey et al 1978, Ellis 1987), and the concept was advanced that a chaperone function is required in all cells (Ellis & Hemmingsen 1989). Thus the original view that proteins fold and assemble to their lowest free energy state inside the cell by a process of spontaneous self-assembly has been replaced by the concept of assisted self-assembly. This replacement does not change the basic concept of protein self-assembly. Proteins contain all the information required for their correct folding and assembly, but their problem is to avoid aggregation inside the cell. The protein aggregation problem We define protein aggregation as the association of two or more polypeptide chains into non-functional structures, where the proteins are not folded into their active conformations. Because aggregation is at least a second-order process, it is dependent on the concentration of interacting chains. All cells require a chaperone function to combat aggregation because two universal features of the intracellular environment where proteins fold and assemble favour aggregation, namely the synthesis of proteins on polyribosomes and the occurrence of macromolecular crowding. Polyribosomes The rate of polypeptide chain synthesis is typically between 5 and 20 amino acid residues added per second, but the rate of protein folding is often much faster, with some proteins able to fold in less than one second. Thus incomplete chains emerging from ribosomes may misfold into non-functional conformations before an entire folding domain has been made. Moreover, one molecule of messenger RNA binds more than one ribosome at a time, forming a polyribosome. This arrangement results in partly folded identical chains coming within touching distance as they are synthesized, enhancing the possibility that these chains may interact to form aggregates. Thus the advantage of making several polypeptide chains simultaneously from one molecule of mRNA is bought at the risk of both misfolding and aggregation.

CHAPERONES

25

Macromolecular crowding This term describes the fact that the total concentration of macromolecules in all intracellular compartments where proteins fold and assemble is very high, in the range 80–400 g/l. Such a degree of crowding increases macromolecular association constants to values that can be two to three orders-of-magnitude greater than those in uncrowded media (Ellis 2001, Ellis & Minton 2006). Such association reactions include protein aggregation. In protein refolding experiments a fraction of the chains commonly aggregate with one another, even though such experiments are almost always done in uncrowded media. Refolding experiments show that aggregation is sensitive to both the protein concentration and the temperature, so protein chemists solve this problem by reducing these parameters. Cells cannot do this but rather employ molecular chaperones to combat the enhanced tendency to aggregate caused by macromolecular crowding.

Origin of the molecular chaperone concept The term ‘molecular chaperone’ first appeared in 1978 to describe the properties of a nuclear protein called nucleoplasmin found in amphibian eggs (Laskey et al 1978), that assists the assembly of nucleosomes from histones and DNA when these eggs divide rapidly after fertilization. The role of nucleoplasmin is transient, because it is not a component of the assembled nucleosome. The term ‘chaperone’ is appropriate because the function of the human chaperone is also to prevent incorrect interactions between the transiently exposed surfaces of human couples, but without being present on the wedding night. In the 1980s two reports demonstrated that some newly synthesised polypeptide chains bind to other pre-existing proteins before assembly into functional complexes. One (Barraclough & Ellis 1980) led later to the more general concept of the molecular chaperone function (Ellis 1987). The second (Ostermann et al 1989) established that at least one family of chaperone functions at the level of protein folding rather than protein assembly, requiring ATP hydrolysis to do so, and initiated the continuing wave of research on folding chaperones. The high degree of conservation of some chaperones between very diverse organisms came to be recognized, and genetic studies on bacteria and yeast showed that some of these conserved chaperones were important or even essential for growth (e.g. Hemmingsen et al 1988, Kusukawa & Yura 1988, Fayet et al 1989, Normington et al 1989, Ursic & Culbertson 1991). The term ‘molecular chaperone’ is now in general use, but is sometimes used uncritically by those unfamiliar with the origins of the concept and the definitions that have been proposed.

26

LUND & ELLIS

Definitions The English philosopher Cyril Joad (1891–1953) was often criticized for always responding to questions with the catchphrase, ‘It all depends on what you mean by. . . . ’ But of course he was right. Science is basically a set of ideas about how the world works, and these ideas are expressed in words. So it is not a semantic quibble to insist on agreed definitions of terms that are used to discuss new ideas. The molecular chaperone function is currently defined as the ability of all cells to prevent and/or reverse incorrect interactions that may result when potentially interactive macromolecular surfaces are exposed to the intracellular environment (Ellis 2006). These surfaces may occur on growing and newly released polypeptide chains, on mature proteins unfolded by certain environmental stresses, and on folded proteins in near-native and native conformations. The same concept applies to other macromolecules that can undergo incorrect interactions, especially RNA. Molecular chaperones are currently defined as a large and diverse group of proteins that share the functional property of assisting the non-covalent folding/unfolding and the assembly/disassembly of other macromolecular structures, but are not components of these structures when these are performing their normal biological functions (Ellis 2006). Some molecular chaperones, but by no means all, are also heat shock proteins (HSPs), because the requirement for chaperone function increases under conditions such as heat shock that cause some proteins to unfold and aggregate. Thus, molecular chaperones and stress response proteins are overlapping sets, not identical sets. It is formally possible that some chaperone functions reside in macromolecules other than protein, but the evidence for this as yet is not compelling (Ellis 1997). The ultimate aim of research on chaperones is to clarify their role in cellular chaperone function, which often requires the study of networks of chaperones rather than individual proteins (Lund 2001). The above definition is entirely functional and contains no constraints on the mechanisms by which different chaperones may act. The term ‘non-covalent’ is used to exclude those proteins that carry out post-translational covalent modifications (though some proteins, such as protein disulphide isomerase, may do both). It is better to think of a molecular chaperone as a function rather than as a molecule; on this basis a chaperone function can be a property of a molecule that has additional and quite different functions. Co-chaperones are proteins that modulate chaperone activity. Some co-chaperones are also chaperones (e.g. HSP40). There are ever-increasing reports of cochaperone functions, especially with respect to HSP70 and HSP90 chaperones, and it is clear that different co-chaperones can modulate the properties of their cognate chaperones in different ways (Song & Morimoto 2001, Smith 2001).

CHAPERONES

27

Categories of chaperones Chaperones can be grouped in many different ways: by size, by mechanism (including whether or not they require ATP hydrolysis), by whether they are involved in folding or in assembly, by whether they act co- or post-translationally, by whether they bind one or several substrates, or phylogenetically. As our appreciation of the importance of chaperone function grows, so does the number of proteins that can be called chaperones; at the same time, the distinctions within and between some of the above categories become fuzzier. On the basis of sequence similarity many of the major known chaperones whose main function is folding fall into five clear groups, which are both broadly distributed and in general highly conserved across all three domains of life (though less so in the Archaea than in the Bacteria and Eukarya). These groups are often referred to as the HSP100, HSP90, HSP60, HSP70 and sHSP groups (derived from the observation that many are heat shock proteins, though importantly this is not the case for all members of these groups). Several of these groups are associated with co-chaperones which have chaperone properties in their own right, such as the HSP40 or DnaJ-proteins, a broad and important group associated with the HSP70 proteins. In addition, there are several other well-known chaperones which are more domain-specific, such as the ribosome-associated chaperones trigger factor (found in bacteria) and nascent-chain associated complex (NAC) (found in archaea and eukaryotes), or the ER-resident chaperones calnexin and calreticulin, unique to eukaryotes. Other chaperones (such as Chs7p and Shr3p) appear to prevent the aggregation of polytopic membrane proteins. Table 1 lists some of the chaperones that assist protein folding in various intracellular compartments in the three domains of life. It is important to note that as well as assisting protein folding, many examples exist in these groups of proteins which can also assist in the disassembly of protein complexes and aggregates, probably using the same mechanism that they use to assist folding. In addition to the chaperones whose main role is to assist protein folding, other chaperones exist whose role is more concerned with assembly of macromolecular complexes, such as chromatin, ribosomes, proteasomes, and bacterial flagella and fimbriae. These chaperones are more restricted in the range of substrates with which they interact, and as they have been recently reviewed (Ellis 2006) and have not to date been considered to have any extracellular role, we will not discuss them further here.

Chaperones involved in protein folding Some chaperones assist the folding of nascent chains bound to ribosomes, newly synthesized chains released from ribosomes, and some intrinsic membrane

ClpA ClpB ClpX

HtpG

DnaK Hsc66

GroEL (also referred to as Cpn60)

DnaJ CbpA DjlA

HSP90

HSP70

Chaperonin/ HSP60

HSP40

Bacteria*

HSP60 (m) Rubisco sub-unit binding protein (c) Ydj1 (m) Cdj1 (c)

SSC1 (m)

TRAP1 (m) HSP90 (c)

HSP78 (m) ClpC (c) ClpB (c)

Mitochondria (m) / chloroplasts (c)

HSP40 (absent from many archaea)

Thermosome, TF55, CCT

HSP70 (absent from many archaea)

Absent

Absent

Archaea

HSP40 Sis1 Zuo1

TRiC, CCT

HSP70, HSC70, SSB, SSA

HSP104; absent from cytosol of animals. HSP90 HSP82 HSP83

Eukaryotic cytosol

Examples of chaperones and co-chaperones that assist protein folding

HSP100

TABLE 1

Sec63

Absent

BiP

GRP94

Absent

Endoplasmic reticulum

Binding of unfolded proteins, stimulation of ATPase activity of HSP70 (HSP70 co-chaperone). Very heterologous group as many proteins possess ‘DnaJ-domain’

In eukaryotes: regulation of assembly of steroid receptors and signal transduction proteins; protein folding and degradation. Prevention of aggregation of unfolded/newly translated/newly imported proteins; many other functions such as clathrin uncoating, regulation of heat shock response Folding of newly translated and newly imported proteins

Disassembly of oligomers and aggregates

28 LUND & ELLIS

HSP10

GroES

Absent

Tig

Absent

Absent

Prefoldin

Trigger factor

Nascent polypeptideassociated complex

ER chaperones

Absent

MTH177

Absent

Prefoldin GimC

Absent

HSP16.5

Absent

NAC

Absent

Prefoldin GimC

Absent

HSP26 Alpha-crystallin

Calnexin Calreticulin ERp53 HSP47

Absent

Absent

Absent

Absent

Absent

Stabilization of unfolded proteins, prevention of protein aggregation Essential for function of type I chaperonins Prevents protein aggregation; functions with type II chaperonins but not essential. Assists folding of nascent ribosome-bound polypeptide Interacts with and probably assists folding of nascent ribosome-bound polypeptide Folding of ER glycosylated (and non-glycosylated?) proteins. HSP47 is specific for collagen chains.

* All examples given are from E. coli, but these proteins are highly conserved in nearly all bacterial genera. This table is highly selective in showing (a) only selected examples of chaperone families and (b) only a few examples of each family in each domain. Names of chaperones are often complicated by the fact that proteins which are clearly evolutionarily related and functionally interchangeable have frequently been given different names depending on the organism or the cell compartment in which they are studied. Many chaperones have multiple copies, which may have diverse functions, even in a single compartment. The description of a chaperone as being absent from a particular domain or location should always be regarded as provisional, due to the difficulty in some cases of determining whether sequence homologies detected by BLAST or FASTA searches are functionally significant.

Absent

Absent

absent (m); present (c)

Absent

HSP22 (m) HSP21(c)

IbpA IbpB

sHSP (small heat shock proteins) Cochaperonin

CHAPERONES 29

30

LUND & ELLIS

FIG. 1. Models for the chaperone-assisted folding of newly synthesized polypeptides in the cytosol. (A) Bacteria. TF, trigger factor; N, native protein. Most nascent chains probably interact with TF, and most small proteins (about 65–80% of total chain types) may fold rapidly upon synthesis without further chaperone assistance. Longer chains (10–20% of total chain types) interact subsequently with DnaK and DnaJ and fold after one or several cycles of ATPdependent binding and release. About 10–15% of total chains fold within the chaperonin GroEL/GroES system. GroEL does not bind to nascent chains and thus is likely to receive its substrates after their release from DnaK. (B) Archaea. PFD, prefoldin; NAC, nascent-chain associated complex. Only some archaeal species contain DnaK/DnaJ. (C) Eukarya. Like TF, NAC probably interacts with many nascent chains. The majority of smaller chains may fold without further chaperone assistance. About 15–20% of chains reach their native states after assistance by HSP70 and HSP40, and a specific fraction of these are then transferred to HSP90. About 10% of chains are passed to the TRiC system in a reaction involving PFD. From Hartl & Hayer-Hartl (2002) reprinted with permission from AAAS.

proteins, as well as the refolding of proteins denatured by stress (Young et al 2004, Hartl & Hayer-Hartl 2002, Kota & Ljungdahl 2005). Figure 1 illustrates some of the key chaperones that assist protein folding in the cytosol. Cytoplasmic chaperones working in either the co- or post-translational modes are distinct, and can be usefully termed small and large chaperones respectively, because size is important for function. Small chaperones include trigger factor, nascent chain-associated complex, and the HSP70 and HSP40 families and their associated co-chaperones. Some of these have homologues in other compartments in the cell. In general,

CHAPERONES

31

these types of chaperone are thought to function by reducing the time that potentially interactive surfaces on neighbouring chains are exposed by cycling on and off these chains until they have folded. Some, but not all, require repeated rounds of ATP binding and hydrolysis for their action. Large chaperones are oligomeric proteins with molecular masses of several hundred kilodaltons. The best studied examples are the evolutionarily related group known as the chaperonins. These include the thermosome in Archaea, GroE proteins in Bacteria, mitochondria and chloroplasts, and the CCT or TRiC complexes and associated co-chaperones in the cytosol of Eukarya. Other examples of chaperones that exist in large complexes are the HSP100 class of proteins, and the small heat shock proteins, the individual sub-units of which are typically 15–30 kDa but which are generally found in oligomers with up to 24 subunits. Small chaperones often act by binding transiently to short hydrophobic regions (typically seven or eight residues long) (Young et al 2004, Hartl & Hayer-Hartl 2002). For example, trigger factor (48 kDa), the first chaperone to bind to nascent chains in prokaryotes (Kaiser et al 2006), contains a hydrophobic groove that binds transiently to regions of the nascent chain enriched in aromatic residues. The eukaryotic cytosol lacks trigger factor but its function may be replaced by the nascent chain-associated complex (NAC), a heterodimeric complex of 33 kDa and 22 kDa subunits. Another small chaperone family, the HSP70 family, has many members distributed in the cytoplasm of bacteria and some, but not all, archaea, the cytosol of eukaryotes, and eukaryotic organelles such as the endoplasmic reticulum, mitochondria and chloroplasts. Some of these members are also stress response proteins. Unlike trigger factor, most of the HSP70 members do not bind to ribosomes, but do bind to short regions of hydrophobic residues exposed on nascent and newly synthesized chains. HSP70 proteins also have other roles such as disassembly of clathrin cages (Ungewickell 1985) and other protein complexes. The large oligomeric chaperones include the chaperonins, which are the most intensively studied chaperones, particularly the GroEL and GroES proteins of E. coli, but increasingly also the Archaeal thermosome and the CCT or TRiC complex of eukaryotic cytosol. They are the only example of a molecular chaperone that is essential in all three domains of life (Fayet et al 1989, Ursic & Culbertson 1991, Kapatai et al 2006). GroEL (800 kDa) consists of two heptameric rings of identical 57 kDa ATPase subunits stacked back to back, containing a cage in each ring (Braig et al 1994). Each subunit contains three domains. The equatorial domain contains the nucleotide binding site and is connected by a flexible intermediate domain with the apical domain. The latter presents several hydrophobic side chains at the top of the ring orientated towards the cavity of the cage, an arrangement that permits either a partly folded polypeptide chain or a molecule of GroES to bind, but prevents binding to another GroEL oligomer.

32

LUND & ELLIS

GroES is a single heptameric ring of 10 kDa subunits that cycles on and off either end of the GroEL in a manner regulated by the ATPase activity of GroEL. At any one time GroES is bound to only one end of GroEL, leaving the other end free to bind a partly folded polypeptide chain after its release from the ribosome or from another chaperone. Only one GroEL ring at a time binds nucleotide, but within each ring the binding of nucleotide is co-operative. When ADP or ATP is bound to one GroEL ring, the GroES sits on top of this ring—now called the cis ring. This binding triggers a rotation and upward movement of the apical domains, resulting in an enlarged cage and a change in its internal surface properties from hydrophobic to hydrophilic. This enlarged cage provides a secluded space that can accommodate a single partly folded compact polypeptide chain up to about 60 kDa in size, and allows that chain to continue its folding free from the danger of aggregating with similar chains (Farr et al 2001, Horwich et al 2006). Figure 2 illustrates the basic mechanism. As with many chaperones, conservation between homologues is high, and function is well preserved, to the extent that HSP60 and HSP10 from human mitochondria can functionally replace GroEL and GroES in E. coli (Nielsen et al 1999). A reasonable question to ask is, what is the evolutionary significance of cells possessing so many different types of chaperone? Even in simple bacterial cells, members of all five of the classes discussed above are present, though only the HSP60 proteins are essential for growth, and in eukaryotes, many more chaperone types are present. One possibility is redundancy of function, a sensible evolutionary strategy for such an important cellular function. However, although there is evidence for this in some cases (the roles of HSP70 and trigger factor overlap, for example), in other cases it is not seen—no other chaperone can replace the HSP60 proteins. The other likely explanation is that different chaperones may act on different substrates and/or at different stages in the process of protein folding. This idea is well supported in the cases where it has been examined; for example, the substrates of HSP70 and HSP60 in bacteria have been defined and show little overlap (Mogk et al 1999, Houry et al 1999, Deuerling et al 1999), and it has been demonstrated that chaperones do not act in isolation but in networks where proteins can be passed from one to the other (Buchberger et al 1996, Goloubinoff et al 1999). However, more work is required to fully address this important question. One important factor links all examples, however: they all perform their chaperone function somewhere inside the cell. Myths As with any field, misconceptions abound. The literature contains statements of the form ‘chaperones fold proteins’, suggesting incorrectly that chaperones possess steric information essential for protein folding. Another common error is to use

CHAPERONES

33

FIG. 2. Mechanism of action of the GroEL/GroES system in E. coli. P, unfolded polypeptide chain. N, native folded chain. The reaction cycle starts with a GroEL-GroES complex containing ADP bound to the cis ring (step 1). The hydrophobic residues on the apical domains on the other ring, now called trans, bind to hydrophobic residues exposed on a partly folded polypeptide chain, presumably after release of small chaperones from this chain. GroES and ATP then bind to this ring, thereby converting it into a new cis ring, and causing the release of GroES and ADP from the old cis ring (step 2). This binding of GroES to the trans ring displaces the bound polypeptide into the cavity of the cage because some of the hydrophobic residues of the apical domains that bind the polypeptide are the same residues that bind GroES. The displaced chain lying free in the cavity of the cage now has 10–15 sec to continue folding, a time set by the slow but co-operative ATPase activity of the seven subunits in each ring (step 3). The chain thus continues its folding sheltered in a hydrophilic environment containing no other folding chain. The binding of ATP and GroES to the new trans ring then triggers the release of GroES and ADP from the cis ring containing the polypeptide chain, allowing the latter to diffuse out of the cage into the cytoplasm. If this chain has internalised its hydrophobic residues, it remains free in the cytoplasm (step 4). But any chain that still exposes hydrophobic residues rebinds back to the same ring for another round of encapsulation (step 5). Rebinding to the same ring rather than the ring of another GroEL oligomer is favoured by the crowding effect created by the high concentration of macromolecules in the cytoplasm and reduces the risk that partly folded chains will meet one another in the cytoplasm in a potentially disastrous encounter. Reprinted from Ellis (2003), with permission from Elsevier.

34

LUND & ELLIS

the terms ‘molecular chaperones’ and ‘stress response proteins’ interchangeably. The term ‘chaperonin’ is sometimes wrongly used as if synonomous with the term ‘chaperone’, but the chaperonins are just one particular family of chaperone defined by sequence similarity. The occasional use of the nonsense term ‘molecular chaperonin’ in some journals suggests that some people use these terms casually without reference to either their meaning or their history. Another common myth is that molecular chaperones are necessarily promiscuous, i.e. that each assists the assembly of many different types of polypeptide chain. Similarly, it is not a universal property of chaperones that they hydrolyse ATP. The term ‘chemical chaperone’ has been proposed to describe small molecules such as glycerol, dimethylsulfoxide and trimethylamine N-oxide that act as protein stabilizing agents. This terminology is unfortunate and confuses students, because proteins are also chemicals. This term should be replaced by the term ‘kosmotrope’ that physical chemists use to describe small molecules that stabilize proteins—it is the opposite of ‘chaotrope’, used to describe small molecules that unfold proteins.

Conclusion In conclusion, we can state that the chaperone function is an essential and ubiquitous function in all organisms, and is sufficiently complex and important that many different types of chaperone protein have evolved to fulfi l it. Chaperones are ancient proteins, whose induction is often stress-related, and although they exert their effects as chaperones inside the cell, it would not be surprising if their presence could not also mediate important biological events outside the cell, though such a role need not necessarily require their chaperone function.

References Anfi nsen CB 1973 Principles that govern the folding of protein chains. Science 181:223–230 Barraclough BR, Ellis RJ 1980 Protein synthesis in chloroplasts. IX. Assembly of newly synthesized large subunits of ribulose bisphosphate carboxylase in isolated intact chloroplasts. Biochim Biophys Acta 608:19–31 Braig K, Otwinowski Z, Hegde R et al 1994 The crystal structure of the bacterial chaperonin GroEL at 2.8 Å. Nature 371:578–586 Buchberger A, Schroder H, Hesterkamp T, Schonfeld HJ, Bukau B 1996 Substrate shuttling between the DnaK and GroEL systems indicates a chaperone network promoting protein folding. J Mol Biol. 261:328–333 Deuerling E, Schulze-Specking A, Tomoyasu A, Mogk A, Bukau B 1999 Trigger factor and DnaK cooperate in folding of newly synthesized proteins. Nature 400:693–696 Ellis RJ 1987 Proteins as molecular chaperones. Nature 328:378–379

CHAPERONES

35

Ellis RJ 1997 Do molecular chaperones have to be proteins? Biophys Res Commun 238: 687–692 Ellis RJ 2001 Macromolecular crowding; obvious but underappreciated. Trends Biochem Sci 26:597–603 Ellis RJ 2003 Protein folding: the importance of the Anfi nsen cage. Curr Biol 13:R881– 883 Ellis RJ 2006 Molecular chaperones: assisting assembly in addition to folding. Trends Biochem Sci 31:395–401 Ellis RJ, Hemmingsen SM 1989 Molecular chaperones: proteins essential for the biogenesis of some macromolecular structures. Trends Biochem Sci 14:339–342 Ellis RJ, Minton AP 2006 Protein aggregation in crowded environments. Biol Chem 387: 485–497 Farr GW, Fenton WA, Rospert S, Horwich AR 2001 Folding with and without encapsulation by cis and trans-only GroEL-GroES complexes. EMBO J 22:3220–3230 Fayet O, Ziegelhoffer T, Georgopoulos C 1989 The groES and groEL heat shock gene products of Escherichia coli are essential for bacterial growth at all temperatures. J Bacteriol 171:1379–1385 Goloubinoff P, Mogk A, Zvi AP, Tomoyasu T, Bukau B 1999 Sequential mechanism of solubilization and refolding of stable protein aggregates by a bichaperone network. Proc Natl Acad Sci USA. 96:13732–13737 Hartl FU, Hayer-Hartl M 2002 Molecular chaperones in the cytosol: from nascent chain to folded protein. Science 295:1852–1858 Hemmingsen SM, Woolford C, van der Vies SM et al 1988 Homologous plant and bacterial proteins chaperone oligomeric protein assembly. Nature 333:330–334 Horwich AL, Farr GW, Fenton WA 2006 GroEL-GroES-mediated protein folding. Chem Rev 106:1917–1930 Houry WA, Frishman D, Eckerskorn C, Lottspeich F, Hartl FU 1999 Identification of in vivo substrates of the chaperonin GroEL. Nature 402:147–154 Kaiser CM, Chang H-C, Agashe VR et al 2006 Real-time observation of trigger factor function on translating ribosomes. Nature 444:455–460 Kapatai G, Large A, Benesch JL et al 2006 All three chaperonin genes in the archaeon Haloferax volcanii are individually dispensable. Mol Microbiol 61:1583–1597 Kota J, Ljungdahl PO 2005 Specialized membrane-localized chaperones prevent aggregation of polytopic proteins in the ER. J Cell Biol 168:79–88 Kusukawa N, Yura T 1988 Heat shock protein GroE of Escherichia coli: key protective roles against thermal stress. Genes Dev 2:874–882 Laskey RA, Honda BM, Finch JT 1978 Nucleosomes are assembled by an acidic protein that binds histones and transfer them to DNA. Nature 275:416–420 Lund PA 2001 Molecular chaperones in the cell. Oxford University Press Mogk A, Tomoyasu T, Goloubinoff P et al 1999 Identification of thermolabile Escherichia coli proteins: prevention and reversion of aggregation by DnaK and ClpB. EMBO J 18:6934– 6949 Nielsen KL, McLennan N, Masters M, Cowan NJ 1999 A single-ring mitochondrial chaperonin (HSP60-HSP10) can substitute for GroEL-GroES in vivo. J Bacteriol 181:5871–5875 Normington K, Kohno K, Kozutsumi Y, Gething MJ, Sambrook J 1989 S. cerevisiae encodes an essential protein homologous in sequence and function to mammalian BiP. Cell 57:1223–1236 Ostermann J, Horwich AL, Neupert W, Hartl F-U 1989 Protein folding in mitochondria requires complex formation with HSP60 and ATP hydrolysis. Nature 341:125–130 Smith DF 2001 Chaperones in signal transduction. In: Lund PA (ed) Molecular chaperones in the cell. Oxford University Press, p 164–179

36

LUND & ELLIS

Song J, Morimoto RI 2001 HSP70 chaperone networks. In: Lund PA (ed) Molecular chaperones in the cell. Oxford University Press, p 142–164 Ungewickell E 1985 The 70-kd mammalian heat shock proteins are structurally and functionally related to the uncoating protein that releases clathrin triskelia from coated vesicles. EMBO J 4:3385–3391 Ursic D, Culbertson MR 1991 The yeast homolog to mouse Tcp-1 affects microtubule-mediated processes. Mol Cell Biol 11:2629–2640 Young JC, Agashe VR, Siegers K, Hartl FU 2004 Pathways of chaperone-mediated protein folding in the cytosol. Nat Rev Cell Biol 5:781–790

DISCUSSION Feder: You distinguished between stress and chaperoning, suggesting that chaperoning evolved in normally functioning cells and the interaction with stress came later. When life was first evolving 3 billion years ago, cells probably functioned under situations that today we would regard as stress. Do you think that at the beginning these were actually distinct situations? Ellis: No, what I mean is that cells may have originated in a high temperature environment in which a temperature gradient, redox gradient and proton gradient coexist, such as in hydrothermal vents. These gradients are the minimum requirements to form a protocell on this hypothesis (Martin & Russell 2003). On this view, all the first proteins would have to be thermostable, and the need for chaperones would be high because aggregation increases with temperature. Mesophiles evolved from these extremophiles so that some chaperones could acquire additional functions, such as acting as stress response proteins. The word ‘stress’ only has meaning in relation to something that is non-stressed. If stress is normal, we are all stressed all the time, so the word means nothing (i.e. the concept of stress implies that there must be a normality against which to compare it). I am suggesting that cells originated under non-stress conditions. My basic point is that under non-stress conditions chaperones are essential because of the problem of protein aggregation created by the intrinsic nature of protein synthesis, i.e. polysomes bring close together identical, partly folded polypeptide chains in a crowded environment. Once chaperones evolved to solve that basic problem they could be co-opted later for other uses, such as in the stress response. You have to remember that most in vitro biochemistry is done at concentrations of macromolecule much lower than those that pertain inside the cell. Integrated metabolism cannot work at 10 mg/ml of protein—you need 200–300 mg/ml protein. This fact is summarized by the statement that the cellular interior is highly crowded with macromolecules (Ellis 2001). The resultant dramatic stimulation of protein aggregation by this crowding generates a need for the chaperone function under normal conditions; this problem is accentuated even further under stress conditions because these conditions cause mature

CHAPERONES

37

proteins to unfold. This is why so many chaperones (but not all) are also stress response proteins. Calderwood: One of your points is that folding is spontaneous. Are there not conditions in the cell in which aggregates form and then are refolded? For example, Hsp104 is said to be able to do this. Ellis: Purified denatured proteins can refold spontaneously in the test tube, but in the cell many proteins require chaperones to avoid aggregation during folding, i.e. protein folding in vivo is not spontaneous but chaperone-assisted. Folding is defined as the collapse and compaction of either a nascent chain or a stress-denatured chain to a point where a stable fold is generated. If the resulting monomer then oligomerizes, it might form either a functional oligomer or an aggregate, an aggregate being defined as a non-functional oligomer. In some cases the assembly of monomers into functional oligomers also requires chaperones, i.e. there are both folding chaperones and assembly chaperones (Ellis 2006). In fact, the very first protein to be called a molecular chaperone assists an assembly process (Laskey et al 1978). In the case of Hsp104, this chaperone can function together with Hsp70 and Hsp40 and ATP to unfold insoluble aggregates. Other chaperones also use the energy of ATP binding and/or hydrolysis to do conformational work. An obvious example would be the uncoating of clathrin cages, which requires Hsp70, Hsp40 and ATP hydrolysis. It is thought that the underlying mechanism involves the recognition of a sequence of exposed hydrophobic residues on the surface of the folded clathrin cage. Calderwood: Could that be described as spontaneous? Ellis: No, because uncoating requires ATP. Aggregates are regarded as highly stable, especially if they are large enough to be insoluble. If you make a solution of any protein and leave it long enough, eventually it will all aggregate. This is because proteins are only marginally stable at physiological temperatures, i.e. there will always be a small proportion of unfolded molecules in equilibrium with folded ones. So in any protein solution, there will be a few unfolded chains among many thousands of folded ones—the exact ratio depends on the free energy of folding. These unfolded molecules then aggregate, because aggregation requires only that identical or very similar chains are unfolded to some extent. The rate of aggregation depends on chain concentration and the temperature. This aggregation displaces the equilibrium between folded and unfolded chains so that more chains unfold. Thus in time, all the protein will end up as an aggregate. This is true for all proteins where the free energy of folding allows some unfolded chains to exist at equilibrium. So the problem is not that aggregation is a rare event—it is the inevitable default state for all proteins in solution. The mystery is how biology has come up with so many globular proteins that manage to avoid this fate. One answer to this mystery is that molecular chaperones combat aggregation. All this has nothing to do with stress.

38

LUND & ELLIS

Macario: Can you integrate the chaperones with protein degradation systems? Ellis: If you have an unfolded protein or a partly folded protein, the cell can take a decision about whether or not to get rid of it. The cell might decide to let it fold by avoiding aggregation, or might decide to eliminate it. This decision is regulated by the co-chaperone machinery. The biggest family of co-chaperones is the DnaJ family, which contains both chaperones and co-chaperones. I define cochaperones as proteins that regulate the activity of chaperones, but do not themselves bind to a polypeptide. However there are some co-chaperones, such as DnaJ from E. coli, that are also chaperones (i.e. they bind to a nascent polypeptide as well as to a chaperone). Some co-chaperones are expressed only in particular tissues e.g. neurons, so there is a whole layer of specific regulation of chaperone function. Part of that regulation is to decide whether to let the protein refold or to direct it to the proteasome. CHIP and BAG are two co-chaperones involved in making this decision. Gupta: You mentioned that the protein aggregation increases with the concentration of the protein. Is there any evidence that in more crowded conditions such as in the red blood cells, more chaperones are present? Ellis: No. If we look at bacterial cells, the figure for total macromolecules is around 300–340 g/l, around the same value for the concentration of haemoglobin inside red blood cells. In eukaryotic cells the numbers vary from 120 to 280 g/l: they tend to be less crowded than prokaryotic cells. All these values lie within the effective range where crowding stimulates aggregation by between one to three orders of magnitude. It is not generally appreciated that the relationship between aggregation and concentration of the crowding molecules is highly non-linear (Ellis 2001). The dramatic effects of crowding have been known for 30 years, but some protein chemists are in denial about crowding effects, because they mess up their simple experiments. They prefer to study a pure protein in a simple buffer, with expensive machines to analyse it. But crowding is a universal feature of the environment where proteins have evolved and function, so in my view all studies on isolated macromolecules should include experiments on the effects of crowding agents—otherwise the biological relevance of such experiments is questionable. The sad fact is that studies on the properties of isolated macromolecules can ignore crowding, but yet be published in respectable journals, with biological inferences drawn from the measurements reported. Gupta: I have a follow-up question. There are so many proteins in the cell that require folding, but not that many chaperones. Do we have any understanding of which proteins require chaperones to fold, and which do not? Ellis: There is a field I call chaperomics. A chaperome is the total spectrum of polypeptides that a given chaperone binds to in the cell. This spectrum has been determined for GroEL in E. coli (Kerner at al 2005) and for Hsp90 in yeast by

CHAPERONES

39

Zhao et al (2005). The E. coli experiments determined how many different proteins bind to GroEL as they are synthesized in vivo, and the answer is around 250. If you then ask how many of them have to bind to GroEL in order to fold, the answer comes out as 85. Of these 85, 13 are essential proteins for the cell, explaining why GroEL is an essential chaperone. If E. coli makes 2500 proteins or so in a normal growth medium, GroEL binds to around 10% of them. Hsp90 in yeast binds to 198 proteins, as determined by two-hybrid and immunoprecipitation screens. Gupta: Can we define the common denominator? Ellis: The Hartl laboratory has attempted to do this. The majority of the proteins are αβ domain proteins that are particularly prone to aggregation, as shown by using algorithms that predict aggregation propensity from sequence. They are not all in this αβ class, however. So it seems that GroEL selects out those proteins that are particularly prone to aggregation but how this is achieved is not clear. These data fit with the view that says that the basic problem that GroEL deals with is protein aggregation. Lund: There are some elegant genetics on this in E. coli. It is possible to turn off all the heat shock proteins by deleting a subunit of the RNA polymerase that regulates their expression. If this is done, the cells won’t grow at above 20 °C. This tells us that these proteins are needed even for normal growth. Under those conditions, some of these proteins are being expressed at a low level. We know that GroEL, for example, is essential for growth at all temperatures. If we now use the power of E. coli genetics to ask for mutants that can grow at higher temperatures, even though their heatshock response is turned off, you can easily get mutations that would grow up to 40 °C. All of these turn out to be up-promoter mutations in the GroEL/ES operon. In other words, the only heat shock protein needed to get growth up to 40 °C is the Hsp60/Hsp10 GroEL/GroES system. The others are not required for growth up to about 40 °C. If you want the cells to grow at higher temperatures, you have to do a second round of mutatgenesis, which yields up-promoter mutations in the DnaK and DnaJ systems. At least in the context of the E. coli cytoplasm, most of the important essential protein folding that takes place under normal growth conditions is done by GroEL/GroES, but if you want to grow things at a higher temperature, DnaK and DnaJK systems are also needed. This is probably because they work in concert with the ClpB chaperone to disagreggate proteins that become unfolded and aggregated at those temperatures. This has only been demonstrated with E. coli, and it may be different in mitochondria or eukaryotic cytosol. Nixon: The chaperones themselves need to fold. Is there a specific mechanism that enables them to withstand the types of stresses under which they become more highly expressed, to retain their function where other proteins are starting to denature?

40

LUND & ELLIS

Ellis: Heat shock proteins will obviously function at the heat shock temperature, but there is nothing peculiar about them in terms of folding. There was an experiment done by Cheng et al (1990) in which they asked whether pre-existing GroEL is required inside the mitochondrion in order for newly synthesized GroEL imported into the mitochondrion to fold correctly. The answer to this question is yes. If the existing GroEL inside the mitochondria is depleted, then that cell dies, even though it still has the gene for GroEL in the nucleus—i.e. pre-existing GroEL is required for new GroEL to avoid aggregation and fold. The pre-existing GroEL is obtained from the parent cell. So there is no problem once the GroEL has appeared—the problem is how it started in the first place, but this is true of most aspects of biochemistry! Nixon: You mentioned there was specificity associated with the aggregation process. Are the most abundant proteins in the cell those that are more prone to aggregation? Ellis: Abundance is related to the function of the protein. Rubisco is the most abundant protein on the planet, which means that the Rubisco chaperonin is the most abundant chaperone on the planet. The reason that this enzyme is so abundant is because it is such an inefficient enzyme, with a very low turnover number and it is not saturated at current levels of atmospheric carbon dioxide. When it first evolved, this poor activity wasn’t a problem because atmospheric CO2 was at a much higher concentration than it is now. The other problem with Rubisco is that it reacts with molecular oxygen, which is always present in chloroplasts as a product of the light reaction. This reaction with oxygen produces phosphoglycollate, which is metabolised by the photorespiratory pathway peculiar to plants. But photorespiration loses carbon as carbon dioxide, so the oxygenase reaction works against photosynthesis. When Rubisco first evolved, there was no oxygen in the air so this oxygenase side reaction did not matter, but now it limits plant productivity. So Rubisco is an excellent example of unintelligent design. As Sydney Brenner said, everything produced by natural selection is bound to be a bit of a botched job, because evolution has no foresight. As regards aggregation, plant Rubisco is one of the very few proteins that cannot be expressed in bacteria or refolded after denauration, despite all we know about chaperones. So it is thought that additional chaperones are required besides a chaperonin, and one called RbcX that functions at the assembly step, rather than the folding step, was identified recently in cyanobacteria (Saschenbrecker et al 2007). Gaston: Can I ask you to speculate about the topic of this meeting, the extracellular function of these proteins? Under these circumstances crowding is much less of an issue. Does this mean that whatever extracellular functions are attributed to chaperones they are unlikely to be due to their folding-type properties?

CHAPERONES

41

Ellis: I don’t think this. If we consider whether extracellular environments are crowded, it depends where you look. Human blood contains about 80 g/l macromolecules, which is sufficient to increase association constants by one order of magnitude. If you look at the surface of neurons, neurobiologists will tell you that this is a highly crowded environment because the plasma membranes of adjacent neurons are very close together. The other crowded extracellular environment is provided by collagen. On the other hand, cerebrospinal fluid is not crowded, nor are the large central vacuoles in plant cells. So you need to consider this question case-by-case. Gupta: Regarding Dr Gaston’s question, I can comment on mitochondrial chaperones. When they are present outside mitochondria, they are not always involved in the protein-folding type activities. Gaston: I’m very happy with the idea of proteins doing different things in different places in different contexts, but I wanted to address the idea of whether something to do with folding was also involved. Ellis: There are extracellular amyloid fibres involved in certain neurodegenerative diseases. I don’t think we can generalize on this. In an animal there are so many compartments. We need to be specific. Cheetham: You were careful to say that aggregates were non-functional. I think this is an important distinction. For example, when there are large polymers of functional components, such as actin or tubulin in the cytoskeleton, then they are functional. Does this require more regulation, even though it is controlled ‘aggregation’? Ellis: It’s partly a question of definitions. Historically, aggregation was regarded as a bad thing that complicated protein refolding experiments. The word ‘aggregate’ is embedded in the literature and is usually defined in purely chemical terms as an insoluble protein, but I use it to refer to any oligomer that does not carry out the function that was selected for in evolution. Some protein chemists do not like this functional definition, but prefer a chemical one. But in my view, structure is not the important criterion in biology in itself, but function is, because it is function that is selected for in evolution, not structure per se. I have proposed that we forget the word ‘aggregation’, and replace it with another term, ‘misassembly’, which was used in the 1950s. I like this term because it suggests that something has gone wrong, i.e. it is a functional definition. It also solves another problem, which is that some amyloid fibrils have been shown recently to have functions. The word ‘aggregate’ will probably persist, though, and most people will mean by it an oligomer that does not carry out its normal biological function. Coates: I want to return to the topic of extracellular aggregates. You would expect that if aggregation extracellularly was important, that there should be a constitutive level of a molecular chaperone in the blood. So why is it that if we measure Hsp60 in the blood of healthy people there are such different levels?

42

LUND & ELLIS

Ellis: I’m not sure what the evidence is for thinking that aggregation is a problem in the blood in a non-diseased condition. What would be the problem? Isn’t it the result of disease, where a given protein accumulates in a large amount? Coates: I would suggest that aggregation is not particularly important outside the cell. Ellis: There is an international commission of nomenclature run by the Amyloidosis Society. They define amyloid fibrils as extracellular, but this criterion is widely disregarded by protein chemists who define amyloid in structural terms. Macario: What do you think about RNA chaperones? Ellis: If you mean by this proteins that deal with RNA misfolding, nucleic acids have much less information content than proteins, therefore the chances of misfolding are much higher for RNA than they are for proteins. There is quite a body of work on the folding of tRNAs showing that an appreciable portion of them will misfold. There are proteins that stop this happening—they are called RNA chaperones. Another question is, are there any RNA molecules that act as chaperones? There is some work suggesting this with respect to ribosomal RNA. If you carry out in vitro protein aggregation assays, you can show that ribosomal RNA will prevent aggregation but I am unaware of any data suggesting that ribosomal RNA in the cell acts as a chaperone. Csermely: What do you think would be a legitimate assay for molecular chaperones? Ellis: In practice, people tend to use in vitro aggregation assays. If this assay is positive, you have a potential chaperone, but at the end of the day what matters is what goes on in the cell, not what goes on in a test-tube and this requires genetic techniques to evaluate. The early GroEL field contains some observations that are probably artefacts of the in vitro conditions used. Triantafilou: In the plasma membrane, do you think that chaperones could bind to receptor assemblies and act to disassemble the oligomerization of the receptor? Ellis: I see no reason to rule this out. The plasma membrane is a highly crowded environment. Once you get into lipid rafts you are in a high concentration environment, so the same sorts of arguments apply, and the disassembly arguments would apply just as well. Multhoff: Heat shock proteins are produced very quickly, and huge amounts are present. Overall, after stress protein synthesis is down-regulated. Do you think that a physiological response to this large amount of crowded heat shock proteins is to bring them into the extracellular space? They are present in huge amounts and are produced very quickly. Perhaps there is the danger of their aggregation inside the cells and thus the cell decides to get rid of them. Huge amounts of cytosolic heat shock proteins inside the cell might not be beneficial. Exosomes with the transferrin B receptor were found to be secreted by reticu-

CHAPERONES

43

locytes: people speculate that exosomes act as a garbage bag for the transferrin receptor. Ellis: Heat shock is a response to a particular circumstance, so once this is over you want to get the level down again. If you up-regulate a whole raft of chaperones there could be many pleiotropic effects, so I can see you might want to get rid of them. You could just degrade them or no longer make them. Lund: It might be quicker to secrete them. Saturating the proteasome might be expensive for the cell. Feder: Sue Lindquist’s group (Feder et al 1992) overexpressed Hsp70 off an exogenous promoter in Drosophila cells. As this went on the cells were sick and wouldn’t divide. Left for a while, Hsp70 was granularized and the cells resumed division and growth when this occurred. Pockley: There has been one paper published (Wang et al 2004) in which Hsp70 was overexpressed in prostate cancer cells. This is associated with increased release of Hsp70 from those cells. Multhoff: You proposed that heat shock proteins could be degraded. But isn’t the function of stress proteins for them to be stable? Reports showed that heat shock proteins aren’t easily degraded and remain quite stable following stress (Chen et al 1990, Ashcroft et al 2002). Ellis: I don’t know the literature on this. Proteins vary enormously in their turnover rates. Calderwood: The lifetime of Hsp70 after the heat shock response depends on when the ubiquitin ligase CHIP switches on (Qian et al 2006). Ellis: Are heat shock proteins in general longer-lived than other proteins? Hightower: There is evidence that levels of both Hsc70 and BiP are closely monitored and regulated by cells. In the early studies of antisense RNA people tried to knock down their levels by making lots of antisense RNA. Sure enough, in those transfection experiments lots of antisense RNA was made but BiP and Hsc70 levels didn’t change (Li et al 1992, Li & Hightower 1995). The cells had up-regulated the mRNA levels tenfold or more. Macario: One thing we shouldn’t forget in regulation of Hsp70 levels in eukaryotic cells is heat-shock factors. Maybe the problem is with the heat-shock factor gene or its product (Macario & Conway de Macario 2005).

References Ashcroft AE, Brinker A, Coyle JE et al 2002 Structural plasticity and noncovalent substrate binding in the GroEL apical domain. A study using electrospray ionization mass spectrometry and fluorescence binding studies. J Biol Chem 277:33115–33126 Chen Q, Lauzon LM, DeRocher AE, Vierling E 1990 Accumulation, stability and localization of major chloroplast heat-shock protein. J Cell Biol 110:1873–1883

44

LUND & ELLIS

Cheng MY, Hartl FU, Horwich AL 1990 The mitochondrial chaperonin hsp60 is required for its own assembly. Nature 348:455–458 Ellis RJ 2001 Macromolecular crowding: obvious but under appreciated. Trends Biochem Sci 26:597–604 Ellis RJ 2006 Molecular chaperones: assisting assembly as well as folding. Trends Biochem Sci 31:395–401 Feder JH, Rossi JM, Solomon J, Solomon N, Lindquist S 1992 The consequences of expressing Hsp70 in Drosophila cells at normal temperatures. Genes Dev 6:1402–1413 Kerner MJ, Naylor DJ, Ishihama Y et al 2005 Proteome-wide analysis of chaperonin-dependent protein folding in Escherichia coli. Cell 122:209–220 Laskey RA, Honda BM, Mills AD, Finch JT 1978 Nucleosomes are assembled by an acidic protein which binds histones and transfers them to DNA. Nature 275:416–420 Li T, Hightower LE 1995 Effects of dexamethasone, heat shock and serum responses on the inhibition of hsc70 synthesis by antisense RNA in NIH 3T3 cells. J Cell Physiol 164:344–355 Li L-J, Li X, Ferrario A et al 1992 Establishment of a Chinese hamster ovary cell line that expresses grp78 antisense transcripts and suppresses A23187 induction of both GRP78 and GRP94. J Cell Physiol 153:575–582 Martin W, Russell MJ 2003 On the origins of cells: a hypothesis for the evolutionary transitions from abiotic geochemistry to chemoautotrophic prokaryotes and from prokaryotes to nucleated cells. Philos Trans R Soc Lond B Biol Sci 358:59–85 Macario AJL, Conway de Macario E 2005 Sick chaperones, cellular stress and disease. New Engl J Med 353:1489–1501 Qian SB, McDonough H, Boellmann F, Cyr DM, Patterson C 2006 CHIP-mediated stress recovery by sequential ubiquitination of substrates and Hsp70. Nature 440:551–555 Saschenbrecker S, Bracher A, Rasudeva Rao K, Rasudeva Rao B, Hartl FU, Hayer-Hartl M 2007 Structure and function of RbcX, a specific assembly chaperone for hexadecameric Rubisco. Cell 129:1189–1200 Wang MH, Grossmann ME, Young CY 2004 Forced expression of heat-shock protein 70 increases the secretion of Hsp70 and provides protection against tumour growth. Br J Cancer 90:926–931 Zhao R, Davey M, Hsu Y-C et al 2005 Navigating the chaperone network: an integrative map of physical and genetic interactions mediated by the Hsp90 chaperone. Cell 120:715–727

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Systems biology of molecular chaperone networks Péter Csermely, Tamás Korcsmáros, István A. Kovács, Máté S. Szalay and Csaba Soti ´´ Department of Medical Chemistry, Semmelweis University, Puskin Street 9, H-1088 Budapest, Hungary

Abstract. Molecular chaperones are not only fascinating molecular machines that help the folding, refolding, activation or assembly of other proteins, but also have a number of functions. These functions can be understood only by considering the emergent properties of cellular networks—and that of chaperones as special network constituents. As a notable example for the network-related roles of chaperones they may act as genetic buffers stabilizing the phenotype of various cells and organisms, and may serve as potential regulators of evolvability. Why are chaperones special in the context of cellular networks? Chaperones: (1) have weak links, i.e. low affi nity, transient interactions with most of their partners; (2) connect hubs, i.e. act as ‘masterminds’ of the cell being close to several centre proteins with a lot of neighbours; and (3) are in the overlaps of network modules, which confers upon them a special regulatory role. Importantly, chaperones may uncouple or even quarantine modules of protein–protein interaction networks, signalling networks, genetic regulatory networks and membrane organelle networks during stress, which gives an additional chaperone-mediated protection for the cell at the network-level. Moreover, chaperones are essential to rebuild inter-modular contacts after stress by their low affi nity, ‘quasi-random’ sampling of the potential interaction partners in different cellular modules. This opens the way to the chaperone-regulated modular evolution of cellular networks, and helps us to design novel therapeutic and anti-ageing strategies. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 45–58

System biology and the utility of the network approach The complexity of cells can be described reasonably well if we use the network approach and catalogue the interactions between cellular molecules, processes or organelles. Here the interacting macromolecules are treated as network elements, and their interactions form the weighted and sometimes directed links of the respective network. We may also assemble a network-set of directed links as representations of signalling, genetic regulatory or metabolic processes of these functional networks in the cell (Barabasi & Oltvai 2004, Boccaletti et al 2006, Csermely 2006). 45

46

CSERMELY ET AL

Cellular networks often form small worlds, where any other element of the network can be reached from a starting element via only a few links. Networks of our cells contain hubs, i.e. elements, which have a large number of neighbours. These networks can be dissected to overlapping modules, which form hierarchical communities. Small world-ness, the importance of hubs and modules in network organization, is a characteristic of most cellular networks, which makes the network approach a highly useful conceptual framework to understand the complexity of the cell. Moreover, the hierarchical modules of cellular networks enables us to treat larger segments of these networks as single entities (elements), which greatly simplifies the multitude of the hundreds of thousands of interactions, and gives us a chance for understanding and visualization. Last but not least, the above features of network organization are much more general than the realm of cellular networks, and can be extended to the networks of cells (such as the nervous system), to social networks, to ecosystems and to cultural networks, such as power grids, the worldwide web or the internet. The general network organization principles pinpoint those features of the special cellular network, which have a key importance in our understanding of the specific cellular functions. This holistic view of network properties greatly helps the interdisciplinary approach of systems biology in the cellular domain (Boccaletti et al 2006, Csermely 2006, Newman 2003, Palla et al 2005). Special roles of chaperones in cellular networks Molecular chaperones are not only fascinating molecular machines that help the folding, refolding, activation or assembly of other proteins, but also have a number of functions. These functions can be understood only by considering the emergent properties of cellular networks—and that of chaperones as special network constituents. What are the special roles of chaperones in the context of cellular networks? • Chaperones have weak links, i.e. low affinity, transient interactions with most of their partners. This feature is actually a consequence of the assistance of chaperones in folding and re-folding of cellular proteins. Chaperones bind and release a large number of newly born or damaged proteins, which requires a large promiscuity and low affinity of their interactions (Csermely 2004, Tsigelny & Nigam 2004). • Chaperones preferentially connect hubs. In other words, chaperones act as ‘masterminds’ of the cell being neighbours of several centre proteins with a lot of secondary neighbours. This may help the chaperone-mediated cross-talk between signalling and gene regulatory pathways enabling chaperones to act as a central switchboard of the cell re-programming cellular functions during and after stress (Korcsmaros et al 2007a). • Chaperones are in the overlaps of network modules. This key position gives them a special regulatory role, since they can easily couple, uncouple or even quarantine

CHAPERONE NETWORKS

47

network communities, i.e. protein complexes, cellular organelles such as damaged mitochondria, signalling pathways, metabolic routes or genetic regulatory circuits (Korcsmaros et al 2007a, Soti et al 2005, Szabadkai et al 2006). Many chaperones are also known as ‘stress’ or ‘heat shock’ proteins, since they are synthesized when the cell experiences stress. During stress, chaperones become increasingly occupied by damaged proteins, and a so-called ‘chaperone overload’ may easily occur (Csermely 2001, Nardai et al 2002). This ‘competitive inhibition’ of molecular chaperones might lead to a de-coupling of all the chaperone-mediated contacts between network modules mentioned above. De-coupling may even be so extensive that the damaged module becomes quarantined, and practically all of its contacts are efficiently severed isolating it from the rest of the cell. Since decoupling of modules stops the propagation of network damage at the modular boundaries, chaperone-induced module de-coupling provides an additional safety measure for the cell (Soti et al 2005, Szabadkai et al 2006, Szalay et al 2007). When the stress is over, and cellular resources slowly start to get back to normal again, cellular networks start to re-establish those links, which ceased to operate during stress. Bridges, local hubs are re-built, modules are re-coupled. As a gross summary of these processes, the cell re-establishes its lost repertoire of weak links, which enable its networks to a large number of dynamic and flexible changes. In this way the re-gaining of the links shed during stress can be envisioned as a purchase of a general ‘insurance’, which enables the stressed cell to recover from its former, rigid state highly specialized to the given form of stress, and to attain a more flexible structure, which will be able to cope with a large number of unexpected changes in the ‘peaceful’ period. (NB, this ‘network-based’ conceptualization of stress defines stress as an environmental change strong and/or repeated enough to exhaust cellular resources requiring an efficient, cheap and simple defence.) Cellular remodelling steps after stress may be greatly helped by the newly synthesized molecular chaperones, since their low affi nity interactions effectively sample a large number of proteins, and allow the re-arrangement of hubs, reformation of bridges and binding of de-coupled modules to each other in a very flexible, partially stochastic manner. Thus, chaperones give the cell a refined and flexible way for the gradual build-up of the complex modular structure and function, when the stress is already over (Szalay et al 2007). One of the best examples of chaperone-mediated emergent network properties was shown by Susannah Rutherford and Susan Lindquist, when they discovered that Hsp90 acts as a buffer to conceal the phenotype of the genetic changes in Drosophila melanogaster (Rutherford & Lindquist 1998). Chaperone-induced genetic buffering is released upon stress, which causes the sudden appearance of the phenotype of previously hidden mutations, helps population survival and gives a

48

CSERMELY ET AL

possible molecular mechanism for fast evolutionary changes. On the other hand, stress-induced appearance of genetic variation at the level of the phenotype cleanses the genome of the population by allowing the exposure and gradual disappearance of disadvantageous mutations by natural selection. After the initial report of Rutherford and Lindquist (1998), the effect was extended to other chaperones and to Escherichia coli, Arabidopsis thaliana and the evolution of resistance in fungi (Cowen & Lindquist 2005, Fares et al 2002, Queitsch et al 2002). In recent years the scientific community has became increasingly aware of the idea that not only chaperones but a large number of other proteins may also regulate the diversity of the phenotype (Bergman & Siegal, 2003, Csermely 2004, 2006). If a general explanation is sought, it is more likely to be related to the network properties of the cell than to a simple bi-molecular interaction. In this context, the weak links of chaperones, their central position linking hubs to each other and their inter-modular links may all help their regulatory role in the evolvability of complex systems. The remodelling of the inter-modular contacts is an especially intriguing idea for the explanation of chaperone-mediated sudden changes in the emergent properties of cellular networks (such as the phenotype of the hosting organism). Different assembly of slightly changed cellular modules may cause profound and abrupt changes of the functional repertoire without major changes of the underlying structure of protein–protein interactions. This gives an exploratory but stable mechanism for the evolution of cellular networks (Korcsmaros et al 2007a). Chaperone networks Chaperones never act alone. They form highly dynamic complexes with each other (forming sometimes truly extensive homo- or hetero-oligomers), with their cochaperones (regulating and modifying their function) and with the plethora of their client proteins (Soti et al 2005, Zhao et al 2005). This large set of primary, secondary and more distant chaperone-neighbours can be easily perceived as a network. Indeed, in yeast, two interrelated, but separated chaperone networks have been reported (Fig. 1). One of them, the CLIPS chaperones (chaperones linked to protein synthesis exemplified by the SSB Hsp70 proteins and by the TriC/CCT complex) operates to help the folding of de novo synthesized proteins, while the other, the HSP chaperone group (containing the SSA Hsp70 and Hsp90) assists in the re-folding of damaged proteins after stress. While the synthesis of CLIPS chaperones becomes repressed during stress, the synthesis of HSP chaperones is grossly activated (Albanese et al 2006). A closer look at these two ‘separate’ chaperone networks uncovers a large set of connecting, overlapping chaperones, such as the SSE1 (Hsp104) chaperone, which acts as a nucleotide exchanger for both key Hsp70 proteins in the different groups (Raviol et al 2006). Additionally,

CHAPERONE NETWORKS

49

CLIPS chaperones CCT GIMc

ZUO1 SSB2

SSZ1

SSE1 YDJ1

Hsc82

SSA1

Cpr7

SSA2-4 SSE2 Hsp26

HSP chaperones

Cpr6 Hch1 Cns1 Pih1 Aha1 Sti1 Cdc37 Tah1 Ppt1

Hsp82

FIG. 1. The two interrelated yeast chaperone networks. CLIPS chaperones (chaperones linked to protein synthesis) operate to help the folding of de novo synthesized proteins, while the HSP chaperone group mostly assists in the refolding of damaged proteins after stress. A large set of chaperones connects the two chaperone networks, such as the SSE1 (Hsp104) chaperone, which acts as a nucleotide exchanger for both key Hsp70 proteins in the different groups (Albanese et al 2006, Raviol et al 2006, Zhao et al 2005).

members of the yeast Hsp90-related chaperone co-factor complex also have extensive contacts with both Hsp70 complexes (Zhao et al 2005). The utility of chaperone-networks is further supported by the bacterial chaperones from the Mycoplasma genus, if compared to those of E. coli. Mycoplasmas evolve 50% faster than related organisms. This high mutation rate allows them an easy escape from the detection mechanisms of the host organism. A likely consequence of this high mutation rate is an increase in the frequency of misfolded Mycoplasma proteins. Indeed, recent estimates using comparative structural genomics resulted in generally lower protein stability of 11 protein families in Mycoplasmas compared to other bacteria. The interesting fact that most Mycoplasmas have lost either the gene or the activity of their central chaperone, GroEL, strongly suggests that protein quality control is mostly mediated by protein degradation in these bacteria. This is even more likely, since most major bacterial proteases are more intimately linked to the GroE system than to the more distant DnaK system (Fig. 2), which makes such a rearrangement plausible (Wong & Houry 2004). As has already been mentioned above, chaperones couple other cellular networks, besides those of their own. The mitochondrial matrix chaperone Hsp78 has

50

CSERMELY ET AL

GroEL ClpB GroES

ClpP

ClpA FtsH

ClpX

Lon

HsIU

RadA

Tig

HsIV CbpA GrpE

DnaK

DnaJ

YegD MreB HscC

HscA

Hsp90

FIG. 2. The bacterial chaperone network of Escherichia coli, and its possible rearrangement in Mycoplasmas. The figure shows a sketch of the chaperone network of E. coli after Wong and Houry (2004). In most Mycoplasmas the key chaperone, GroEL is either missing or inactivated. This may trigger a shift towards protein degradation in the quality control of damaged proteins (Wong & Houry 2004), which is in agreement with the structure of the chaperone network.

been shown to mediate the restoration of the mitochondrial network after stress in yeast (Lewandowska et al 2006). The involvement of chaperones in the coupling of mitochondria and the endoplasmic reticulum was also observed in higher eukaryotes (Fig. 3, Szabadkai et al 2006). A member of the small heat shock protein family, αB-crystallin was shown to regulate the dynamics of actin fi lament networks protecting the remaining network integrity after stress (Launay et al 2006). The role of chaperone networks at the emergent properties of the whole-cell level was further supported by a genetic screen searching for synthetically damaging mutations with an inactive Hsp70/SSB system in yeast. Surprisingly, the uncovered synthetically damaging mutants could not be associated with protein damage needing a direct assistance of the damaged molecular chaperones. This showed that chaperones may be indirectly stabilized by the damage of many mutant proteins, by opening alternative routes in various cellular networks (Bobula et al 2006, Csermely 2004, 2006).

CHAPERONE NETWORKS

A

endoplasmic reticulum

Crt

51

B mitochondrium

Cnx Grp78 IP3 receptor ryanodine receptor Grp75 voltagedependent anion channel (VDAC) Hsp60 complex Cyp-D

MDJ1 SSC1 Hsp78 MGE1 mitochondrium

mitochondrium

FIG. 3. Chaperone complexes help the coupling of mitochondria to the endoplasmic reticulum (A) and the re-assembly of the mitochondrial network after stress (B). The chaperone complex including calnexin (Cnx), calreticulin (Crt), Grp78, Grp75, Hsp60 and Cyclophylin D (Cyp-D) is involved in the coupling of mitochondria to the endoplasmic reticulum (Panel A, Szabadkai et al 2006). The mitochondrial chaperone complex, Hsp70(SSC1)/MDJ1 plays a key role in the maintenance of the yeast mitochondrial network, while the co-operating mitochondrial chaperone complex, Hsp78/MGE1, plays a key role in the re-assembly of the mitochondrial network after stress (Panel B, Lewandowska et al 2006).

Extracellular chaperones and networks Though recent studies uncovered a number of highly important roles for extracellular chaperones (see Asea 2003, Eustace et al 2004, Schmitt et al 2007, Yerbury et al 2005 and other chapters in this volume), our understanding of extracellular chaperone networks is still lagging behind. Components of the Hsp90 chaperone complex, such as Hop or p23, have already been identified as extracellular chaperones (Eustace & Jay 2004), and the concerted tyrosine phosphorylation of both Grp94 (endoplasmin) and Hsp60 were shown to mediate sperm-zona recognition (Asquith et al 2004). The extraordinarily dynamic nature of the plasma membrane and its special permeability at the boundaries of its microdomains (Marguet et al 2006) suggests that we will discover many more elements of intracellular chaperone networks at the extracellular space. During this build-up of the databases for extracellular chaperone networks we will have to consider special, extracellular chaperones as well, such as the SPARC (secreted protein, acidic and rich in

52

CSERMELY ET AL

cysteine), a matricellular glycoprotein modulating cell proliferation, adhesion, migration and extracellular matrix production (Emerson et al 2006). Conclusions and open questions Recent progress in network science and, especially, our emerging knowledge of network dynamics provides a unique chance to understand chaperone function at a novel level. However, the lack of adequate and verified information on the several hundred thousand interactions in cellular networks, the incomplete accuracy and sensitivity of currently available analytical methods, as well as inadequate methods for network analysis raise more open questions than satisfying answers at the moment. We list a few of these questions in the following: • What can be the in vivo distribution of chaperone function between the folding and refolding of single proteins, versus the assistance in the assembly of protein– protein, protein–RNA and protein–DNA complexes? How are these two functions related to the role of chaperones in regulating protein degradation? What mechanisms regulate the shift between these functions during and after stress, in the onset and propagation of disease and during ageing? • Do we have extracellular chaperone networks? How are these chaperone networks regulated by the local fluctuations of extracellular ATP levels? What are the chaperones assisting in the assembly of the extracellular matrix? Do we have polysaccharide chaperones? We are quite certain that chaperone networks will give a lot of excitement and pleasure for systems biologists, who would like to understand and modify the function of our cells in health, stress, disease and ageing. As a result of these studies a renaissance of network-based and chaperone-based therapies is expected, where target sets of multi-target drugs will be identified using our knowledge of the vulnerable points (hot spots) of cellular networks, including molecular chaperones and their sub-networks (Csermely et al 2005, Korcsmáros et al 2007b). Acknowledgements The authors would like to thank members of the LINK-group (www.weaklinks.sote.hu) for helpful discussions. Work in the authors’ laboratory was supported by research grants from the EU (PROTEOMAGE, FP6–518230,) and by the Hungarian National Research Initiative (NKFP1A/056/2004 and KKK-0015/3.0).

References Albanese V, Yam AY, Baughman J, Parnot C, Frydman J 2006 Systems analyses reveal two chaperone networks with distinct functions in eukaryotic cells. Cell 124:75–88

CHAPERONE NETWORKS

53

Asea A 2003 Chaperokine-induced signal transduction pathways. Exerc Immunol Rev 9:25–33 Asquith KL, Baleato RM, McLaughlin EA, Nixon B, Aitken RJ 2004 Tyrosine phosphorylation activates surface chaperones facilitating sperm-zona recognition. J Cell Sci 117:3645– 3657 Barabasi AL, Oltvai ZN 2004 Network biology: understanding the cell’s functional organization. Nat Rev Genet 5:101–113 Bergman A, Siegal ML 2003 Evolutionary capacitance as a general feature of complex gene networks. Nature 424:549–552 Bobula J, Tomala K, Jez E, Wloch DM, Borts RH, Korona R 2006 Why molecular chaperones buffer mutational damage: a case study with a yeast Hsp40/70 system. Genetics 174:937–944 Boccaletti S, Latora V, Moreno Y, Chavez M, Hwang D-U 2006 Complex networks: structure and dynamics. Physics Rep 424:175–308 Cowen LE, Lindquist S 2005 Hsp90 potentiates the rapid evolution of new traits: drug resistance in diverse fungi. Science 309:2185–2189 Csermely P 2001 Chaperone-overload as a possible contributor to “civilization diseases”: atherosclerosis, cancer, diabetes. Trends Genet 17:701–704 Csermely P 2004 Strong links are important—but weak links stabilize them. Trends Biochem Sci 29:331–334 Csermely P 2006 Weak links: a universal key for network diversity and stability. Springer Verlag, Heidelberg Csermely P, Agoston V, Pongor S 2005 The efficiency of multi-target drugs: the network approach might help drug design. Trends Pharmacol Sci 26:178–182 Emerson RO, Sage EH, Ghosh JG, Clark JI 2006 Chaperone-like activity revealed in the matricellular protein SPARC. J Cell Biochem 98:701–705 Eustace BK, Jay DG 2004 Extracellular roles for the molecular chaperone, hsp90. Cell Cycle 3:1098–1100 Eustace BK, Sakurai T, Stewart JK et al 2004 Functional proteomic screens reveal an essential extracellular role for hsp90 alpha in cancer cell invasiveness. Nat Cell Biol 6:507–514 Fares MA, Ruiz-González MX, Moya A, Elena SF, Barrio E 2002 Endosymbiotic bacteria: GroEL buffers against deleterious mutations. Nature 417:398 Korcsmaros T, Kovacs IA, Szalay MS, Csermely P 2007a Molecular chaperones: the modular evolution of cellular networks. J Biosci 32:441–446 Korcsmáros T, Szalay MS, Böde C, Kovács IA, Csermely P 2007b How to design multi-target drugs: target-search options in cellular networks. Expert Opin Drug Discov 2:1–10 Launay N, Goudeau B, Kato K, Vicart P, Lilienbaum A 2006 Cell signaling pathways to alphaBcrystallin following stresses of the cytoskeleton. Exp Cell Res 312:3570–3584 Lewandowska A, Gierszewska M, Marszalek J, Liberek K 2006 Hsp78 chaperone functions in restoration of mitochondrial network following heat stress. Biochim Biophys Acta 1763:141–151 Marguet D, Lenne PF, Rigneault H, He HT 2006 Dynamics in the plasma membrane: how to combine fluidity and order. EMBO J 25:3446–3457 Nardai G, Csermely P, Soti ´´ C 2002 Chaperone function and chaperone overload in the aged. Exp Gerontol 37:1255–1260 Newman MEJ 2003 The structure and function of complex networks. SIAM Rev 45: 167–256 Palla G, Derenyi I, Farkas T, Vicsek T 2005 Uncovering the overlapping community structure of complex networks in nature and society. Nature 435:814–818 Queitsch C, Sangster TA, Lindquist S 2002 Hsp90 as a capacitor of phenotypic variation. Nature 417:618–624

54

CSERMELY ET AL

Raviol H, Sadlish H, Rodriguez F, Mayer MP, Bukau B 2006 Chaperone network in the yeast cytosol: Hsp110 is revealed as an Hsp70 nucleotide exchange factor. EMBO J 25:2510–2518 Rutherford SL, Lindquist S 1998 Hsp90 as a capacitor for morphological evolution. Nature 396:336–342 Schmitt E, Gehrmann M, Brunet M, Multhoff G, Garrido C 2007 Intracellular and extracellular functions of heat shock proteins: repercussions in cancer therapy. J Leukoc Biol 81:15–27 Soti C, Pal C, Papp B, Csermely P 2005 Chaperones as regulatory elements of cellular networks. Curr Opin Cell Biol 17:210–215 Szabadkai G, Bianchi K, Varnai P et al 2006 Chaperone-mediated coupling of endoplasmic reticulum and mitochondrial Ca2+ channels. J Cell Biol 175:901–911 Szalay MS, Kovács IA, Korcsmáros T, Böde C, Csermely P 2007 Stress-induced rearrangements of cellular networks: consequences for protection and drug design. FEBS Lett 581:3675– 3680 Tsigelny IF, Nigam SK 2004 Complex dynamics of chaperone-protein interactions under cellular stress. Cell Biochem Biophys 40:263–276 Wong P, Houry WA 2004 Chaperone networks in bacteria: analysis of protein homeostasis in minimal cells. J Struct Biol 146:79–89 Yerbury JJ, Stewart EM, Wyatt AR, Wilson MR 2005 Quality control of protein folding in extracellular space. EMBO Rep 6:1131–1136 Zhao R, Davey M, Hsu YC et al 2005 Navigating the chaperone network: an integrative map of physical and genetic interactions mediated by the hsp90 chaperone. Cell 120:715–727

DISCUSSION Asea: How would you design a network for the extracellular molecular chaperones? Csermely: We know very little about the chaperone interactions outside the cell. On the one hand, the extracellular space is much less crowded than the intracellular space and, therefore, we would expect fewer interactions, especially in the low-affinity range that is typical for chaperones. However, the extracellular matrix adds a lot of possibilities for a highly dynamic organization. This may put the networking features back on the stage. There is an additional trick here. The intracellular chaperone network might simply be continued in the extracellular space, due to the high dynamism of the plasma membrane. I am sorry for not being able to give an exact answer to the question, but this field is in its infancy. Macario: At a much simpler level, now we know that chaperones interact with other chaperones to form a team, such as the Hsp70(DnaK) team formed by it and Hsp40(DnaJ) and the nucleotide-exchange factor (e.g. GrpE in bacteria) and that some of these teams interact with other teams, for example the Hsp70(DnaK) team interacts with the GroEL/S team (Macario & Conway de Macario 2007), how do you integrate this simple knowledge with the more complex picture provided by systems biology?

CHAPERONE NETWORKS

55

Csermely: This approach is relatively young. We have data sets for six years good enough for analysis, and the scope of these data sets is relatively limited even today. Regarding the protein–protein interaction networks, so far the bacterial system has not been adequately addressed. The example you mentioned, GroEL and DnaK, cannot be put into a bacterial network because of the paucity of the data. However, in the yeast system, we can work with good reliability and efficiency. I hope in two or three years it will change because the other systems will have enough data. These molecular chaperone networks can be put in as integrators of the other complexes in the cell. That’s our current view. Hightower: One thing I thought was missing was work on integrative cellular functions. I think more of this is now happening, but more needs to come. Network theory will be helpful there. Recently, Juliet Lee and I have studied cell movement as an integrated cellular activity. Juliet has studied cell movement in fish keratocytes for many years. Recently, we have been studying human colonocytes. We can paint actin microfi laments red and Hsp27 green. As we watch the cells move around, when they make their projections from their surface they flash yellow indicating colocalization of actin and Hsp27. This really made me think of your network theory. Hsp27 may be linking signalling domains with cytoskeletal domains. Csermely: When people measure these integrative cellular responses, what is usually neglected is measuring the scatter. If the scatter is large, it is not always the problem of the graduate student: it can tell you that there is an imbalance in the system, or the system is not so well buffered by molecular chaperones or other proteins. Let me stress that we have an important take-home message here: the scatter does have a meaning. Nixon: It is becoming obvious that some extracellular chaperones are spatially constrained in lipid raft structures. What role do you think lipid rafts play in the emergence of these networks within the extracellular domain of the cell? Csermely: In the future I think the lipid raft structures will assume much more importance, along with microheterogeneities of the membrane. Hsp90, Hsp70 and other molecular chaperones are already part of lipid rafts (Triantafilou et al 2004), but at the moment we don’t know much about how rafts and microheterogeneities change in response to various stresses, and disease and ageing. It is an unexplored field. Calderwood: I was interested in what you said about comparing the chaperones with signalling molecules. You said they were enriched in unstructured disordered regions. When we think of signalling molecules we think of molecules that are in a precisely defined place in the pathway. If you were to look at the same pathway with Hsp90, it interacts with numerous enzymes along the pathway. Csermely: The reason for this enrichment in segments may not have a functional role at the level of the whole cell, but may participate in a functional role at the

56

CSERMELY ET AL

level of the protein–protein interactions. Unstructured regions expand the capture radius of a particular protein, because this segment can point well out of the main bulk of the protein, and can capture proteins that are outside the average radius of the original protein. What often happens, both in signalling and in molecular chaperones, is that these proteins expand the space where they can interact with other proteins by these features. In signalling, this helps the assembly of the pathway under conditions where it hasn’t been assembled before. In terms of molecular chaperones, they have to connect those proteins which need assistance to assemble. Therefore unstructured regions in both protein classes are probably more important in the ‘local’ functional sense than in the general functional sense meaning their function in the whole cellular network. Henderson: For obvious reasons you have had to present these networks as static structures. Clearly, networks are dynamic. Are you utilizing mathematical modelling to try to capture network dynamics? Csermely: Yes. At the moment the topology of the networks can be elucidated by relatively simple mathematics. We are just at the beginning. At the moment people working on these structural networks are at the point, where we start to build in the weights of the links (whether they are weak or strong), their direction and also the colouring of the links (whether they are activating or inhibiting). Mathematically, the theory isn’t good enough that we could accommodate all these link/edge properties at the moment. The perturbations are currently used in a simplistic form: we see how a single perturbation to a small degree is propagating through the network. We haven’t reached the point where we can do a gross differential equation treatment. This would be a major advance, but if you have a network with 100 000 elements, then this many differential equations will be obviously too much. Henderson: You don’t know how many interacting links you have. Perhaps there are hundreds of proteins in the network. Csermely: If it is the yeast, you have 6000, and in the human perhaps 30 000, but take into account post-translational modifications and you have 100 000. Henderson: You have to curtail the modelling to a subnetwork of a subnetwork. Csermely: Yes. These are two different approaches, and both are fruitful. The usual modelling approach is to confi ne the network to a small subset, say of 10 proteins, and find out all you can from this. This was fruitful in the elucidation of the role of molecular clocks and circadian rhythms (Elowitz & Leibler 2000, Ueda et al 2005), for example. Currently, the sort of network theory I introduced applies to the whole cell, and gives up a little of the precision of the interactions, because you can’t have both. Henderson: The network theory allows you to posit a hypothesis, but the big problem is testing it biologically. For years I have been trying to model cytokine

CHAPERONE NETWORKS

57

networks, and our big problem is that we can model them in a couple of days, but it takes the biologist a couple of years to test the hypothesis. Can you see yourself testing hypotheses generated by your network modelling systems? Csermely: Sometimes testing is not that difficult. One example of what we are doing in the lab is that we are elucidating signalling networks, and we assembled a general signalling network compiled from three different species. Along this process we found a lot of cross-talk candidates between various signalling pathways. In C. elegans it is relatively easy to test whether a protein is a cross-talk protein or not. Multhoff: Returning to the topic of lipid rafts, you were talking about protein– protein interactions, but for cell signalling protein–lipid interactions might be important, or protein–carbohydrate interactions, also. Csermely: Absolutely. The problem is twofold. First, we don’t have a big database on protein–lipid interactions: the lipidome is just emerging (Lu et al 2005). Secondly, many of these interactions are not that specific. I agree this is important, but there are conceptual problems at the moment. Gupta: You showed the largest number of interactions for Hsp70. There are many different forms of this protein. Are you distinguishing them in your analyses? Once you do that, do you still see a large number of interactions? Csermely: We do have data individually on all the isoforms of the HSPs. Interestingly, most of the interactions are coming from the cytosolic Hsp70. This doesn’t mean that this is disproportionally having a large number of interactions. It is probably the bias of the experimental methods, which provided the databases for the interactions (Korcsmaros et al 2007). Gupta: The chaperone protein is supposed to interact with a large number of proteins. So how can you distinguish between a network-type interaction and a normal functional interaction? Csermely: One of the problems with the current data on chaperones is that we don’t have a good discrimination between highly transient chaperone–client interactions and network-type, regularly occurring chaperone–protein interactions. There might be later on a basis of discrimination by means of the affinity constants, but we only have a few of these now. But a warning is appropriate here: I’m sure we will have a continuum of binding affinities, and it will be difficult to dissect from which affinity value on a client–chaperone interaction is ‘just a normal functional interaction’ and not a regularly occurring network-type interaction. Lund: I was intrigued by the idea that when a stress occurs, chaperones are wrapped up with dealing with this problem, and this causes a separate module to become disconnected. When the stress is finished these modules become rewired, but may be rewired in subtly different ways. This is a mechanism for epigenetic change, because you are not changing the protein, but the way in which the network is put together. Is there any evidence for this that you are aware of?

58

CSERMELY ET AL

Csermely: Not direct evidence, just a bit of indirect evidence (Korcsmaros et al 2007, Bobula et al 2006). The basic assumption does not seem to be completely valid that if you inhibit molecular chaperones and the system is beginning to behave differently in this epigenetic way, then the reason is that molecular chaperones cannot fold a certain number of proteins, and therefore those proteins will be in a different form in the cell. This may still be valid, but it is not the complete picture. Hightower: There are interesting observations about the properties of the cytoprotected state of cells in tissues being different from normal cell and tissue functions. Cells are less responsive to extracellular cues, i.e. they are non-mitotic and non-apoptotic, when they are in the cytoprotected state, suggesting the kinds of changes you might expect a different network reassembly to be able to accomplish. Lund: This is a mechanism of cellular learning or cellular memory, then. Csermely: Learning in the sense of cells or other unconscious assemblies can be rationalized as a topological change in the network system. This may be actually the underlying mechanism of learning in our brains, where neural cells change their network assembly (both in the form of changing the topology of their anatomical contacts in the long-term, as well as changing the emerging oscillation networks in the short-term). References Bobula J, Tomala K, Jez E, Wloch DM, Borts RH, Korona R 2006 Why molecular chaperones buffer mutational damage: a case study with a yeast Hsp40/70 system. Genetics 174:937–44 Elowitz MB, Leibler S 2000 A synthetic oscillatory network of transcriptional regulators. Nature 403:335–338 Korcsmaros T, Kovacs IA, Szalay MS, Csermely P 2007 Molecular chaperones: the modular evolution of cellular networks. J Biosci 32:441–446 Lu Y, Hong S, Tjonahen E, Serhan CN 2005 Mediator-lipidomics: databases and search algorithms for PUFA-derived mediators. J Lipid Res 46:790–802 Macario AJL, Conway de Macario E 2007 Molecular chaperones: multiple functions, pathologies, and potential applications. Front Biosci 12:2588–2600 Triantafi lou M, Miyake K, Golenbock DT, Triantafi lou K 2002 Mediators of innate immune recognition of bacteria concentrate in lipid rafts and facilitate lipopolysaccharide-induced cell activation. J Cell Sci 115:2603–2611 Ueda HR, Hayashi S, Chen W et al 2005 System-level identification of transcriptional circuits underlying mammalian circadian clocks. Nat Genet 37:187–192

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Unusual cellular disposition of the mitochondrial molecular chaperones Hsp60, Hsp70 and Hsp10 Radhey S. Gupta, Nallur B. Ramachandra, Timothy Bowes and Bhag Singh Department of Biochemistry, McMaster University, Hamilton, Ontario L8N 3Z5, Canada

Abstract. A number of molecular chaperones in mammalian cells are localized in mitochondria and they are presumed to function mainly within this organelle. However, there is now compelling evidence that these chaperones are also localized at a variety of other sites/compartments in cells where they perform important functions. These proteins include: (i) the major chaperonin Hsp60 (or P1), which was identified in mammalian cells as a protein altered in mutants resistant to microtubule inhibitors and is involved in numerous functions at the cell surface and in other compartments; (ii) the Hsp10 or Cpn10 protein, which is a co-chaperone for Hsp60 in protein folding but also serves as an early pregnancy factor in maternal serum; and (iii) the mHsp70 protein, which plays a central role in mitochondrial protein import but is also important for cellular senescence (mortalin) and antigen presentation processes. The presence of these mitochondrial chaperones at specific extramitochondrial locations greatly broadens the range of functions that they can carry out in cells. However, these observations also raise important questions regarding the mechanisms by which these proteins reach these extramitochondrial locations. My paper will review some work in this area and discuss the significance of these results. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 59–73

A number of major molecular chaperones in eukaryotic cells (namely Hsp60, Hsp10 and mHsp70) are bona fide mitochondrial (m) proteins, which are initially targeted to and localized in this compartment. However, it is now well established that these chaperone proteins, in addition to mitochondria, are also localized and perform important functions at a variety of other sites/compartments within cells (Soltys & Gupta 1999, 2000, Gupta et al 2005). These nuclear DNA encoded proteins are translated in the cytosol prior to their import into mitochondria. The import of protein into mitochondria is generally a highly efficient process that occurs rapidly and mostly to completion (Neupert 1997). The proteins, once imported into mitochondria, are not known to exit from this compartment under 59

60

GUPTA ET AL

normal physiological conditions. Hence, the presence or functioning of these molecular chaperones at extramitochondrial sites raises important questions regarding the possible mechanisms by which they are reaching these locations (Soltys & Gupta 1999, 2000). Despite extensive work on the involvement of these proteins in extramitochondrial functions, the question as to how they might be reaching these sites has not received serious attention. Our work on these molecular chaperones and other mitochondrial proteins is leading to the understanding that the presence of mitochondrial proteins at extramitochondrial locations is a widespread phenomenon and it likely occurs via novel, as of yet, unidentified mechanism(s) for protein export from mitochondria to other specific locations. Importantly, many of these proteins are involved in functions at these extramitochondrial locations, which are very different from their established functions within mitochondria. This indicates that in cases where mitochondrial proteins are detected at other cellular locations (i.e. moonlighting proteins), one cannot assume that they are performing the same traditional chaperone functions (e.g. protein folding or unfolding) whose mechanisms have been well studied in simpler model systems. The involvement of mitochondrial proteins in a wide variety of cellular functions in different locations also indicates that mitochondria play a much broader and central role in the cell than is presently recognized. The story of Hsp60 The 60 kDa heat shock protein (Hsp60), also known as chaperonin 60 or Cpn60, is the eukaryotic homologue of the bacterial GroEL protein (Jindal et al 1989). It is one of the major molecular chaperones in both stressed and unstressed cells, and in mammalian cells it is primarily found in mitochondria (Soltys & Gupta 1996, Gupta 1990). This protein is synthesized as a larger precursor containing an N-terminal mitochondrial targeting sequence (MTS), which is cleaved during import of the protein into the matrix compartment (Jindal et al 1989, Gupta 1990). The Hsp60 found in mitochondria and various other locations is the mature form of the protein that lacks the MTS sequence (Ikawa & Weinberg 1992, Jones et al 1994, Khan et al 1998). Hsp60 was discovered in mammalian cells as a protein (P1) that was specifically altered in Chinese hamster cell mutants resistant to the microtubule inhibitor podophyllotoxin (Gupta et al 1982, Jindal et al 1989, Gupta 1990). These mutants affected in the P1 (or Hsp60) protein exhibited highly specific cross-resistance to various anti-tubulin drugs. The co-release of P1 with tubulin under a number of conditions indicated that it was behaving as a tubulin-associated protein (Gupta et al 1982). Furthermore, crude microtubule preparations made from these mutants exhibited reduced binding of [3H] colchicine and [3H] podophyllotoxin to tubulin,

MITOCHONDRIAL CHAPERONES IN THE CELL

61

indicating that the genetic alterations in them were affecting drug binding to tubulin (Gupta 1983). However, subsequent cell fractionation and immunofluorescence studies indicated that this protein was localized in the mitochondrial matrix compartment (Gupta & Austin 1987). This observation was totally unexpected in view of our genetic and biochemical studies and the fact that tubulin, with which this protein interacted, is not found in mitochondria. The identity of the P1 protein as Hsp60 became clear to us after several years, when its cDNA was sequenced in our lab (Jindal et al 1989). The mutation in Hsp60 in these mutants has been identified as a G → A substitution in Chinese hamster Hsp60 cDNA leading to change of a conserved Gly residue (GGA) to Asp (GAA) (unpublished results). To understand how a mutation in this protein could lead to resistance to antimitotic drugs, it became important for us to understand how and where this mitochondrial matrix protein interacted with tubulin (Gupta 1990). This has led us to examine, in detail, the subcellular localization of Hsp60 in many different types of cells and tissues using high-resolution techniques. These studies have revealed that Hsp60, in addition to mitochondria, is also present at many other specific locations, including on the cell surface (Soltys & Gupta 1999, 2000, Gupta et al 2005). The presence of Hsp60 on the cell surface was fi rst inferred when murine and human T cells reactive with mycobacterial Hsp60 were stimulated by a protein present on the surface of stressed macrophages and certain tumour cells. This stimulation was blocked by antibodies to Hsp60 (Fisch et al 1990, Kaur et al 1993). This inference has been confirmed by many other observations including the direct demonstration of Hsp60 on the cell surface by surface labelling and backscattered electron imaging of intact cells (Soltys & Gupta 1997). The ability of Hsp60 to interact with a wide variety of proteins in the plasma membrane (e.g. p21ras, amino acid transporters, protein kinase A, high-density lipoproteins, prion protein PrPc, histone 2B, HIV transmembrane glycoprotein gp41, etc.) (see Table 1) indicates that it is likely functioning as a membrane chaperone, enabling association and interaction of various proteins with their respective ligands (Soltys & Gupta 2000). It is of particular interest that free tubulin, to which the various antimitotic drugs bind with high specificity and avidity, is also an important component of the plasma membrane (Stephens 1986, Soltys & Gupta 2000), although its function in this location is not understood. The presence of both Hsp60 and tubulin in the plasma membrane provides a plausible explanation as to how mutational changes in the Hsp60 protein can cause resistance to tubulin-binding antimitotic drugs. Based upon our finding that Hsp60 (P1) showed association with tubulin, it is likely that tubulin in the plasma membrane is associated with Hsp60, such that mutational changes in Hsp60 can indirectly alter drug binding to the tubulin (Soltys & Gupta 2000). Drug binding and uptake studies with these mutants

62

GUPTA ET AL

TABLE 1 Involvement of Hsp60 proteins in many extramitochondrial functions and locations Proposed function or association with other proteins

Subcellular localization (if known)

Altered in CHO cell mutants resistant to antimitotic drugs and identified as a tubulin-associated protein Immune response

Mitochondria and plasma membrane

Associated with p21Ras Associated with A and L systems of amino acid transporters Plasma membrane chaperone for histone 2B High density lipoprotein binding protein Increases susceptibility to complement-dependent lysis Assembly of proteins into different types of secretory granules and peroxisomes Involvement in cell-cell signalling Receptor for Listeria adhesion protein Suppression Hsp60-ts mutant phenotype in yeast by cytosolic SCS1 protein Interaction with immunosupressant drug mizoribidine Interaction with prion protein PrPc

Plasma membrane of stressed macrophages and certain tumour cells Plasma membrane of various cell lines Plasma membrane of CHO cells Plasma membrane of human leukaemic CD4(+) CEM-SS T cell line Primary cultured hepatocytes Stressed aortic endothelial cells Various pancreatic and pituitary secretory granules and peroxisomes Extracellular Plasma membrane of Caco-2 cells Unknown

Reference Gupta et al 1982, Jindal et al 1989, Gupta 1990, Soltys & Gupta 1997 Fisch et al 1990, Kaur et al 1993 Ikawa & Weinberg 1992 Jones et al 1994, Woodlock et al 1997 Khan et al 1998

Bocharov et al 2000 Xu et al 1994 Brudzynski et al 1992, Soltys & Gupta 1996, Cechetto et al 2000 See other chapters in this volume Wampler et al 2004 Shu & Hallberg 1995

Unknown

Itoh et al 1999

Unknown

Edenhofer et al 1996

indeed provide evidence that binding of the antimitotic drugs to tubulin is affected at the plasma membrane level (unpublished results). It is also interesting that a number of our antimitotic drug resistant mutants display alterations in the other major molecular chaperone protein, Hsc70, which also behaves as a tubulin-associated protein (Gupta 1990, Ahmad et al 1990).

MITOCHONDRIAL CHAPERONES IN THE CELL

63

Subcellular localization of Hsp60 in cultured cell lines by high resolution immunogold labelling techniques has revealed that while the majority of Hsp60 labelling in these cells is within mitochondria, 15–20 % of the reactivity is consistently observed at discrete extramitochondrial sites, including unidentified cytoplasmic vesicles and granules, sites on the endoplasmic reticulum (ER), and at the cell surface (Soltys & Gupta 1996). In rat tissue sections derived from heart, kidney (proximal and distal tubules), skeletal muscle, adrenal gland and spleen, reactivity to Hsp60 antibody was mainly restricted to mitochondria. However, in a number of other tissues, strong and specific labelling due to Hsp60 antibody was also observed in other compartments (Cechetto et al 2000). In pancreatic β cells, strong reactivity with Hsp60 antibodies was seen in the central core of the mature insulin secretory granules (ISGs) (Brudzynski et al 1992). In rat liver, Hsp60 antibodies labelled both mitochondria and peroxisomes and the reactivity in the latter case was primarily associated with the urate oxidase crystalline core (Soltys & Gupta 1996). In pancreatic acinar cells and pituitary cells, strong reactivity with Hsp60 antibodies was observed in zymogen granules (ZGs) and growth hormone granules (GHGs), respectively (Cechetto et al 2000). The labelling of these compartments was completely abolished upon preadsorbtion of the antibodies with recombinant Hsp60, providing evidence that it was due to Hsp60. It is of interest that although strong labelling due to Hsp60 antibodies was observed in ZGs, GHGs and ISGs, there was negligible labelling seen in immature ISGs, as well as in the ER and Golgi compartments (Cechetto et al 2000). These results are in marked contrast to those obtained with antibodies to other proteins such as insulin and amylase, which showed strong labelling of the entire ER–Golgi pathway (Rothman 1975, Sadacharan et al 2001). These results suggest that Hsp60 could be reaching these granules/compartments via a novel mechanism that is different from the classical ER–Golgi pathway. The possible function of Hsp60 in these compartments is currently not known. In ISGs, Hsp60 is associated with the central core of the mature ISGs, but it is not present in immature ISGs. The main difference between these two types of granules is that during transition from immature ISGs to mature ISGs, proinsulin is enzymatically cleaved to form insulin, which is then extensively condensed to form a highly compacted core (Orci et al 1988). The central core of the mature ISGs thus represents a highly organized, supramolecular structure, which contains highly concentrated insulin in a functional form that is ready to be secreted. In a similar manner, in peroxisomes, Hsp60 is associated with the urate oxidase crystalline cores, which likely require a chaperone for their assembly. In accordance with the established functions of Hsp60 in the formation of oligomeric protein complexes and in protein secretion, it is likely that this protein is playing a similar chaperone role in the condensation of proteins

64

GUPTA ET AL

within these granules and in maintaining the highly compacted proteins within them in a functional state (Brudzynski et al 1992, Cechetto et al 2000, Soltys & Gupta 2000). Subcellular distribution of Hsp10 (Cpn10) Cpn10, which serves as a co-chaperone for Hsp60, is also localized in the matrix compartment of mammalian mitochondria (Sadacharan et al 2001). Unlike Hsp60 or most other mitochondrial matrix proteins, Cpn10 does not contain a cleavable MTS, but instead its N-terminal sequence has the ability to form an amphipathic alpha helix, which enables it to cross the mitochondrial membrane. Surprisingly, a protein corresponding to Cpn10 was independently identified as early pregnancy factor (EPF), which appears in maternal serum within 24 hours of fertilization (Cavanagh & Morton 1994). The evidence that these two proteins are identical is provided by several observations including the complete identity of their amino acid sequences and the ability of purified rat Hsp10 to serve as EPF in an EPF bioactivity assay (Cavanagh 1996). The subcellular distribution of Cpn10 in rat tissues has been examined by means of a high-resolution immuno-electron microscopy technique employing polyclonal antibodies raised against different regions of human Hsp10 (Sadacharan et al 2001). In all rat tissues, antibodies to Hsp10 showed strong labelling of mitochondria. However, in a number of tissues, in addition to mitochondria, strong and specific labelling with the Hsp10 antibodies was also observed in several extramitochondrial compartments. These sites included zymogen granules (ZGs) in pancreatic acinar cells, growth hormone granules (GHGs) in anterior pituitary and pancreatic polypeptide granules in islet cells (Sadacharan et al 2001). These granules likely provide the pathway for secretion of this protein into the blood stream, where it can serve the EPF function. In contrast to these secretory granules, the labelling in cytoplasm, nucleus and ER was at near background levels. Additionally, strong reactivity of the Cpn10 antibodies has also been observed within mature red blood cells (Sadacharan et al 2001). This observation is surprising because red blood cells, or erythrocytes, are believed to be devoid of mitochondria and various other organelles. However, it is possible that although the mature erythrocytes extrude all mitochondria, they retain specific mitochondrial proteins that may be required for particular functions. Our studies also indicate that similar to Cpn10, cross-reactive proteins to Hsp60 antibodies are also present in the red blood cells (unpublished results). The possible functions of Hsp60 and Hsp10 in the red blood cells is currently unclear, but it is possible that they may be involved in the assembly or functioning of haemoglobin, which is the main constituent of red blood cells.

MITOCHONDRIAL CHAPERONES IN THE CELL

65

Subcellular distribution of mitochondrial Hsp70 mHsp70 is the mitochondrially targeted member of the highly conserved Hsp70/ DnaK family of proteins. This protein has been independently identified as mortalin (Kaul et al 2002), as the 74 kDa peptide binding protein (PBP74) (VanBuskirk et al 1991), and as the 75 kDa glucose regulating protein. Similar to Hsp60, it contains an N-terminal MTS, which is responsible for its targeting to the mitochondrial matrix (Bhattacharyya et al 1995, Singh et al 1997). Within mitochondria, it plays a central role in the protein import process by binding to and pulling in unfolded polypeptide chains entering the matrix compartment through the TIM/TOM translocon (Hendrick & Hartl 1993). In an earlier study, where mHsp70 was identified as the PBP74, it was localized in the endocytic vesicles of B cells containing internalized antigen, ER and plasma membrane (VanBuskirk et al 1991). Studies on its localization by high-resolution immuno-electron microscopy show that in addition to mitochondria, this protein is also present at discrete subcellular locations including the plasma membrane, endocytic vesicles, and unidentified cytoplasmic granules in both Chinese hamster and BSC1 cells (Singh et al 1997). Furthermore, similar to Hsp60 and Hsp10, antibodies to mHsp70 also show strong reactivity towards a number of different types of granules including ZGs (Gupta et al 2005). Wadhwa and Kaul’s groups have implicated mHsp70 or ‘mortalin’ in conferring the senescence phenotype on cultured mammalian cells (Deocaris et al 2006, Kaul et al 2002). Their confocal immunofluorescence studies using antibodies to mortalin show a pancytosolic labelling of this protein in normal fibroblasts, whereas in immortal cells, its localization varied from a fibrous to granular staining of the perinuclear region (Wadhwa et al 1993). However, due to the limited resolution of confocal microscopy, it is difficult to determine from their studies whether the observed differences are due to altered mitochondrial distributions (or morphology) in different cells, or due to the differences in the subcellular localization of mHsp70 in these cells.

Significance of the presence of mitochondrial chaperones in other compartments The presence of the mitochondrial chaperones Hsp60, Hsp10 and mHsp70 in different locations outside of mitochondria raises the obvious question of how these proteins are reaching these sites. Recent studies show that the presence of mitochondrial proteins at extramitochondrial locations/sites is a widespread phenomenon and that many of these proteins carry out unique functions at these locations. For example, mitochondrial aspartate aminotransferase, which is responsible for the malate–aspartate shuttle in mitochondria, functions as a fatty acid transporter

66

GUPTA ET AL

when it is present on the cell surface (Isola et al 1995). Numerous other proteins, which were earlier thought to be exclusively localized in mitochondria under normal physiological conditions (namely cytochrome c) are also now shown to be present in other compartments (Soltys & Gupta 1999, 2000, Gupta et al 2005). Despite the importance of these findings for the cellular functions of mitochondria and these chaperones, the question of how these proteins are reaching these alternate locations has not yet received serious consideration. The main reason for this apparent apathy is that most mitochondrial proteins (nuclear DNA encoded) are translated in the cytosol before they are targeted to mitochondria. Hence, it is generally assumed that they can easily move from cytosol to other extramitochondrial locations. However, the available facts do not support this simplistic view. The m-proteins that are found at extramitochondrial locations are all mature forms of the proteins that are lacking the MTS. The cleavage of the MTS is carried out in a highly specific manner by a mitochondrial matrix resident protease. For both Hsp60 and mHsp70 (as well as numerous other m-proteins), it has been shown that the conversion of the precursor form into the mature form requires entry of the precursor (or at least its MTS) into mitochondria (Singh et al 1997, Soltys & Gupta 1996). In cells treated with potassium ionophores, nonactin or valinomycin, which inhibit mitochondrial import of the precursor, no conversion of the precursor to the mature form was observed. To account for the presence of mature forms of these proteins outside of mitochondria, two possible mechanisms have been suggested: according to the first, the proteins are partially imported into mitochondria so that their MTS is cleaved by the processing peptidase. This is then followed by retrograde translocation of the protein into the cytosol (Karniely & Pines 2005). In the alternate mechanism that we have suggested, the proteins are initially fully imported into mitochondria, and this is then followed by their export to other locations by, as of yet, unknown mechanism(s) (Soltys & Gupta 1999, 2000). Our recent studies on the subcellular targeting of yeast and human fumarase in live mammalian cells strongly indicate that the retrograde translocation mechanism does not play a significant role in the extramitochondrial presence of mammalian fumarase (Singh & Gupta 2006). Moreover, it has now been shown that a number of proteins that are encoded by m-DNA, which are transcribed and translated within mitochondria, are also present at other specific locations in cells (Gingrich et al 2004, Sadacharan et al 2005), where they perform novel functions (Gingrich et al 2004). These results strongly support the existence of mechanisms for the transport of proteins from mitochondria to other cellular compartments (Soltys & Gupta 1999, 2000). Although the mechanisms responsible for the translocation of proteins from mitochondria to other specific destinations remain to be characterized, our work on the extramitochondrial localization of various chaperone proteins has opened up the possibility that these proteins (as well as others) and mitochondria themselves play much broader roles in diverse cellular processes (Soltys & Gupta 1999,

MITOCHONDRIAL CHAPERONES IN THE CELL

67

2000). Furthermore, the multiple functional roles that are performed by a number of mitochondrial proteins in different compartments/locations indicate that these chaperone proteins could be involved in novel functions at these locations that could be entirely different from those known at present. Acknowledgments This work was supported by a research grant from the Canadian Institute of Health Research.

References Ahmad S, Ahuja R, Venner TJ, Gupta RS 1990 Identification of a protein altered in mutants resistant to microtubule inhibitors as a member of the major heat shock protein (hsp70) family. Mol Cell Biol 10:5160–5165 Bhattacharyya T, Karnezis AN, Murphy SP et al 1995 Cloning and subcellular localization of human mitochondrial hsp70. J Biol Chem 270:1705–1710 Brudzynski K, Martinez V, Gupta RS 1992 Immunocytochemical localization of heat-shock protein 60-related protein in beta-cell secretory granules and its altered distribution in nonobese diabetic mice. Diabetologia 35:316–324 Cavanagh AC 1996 Identification of early pregnancy factor as chaperonin 10: implications for understanding its role. Rev Reprod 1:28–32 Cavanagh AC, Morton H 1994 The purification of early-pregnancy factor to homogeneity from human platelets and identification as chaperonin 10. Eur J Biochem 222:551–560 Cechetto JD, Soltys BJ, Gupta RS 2000 Localization of mitochondrial 60-kD heat shock chaperonin protein (Hsp60) in pituitary growth hormone secretory granules and pancreatic zymogen granules. J Histochem Cytochem 48:45–56 Deocaris CC, Kaul SC, Wadhwa R 2006 On the brotherhood of the mitochondrial chaperones mortalin and heat shock protein 60. Cell Stress Chaperones 11:116–128 Fisch P, Malkovsky M, Kovats S et al 1990 Recognition by human V gamma 9/V delta 2 T cells of a GroEL homolog on Daudi Burkitt’s lymphoma cells. Science 250:1269–1273 Gingrich JR, Pelkey KA, Fam SR et al 2004 Unique domain anchoring of Src to synaptic NMDA receptors via the mitochondrial protein NADH dehydrogenase subunit 2. Proc Natl Acad Sci USA 101:6237–6242 Gupta RS 1983 Podophyllotoxin-resistant mutants of Chinese hamster ovary cells: crossresistance studies with various microtubule inhibitors and podophyllotoxin analogues. Cancer Res 43:505–512 Gupta RS 1990 Mitochondria, molecular chaperone proteins and the in vivo assembly of microtubules. Trends Biochem Sci 15:415–418 Gupta RS, Austin RC 1987 Mitochondrial matrix localization of a protein altered in mutants resistant to the microtubule inhibitor podophyllotoxin. Eur J Cell Biol 45:170–176 Gupta RS, Ho TK, Moffat MR, Gupta R 1982 Podophyllotoxin-resistant mutants of Chinese hamster ovary cells. Alteration in a microtubule-associated protein. J Biol Chem 257: 1071–1078 Gupta RS, Bowes T, Sadacharan SK, Singh B 2005 Intracellular disposition of mitochondrial molecular chaperones: Hsp60, mHsp70, Cpn10 and TRAP-1. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, New York, p 22–41 Hendrick JP, Hartl FU 1993 Molecular chaperone functions of heat-shock proteins. Annu Rev Biochem 62:349–384

68

GUPTA ET AL

Ikawa S, Weinberg RA 1992 An interaction between p21ras and heat shock protein hsp60, a chaperonin. Proc Natl Acad Sci USA 89:2012–2016 Isola LM, Zhou SL, Kiang CL, Stump DD, Bradbury MW, Berk PD 1995 3T3 fibroblasts transfected with a cDNA for mitochondrial aspartate aminotransferase express plasma membrane fatty acid-binding protein and saturable fatty acid uptake. Proc Natl Acad Sci USA 92:9866–9870 Jindal S, Dudani AK, Singh B, Harley CB, Gupta RS 1989 Primary structure of a human mitochondrial protein homologous to the bacterial and plant chaperonins and to the 65kilodalton mycobacterial antigen. Mol Cell Biol 9:2279–2283 Jones M, Gupta RS, Englesberg E 1994 Enhancement in amount of P1 (hsp60) in mutants of Chinese hamster ovary (CHO-K1) cells exhibiting increases in the A system of amino acid transport. Proc Natl Acad Sci USA 91:858–862 Karniely S, Pines O 2005 Single translation-dual destination: mechanisms of dual protein targeting in eukaryotes. EMBO Rep 6:420–425 Kaul SC, Taira K, Pereira-Smith OM, Wadhwa R 2002 Mortalin: present and prospective. Exp Gerontol 37:1157–1164 Kaur I, Voss SD, Gupta RS, Schell K, Fisch P, Sondel PM 1993 Human peripheral gamma delta T cells recognize hsp60 molecules on Daudi Burkitt’s lymphoma cells. J Immunol 150:2046–2055 Khan IU, Wallin R, Gupta RS, Kammer GM 1998 Protein kinase A-catalyzed phosphorylation of heat shock protein 60 chaperone regulates its attachment to histone 2B in the T lymphocyte plasma membrane. Proc Natl Acad Sci USA 95:10425–10430 Neupert W 1997 Protein import into mitochondria. Annu Rev Biochem 66:863–917 Orci L, Vassalli J-D, Perrelet A 1988 The insulin factory. Sci Am 259:85–94 Rothman SS 1975 Protein transport by the pancreas. Science 190:747–753 Sadacharan SK, Cavanagh AC, Gupta RS 2001 Immunoelectron microscopy provides evidence for the presence of mitochondrial heat shock 10-kDa protein (chaperonin 10) in red blood cells and a variety of secretory granules. Histochem Cell Biol 116:507–517 Sadacharan SK, Singh B, Bowes T, Gupta RS 2005 Localization of mitochondrial DNA encoded cytochrome c oxidase subunits I and II in rat pancreatic zymogen granules and pituitary growth hormone granules. Histochem Cell Biol 124:409–421 Singh B, Gupta RS 2006 Mitochondrial import of human and yeast fumarase in live mammalian cells: retrograde translocation of the yeast enzyme is mainly caused by its poor targeting sequence. Biochem Biophys Res Commun 346:911–918 Singh B, Soltys BJ, Wu ZC, Patel HV, Freeman KB, Gupta RS 1997 Cloning and some novel characteristics of mitochondrial Hsp70 from Chinese hamster cells. Exp Cell Res 234: 205–216 Soltys BJ, Gupta RS 1996 Immunoelectron microscopic localization of the 60-kDa heat shock chaperonin protein (Hsp60) in mammalian cells. Exp Cell Res 222:16–27 Soltys BJ, Gupta RS 1997 Cell surface localization of the 60 kDa heat shock chaperonin protein (hsp60) in mammalian cells. Cell Biol Int 21:315–320 Soltys BJ, Gupta RS 1999 Mitochondrial-matrix proteins at unexpected locations: are they exported? Trends Biochem Sci 24:174–177 Soltys BJ, Gupta RS 2000 Mitochondrial proteins at unexpected cellular locations: export of proteins from mitochondria from an evolutionary perspective. Int Rev Cytol 194:133–196 Stephens RE 1986 Membrane tubulin. Biol Cell 57:95–109 VanBuskirk AM, DeNagel DC, Guagliardi LE, Brodsky FM, Pierce SK 1991 Cellular and subcellular distribution of PBP72/74, a peptide-binding protein that plays a role in antigen processing. J Immunol 146:500–506 Wadhwa R, Kaul SC, Mitsui Y, Sugimoto Y 1993 Differential subcellular distribution of mortalin in mortal and immortal mouse and human fibroblasts. Exp Cell Res 207:442–448

MITOCHONDRIAL CHAPERONES IN THE CELL

69

DISCUSSION Panayi: I can visualize the idea that you have tubulin on the plasma membrane and this is associated with the two stress proteins, but I can’t visualize how mutation in either one or both proteins leads to the drug not being able to bind. Is it because the tubulin in the plasma membrane is now not present at a high enough concentration, or is it present in a different configuration? Gupta: We envisage that tubulin in the plasma membrane is present in a complex form with Hsp60 and Hsp70 proteins. When Hsp60 or Hsp70 is mutated, the complex conformation is altered such that binding of the antimitotic drug(s) to tubulin is reduced or prevented. Csermely: After the mutations, does the complex still form, or is the tubulin released from the chaperones? Gupta: In the mutant cell there is no change in the cell apart from the fact that it has become resistant to the drug. In other respects, the cells are normal. Our results indicate that mutations in these proteins do not affect their interaction with tubulin, but only the drug binding to tubulin. Panayi: If you were to transfect these cells with wild-type Hsp60, what would be the outcome? Gupta: We have done an experiment of this nature a long time ago. Most people aren’t aware of our finding because these proteins weren’t then known as Hsp60 and Hsp70. In these experiments, cell hybrids were made between the resistant and sensitive cells, and their resistance to the drug was examined (Gupta et al 1982). These hybrids, in which the ratio of the wild-type to the mutant proteins was 3:1, showed an intermediate level of resistance to the drug, indicating that the degree of drug-resistance is dependent upon the relative amounts of the wild-type vs. mutant proteins. Asea: You mentioned that the mitochondrial Hsp60 is transported via the nonclassical protein pathway. Have you looked further to find out which pathway this is? Gupta: That has proven to be a difficult problem to investigate. There are no good systems to study this problem. The difficulty in this regard lies in the fact that most of the nuclear encoded proteins are translated in the cytosol. In the cytosol, we need to clearly distinguish the newly synthesized protein from those that are coming out from mitochondria, which has proven difficult. Although some proteins are encoded by mitochondrial DNA, they cannot be suitably tagged because of the difficulties of introducing genes into mitochondria. We have preliminary evidence that certain kinds of secretory granules containing Hsp60 are budding as well as fusing with mitochondria. However, we are still looking for ways to answer this question more definitively. Because some of these problems challenge the conventional thinking, it has proven difficult to obtain grant support

70

GUPTA ET AL

for this work. However, an understanding of the mechanisms responsible for trafficking of these proteins from mitochondria to other cellular destinations requires sustained efforts and funding. Asea: Maybe one way of overcoming the funding problem is to show that there is localization into the plasma membrane. Do you see any release to the extracellular space? This might be the mature protein that you could follow. Gupta: I am sure there will be some release. Many of these secretory granules release their content to the outside by fusion with the plasma membrane. If Hsp60 or other proteins are found in the various granules—insulin secretory granules, zymogen granules, growth hormone granules—this provides a natural pathway for release of this protein. The problem is that the amount seen outside is very small. Sometimes the techniques that are available aren’t sensitive enough. Asea: There are some methods, such as serial ultracentrifugation, that can be used for the recovery of exosomes. We have used this technique to study the release of Hsp70 within exosomes. Once you get the method going you can recover relatively high concentration of exosomes. Gupta: That is what we need. Multhoff: We have also worked with taxol and vincristine, which react with tubulin. What we see is in line with your results: Hsp70 is increased in the cytosol but also in the plasma membrane. We also find exosomal release of Hsp70. Apart from cytosolic and membrane-spanning proteins another characteristic of exosomes is that they contain a lot of the small GTPase Rab4, which is characteristic for a protein transport pathway from the early endosomes directly to the plasma membrane. Coates: Has this mechanism of resistance been looked at in bacteria? Gupta: The antimitotic drugs don’t act on bacteria. Tubulin is not found in bacteria. Coates: Bacteria have many antibiotics that inhibit their replication. They have hundreds of different mechanisms of resistance. Mitochondria and bacteria, of course, are connected. With something like chaperones, which have been around for 3 billion years, you would expect to see some mechanism of antibiotic resistance in bacteria that was linked. Gupta: We stumbled on the involvement of Hsp60 and Hsp70 in resistance to antimitotic drugs only by chance. Hence, it is possible that if you look at enough antibiotics, you may eventually find some where Hsp60 or Hsp70 are involved in a similar role in bacteria. Nixon: It almost seems counterproductive that the chaperones produced in the cytoplasm move into the mitochondria, and then from there go to the plasma membrane. Is there any evidence that the chaperones are binding to client proteins within the mitochondria and then facilitating their delivery to the plasma membrane?

MITOCHONDRIAL CHAPERONES IN THE CELL

71

Gupta: Mitochondria are effectively bacteria living in eukaryotic cells. We have to understand how that the ancestral eukaryotic cell came into formation (Gupta & Golding 2000). At the time when the endosymbiont was taken up by the preeukaryotic cell that did not have mitochondria, the functions of the two cells had to be integrated. Bacteria have many different mechanisms for protein secretion across membranes (Soltys & Gupta 2000). Hsp60 is bacterially derived in the sense that it comes from mitochondria. Although this gene has been transferred to the nucleus, the cells still retain the original mechanism by which this protein was functioning in the original endosymbiont. Although the initial import of this protein into mitochondria and subsequent export to other locations may seem counterproductive, it probably has to do with the manner in which the mitochondria and eukaryotic cells have evolved. Feder: Is the glycine to glutamate mutation you see a conserved residue in other GroELs? Where is it in the structure? Gupta: It is in the N-terminal half, in a region close to the ATPase domain. We have looked at the ATPase activity and it is not affected. The mutation is such that other essential cellular functions of this protein are preserved; if they are compromised then the mutation will be lethal and the mutants won’t be selected. Specifically, the mutation alters tubulin binding and apparently nothing else. Macario: Is it possible to absolutely exclude the possibility that some of the Hsp60 in the cytosol is synthesized without the mitochondrial retention signal? Gupta: We have done studies on a number of proteins tagged with EGFP (enhanced green fluorescent protein) at the C-terminal end (Singh & Gupta 2006). We also added a nuclear localization signal after the EGFP so that any protein made without the mitochondrial targeting sequence or not imported into mitochondria is targeted to nucleus, where it can be sensitively detected. Then we asked the question where does the newly synthesized protein with a mitochondrial targeting sequence on one end and a nuclear localization signal at the other end localize? The tagged-protein is detected primarily in mitochondria and is very efficiently transported to this organelle. These studies have been carried out with a number of proteins. These results provide evidence that there is only a single transcription initiation, and no alternative splicing or internal translation leading to synthesis of the protein without the mitochondrial targeting sequence. Macario: If they go to the mitochondria and then are released back into the cytosol, you can predict that there should be a mechanism for this reverse translocation, so to speak, and you can also predict that the molecules participating in this mechanism may occasionally be affected by structural or functional abnormalities, thus generating a pathological condition, a chaperonopathy by defect (Macario & Conway de Macario 2005, 2007). Do you predict that something like that may happen?

72

GUPTA ET AL

Gupta: That should be possible. There should be some diseases where this occurs, but I’m not aware of any at the moment. Ellis: With respect to organelles having evolved from bacteria, there is a protein transport system found in both bacteria and plastids that transports the protein across membranes in a fully folded state. This is called the TAT system, because it depends on twin arginines in the signal sequence. It transports fully folded proteins into the periplasmic space in bacteria and into the thylakoid lumen of chloroplasts, so it would be interesting to see whether this system is also retained in mitochondria. In light of the claim that your findings change the paradigm about protein transport, we have to ask how good the quality of the data is. The data you showed were based on antibodies and you are detecting whatever the antibodies recognize, but this might be only a fragment of the protein. There are methods whereby you can label the surface of cells with a membraneimpermeant label, and then analyse those proteins that are labelled in a mass spectrometer. What is the state of the evidence with respect to mitochondria or Cpn60? Is there evidence that there is a full length Cpn60 on the surface of the cells by that method? Gupta: We and others have done a number of studies of this kind (reviewed in Soltys & Gupta 2000). In these studies intact cells were radioiodinated or biotinylated and subsequently immunoprecipitated using Hsp60 antibody. These studies revealed that the full length protein was present on the cell surface. Another study which employed cross-linking of cell surface proteins found the full-length Hsp60 cross-linked to the p21ras—a membrane-associated protein. In yet another study, full length Hsp60 on the plasma membrane in the lymphocyte cell line CEM-SS interacted with the histone 2B (Khan et al 1998). Again, you wouldn’t expect histone to be in the plasma membrane or interacting with a membrane-associated protein. Ellis: What about tubulin? Gupta: There is extensive evidence showing the presence of tubulin in plasma membrane (see Soltys & Gupta 2000, Stephens 1986), however, its cellular function is not known. There have also been studies showing that antimitotic drugs bound to a solid support such as sepharose can inhibit cells in mitosis (see Gupta & Dudani 1989). The drug in this case likely binds to the tubulin on the cells surface, and then through a signalling mechanism leads to cell growth arrest in mitosis. References Gupta RS, Dudani AK 1989 Mechanism of action of antimitotic drugs: a new hypothesis based on the role of cellular calcium. Med Hypotheses 28:57–69 Gupta RS, Golding GB 1996 The origin of the eukaryotic cell. Trends Biochem Sci 21:166–171

MITOCHONDRIAL CHAPERONES IN THE CELL

73

Gupta RS, Ho TK, Moffat MR, Gupta R 1982 Podophyllotoxin-resistant mutants of Chinese hamster ovary cells. Alteration in a microtubule-associated protein. J Biol Chem 257: 1071–1078 Khan IU, Wallin R, Gupta RS, Kammer GM 1998 Protein kinase A-catalyzed phosphorylation of heat shock protein 60 chaperone regulates its attachment to histone 2B in the T lymphocyte plasma membrane. Proc Natl Acad Sci USA 95:10425–10430 Macario AJL, Conway de Macario E 2005 Sick chaperones, cellular stress and disease. New Engl J Med 353:1489–1501 Macario AJL, Conway de Macario E 2007 Chaperonopathies by defect, excess, or mistake. Ann N Y Acad Sci 1113:178–191 Singh B, Gupta RS 2006 Mitochondrial import of human and yeast fumarase in live mammalian cells: retrograde translocation of the yeast enzyme is mainly caused by its poor targeting sequence. Biochem Biophys Res Commun 346:911–918 Soltys BJ, Gupta RS 2000 Mitochondrial proteins at unexpected cellular locations: export of proteins from mitochondria from an evolutionary perspective. Int Rev Cytol 194:133–196 Stephens RE 1986 Membrane tubulin. Biol Cell 57:95–109

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Cell surface molecular chaperones as endogenous modulators of the innate immune response Martha Triantafi lou, Daniel Sawyer, Abdiaziz Nor, Emmanouil Vakakis and Kathy Triantafi lou1 Infection and Immunity Group, School of Life Sciences, University of Sussex, Falmer, Brighton, BN1 9QG, UK

Abstract. Mammalian responses to bacterial products can lead to an uncontrolled inflammatory response that can be deadly for the host. It has been shown that the innate immune system employs at least three cell surface receptors, TLR4, CD14 and MD2, in order to recognize bacterial products. We have previously shown that heat shock proteins (HSPs) are also involved in the innate immune recognition. HSPs are a family of highly conserved proteins that act as molecular chaperones and assist in proper folding, assembly and intracellular trafficking of proteins. How HSPs reach the cell surface and how they are involved in the innate immune response still remain unclear. In the present study we investigated their association with the TLR4/CD14/MD2 complex in response to bacterial products and provide evidence that the Hsp70 and Hsp90 associate with TLR4 on the cell surface in response to stimulation by bacterial products. These associations seem to take place within lipid rafts. The addition of exogenous recombinant Hsp70 to cells in vitro results in a dose-responsive inhibition of the inflammatory signal cascade and cytokine production. Our studies reveal that HSPs may play an important role as endogenous regulators of the innate immune response. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 74–85

Heat shock proteins (HSPs) are a family of highly conserved proteins found in cells that have traditionally been described as intracellular chaperones, involved in facilitating protein folding, degradation, translocation across the membrane and disassembly of oligomers (Slavotinek & Biesecker 2001). While these properties suggest HSPs are intracellular chaperones, it is becoming increasingly apparent that HSPs participate in the generation of immune responses. Recent reports have 1

This paper was presented at the symposium by Kathy Triantafi lou, to whom correspondence should be addressed. 74

CHAPERONES AND INNATE IMMUNITY

75

shown that a number of stressful conditions up-regulate their expression, including cytokine stimulation and oxidized low density lipoprotein exposure (Ferm et al 1992, Frostegard et al 1996, Pockley et al 2003), as well as pathological conditions in a variety of human disease states, including ischaemia, inflammation and infection (Santoro 2000). HSPs seem to participate in the acquired immune responses by enhancing processing and presentation of antigens via the MHC class II pathway (Wang et al 2006), but also seem to be involved in the innate immune response, since they have been found to be involved in the recognition of bacterial lipopolysaccharide (LPS) (Triantafi lou et al 2001). At least three cell surface molecules have been recognised as components of the mammalian signalling receptor for LPS: CD14 (Wright et al 1990), TLR4 (Poltorak et al 1998, Qureshi et al 1999, Chow et al 1999) and MD2 (Nagai et al 2002). HSPs (Triantafi lou et al 2001, Byrd et al 1999) have been suggested to be part of this activation cluster, possibly acting as additional LPS transfer molecules. Toll-like receptors (TLRs) are the primary sensors of pathogen invasion that signal the initiation of both an innate immune response and the development of antigenspecific adaptive immunity. The TLR family includes 10–11 known members in mammals that have evolved to recognize and respond to conserved motifs present in viruses, bacteria, fungi and parasites. These conserved motifs are termed pathogen-associated molecular patterns (PAMPs). Although HSPs have also been shown to interact with TLRs intracellularly (Vabulas et al 2001, 2002a, b), the mechanism by which HSPs might be interacting with the TLR4/MD2 complex or other TLRs on the cell surface has not been revealed. We have previously shown that the innate recognition of LPS takes place within lipid rafts, and Hsp70 interacts with TLR4 within membrane microdomains (Triantafi lou & Triantafi lou 2003, 2004, Triantafi lou et al 2002, 2004). Although HSPs have been shown to be able to bind LPS (Gao & Tsan 2003) and to be involved in the innate recognition of bacterial products (Triantafi lou et al 2001, Byrd et al 1999), the possible mechanism for their participation in innate immunity has not been elucidated yet. Important questions still remain: which domain of Hsp70 is involved in binding LPS? Does Hsp70 act as an LPS-transfer molecule for the TLR4/MD2 complex? Can Hsp70 modulate LPS responses? Is Hsp70 the only chaperone involved in innate immune responses, or is this a common feature of all HSPs? And most importantly are they only involved in the innate recognition of bacterial products, or do they modulate immune responses for different microbial products? Results and discussion Recently, we have attempted to answer all these questions. Initially we investigated whether the addition of exogenous soluble Hsp70 was able to diminish LPS-

76

TRIANTAFILOU ET AL 1800 1600

LPS

1400

LPS + Hsp70

1200 1000 800 600 400 200 0 TNF-a

IL-6

FIG. 1. Hsp70 inhibits LPS-induced inflammatory responses. Human monocytes were either pre-incubated (grey bar charts) or not (white bar charts) with 100 µg/ml of human recombinant Hsp70 followed by stimulation with 100 ng/ml of LPS for 4 hours at 37 °C under 5% CO2 . Supernatants were subsequently collected and analysed for TNFα and IL6 secretion using CBA Flow Cytometric Bead Array from Becton Dickinson. The mean and standard deviation from a number of independent experiments is presented.

induced cytokine responses. It was shown that exogenous Hsp70 could inhibit LPS-induced tumour necrosis factor (TNF) α responses (Fig. 1), suggesting that Hsp70 plays a crucial role in modulating the innate recognition of LPS. Furthermore, in order to determine the molecular basis of LPS recognition by Hsp70, we generated recombinant truncated products of Hsp70 (C- and N-terminal fragments), as well as site-directed mutagenesis within the N-terminal domain. Our data suggest that the C-terminus of Hsp70 does not play any role in LPS-binding and that LPS binds to the N-terminal ATPase domain of Hsp70. In particular it seems that the base of the ATP-binding cleft is crucial for LPS binding, since mutations within the ATP-binding cleft inhibited LPS-binding to Hsp70. Since we had found that Hsp70 is able to associate with LPS and TLR4 and to modulate LPS-induced responses, the question that remains is whether this is a common feature of HSPs and whether they are able to interact and modulate all TLR-induced responses. This hypothesis seems to be supported by recent findings that Hsp10, similarly to Hsp70, is able to modulate LPS-induced responses (Johnson et al 2005, Vanags et al 2006), thus suggesting that HSPs have the capacity to interact with TLRs and interfere with ligand binding, as well as the subsequent inflammatory response. Thus we proceeded to investigate whether Hsp10 could interact with TLRs. Using fluorescence resonance energy transfer (FRET), a biophysical method that measures molecular associations, we showed that Hsp10 could associate with different TLRs in response to different microbial ligands (LPS, LTA, ssRNA, CpG DNA). When similar experiments were performed with Hsp70 we showed that Hsp70 could also interact with different TLRs in response

CHAPERONES AND INNATE IMMUNITY

77

40

E (%)

35 30

TLR2

25

TLR3

20

TLR4

15

TLR7

10

TLR9

5 0 LTA

Poly I:C

LPS

Viral ssRNA

Imiquinod ODN CpG

FIG. 2. Hsp70 heterotypic associations in response to different microbial ligands. Human monocytes were stimulated with different microbial structures for 10 min. Energy transfer between Hsp70 (Cy3) and the different receptors was measured from the increase in donor (Cy3) fluorescence after acceptor (Cy5) photobleaching. The percentage of energy transfer and standard deviation was calculated from three independent experiments.

to different microbial ligands (Fig. 2), thus suggesting that it is a common feature for HSPs to associate with TLRs, and that they are involved in the innate immune sensing of different microbial ligands. This association seems to be important since pre-incubation with either Hsp10 or Hsp70 could inhibit LPS/LTA/ssRNA/CpG DNA-induced responses. In addition, when we investigated the intracellular targeting and trafficking of the HSP-TLR complexes, we found that HSPs were targeted to different intracellular compartments depending on the microbial ligand along with their corresponding TLRs. In response to LPS and LTA, Hsp10 and Hsp70 were targeted to the Golgi apparatus whereas in response to ssRNA and CpG DNA, Hsp10 and Hsp70 were targeted to the endosomes. Overall the findings of our recent studies suggest that HSPs must play a role as immunomodulators of the innate immune response. In particular, it seems that they are part of a negative feedback pathway for TLRs that seems to be central in regulating an efficient immune response. Activation of TLRs is essential for protecting the host against invading pathogens, but unfortunately severe or chronic TLR signalling is involved in the pathogenesis of a number of autoimmune and chronic inflammatory diseases (Cristofaro & Opal 2006). Limitation of TLR signalling is therefore paramount to immune balance, and several types of negative regulators have been identified in recent years, including those acting as soluble decoy receptors and others that inhibit TLR signalling pathways. Thus we are left thinking that HSPs, in addition to their activity as chaperones, are essential regulators of the TLR-pathogen-sensing machinery.

78

TRIANTAFILOU ET AL

Acknowledgements This work was partially supported by CBio Ltd (Australia).

References Byrd CA, Bornmann W, Erdjument-Bromage H et al 1999 Heat shock 90 mediates macrophage activation by Taxol and bacterial lipopolysaccharide. Proc Natl Acad Sci USA 96:5645– 5650 Chow JC, Young DW, Golenbock D, Christ WJ, Gusovsky F 1999 Toll-like receptor-4 mediates lipopolysaccharide-induced signal transduction. J Biol Chem 274:10689–10692 Cristofaro P, Opal SM 2006 Role of Toll-like receptors in infection and immunity: clinical implications. Drugs 66:15–29 Ferm MT, Soderstrom K, Jindal S et al 1992 Induction of human hsp60 expression in monocytic cell lines. Int Immunol 4:305–311 Frostegard J, Kjellman B, Gidlund M, Andersson B, Jindal S, Kiessling R 1996 Induction of heat shock protein in monocytic cells by oxidized low density lipoprotein. Atherosclerosis 121:93–103 Gao B, Tsan MF 2003 Endotoxin contamination in recombinant human heat shock protein 70 (Hsp70) preparation is responsible for the induction of tumor necrosis factor a release by murine macrophages. J Biol Chem 278:174–179 Johnson BJ, Le TT, Dobbin CA et al 2005 Heat shock protein 10 inhibits lipopolysaccharideinduced inflammatory mediator production. J Biol Chem 280:4037–4047 Nagai Y, Akashi S, Nagafuku M et al 2002 Essential role of MD-2 in LPS responsiveness and TLR4 distribution. Nat Immunol 3:667–672 Pockley AG, Georgiades A, Thulin T, de Faire U, Frostegard J 2003 Serum heat shock protein 70 levels predict the development of atherosclerosis in subjects with established hypertension. Hypertension 42:235–238 Poltorak A, He XL, Smirnova I et al 1998 Defective LPS signaling in C3H/Hej and C57BL/ 10ScCr mice: mutations in TLR4 gene. Science 282:2085–2088 Qureshi ST, Lariviere L, Leveque G et al 1999 Endotoxin-tolerant mice have mutations in toll-like receptor 4 (TLR4). J Exp Med 189:615–625 Santoro MG 2000 Heat shock factors and the control of the stress response. Biochem Pharmacol 59:55–63 Slavotinek AM, Biesecker LG 2001 Unfolding the role of chaperones and chaperonins in human disease. Trends Genet 17:528–535 Triantafi lou M, Triantafi lou K 2003 Receptor cluster formation during activation by bacterial products. J Endotoxin Res 9:331–335 Triantafi lou M, Triantafi lou K 2004 Heat-shock protein 70 and heat-shock protein 90 associate with Toll-like receptor 4 in response to bacterial lipopolysaccharide. Biochem Soc Trans 32:636–639 Triantafi lou K, Triantafi lou M, Dedrick RL 2001 A CD14-independent LPS receptor cluster. Nat Immunol 4:338–345 Triantafi lou M, Brandenburg K, Gutsmann T, Seydel U, Triantafi lou K 2002 Innate recognition of bacteria: engagement of multiple receptors. Crit Rev Immunol 22:251–68 Triantafi lou M, Brandenburg K, Kusumoto S et al 2004 Combinational clustering of receptors following stimulation by bacterial products determines LPS responses. Biochem J 381:527–536 Vabulas RM, Ahmad-Nejad P, da Costa C et al 2001 Endocytosed HSP60s use Toll-like receptor 2 (TLR2) and TLR4 to activate the Toll/Interleukin-1 receptor signalling pathway in innate immune cells. J Biol Chem 276:31332–31339

CHAPERONES AND INNATE IMMUNITY

79

Vabulas RM, Braedel S, Hilf N et al 2002a The endoplasmic reticulum-resident heat shock protein Gp96 activates dendritic cells via the Toll-like receptor 2/4 pathway. J Biol Chem 277:20847–20853 Vabulas RM, Ahmad-Nejad P, Ghose S, Kirschning CJ, Issels RD, Wagner H 2002b Hsp70 as endogenous stimulus of the Toll/Interleukin-1 receptor signal pathway. J Biol Chem 277:15107–15112 Vanags D, Williams B, Johnson B et al 2006 Therapeutic efficacy and safety of chaperonin 10 in patients with rheumatoid arthritis: a double-blind randomised trial. Lancet 368: 855–863 Wang R, Kovalchin JT, Muhlenkamp P, Chandawarkar RY 2006 Exogenous heat shock protein 70 binds macrophage lipid raft microdomain and stimulates phagocytosis, processing, and MHC-II presentation of antigens. Blood 107:1636–1642 Wright SD, Ramos RA, Tobias PS, Ulevitch RJ, Mathison JC 1990 CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding-protein. Science 249:1431–1433

DISCUSSION Henderson: There are two rheumatologists in the audience: what do you think of the Hsp10 rheumatic study (Vanags et al 2006)? Gaston: It is topical, but viewed as a therapeutic trial it leaves a lot to be desired. There is only the merest hint that Hsp10 might do something. They had to subdivide their responding groups to show an effect. If it were being assessed as a standard disease modifying drug I don’t think the report would have been published. Panayi: They used a lot; it was repetitive stimulation; there were only 23 patients; there wasn’t a placebo group. And some of the patients actually had a flare of the disease. This scenario would imply that there must be a time element to this. The idea would be that at the beginning you would have a stimulus which uses the lipid raft as a way of stimulating an important event, namely the innate immune response. Subsequently, other events lead to the release of Hsp10 to dampen it down. Do you have any evidence that this sequence of events actually takes place? You are just adding exogenous stress protein and looking at the result. Triantafilou: The data with the fluorescent resonance energy transfer (FRET) were obtained without adding exogenous stress protein. We find a small amount of Hsp70 and Hsp90 in the lipid raft beforehand, and more is recruited after about 30 min of incubation with LPS. There seems to be a rapid appearance of more Hsps within the lipid rafts. We think this is the time frame. For the exogenous addition of Hsps, we preincubate for about 15 min and then add the ligand: LPS or ssRNA. We monitor the effects. We are giving the system a window of time. Panayi: If you took cells, stimulated them and followed the supernatant, do you see amounts of Hsp10 that are relevant to the kinds of concentrations that you are using to inhibit the reaction? Triantafilou: No.

80

TRIANTAFILOU ET AL

Csermely: I was rationalizing to myself the biological meaning of the sort of negative feedback that you described. It would make even more sense for me, if the whole phenomenon could be inhibited by the presence of damaged proteins. Have you ever added Hsp70 or Hsp10 in your experiments with a substrate-peptide binding to these chaperones quite tightly? Under these conditions you might expect an abrogation of your effect. This would make sense, because it is a titration of how the cell is feeling—whether it is feeling bad, and then this whole effect is not working, or whether it is feeling good, and then this effect is working. Triantafilou: That is what we are doing at the moment. Quintana: You have these effects with Hsp10, 70 and 90. What about Hsp60 which is also in the membrane? Triantafilou: We haven’t tested this at all. It is the next thing we wanted to try. Quintana: People claim that Hsp60 and 70 might be doing different things, activating macrophages. Is this correct? Triantafilou: We really don’t know. There is work suggesting that HSPs are immunostimulatory and other work suggesting they are anti-inflammatory. We don’t know why there is this disparity between the different studies. Perhaps it is a receptor-associated result, depending on what receptor it is associating with. Macario: Could it be related to different subpopulations of cells? Triantafilou: Yes, that’s why we want to try different cell types. It could be celltype specific. Miller-Graziano: We have the same type of results with Hsp27. Hsp27 can inhibit TLR4 receptor responses by being associated with, but not binding to, TLR4. In contrast, we can’t inhibit Hsp27 induced responses with CD14. We have also done FRET: it indicates close association of Hsp27 with TLR4 but not as close as expected if Hsp27 were actually binding TLR4. If we stimulate macrophage with LPS we can see Hsp27 in the supernatant. There is a commercial Hsp27 ELISA. We have also shown macrophage Hsp27 secretion at the western blot level. There is a separate pathway for T cell down-regulation by Hsp10. One publication says that Hsp10 down-regulates CD3 in a mouse model where TLR4 is expressed on the surface of mouse T cells (Akyol et al 2006). However, TLR4 is poorly expressed on human T cells and this Hsp10 pathway may be less functional there. Nevertheless, this mechanism of surface receptor modulation may be fairly universal for some of the heat shock proteins associated with the membranes in lipid rafts. If there is a pattern of association between TLR4 and some other receptor (like Hsp10-TLR4 and mouse T cells’ CD3), multiple TLR-HSP effects might exist. We have looked at TLR3 alterations by Hsp27 on dendritic cells (DCs). We chose TLR3 because of its strong stimulation of human DCs, and down-regulation of TLR3 might be unique to DCs. The current paradigm for human DCs is that TLR3 may be up-regulated onto

CHAPERONES AND INNATE IMMUNITY

81

the surface only on DCs, not on monocytes or macrophages, as part of the differentiation process. Only after TLR3 up-regulation, can you see some modulation of the receptors by HSP. This may be a protein association effect that is more typical of HSPs when they are on the membrane and associated with various receptors in the rafts. Pockley: We did some work with another stress protein, gp96, looking for cell surface receptors with flow cytometry at 4 °C. There is a body of evidence to indicate that many proteins are continually being sampled by DCs, and that they are being internalised by non-specific endocytosis (Norbury 2006). Within the cell culture situation, we can propose that these proteins are going into the cell via non-specific mechanisms. Some of the proteins could interact with intracellular receptors such as the TLRs. In such instances, the modulation of the cell surface receptors is mediated from the inside out, rather than the outside in. This is pure speculation: we haven’t done the experiments. When we incubated gp96 with DCs we couldn’t get any phenotypic changes, nor could we get them to secrete cytokines. However, when we looked at the morphology of the cells we saw changes in the absence of any surface phenotypic changes or cytokine secretion. We have also demonstrated that the gp96 is internalised when the experiment is done at 37 °C. Asea: It is important to consider the cell type. There is some interesting work from Willem van Eden’s lab showing that Tregs are stimulated by HSPs (van Eden 2007). Carol Miller-Graziano, haven’t you worked with Tregs? Miller-Graziano: Yes, we see a secondary induction by altering DCs with Hsp27. This is one of the problems of employing a dual DC–T cell system: you don’t know whether your effect is because you have changed the DC receptors, or directly inducted Tregs. The studies you are talking about use a murine system. The evidence is overwhelming that TLR4 signalling to mouse T cells is a very strong signal for Tregs. If HSP affects TLR4, it could trigger Tregs. Unfortunately, this may not work in humans. Pockley: Irun Cohen’s group (Zanin-Zhorov et al 2006) has published work which indicates that Hsp60 can activate regulatory T cell populations via TLR2 signalling. Also, there is a recent paper from Eddy Liew’s group (Liu et al 2006) showing that if regulatory T cells are stimulated with an agonist, then the initial response is their expansion, but with a concomitant reduction in their anti-inflammatory effects. At a later time point, this expanded population then acquires the ability to inhibit. There is a timing issue: with an immune response you want it quickly, but you also want to have a system in place to switch it off at the appropriate point. We also need to look at the kinetics. This paper demonstrated that you can activate these cells without inducing anti-inflammatory effects, and then at a later stage they acquire the ability to switch off the immune response. These results were obtained using a synthetic TLR ligand.

82

TRIANTAFILOU ET AL

Hightower: I have a comment about sepsis. You mentioned that in addition to bacterial sepsis, this might also be applicable to viral sepsis. A colleague, George Perdrizet, is a trauma surgeon. Trauma surgery has become so good they save most of the patients, only to see a large percentage of them die in critical care due to a phenomenon that starts out as systemic inflammatory response syndrome (SIRS), and then often deteriorates into multiple organ system failure. He showed me a chart to determine whether a condition is SIRS or sepsis: the only difference is that with sepsis there is the suspected or overt presence of a pathogen. If this isn’t there, they say it is SIRS. It is described as a cytokine storm that takes over patients systemically and then begins destroying their organs and they die. It may be that this is innate immunity out of control. Something like Hsp70 administered to these patients might help them regain control by down-regulating the response. Currently there is no good treatment for SIRS. Panayi: A few years ago there was a clinical trial using the TNFR:Fc fusion protein in septic shock (Fisher et al 1996). At higher doses of construct patients died rather than being saved. One hypothesis was that the construct was binding the endotoxin and then releasing it later on. All attempts to treat Gram-negative sepsis, to my knowledge, have not been very successful. There is a late phase of endotoxaemic shock which is the lethal phase, and most of the mouse models are acute ones where a sensitive strain of mouse is given endotoxin and it dies. It may be that looking at endotoxin and the receptor system is just part of what happens in the clinical setting. Hightower: George Perdrizet has some interesting ideas on this. They go back to Hans Selye, the pioneering physiologist of generalized stress responses. He had a general stress response syndrome, and at the end of it was exhaustion. George thinks the end game is this exhaustion phase where a key organ becomes so dysfunctional that it fails. As you point out, this goes well beyond the initial endotoxaemia. Feder: I would like to play devil’s advocate. In 1997 Rudiger and Bukau published a paper (Rudiger et al 1997) on Hsp70’s binding sites, suggesting that Hsp70 was capable of binding every known protein approximately every 40 amino acids or so. One could posit that there is no specificity here. Hsp70, if it happens to make it outside the cell, could interact with any protein present. It could associate and simply interfere with the normal interactions of the client protein with its normal client. This sort of tallies with Radhey Gupta’s observations about the various drugs that no longer bind to tubulin. What might be happening is not Hsp70 being co-opted as an anti-inflammatory or an anticarcinogen, but it simply behaving as a generalized polyinteractive protein. Calderwood: In our search for HSP receptors we use a system in which we transfect expression vectors with suspected receptors into CHO cells which are null for Hsp70 binding. There is some specificity in the interaction. There are some

CHAPERONES AND INNATE IMMUNITY

83

cells which do not bind Hsp70 and these only bind when we transfect in the receptor. Hightower: There are rules. Some will remember the phage display libraries where it was possible to display 7-mers and 9-mers of a random selection of peptides. Mary Jane Gething did this for BiP (Blond-Elguindi et al 1993) and we did it for Hsc70 (Takenaka et al 1995). In fact, they don’t bind all the peptides but rather small groups of peptides with different affinities. Different chaperones bind different subgroups. They have rules. Feder: I like Stuart’s and John’s response better than that! There are rules, but the binding sites occur frequently in proteins. The fact they have rules doesn’t exclude interaction with most known proteins. Hightower: It does, because most proteins aren’t unfolded and displaying all their hydrophobic stretches and potentially Hsp70 binding sites. They may have a loop here or there, but if that loop doesn’t have one of these binding sites then your chaperone won’t touch it. Ellis: Isn’t there a much simpler argument? If what each of you suggest is true, there should be many proteins which, when purified, will allow Hsp70 to bind to them. Lund: How do we know that the proteins on the outside of the cell are fully folded? Hightower: There probably are some that have loops flopping around. It is a nasty environment for proteins outside of the cell, so I would guess there is some denaturation. Most proteins are going from states of partially unfolded to folded. The energy differences are not great. Miller-Graziano: On the SIRS versus CARS (compensatory anti-inflammatory response syndrome) argument, the current paradigm is that there isn’t a linear progression of A to B, i.e. SIRS to CARS. Instead, there is repeated cycling between these two post injury dysfunctional states. Patients cycle along going into immunosuppression and then into cytokine storms. There is a continual alteration of the inflammatory versus immunosuppressed state until they resolve or not resolve. The typical time of death for burn patients is about 10–12 days post-injury. After about day 4, we start seeing this cycling of immunosuppression alternating with hypo-inflammation, but then we see hyperinflammation too. It goes both ways. The idea is that there is complete dysregulation of the whole immune response because the appropriate connection between the innate and adaptive immune system is broken. I would also like to remind everyone that the most recent example of SIRS and multiple organ failure death was when the CD28 trials were done. Antibody to CD28 worked well to induce T regulator cells in mice, but when given to healthy young male volunteers, this treatment sent them all into the ICU with multiple organ failure. One was in there for a month and almost died.

84

TRIANTAFILOU ET AL

Panayi: They recovered, though. Miller-Graziano: Yes, because they were young and healthy. Panayi: The fact that they recovered is important. We must all get insults from time to time that stimulate part of the innate immune response and we dampen it down. The question is: why don’t we dampen it down in every circumstance? We heard earlier that as we age we don’t produce as many HSPs. It could be that if these are truly regulators of inflammation and you are older, you are more likely to have a little problem in regulating it. The cycling is telling me that you get a positive response during the process, negative regulators come out but don’t quite manage it, the positive response comes and in the end for all sorts of reasons, a proportion of the patients succumb. How else do you explain cycling? Miller-Graziano: It is an age-based effect. The volunteers in the CD28 trial were young men. In a group of trauma patients it is unusual for a young person to die. Most of the patients who succumb are over 30. Those in their 80s are most vulnerable. Lund: Isn’t this typical network behaviour? There are many examples in networks where if you change parameters they will oscillate between two stable states. Gupta: What was the source of your Hsp70 that you applied exogenously? Triantafilou: It was human recombinant Hsc70. Gupta: If this was a response specific to Hsc70, then it is likely that the mitochondrial Hsp70 will not produce that response. It could be an interesting control. Henderson: This raises an interesting point. The human Hsp70 and the Mycobacterium tuberculosis Hsp70 bind to different parts of CD40. There is a danger in saying that all Hsp70s are the same. References Akyol S, Gercel-Taylor C, Reynolds LC, Taylor DD 2006 HSP-10 in ovarian cancer: expression and suppression of T-cell signaling. Gynecol Oncol 101:481–486 Blond-Elguindi S, Cwirla SE, Dower WJ et al 1993 Affi nity panning of a library of peptides displayed on bacteriophages reveals the binding specificity of BiP. Cell 75:717–728 Fisher CJ Jr, Agosti JM, Opal SM et al 1996 Treatment of septic shock with the tumor necrosis factor receptor:Fc fusion protein. The Soluble TNF Receptor Sepsis Study Group. N Engl J Med 334:1697–1702 Liu H, Komai-Koma M, Xu D, Liew FY 2006 Toll-like receptor 2 signaling modulates the functions of CD4+ CD25+ regulatory T cells. Proc Natl Acad Sci USA 103:7048–7053 Norbury CC 2006 Drinking a lot is good for dendritic cells. Immunology 117:443–451 Rudiger S, Germeroth L, Schneider-Mergener J, Bukau B 1997 Substrate specificity of the DnaK chaperone determined by screening cellulose-bound peptide libraries. EMBO J 16:1501–1507 Takenaka IM, Leung SM, McAndrew SJ, Brown JP, Hightower LE 1995 Hsc70-binding peptides selected from a phage display peptide library that resemble organellar targeting sequences. J Biol Chem 270:19839–19844

CHAPERONES AND INNATE IMMUNITY

85

Vanags D, Williams B, Johnson B et al 2006 Therapeutic efficacy and safety of chaperonin 10 in patients with rheumatoid arthritis: a double-blind randomised trial. Lancet 368: 855–863 van Eden W 2007 Heat shock proteins are targets for T cell regulation: how microbial HSP induce IL-10 producing anti-inflammatory T cells. In: Asea A, De Maio A (eds) Heat shock proteins: potent mediators of inflammation and immunity, Vol. 1. Springer, Dordrecht, The Netherlands, p 289–308 Zanin-Zhorov A, Cahalon L, Tal G, Margalit R, Lider O, Cohen IR 2006 Heat shock protein 60 enhances CD4+CD25+ regulatory T cell function via innate TLR2 signaling. J Clin Invest 116:2022–2032

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Cell stress proteins in extracellular fluids: friend or foe? A. Graham Pockley*† and Gabriele Multhoff† *Immunobiolog y Research Unit, School of Medicine and Biomedical Sciences, University of Sheffield, L Floor, Royal Hallamshire Hospital, Sheffield S10 2JF, UK and †Klinikum rechts der Isar, Technische Universität München, Department of Radiotherapy and Radiooncolog y and GSF Institute of Patholog y, Munich, Germany

Abstract. For many years the perception has been that mammalian stress proteins are intracellular molecules that are only present in the extracellular environment as a consequence of pathological situations such as necrotic cell death. However, many investigators have now shown that these proteins can be released from a variety of viable (non-necrotic) cell types in vitro, by a mechanism which has yet to be fully established. Moreover, we and a number of others have reported Hsp60 and/or Hsp70 to be present in the peripheral circulation of normal individuals. These observations have profound implications for the perceived role of these proteins as universal pro-inflammatory intercellular ‘danger’ signalling molecules, and the functional significance and role(s) of these ubiquitously expressed and highly conserved families of molecules must therefore be critically re-evaluated. This paper reviews the evolving evidence which indicates that stress proteins such as Hsp60 and Hsp70 are present in, and can be released into the extracellular compartment under normal physiological conditions, and puts into context their pro- and anti-inflammatory potential. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 86–100

It was in 1989 that the seminal study reporting that heat shock/stress proteins could be released by cells was published by Hightower and Guidon (Hightower & Guidon 1989). In this study, heat treatment was shown to broaden the spectrum of proteins released from cultured rat embryo cells, from a small set of proteins including the constitutively-expressed member of the 70 kDa family of molecules, Hsc70, to include its inducible counterpart, Hsp70, and Hsp110. It was thought that the release of what had hitherto been regarded as being exclusively intracellular molecules might have resulted from changes in pH and gas tension, disruption of the diffusion layer at the cell surface or mechanical stresses associated with the in vitro manipulations that are an inevitable consequence of cell culture procedures (Hightower & Guidon 1989). The release of these proteins was shown not 86

EXTRACELLULAR STRESS PROTEINS

87

to be mediated via the common secretory pathway, as inhibitors of this pathway, colchicine and monensin, did not block it (Hightower & Guidon 1989). Exposure of cells to low concentrations of non-ionic detergents indicated that Hsp70 was not readily released from damaged cells. The observation that Hsp70 which had been synthesised in the presence of the lysine amino acid analogue aminoethyl cysteine was not released from cells, suggested that an altered structure or function prevented its correct interaction with a specific release mechanism (Hightower & Guidon 1989). Another of the earliest papers documenting an apparent physiological release of stress proteins came from Tytell and colleagues in 1986. In this, the authors reported the transfer of glia-axon transfer proteins, which include Hsp70, Hsc70 and Hsp100, from adjacent glial cells into the squid giant axon (Tytell et al 1986). This finding prompted the suggestion that the release of such proteins might be a mechanism by which glial cells, which are capable of generating effective stress protein-mediated resistance to physical and metabolic insults, can protect adjacent neuronal cells which exhibit a deficient response to stress. Stress proteins have now been reported to be released from a wide range of cells such as human islet cells, cultured rat cortical astrocytes, a human neuroblastoma cell line, a human keratinocyte-derived cell line, cultured vascular smooth muscle cells, murine and human prostate cancer cells and human peripheral blood mononuclear cells, to highlight but a few (Frostegård & Pockley 2005). The concept that stress proteins can be released in the absence of cellular necrosis continues to be viewed with scepticism by the wider scientific community. One of the prime conceptual difficulties with accepting such a concept is that stress proteins do not express the typical N-terminal signal peptide sequences that are typically required for secretion. However, ‘non-classical’ secretion of proteins that lack such sequences has been observed for a number of proteins such as fibroblast growth factors 1 and 2, interleukin (IL)-1 as well as viral proteins. The mechanisms that are involved in such a process have been reviewed in detail elsewhere (Chimini & Rubartelli 2005). The paucity of information concerning the mechanism(s) should not therefore necessarily cast doubt on their capacity to be released/secreted by cells under normal physiological conditions. Little is known about protein secretion and it is only within the last decade or so that a number of extremely important and novel pathways of protein secretion in bacteria have been identified. Indeed, the secretion pathways by which key mediators such as IL-1 are released from eukaryotic cells are still not fully elucidated. Wider acceptance that there is a physiological role for stress proteins will to some extent depend on the elucidation of pathways via which release/secretion is achieved. Work from a number of laboratories is now providing insight into such mechanisms. The Multhoff laboratory has demonstrated that the spontaneous release of soluble Hsp70 by viable human colon and pancreatic carcinoma cells

88

POCKLEY & MULTHOFF

tumour sublines differentially expressing Hsp70 on their cell surface is low, whereas detergent-soluble vesicles actively released by tumours contain high amounts of Hsp70/Bag-4 and Hsp70/Hsp40 (Gastpar et al 2005). The biochemical and biophysical properties of these vesicles (density, acetylcholine esterase activity, protein composition) indicate that they are exosomes (Gastpar et al 2005). Exosomes are small membrane vesicles that form within late endocytic compartments (multivesicular bodies) and they are distinct from apoptotic vesicles, in that their mode of production and protein composition differ (Théry et al 2001). An increased exosomal export of Hsp70 from tumour cells after treatment with interferon- γ has also been reported by Asea’s group (Bausero et al 2005), and Hsp70-containing exosomes have been shown to be released by peripheral blood mononuclear cells (Lancaster & Febbraio 2005). Although cytosolic stress proteins do not contain the leader peptide sequences that typically enable membrane localization, one of their major tasks is the transport of other proteins across lipid membranes. It is therefore conceivable that cytosolic stress proteins are transported to the plasma membrane in concert with proteins that do possess transmembrane domains with such a capacity to shuttle molecules. It might therefore be that lipid rafts, which are specialised membrane domains enriched in sphingolipids, cholesterol and proteins, are involved in the localization of Hsp70 to the cell surface and its secretion into the extracellular environment (Broquet et al 2003). Many functions have been attributed to lipid rafts, including cholesterol transport, membrane sorting, endocytosis and signal transduction, and they can be isolated as detergent-resistant microdomains (DRMs). In unstressed Caco-2 human colonic adenocarcinoma epithelial cells, stress proteins (especially Hsp70) are present in a major Triton X-100 soluble form and a minor detergent insoluble form that is associated with DRMs. The translocation of Hsp70 into DRMs can be enhanced by heat shock or by increasing intracellular Ca2+ levels (Broquet et al 2003). Although the incorporation of Hsp70 into the DRMs and the release of Hsp70 from Caco-2 cannot be inhibited by blockade of the common secretory pathway using brefeldin A or monensin, Hsp70 release has been shown to be blocked by disrupting lipid rafts using methyl- β -cyclodextrin (Broquet et al 2003). The ability of extracellular Hsp70 and Hsc70 to selectively interact with phosphatidylserine (PS) moieties on the surface of cells suggests that these proteins might act as accelerators of apoptosis, as cell surface expressed PS is an early indicator of this process and the addition of these proteins to cells expressing PS leads to their incorporation into the lipid bilayer and a loss in their viability (Arispe et al 2002). The release of stress proteins might also involve their entry into secretary lysosomal endosomes and their cell surface and release of the contents of the endolysosome into the extracellular space. Certainly, febrile conditions induce the entry of

EXTRACELLULAR STRESS PROTEINS

89

Hsp70 into endolysosomes and their secretion in an extracellular ATP-dependent manner. The release of Hsp70 from tumour cells and macrophages has been shown to be mediated via such a route (Mambula & Calderwood 2006). One study showing that a pharmacological inhibitor of phospholipase C activity (U731222) induces the release of Hsp70 from the A431 human carcinoma cell line suggests that phospholipase C inhibition might also be one of the mechanisms by which stress proteins are released (Evdonin et al 2004). The release of Hsp70 induced by inhibiting phospholipase C activity might occur via vesicular transport, as the inhibition of vesicular transport with brefeldin A prevents it (Evdonin et al 2006). Another observation has been that a large proportion of the Hsp70 released by A431 human carcinoma cells is in a ubiquitinated form (Evdonin et al 2004). In addition to signalling for proteosome-dependent degradation, ubiquitination has been shown to serve as a trigger for different transport events (Katzmann et al 2002). But what is the physiological role for extracellular stress proteins? As indicated above, the very early studies in which the release of stress proteins was first reported demonstrated that glia-axon transfer proteins, which include Hsp70, Hsc70 and Hsp100, are transferred from adjacent glial cells to the squid giant axon (Tytell et al 1986). These findings and those demonstrating that exogenous Hsp70 can enhance the tolerance of neuronal cells to stress (Guzhova et al 2001) suggest that the release of such proteins might be an altruistic response on the part of one cell for the protection of adjacent cells. Exogenous Hsp70 has also been shown to protect heat-stressed cynomolgus macaque aortic cells (Johnson et al 1990) and serum-deprived rabbit arterial smooth muscle cells (Johnson & Tytell 1993) from cytotoxic damage, the latter by a mechanism which involves cell association, but not internalization. The mechanism by which such protection is afforded remains unclear, and the cell surface receptors involved have not been identified. However, the protective capacity of extracellular Hsp70 might result from the fact that exogenous Hsp70 increases intracellular Hsp70 levels, which in turn could delay the decline in viability of stressed arterial cells (Berberian et al 1990). Extracellular Hsp70 has also been shown to have a number of neuroprotective properties and the capacity to inhibit motor and sensory neuron degeneration (Tidwell et al 2004, Robinson et al 2005, Tytell 2005). It is also interesting to note that the bacterial 60 kDa stress protein GroEL induces the release of endogenous Hsp60 and Hsp70 from cultured human peripheral blood mononuclear cells (Davies et al 2006). These findings prompt questions regarding the physiological role of extracellular stress proteins. Much attention has also been focused on the immunological properties of stress proteins, to which have been attributed both pro-inflammatory (e.g. members of the 60 kDa, 70 kDa and 90 kDa families) and anti-inflammatory (e.g. Hsp10, Hsp27, grp78/ BiP) properties.

90

POCKLEY & MULTHOFF

The anti-inflammatory properties of Hsp27 and BiP are discussed elsewhere in this volume. Attention herein will therefore primarily be focused on the apparently counterintuitive observation that Hsp60 and Hsp70 are typically regarded as being potent pro-inflammatory molecules, yet they have been found to be present in the peripheral circulation of normal individuals at levels that are similar to those which have been shown to potently activate innate immune cells in vitro (Pockley et al 1998, 1999, Xu et al 2000, Lewthwaite et al 2002). It certainly appears that the circulating levels of such proteins provide some insight into ongoing physiological processes, as elevated levels of circulating Hsp60 are associated with the presence of early atherosclerosis in humans (Pockley et al 2000, Xu et al 2000), whereas elevated levels of Hsp70 are associated with a reduced risk of developing cardiovascular disease, at least in subjects with established hypertension (Pockley et al 2003). Such counterintuitive fi ndings stress the importance of considering the immunobiological properties of extracellular stress proteins not in isolation, but in the context of the systems biology of the organism. For example, although there is a wealth of evidence to indicate that stress proteins such as Hsp60, Hsp70 and gp96 are potent activators of innate immune cells such as monocytes/macrophages and dendritic cells (DCs), these proteins have also been shown to exhibit anti-inflammatory properties in a number of different experimental model systems and clinical conditions such as autoimmune disease (Todryk et al 2003, Chandawarkar et al 2004, Quintana & Cohen 2005, van Eden et al 2005, Slack et al 2007). It has been known for some considerable time that the induction of T cell reactivity to self-Hsp60 and Hsp70 epitopes down-regulates inflammatory disease in experimental models by mechanisms which involve the generation of Th2 cells producing the regulatory cytokine IL-10 (Kingston et al 1996, Tanaka et al 1999, Wendling et al 2000, van Eden et al 2005). Although the precise mechanism via which exogenous stress proteins such as Hsp60, Hsp70 and gp96 might influence inflammatory events has yet to be elucidated, some insight might be provided by the observation that naturally-occurring CD4 + CD25 + immunoregulatory T cell populations express Toll-like receptors (TLRs) (Caramalho et al 2003), for which stress proteins are ligands (Vabulas & Wagner 2005). Triggering of TLRs on CD4 + CD25 + T cells by lipopolysaccharide (LPS) has been shown to induce a 10-fold increase in their suppressive activity (Caramalho et al 2003) and it might therefore be that stress proteins such as Hsp60 and Hsp70 also influence the presence and/or functional activities of such cells. Certainly it has been shown that human Hsp60 can activate CD4 + CD25 + regulatory T cell populations via an interaction with, and signalling via, TLR2 (ZaninZhorov et al 2006). This immunoregulatory activity appears to involve cell–cell contact and the secretion of the immunoregulatory cytokines IL-10 and transforming growth factor (TGF) β (Zanin-Zhorov et al 2006).

EXTRACELLULAR STRESS PROTEINS

91

Another study has shown that although a synthetic bacterial lipoprotein, Pam3Cys-SK4 (a TLR2 agonist) can promote the expansion of CD4 + CD25 + regulatory T cells. Following polyclonal stimulation in vitro, there is a transient loss of suppressive activity which is further compounded by the concomitant induction of IL-2 secretion from CD4 + CD25− effector cells (Liu et al 2006). Interestingly, Pam3Cys-SK4 treated CD4 + CD25 + regulatory T cells regain their suppressive activity following a period in culture (Liu et al 2006). As important to the induction of an inflammatory response is its timely resolution. These findings suggest that TLR ligands can both induce and contribute to the resolution of inflammatory responses, the former by inducing the expansion of CD4 + CD25− effector cells via the induction of IL-2 and the attenuation of CD4 + CD25 + regulatory T cell function, and the latter by the release of inhibitory effects on what has become an expanded population of CD4 + CD25 + regulatory T cells (Liu et al 2006). These are key findings and further highlight the complexities of endogenous immunoregulatory networks. Although the capacity of Hsp70 to influence the activity of CD4 + CD25 + regulatory T cells has yet to be demonstrated, it is possible that Hsp70 might also influence the immunoregulatory effects of such naturally occurring regulatory T cells. These studies are currently underway in the authors’ laboratories. Further insight into the immunoregulatory properties of Hsp60 and their underlying mechanisms are provided by Francisco Quintana and Irun Cohen elsewhere in this volume. It is important to keep in mind that extracellular Hsp70 is also present as a membrane-bound form on a large proportion of different tumour cells, but not on their normal tissue and cellular counterparts (Radons & Multhoff 2005). Membrane-bound Hsp70 acts as a tumour-specific recognition structure for NK cells, and mediates NK cell-mediated lysis via granzyme B-mediated, perforin-independent apoptosis. This involves the uptake of active granzyme B via a pathway which is facilitated by membrane-bound Hsp70 (Gross et al 2003). This discovery adds an additional dimension to the concept of an intercellular signalling role for extracellular stress proteins and this finding is being used to underpin the development of innovative approaches for the treatment of patients with membrane Hsp70positive leukaemias and solid tumours. Conclusions Although the concept that stress proteins such as Hsp60 and Hsp70 can be released from viable cells under normal physiological conditions has yet to receive universal acceptance, there is much evidence to support such a concept. Molecular chaperones are known to have a number of cytoprotective properties when in the intracellular compartment and it might be that stress proteins with chaperone function have a similar role when in the extracellular compartment. Indeed, the wealth of

92

POCKLEY & MULTHOFF

evidence which indicates that these proteins can protect against neural cell damage suggests that they might be used for the development of neuroprotective therapeutic strategies (Robinson et al 2005, Tytell 2005). To date, work on the function of extracellular stress proteins has primarily focused on their immunobiology, particularly their pro-inflammatory properties. In this respect there is a wealth of data to indicate that these proteins are indeed potent inducers of inflammatory events, and these aspects of stress protein immunobiology are reviewed by Alexzander Asea elsewhere in this volume. The capacity of these proteins to activate immune responses, to chaperone antigenic peptides and to deliver these into antigen-presentation pathways within antigen presenting cells (APCs) has prompted consideration and studies into their potential use as immunological adjuvants and immunotherapeutics for cancer and infectious disease. The release of Hsp70 from tumour cells might serve to induce protective antitumour immunity and the immunotherapeutic potential of stress proteins is reviewed by Stuart Calderwood elsewhere in this volume. However, although these molecules are archetypal pro-inflammatory ‘danger’ signals, they can also act as anti-inflammatory molecules with the capacity to attenuate inflammatory disease. The evidence for both of these capacities is strong and it is therefore necessary to rationalise the evidence into a model system which can encompass these contrasting functions. Although such a rationalization has been difficult, recent evidence that TLRs are expressed on naturally-occurring regulatory T cell populations and that their triggering can activate such regulatory properties have provided some explanation. The elegant experiments from Liu et al (2006), which have demonstrated that ligation of a TLR on CD4 + CD25 + regulatory T cells can both promote and resolve inflammation, further helps to explain the potential physiological basis to the pro- and antiinflammatory properties of these proteins, and gives some insight into the likely complexities of models that attempt to explain the immunobiology of stress proteins. Evidence that molecules such as Hsp27 and BiP can elicit anti-inflammatory consequences by directly influencing the function and phenotype of APCs and T cells is presented elsewhere by Carol Miller-Graziano and Gabriel Panayi respectively. The concept that prokaryotic and eukaryotic 60 kDa stress proteins can elicit anti-inflammatory effects by blocking, in a dose-dependent manner, LPS-induced macrophage activation, which is introduced by Anthony Coates, adds another potential mechanistic insight into the anti-inflammatory properties of these molecules. Work in the Pockley laboratory has demonstrated that although the stress protein gp96 does not bind to a population of rat bone marrow-derived DCs via cell surface receptors, it is internalised by pinocytosis and influences the morphology of these cells (Mirza et al 2006). This is an important observation as it is known that DCs can subsequently present antigenic peptides that are derived from

EXTRACELLULAR STRESS PROTEINS

93

pinocytosed exogenous material to T cells (reviewed in Norbury 2006). These findings suggest that antigenic peptides derived from pinocytosed gp96 might be presented to appropriate T cell populations by immature DCs in the absence of essential co-stimulatory signals, as the same laboratory has shown that gp96 has no effect on the maturational status of rat bone marrow-derived DCs (BMDCs) (Mirza et al 2006). This could result in the generation/recruitment of a population of immunoregulatory T cells having the capacity to influence and control inflammatory responses. These findings provide some insight into the anti-inflammatory properties of gp96 and its capacity to induce a state of peripheral T cell hyporesponsiveness (Chandawarkar et al 2004, Slack et al 2007). Further studies aimed at testing this and associated hypotheses regarding the anti-inflammatory effects of gp96 are underway. In summary, further insight into the functionality and significance of actively released, exogenous and circulating stress proteins might reveal hitherto unknown physiological, pathophysiological and immunoregulatory roles for these ubiquitously expressed and highly conserved families of molecules. From an immunological perspective, the apparently opposing functional properties of these molecules must be rationalised into a model in which both of these functional capacities are encompassed. This rationalisation and understanding will require a wider appreciation that of equal importance to the induction of inflammatory responses is their effective resolution. It is not too conceptually difficult to envisage a model in which the initial inflammatory response to an infective episode is induced, in part at least, by the effects of certain pathogen-derived stress proteins on innate immune cells and that at later stages, either pathogen-derived or hostderived stress proteins elicit and/or promote anti-inflammatory responses via the activation of immunoregulatory cell populations that have been recruited to the inflammatory site. Further studies aimed at better understanding the role(s) for extracellular stress proteins are undoubtedly required. References Arispe N, Doh M, De Maio A 2002 Lipid interaction differentiates the constitutive and stressinduced heat shock proteins Hsc70 and Hsp70. Cell Stress Chaperones 7:330–338 Bausero MA, Gastpar R, Multhoff G, Asea A 2005 Alternative mechanism by which IFNgamma enhances tumor recognition: active release of heat shock protein 72. J Immunol 175:2900–2912 Berberian P, Johnson A, Bond M 1990 Exogenous 70kD heat shock protein increases survival of normal and atheromatous arterial cells. FASEB J 4:A1031 Broquet AH, Thomas G, Masliah J, Trugnan G, Bachelet M 2003 Expression of the molecular chaperone Hsp70 in detergent-resistant microdomains correlates with its membrane delivery and release. J Biol Chem 278:21601–21606 Caramalho I, Lopes-Carvalho T, Ostler D, Zelenay S, Haury M, Demengeot J 2003 Regulatory T cells selectively express toll-like receptors and are activated by lipopolysaccharide. J Exp Med 197:403–411

94

POCKLEY & MULTHOFF

Chandawarkar RY, Wagh MS, Kovalchin JT, Srivastava P 2004 Immune modulation with highdose heat-shock protein gp96: therapy of murine autoimmune diabetes and encephalomyelitis. Int Immunol 16:615–624 Chimini C, Rubartelli A 2005 Novel pathways of protein secretion. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, New York, p 45–60 Davies EL, Bacelar MM, Marshall MJ et al 2006 Heat shock proteins form part of a danger signal cascade in response to lipopolysaccharide and GroEL. Clin Exp Immunol 145:183– 189 Evdonin AL, Guzhova IV, Margulis BA, Medvedeva ND 2004 Phospholipase C inhibitor, U73122, stimulates release of hsp-70 stress protein from A431 human carcinoma cells. Cancer Cell Int 4:2 Evdonin AL, Martynova MG, Bystrova OA, Guzhova IV, Margulis BA, Medvedeva ND 2006 The release of Hsp70 from A431 carcinoma cells is mediated by secretory-like granules. Eur J Cell Biol 85:443–455 Frostegård J, Pockley AG 2005 Heat shock protein release and naturally-occurring, exogenous heat shock proteins. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, New York, p 195–219 Gastpar R, Gehrmann M, Bausero MA et al 2005 Heat shock protein 70 surface-positive tumor exosomes stimulate migratory and cytolytic activity of natural killer cells. Cancer Res 65:5238–5247 Gross C, Koelch W, DeMaio A, Arispe N, Multhoff G 2003 Cell surface-bound heat shock protein 70 (Hsp70) mediates perforin-independent apoptosis by specific binding and uptake of granzyme B. J Biol Chem 278:41173–41181 Guzhova I, Kislyakova K, Moskaliova O et al 2001 In vitro studies show that Hsp70 can be released by glia and that exogenous Hsp70 can enhance neuronal stress tolerance. Brain Res 914:66–73 Hightower LE, Guidon PT 1989 Selective release from cultured mammalian cells of heat-shock (stress) proteins that resemble glia-axon transfer proteins. J Cell Physiol 138:257–266 Johnson AD, Tytell M 1993 Exogenous Hsp70 becomes cell associated, but not internalised by stressed arterial smooth muscle cells. In Vitro Cell Dev Biol 29A:807–812 Johnson AD, Berberian PA, Bond MG 1990 Effect of heat shock proteins on survival of isolated aortic cells from normal and atherosclerotic cynomolgus macaques. Atherosclerosis 84:111–119 Katzmann DJ, Odorizzi G, Emr SD 2002 Receptor downregulation and multivesicular-body sorting. Nat Rev Mol Cell Biol 3:893–905 Kingston AE, Hicks CA, Colston MJ, Billingham MEJ 1996 A 71-kD heat shock protein (hsp) from Mycobacterium tuberculosis has modulatory effects on experimental rat arthritis. Clin Exp Immunol 103:77–82 Lancaster GI, Febbraio MA 2005 Exosome-dependent trafficking of HSP70: a novel secretory pathway for cellular stress proteins. J Biol Chem 280:23349–23355 Lewthwaite J, Owen N, Coates A, Henderson B, Steptoe A 2002 Circulating human heat shock protein 60 in the plasma of British civil servants. Circulation 106:196–201 Liu H, Komai-Koma M, Xu D, Liew FY 2006 Toll-like receptor 2 signaling modulates the functions of CD4 + CD25 + regulatory T cells. Proc Natl Acad Sci USA 103:7048– 7053 Mambula SS, Calderwood SK 2006 Heat shock protein 70 is secreted from tumor cells by a nonclassical pathway involving lysosomal endosomes. J Immunol 177:7849–7857 Mirza S, Muthana M, Fairburn B, Slack LK, Hopkinson K, Pockley AG 2006 The stress protein gp96 is not an activator of resting rat bone marrow-derived dendritic cells, but is a costimulator and activator of CD3 + T cells. Cell Stress Chaperones 11:364–378

EXTRACELLULAR STRESS PROTEINS

95

Norbury CC 2006 Drinking a lot is good for dendritic cells. Immunology 117:443–451 Pockley AG, Shepherd J, Corton J 1998 Detection of heat shock protein 70 (Hsp70) and anti-Hsp70 antibodies in the serum of normal individuals. Immunol Invest 27:367– 377 Pockley AG, Bulmer J, Hanks BM, Wright BH 1999 Identification of human heat shock protein 60 (Hsp60) and anti-Hsp60 antibodies in the peripheral circulation of normal individuals. Cell Stress Chaperones 4:29–35 Pockley AG, Wu R, Lemne C, Kiessling R, de Faire U, Frostegård J 2000 Circulating heat shock protein 60 is associated with early cardiovascular disease. Hypertension 36:303–307 Pockley AG, Georgiades A, Thulin T, de Faire U, Frostegård J 2003 Serum heat shock protein 70 levels predict the development of atherosclerosis in subjects with established hypertension. Hypertension 42:235–238 Quintana FJ, Cohen IR 2005 Heat shock proteins regulate inflammation by both molecular and network cross-reactivity. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, New York, p 263–287 Radons J, Multhoff G 2005 Immunostimulatory functions of membrane-bound and exported heat shock protein 70. Exerc Immunol Rev 11:17–33 Robinson MB, Tidwell JL, Gould T et al 2005 Extracellular heat shock protein 70: a critical component for motoneuron survival. J Neurosci 25:9735–9745 Slack LK, Muthana M, Hopkinson K et al 2007 Administration of the stress protein gp96 prolongs rat cardiac allograft survival, modifies rejection-associated inflammatory events and induces a state of peripheral T cell hyporesponsiveness. Cell Stress Chaperones 12:71–82 Tanaka S, Kimura Y, Mitani A et al 1999 Activation of T cells recognizing an epitope of heat-shock protein 70 can protect against rat adjuvant arthritis. J Immunol 163:5560– 5565 Théry C, Boussac M, Véron P et al 2001 Proteomic analysis of dendritic cell-derived exosomes: a secreted subcellular compartment distinct from apoptotic vesicles. J Immunol 166:7309– 7318 Tidwell JL, Houenou LJ, Tytell M 2004 Administration of Hsp70 in vivo inhibits motor and sensory neuron degeneration. Cell Stress Chaperones 9:88–98 Todryk SM, Gough MJ, Pockley AG 2003 Facets of heat shock protein 70 show immunotherapeutic potential. Immunology 110:1–9 Tytell M 2005 Release of heat shock proteins (Hsps) and the effects of extracellular Hsps on neural cells and tissues. Int J Hyperthermia 21:445–455 Tytell M, Greenberg SG, Lasek RJ 1986 Heat shock-like protein is transferred from glia to axon. Brain Res 363:161–164 Vabulas RM, Wagner H 2005 Toll-like receptor-dependent activation of antigen presenting cells by Hsp60, Hsp70 and gp96. In: Henderson B, Pockley AG (eds), Molecular chaperones and cell signalling. Cambridge University Press, New York, p 113–132 van Eden W, van der Zee R, Prakken B 2005 Heat shock proteins induce T-cell regulation of chronic inflammation. Nat Immunol 5:318–330 Wendling U, Paul L, van der Zee R, Prakken B, Singh M, van Eden W 2000 A conserved mycobacterial heat shock protein (hsp) 70 sequence prevents adjuvant arthritis upon nasal administration and induces IL-10-producing T cells that cross-react with the mammalian self-hsp70 homologue. J Immunol 164:2711–2717 Xu Q, Schett G, Perschinka H et al 2000 Serum soluble heat shock protein 60 is elevated in subjects with atherosclerosis in a general population. Circulation 102:14–20 Zanin-Zhorov A, Cahalon L, Tal G, Margalit R, Lider O, Cohen IR 2006 Heat shock protein 60 enhances CD4 + CD25 + regulatory T cell function via innate TLR2 signaling. J Clin Invest 116:2022–2032

96

POCKLEY & MULTHOFF

DISCUSSION Henderson: I’d like to ask Peter Lund a question. Why do you think it is necessary to have more than one copy of Hsp60? As an example, Drosophila has four Hsp60 genes. People have looked at individual proteins, and there seem to be different functions, and functions over and above protein folding. Lund: You can ask the same question of Hsp70. Even yeast has multiple Hsp70s. We would like to believe that this has evolutionary significance, and that they have diverged to fold different subclasses of proteins, but there aren’t convincing data to support this. It is a complicated field and there aren’t many people looking at this specifically. The proteins usually have broadly similar functions. They can replace each other to a degree in complementation assays. It doesn’t look as though there has been a huge amount of functional divergence. Having said this, there is a classic case in Mycobacterium smegmatis where one of the GroEL proteins appears to be important for mycolic acid synthesis in the cell wall. One would like to believe that this is because it has some specificity to hold the particular substrate which is the FAS2 complex. I don’t think there is a simple answer. It is interesting that the organisms that have multiple GroEL genes tend to cluster. Feder: In Drosophila five or six genes are in two different clusters. This is significant in two ways. First, quantitative: the more gene copies the more protein. Second, they are distinctively regulated. Lakhotia & Prasanth (2002) examined the domains of expression of cluster 1 and cluster 2. They show very different cell type specificity, kinetics and organ specificity. Apparently, this provides the organism with the ability to fine-tune this. In an evolutionary sense, the notion is that whenever there is gene duplication, the second copy is then free to take on new functions. Gupta: The bacteria that have multiple GroEL genes have more complex life cycles. In the case of rhizobiae, they can grow either as free-living bacteria or in the root nodules of legume plants where they are involved in nitrogen fixation. In each condition they may require a different kind of chaperone that is specialized for this function. Similarly, both streptomyces and mycobacteria, which contain multiple GroEL homologues, have complex life cycles. Hightower: In the human Hsp70s, Hsp70B’ is the true stress Hsp70. It has no detectable basal levels in virtually all cells, and is strictly stress inducible. Ironically, it is one of the most infrequently studied of the human Hsp70s and is regulated differently from the others. My last student, Emily Noonan, recently demonstrated that it is present on the surface of human colonocytes (Noonan et al 2008). Macario: Have nucleotide exchange factors, such as the BAG type of factors (Macario & Conway de Macario 2001, 2005), and some of the other co-chaperones such as Hsp40, been found in the extracellular space?

EXTRACELLULAR STRESS PROTEINS

97

Multhoff: We published a paper that showed that the membrane Hsp70 is actually associated with Hsp40 in the membrane (Gehrmann et al 2005). It is always present as a complex. Macario: If those co-factors are found, one may begin to infer some functions for the extracellular proteins, such as folding or refolding of polypeptides in a manner similar to that occurring inside cells (Macario & Conway de Macario 2007). Multhoff: We should distinguish not only among the different species, but also between the constitutive and inducible forms. There is strong evidence that although the peptide binding cleft might be similar, they might bind distinct repertoires of peptides. An example is the repertoire of peptides that are bound to the constitutive and the inducible forms of Hsp70. This might also explain the different functions. Corrigall: When you were measuring Hsp70 and Hsp60 in the serum, did you look at all forms of the HSP or just the constitutively expressed or inducible ones? Pockley: In our Hsp70 assay we can only go by the specificity that the supplier of the monoclonal antibody indicates. The antibody we used is specific for the inducible form and does not react with the constitutive form. The way the assay was developed is that there is a monoclonal coating antibody, and this is the one that is specific and doesn’t react with Hsc70. Bound protein is detected using a polyclonal antibody. Corrigall: Did you look in the same patient at Hsp60 and Hsp70 levels? If you have high Hsp60 do you also have high Hsp70? Pockley: We analyze all of our samples for both Hsp60 and Hsp70 and yes, there is a positive relationship. There is a big difference in the data in the literature by individuals who are using a commercial kit, because there are commercial ELISA kits available for Hsp60 and 70. The commercial kit for Hsp70 and our own assay show a large divergence in measured levels. The in-house assays from different labs are pretty consistent. Csermely: You mentioned the altruistic response, with one cell releasing an HSP to benefit another cell. Is this well established? Pockley: This came from the early work using neuronal cells which originated from Larry Hightower and Mike Tytell and it is a hypothesis which was originally proposed by Larry. Everyone is focused on the immunology of these proteins, however I think there are other aspects. I am not sure why the early Johnson work (Johnson et al 1990) has not been followed up on because it is quite an important observation, especially in the context of Mike Tytell’s and Boris Margulis’ work, which has clearly demonstrated that extracellular Hsp70 has cytoprotective effects.

98

POCKLEY & MULTHOFF

Csermely: During the active process some of these proteins become ubiquitinated. How general is this? You may find differences through post-translational modifications in proteins secreted by an active process compared with those just getting out without any active process. Pockley: This was related to Boris’ work (Evdonin et al 2004). The work relating to the potential release via exosomes has come since that date. We now have more information regarding potential mechanisms. It was a proposition from Boris that this might be one of the mechanisms. Nixon: Yesterday we heard a lot of information about how the levels of Hsp70 vary with age. In one of your papers you have identified differences in extracellular chaperone expression between males and females, particularly in circulating Hsp60, suggesting it is 3.5-fold higher in females. Does this mean females live more stressful lives? Pockley: In our first paper we demonstrated a marked difference. Since that time we have not been able to demonstrate such a marked difference between males and females. There is no convincing difference except for the fi rst paper from the Whitehall study, which demonstrated that Hsp60 levels were related to psychosocial stress, but only in women. Henderson: We have done a larger study and it involved both males and females, showing a relationship between psychological stress and levels of Hsp60 (ShamaeiTousi et al 2008). Pockley: The other issue with these measurements is that the levels differ a lot among individuals so you need to analyse a large number of samples in order to identify any distinct trends or differences. Cheetham: In your ELISA, would client protein binding or substrate binding block the antibody accessibility? When we tried to do ELISA-based studies with client protein binding, we found that this blocked accessibility for a lot of the antibodies that bind the C-terminus in general. If Hsp70 is occupied or free it could create an apparent difference in levels. Pockley: This is one of the problems with the ELISA approach. We haven’t done those experiments, but we have shown that in the peripheral circulation there are circulating antibodies to these stress proteins. This was always a big issue when we were trying to work on the ELISA. We can certainly identify immune complexes in the circulation. If you coat an ELISA plate with an antibody to Hsp70 or 60 and then detect binding with an antibody to immunoglobulin, you can get a signal. There is a complex relationship between protein and antibody. Miller-Graziano: What was your source of dendritic cells (DCs)? I noticed you saw no binding of gp96. Pockley: These were rat bone marrow-derived DCs generated using recombinant rat IL-4 and GM-CSF. Miller-Graziano: Were the regulatory T cells (Tregs) also rat?

EXTRACELLULAR STRESS PROTEINS

99

Pockley: The Treg work I cited is not ours, but it is in human. We are currently trying to set up a system in human at the moment. Conway de Macario: When you talk about high levels of Hsp70, what would you consider to be a normal level? Pockley: If you look in normal individuals there are 30% where you can’t detect Hsp70. Levels will range from barely detectable to tens of microgrammes per ml. The latter is a very high level for a protein. The problem with identifying a ‘normal’ level is the fact that the levels are so variable. Panayi: This business about the varying concentration of stress proteins in healthy individuals raises the question as to whether the level is under genetic control. Have you looked within families? It is an important topic and it would advance the question of the role of stress proteins in the circulation if we could show that they were genetically regulated. Pockley: We haven’t. Brian and Tony have generated some nice longitudinal data looking at the stability of levels. Henderson: I’ll mention some data in my paper on a polymorphism in the Hsp60 promoter that we have identified. Lund: In the cases in which you can see an effect of adding exogenous Hsp70 to a culture, is de novo protein synthesis or gene expression needed for the effect? Pockley: Some of the early work relating to the aortic protection suggested that the endogenous Hsp70 induced the expression of Hsp70 (Johnson et al 1990). One of the authors’ propositions was that it involved an induction of the intracellular levels of Hsp70, and that this elicited the protective effect. I am not aware that this study has been followed up. Lund: Do you need intact Hsp60 protein, or can you put in a hydrolysed protein and still get an effect? Cheetham: If Hsp70 is heat inactivated, it loses its cytoprotective activity (Guzhova et al 2001, Novoselova et al 2005).

References Evdonin AL, Guzhova IV, Margulis BA, Medvedeva ND 2004 Phospholipase C inhibitor, U73122, stimulates release of hsp-70 stress protein from A431 human carcinoma cells. Cancer Cell Int 4:2 Gehrmann M, Marienhagen J, Eichholtz-Wirth H et al 2005 Dual function of membranebound Hsp70, Bag-4, and Hsp40: protection against radiation-induced effects and target structure for NK cells. Cell Death Differ 12:38–51 Guzhova I, Kislyakova K, Moskaliova O et al 2001 In vitro studies show that Hsp70 can be released by glia and that exogenous Hsp70 can enhance neuronal stress tolerance. Brain Res 914:66–73 Johnson AD, Berberian PA, Bond MG 1990 Effect of heat shock proteins on survival of isolated aortic cells from normal and atherosclerotic cynomolgus macaques. Atherosclerosis 84:111–119

100

POCKLEY & MULTHOFF

Lakhotia SC, Prasanth KV 2002 Tissue- and development-specific induction and turnover of hsp70 transcripts from loci 87A and 87C after heat shock and during recovery in Drosophila melanogaster. J Exp Biol 205:345–358 Macario AJL, Conway de Macario E 2001 The molecular chaperone system and other antistress mechanisms in archaea. Front Biosci 6:d262–283 Macario AJL, Conway de Macario E 2005 Sick chaperones, cellular stress and disease. New Engl J Med 353:1489–1501 Macario AJL, Conway de Macario E 2007 Molecular chaperones: multiple functions, pathologies, and potential applications. Front Biosci 12:2588–2600 Noonan EJ, Fournier G, Hightower LE 2008 Surface expression of Hsp70B’ in response to proteasome inhibition in human colon cells. Cell Stress Chaperones, in press Novoselova TV, Margulis BA, Novoselov SS et al 2005 Treatment with extracellular HSP70/ HSC70 protein can reduce polyglutamine toxicity and aggregation. J Neurochem 94:597– 606 Shamaei-Tousi A, Steptoe A, O’Donnell K et al 2008 Plasma heat shock protein 60 and cardiovascular disease risk: the role of psychological, genetic, and biological factors. Cell Stress Chaperones, in press

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

HSP60 speaks to the immune system in many voices Francisco J. Quintana and Irun R. Cohen* Center for Neurologic Diseases, Harvard Medical School, Boston, MA 02115, USA and *Department of Immunolog y, The Weizmann Institute of Science, Rehovot 76100, Israel

Abstract. Heat shock proteins (HSP) were initially identified as a family of stress-induced proteins characterized by their chaperone activity. HSP, however, are also important players in the control of the immune response: HSP are targeted by HSP-specific T cells and antibodies in healthy subjects and also during the course of autoimmune disorders and, conversely, HSP influence the activity of several immune cell types via innate receptor signalling pathways. In addition, the immune response to HSP can be mined for information on the state of the immune system. Newborns carry autoantibodies to HSP. However, this natural autoreactivity to HSP is modified by inflammation, and these changes can be monitored as biomarkers during postnatal life. Using antigen microarrays, we have shown that autoantibody patterns identify individuals prone to develop autoimmune diabetes before disease onset. Moreover, changes in the inflammatory process in multiple sclerosis are also reflected in the antibody response to self-HSP. Herein, we discuss the relevance of HSP and their immune activities for the regulation and monitoring of inflammation and autoimmune disease. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 101–114

Heat shock proteins (HSPs) were initially identified as a heterogeneous family of stress-induced proteins required to assist the folding of polypeptides into mature proteins in routine protein synthesis, in normal transport of proteins across membranes and in response to protein denaturation during cell stress (Hartl & Hayer-Hartl 2002). However, in addition to their intracellular chaperone function, HSPs establish a two-way dialogue with the immune system: HSPs are targeted by HSP-specific T cells and antibodies in healthy subjects and also during the course of autoimmune disorders and, conversely, HSPs influence the activity of several immune cell types. In this chapter, we discuss several aspects of this dialogue and postulate that its comprehension is of importance for our ability to prevent, treat and monitor human autoimmune disorders. 101

102

QUINTANA & COHEN

The immune system talks to HSPs: HSPs as targets of adaptive immunity HSPs are targets of the physiological immune response. Microbial HSPs were initially identified as a common dominant antigen recognized by the immune system of subjects infected with different micro-organisms (Young et al 1987). Soon afterwards, self-HSPs were found to be targeted by the immune system in response to inflammation (van Eden et al 1988, Anderton et al 1993). These initial studies paved the way to the identification of HSPs as a target of pathogenic and regulatory immune responses.

HSP as a target of pathogenic T and B cell responses Adjuvant arthritis (AA) is an autoimmune disease inducible in Lewis rats by immunization with killed Mycobacterium tuberculosis (Wauben et al 1994). T cell clones cross-reactive with cartilage and mycobacteria transfer arthritis to irradiated rats (Holoshitz et al 1984, van Eden et al 1985). The antigen recognized by the pathogenic clones was identified as the 180–188 region (mt180) of the mycobacterial 65 kDa HSP (Hsp65) (van Eden et al 1988), suggesting that Hsp65 bore an epitope cross-reactive with a joint self-epitope (van Eden et al 1985). Autoimmune T cell responses directed against the mammalian counterpart of Hsp65, the 60 kDa HSP (Hsp60) and also against other HSPs, have also been characterized in human arthritis (Blass et al 2001, Corrigall et al 2001, de Kleer et al 2003, Macht et al 2000). NOD mice develop diabetes as a consequence of a spontaneous T cell-mediated autoimmune process that destroys the insulin-producing β cells (Tisch & McDevitt 1996). NOD mice have a high frequency of self-reactive T cells, which is reflected by a highly self-reactive B cell repertoire (Quintana & Cohen 2001, Quintana et al 2004c). Several antigens are targeted by diabetogenic T cells including insulin (Wegmann et al 1994) and glutamic acid decarboxylase (GAD) (Tisch et al 1993). The connection of HSPs to autoimmune disease was confirmed when it was discovered that Hsp60 is targeted by T and B cells in spontaneous autoimmune NOD diabetes (Elias et al 1990). Moreover, anti-Hsp60 T-cell clones induce diabetes in irradiated NOD mice (Elias et al 1990), and immunization with the 437–460 immunodominant T cell epitope of Hsp60 (p277) can trigger transient insulitis and hyperglycaemia in standard strains of mice (Elias et al 1995). The relevance of HSPs for human autoimmune diabetes was highlighted by the finding of B and T cell responses to Hsp60 and p277 in autoimmune diabetes patients (AbulafiaLapid et al 1999, 2003, Sobel & Creswell 2006, Szebeni et al 2005). Autoimmunity to HSPs also characterizes atherosclerosis (Wick 2000), Bechet’s disease (Lehner 1997), systemic lupus erythematosus (Dhillon et al 1991), and

IMMUNE SYSTEM AND HSP60

103

other inflammatory and non-inflammatory disorders. These findings point to the importance of HSPs in autoimmune disorders, and suggest that HSP-specific T cell responses might be targeted for the management of pathogenic autoimmunity. HSPs as a target of regulatory responses Although it was shown that HSP-specific T cells can transfer autoimmune arthritis (AA) (Holoshitz et al 1984, van Eden et al 1985) or diabetes (Elias et al 1990), it was soon realized that immunity to HSPs is a double-edged sword with the power not only to drive but also to dampen autoimmune disorders. Hsp65 administered as a protein (Billingham et al 1990), encoded in a recombinant vaccinia virus (Hogervorst et al 1991) or as a DNA vaccine (Ragno et al 1997) can inhibit AA. The administration of HSP65 can also regulate other experimental models of arthritis, triggered either by the lipoidal amine CP20961 (Anderton et al 1995) or by pristane (Thompson et al 1998). Inhibition of AA by Hsp65 involves cross-reactivity with self-Hsp60 (van Eden et al 2005). We studied the specificity of the regulatory immune response that controls AA using DNA vaccines coding for human HSP60 (pHSP60) or mycobacterial Hsp65 (pHsp65) (Quintana et al 2002). Although both pHSP60 or pHsp65 protected against AA, pHSP60 was significantly more effective (Quintana et al 2002). We used DNA vaccines encoding fragments of HSP60 to identify immunoregulatory regions within HSP60. The anti-arthritogenic effects of the pHSP60 construct resided in the region of amino acids 1–260 of HSP60 (Quintana et al 2003a). Using HSP60derived overlapping peptides, we identified peptide Hu3 (amino acids 31–50 of HSP60) which was specifically recognized by the T cells of rats protected from AA by DNA-vaccination (Quintana et al 2003a). Vaccination with Hu3, or transfer of splenocytes from Hu3-vaccinated rats, prevented the development of AA, while vaccination with the mycobacterial homologue of Hu3 had no effect (Quintana et al 2003a). Prevention of AA by vaccination with pHSP60, DNA vaccines encoding the N-terminus of HSP60, or Hu3 was associated with the induction of T cells that secreted IFNγ, IL10 and TGFβ1 upon stimulation with HSP60 (Quintana et al 2002, 2003a). Thus, HSP60-specific T cells can control the progression of AA. Moreover, using DNA vaccines encoding the human 70 kDa heat shock protein (HSP70) or the human 90 kDa heat shock protein (HSP90), we showed that HSPs other than HSP60 can also inhibit AA. We found that DNA vaccination with HSP70 or HSP90 shifted the specific arthritogenic T-cell response from a Th1 to a Th2/3 phenotype and inhibited AA (Quintana et al 2004b). Thus, HSP70 and HSP90, and not only HSP60, can modulate arthritogenic T cell responses in AA. In humans, HSP60-specific immune responses in rheumatoid arthritis and juvenile chronic arthritis (Macht et al 2000, Prakken et al 1996) patients are associated

104

QUINTANA & COHEN

with a better prognosis and milder symptoms. Indeed human HSP-specific regulatory T cells have been characterized in arthritis patients (Kamphuis et al 2005, Massa et al 2007). All in all, these studies demonstrated that immunity to HSP is beneficial for the regulation of AA. Work with HSP60 in autoimmune diabetes strengthened this view. Vaccination of NOD mice with the p277 peptide of HSP60 arrested the development of diabetes (Elias et al 1991) and even induced the remission of overt hyperglycaemia (Elias & Cohen 1994). Successful p277 treatment was associated with the down-regulation of spontaneous T cell reactivity to p277 and with the induction of antibodies to p277 displaying Th2-like isotypes IgG1 and IgG2b (Elias et al 1997). Other peptides of HSP60 could also inhibit the development of spontaneous diabetes in NOD mice (Bockova et al 1997). NOD mice can also develop a more robust form of diabetes upon the administration of cyclophosphamide—cyclophosphamide-accelerated diabetes (CAD) (Yasunami & Bach 1988). We used DNA vaccination with constructs encoding human HSP60 (pHSP60) or mycobacterial Hsp65 (pHsp65) to explore the regulatory role of HSP60 (Quintana et al 2003a). Vaccination with pHSP60 protected NOD mice from CAD. In contrast, vaccination with pHsp65, with an empty vector or with a CpG-positive oligonucleotide was not effective, suggesting that efficacy of the pHSP60 construct might be based on regulatory HSP60 epitopes not shared with its mycobacterial counterpart, Hsp65 (Quintana et al 2003a). Vaccination with pHSP60 modulated the T cell responses to HSP60, and also to GAD and insulin: T cell proliferative responses were significantly reduced and the pattern of cytokine secretion to HSP60, GAD and insulin showed an increase in IL10 and IL5 secretion and a decrease in IFNγ secretion, compatible with a shift from a Th1-like towards a Th2-like autoimmune response (Quintana et al 2003a). Thus, immunoregulatory networks activated by vaccination with pHSP60 or p277 can spread to other B cell antigens such as insulin and GAD, and can control NOD diabetes. Human autoimmune diabetes was susceptible to immunomodulation by p277 therapy. A double-blind, phase II clinical trial was designed to study the effects of p277 therapy on newly diagnosed patients (Raz et al 2001, 2007). The administration of p277 after the onset of clinical diabetes preserved the endogenous levels of C-peptide (which fell in the placebo group) and was associated with lower requirements for exogenous insulin, revealing the arrest of inflammatory β -cell destruction (Raz et al 2001, 2007). Treatment with p277 was associated with an enhanced Th2 response to HSP60 and p277 (Raz et al 2001). Encouraging results of treatment with p277 were later obtained in an independent clinical trial (Huurman et al 2007). In conclusion, the administration of HSP60 peptides, or of whole HSP60 as a recombinant protein or a DNA vaccine, can halt autoimmune diabetes. Several

IMMUNE SYSTEM AND HSP60

105

antigens are targeted during the progression of diabetes (Tisch & McDevitt 1996). Thus, it is remarkable that HSP60-triggered immunoregulation controls the immune reactivity to non-cross-reactive autoantigens such as insulin and GAD. HSP60 as a target of anti-ergotypic regulatory T cells To exploit the immune properties of HSP60 for the treatment of autoimmune disorders, we have first to understand the mechanisms by which HSP60 can affect autoimmunity to other molecules. The identification of HSP60 as a target of antiergotypic regulatory T cells started to unveil these mechanisms. T cells reactive to activated T cells (but not to resting T cells) can control experimental autoimmune disease (Lohse et al 1989, Mimran et al 2004, 2005, Mor et al 1996). The T cell receptor of these regulatory T cells recognizes peptides derived from activation markers (ergotopes), such as the α -chain of the IL2 receptor (Mimran et al 2004, Mor et al 1996) or the TNFα receptor (Mor et al 1996). These cells are termed anti-ergotypic (Lohse et al 1989, Quintana & Cohen 2006). HSP expression is also up-regulated upon T cell activation (Ferris et al 1988). Thus, HSP60 might serve as an ergotope. We studied whether vaccination with DNA vaccines encoding HSP60, or with the regulatory peptide Hu3, might induce anti-ergotypic responses. To serve as an ergotope, HSP60 would have to fulfil two requirements. First, HSP60 must be up-regulated in activated T cells. Second, activated T cells must present HSP60-derived peptides to HSP60-specific regulatory T cells. The activation of T cells by mitogenic Con A or by specific antigen up-regulates intracellular levels of HSP60 (Cohen et al 2004, Quintana & Cohen 2006). Thus, the first condition is fulfi lled: T cell activation triggers HSP60 expression. Moreover, activated T cells can present HSP60. HSP60-specific T cells proliferated to activated T cells and secreted both IFNγ and TGFβ1 (Cohen et al 2004, Quintana & Cohen 2006). The activation of HSP60-specific T cells was MHC class II (RT1. B) restricted, since it could be inhibited with the OX6 (Wick 2000) monoclonal antibody (Cohen et al 2004, Quintana & Cohen 2006). Thus, HSP60 could function as an ergotope in vitro; however, can functional HSP60-specific anti-ergotypic responses be induced in vivo? DNA vaccination with pHSP60 was found to induce anti-ergotypic T cell responses, MHC class II (RT1.B) and MHC class I restricted. Vaccination with Hu3 induced only an MHC class II restricted (RT1.B) anti-ergotypic T cell response (Cohen et al 2004, Quintana & Cohen 2006). Thus, Hsp60-specific CD4 + and CD8 + anti-ergotypic T cells can be induced in vivo. Lymph node cells (LNCs) of AA rats stimulated with the immunodominant 180-88 T cell epitope of Hsp65 (mt180) secrete high levels of IFNγ (Cohen et al 2004, Quintana & Cohen 2006). Since T cells specific for this epitope have been

106

QUINTANA & COHEN

shown to transfer AA (van Eden et al 1988), the reactivity of LNC of AA to mt180 is thought to reflect the behaviour of the arthritogenic T cells. LNCs of AA rats stimulated with mt180 in the presence of Hsp60-specific anti-ergotypic T cells (but not with a control anti-MBP line) secreted significantly less IFNγ. Moreover, the transfer of Hsp60-specific anti-ergotypic T cells led to a significant decrease in the severity of AA. Thus, anti-ergotypic responses can control the arthritogenic response in vitro and in vivo.

HSP talks to the immune system: innate functions of HSP Several lines of research suggest that HSP can have a broad range of effects on the innate immune response (reviewed in Quintana & Cohen 2005). HSP60 and p277 can directly inhibit chemotaxis and activate anti-inflammatory programmes in human T cells, via toll-like receptor (TLR)2 (Zanin-Zhorov et al 2003), and can also modify B cell behaviour (Cohen-Sfady et al 2005). However, due to space limitations, we will focus our discussion on the direct effects of HSP on CD4 + CD25 + regulatory T cells (Treg). HSPs are endogenous ligands for innate receptors. HSP60 and HSP70 activate TLR4 and TLR2 (Vabulas et al 2002). Caramalho and colleagues reported that Treg can be activated via TLR4 by bacterial lipopolysaccharide (Caramalho et al 2003). Thus, it was conceivable that self-HSP might activate Treg through TLR. Indeed, we found that HSP60 acts as a costimulator of human Treg (Zanin-Zhorov et al 2006). Treatment of Treg with HSP60 or p277 significantly enhanced their ability to inhibit the activation of CD4 + CD25− or CD8 + effector T cells in terms of proliferation and IFNγ and TNFα secretion. The effects of HSP60 on Treg involved TLR2-dependent signalling pathways that led to the activation of PKC, PI3K and p38. HSP60-treated Treg suppressed effector T cells both by cell-to-cell contact and by secretion of TGFβ1 and IL10. Thus, HSP60 boosts the suppressive activity of Treg via TLR2 signalling.

Immune system and HSP dialogue as an inflammation biomarker Antibodies are of interest for the study of autoimmunity because they might play pathological roles, and also because IgG antibodies can reflect the activity of the T cell compartment. Since it is easier to assay antibodies than antigen-specific T cells, efforts have been invested to develop new technologies to monitor antibody repertoires (Quintana et al 2004c, Robinson et al 2002). Antigen microarrays are newly developed tools for the high-throughput characterization of the immune response that have been shown to be useful for the study of autoimmunity (Quintana et al 2004c, 2006, Robinson et al 2002). Healthy humans and mice are born with low levels of self-reactive antibodies (Coutinho

IMMUNE SYSTEM AND HSP60

107

et al 1995, Quintana & Cohen 2004). These natural autoantibody repertoires can be mined for information regarding the state of the immune response (Cohen 2007). By studying different mouse strains, we found that natural autoantibodies are indicators of susceptibility to the future development of various autoimmune diseases (Quintana et al 2003b, 2004a, 2004c, Quintana & Cohen 2001, 2004). Moreover, patterns of natural antibodies can be used to identify individuals prone to develop autoimmune disorders later in life. Approximately 50% of the male NOD mice in our colony develop diabetes following administration of cyclophosphamide (Quintana et al 2004c). We have used antigen microarrays to identify the individual NOD mice that will develop diabetes by studying serum samples taken before treatment with cyclophosphamide (Quintana et al 2004c). Based on the role played by Hsp60 in the immunoregulation of NOD diabetes (Cohen 1997, Cohen et al 2003), we included along with other diabetes-associated antigens, recombinant heat shock proteins, and peptide libraries covering the whole sequence of the 60 kDa and the 70 kDa heat shock proteins (HSP60 and HSP70, respectively). By analysing the antibody repertoire of each mouse, we could identify healthy mice that later developed autoimmune diabetes, (Quintana et al 2004c). Interestingly, the presence of antibodies to HSP60 and HSP70 were central in the identification of the individuals that would later become diabetic. We have recently used antigen arrays to study the immune response in multiple sclerosis (MS). MS is a chronic inflammatory disease of the CNS of presumed autoimmune aetiology (Sospedra & Martin 2005, Weiner 2004). Approximately 85–90% of patients begin with a relapsing-remitting (RRMS) course; in the remaining patients, MS presents a primary progressive course (PPMS). There is compelling evidence that immune processes play a major role in MS pathogenesis and progression (Sospedra & Martin 2005, Weiner 2004), and that inflammation outweighs neurodegeneration in the RRMS, but not in the progressive forms of the disease (Hauser & Oksenberg 2006). However, to date there has been no reliable method to characterize the complex immune processes that are unique to MS. Reactivity to HSP, however, has been described in MS (Aquino et al 1997, Gao et al 1995 Ousman et al 2007 Selmaj et al 1992) and experimental autoimmune encephalomyelitis (EAE) (Mor & Cohen 1992). We used antigen arrays to analyse the antibody response in MS, and we found specific immune responses directed against CNS proteins and lipids (Quintana et al, submitted); strikingly HSP-specific antibodies characterized RRMS but not PPMS (Quintana et al, submitted). Thus, antibodies to HSPs might be useful biomarkers to monitor inflammation in MS. Conclusions All in all, the results discussed here point to an ongoing dialogue between the immune system and HSPs, which results in the regulation of the immune response.

108

QUINTANA & COHEN

Our characterization of this dialogue provides us with a tool to monitor the immune state of a subject. Moreover, strategies aimed at exploiting this ongoing dialogue have the potential to constitute new strategies for the management of autoimmune disease; our successful initial results on the treatment of autoimmune diabetes supports this view. References Abulafia-Lapid R, Elias D, Raz I, Keren-Zur Y, Atlan H, Cohen IR 1999 T cell proliferative responses of type 1 diabetes patients and healthy individuals to human hsp60 and its peptides. J Autoimmun 12:121–129 Abulafia-Lapid R, Gillis D, Yosef O, Atlan H, Cohen IR 2003 T cells and autoantibodies to human HSP70 in type 1 diabetes in children. J Autoimmun 20:313–321 Anderton SM, van der Zee R, Goodacre JA 1993 Inflammation activates self hsp60-specific T cells. Eur J Immunol 23:33–38 Anderton SM, van der Zee R, Prakken B, Noordzij A, van Eden W 1995 Activation of T cells recognizing self 60-kD heat shock protein can protect against experimental arthritis. J Exp Med 181:943–952 Aquino DA, Capello E, Weisstein J et al 1997 Multiple sclerosis: altered expression of 70- and 27-kDa heat shock proteins in lesions and myelin. J Neuropathol Exp Neurol 56:664–672 Billingham ME, Carney S, Butler R, Colston MJ 1990 A mycobacterial 65-kD heat shock protein induces antigen-specific suppression of adjuvant arthritis, but is not itself arthritogenic. J Exp Med 171:339–344 Blass S, Union A, Raymackers J et al 2001 The stress protein BiP is overexpressed and is a major B and T cell target in rheumatoid arthritis. Arthritis Rheum 44:761–771 Bockova J, Elias D, Cohen IR 1997 Treatment of NOD diabetes with a novel peptide of the hsp60 molecule induces Th2-type antibodies. J Autoimmun 10:323–329 Caramalho I, Lopes-Caravalho T, Ostler D, Zelenay S, Haury M, Demengeot J 2003 Regulatory T cells selectively express toll-like receptors and are activated by lipopolysaccharide. J Exp Med 197:403–411 Cohen IR 1997 Questions about NOD mouse diabetes. Res Immunol 148:286–291 Cohen IR 2007 Real and artificial immune systems: computing the state of the body. Nat Rev Immunol 7:569–574 Cohen IR, Quintana FJ, Nussbaum G, Cohen M, Zanin A, Lider O 2003 HSP60 and the regulation of inflammation: physiological and pathological. In: van Eden W (ed) Heat shock proteins and inflammation. Basel: Birkhauser Verlag AG, p 1–13 Cohen IR, Quintana FJ, Mimran A 2004 Tregs in T cell vaccination: exploring the regulation of regulation. J Clin Invest 114:1227–1232 Cohen-Sfady M, Nussbaum G, Pevsner-Fischer M et al 2005 Heat shock protein 60 activates B cells via the TLR4-MyD88 pathway. J Immunol 175:3594–3602 Corrigall VM, Bodman-Smith MD, Fife MS et al 2001 The human endoplasmic reticulum molecular chaperone BiP is an autoantigen for rheumatoid arthritis and prevents the induction of experimental arthritis. J Immunol 166:1492–1498 Coutinho A, Kazatchkine MD, Avrameas S 1995 Natural autoantibodies. Curr Opin Immunol 7:812–818 de Kleer IM, Kamphuis SM, Rijkers GT et al 2003 The spontaneous remission of juvenile idiopathic arthritis is characterized by CD30+ T cells directed to human heat-shock protein 60 capable of producing the regulatory cytokine interleukin-10. Arthritis Rheum 48:2001– 2010

IMMUNE SYSTEM AND HSP60

109

Dhillon V, Latchman D, Isenberg D 1991 Heat shock proteins and systemic lupus erythematosus. Lupus 1:3–8 Elias D, Cohen IR 1994 Peptide therapy for diabetes in NOD mice. Lancet 343:704–706 Elias D, Markovits D, Reshef T, van der Zee R, Cohen IR 1990 Induction and therapy of autoimmune diabetes in the non-obese diabetic (NOD/Lt) mouse by a 65-kDa heat shock protein. Proc Natl Acad Sci USA 87:1576–1580 Elias D, Reshef T, Birk OS, van der Zee R, Walker MD, Cohen IR 1991 Vaccination against autoimmune mouse diabetes with a T-cell epitope of the human 65-kDa heat shock protein. Proc Natl Acad Sci USA 88:3088–3091 Elias D, Marcus H, Reshef T, Ablamunits V, Cohen IR 1995 Induction of diabetes in standard mice by immunization with the p277 peptide of a 60-kDa heat shock protein. Eur J Immunol 25:2851–2857 Elias D, Meilin A, Ablamunits V et al 1997 Hsp60 peptide therapy of NOD mouse diabetes induces a Th2 cytokine burst and downregulates autoimmunity to various beta-cell antigens. Diabetes 46:758–764 Ferris DK, Harel-Bellan A, Morimoto RI, Welch WJ, Farrar WL 1988 Mitogen and lymphokine stimulation of heat shock proteins in T lymphocytes. Proc Natl Acad Sci USA 85:3850–3854 Gao YL, Brosnan CF, Raine CS 1995 Experimental autoimmune encephalomyelitis. Qualitative and semiquantitative differences in heat shock protein 60 expression in the central nervous system. J Immunol 154:3548–3556 Hartl FU, Hayer-Hartl M 2002 Molecular chaperones in the cytosol: from nascent chain to folded protein. Science 295:1852–1858 Hauser SL, Oksenberg JR 2006 The neurobiology of multiple sclerosis: genes, inflammation, and neurodegeneration. Neuron 52:61–76 Hogervorst EJ, Schouls L, Wagenaar JP et al 1991 Modulation of experimental autoimmunity: treatment of adjuvant arthritis by immunization with a recombinant vaccinia virus. Infect Immun 59:2029–2035 Holoshitz J, Matitiau A, Cohen IR 1984 Arthritis induced in rats by cloned T lymphocytes responsive to mycobacteria but not to collagen type II. J Clin Invest 73:211–215 Huurman VA, Decochez K, Mathieu C, Cohen IR, Roep BO 2007 Therapy with the hsp60 peptide DiaPep277 in C-peptide positive type 1 diabetes patients. Diabetes Metab Res Rev 23:269–275 Kamphuis S, Kuis W, de Jager W et al 2005 Tolerogenic immune responses to novel T-cell epitopes from heat-shock protein 60 in juvenile idiopathic arthritis. Lancet 366:50– 56 Lehner T 1997 The role of heat shock protein, microbial and autoimmune agents in the aetiology of Behcet’s disease. Int Rev Immunol 14:21–32 Lohse AW, Mor F, Karin N, Cohen IR 1989 Control of experimental autoimmune encephalomyelitis by T cells responding to activated T cells. Science 244:820–822 Macht LM, Elson CJ, Kirwan JR et al 2000 Relationship between disease severity and responses by blood mononuclear cells from patients with rheumatoid arthritis to human heat-shock protein 60. Immunology 99:208–214 Massa M, Passalia M, Manzoni SM et al 2007 Differential recognition of heat-shock protein dnaJ-derived epitopes by effector and Treg cells leads to modulation of inflammation in juvenile idiopathic arthritis. Arthritis Rheum 56:1648–1657 Mimran A, Mor F, Carmi P, Quintana FJ, Rotter V, Cohen IR 2004 DNA vaccination with CD25 protects rats from adjuvant arthritis and induces an antiergotypic response. J Clin Invest 113:924–932 Mimran A, Mor F, Quintana FJ, Cohen IR 2005 Anti-ergotypic T cells in naive rats. J Autoimmun 24:191–201

110

QUINTANA & COHEN

Mor F, Cohen IR 1992 T cells in the lesion of experimental autoimmune encephalomyelitis. Enrichment for reactivities to myelin basic protein and to heat shock proteins. J Clin Invest 90:2447–2455 Mor F, Reizis B, Cohen IR, Steinman L 1996 IL-2 and TNF receptors as targets of regulatory T-T interactions: isolation and characterization of cytokine receptor-reactive T cell lines in the Lewis rat. J Immunol 157:4855–4861 Ousman SS, Tomooka BH, van Noort JM et al 2007 Protective and therapeutic role for alphaBcrystallin in autoimmune demyelination. Nature 448:474–479 Prakken AB, van Eden W, Rijkers GT et al 1996 Autoreactivity to human heat-shock protein 60 predicts disease remission in oligoarticular juvenile rheumatoid arthritis. Arthritis Rheum 39:1826–1832 Quintana FJ, Cohen IR 2001 Autoantibody patterns in diabetes-prone NOD mice and in standard C57BL/6 mice. J Autoimmun 17:191–197 Quintana FJ, Cohen IR 2004 The natural autoantibody repertoire and autoimmune disease. Biomed Pharmacother 58:276–281 Quintana FJ, Cohen IR 2005 Heat shock proteins as endogenous adjuvants in sterile and septic inflammation. J Immunol 175:2777–2782 Quintana FJ, Cohen IR 2006 Anti-ergotypic immunoregulation. Scand J Immunol 64:205–210 Quintana FJ, Carmi P, Mor F, Cohen IR 2002 Inhibition of adjuvant arthritis by a DNA vaccine encoding human heat shock protein 60. J Immunol 169:3422–3428 Quintana FJ, Carmi P, Mor F, Cohen IR 2003a DNA fragments of human HSP60 vaccinate against adjuvant arthritis: identification of a regulatory HSP60 peptide. J Immunol 171: 3533–3541 Quintana FJ, Pitashny M, Cohen IR 2003b Experimental autoimmune myasthenia gravis in naive non-obese diabetic (NOD/LtJ) mice: susceptibility associated with natural IgG antibodies to the acetylcholine receptor. Int Immunol 15:11–16 Quintana FJ, Buzas E, Prohaszka Z et al 2004a Knock-out of the histidine decarboxylase gene modifies the repertoire of natural autoantibodies. J Autoimmun 22:297–305 Quintana FJ, Carmi P, Mor F, Cohen IR 2004b Inhibition of adjuvant-induced arthritis by DNA vaccination with the 70-kd or the 90-kd human heat-shock protein: immune cross-regulation with the 60-kd heat-shock protein. Arthritis Rheum 50:3712–3720 Quintana FJ, Hagedorn PH, Elizur G, Merbl Y, Domany E, Cohen IR 2004c Functional immunomics: microarray analysis of IgG autoantibody repertoires predicts the future response of mice to induced diabetes. Proc Natl Acad Sci USA 101 Suppl 2:14615–14621 Quintana FJ, Merbl Y, Sahar E, Domany E, Cohen IR 2006 Antigen-chip technology for accessing global information about the state of the body. Lupus 15:428–430 Ragno S, Colston MJ, Lowrie DB, Winrow VR, Blake DR, Tascon R 1997 Protection of rats from adjuvant arthritis by immunization with naked DNA encoding for mycobacterial heat shock protein 65. Arthritis Rheum 40:277–283 Raz I, Elias D, Avron A, Tamir M, Metzger M, Cohen IR 2001 Beta-cell function in new-onset type 1 diabetes and immunomodulation with a heat-shock protein peptide (DiaPep277): a randomised, double-blind, phase II trial. Lancet 358:1749–1753 Raz I, Avron A, Tamir M, Metzger M et al 2007 Treatment of new-onset type 1 diabetes with peptide DiaPep277 is safe and associated with preserved beta-cell function: extension of a randomized, double-blind, phase II trial. Diabetes Metab Res Rev 23:292–298 Robinson WH, DiGennaro C, Hueber W et al 2002 Autoantigen microarrays for multiplex characterization of autoantibody responses. Nat Med 8:295–301 Selmaj K, Brosnan CF, Raine CS 1992 Expression of heat shock protein-65 by oligodendrocytes in vivo and in vitro: implications for multiple sclerosis. Neurology 42:795–800 Sobel DO, Creswell K 2006 Characterization of anti-islet cytotoxic human T-cell clones from patients with type 1 diabetes mellitus. Autoimmunity 39:323–332

IMMUNE SYSTEM AND HSP60

111

Sospedra M, Martin R 2005 Immunology of multiple sclerosis. Annu Rev Immunol 23:683–747 Szebeni A, Schloot N, Kecskemeti V et al 2005 Th1 and Th2 cell responses of type 1 diabetes patients and healthy controls to human heat-shock protein 60 peptides AA437–460 and AA394–408. Inflamm Res 54:415–419 Thompson SJ, Francis JN, Siew LK et al 1998 An immunodominant epitope from mycobacterial 65-kDa heat shock protein protects against pristane-induced arthritis. J Immunol 160:4628–4634 Tisch R, McDevitt H 1996 Insulin-dependent diabetes mellitus. Cell 85:291–297 Tisch R, Yang XD, Singer SM, Liblau RS, Fugger L, McDevitt HO 1993 Immune response to glutamic acid decarboxylase correlates with insulitis in non-obese diabetic mice. Nature 366:72–75 Vabulas RM, Braedel S, Hilf N et al 2002 The endoplasmic reticulum-resident heat shock protein Gp96 activates dendritic cells via the Toll-like receptor 2/4 pathway. J Biol Chem 277:20847–20853 van Eden W, Holoshitz J, Nevo Z, Frenkel A, Klajman A, Cohen IR 1985 Arthritis induced by a T-lymphocyte clone that responds to Mycobacterium tuberculosis and to cartilage proteoglycans. Proc Natl Acad Sci USA 82:5117–5120 van Eden W, Thole JE, van der Zee R et al 1988 Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature 331:171–173 van Eden W, van der Zee R, Prakken B 2005 Heat-shock proteins induce T-cell regulation of chronic inflammation. Nat Rev Immunol 5:318–330 Wauben MHM, Wagenaar-Hilbers JPA, van Eden W 1994 Adjuvant arthritis. In: Cohen IR, Miller A (eds) Autoimmune disease models. Academic Press Inc. p 201–216 Wegmann DR, Norbury-Glaser M, Daniel D 1994 Insulin-specific T cells are a predominant component of islet infi ltrates in pre-diabetic NOD mice. Eur J Immunol 24:1853–1857 Weiner HL 2004 Multiple sclerosis is an inflammatory T-cell-mediated autoimmune disease. Arch Neurol 61:1613–1615 Wick G 2000 Atherosclerosis—an autoimmune disease due to an immune reaction against heat-shock protein 60. Herz 25:87–90 Yasunami R, Bach JF 1988 Anti-suppressor effect of cyclophosphamide on the development of spontaneous diabetes in NOD mice. Eur J Immunol 18:481–484 Young DB, Ivanyi J, Cox JH, Lamb JR 1987 The 65 kDa antigen of mycobacteria—a common bacterial protein? Immunol Today 8:215–219 Zanin-Zhorov A, Nussbaum G, Franitza S, Cohen IR, Lider O 2003 T cells respond to heat shock protein 60 via TLR2: activation of adhesion and inhibition of chemokine receptors. FASEB J 17:1567–1569 Zanin-Zhorov A, Cahalon L, Tal G, Margalit R, Lider O, Cohen IR 2006 Heat shock protein 60 enhances CD4+ CD25+ regulatory T cell function via innate TLR2 signaling. J Clin Invest 116:2022–2032

DISCUSSION Gaston: I am interested in whether endogenous Hsp60, which you showed to be up-regulated, is actually presented by these T cells. In your experiments you demonstrated the ability of the class II-expressing T cell to present exogenous HSP. Quintana: There wasn’t any exogenous HSP. They were just activated T cells that were up-regulating their endogenous levels of Hsp60, used as APCs (antigen

112

QUINTANA & COHEN

presenting cells). I raised the clones by immunization with the same antigen, but then I used the clones as probes to see whether the endogenous Hsp60 was being presented in the membrane. Gaston: How does that endogenous Hsp60 get into the class II pathway? Does it have to be excreted first? Quintana: Which one is the HSP that is being presented? Is it the one that we saw (up-regulated by western blot) that might be in the membrane? I have no idea. What I do know is that if you treat those T cells with different cytokines, these are known to affect antigen processing, and this leads to the presentation of different epitopes of Hsp60. We still have no clue about the molecular mechanism for this. Macario: Is the proteasome a mediator of this antigen processing and presentation for Hsp60? We know that Hsp70 can interact with the proteasomal system (Macario & Conway de Macario 2007), but the situation with Hsp60 seems to be less clear, particularly in what pertains to the immunoproteasome. Quintana: Yes, there are indications that cytokines can change the composition of some of the members of the proteasome. This is linked to the presentation of different self peptides from Hsp60. Coates: I’d like to return to the blood. You are measuring antibodies in blood, and other people have been measuring heat shock proteins in blood, but what do we know about the effects of injection of high levels of HSPs into the blood, with measurable amounts of HSPs and antibodies? You’d have to inject a lot of HSP to get high levels. Quintana: I don’t recall any reports of this in the literature. We tried to generate a TLR4 knockout on the NOD background. HSPs are very important for diabetes, and one of the easiest ways to screen for the knockouts was to give them lipopolysaccharide (LPS) and get the survivors. Initially, we wanted to see whether injection of high levels of HSP would lead to some kind of endotoxin shock. We didn’t get to high levels. We didn’t really see anything, but we were looking for gross effects. Miller-Graziano: The presentation highlights the classical way of getting anergic T cells. This alteration in humans induces adaptive Treg. These adaptive Treg are a little bit different from the CD4 + CD25 + ‘natural’ Treg seen in a mouse or rat system. The classical way to block these T cells’ suppressive activity is to block their co-inhibitory molecules. This is what makes the TT antigen presentation possible: they up-regulate CD86, which binds preferentially to CTLA4 on the secondary T cells. I am interested in the idea that HSPs can up-regulate and down-regulate receptors. Did you see any up-regulation of CD86 or CTLA4 as a result of your HSP system? If you did assess receptor expression, did you assess blocking it?

IMMUNE SYSTEM AND HSP60

113

Quintana: Simplistically speaking, it is possible to take this anti-idiotypic antiHsp60 T cell and activate it either with a T cell or an APC. We did this. When we activate with DCs it leads to an expansion of the regulatory population, but if we activate it with a regulatory cell, on the one hand it will shut off the cell that is to be regulated but on the other hand it becomes anergic. In effect, it is regulating the regulators. It is shutting off what it has to shut off, then it is shutting itself off. Regarding the CD80/CD86, when T cells are activated they up-regulate their surface levels of co-stimulatory molecules. We blocked with these antibodies and then the whole system doesn’t work. Just to complicate things a bit more, human Tregs (CD4 + CD25 + ) are characterized by high expression of MHC class II. At a certain point the two systems are interconnected and cross-regulate each other. Asea: In your system the anti-Hsp60 Treg express TLR2. Did you try to generate any anti-Hsp70 Treg? If so, would you predict that they would express TLR4 on their surface? Quintana: This is an extension of some work we did with DNA vaccination. We got nice responses in terms of regulation of the disease by vaccination with Hsp70. We have some data showing that we also induced some anti-idiotypic regulators. I never got a good anti-Hsp70 cell line. Theoretically I would expect it to work. Pockley: A lot of the early work relating to the measurement of circulating antibodies derived from their relationship to both autoimmune disease and particularly cardiovascular disease/atherosclerosis. Georg Wick’s group in Innsbruck drove these studies. Qingbo Xu, who is now in London showed in a mouse model that if anti-HSP antibodies are administered to an ApoE knockout mouse (an animal model of cardiovascular disease) this markedly accelerates the development of disease (Foteinos et al 2005). This adds to some of the previous data which suggest in vitro at least that if these antibodies are added to cultured endothelial cells they induce cytotoxicity of these endothelial cells. The inference from this is that while it has been known for some time that stress protein expression within vascular bifurcations is up-regulated, the fact that antibodies can induce cytotoxicity suggests that there is some membrane or surface expression of these proteins to allow the antibodies to elicit their effect. It is also worth trying to differentiate between the baseline antibodies and their affinity levels, and then elevated levels of antibodies, to see whether the latter have different affinities to the naturally occurring ones. It is also interesting to look at the isotype profi le of the antibodies: are they IgG1? This would provide insight into the nature of the immune response that is driving the generation of these antibodies. Quintana: Obviously, the antibodies we are measuring are low affinity. They are not diluted, so even if they are of low affinity they might be relevant. The NOD mice are rich in natural antibodies to Hsp60. I did a fusion of spleens from naïve NOD mice to generate monoclonals. We raised many HSP-specific monoclonals

114

QUINTANA & COHEN

which are IgMs. One of the controls we use, to show that we can inhibit the activation of HSPs with TLR2, is one of those monoclonals. We don’t know how many of those antibodies in this Hsp60-specific repertoire are inhibitory, but we know that at least one of them is. In diabetes there is some good literature showing that antibodies of Th1-associated isotypes are linked with the progression of diabetes. Pockley: I think it is important to differentiate between types of antibodies, which is what you are doing. There is some work from Zoltan Prohászka which has demonstrated that antibodies against human HSP60 and mycobacterial Hsp65 differ in their complement-activating ability (Prohászka et al 1999). Quintana: There was some work by him on HSP-specific antibodies (Veres et al 2002). They have shown that there are complement-fixing isotypes associated with disease expression. Pockley: We did some work in some subjects with established hypertension. We found much higher levels of anti-HSP antibody in those individuals. When we refer these back and look at the immunoglobulin levels in individuals with established hypertension, these are elevated as well. If we correct everything for the non-specific effect of the disease process on immunoglobulin levels, then the levels of the anti-stress protein antibodies were no different to the controls. We also need a baseline against which we can compare antibody levels. Quintana: In the case of the NOD work we’ve done, it is interesting that we had the whole library of peptides for Hsp70, and none of these were picked up. References Foteinos G, Afzal AR, Mandal K, Jahangiri M, Xu Q 2005 Anti-heat shock protein 60 autoantibodies induce atherosclerosis in apolipoprotein E-deficient mice via endothelial damage. Circulation 112:1206–1213 Macario AJL, Conway de Macario E 2007 Molecular chaperones: multiple functions, pathologies, and potential applications. Front Biosci 12:2588–2600 Prohászka Z, Duba J, Lakos G et al 1999 Antibodies against human heat-shock protein (hsp) 60 and mycobacterial hsp65 differ in their antigen specificity and complement-activating ability. Int Immunol 11:1363–1370 Veres A, Szamosi T, Ablonczy M et al 2002 Complement activating antibodies against the human 60 kDa heat shock protein as a new independent family risk factor of coronary heart disease. Eur J Clin Invest 32:405–410

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Cell stress proteins: novel immunotherapeutics Stuart K. Calderwood, Jianlin Gong, Jimmy R. Theriault, Salamatu S. Mambula and Philip J. Gray Jnr Department of Medicine, Boston University School of Medicine, Boston, MA 02118 and Molecular and Cellular Radiation Oncolog y, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA 02215, USA

Abstract. Heat shock proteins (HSPs) play important roles in the immune system as carriers of tumour antigens and inflammatory agents. The HSPs are abundantly expressed stress proteins intrinsic to all cellular life, permitting proteins to carry out essential enzymic, signalling and structural functions within the tightly crowded milieu of the cell. To carry out these tasks, HSPs are equipped with a domain that binds unstructured sequences in polypeptides and a N-terminal ATPase domain that controls the opening and closing of the peptide-binding domain. HSPs can, using these domains, capture antigens processed by partial proteolysis in the cytoplasm of cancer cells. HSP/peptide complexes formed in the cytoplasm can then be secreted to take part in immune surveillance. Extracellular Hsp70 interacts with receptors on antigen presenting cells (APCs) either during episodes of cell death and lysis in vivo or during vaccination. A number of candidate receptors for Hsp70 on APCs have been proposed to take part in the antitumour immune function including the α 2 macroglobulin receptor CD91, Toll-like receptors, the signalling receptor CD40 and a number of scavenger receptors. Finally, Hsp70 complexes are able to deliver antigens to MHC class I and II molecules on the APC cell surface and lead to the presentation of tumour antigens to T lymphocytes. HSP–antigen complexes have proven effective in the treatment of rodent tumours in preclinical studies and are now undergoing clinical trials for treatment of human cancer. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 115–136

Extracellular Hsp70 and its role in antitumour immunity The Hsp70, Hsp90, Hsp110, Grp96 and Grp170 gene families encode molecular chaperones intrinsic to mammalian cellular life and important for cell survival (Georgopolis & Welch 1993, Lindquist & Craig 1988). The essential intracellular functions of these genes are in protein folding and protection from cell death pathways. It is now apparent that these proteins also possess extracellular properties and play 115

116

CALDERWOOD ET AL

significant roles in the immune response. Each protein contains at least two major functional domains including: (1) a peptide binding domain and (2) an ATPase domain that controls the opening and closing of the peptide binding domain (see reviews Bukau & Horwich 1998, Bukau et al 2006). These two domains mediate the immune properties of heat shock proteins (HSPs), permitting the acquisition of cellular antigens and their delivery to immune effector cells (Calderwood et al 2005, Noessner et al 2002, Srivastava & Amato 2001). In addition, most HSP gene families contain more than one member. For example, there are at least 12 members of the human HSP70 gene family members, including proteins expressed in cytoplasm, endoplasmic reticulum and mitochondria (Bukau & Horwich 1998, Lindquist & Craig 1988, Tang et al 2005). HSP expression becomes dysregulated in cancer and the elevated HSP levels that ensue protect emerging cancer cells from the triggering of programmed cell death that accompanies many of the steps in transformation, while also creating an opportunity for immune attack (Calderwood et al 2006, 2007, Ciocca & Calderwood 2005, Clark & Menoret 2001, Cornford et al 2000, Nylandsted et al 2000, Tang et al 2005). Current investigations suggest that HSPs, through their ability to bind peptide antigens, are major factors linking the tumour phenotype, with its accompanying expression of mutated and overexpressed oncoproteins, to immunity. A pro-immune function for Hsp70 family members is also inferred from the fact that the immunosuppressive drug 15-deoxyspergualin binds with high affinity to Hsp70 proteins (Nadler et al 1992). In addition, expression of HSPs in the execution stages of tumour cell killing directly stimulates tumour immunogenicity (Daniels et al 2004, Melcher et al 1998). Role of Hsp70 family members in antigen processing and cross presentation Mice and human subjects are able to mount a T cell-mediated response to cancer based on their recognition of epitopes from mutated proteins, the products of overexpressed oncoproteins or ectopically expressed developmental genes (Pardoll 2003). However, it is evident that most human tumours of non-viral origin are not rejected and that tumour cells often evolve a range of strategies that lead to suppression of the immune responses, including complete ablation of MHC class I gene expression, that may be stimulated (Chouaib et al 2002, Moller & Hammerling 1992, Pardoll 2003). In addition, as tumour cells with intact MHC class I expression can present tumour antigens to antigen presenting cells (APCs) in the absence of a co-stimulatory molecule, tolerance to tumour antigens is a frequent occurrence (Zou 2006). However, approaches to enhancing the immunogenicity of human cancer have been demonstrated (Pardoll 2003). Molecular chaperones offer a novel approach to enhancing tumour immuno-

HSP CANCER VACCINES

117

genicity as they are powerful multi-functional agents in tumour immunotherapy with talents as (1) immune adjuvants for activating APCs and breaking tolerance to tumour antigens and (2) antigen chaperones that can deliver tumour antigens to APCs (Belli et al 2002, Calderwood 2005b, Daniels et al 2004, Huang et al 2003, Enomoto et al 2006b, Melcher et al 1998, Srivastava 2002, Srivastava & Amato 2001). The immunological properties of HSPs derive initially from their roles in antigen processing. Tumour antigens are peptides originating in the damaged, denatured or prematurely terminated intracellular proteins that are degraded through the ubiquitin proteosome pathway to small peptides (Goldberg et al 2002, Rock et al 2002). A fraction of the peptides released from the proteasome are then used for immune surveillance (Goldberg et al 2002, Gromme & Neefjes 2002, Rock et al 2002). Cytoplasmic peptides are taken up into the endoplasmic reticulum (ER) through the ABC family transport system that involves the transporters associated with antigen processing (TAP); TAP1 and TAP2 form a complex in the ER membrane that transports peptides and delivers them to major histocompatibility class I (MHC I) protein complexes (Schumacher et al 1994, Shepherd et al 1993). Peptides of suitable size and sequence are then bound by MHC I, transported via the vesicular system and displayed on the cell surface where they are subject to surveillance by the T cell receptors of CD8 + lymphocytes (Neefjes et al 1993). Such cell surface display of antigens on MHC class I molecules permits identification of non-self foreign antigens in micro-organism-infected cells, which are then targeted for lysis by cytotoxic T lymphocytes (CTLs). However, this pathway can also be used to detect malignant cells if tumour antigens can be discriminated from normal self antigens (Germain 1994, Monaco 1992, Noessner et al 2002, Topalian et al 1994). Hsp70 family members participate in antigen processing through their ability to bind peptides using their polypeptide binding domain (Bukau & Horwich 1998, Flaherty et al 1990, Flynn et al 1989, Wilbanks et al 1995). It has been proposed that Hsp70 and Hsp90 might bind peptides released into the cytoplasm from the proteasome in a similar way to their acquisition by the TAP1/TAP2 complex, while ER-resident molecular chaperones such as Grp94/Gp96 and Grp170 might acquire such peptides in the ER. A number of studies have addressed the peptide sequence binding preferences of Hsp70 and indicated a similar peptide sequence preference as for MHC I, with roles for hydrophobic and basic amino acids (Flynn et al 1989, Fourie et al 1994, Gragerov & Gottesman 1994, Wu & Wang 1999). It is not known whether Hsp70 binds preferentially to any class of intracellular peptides although in an in vivo proteomic study conducted by Grossmann et al (2004) Hsp70 was shown to bind to peptides 8–26 amino acids in length. Hsp70 interacts selectively with a five amino acid core sequence containing some acidic residues (Grossmann et al 2004). Sequence preferences for the other HSPs are largely unknown.

118

CALDERWOOD ET AL

In addition, Hsc70 takes part in the processing of external antigens as well as endogenous antigens undergoing autophagy in the lysosome proximal endosome compartments of APCs, and is thus involved in peptide loading of MHC class II and presentation to CD4 + T lymphocytes (Panjwani et al 1999). Intrinsic cell proteins containing a consensus sequence (KFERQ) are bound by Hsc70 and then taken up into lysosomal endosomes by the receptor LAMP2a (Agarraberes et al 1997, Chiang et al 1989, Cuervo & Dice 1998, Strawbridge & Blum 2007). Hsc70 co-associates with MHC II in spherical organelles in macrophages and participates in the presentation of external antigens to MHC class II-restricted T lymphocytes (Panjwani et al 1999). A role for Hsc70 in protecting peptides generated by proteolysis in MHC II containing endosomes similar to the role proposed for the Hsp70 family in chaperoning cytoplasmic peptides has been suggested (Agarraberes et al 1997, Chiang et al 1989, Cuervo & Dice 1998). A common theme for HSPs— mobilizing target proteins towards the sites of degradation while sparing a fraction of the partially digested peptide for immune surveillance purposes—is indicated for the main pathways of protein degradation in both the cytoplasm and lysosome. As the lysosomal/autophagic pathway can be stimulated by serum starvation of tissue culture cells, it may thus play a role in generation of tumour immunity in nutritionally-deprived tumour cells, and could potentially be captured in the production of antitumour Hsp70-based vaccines (Asea et al 2000b, Cuervo & Dice 1998, Strawbridge & Blum 2007). HSPs thus have ready access to tumour antigens from both intracellular and external processing pathways. These proteins, devoid of a leader sequence for secretion, would seem doomed to remain trapped along with their cargo in the cytoplasm of the tumour. Hsp60 and Hsp70 can, however, escape from the cytoplasm as indicated by their appearance in the serum of many patients (Pockley 2003). It has also been shown that Hsp70 is released from tumour cells undergoing necrosis and such Hsp70 is highly immunogenic (Mambula & Calderwood 2006a, Srivastava 2003, Daniels et al 2004). Hsp70 can also be secreted from intact tumour cells and other cell types by a non-classical pathway involving lysosomal endosomes (Mambula & Calderwood 2006b). HSPs, with or without tumour antigen cargo, therefore have access to the extracellular spaces, the bloodstream and targets in the immune system. Interaction of HSPs with immune cells Extracellular HSP–peptide complexes (HSP–PC) can therefore exit viable or compromised cells and interact with the host immune system. Such complexes can bind in a saturable manner to the cell surfaces of macrophages, B cells, dendritic cells (DCs), natural killer cells and T lymphocytes, as well as some non-immune cells (Srivastava 2002, Theriault et al 2005, 2006). As DCs are the most potent

HSP CANCER VACCINES

119

APC

T cell Chaperoned antigen ( ) presented to CD8+ cells

HSP-peptide complex

T cell receptor

T cell activation

R1

R2

Receptors involved in antigen presentation by HSPs

Co-stimulatory signal induced by R2

FIG. 1. Mechanisms of T cell activation by Hsp70 peptide complexes. HSP complexed to antigenic peptide (HSP–PC) binds to receptor R1 on the APC, which triggers internalization of the complex. A series of intracellular interactions leads to HSP–PC dissociation from R1, peptide dissociation from Hsp and with or without further proteolytic processing, peptide binding to MHC class I and presentation on the cell surface. The MHC class I–peptide complex is then recognized by T cell receptors on CD8 + T lymphocytes, leading to T cell activation. Full activation of APCs to a form that can interact effectively with T cells and lead to killing of target cells requires the receipt of a second signal through receptor R2. Occupation of R2 by HSP leads to the expression of co-stimulatory molecules, which migrate to the APC surface adjacent to MHC class I–peptide complexes and interact with counter-receptors on the T cell surface.

antigen processing cells (APCs) in tumour immunity, we will mainly discuss the interaction of HSPs with this cell type (Larsson et al 2001). In order for DCs to stimulate a potent T cell response, at least two activating events need to occur: (1) the acquisition of antigen by DCs, presentation on cell surface MHC class I molecules and binding of MHC class I–antigen complexes to the T cell receptors of CD8 + T cells; and (2) the transmission of a second signal that leads to induction of co-stimulatory molecules such as B7 and CD40 that bind counter-receptors on the T cells and lead to a robust immune response (Fig. 1). These co-ordinated signals lead to MHC class I-restricted activation of T cells, T cell proliferation and the ability to kill target cells in tumours. The second, MHC class I-independent signal may be provided by adjuvants that stimulate innate immunity or by helper T cells (Engleman et al 2004). HSPs, however, have been proposed to offer ‘one-stop shopping’ with regard to tumour antigen presentation, providing antigen portage plus conferring the second

120

CALDERWOOD ET AL

signal (Srivastava & Amato 2001). Several HSPs, including Hsp70, Hsp90, Gp96 and calreticulin confer the ability of cell extracts to cross prime CD8 + T cells; this group of stress proteins is evidently necessary and sufficient for this purpose (Binder & Srivastava 2005). Cross priming involves peptide binding to Hsp, internalization of the Hsp-antigen complex and delivery of the peptide to MHC class I molecules to form MHC I/peptide complexes expressed on the APC surface (Fig. 1). Previous studies indicated that HSP-mediated uptake of tumour antigens occurs through receptor-mediated endocytosis because of the low (nanomolar) concentrations of HSP involved and the saturability of the process, a property of binding to surfaces with limited numbers of high affinity receptors (Singh-Jasuja et al 2000a, b). Hsp60, Hsp70, Gp96 and Hsp110 are able to directly confer a second, adjuvant-like signal and lead to expression of inflammatory cytokines (interleukin [IL]1β, tumour necrosis factor [TNF] α , IL6, IL12) and co-stimulatory molecules (Calderwood et al 2005). Whether this dual stimulus is achieved through binding to one receptor or multiple cell surface receptors is not clear. Four classes of cell surface structure have been proposed as HSP receptors, including: (1) the CDI4/ TLR2/4 pattern recognition receptor (PRR) complex known to play a key role in innate immunity; (2) the α 2 macroglobulin receptor LRP/CD91 involved in internalizing a number of extracellular proteins; (3) the TNF receptor family protein CD40; and (4) members of the c-type lectin/scavenger receptor (SR) families (Asea et al 2000c, Becker et al 2002, Binder et al 2000, Delneste et al 2002, Srivastava 2002). We have evaluated these receptor types by cloning expression vectors encoding each receptor into CHO cells (normally null for Hsp70 binding activity) as described (Theriault et al 2005, 2006). The results of the experiments on Hsp70 binding are summarized in Table 1. Our current findings indicate that Hsp70 binds with relatively high avidity to SR family members LOX1, SREC1 and FEEL1 and to the c-type lectins NKG2D and NKG2A (Theriault et al 2005, 2006). We observed minimal binding to c-type lectins DC-SIGN, Dectin1, CLEC1 and CLEC2 (Table 1). In addition, we were, under the conditions of the assay, unable to detect high avidity binding to CD14, TLR2, TLR4, CD91 or CD40, cell surface proteins that were previously reported as Hsp70 receptors (Table 1) (Theriault et al 2005, 2006). Nonetheless, there is little doubt that most of these receptors play key roles in the immune effects of HSPs (Becker et al 2002, Srivastava 2002). These considerations therefore imply the possibilities of (1) kinetically rapid but reversible HSP–receptor binding not detected by the equilibrium binding assay or (2) hierarchical coupling of receptors with a primary receptor such as SR providing a platform for secondary, tertiary, etc. receptors that may mediate signalling. An illustration of hierarchical regulation of signalling receptors after SR occupation by ligand is that SREC1 can act in combination with TLR2 to mediate signalling in response to Escherichia coli membrane protein OMPA (Jeannin et al 2005).

HSP CANCER VACCINES TABLE 1

121

Candidate receptors for Hsp70

Receptor

Type

Cells

Hsp70 binds?

TLR2 TLR4 CD14 CD40 CD91 LOX1 DC-SIGN Dectin 1 CLEC1 CLEC2 SREC1 FEEL1 NKG2A NKG2C NKG2D

signalling signalling signalling signalling adaptor protein scavenger scavenger scavenger scavenger scavenger scavenger scavenger CTL CTL CTL

APC etc APC etc APC etc APC many APC, endo APC APC APC APC APC ? NK NK NK, T cell

− − − − − + − − − − ++ ++ ++ ++ ++

Binding to receptors was assayed in CHO transfectants with two exceptions, which are DC-SIGN, which was in K-562 (also HSP binding null) and the TLR, which were in HEK293. Binding is indicated by +, lack of binding by −.

Uptake of HSP–PC by APCs, activation of transmembrane signalling cascades and presentation of peptides to cell surface MHC molecules Binding of Hsp70 is followed by the activation of transmembrane signalling mechanisms, internalization and the delivery of the peptide cargo of HSP to MHC molecules. HSP-induced APC maturation and pro-inflammatory signalling When HSPs are released from cells they can initiate a potent innate immune/ inflammatory response resulting in APC maturation (Noessner et al 2002, Srivastava 2002, Srivastava & Amato 2001). APC maturation requires the activation of a gene expression programme including co-stimulatory molecules such as CD40, OX40L, B7.1 and B7.2 on the cell surface in order for the APC to interact more efficiently with CD8+ T lymphocytes (Fig. 2) (Ohashi & DeFranco 2002). Indeed, Hsp70 plays a key functional role for CD40 in breaking tolerance in an autoimmune form of diabetes by Hsp70 (Millar et al 2003). It seems likely that that Hsp70-induced DC maturation involves CD40 and/or CD40L up-regulation and probably requires activation of the transcription factor NF-κ B (Pulendran 2004) (Fig. 2).

122

CALDERWOOD ET AL

Heat shock protein

LOX1

CD14 CD91

CD40 TLR2

NF-κB complex

SSR5 Ca2+

Adhesion molecules; Immunological synapse Cytokines, Co-stimulatory molecules Adhesion factors

Receptor-mediated endocytosis

HSP-mediated transcription

FIG. 2. HSP receptors. Receptors known to interact with HSPs are shown. These can be divided into known internalizing receptors such as the scavenger receptor LOX1 and CD91 that bind multiple HSPs, and receptors with powerful transmembrane signalling activity such as TLR2, CD40 and SSR5. Receptors may act in parallel as shown here or may be organized into hierarchical interaction patterns. The SR can form large structures on the cell surface, which could act as platforms for co-ordinating Hsp70-mediated signalling.

Many of the effector genes for DC maturation require transcriptional activation by NF-κ B and therefore receptors that are involved in DC maturation by HSPs likely cause downstream activation of NF-κ B (Fig. 2) (Pulendran 2004). Some intriguing studies have shown that extracellular Hsp60 and Hsp70 induce NF-κ B through the activation of the CD14/TLR signalling pathway in a CD14-dependent manner in APCs (Asea et al 2000c, Vabulas et al 2002). This results in the expression of TNFα , IL1β and IL6 as well as co-stimulatory molecules (Asea et al 2000c, Vabulas et al 2002). Although the role of this pathway in the innate immune response to molecular chaperones is not clear, it can be activated equally effectively by Hsp70–PC and free Hsp70 (Asea et al 2000a). However, these signalling properties of HSPs may be indirect since we could fi nd no evidence for direct binding of extracellular Hsp70 to TLR 2/4 or CD14 (Theriault et al 2005). In addition, some scepticism has been directed at the role of CD14 or TLR4 due to the possibility of Hsp70 contamination by endotoxins (Gao & Tsan 2003). However, the evidence for HSP activation of TLR signalling now seems almost overwhelming (Asea et al 2002, Hasan et al 2005, Quintana & Cohen 2005, Takenaka et al 2004).

HSP CANCER VACCINES

123

In addition, other HSP receptors are coupled to NF-κ B, and LOX1 has recently been shown to activate the NF-κ B pathway in endothelial cells, as can ligand binding to SREC1 (Cominacini et al 2000, Jeannin et al 2005, Ouaaz et al 2002). The SR ligand oxidized-low density lipoprotein can also induce CD40 expression through LOX1 suggesting a putative role for ligand-bound LOX1 in APC maturation (Delneste et al 2002, Li et al 2003). Recent studies also show that mycobacterial Hsp70 binds to CCR5 chemokine receptors on DCs promoting DC aggregation, immune synapse formation between DCs and T cells and enhanced immune responses (Floto et al 2006). It is not clear whether human HSP70 binds to CCR5 although experiments showing increased Ca2+ flux after Hsp70 binding suggest such an interaction as chemokine receptors signal through Ca2+ transients (Asea et al 2000c). Hsp70–PC internalization by APCs and peptide association with cell surface MHC molecules Following receptor-induced HSP internalization by APCs, bound peptides may proceed through a number of potential pathways prior to re-presentation on the cell surface by MHC receptors. HSP–PC could be taken up by the endocytic pathway and presented to MHC class II molecules by the standard extracellular pathway of antigen presentation. In addition, Hsp70 can also present peptides to MHC class I molecules by a process known as antigen cross-presentation (Baker-LePain et al 2003, Srivastava 2002). Gp96 has been shown to be rapidly internalized after interacting with surface receptor(s) in a pre-endosomal compartment (Arnold-Schild et al 1999, Berwin & Nicchitta 2001). Endocytosed Gp96 co-localizes with FcR and MHC class I but not with other receptors such as CD91, transferrin, Rab5a or lysosomal markers such as LAMP1 or LAMP2 (Berwin & Nicchitta 2001, Singh-Jasuja et al 2000a). It has been shown recently that Hsp70–PC are rapidly internalized by each of the SR family members LOX1, SREC1 and FEEL1, strongly suggesting a role for these structures in pathways involving HSP–PC uptake and cross-priming (Fig. 2) (Theriault et al 2006). As mentioned above, HSPs may also bind to other immune cells in addition to APCs. For instance, Hsp60 acts as a co-stimulator of regulatory T cells by interacting with TLR2 (Zanin-Zhorov et al 2006) and can also interact with TLR4 in B cells (Cohen-Sfady et al 2005). These interactions, which are anti-inflammatory in nature, may be involved in regulating the response to HSPs and in preventing autoimmunity in vaccine-treated patients (Daniels et al 2004). In addition, NK cells associate with Hsp70 on the surface of tumour cells and this mediates tumour cell killing (Multhoff 2002, Multhoff & Hightower 1996). We have recently found that Hsp70 binds to the NK receptors NKG2A and NKG2D, and these

124

CALDERWOOD ET AL

interactions could be involved in NK killing of Hsp70-expressing tumour cells (Theriault et al 2006). Strategies for Hsp70-based vaccine design The major pioneer in molecular chaperone-based anticancer vaccine design has been Pramod Srivastava who has prepared autologous vaccines in mice and in human patients with the aim of directly targeting the unique, often mutated antigens that characterize each individual neoplasm (Belli et al 2002, Mazzaferro et al 2003, Srivastava 2000, 2002, 2003). In this approach, HSP–PC (Hsp70 and Gp96) are isolated from the patients’ tumours by affinity chromatography and formulations of such purified HSPs applied in a multi-dose regimen. The aim is for tumour antigens to be cross-presented to the patient’s APCs and for the unique mixture of peptides from the individual tumour to induce immunity. These studies build on the pioneering work of Srivastava and Old who showed that experimentally induced cancers of inbred mice and rats elicit stronger individually tumour-specific immunity (which is not tumour-type specific) (Srivastava & Old 1988). These studies are currently undergoing Phase III trial and recent results suggest a trend towards increased response particularly in patients receiving longer courses of vaccination (Srivastava 2006). Success may thus be limited by the amount of tumour available after resection as response to treatment evidently is related to the number of treatments with HSP vaccine (Srivastava 2006). These results may be viewed as encouraging when compared with the recent NCI trials of metastatic cancer treatment in the USA with synthetic peptides, DNA vaccines, DC vaccines and viruses in which an objective response rate of 2.6% was reported (Rosenberg et al 2004). Factors that limit the effects of vaccines may be structural and relate to the avidity with which individual HSPs bind peptides, the nature of their peptide repertoires and ability to induce a co-stimulatory response. For instance, Hsp70 gene family members Hsp110 and Grp170 possess a greatly enhanced ability to bind peptides compared with other HSPs, can bind avidly to larger polypeptides and are superior agents in cancer vaccine production than smaller HSPs (Manjili et al 2003, Segal et al 2006). Recent studies have suggested that larger polypeptides are superior to smaller peptides in inducing immunity (Rock et al 2005, Shen & Rock 2006). One recent approach emphasizes use of combinations of chaperones with the hope of increasing the repertoire of peptides presented to APC. Indeed the studies of Binder & Srivastava (2005) indicate that Hsp70, Gp96, Hsp90 and calreticulin carry between them the whole of the intracellular peptide repertoire required for cross priming T cells against ovalbumin or β -galactosidase. An approach using what are described as ‘chaperone-rich lysates’ has been described in which tumour cell lysates are partially purified by isoelectric focusing (IEF) to

HSP CANCER VACCINES

125

enrich the above chaperones (Zeng et al 2006). Such preparations are effective against mouse tumours in a prophylactic context. Curiously, although the lysates are prepared under highly denaturing conditions for IEF using molar concentrations of urea, the preparations evidently retain peptides and are able to induce antigen specific responses (Zeng et al 2006). These studies are supported by a recent publication showing that the effectiveness of an Hsp70-based vaccine derived from tumour–DC fusion cells was partially dependent on co-isolation of Hsp90 (Enomoto et al 2006). Use of multi-chaperone formulations may thus be indicated. This study also showed that gentle and rapid isolation of the Hsp70 is optimal for retaining Hsp70–peptide interactions and immunity, and that peptides are lost when Hsp70 purification by affinity chromatography is employed (Enomoto et al 2006). These studies used MUC1, a normal antigen whose expression and processing is altered in cancer. Mice expressing MUC1 became tolerant to this antigen while the Hsp70-based vaccine was able to overcome tolerance and cause tumour rejection, suggesting that non-mutated antigens may be targeted by Hsp70based vaccines (Enomoto et al 2006). The effectiveness of the Hsp70-based vaccine derived from tumour–DC fusion cells was dependent on stimulation of innate immunity and was ineffective in MyD88 knockout cells further indicating a role for TLR. It has been suggested that there are three requirements for effective cancer immunotherapy: (1) a sufficient number of avid tumour reactive lymphocytes present in the tumour-bearing host, (2) these must be capable of reaching the tumour and extravasating, and (3) lymphocytes that penetrate the tumour must have appropriate effector mechanisms to destroy the cancer (Rosenberg 2004). One innovative approach described recently emphasizes the inflammatory nature of HSPs; local inflammation may lead to increased influx of T lymphocytes and enhanced cell killing due to APC maturation. The approach devised by Vile and colleagues involves targeted killing of normal melanocytes overexpressing HSP70 to generate an antigen-specific CD8 + T lymphocyte response against established melanoma (Calderwood 2005, Daniels et al 2004). The rationale behind this unorthodox approach is that the majority of peptides that are presented by melanoma cells and recognized by T cells from patients arise from developmental proteins that are also expressed in normal melanocytes (Engelhard et al 2002). The existence of shared antigens between the proliferating melanocytes and the melanoma cells suggested that if a CD8 + T cell response could be generated against the dying melanocytes it could also target the tumour cells (Daniels et al 2004). This approach requires the specific targeting of proliferating melanocytes using local expression of a cell suicide gene (HSVtk) under the control of a promoter from the tyrosinase gene that is specifically active in this cell population. For tumour rejection, the critical requirement was that the killing took place in melanocytes engineered to overexpress HSP70 (Daniels et al 2004). Cell death in

126

CALDERWOOD ET AL

the sacrificed population leads to the extracellular release of the HSP70, modulation of DC function, and generation of a CD8 + T cell response against melanocytic TSTA that eradicates primary and metastatic melanoma (Daniels et al 2004). The preclinical studies in mice show the feasibility of this approach and its potential for translation into clinical treatment of malignant melanoma (Daniels et al 2004). One further question is whether such an approach could be generalized to the treatment of other tumours at different sites. This approach would require the targeted sacrifice of a population of normal cells with an antigenic repertoire similar to the tumour under treatment. A related approach, involves local killing of tumour cells with forced overexpression of Hsp70, obviating the need to find a related normal tissue for priming antitumour immunity (Huang et al 2003, Todryk et al 1999). Of course, the question of potential autoimmune destruction of normal tissues, especially as this approach depends on the generation of CD8 + T lymphocytes directed against antigens common to normal cells and tumour cells, is a concern with this approach (Daniels et al 2004). The targeting of the small fraction of proliferating melanocytes may, however, limit the extent of normal cell targets. In addition, the studies of Vile and colleagues show that the effects of the treatment are self-regulatory through the activation of CD4 + CD25 + lymphocytes, which inhibited the activity of tumour-specific CD8 + T lymphocytes and rapidly attenuated the response (Daniels et al 2004). It is interesting to speculate that this may be due to delayed triggering of the anti-inflammatory effects of HSPs which can activate the CD4 + CD25 + population of lymphocytes (Chandawarkar et al 1999, Quintana et al 2004). The outcome of immunotherapy approaches involving HSPs may therefore depend on the relative potencies of their pro-immune and anti-inflammatory effects. A further refinement is to combine chaperone-based immunotherapy with low temperature hyperthermia (Repasky & Issels 2002). Fever-range hyperthermia has the advantage of causing HSP induction and release and inducing highly efficient homing of APC and T lymphocytes to tumours and activation (Evans et al 2001, Mambula & Calderwood 2006b). Recent studies show that hyperthermia also causes lymphocyte trafficking across the high endothelial venules through a mechanism involving enhanced IL6 signalling (Chen et al 2006). This approach therefore is of high promise in combination with HSP-based immunotherapy. Conclusions Anticancer vaccines prepared using heat shock proteins as primary ingredients are beginning to show promise in animal studies and clinical trials, although it is clear that further advances may be required to increase their effectiveness in the clinic. These advances depend partially on increasing our understanding of the immunobiological properties of the HSP. The potential of HSPs as vaccines depends on

HSP CANCER VACCINES

127

their ability to function in cross-priming of T cells and as immunological adjuvants and pro-inflammatory agents. Understanding of these effects has been delayed by the open question concerning the nature of the key receptors that mediate HSP interaction with immune cells. However, discovery of the SR as the primary Hsp70 receptor and TLR and CCR5 as signalling receptors may lead the way for rapid development of this approach. Acknowledgements We acknowledge the support of the Department of Radiation Biology at Beth Israel Deaconess Medical Center. This work was supported by grants 5RO1CA047407 and 3RO1CA094397.

References Agarraberes FA, Terlecky SR, Dice JF 1997 An intralysosomal hsp70 is required for a selective pathway of lysosomal protein degradation. J Cell Biol 137:825–834 Arnold-Schild D, Hanau D, Spehner D et al 1999 Cutting edge: receptor-mediated endocytosis of heat shock proteins by professional antigen-presenting cells. J Immunol 162:3757–3760 Asea A, Kabingu E, Stevenson MA, Calderwood SK 2000a HSP70 peptide-bearing preparations act as chaperokines. Cell Stress Chaperones 5:425–431 Asea A, Kraeft SK, Kurt-Jones EA et al 2000b HSP70 stimulates cytokine production through a CD14-dependant pathway, demonstrating its dual role as a chaperone and cytokine. Nat Med 6:435–442 Asea A, Rehli M, Kabingu E et al 2002 Novel signal transduction pathway utilized by extracellular HSP70: role of toll-like receptor (TLR) 2 and TLR4. J Biol Chem 277:15028–15034 Baker-LePain JC, Reed RC, Nicchitta CV 2003 ISO: a critical evaluation of the role of peptides in heat shock/chaperone protein-mediated tumor rejection. Curr Opin Immunol 15:89–94 Becker T, Hartl FU, Wieland F 2002 CD40, an extracellular receptor for binding and uptake of Hsp70-peptide complexes. J Cell Biol 158:1277–1285 Belli F, Testori A, Rivoltini L et al 2002 Vaccination of metastatic melanoma patients with autologous tumor-derived heat shock protein gp96-peptide complexes: clinical and immunologic fi ndings. J Clin Oncol 20:4169–4180 Berwin B, Nicchitta CV 2001 To fi nd the road traveled to tumor immunity: the trafficking itineraries of molecular chaperones in antigen-presenting cells. Traffic 2:690–697 Binder RJ, Srivastava PK 2005 Peptides chaperoned by heat-shock proteins are a necessary and sufficient source of antigen in the cross-priming of CD8+ T cells. Nat Immunol 6:593–599 Binder RJ, Han DK, Srivastava PK 2000 CD91: a receptor for heat shock protein gp96. Nat Immunol 1:151–155 Bukau B, Horwich AL 1998 The Hsp70 and Hsp60 chaperone machines. Cell 92:351–66 Bukau B, Weissman J, Horwich A 2006 Molecular chaperones and protein quality control. Cell 125:443–451 Calderwood SK 2005 Chaperones and slow death—a recipe for tumor immunotherapy. Trends Biotechnol 23:57–59 Calderwood SK, Theriault JR, Gong J 2005 Message in a bottle: role of the 70-kDa heat shock protein family in anti-tumor immunity. Eur J Immunol 35:2518–2527 Calderwood SK, Khaleque MA, Sawyer DB, Ciocca DR 2006 Heat shock proteins: chaperones of tumorigenesis. Trends Biochem Sci 31:164–172

128

CALDERWOOD ET AL

Calderwood SK, Khaleque MA, Ciocca DR 2007 Heat shock proteins in the progression of cancer, Vol. 7. Springer, New York Chandawarkar RY, Wagh MS, Srivastava PK 1999 The dual nature of specific immunological activity of tumor-derived gp96 preparations. J Exp Med 189:1437–1442 Chen Q, Fisher DT, Clancy KA et al 2006 Fever-range thermal stress promotes lymphocyte trafficking across high endothelial venules via an interleukin 6 trans-signaling mechanism. Nat Immunol 12:1299–1308 Chiang HL, Terlecky SR, Plant CP, Dice JF 1989 A role for a 70-kilodalton heat shock protein in lysosomal degradation of intracellular proteins. Science 246:382–385 Chouaib S, Thiery J, Gati A et al 2002 Tumor escape from killing: role of killer inhibitory receptors and acquisition of tumor resistance to cell death. Tissue Antigens 60:273–281 Ciocca DR, Calderwood SK 2005 Heat shock proteins in cancer: diagnostic, prognostic, predictive, and treatment applications. Cell Stress Chaperones 10:86–103 Clark PR, Menoret A 2001 The inducible Hsp70 as a marker of tumor immunogenicity. Cell Stress Chaperones 6:121–125 Cohen-Sfady M, Nussbaum G, Pevsner-Fischer M et al 2005 Heat shock protein 60 activates B cells via the TLR4-MyD88 pathway. J Immunol 175:3594–3602 Cominacini L, Pasini AF, Garbin U et al 2000 Oxidized low density lipoprotein (ox-LDL) binding to ox-LDL receptor-1 in endothelial cells induces the activation of NF-kappaB through an increased production of intracellular reactive oxygen species. J Biol Chem 275:12633–12638 Cornford PA, Dodson AR, Parsons KF et al 2000 Heat shock protein expression independently predicts clinical outcome in prostate cancer. Cancer Res 60:7099–7105 Cuervo AM, Dice JF 1998 Lysosomes, a meeting point of proteins, chaperones, and proteases. J Mol Med 76:6–12 Daniels GA, Sanchez-Perez L, Diaz RM et al 2004 A simple method to cure established tumors by inflammatory killing of normal cells. Nat Biotechnol 22:1125–1132 Delneste Y, Magistrelli G, Gauchat J et al 2002 Involvement of LOX-1 in dendritic cellmediated antigen cross-presentation. Immunity 17:353–362 Engelhard VH, Bullock TN, Colella TA, Sheasley SL, Mullins DW 2002 Antigens derived from melanocyte differentiation proteins: self-tolerance, autoimmunity, and use for cancer immunotherapy. Immunol Rev 188:136–146 Engleman EG, Brody J, Soares L 2004 Using signaling pathways to overcome immune tolerance to tumors. Sci STKE pe28 Enomoto Y, Bharti A, Khaleque AA et al 2006 Enhanced immunogenicity of heat shock protein 70 peptide complexes from dendritic cell-tumor fusion cells. J Immunol 177:5946– 5955 Evans SS, Wang WC, Bain MD, Burd R, Ostberg JR, Repasky EA 2001 Fever-range hyperthermia dynamically regulates lymphocyte delivery to high endothelial venules. Blood 97:2727–2733 Flaherty KM, DeLuca-Flaherty C, McKay DB 1990 Three-dimensional structure of the ATPase fragment of a 70K heat-shock cognate protein. Nature 346:623–628 Floto RA, MacAry PA, Boname JM et al 2006 Dendritic cell stimulation by mycobacterial Hsp70 is mediated through CCR5. Science 314:454–458 Flynn GC, Chappell TG, Rothman JE 1989 Peptide binding and release by proteins implicated as catalysts of protein assembly. Science 245:385–390 Fourie AM, Sambrook JF, Gething MJ 1994 Common and divergent peptide binding specificities of hsp70 molecular chaperones. J Biol Chem 269:30470–30478 Gao B, Tsan MF 2003 Endotoxin contamination in recombinant human heat shock protein 70 (Hsp70) preparation is responsible for the induction of tumor necrosis factor alpha release by murine macrophages. J Biol Chem 278:174–179

HSP CANCER VACCINES

129

Georgopolis C, Welch WJ 1993 Role of the major heat shock proteins as molecular chaperones. Annu Rev Cell Biol 9:601–634 Germain RN 1994 MHC-dependent antigen processing and peptide presentation: providing ligands for T lymphocyte activation. Cell 76:287–299 Goldberg AL, Cascio P, Saric T, Rock KL 2002 The importance of the proteasome and subsequent proteolytic steps in the generation of antigenic peptides. Mol Immunol 39:147– 164 Gragerov A, Gottesman ME 1994 Different peptide binding specificities of hsp70 family members. J Mol Biol 241:133–135 Gromme M, Neefjes J 2002 Antigen degradation or presentation by MHC class I molecules via classical and non-classical pathways. Mol Immunol 39:181–202 Grossmann ME, Madden BJ, Gao F et al 2004 Proteomics shows Hsp70 does not bind peptide sequences indiscriminately in vivo. Exp Cell Res 297:108–117 Hasan A, Sadoh D, Palmer R, Foo M, Marber M, Lehner T 2005 The immune responses to human and microbial heat shock proteins in periodontal disease with and without coronary heart disease. Clin Exp Immunol 142:585–594 Huang XF, Ren W, Rollins L et al 2003 A broadly applicable, personalized heat shock proteinmediated oncolytic tumor vaccine. Cancer Res 63:7321–7329 Jeannin P, Bottazzi B, Sironi M et al 2005 Complexity and complementarity of outer membrane protein A recognition by cellular and humoral innate immunity receptors. Immunity 22:551–560 Larsson M, Fonteneau JF, Bhardwaj N 2001 Dendritic cells resurrect antigens from dead cells. Trends Immunol 22:141–148 Li D, Liu L, Chen H, Sawamura T, Mehta JL 2003 LOX-1, an oxidized LDL endothelial receptor, induces CD40/CD40L signaling in human coronary artery endothelial cells. Arterioscler Thromb Vasc Biol 23:816–821 Lindquist S, Craig EA 1988 The heat shock proteins. Annu Rev Genet 22:631–637 Mambula SS, Calderwood SK 2006a Heat induced release of Hsp70 from prostate carcinoma cells involves both active secretion and passive release from necrotic cells. Int J Hyperthermia 22:575–585 Mambula SS, Calderwood SK 2006b Heat shock protein 70 is secreted from tumor cells by a nonclassical pathway involving lysosomal endosomes. J Immunol 177:7849–7857 Manjili MH, Wang XY, Chen X et al 2003 HSP110-HER2/neu chaperone complex vaccine induces protective immunity against spontaneous mammary tumors in HER-2/neu transgenic mice. J Immunol 171:4054–4061 Mazzaferro V, Coppa J, Carrabba MG et al 2003 Vaccination with autologous tumor-derived heat-shock protein gp96 after liver resection for metastatic colorectal cancer. Clin Cancer Res 9:3235–3245 Melcher A, Todryk S, Hardwick N, Ford M, Jacobson M, Vile RG 1998 Tumor immunogenicity is determined by the mechanism of cell death via induction of heat shock protein expression. Nat Med 4:581–587 Millar DG, Garza KM, Odermatt B et al 2003 Hsp70 promotes antigen-presenting cell function and converts T-cell tolerance to autoimmunity in vivo. Nat Med 9:1469–1476 Moller P, Hammerling G 1992 The role of surface HLA A, B, C molecules in tumor immunity. Cancer Surv 13:101–127 Monaco JJ 1992 A molecular model of MHC class-I-restricted antigen processing. Immunol Today 13:173–179 Multhoff G 2002 Activation of natural killer cells by heat shock protein 70. Int J Hyperthermia 18:576–585 Multhoff G, Hightower LE 1996 Cell surface expression of heat shock proteins and the immune response. Cell Stress Chaperones 1:167–176

130

CALDERWOOD ET AL

Nadler SG, Tepper MA, Schacter B, Mazzucco CE 1992 Interaction of the immunosuppressant deoxyspergualin with a member of the Hsp70 family of heat shock proteins. Science 258:484–486 Neefjes JJ, Momburg F, Hammerling GJ 1993 Selective and ATP-dependent translocation of peptides by the MHC-encoded transporter. Science 261:769–771 Noessner E, Gastpar R, Milani V et al 2002 Tumor-derived heat shock protein 70 peptide complexes are cross-presented by human dendritic cells. J Immunol 169:5424–5432 Nylandsted J, Brand K, Jaattela M 2000 Heat shock protein 70 is required for the survival of cancer cells. Ann N Y Acad Sci 926:122–125 Ohashi PS, DeFranco AL 2002 Making and breaking tolerance. Curr Opin Immunol 14:744–759 Ouaaz F, Arron J, Zheng Y, Choi Y, Beg AA 2002 Dendritic cell development and survival require distinct NF-kappaB subunits. Immunity 16:257–270 Panjwani N, Akbari O, Garcia S, Brazil M, Stockinger B 1999 The HSC73 molecular chaperone: involvement in MHC class II antigen presentation. J Immunol 163:1936–1942 Pardoll D 2003 Does the immune system see tumors as foreign or self? Annu Rev Immunol 21:807–839 Pockley AG 2003 Heat shock proteins as regulators of the immune response. Lancet 362:469–476 Pulendran B 2004 Modulating vaccine responses with dendritic cells and Toll-like receptors. Immunol Rev 199:227–250 Quintana FJ, Cohen IR 2005 Heat shock proteins regulate inflammation by both molecular and network cross-reactivity. Cambridge University Press, Cambridge Quintana FJ, Carmi P, Mor F, Cohen IR 2004 Inhibition of adjuvant-induced arthritis by DNA vaccination with the 70-kd or the 90-kd human heat-shock protein: immune cross-regulation with the 60-kd heat-shock protein. Arthritis Rheum 50:3712–3720 Repasky E, Issels R 2002 Physiological consequences of hyperthermia: heat, heat shock proteins and the immune response. Int J Hyperthermia 18:486–489 Rock KL, York IA, Saric T, Goldberg AL 2002 Protein degradation and the generation of MHC class I-presented peptides. Adv Immunol 80:1–70 Rock KL, Hearn A, Chen CJ, Shi Y 2005 Natural endogenous adjuvants. Springer Semin Immunopathol 26:231–246 Rosenberg SA 2004 Shedding light on tumor immunotherapy of cancer. N Engl J Med 350:1461–1463 Rosenberg SA, Yang JC, Restifo NP 2004 Cancer immunotherapy: moving beyond current vaccines. Nat Med 10:909–915 Schumacher TN, Kantesaria DV, Heemels MT et al 1994 Peptide length and sequence specificity of the mouse TAP1/TAP2 translocator. J Exp Med 179:533–540 Segal BH, Wang XY, Dennis CG et al 2006 Heat shock proteins as vaccine adjuvants in infections and cancer. Drug Discov Today 11:534–540 Shen L, Rock KL 2006 Priming of T cells by exogenous antigen cross-presented on MHC class I molecules. Curr Opin Immunol 18:85–91 Shepherd JC, Schumacher TN, Ashton-Rickardt PG et al 1993 TAP1-dependent peptide translocation in vitro is ATP dependent and peptide selective. Cell 74:577–584 Singh-Jasuja H, Scherer HU, Hilf N et al 2000a The heat shock protein gp96 induces maturation of dendritic cells and down-regulation of its receptor. Eur J Immunol 30:2211–2215 Singh-Jasuja H, Toes RE, Spee P et al 2000b Cross-presentation of glycoprotein 96-associated antigens on major histocompatibility complex class I molecules requires receptor-mediated endocytosis. J Exp Med 191:1965–1974 Srivastava PK 2000 Immunotherapy of human cancer: lessons from mice. Nat Immunol 1:363–366

HSP CANCER VACCINES

131

Srivastava P 2002 Interaction of heat shock proteins with peptides and antigen presenting cells: chaperoning of the innate and adaptive immune responses. Annu Rev Immunol 20:395–425 Srivastava P 2003 Hypothesis: controlled necrosis as a tool for immunotherapy of human cancer. Cancer Immun 3:4 Srivastava PK 2006 Therapeutic cancer vaccines. Curr Opin Immunol 18:201–205 Srivastava PK, Amato RJ 2001 Heat shock proteins: the ‘Swiss Army Knife’ vaccines against cancers and infectious agents. Vaccine 19:2590–2597 Srivastava PK, Old LJ 1988 Individually distinct transplantation antigens of chemically induced mouse tumors. Immunol Today 9:78–83 Strawbridge AB, Blum JS 2007 Autophagy in MHC class II antigen prpcessing. Curr Opin Immunol 19:87–92 Takenaka R, Yokota K, Ayada K et al 2004 Helicobacter pylori heat-shock protein 60 induces inflammatory responses through the Toll-like receptor-triggered pathway in cultured human gastric epithelial cells. Microbiology 150:3913–3922 Tang D, Khaleque AA, Jones ER et al 2005 Expression of heat shock proteins and HSP messenger ribonucleic acid in human prostate carcinoma in vitro and in tumors in vivo. Cell Stress Chaperones 10:46–59 Theriault JR, Mambula SS, Sawamura T, Stevenson MA, Calderwood SK 2005 Extracellular HSP70 binding to surface receptors present on antigen presenting cells and endothelial/ epithelial cells. FEBS Lett 579:1951–1960 Theriault JR, Adachi H, Calderwood SK 2006 Role of scavenger receptors in the binding and internalization of heat shock protein 70. J Immunol 177:8604–8611 Todryk S, Melcher AA, Hardwick N et al 1999 Heat shock protein 70 induced during tumor cell killing induces Th1 cytokines and targets immature dendritic cell precursors to enhance antigen uptake. J Immunol 163:1398–1408 Topalian SL, Rivoltini L, Mancini M et al 1994 Human CD4+ T cells specifically recognize a shared melanoma-associated antigen encoded by the tyrosinase gene. Proc Natl Acad Sci USA 91:9461–9465 Vabulas RM, Ahmad-Nejad P, Ghose S, Kirschning CJ, Issels RD, Wagner H 2002 HSP70 as endogenous stimulus of the Toll/interleukin-1 receptor signal pathway. J Biol Chem 277:15107–15112 Wilbanks SM, Chen L, Tsuruta H, Hodgson KO, McKay DB 1995 Solution small-angle X-ray scattering study of the molecular chaperone Hsc70 and its subfragments. Biochemistry 34:12095–12106 Wu SJ, Wang C 1999 Binding of heptapeptides or unfolded proteins to the chimeric C-terminal domains of 70kDa heat shock cognate protein. Eur J Biochem 259:449–455 Zanin-Zhorov A, Cahalon L, Tal G, Margalit R, Lider O, Cohen IR 2006 Heat shock protein 60 enhances CD4+ CD25+ regulatory T cell function via innate TLR2 signaling. J Clin Invest 116:2022–2032 Zeng Y, Graner MW, Katsanis E 2006 Chaperone-rich cell lysates, immune activation and tumor vaccination. Cancer Immunol Immunother 55:329–338 Zou W 2006 Regulatory T cells, tumour immunity and immunotherapy. Nat Rev Immunol 6:295–307

DISCUSSION Feder: I thought you said that only Hsp70 that was specifically secreted or at least expressed by the tumour was effective, and that generalized Hsp70 was not. Is that correct?

132

CALDERWOOD ET AL

Calderwood: No. This is Hsp70 extracted from tumours and purified by affinity chromatography. People are using these approaches of trying to modify HSPs, tranfect them into tumours, and have them released in vivo. It may be more effective this way. When affinity chromatography is used, a good deal of the peptide will be lost. Feder: So there is nothing special about the Hsp70 that is being expressed by the tumour itself. Is there any post-translational modification that makes it more effective? Calderwood: We don’t really know. Hsp70 is not phosphorylated. It may be ubiquitinated. Most proteins that are normally intracellular proteins but which then become secreted require a sorting signal or some kind of modification. But we don’t know for Hsp70. Macario: Has it been possible to characterize the antigenic moiety in the HSP– antigen complex? This is a very interesting point since the effects of HSPs on tumour growth are varied and, most likely, they occur through distinct mechanisms, some involving tumour antigen from the start, i.e. tumour antigen complexed to the HSP molecule, and some seem to proceed through other immunomodulatory pathways in which the tumour antigen participates late in the process (Capello et al 2007). Calderwood: People have tried this but it is hard to get at. There is a disparate population of peptides, so it is hard to get a discernable signal out. Macario: We know that the same tumours are heterogeneous in different individuals, but it is puzzling that only the complex isolated from the same individual is effective against the tumour in that individual but not in others, even if the tumour is of the same type in all individuals (Cappello et al 2007). Does this mean that the antigenic moiety is unique for each individual? Calderwood: That’s the theory. But it has been possible to use common antigens. For instance, John Subjeck and his group, working on the large HSPs used c-ERBB2/neu, which is a common antigen overexpressed in many breast cancers, and it works (Manjili et al 2003). Asea: You said that the scavenging receptor might be able to send downstream signals to the TLRs. There are some TLRs which are intracellular receptors. How do you think this might work? Calderwood: The only data available on this comes from a group in France. They found that the TLRs can function downstream of SREC1 and LOX1 (Jeannin et al 2005). Multhoff: You mentioned C-type lectin receptors NKG2A and C. Normally in NK cells these receptors are associated with CD94, which is also a C-type lectin receptor. We showed binding of Hsp70 to the C-type lectin receptor CD94, and we also could block this with the antibody. Calderwood: We use cells overexpressing both CD94 and the NKG receptors, so it could be difficult to know if Hsp70 binds CD94, NKG2A or NKG2C or the

HSP CANCER VACCINES

133

heterodimer. However, Hsp70 also binds to NKG2D which is a homodimer, suggesting that the NGKG2 receptors can bind to Hsp70. Cheetham: There is a nice situation with two families, some members of which bind Hsp70 and some members of which don’t. Does this give you any hint as to what feature of Hsp70 might be being seen? Calderwood: The scavenger receptors are thought to recognize patterns of charge on the derivatized protein. The LOX1 receptor has been looked at carefully, and it would appear that ligands bind to bands of negative charge. I believe that there are domains like this in Hsp70. Nixon: This might seem naïve, but it seems to me that elements of what you presented echo the sort of work going on in the reproductive field where they are trying to develop immunocontraceptives. The goal of this research is to raise a specific immune response against reproductive antigens that render the target animals infertile. From our experience it is apparent that the use of different adjuvants and different modes of antigen delivery can dramatically influence the immune response generated, and therefore the level of fertility inhibition that we are able to achieve. I am wondering whether similar approaches, incorporating different adjuvants and different modes of delivery, are also being investigated in your field? Calderwood: People are looking at the effect of different sites of injection of vaccine. We do get different responses at different sites. Pockley: Some of the first experimental work from Chandarwarkar and Pramod Srivastava with regards to the immunological properties of Gp96/Grp94 demonstrated that if it is administered subcutaneously it is 10 times less biologically effective than if it is administered intradermally. The mode of immunization does seem to play a part. Coates: The new therapies were really interesting, with multiple chaperones and cytokine release, together with necrotic killing of tissue. This reminded me of tuberculosis (TB). There is a current therapy for bladder cancer with BCG, which is a TB vaccine. Is there any ongoing work on mimicking aspects of TB for cancer therapy? Henderson: The Coley vaccine was used back in the 19th century for treating cancer. This is simply a preparation of a bacterium such as Mycobacterium tuberculosis. Hightower: Some of the phage display libraries that I mentioned earlier found not only the expected aromatic and hydrophobic peptide, but also some unusual ones. For example, Ivone Takanaka (Takanaka et al 1995) found that NIVRKKK was a strong binder. Those results were quickly buried in a shallow grave! I think it is time to dig some of these unusual peptides up and take another look at them. Cheetham: Different tumours seem to have different chaperones as their drug of choice. Is there any correlation between high levels of a particular chaperone and

134

CALDERWOOD ET AL

its efficiency in your therapy, or is it just these chaperones involved in antigen presentation that are the best ones to use for therapy? Calderwood: It would appear that the best ones to use are the large HSPs. I think the reason they haven’t been in clinical trials may be related to patent issues. Cheetham: Has anyone tried the small HSPs? They are highly overexpressed in some tumours. Calderwood: They are thought to be immunosuppressive. Hsp27 is the protein found most often to be overexpressed in cancer, followed by Hsp70. Miller-Graziano: You said you weren’t interested in some of the NK receptors, but I know that the old LIR inhibitory receptors which were traditionally NK receptors have now been renamed the ILT receptors, and they are prominently found on DCs. I was interested to hear you say that HSP doesn’t bind MARCO, because according to Siamon Gordon this is an inhibitory receptor for antigenpresenting cells. I have seen some evidence that Hsp10 may bind MARCO. I thought this was an interesting discrimination. Calderwood: We didn’t do the MARCO experiment: we took this from the literature. The early literature also said that SRA doesn’t bind HSP, so we will go back and look at MARCO again. Gupta: You mentioned that Hsp70 binds to the SLC receptor, which also binds to oxidized LDL. Oxidized LDL is important in the development of atherosclerosis. Is it possible that Hsp70 may have an effect on the development of this process through binding to this receptor? Calderwood: Potentially, it could compete with oxidized LDL. Graham Pockley’s work suggests there is a negative correlation between Hsp70 levels and atherosclerosis. Pockley: Johan Frostegård at the Karolinska Institute used a cultured monocytic cell line (Frostegård et al 1996) and showed that if these cells are incubated with oxidized LDL then the cells up-regulate the expression of Hsp70. There is a relationship between the inflammatory consequences of oxidized LDL and Hsp70 induction. Gupta: You mentioned that Hsp70 with bound peptides extracted from tumours is helpful in immunotherapy. If you strip the peptides from the Hsp70 are they still effective? Calderwood: No. The peptide needs to be bound with Hsp70 for it to work. You can purify with either ADP agarose, in which case you get Hsp70 peptide complexes, or ATP agarose, in which case you lose the peptides. Gupta: So it is the complex that has the immunotherapeutic properties, not the peptides alone. Calderwood: Yes, in in vitro cross-presentation assays the complex is about 100fold more effective than unconjugated peptide.

HSP CANCER VACCINES

135

Multhoff: I agree that Hsp70 devoid of peptide doesn’t stimulate any T cells. But there is activation of NK cells. In Srivastava’s models (Janetzki et all 2000) he showed that if he completely depletes NK cells the immunogenic activity goes down. There is an impact mediated by NK cells and an impact of T cells. The NK effect might require just HSPs without peptides; the T cell effect needs the HSPchaperoned tumour-specific peptides. If you look at an immune response, NK cells are the first line of defence. It makes sense to do NK cell therapy. You start an immune response, then there is necrosis or apoptosis, and then you go on with the T cell response. Panayi: Tumours express HSPs. Presumably, this is just a quantitative effect, by delivering HSPs that you have extracted from a tumour that are loaded with peptide. Or is there something different with these HSPs that are present in the tumour, that they can’t do this cross-priming? Let me put this another way. You have a tumour and you extract the HSPs from it which are loaded with peptides. You use these in immunotherapy. They are already present in the original tumour, so why aren’t they able to exercise this cross-priming and enhancement of tumour immunity? Is it a quantitative effect? Calderwood: I’m not sure why HSP isn’t released from the tumour. Pockley: It might also depend on the availability of the relevant immune cells. If you immunize subcutaneously then you have the full array of APCs. Within the tumour microenvironment you will have a very different potential profi le of cells that are capable of generating a response. Lund: It is possible that it does occur, and the therapeutic approach is just enhancing the natural approach, which might be quite effective. Calderwood: It might also be tolerogenic, at least in small quantities. Gaston: A lot of cancer immunotherapy has been focused on trying to generate an immune response to tumours. But it is clear that for a number of tumours there is an excellent response that is entirely turned off by regulatory T cells. If you have markers for antigen-specific T cells you can find lymph nodes and tumours stuffed with these, but entirely unable to work.

References Cappello F, Czarnecka AM, La Rocca GL, Di Stefano A, Zummo G, Macario AJL 2007 Hsp60 and Hsp10 as antitumor molecular agents. Cancer Biol Ther 6:487–489 Frostegård J, Kjellman B, Gidlund M, Andersson B, Jindal S, Kiessling R 1996 Induction of heat shock protein in monocytic cells by oxidised low density lipoprotein. Atherosclerosis 121:93–103 Janetzki S, Palla D, Rosenhauer V, Lochs H, Lewis JJ, Srivastava PK 2000 Immunization of cancer patients’s with autologous cancer-derived HSP gp96 preparations: a pilot study. Int J Cancer 88:232–238

136

CALDERWOOD ET AL

Jeannin P, Bottazzi B, Sironi M et al 2005 Complexity and complementarity of outer membrane protein A recognition by cellular and humoral innate immunity receptors. Immunity 22:551–560 Manjili MH, Wang XY, Chen X et al 2003 HSP110-HER2/neu chaperone complex vaccine induces protective immunity against spontaneous mammary tumors in HER-2/neu transgenic mice. J Immunol 171:4054–4061 Takenaka IM, Leung SM, McAndrew SJ, Brown JP, Hightower LE 1995 Hsc70-binding peptides selected from a phage display peptide library that resemble organellar targeting sequences. J Biol Chem 270:19839–19844

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

General discussion Ellis: I have a concern about blood protein levels. My only knowledge of this is from my annual check-up, which measures a dozen or so proteins in my blood. The report lists a ‘normal range’ for each protein. I assume this range reflects a homeostatic mechanism that controls blood composition. Against this background, I am not sure how to view what we are being told about the level of heat shock proteins (HSPs) in the blood, which I understand varies from below detection to up to 10 mg/ml. This doesn’t seem to fit into a homeostatic view of blood as regards HSPs. Pockley: I have the same concerns. I think what we find is that in early atherosclerosis, there is a relationship between high levels of circulating Hsp60 and atherosclerosis. There is a correlation, but it is difficult to understand how such high ranges can be possible. In the context of the fi nding that these high levels are stable within individuals, there is regulatory control. Whether this is the regulatory control of induction or release of these proteins into the circulation, or control at the level of removal, is unknown. Brian Henderson has immunoprecipitated these proteins from the blood and identified them as being Hsp60. In some instances these proteins might be generating some sort of complex which is elevating the signal within the ELISA to give a heightened response. Individual groups who use slightly different protocols come in with similar results. Ellis: Is anything known about the physical state of chaperones in the blood? Henderson: When we immunoprecipitate we get what looks like the monomer with higher oligomers. We have done MALDI-TOF with the 60 kDa band and higher bands, and it appears to be genuine Hsp60. Hightower: I’d like to make a pitch for the vascular endothelium as a source of HSPs in the blood. If we go back to the late 1970s and the work of Fredric P. White (Currie & White 1983, White 1981), he was one of the first to heat shock animals. He used rats, and then sliced virtually every organ, ran gels and did autoradiography on the slices. They all lit up. When he looked at the slices, the place where the HSPs were most highly concentrated was in and around the microvasculature. In later studies of heat-shocking rats at rather high temperatures, it was possible to make the vascular endothelium cytoprotected. One of the results is that extravasation is stopped (House et al 2001). Panayi: Francisco Quintana, I think I got the data wrong, but in the multiple sclerosis (MS) patients that you studied, one of the things that disturbed me is that the group with persistent MS was a very small column. 137

138

GENERAL DISCUSSION

Quintana: The way we do this is by using a learning set and a test set. Each group has at least 30 patients and 30 controls. Whatever we fi nd in these is validated in 50 new patients and 50 new controls. In the case of the Nurses’ Health Study this is more difficult because we didn’t have access to so many samples. When I talk about 30 samples, we use 20 for a learning set. What I showed was the frequency in all those groups of the reactivity of Hsp60. In other words, how many epitopes were recognized consistently by all the patients in the group, and then validated in a test study. Panayi: This raises the question of whether the set that you are determining has specificity and sensitivity for the outcome you want. I am a clinician and I see a patient presenting for the first time with a disease that I diagnose as MS. How sensitive and specific is my diagnosis that on the basis of that set, this patient will have one of the three forms of MS over the next 10 years? Quintana: First I will tell you how we did this experiment. It was with blind samples. We analysed 120 samples. We used 80 of them to learn, and then the other 40 we used to predict what we thought they were. Our success rate was about 80%, which is not perfect. We are trying to use the same type of approach to see who will respond to interferon, for example, and we didn’t get any predictive ability. Lund: What is critical in these kinds of analyses is the size of the training set. If this were rolled out as a national programme with thousands of data sets, how much could you improve this 80%? Is this an upper limit because of the variability inherent in the system? Quintana: We have run similar experiments in Alzheimer’s disease. It is a sampling problem. The challenge we are facing now is how to analyse thousands of samples. We are printing our own chips. How can you make 1000 chips where you don’t have a batch effect? It’s a technical issue. Conway de Macario: When we talk about Hsp70 in human fluids, we talk about blood. But there are reports of the presence of Hsp70 and other HSPs in other biological fluids, for example cerebrospinal fluid (Mueller et al 2003), urine (Mudaliar et al 2006), vaginal secretion (Giraldo et al 1999), and saliva (Fabian et al 2003, 2004, 2007). Does anyone know more about this? Nixon: We have found Hsp60 and Hsp90b1 in the fluids of the male reproductive tract. They are present in amorphous masses we term dense bodies in the epididymal fluid. Conway de Macario: When you need to fi ll out all these forms to extract fluids from patients, getting saliva is a lot simpler than getting blood, so if we could find a correlation between Hsp70 in saliva and blood this would be useful. Corrigall: We have measured BiP in synovial fluid and blood. We find BiP at a slightly higher level in synovial fluid.

GENERAL DISCUSSION

139

Asea: We have worked with Steve Witkin at the Weill Medical College of Cornell University to look at HSPs in amniotic fluid of pregnant women. Panayi: We need to make a distinction between serum and plasma. Serum is where the whole blood clots, and could very well be a stress situation. There are several proteins useful in medical diagnostics where plasma is needed. People who have done these experiments on serum may have to repeat their experiments in plasma. Henderson: We measure in plasma. Pockley: Our first study was plasma. The large studies we have done on historical samples from the Karolinska Institute are all on serum. We did a lot of work, however, where we collected blood samples, prepared both serum and plasma from the same sample, and then measured the levels of the stress proteins. There was no difference. As a default, we would always analyse plasma though. Gupta: It has been reported that Hsp60/70 in an isolated form can have effects on different types of cells. We are talking about presence in the extracellular fluid. Are the concentrations of these proteins used to produce this effect in the physiological range? Pockley: If we look at the upper levels of Hsp60 and 70 in the human circulation, they are consistent with what has been shown to have an effect on cultured monocytes and dendritic cells (DCs). One of the issues that needs to be considered is the potential influence of soluble immune complexes in the circulation. There is evidence from other systems that putting soluble immune complexes on DCs in culture can have anti-inflammatory consequences—by inducing the secretion of interleukin 10. Calderwood: The concentrations that we normally use are mid-nanomolar, which is easily achievable in vivo. References Currie RW, White FP 1983 Characterization of the synthesis and accumulation of a 71-kilodalton protein induced in rat tissues after hyperthermia. Can J Biochem Cell Biol 61:438–446 Fabian TK, Gaspar J, Fejerdy L et al 2003 Hsp70 is present in human saliva. Med Sci Monit 9: BR62–65 Fabian TK, Toth Z, Fejerdy L, Kaan B, Csermely P, Fejerdy P 2004 Photo-acoustic stimulation increases the amount of 70 kDa heat shock protein (Hsp70) in human whole saliva. A pilot study. Int J Psychophysiol 52:211–216 Fabian TK, Fejerdy P, Nguyen MT, Soti C, Csermely P 2007 Potential immunological functions of salivary Hsp70 in mucosal and periodontal defense mechanisms. Arch Immunol Ther Exp (Warsz) 55:91–98 Giraldo P, Neuer A, Ribeiro-Filho A, Linhares I, Witkin SS 1999 Detection of the human 70kD and 60-kD heat shock proteins in the vagina: relation to microbial flora, vaginal pH, and method of contraception. Infect Dis Obstet Gynecol 7:23–25 House SD, Guidon PT Jr, Perdrizet G et al 2001 Effects of heat shock, stannous chloride, and gallium nitrate on the rat inflammatory response. Cell Stress Chaperones 6:164–171

140

GENERAL DISCUSSION

Mudaliar AV, Kashyap RS, Purohit HJ, Taori GM, Daginawala HF 2006 Detection of 65 kD heat shock protein in cerebrospinal fluid of tuberculous meningitis patients. BMC Neurol 6:34 Mueller T, Bidmon B, Pichler P et al 2003 Urinary heat shock protein-72 excretion in clinical and experimental renal ischemia. Pediatr Nephrol 18:97–99 White FP 1981 Protein and RNA synthesis in cerebral microvessels: a radioautographic study. Brain Res 229:43–52

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Cell stress proteins as modulators of bacteria–host interactions Brian Henderson Division of Microbial Diseases, UCL Eastman Dental Institute, University College London, 256 Gray’s Inn Road, London WC1X 8LD, UK

Abstract. Multicellular organisms exist in dynamic equilibrium with bacterial populations, either in the form of the microbiota of the organism or as pathogens. A challenge to 21st century systems biology is to determine the networks of interactions that exist between the prokaryotic and eukaryotic components of multicellular organisms. Bacterial colonization is stressful for both the prokaryotic and eukaryotic components of an organism and it is emerging that the cell stress proteins (CSPs) of both bacteria and host play roles in the interaction between both Kingdoms. In addition to acting intracellularly to fold proteins, it is being established that CSPs have a wide range of moonlighting functions that are relevant to controlling bacterial colonization. Thus host CSPs can act as cell surface receptors to bind bacteria or to respond to their components. Host CSPs are also secreted into the extracellular fluid where they modulate leukocyte function, potentially to activate antibacterial defences. Bacteria, in turn, have evolved CSPs with adhesive properties for the host. Bacterial CSPs can also modulate host leukocyte function and can induce cellular apoptosis. In insects, endosymbiotic bacteria provide bacterial CSPs which are utilised by the host. Bacterial CSPs have also been shown to be antibacterial targets. These fi ndings establish a range of roles for CSPs in bacteria–host interactions. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 141–159

Bacteria are generally perceived as purveyors of disease, to be killed as rapidly as possible. However, the major life-form on our planet is the bacterium, which can be found in every part of the planet from the deep rocks in the Earth’s crust to the top of the atmosphere (McFall-Ngai et al 2005). Moreover, it is now established that all multicellular organisms are colonized by bacteria. Such colonization generally takes two forms—binary or consortial. In the former, the host organism (generally invertebrate) associates with a single bacterium, which can exist either extracellularly or intracellularly. Consortial colonization refers to the association of the host (generally vertebrate) with large numbers of individual bacterial species (McFall-Ngai et al 2005). Homo sapiens are likely to be the acme of consortial 141

142

HENDERSON

colonisation. It is estimated that humans are colonized by well over 1000 bacterial species, termed commensal bacteria (Wilson 2005). Surprisingly, if we count all the cells in an average human, 90% are bacteria. Of note, only around 40 bacterial species routinely cause disease. The study of the interactions between bacteria and their hosts has evolved into a new branch of science, termed Cellular Microbiology (Henderson et al 1999) and it is becoming clear that such interactions are many and complex. Colonization by pathogenic or commensal bacteria is stressful for both the host and the bacteria and both life-forms must maintain some form of dynamic balance which, if disturbed, can lead to the demise or damage of one or the other. Indeed, many bacterial diseases are caused by organisms that normally live in harmony with their host until some disturbance changes the equilibrium (Henderson et al 1999). This chapter continues the theme of this book on the signalling actions of chaperones and stress proteins and proposes the hypothesis that these proteins function as homeostatic signals both for bacteria and multicellular organisms. Furthermore, as prokaryotes and eukaryotes share many stress protein homologues (Table 1) it is further hypothesized that these homologues can interact with each other’s signalling actions. This chapter will explore the many biological

TABLE 1

Homologous cell stress proteins of prokaryotes and eukaryotes Alternative names

Family

Prokaryotes

Eukaryotes

Chaperonins

GroEL, GroES, Cpn10/60 Trx, DsbA to DsbE, glutaredoxin

Hsp60, Tric, CCT

Thioredoxin Small Hsps

IbpA, IbpB

Peptidyl-prolyl isomerases GrpE Hsp40/DnaJ Hsp70

Cyclophilins, FKBPs*, parvulins GrpE DnaJ, CbpA, RcsG DnaK, Hsc66

Hsp90 Hsp100 Prefoldin

HtpG ClpA,B,C,X,Y Prefoldin

*FKBPs—FK506-binding proteins; NAC—nascent chain associated complex.

Trx, glutaredoxin, protein disulphide isomerise (PDI) α -crystallins, Hsp25/27 Cyclophilins, FKBPs, parvulins GrpE Hsp40, Hdj2, Mtj1 Many e.g. Hsp70, Hsc70, BiP, mitochondrial Hsp70, etc. Hsp90, Gp96 Hsp100 NAC*

CELL STRESS PROTEINS IN BACTERIAL INFECTION

143

actions that bacterial and human molecular chaperones and cell stress proteins have and their known and potential interactions, with the consequences for bacteria–host interactions. To cut down on terminology and word usage, molecular chaperones and cell stress proteins will be lumped together in this chapter under the term cell stress proteins (CSPs) although it is realised that with some proteins this terminology is incorrect.

Cell stress and protein moonlighting A theme underlying this symposium, and an explicit focus of this chapter, is the concept of protein moonlighting (Jeffery 2005). The one-gene-one-protein-onefunction paradigm has had to give way to the growing number of examples of proteins which exhibit multiple functions. Such proteins are termed moonlighting proteins—moonlighting meaning having more than one job. Many proteins, including almost all of the glycolytic enzymes, and many in the Krebs cycle, have multiple functions. For example, the glycolytic enzyme, phosphoglucoisomerase, functions as three separate cytokines and as an implantation factor (Jeffery 2005). Surprisingly, there is no mention of CSPs in the literature on moonlighting proteins, which is strange given the fact that CSPs were shown to have cell signalling functions as early as the late 1980s, around the time that the first examples of protein moonlighting appeared. In a recent review, Connie Jeffery, the pioneer in this subject, has suggested the benefits of the evolution of protein moonlighting (Jeffery 2005). These include: (i) co-ordination of protein pathways (metabolic, signalling etc); (ii) optimising the switching between protein pathways; and (iii) optimising feedback control systems (Jeffery 2005). The various moonlighting functions of selected CSPs are provided in Table 2.

Moonlighting functions of mammalian host CSPs Although the terminology had not been invented at the time, the first moonlighting CSP was human chaperonin 10 (Hsp10) identified in the circulation in early pregnancy and so termed early pregnancy factor (EPF) (Morton et al 1977). This started a trend for finding both new functions for CSPs and their presence on the cell surface or in extracellular fluid (Table 3). Surprisingly, as other chapters delineate, mammalian CSPs are able to activate or inhibit myeloid cells, lymphocytes and vascular endothelial cells and thus function like pro- or anti-inflammatory cytokines. More surprisingly, a number of these CSPs activate cells by binding to receptors normally used to detect bacteria and their products (reviewed in Henderson et al 2006).

144 TABLE 2

HENDERSON Moonlighting functions of selected eukaryotic cell stress proteins

Protein

Moonlighting functions

Hsp10

early pregnancy factor, immunosuppressive/anti-inflammatory protein in vitro and in vivo induces alternative activation state in human macrophages demonstrates cytokine and novel chemokine activities cytokine and chemokine anti-inflammatory effects on monocytes (induces IL-10 synthesis) induces alternative macrophage activation state myeloid cell activator but inhibitory to lymphocytes, sperm capacitation factor, insect toxin, receptor for various bacteria myeloid cell activator, co-receptor for LPS and viruses anti-inflammatory protein in vitro and in vivo, receptor for Dengue virus, receptor for α2-macroglobulin receptor for OmpA protein of E. coli myeloid cell activator, co-receptor for LPS and viruses

Thioredoxin (Trx)80 Thioredoxin PPI Hsp27 Peroxiredoxin Hsp60 Hsp70 BiP Gp96 Hsp90

TABLE 3

Cell surface and secreted host CSPs

Hsp10 Thioredoxin (Trx)80 Thioredoxin Thioredoxin reductase Glutaredoxin TRANK* Peroxiredoxin Cyclophilin A (A PPI) Cyclophilin B Hsp27 Protein disulphide isomerase Calnexin Calreticulin Hsp54 Hsp60 Hsp70 Hsp72 Grp74 BiP Hsp90 Gp96 Clusterin * TRANK—thioredoxin peroxidase-related activator of NF-κB.

CELL STRESS PROTEINS IN BACTERIAL INFECTION

145

Host CSPs as receptors for pathogens Not only do host CSPs act as agonists, able to modulate cell function, but there is increasing evidence that they also behave as receptors able to bind a wide variety of host, bacterial, parasite and viral proteins (Table 4). Kathy Triantafi lou has described her work identifying the cell surface receptor for the major proinflammatory Gram-negative cell wall component, lipopolysaccharide (LPS), as containing, among other proteins, Hsp70 and Hsp90 (Triantafi lou et al 2001). The combination of Hsp70 and Hsp90 also functions as a receptor for Dengue virus (Reyes-del Valle et al 2005). Hsp90 α is a receptor for the Campylobacter jejuni surface lipoprotein JlpA and binding results in activation of NF-κ B and MAP kinases (Jin et al 2003). Another HSP family member, Gp96 is a receptor for the OmpA protein of E. coli and is involved in the invasion of this bacterium into cells (Prasadarao et al 2003). Human chaperonin (Hsp)60 has been reported to be on the surface of a variety of cells and is also able to bind LPS (Habich et al 2005) and acts as a cell surface receptor for Staphylococcus aureus (Dziewanowska et al 2000) and Listeria monocytogenes (Wampler et al 2004) by binding to the fibronectin receptor and Listeria adhesion protein respectively. This latter protein has recently been identified as alcohol acetaldehyde dehydrogenase, which binds to Hsp60 with a high affinity (ka = 5.4 × 10−8 M) (Kim et al 2006). This is the first example of a bacterial moonlighting protein interacting with a human moonlighting protein. Protein disulphide isomerase has recently been found to be required for the adhesion and invasion of mammalian cells by the obligate intracellular bacterium Chlamydia (Conant & Stevens 2007).

TABLE 4

Host CSPs function as receptors for self and pathogen proteins

Host CSP

Binding ligands

Hsp60

HIV glycoprotein gp41, α 3 β1 integrin, high density lipopoprotein, LPS, S. aureus FNBP*, alcohol acetaldehyde dehydrogenase of L. monocytogenes C. trachomatis, diphtheria toxin

Protein disulphide isomerase Hsp70 BiP Gp96 Hsp90

LPS, Dengue virus Dengue virus, α2-macroglobulin OmpA of E. coli LPS, taxol, C. jejuni JlpA, components of H. pylori, bacterial toxins, Dengue virus

* FNBP—fibronectin binding protein.

146

HENDERSON

Circulating host CSPs: a new form of endocrine control? Increasing numbers of CSPs are being detected in human blood. To a large extent, such reports have been ignored because of the apparent lack of a mechanism for releasing these stress proteins. This largely reflects the ignorance of the fact that there are now many non-classical pathways of protein secretion known to be used by eukaryotic and prokaryotic cells. As an example, it has been reported that human Hsp70 can be released via 50–100 µm cell vesicles known as exosomes (Lancaster & Febbraio 2005) and by a pathway similar to that utilized by interleukin (IL)1, which involves lysosomal endosomes (Mambula & Calderwood 2006). It is expected that similar mechanisms will be discovered for the release of the other CSPs found in human blood (Table 3). Most protein agonists bind to a single receptor. However, the evidence to date suggests that CSPs bind cells through a diversity of receptors. To some extent this is a reflection of the poorly comprehended fact that the same CSP from different species may have very different actions. Thus the author’s introduction to CSPs was the finding that a potent bone resorbing protein released by an oral bacterium was the Hsp60 protein of this organism (Kirby et al 1995). The E. coli protein also stimulated bone resorption but the homologous proteins from Mycobacterium tuberculosis had no bone resorbing activity (Kirby et al 1995), and indeed the Hsp60.1 protein from the latter organism actually inhibits bone resorption (Henderson & Coates, unpublished). Having said this, the literature on Hsp70 binding reports only on the human or the M. tuberculosis protein. To date, it has been found that the human protein binds to CD14, TLR2/4, CD40, CD91 and LOX1. The M. tuberculosis protein has been reported to bind to CD40 and, more recently, to the HIV1 co-receptor CCR5 (see Whittall et al 2006). The true biological function of the binding of Hsp70 to these various receptors still awaits elucidation. A review of the properties of secreted CSPs is provided by Graham Pockley. The author has concentrated on measuring Hsp60, BiP and Hsp10 in various clinical samples. Using participants in the Whitehall II study (a prospective study of cardiovascular risk factors in British civil servants) we have found that levels of Hsp60 in human plasma range from below assay detection to mg/ml levels. No other circulating protein shows such an enormous range. The circulating protein is remarkably stable and exists largely as the intact protein minus the mitochondrial import peptide, signifying that this protein emanates from the mitochondrion. In a population of periodontal patients bled for 30 days after dental treatment, which involves removal of the bacteria causing disease, there were absolutely no changes in Hsp60 levels in the blood (Lewthwaite et al 2002, Shamaei-Tousi et al 2006, Coates & Henderson, unpublished). Curiously, levels of Hsp60 in the Whitehall cohort correlate with measures of psychological distress. Our general finding is that levels of Hsp60 in plasma relate to measures of vas-

CELL STRESS PROTEINS IN BACTERIAL INFECTION

147

cular dysfunction or to cardiovascular pathology (Halcox et al 2005, Shamaei-Tousi et al 2006). As discussed by Panayi (this volume), Hsp10 and BiP have been found to have anti-inflammatory actions and both proteins are being tested for their therapeutic effects in rheumatoid arthritis. The author’s group has just finished measuring Hsp10, Hsp60 and BiP levels in the blood of periodontal patients undergoing therapy. This is a fascinating human model which treats bacterial-driven inflammation in the mouth by mechanical debridement to remove bacteria. The consequence is an acute systemic inflammatory response with a measurable acute phase response due to the released bacteria. What is fascinating is that in this model there were: (i) no changes in Hsp60 levels; (ii) a rapid increase in Hsp10 levels to those present in the control population; and (iii) a smaller increase in BiP concentrations. Thus bacterially induced inflammation had markedly decreased the levels of circulating Hsp10. As this protein has anti-inflammatory actions the decreased levels may imply increased usage (binding, uptake, etc.) or a decreased synthesis or, possibly, elevated excretion. These findings, and those from other groups, are still preliminary, but in a world busy developing systems biology it would be sensible to view circulating CSPs as another level of homeostatic control, possibly reflecting and integrating the state of the internal milieu of relevant cell populations and acting to control systems involved in managing the inflammatory response to the normal microbiota and infectious agents (Fig. 1).

Bacterial CSPs The stress of infection We know an enormous amount about the response of bacteria to certain stresses, such as elevated temperature. However, much less is known about the role of the stress response in bacteria colonizing their hosts. The first stage in colonization generally involves contact with epithelial cells. Studies of the oral Gram-negative organism Porphyromonas gingivalis cultured on a human epithelial cell line revealed significant increases in mRNA for Hsp40, Hsp60, Hsp70, Hsp90, various PPIs and thioredoxin (Hosogi & Duncan 2005). The interaction of Neisseria gonorrhoeae with epithelial cells induces rpoH, a homologue of σ32 , whose regulon includes Hsp60 and Hsp10. Construction of a mutant conditionally expressing rpoH revealed that this regulatory gene is required for the invasion of the epithelial cells by the bacterium (Du et al 2005). Once past the epithelial barrier, bacteria interact with host immune factors and cells. Some bacteria, such as M. tuberculosis, actually live within the defending macrophages, and it has been reported that the Hsp20 gene, acr2, in this organism increases 800-fold when the bacterium invades the

148

HENDERSON endocrine signalling

paracrine signalling

stressor autocrine signalling?

etion sec r Ps S C of

upregulation of CSPs

r llula ace intr nalling sig

circulating CSPs ‘endocrine’ CSPs?

or t ion ibit em en ks h n i nc twor a h en s t n e ho of

CSPs interacting with and modulating bacterial behaviour

secreted bacterial CSPs

virus/bacterium, or components of, bind to cell surface CSPs

INTRACRINE FIG. 1.

b c o a c te m m ri u en m sa l

CSPs acting as receptors

ba p a c te r th o ium ge n

r llula ac e ling r t n i al sign

viruses

PARACRINE

ENDOCRINE

The potential network of interactions between prokaryotic and eukaryotic CSPs.

macrophage (Schnappinger et al 2003). Obviously, much more work is required to define how bacteria respond to contact with their host species. Bacterial CSPs as host cell modulating agonists In other chapters in this volume, the ability of human CSPs to modulate leukocyte function is described. Most workers have studied the effect of CSPs on monocytes/macrophages and in recent years a new hypothesis has developed concerning the activation of these cells. It is now conceived that macrophages are phenotypically plastic and can enter into a range of ‘activation’ states. Response to bacterial LPS or γ -interferon induces what is termed the classically activated (ca) macrophage able to ingest and kill pathogens. Exposure to IL4, IL10, IL13 or immune complexes/TLR4 agonist induces alternative states of activation (aamacrophages) in which IL10 or specific chemokine networks are produced (Noel

CELL STRESS PROTEINS IN BACTERIAL INFECTION

149

TABLE 5 Macrophage activation states induced by eukaryotic CSPs Hsp10 Thioredoxin (Trx)80 Trx Cyclophilin A Peroxiredoxin* Hsp27 Hsp60 Hsp70 BiP Gp96 Clusterin

Human protein inhibits caMØ activation, Induces novel macrophage activation state Activates ca-like macrophage state Induces caMØ state Induces aaMØ Induces IL-10 synthesis but not TNFα (similar to aa macrophage) Most proteins induce a caMØ-like state Induces caMØ state Inhibits monocyte activation and may induce aaMØ Modified form of macrophage activation Induces what appears to be caMØ

caMØ, classically activated macrophage; aaMØ, alternatively activated macrophage. * Protein is from Fasciola hepatica. Note: the terms classically and alternatively activated imply multiple changes in cell phenotypes. In studies with CSPs only a few activation state changes are measured and therefore it is difficult to state accurately that any one CSP induces a pure ca versus a pure aa state. Indeed some CSPs may induce completely novel macrophage activation states.

et al 2004). Studies of human CSPs can be interpreted largely in terms of the induction of alternative macrophage activation states with proteins such as Hsp10, Hsp27 and BiP clearly inducing an IL10 response in stimulated macrophages (Table 5). Unfortunately, there have been no comparative studies of human and bacterial CSPs in this respect, so comparisons are difficult. Bacterial Hsp60 proteins can exhibit similar, or distinct, responses to the human protein and the one comparative study reported revealed that the human protein was significantly less active as a stimulator of human monocytes and vascular endothelial cells than were bacterial proteins (Maguire et al 2005). The only other bacterial/human comparison has been with Hsp70. As reviewed earlier, the human protein binds to a range of receptors (CD14, TL2/4, CD40, LOX1) while the mycobacterial equivalent is reported to bind only to CD40 and CCR5. Moreover, the binding site for CD40 on the human and mycobacterial protein is different (see Whittall et al 2006). This is an obvious area for future comparative research. Bacterial CSPs as virulence factors Any bacterial protein that enhances bacterial colonization is termed a virulence factor. Of course, as most of the bacteria colonizing multicellular organisms are harmless commensals, the term ‘virulence’ is incorrect and some other term like ‘colonization factor’ may be more appropriate. There are now a substantial number of reports of bacteria with CSPs either associated with the cell surface or secreted into the extracellular milieu (Table 6). A growing number of these CSPs appear to act as adhesins. Such proteins allow bacteria to bind to host cells or matrix and

150

HENDERSON

TABLE 6 Bacterial cell surface CSPs and those acting as adhesins Bacterium

Molecular chaperone

Host ligand

Biological activity

Aa* Borrelia burgdorferi Clostridium difficile Coxiella burnetti EPEC Haemophilus ducreyi Helicobacter pylori Helicobacter pylori Helicobacter pylori Legionella pneumophila Mycobacterium avium Mycobacterium avium Mycobacterium smegmatis Rickettsia prowazekii Salmonella typhimurium

Cpn60 Cpn60 Cpn60 Hsp70 Hsp70 Cpn60 Hsp20 Cpn60 Hsp70 Cpn60, Hsp70 Cpn60 Hsp70 Cpn60 PPI** Cpn60

? glycosphingolipid ? ? sulphogalactosylceramide ? ? ? sulphatides ?

not yet defi ned adhesin adhesin not yet defi ned adhesin adhesin not yet defi ned adhesin adhesin adhesion

? ? ? ? mucus

adhesion possible adhesion not yet defi ned not yet defi ned adhesin

* Aa, Actinobacillus actinomycetemcomitans. ** PPI, peptidyl prolyl isomerase.

are absolutely required for successful colonization. Thus we have a curious mirroring of the biological use to which eukaryotes and prokaryotes utilize CSPs. As reviewed, eukaryotic CSPs like Hsp60, Hsp70, BiP, Gp96 and Hsp90 can function to bind to bacteria or to their specific components. Likewise, bacteria can use their CSPs to facilitate binding to host components. Unfortunately, we are largely ignorant of the host ligands to which these CSPs bind. Those that have been identified are low molecular mass components of the cell membrane. However, it is expected that bacterial CSPs will be found that bind with the high affinity that the Listeria alcohol acetaldehyde dehydrogenase binds to human Hsp60 (Kim et al 2006). In addition to the CSPs already discussed (Hsp10, Hsp60, Hsp70, Hsp90), and their effects on immune regulation, there is emerging evidence for a role of other bacterial CSPs in pathogenesis. Bacteria express a number of peptidyl prolyl isomerases (PPIs), and the first evidence for their role in bacterial virulence was the identification that the macrophage infectivity promoter/potentiator (Mip) protein of Legionella pneumophila, which enhances survival of these bacteria within macrophages, was a PPI (Fischer et al 1992). Trigger factor (TF) is a key bacterial PPI, and the first molecular chaperone that bacterial nascent polypeptide chains meet. Surprisingly, the gene encoding this protein in E. coli can be inactivated with minimal phenotype (Huang et al 2000). Inactivation of the TF gene in the dental pathogen, Streptocooccus mutans, well known for forming biofi lms on tooth surfaces, influenced responses to stress, decreased competence and inhibited the ability of

CELL STRESS PROTEINS IN BACTERIAL INFECTION

151

the bacterium to form biofilms (Wen et al 2005). Perhaps one of the most intriguing roles for a bacterial PPI is HP0175 of the gastric ulcer-causing bacterium, Helicobacter pylori. This is a major immunogen of the bacterium and also a secreted protein able to induce apoptosis of gastric epithelial cells (Basak et al 2005). Other bacterial CSPs that have been shown to contribute to bacterial virulence include members of the Hsp100 family, DnaK/DnaJ and the α -crystallins. Space precludes their discussion, but this information can be found in Henderson et al (2006). Bacterial CSPs as therapeutic targets Stress proteins represent an Achilles’ heel of all living forms. Some human antibacterial peptides bind specifically to bacterial DnaK and inhibit activity, in the process killing the bacterium (Cudic et al 2002). The most promising avenue in this field is the discovery of acyldepsipeptides produced by Streptococcus hawaiiensis, which bind to the Hsp100 protein ClpP leading to uncontrolled proteolysis in the bacterium and cell death. Synthetic depsipeptides have been produced that kill Gram-positive bacteria both in vitro and in vivo (Hinzen et al 2006). CSPs and bacterial symbiosis The human genome project, hailed as a great advance, is in fact, just the beginning of our understanding of the genes in the human body. This project needs to encompass the 1000 plus genomes of our commensal bacteria. The role of CSPs in controlling microbiota–host interactions can be glimpsed from some recent studies of the endosymbiotic bacteria of insects. These obligate insect symbionts are vertically transmitted and it is proposed that Hsp60, which is generally produced at very high levels in these bacteria, has evolved to limit the effects of deleterious genetic variation and loss of ‘protein fitness’ (Ali Fares et al 2004). In addition, evidence exists that the Hsp60 protein of such endosymbiotic bacteria is released in large quantities into the insect extracellular fluid (Haines et al 2002). In certain plant-sucking insects, the endosymbiont Hsp60 protein (termed symbionin) is involved in the transmission of viruses to the plant. In addition, in the mustard aphid, the Hsp60 protein of the endosymbiont is found on the outer face of the gut brush border membrane where it also acts as a receptor for a mannose binding lectin (Bannerjee et al 2004). The most fascinating role for an endosymbiotic Hsp60 protein is in the Antlion. This insect was known to paralyse its prey using a salivary toxin. When the toxin was isolated and sequenced it turned out to be the Hsp60 protein of an endosymbiotic bacterium, Enterobacter aerogenes. The sequence differed from the E. coli GroEL by only 11 residues and single amino acid substitutions in GroEL could turn this non-toxic protein into a potent insect toxin (Yoshida et al 2001).

152

HENDERSON

Summary A new paradigm is emerging in which CSPs have correlated intra- and extracellular functions. For multicellular organisms it is hypothesized that secreted CSPs form an, as yet unrecognized, layer of homeostatic control by acting as stress controllable cellular signals. It is assumed that bacteria also employ CSPs for homeostatic regulation. When networks of host and bacterial CSPs interact the result will depend on the nature of the prokaryotic organism. With pathogens the bacterial CSPs may interfere with normal host CSP signalling networks. With commensal bacteria their CSPs are more likely to interact in a positive sense with the CSP networks of the host. Unravelling these host–bacterial CSP interactions will be a major biological challenge for the 21st century.

References Ali Fares M, Moya A, Barrio E 2004 GroEL and the maintenance of bacterial endosymbiosis. Trends Genet 20:413–416 Banerjee S, Hess D, Mujumder P, Roy D, Das S 2004 The interactions of Allium sativum agglutinin with a chaperonin group of unique receptor protein isolated from a bacterial endosymbiont of the mustard aphid. J Biol Chem 279:23781–23789 Basak C, Pathak SK, Bhattacharyya A, Pathak S, Basu J, Kundu M 2005 The secreted peptidyl prolyl cis,trans-isomerase HP0175 of Helicobacter pylori induces apoptosis of gastric epithelial cells in a TLR4- and apoptosis signal-regulating kinase 1-dependent manner. J Immunol 174:5672–5680 Conant CG, Stephens RS 2007 Chlamydia attachment to mammalian cells requires protein disulphide isomerase. Cell Microbiol 9:222–232 Cudic M, Condie BA, Weiner DJ et al 2002 Development of novel antibacterial peptides that kill resistant isolates. Peptides 23:2071–2083 Du Y, Lenz J, Arvidson CG 2005 Global gene expression and the role of sigma factors in Neisseria gonorrhoeae in interactions with epithelial cells. Infect Immun 73:4834–4845 Dziewanowska K, Carson AR, Patti JM, Deobald CF, Bayles KW, Bohach GA 2000 Staphylococcal fibronectin binding protein interacts with heat shock protein 60 and integrins: role in internalisation by epithelial cells. Infect Immun 68:6321–6328 Fischer G, Bang H, Ludwig B, Mann K, Hacker J 1992 Mip protein of Legionella pneumophila exhibits peptidyl-prolyl-cis/trans isomerase (PPIases) activity. Mol Microbiol 6:1375–1383 Habich C, Kempe K, van der Zee R et al 2005 Heat shock protein 60: specific binding of lipopolysaccharide. J Immunol 174:1298–1305 Halcox JP, Deanfield J, Shamaei-Tousi A et al 2005 Circulating human heat shock protein 60 in the blood of healthy teenagers: a novel determinant of endothelial dysfunction and early vascular injury? Arterioscler Thromb Vasc Biol 25:e141–e142 Haines LR, Haddow JD, Aksoy S, Gooding RH, Pearson TW 2002 The major protein in the midgut of teneral Glossina morsitans morsitans is a molecular chaperone from the endosymbiotic bacterium Wigglesworthia glossinida. Insect Biochem Mol Biol 32:1429–1438 Henderson B, Wilson M, McNab R, Lax AJ 1999 Cellular microbiology; bacteria-host interactions in health and disease. Wiley, Chichester Henderson B, Allan E, Coates ARM 2006 Stress wars: the direct role of host and bacterial molecular chaperones in bacterial infection. Infect Immun 74:3693–3706

CELL STRESS PROTEINS IN BACTERIAL INFECTION

153

Hinzen B, Raddatz S, Paulsen H et al 2006 Medicinal chemistry optimization of acyldepsipeptides of the enopeptin class antibiotics. Chem Med Chem 1:689–693 Hosogi Y, Duncan MJ 2005 Gene expression in Porphyromonas gingivalis after contact with human epithelial cells. Infect Immun 73:2327–2335 Huang GC, Li ZY, Zhou JM, Fischer G 2000 Assisted folding of D-glyceraldehyde-3-phosphate dehydrogenase by trigger factor. protein Sci 9:1254–1261 Jeffery CJ 2005 Moonlighting proteins: proteins with multiple functions. In: Henderson B, Pockley AG (eds) Molecular chaperones and cell signalling. Cambridge University Press, p 61–77 Jin S, Song YC, Emili A, Sherman PM, Loong Chan V 2003 JlpA of Campylobacter jejuni interacts with surface-exposed heat shock protein 90 α and triggers signalling pathways leading to the activation of NF-κ B and p38 MAP kinases in epithelial cells. Cell Microbiol 5:165–174 Kim K-P, Jagadeesan B, Burkholder KM et al 2006 Adhesion characteristics of Listeria adhesion protein (LAP)-expressing Escherichia coli to Caco-2 cells and of recombinant LAP to eukaryotic receptor Hsp60 as examined in a surface plasmon resonance sensor. FEMS Microbiol Lett 256:324–332 Kirby AC, Meghji S, Nair SP et al 1995 The potent bone resorbing mediator of Actinobacillus actinomycetemcomitans is homologous to the molecular chaperone GroEL. J Clin Invest 96:1185–1194 Lancaster GI, Febbraio MA 2005 Exosome-dependent trafficking of HSP70. J Biol Chem 280:23349–23355 Lewthwaite J, Owen N, Coates A, Henderson B, Steptoe A 2002 Circulating heat shock protein (Hsp)60 in the plasma of British civil servants: relationship to physiological and psychosocial stress. Circulation 106:196–201 Maguire M, Poole S, Coates ARM, Tormay P, Wheeler-Jones C, Henderson B 2005 Comparative cell signalling activity of ultrapure recombinant chaperonin 60 proteins from prokaryotes and eukaryotes. Immunology 115:231–238 Mambula SS, Calderwood SK 2006 Heat shock protein 70 is secreted from tumor cells by a non-classical pathway involving lysosomal endosomes. J Immunol 177:7849–7857 McFall-Ngai M, Henderson B, Ruby N 2005 The influence of cooperative bacteria on animal host biology. Cambridge University Press Morton H, Rolfe B, Clunie GJ 1977 An early pregnancy factor detected in human serum by the rosette inhibition test. Lancet 1:394–397 Noel W, Raes G, Hassanzadeh G, De Baetselier P, Beschin A 2004 Alternatively activated macrophages during parasite infections. Trends Parasitol 20:126–133 Prasadarao NV, Srivastava PK, Rudrabhatla RS, Kim KS, Huang S, Sukamaran SK 2003 Cloning and expression of the Escherichia coli K1 outer membrane protein A receptor, a gp96 homologue. Infect Immun 71:16980–16988 Reyes-del Valle J, Chavez-Salinas S, Medina F, Del Angel RM 2005 Heat shock protein 90 and heat shock protein 70 are components of Dengue virus receptor complex in human cells. J Virol 79:4557–4567 Shamaei-Tousi A, Stephens JW, Bin R et al 2006 Association between plasma levels of heat shock protein 60 and cardiovascular disease in patients with diabetes mellitus. Eur Heart J 27:1565–70 Schnappinger D, Ehrt S, Voskuil MI et al 2003 Transcriptional adaptation of Mycobacterium tuberculosis within macrophages: insights into the phagosomal environment. J Exp Med 198:693–704 Triantafi lou K, Triantafi lou M, Dedrick RL 2001 A CD14-independent LPS receptor cluster. Nat Immunol 2:338–345 Wampler JL, Kim K-P, Jaradat Z, Bhunia AK 2004 Heat shock protein 60 acts as a receptor for the Listeria adhesion protein in Caco-2 cells. Infect Immun 72:931–936

154

HENDERSON

Wen ZT, Suntharaligham P, Cvitkovitch DC, Burne RA 2005 Trigger factor in Streptococcus mutans is involved in stress tolerance, competence development, and biofi lm formation. Infect Immun 73:219–225 Whittall T, Wang Y, Younson J et al 2006 Interaction between the CCR5 chemokine receptors and microbial HSP70. Eur J Immunol 36:2304–2314 Wilson M 2005 Microbial inhabitants of humans: their ecology and role in heath and disease. Cambridge University Press Yoshida N, Oeda K, Watanabe E et al 2001 Chaperonin turned insect toxin. Nature 411:44

DISCUSSION Ellis: I want to make a couple of points about the reported neurotoxic property of some GroEL mutants. When I see a novel result that sounds interesting but is unexpected, my rule of thumb is to see whether anyone else can repeat it before I accept it. Has anyone repeated this observation? Henderson: We have tried to repeat it but with negative results. But we were using different insects and different recombinant chaperonin 60 proteins as the Japanese group that reported the effect were unwilling to provide their active recombinant proteins. Ellis: Another interesting result was the early pregnancy factor story. This came out of one laboratory in Queensland, but so far as I know has not been reported by other laboratories. I would have thought that any protein that was a marker for human conception within 24 h would be of some interest. Why was this not picked up? So I have to reserve judgment on both these results on the same grounds. I mentioned the neurotoxin work to a microbiological colleague who is the chairman of our safety committee. He became concerned: he said that when they sit down to calculate potential problems with using various micro-organisms in the laboratory, they don’t consider for a moment the possibility that a single amino acid mutation in a well-studied harmless protein can change it into a neurotoxin. So he was sceptical about this report. Lund: I have a question about the polymorphism in the promoter field. Have you looked at expression in these individuals inside the cells? Henderson: These are historical samples. We have the DNA and plasma but not the cells. Coates: The early pregnancy factor John Ellis referred to was the work of Morton and colleagues (Morton et al 1977). Eventually it was sequenced and found to be human chaperonin 10 (Cavanagh & Morton 1994). This was found in a number of experiments to be relatively immunosuppressive (Noonan et al 1979, Harness et al 2003), and a company has now gone into clinical trials using chaperonin 10 as an immunomodulator. Ellis: Is it not of interesting potential medical importance to have a maker in the blood for early conception?

CELL STRESS PROTEINS IN BACTERIAL INFECTION

155

Coates: The problem with the original assay is that it was extremely complicated. This was probably the main reason their work didn’t catch on. Ellis: We get told unusual stories all the time and we have to make some sort of judgement. I try to have an open mind, and I like the unexpected, because novelty can sometimes open up new fields. But if no other researchers report the finding, is it because they tried and failed, or because no one has tried to repeat it? Feder: In general, if you have chaperone signalling between host and symbiont, is this used to stabilize the normal healthy mutualistic relationship, or is this often used to support antagonistic relationships, or a mix of both? Henderson: There are only about three examples. There is an interesting example where the Hsp60 from the endosymbiotic bacteria acts as a receptor for a toxin that would kill the insect. Feder: Let’s talk about human microbial flora. Is Hsp60 stabilizing a mutualistic relationship or an antagonistic one? Henderson: I think there must be a mutualistic relationship between ourselves and our microflora, or we would get sick. There must be a healthy relationship between ourselves and our skin and oral flora. Csermely: A little change may have big implications. One reason suggested for chronic obesity might be that the flora was changing in the gut of these people (Backhed et al 2005). Gupta: You mentioned that the chaperonin 10 assay is complicated and this is why this work has not been repeated, but it should be easy to measure whether chaperonin 10 concentration increases in early pregnancy or not. Has anyone done this? Henderson: No. I don’t know why. Gupta: Chaperonin 10 is identical to the early pregnancy factor, and so I presume that this activity can be measured easily. Henderson: I would imagine that CBio (Australian Biotech Company) who have shown that chaperonin 10 has some minor effect on rheumatoid arthritis, will be looking at this. Gupta: I have antibodies to chaperonin 10 in my fridge. If we had samples from women in early pregnancy we can measure this. Panayi: You never find a woman who is 24 h pregnant! These must all be mouse experiments. Gupta: I thought some of these were human studies. Macario: I heard once that the detection of early pregnancy factor (or early conception factor), which turns out to be a close relative of HSP10, is used in veterinary environments to detect females that might be pregnant (Cordoba et al 2001, Fletcher et al 2001). Now turning back to chaperones and microbiota: in your hypothesis, Dr Henderson, you proposed that chaperones are providing some kind of link between the components of the microbiota and the host. Do

156

HENDERSON

you assume that there has to be more than one species of chaperone in that hypothesis? Henderson: In the infection system, individual molecular chaperones would form part of the virulence network. The more interesting aspect is the normal microflora/microbiota, as most humans survive quite happily with several hundred different bacteria species in their bodies. Coming back to Francisco’s point about recognition of Hsp60, can we recognize all the Hsp60s from all the bacteria in our bodies? Quintana: There was some interesting work by Eli Sercarz in which he induced arthritis in rats who were or were not on antibiotics. He found out that the rats grown with no flora did worse. They did nice peptide mapping of T cell epitopes (Moudgil et al 2001) and showed that the flora were inducing T cell responses to some of the macrobiota bacterial heat shock proteins that were cross-reactive. This works in two ways. On the one hand it strengthens the immune regulation that was controlling the arthritis, and on the other hand the bacteria want to get rid of the immune response. What we tend to be tolerant to is whatever we have in our guts. Henderson: Gnotobiotic animals have almost no immune system. If you repopulate them with the microbiota you get high levels of antibodies and large numbers of reactive T cells. Ellis: I can comment on the question of whether the immune system can see chaperone proteins from different bacteria. When we sequenced the chloroplast chaperonin in 1987, we looked at the protein sequence database, which did not contain GroEL at that time. We found that there were a lot of bacterial proteins in the database that were 50% identical or so with our sequence. We got excited by this, because we thought these proteins might give us a clue about the function of the chloroplast protein, but when we looked at the details we found that all these bacterial proteins had the same name—the common bacterial antigen. So I spoke to Doug Young, an immunologist, and he told us that this powerful antigen was found in all human bacterial infections, but that there was no information about its function in bacteria. This story suggests that the human immune system can generate antibodies against lots of different versions of GroEL. Henderson: Yes, the antibody response you get to the chaperonins is very ‘crossreactive’. GroEL is highly immunogenic and so it is easy to make monoclonals against it. Panayi: 80% of women with rheumatoid arthritis go into complete remission during pregnancy. They all relapse within about three months of delivery. The mechanism is unknown, but it may be due to the up-regulation of the regulatory T cells which have been described in mouse pregnancy. Does chaperonin 10 induce the regulatory cells seen around pregnancy, and if it is not this, what is it?

CELL STRESS PROTEINS IN BACTERIAL INFECTION

157

Henderson: Until we measure Hsp10 in pregnant women we won’t know. Pockley: My PhD was in reproductive immunology. At that time there was a great deal of interest in this fetal–maternal relationship, especially in the human situation where the fetal tissue is bathed with maternal blood. At the other end of the spectrum in cows and horses there isn’t any intimate contact because of the different placental structures. There has always been this query about how such tissue can survive in a potentially immunologically hostile environment. I worked on a protein called placental protein 14, which is produced in human decidua. It peaks within the first 9 weeks of pregnancy, so it is an early pregnancy protein. It is now called glycodelin A in the literature. There was a great deal of scepticism at the time about the concept of immunosuppressive proteins in early pregnancy, but it was about this time that the issue of early pregnancy factor was raised. It was proposed as a potential immunosuppressive molecule: this is the way it was marketed. Because it came at a time when there was a lot of scepticism surrounding immunosuppressive proteins, combined with this issue of the difficulty in measuring it, interest waned. Panayi: Since this is a meeting on the extracellular functions of chaperones, and if Hsp10 is an early pregnancy factor, it could be linked to this phenomenon of the survival of the fetus plus the improvement of rheumatoid arthritis. We had this discussion about symbionts: if you are a symbiont it must be a good relationship. What do we learn from the fact that if we give a broad-spectrum antibiotic to a patient, you kill the symbiont and all sorts of horrible bacteria come in and colonize? Should we think of symbiosis as some kind of cartel that is directed against pathogenic bacteria? Henderson: We are talking about hundreds of bacteria interacting with individual cell populations. I have no idea how it works. Habich: You mentioned that you have also measured Hsp60 levels in diabetic individuals. Which type of diabetes did they have, type 1 or type 2 diabetes? Henderson: This was a cohort of mainly type 2, but there were some type 1s. There seems to be no difference between these two types of diabetes in respect to the Hsp60 levels and tissue pathology. Nixon: Early in your talk you mentioned a number of interacting partners for Hsp60 that you have identified through affinity purification. I was particularly interested in your identification of an ADAM protease because, although we have not had much success with our own affinity purifications, one of the few proteins we have identified that interacts with Hsp60 on the mouse sperm surface is also a member of the ADAM family. It is also of interest that other chaperones such as Hsp90 have also been shown to interact with matrix metalloproteinases on the surface of cancer cells (Eustace & Jay 2004). Henderson: This is encouraging, because I’d given up on this work. We followed up on these studies by setting up cross-linking experiments. After two years of

158

HENDERSON

trying we got nowhere. Val Corrigall is probably in the same situation that I am, looking for the BiP receptor. There must be a receptor, but possibly not on the cell surface. Nixon: Were you actually looking specifically for a receptor, or interacting partners? Henderson: What’s the difference? Csermely: You mentioned that the number of species of colonizing bacteria in humans is increasing. Is this established? Henderson: What I meant is that the number of bacteria being discovered is increasing. As more and more people are analysed this number rises. Csermely: You mentioned that the stable levels of Hsp60 are mainly because of the genetic background of the individuals. But I don’t think that the genetic background can be exclusively responsible for these levels. There are other possibilities, such as prenatal stress. The fetus could be preparing for a life of scarcity of food, for example, because the mother had a low calorie diet. Then after birth the food could have been increased, which could be a big stress. Henderson: The thing that worries us is that there seems to be no responsivity of the circulating Hsp60 levels to exogenous cues. If you are high you are high; if you are low you are low. Csermely: There was a recent paper (Cleal et al 2007) showing that a shift from prenatal to postnatal conditions has good predictive value for cardiac problems later in life. This will not change either, and it is not exclusively genetic. Feder: Without taking away from your suggestion, why do you think that the genetic background ought to be insufficient to demonstrate this? This is the whole basis for personalized medicine. In 1999 Rick Moromoto’s group published on blood samples from human subjects (Wang et al 1999). Via multiplex RT-PCR of HSP mRNAs, he discovered that healthy humans varied enormously in concentrations of HSP mRNAs—and by implication in heat shock proteins and their cognates. Genetic variability among humans is large. Csermely: I think we are talking about two levels of differences. The differences in starting conditions (the genetics and genetic background) and on top of these differences in personal life histories. Feder: These are not mutually exclusive. Panayi: The Southampton group led by David Barker showed that birthweight was a significant determinant of cardiovascular disease later in life. All sorts of mechanisms have been postulated. The question that you have raised could perhaps be solved by people alive today whose birthweight is known. They have serum samples from some of these individuals. Henderson: We just put in a British Heart Foundation grant to look at this, but it wasn’t funded.

CELL STRESS PROTEINS IN BACTERIAL INFECTION

159

References Backhed F, Ley RE, Sonnenburg JL, Peterson DA, Gordon JI 2005 Host-bacterial mutualism in the human intestine. Science 307:1915–1920 Cavanagh AC, Morton H 1994 The purification of early-pregnancy factor to homogeneity from human platelets and identification as chaperonin 10. Eur J Biochem 222:551–560 Cleal JK, Poore KR, Boullin JP et al 2007 Mismatched pre- and postnatal nutrition leads to cardiovascular dysfunction and altered renal function in adulthood. Proc Natl Acad Sci USA 104:9529–9533 Cordoba MC, Sartori R, Fricke PM 2001 Assessment of a commercially available early conception factor (ECF) test for determining pregnancy status of dairy cattle. J Dairy Sci 84:1884–1889 Eustace BK, Jay DG 2004 Extracellular roles for the molecular chaperone, hsp90. Cell Cycle 3:1098–1100 Fletcher BH, Cassady AI, Summers KM, Cavanagh AC 2001 The murine chaperonin 10 gene family contains an intronless, putative gene for early pregnancy factor, Cpn10-rs1. Mamm Genome 12:133–140 Harness J, Cavanagh A, Morton H, McCombe P 2003 A protective effect of early pregnancy factor on experimental autoimmune encephalomyelitis induced in Lewis rats by inoculation with myelin basic protein. J Neurol Sci 216:33–41 Morton H, Rolfe B, Clunie GJ 1977 An early pregnancy factor detected in human serum by the rosette inhibition test. Lancet 1:394–397 Moudgil KD, Kim E, Yun OJ, Chi HH, Brahn E, Sercarz EE 2001 Environmental modulation of autoimmune arthritis involves the spontaneous microbial induction of T cell responses to regulatory determinants within heat shock protein 65. J Immunol 166:4237–4243 Noonan FP, Halliday WJ, Morton H, Clunie GJ 1979 Early pregnancy factor is immunosuppressive. Nature 278:649–651 Wang SM, Khandekar JD, Kaul KL, Winchester DJ, Morimoto RI 1999 A method for the quantitative analysis of human heat shock gene expression using a multiplex RT-PCR assay. Cell Stress Chaperones 4:153–161

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Chaperonin 60 and macrophage activation Anthony R. M. Coates, Ana Cehovin and Yanmin Hu Medical Microbiolog y, Centre for Infection, Department of Cellular and Molecular Medicine, St George’s, University of London, Cranmer Terrace, London SW17 0RE, UK

Abstract. Eukaryotic and prokaryotic chaperonin 60s (Cpn60s) activate macrophages to produce pro-inflammatory cytokines. CD14 and TLR4 have been proposed as potential Cpn receptors. In addition, Cpn60s can block LPS-induced activation. This is a doserelated effect, low concentrations block, and high concentrations activate. This may relate to the ability of Cpn60s to block inflammatory disease. Cpns are multiplex or moon-lighting proteins, with functions as molecular chaperones, in stress survival and as inflammatory modulators. A cpn60.1 knockout mutant does not induce a granulomatous response and cytokine levels, such as tumour necrosis factor are reduced in the tissues. These data suggest that Cpn60.1 may also function as a virulence factor. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 160–172

Although Chaperonin (Cpn) 60s (Hsp60s) are generally thought to be intracellular molecules, it has been known for many years that they activate macophages. Much of the early work in this field was performed on Mycobacterium tuberculosis (Mt) heat shock protein 65 (Friedland et al 1993, Peetermans et al 1994), also called GroEL, and Cpn60.2 (Kong et al 1993). The rationale for presuming that Mt Cpn60.2 exists both inside and outside the cell derives from the observation that it can be detected in the supernatant of liquid cultures (Andersen et al 1991) and in the serum of tuberculosis patients (Sethna et al 1998). This chapter reviews the evidence that Cpn60s from many species activate macrophages. We try to relate this to the pathogenesis of one disease, tuberculosis, in the full knowledge that these remarkable molecules probably play many different roles in numerous diseases. Tuberculosis is still one of the major causes of death in the world, currently epidemic in many developing countries, and accounts for about 8% of all deaths each year (WHO 2007). Usually during childhood, M. tuberculosis is inhaled into the lungs (Fig. 1) within droplets where it is phagocytosed by alveolar macrophages. From the lungs, it can spread into the lymph nodes and the blood, and is seeded in many organs of the body, including the lungs. The 160

CHAPERONIN 60 AND MACROPHAGES

161

FIG. 1. The life cycle of M. tuberculosis. During childhood, M. tuberculosis is inhaled and seeds into the lungs where it is surrounded by cells in a granuloma. Blood-borne and lymphatic spread occurs. Later, sometimes after decades or immunosuppression due to AIDS, overt disease occurs; bacteria increase in numbers leading to cavitation. The lesion breaks into the bronchus and the patient coughs up the bacilli. These are then inhaled by the next person and the cycle is repeated.

organism is isolated by macrophages and other cells in a granuloma in which the bacteria are surrounded by cells. Presumably this helps to prevent further spread, whilst at the same time enabling macrophages, in concert with T lymphocytes, to suppress the growth of the bacterium (Russell 2007). In most people, the acquisition of the infection is symptomless, and the organism lies dormant for the remainder of the person’s life. In 5–10% of infected people, overt disease occurs, often associated with failing immunity due to, for example, advancing years or AIDS. Tuberculosis of the lungs (the commonest site) is associated with the appearance of bacteria in the sputum, and coughing, which disseminates droplets laden with organisms into the air, where they are inhaled by uninfected people, thus spreading the disease. Macrophage activation is a key part of this infectious cycle, and suppresses the growth of the bacteria. Chaperonin 60s activate macrophages Activation of macrophages occurs not only with MtCpn60.2, but also with MtCpn60.1 (Lewthwaite et al 2001, Tormay et al 2005) and Cpn60s from other

162

COATES ET AL

species. These include Escherichia coli, Helicobacter pylori, Chlamydia pneumoniae, Rhizobium leguminosarum and human (Chen et al 1999, Lewthwaite et al 2002, Ueki et al 2002, Zheng et al 2004, Maguire et al 2005). Cytokines such as tumour necrosis factor alpha (TNFα ), interleukin 1β (IL1β ), IL4, IL6, IL8, IL10, IL12, gamma interferon (IFNγ ) and granulocyte–macrophage colony-stimulating factor (GMCSF) are up-regulated in response to Cpn60s (Lewthwaite et al 2001). IFNγ activated macrophages suppress the growth of M. tuberculosis in macrophages (Kaufmann 2001). In the context of an M. tuberculosis infection, it makes sense for the immune system to suppress bacterial growth, because it prevents overt disease. However, it is also advantageous for the bacterium, because it triggers entry into the dormant phase, an essential part of the bacterial life cycle. This prolongs the life of the bacterium, which waits until the immune system begins to fade, for instance, in old age. When microbial growth takes off, abscesses form and break into the airways, the patient coughs up bacteria laden droplets, and the cycle of transmission begins once again (Fig. 1).

Potential chaperonin 60 receptors An array of receptors for molecular chaperones exists on myeloid and lymphoid cells. Cpn60s from different species use different receptors on the macrophage surface (Habich et al 2003); however, Cpn60s have so far been mainly associated with Toll-like receptors (TLRs) as mediators of signalling (Fig. 2). M. tuberculosis Cpn60.2 activates the nuclear factor κ B (NF-κ B) through TLR4 (Bulut et al 2005). The same receptor was associated with signalling of human Cpn60 in myeloid cells (Ohashi et al 2000, Flohe et al 2003) and in mouse B cells (Cohen-Sfady et al 2005). Human Cpn60 also uses TLR2 on the surface of vascular smooth muscle cells (de Graaf et al 2006) and T cells (Zanin-Zhorov et al 2005a, 2006). In species such as Porphyromonas gingivalis (Argueta et al 2006), Cpn60.2 binds to TLR2 and 4. C. pneumoniae Cpn60 (Bulut et al 2002) activates macrophages through TLR4 and MD2 in a MyD88-dependent pathway. TLR2 and TLR4 are, however, not necessary for the induction of the innate immune responses by H. pylori Cpn60 despite its high sequence homology with Cpn60s from other species (Gobert et al 2004, Maguire et al 2005). The induction of pro-inflammatory cytokines by M. tuberculosis Cpn60.1 depends on the presence of CD14 on the surface of monocytes (Lewthwaite et al 2001, Tormay et al 2005). The activity of this chaperonin resides in the equatorial domain of the protein and it is blocked by the addition of anti-CD14 antibodies (Tormay et al 2005), while the activity of M. tuberculosis Cpn60.2 is CD-14 independent (Lewthwaite et al 2001).

CHAPERONIN 60 AND MACROPHAGES

163

FIG. 2. CD14, TLR2 and TLR4 are potential receptors for Cpn60s on the cell surface. The binding of Cpn60 to TLR is followed by the recruitment of MyD88 and activation of TNF receptor-associated factor 6 (TRAF6). Through a series of phosphorylations and activations of several other intracellular proteins, the cascade ends with the activation of NF-κ B and MAP kinases. The transcription factors translocate to the nucleus and induce the expression of target cytokine genes.

Immunostimulatory effects of chaperonin 60s High levels of human Cpn60 (Steptoe et al 2007) in the blood are associated with a reduced number of markers of infectious disease, which suggests that this molecule may protect against invasive infections. This concept is supported by the finding that vaccines of bacterial Cpn60s protect against infectious disease. For example, mice that have been vaccinated with DNA vaccine encoding MtCpn60.2 are protected against subsequent challenge with virulent M. tuberculosis (Lowrie 2006). Whilst T lymphocytes are clearly implicated in this type of protection, macrophages are activated by T lymphocyte-derived cytokines and are thought to play an important role in the suppression of growth of M. tuberculosis. Protection against infection with Cpn60s has also been demonstrated for microbial pathogens

164

COATES ET AL

such as Yersinia enterocolitica (Noll & Autenrieth 1996) and Chlamydia pneumoniae (Svanholm et al 2000). Preliminary data (A. R. M. Coates and Y. Hu, unpublished) suggest that mutant M. tuberculosis which lacks Cpn60.1, is unable to stimulate granulomatous-like formation in the mouse or the guinea pig, even though these animals contain the same numbers of the mutant as the wild-type strain in the lungs. This suggests that MtCpn60.1 stimulates the formation of granulomas. Since granulomas are mainly composed of macrophage-derived cells, this demonstrates that MtCpn60.1 stimulates macrophages. Immunosuppressive actions of chaperonin 60s Paradoxically, although Cpn60s clearly activate macrophages, albeit indirectly through T lymphocyte-derived cytokines, they also suppress immune responses in certain situations. For instance, human Cpn60 down-regulates immune responses by up-regulating T regulatory cells (Zanin-Zhorov et al 2006). It is thought that this effect is mediated through TLR2 signalling. Suppression develops in a number of inflammatory conditions, such as autoimmune diabetes and arthritis (van Eden et al 2005). In the case of MtCpn60.2, it can suppress the autoimmune disease, adjuvant arthritis in rats (Quintana et al 2002). In addition, MtCpn60.1 can suppress asthma in the mouse (Riffo-Vasquez et al 2004). It is not known why M. tuberculosis Cpn60s both suppress and activate immune responses. However, it is possible that local suppression of macrophage activation, perhaps in the granuloma, might allow non-multiplying bacteria to survive for long periods of time. This situation is characteristic of the common asymptomatic form of this disease in which bacteria survive in dormant form for life. Granuloma activation–suppression (GAS) cycle Heat shock proteins have been shown in the past to block the production of proinflammatory cytokines. For example, human Cpn60 causes down-regulation of NF-κ B activation in T cells, leading to decreased inflammatory mediator production (Zanin-Zhorov et al 2005a), and up-regulates SOCS3, an intracellular negative regulator of the JAK/STAT pathway which induces cytokine production (ZaninZhorov et al 2005b). Another human chaperonin, Cpn10, also has anti-inflammatory properties (Johnson et al 2005). Pre-treatment of human mononuclear cells with Cpn10 inhibited the LPS-induced TNFα and IL6 production (Johnson et al 2005). In a recently conducted clinical trial, human Cpn10 had an anti-inflammatory effect in patients with rheumatoid arthritis (Vanags et al 2006). We have observed the inhibitory effects of mycobacterial Cpn60s in human monocytes. If the cells were pre-treated with Cpn60s, significantly less TNFα was

CHAPERONIN 60 AND MACROPHAGES

165

released in response to LPS than in non pre-treated cells. However, this inhibitory effect was lost after treating the cells with very high or very low concentrations of chaperonins. This phenomenon leads us to hypothesize that the activity of Cpn60s might involve two receptors on the surface of a human monocyte: one receptor that binds the chaperonins at high concentrations and induces the release of proinflammatory mediators, such as TNFα , and a second receptor to which the chaperonins bind at low concentrations. The binding to the second receptor does not induce the production of cytokines, but it inhibits the cytokine production by other inflammatory stimuli, such as LPS. We suggest that, in the tuberculous granuloma, where, presumably, the concentration of Cpn60s is high, macrophages are activated. This suppresses the growth of the bacteria and prevents them from spreading. However, once growth has been suppressed for some time, it is possible that only low amounts of Cpn60 remain in the granuloma, blocking macrophage activation and allowing growth to resume, resulting in small numbers of bacteria to escape into the blood (Fig. 3). Meanwhile, low amounts of Cpn60 seep out of the granuloma into the blood (Sethna et al 1998). If systemic immunosuppression resulted as a result of low concentrations of Cpn60, small numbers of bacteria could survive in the blood and seed the organs without an immune response being triggered (Fig. 4). Once settled in their

FIG. 3. A proposed model for the granuloma activation–suppression (GAS) cycle. In a tuberculous granuloma, high concentrations of Mt Cpn60s activate the macrophages, which in turn suppress the growth of bacteria (A). After some time, the concentration of Mt Cpn60s falls and the macrophages are suppressed, allowing the bacteria to multiply (B). The multiplying bacteria produce increased amounts of Mt Cpn60s which then activate the macrophages thus completing the cycle.

166

COATES ET AL

FIG. 4. Low amounts of Mt Cpn60 circulating in the blood results in systemic immunosuppression that allows the bacteria to settle and multiply in the organs where the cycle of granuloma activation–suppression (GAS cycle) starts.

new home, for example the kidney, they would begin to multiply, the local concentration of Cpn60 would rise to a critical level when macrophages became activated and the growth of the bacteria would be suppressed. This could explain the well known ‘miliary’ or blood-borne spread of the bacteria, so the low–high phenomenon (GAS cycle) could be part of the pathogenesis of tuberculosis. Conclusions In conclusion, Cpn60s both activate and suppress macrophages. A possible explanation for this paradox is that high levels of Cpn60 stimulate macrophage activation whilst low levels suppress activation. How might this affect the pathogenesis of tuberculosis? This disease is characterised by the formation of a macrophagerich granuloma surrounding M. tuberculosis. It is proposed that high levels of MtCpn60 occur locally in the granuloma, which induce macrophage activation and granuloma formation. Bacterial mutants that lack MtCpn60.1 do not induce granu-

CHAPERONIN 60 AND MACROPHAGES

167

lomas. Macrophage activation suppresses the growth of the bacteria, and reduces the level of MtCpn60.1. Once low levels of MtCpn60.1 are reached, macrophage activation is suppressed, bacteria grow once more and small numbers are released into the blood whence they seed other organs where they induce the formation of granulomas. We suggest that the granuloma activation–suppression (GAS) cycle is a key component of the pathogenesis of tuberculosis. Acknowledgements We acknowledge fi nancial support from the Burton Trust (programme grant), and a CASE award to A. Cehovin funded by the Medical Research Council and Helperby Therapeutics Group plc.

References Andersen P, Askgaard D, Ljungqvist L, Bennedsen J, Heron I 1991 Proteins released from Mycobacterium tuberculosis during growth. Infect Immun 59:1905–1910 Argueta JG, Shiota S, Yamaguchi N, Masuhiro Y, Hanazawa S 2006 Induction of Porphyromonas gingivalis GroEL signaling via binding to Toll-like receptors 2 and 4. Oral Microbiol Immunol 21:245–251 Bulut Y, Faure E, Thomas L et al 2002 Chlamydial heat shock protein 60 activates macrophages and endothelial cells through Toll-like receptor 4 and MD2 in a MyD88-dependent pathway. J Immunol 168:1435–1440 Bulut Y, Michelsen KS, Hayrapetian et al 2005 Mycobacterium tuberculosis heat shock proteins use diverse Toll-like receptor pathways to activate pro-inflammatory signals. J Biol Chem 280:20961–20967 Chen W, Syldath U, Bellmann K, Burkart V, Kolb H 1999 Human 60-kDa heat-shock protein: a danger signal to the innate immune system. J Immunol 162:3212–3219 Cohen-Sfady M, Nussbaum G, Pevsner-Fischer M et al 2005 Heat shock protein 60 activates B cells via the TLR4-MyD88 pathway. J Immunol 175:3594–3602 de Graaf R, Kloppenburg G, Kitslaar PJ, Bruggeman CA, Stassen F 2006 Human heat shock protein 60 stimulates vascular smooth muscle cell proliferation through Toll-like receptors 2 and 4. Microbes Infect 8:1859–1865 Flohe SB, Bruggemann J, Lendemans S et al 2003 Human heat shock protein 60 induces maturation of dendritic cells versus a Th1-promoting phenotype. J Immunol 170:2340–2348 Friedland JS, Shattock R, Remick DG, Griffi n GE 1993 Mycobacterial 65-kD heat shock protein induces release of proinflammatory cytokines from human monocytic cells. Clin Exp Immunol 91:58–62 Gobert AP, Bambou JC, Werts C et al 2004 Helicobacter pylori heat shock protein 60 mediates interleukin-6 production by macrophages via a toll-like receptor (TLR)-2-, TLR-4-, and myeloid differentiation factor 88-independent mechanism. J Biol Chem 279:245–250 Habich C, Kempe K, van der Zee R, Burkart V, Kolb H 2003 Different heat shock protein 60 species share pro-inflammatory activity but not binding sites on macrophages. FEBS Lett 533:105–109 Johnson BJ, Le TT, Dobbin CA et al 2005 Heat shock protein 10 inhibits lipopolysaccharideinduced inflammatory mediator production. J Biol Chem 280:4037–4047 Kaufmann SH 2001 How can immunology contribute to the control of tuberculosis? Nat Rev Immunol 1:20–30

168

COATES ET AL

Kong TH, Coates AR, Butcher PD, Hickman CJ, Shinnick TM 1993 Mycobacterium tuberculosis expresses two chaperonin-60 homologs. Proc Natl Acad Sci USA 90:2608–2612 Lewthwaite JC, Coates AR, Tormay P et al 2001 Mycobacterium tuberculosis chaperonin 60.1 is a more potent cytokine stimulator than chaperonin 60.2 (Hsp 65) and contains a CD14binding domain. Infect Immun 69:7349–7355 Lewthwaite J, George R, Lund PA et al 2002 Rhizobium leguminosarum chaperonin 60.3, but not chaperonin 60.1, induces cytokine production by human monocytes: activity is dependent on interaction with cell surface CD14. Cell Stress Chaperones 7:130–136 Lowrie DB 2006 DNA vaccines for therapy of tuberculosis: where are we now? Vaccine 24:1983–1989 Maguire M, Poole S, Coates AR, Tormay P, Wheeler-Jones C, Henderson B 2005 Comparative cell signalling activity of ultrapure recombinant chaperonin 60 proteins from prokaryotes and eukaryotes. Immunology 115:231–238 Noll A, Autenrieth IB 1996 Immunity against Yersinia enterocolitica by vaccination with Yersinia HSP60 immunostimulating complexes or Yersinia HSP60 plus interleukin-12. Infect Immun 64:2955–2961 Ohashi K, Burkart V, Flohe S, Kolb H 2000 Cutting edge: heat shock protein 60 is a putative endogenous ligand of the toll-like receptor-4 complex. J Immunol 164:558–561 Peetermans WE, Raats CJ, Langermans JA, van Furth R 1994 Mycobacterial heat-shock protein 65 induces proinflammatory cytokines but does not activate human mononuclear phagocytes. Scand J Immunol 39:613–617 Quintana FJ, Carmi P, Mor F, Cohen IR 2002 Inhibition of adjuvant arthritis by a DNA vaccine encoding human heat shock protein 60. J Immunol 169:3422–3428 Riffo-Vasquez Y, Spina D, Page C et al 2004 Effect of Mycobacterium tuberculosis chaperonins on bronchial eosinophilia and hyper-responsiveness in a murine model of allergic inflammation. Clin Exp Allergy 34:712–719 Russell DG 2007 Who puts the tubercle in tuberculosis? Nat Rev Microbiol 5:39–47 Sethna KB, Mistry NF, Dholakia Y, Antia NH, Harboe M 1998 Longitudinal trends in serum levels of mycobacterial secretory (30 kD) and cytoplasmic (65 kD) antigens during chemotherapy of pulmonary tuberculosis patients. Scand J Infect Dis 30:363–369 Steptoe A, Shamaei-Tousi A, Gylfe A et al 2007 Protective effect of human heat shock protein 60 suggested by its association with decreased seropositivity to pathogens. Clin Vaccine Immunol 14:204–207 Svanholm C, Bandholtz L, Castanos-Velez E, Wigzell H, Rottenberg ME 2000 Protective DNA immunization against Chlamydia pneumoniae. Scand J Immunol 51:345–353 Tormay P, Coates AR, Henderson B 2005 The intercellular signaling activity of the Mycobacterium tuberculosis chaperonin 60.1 protein resides in the equatorial domain. J Biol Chem 280:14272–14277 Ueki K, Tabeta K, Yoshie H, Yamazaki K 2002 Self-heat shock protein 60 induces tumour necrosis factor-alpha in monocyte-derived macrophage: possible role in chronic inflammatory periodontal disease. Clin Exp Immunol 127:72–77 van Eden W, van der Zee R, Prakken B 2005 Heat-shock proteins induce T-cell regulation of chronic inflammation. Nat Rev Immunol 5:318–330 Vanags D, Williams B, Johnson B et al 2006 Therapeutic efficacy and safety of chaperonin 10 in patients with rheumatoid arthritis: a double-blind randomised trial. Lancet 368:855– 863 WHO 2007 Global tuberculosis control: surveillance, planning, fi nancing. Geneva, Switzerland Zanin-Zhorov A, Bruck R, Tal G et al 2005a Heat shock protein 60 inhibits Th1-mediated hepatitis model via innate regulation of Th1/Th2 transcription factors and cytokines. J Immunol 174:3227–3236

CHAPERONIN 60 AND MACROPHAGES

169

Zanin-Zhorov A, Tal G, Shivtiel S et al 2005b Heat shock protein 60 activates cytokineassociated negative regulator suppressor of cytokine signaling 3 in T cells: effects on signaling, chemotaxis, and inflammation. J Immunol 175:276–285 Zanin-Zhorov A, Cahalon L, Tal G, Margalit R, Lider O, Cohen IR 2006 Heat shock protein 60 enhances CD4+ CD25+ regulatory T cell function via innate TLR2 signaling. J Clin Invest 116:2022–2032 Zheng L, He M, Long M, Blomgran R, Stendahl O 2004 Pathogen-induced apoptotic neutrophils express heat shock proteins and elicit activation of human macrophages. J Immunol 173:6319–6326

DISCUSSION Ellis: I remember you did some experiments on the physical state of the chaperone proteins. Can you summarize the position on this? Coates: We did this initially to try to see whether LPS might be involved in any of these biological effects. The best experiment was that we can obliterate the whole of the effect using a protease. Boiling does the same thing, but I don’t think it is as convincing. Gaston: The low dose–high dose thing is very interesting. Does the 195 peptide show that dose response? Coates: We haven’t looked at peptides yet. We are just about to start. Gaston: How would you conceptualize this at the moment? Is there a low affinity receptor that signals one way and a high affinity one that goes the other? Coates: Yes. There are other possibilities, but that’s the most attractive one. Macario: Is the granuloma good or bad? Coates: That is an interesting question. It is the hallmark of this disease. What would happen without a granuloma? There are suggestions that there are uncontrolled infections in the absence of a granuloma. In the mutant there isn’t uncontrolled infection, but it might be crippled in some other way, so we can’t exclude that. We have done microarrays with all 4000 mycobacterial genes and can’t see any other proteins that are closely involved or altered. But that is not a perfect experiment. All I can think is that the jury is out regarding what a granuloma does. I thought if you had a bug, a lot of cells around it would contain it. Curiously, in mice there aren’t good granulomas, but the mouse is good at controlling tuberculosis (TB). Macario: Then there is the story of the colonization of the regional lymph nodes, which are another reservoir. Coates: Yes, particularly in early infection it is thought that they go to the draining lymph nodes and form a granulomatous early infection. It is thought that, over time, the lymph node suppresses the bacteria which become dormant. Panayi: Some patients given anti-TNF get reactivation of their TB. How does this fit into the GAS cycle? The image I have is that there is a granuloma and

170

COATES ET AL

somehow anti-TNF breaks the granuloma and off you go. But this doesn’t seem to fit in with these models you have described. What is your explanation? Coates: I don’t know. You shouldn’t get comparable growth of this sort of mutant if it doesn’t induce TNF. It induced a bit of TNF. This mutant does still have the other chaperonins, and another 4000 or so proteins of which at least some will stimulate some immunoreactivity. What about Hsp70s? It would be interesting to know what they were doing in the relationship of the host cell within this nongranuloma. I guess the answer is that it is probably the balance: the mutant is just able to grow and produce just enough of TNF to do the job. Gupta: You are ascribing a central role for Hsp60 in TB. There are so many other genes in M. tuberculosis, and particularly for M. smegmatis gene knockouts have been made for most of these. What is their effect on granuloma formation? Coates: M. smegmatis is not a pathogen so it won’t induce granulomas. Gupta: Was the knockout experiment you described with M. tuberculosis? Coates: Yes, a virulent form. There are quite a few gene knockouts in M. tuberculosis and none of them do this. Gupta: Do the experiments you describe establish an important role for Hsp60 in TB? Coates: Science is the art of the feasible. We have been a bit disappointed with microarrays. We have tried dozens of knockouts. With bacterial knockouts one can make slightly firmer conclusions than from microarrays. Is it certain? No, but it is as certain as we can get as long as this is taken in context. There are a number of experiments that remain to be done. For example, what would happen if we put in the mutant and then intravenously injected different concentrations of Cpn60.1? Gupta: The genomes have been sequenced for different mycobacteria, and M. tuberculosis has a large number of unique genes. I would have thought that some of these M. tuberculosis-specific genes will be more important in this regard. Have any of these genes been looked at? Coates: This is why the question about the role of the granuloma is so interesting. The tuberculosis disease doesn’t just depend on the granuloma: there are many other issues involved. As far as the granuloma is concerned, the evidence from these data is that it may be involved. If you don’t have the granuloma you won’t get the life cycle of TB, because you will not get the caseating tuberculoma which breaks into the bronchi and allows bacteria to access the bronchi, from where they are expelled by coughing. The bacteria-laden aerosol is inhaled by another person, and the cycle begins again. Calderwood: Can you speculate on what the high and low affinity receptors are? Coates: We don’t know yet. We are particularly focusing on TLRs, looking at a range of them using different techniques.

CHAPERONIN 60 AND MACROPHAGES

171

Macario: I have a comment on what appear to be specific sets of genes for a given organism. The comment was made that if we found a set of genes specific for M. tuberculosis these would be the ones that are mostly responsible for the disease, rather than those shared with other organisms. However, we have to think in terms of context of genes, and network of gene products, that will cause the final biological effect. It is difficult to extrapolate from the identification of a specific set of genes in a given genome to a defined biological effect. Gupta: That point is well taken, but at the same time if genes are specific for a given micro-organism, the chances are greater that many of these will be involved in its virulence or specific properties including pathogenicity. Coates: You could say that there is specificity in the chaperonins. There is abundant evidence that each one is slightly different from the next. We have one that might be essential for folding and another which is perhaps less essential, but is an expert in doing other things. Looking at the different sorts of specificity in the genome is a little crude. Perhaps one should be looking at fine differences between old genes like chaperonins. Multhoff: Has anyone measured antibodies to human chaperonins at different stages of disease? Coates: That’s an interesting question. I don’t think so. Lund: If you take any animal, generally speaking they have a lot of antibodies against Hsp60 just because of the infections they have had during their lifetime. These will be cross-reactive. Coates: There are antibodies against the bacterial chaperonins. Csermely: Is a direct involvement of surface-expressed Hsp60.1 excluded in granuloma formation? Coates: That’s an interesting thought. We are doing some confocal microscopy, which should shed light on this. We know we have some binding to cell surfaces. Triantafilou: I was interested in the differences between 60.1 and 60.2. What is the sequence identity between the two and why would they be so different as far as function goes? Coates: They are about 60% amino acid identical. They have certain molecular features that are different. The 60.1 has a polyhistidine tail which is rather curious. We don’t know the reason for these subtle functional differences. We just know that they exist. Triantafilou: You showed that the 60.2 wasn’t inducing much cytokine production. Coates: It induces some, but less. Corrigall: In your mutant in which TNF levels were down regulated, have you looked at other cytokines? Is IL10 up-regulated?

172

COATES ET AL

Coates: We didn’t look at IL10. Panayi: These macrophages are making IFNγ. Were these pure macrophage cultures? Coates: I think there were T cells present. Multhoff: Is there any impact of NK cells in this disease? You could compare SCID (containing active NK cells but lacking T and B cells) with SCID/beige mice that lack active T, B and NK cells. Coates: That’s a good question. We don’t know yet. Feder: When you have mycobacteria in human patients, does the Cpn60.1 and possibly 60.2 ever make its way out of the granulomas into the blood of the patients? If so, does this secreted CPN60 have any impact on the human patient? Coates: Yes, it is detectable in the blood of humans. We have measured this with SELDI mass spectrometry and ELISA. There isn’t any doubt that 60.1 and 60.2 get into the blood. Morten Harboe and colleagues measured 60.2 in the blood and found low levels which persisted (Sethna et al 1998). In patients with TB it lasts for months. We looked at some healthy controls, and a few of them had 60.1/60.2 in their blood. I have no idea what effect this has on their health, but there are a number of possibilities. Is their response to allergy different? Do they get autoimmune diseases at a different rate? We don’t know. Reference Sethna KB, Mistry NF, Dholakia Y, Antia NH, Harboe M 1998 Longitudinal trends in serum levels of mycobacterial secretory (30 kD) and cytoplasmic (65 kD) antigens during chemotherapy of pulmonary tuberculosis patients. Scand J Infect Dis 30:363–369

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Hsp70: a chaperokine Alexzander Asea Division of Investigative Patholog y, Scott & White Memorial Hospital and Clinic and Texas A&M University System Health Science Center, College of Medicine, 1901 South 1st Street, Temple, TX 76508, USA

Abstract. Chaperokine is a term recently coined to better describe the dual role of extracellular heat shock protein as both chaperone and cytokine. The augmentation of intracellular Hsp72 expression has clearly been demonstrated to be cytoprotective; with the ability to activate anti-apoptotic and anti-inflammatory pathways, exerting inhibitory effects on cell cycle progression and suppressing genes important in proliferation and differentiation. On the other hand, the role of extracellular Hsp72 is only now being elucidated and has been demonstrated to play a cytostimulatory role by enhancing proinflammatory cytokine and chemokine synthesis, up-regulate co-stimulatory molecule expression, enhance the maturation of dendritic cells and promote antitumour surveillance. This paper covers the most recent advances in elucidation of the mechanism by which the chaperokine activity of Hsp72 is transduced, addresses its biological significance and, fi nally, covers how it is being harnessed to produce novel therapeutic agents. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 173–183

Chaperokine, is a term recently coined to better describe the unique function of exogenous Hsp72 (eHsp72) as both a chaperone and cytokine (for review, see Asea 2003, 2007). The consequence of binding and signalling is the stimulation of a potent and long lasting immune response. eHsp72 induces a plethora of immune responses and the list continues to grow. Table 1 lists some immune effector functions that are attributed to various sources of Hsp70 including recombinant bacterial, viral and parasitic sources, Hsp70 from human, mouse and non-human primate tissue and their targets (Table 1). Chaperokine activity of eHsp72 As early as 2–4 hours post exposure of antigen presenting cells (APCs) to eHsp72, there is significant release of cytokines including tumour necrosis factor (TNF) α , interleukin (IL)1β, IL6 and IL12 (Asea et al 2000b, 2002) and GM-CSF (Srivastava 173

murine encephalitis mouse measles virus murine tumors mouse immunization Plasmodium yoelii sporozoiteinfected mice

protective immunity protective immunity protective immunity serum Ab, cytokine release protective immunity

(Chen et al 2000) (Oglesbee et al 2002) (Huang et al 2003) (Qazi et al 2005) (Kawabata et al 2002)

co-stimulatory molecules maturation signal antigen presentation, cytokine release humoral and cellular immune response

mouse immunization

(Asea et al 2002) (Asea et al 2002) (Asea et al 2000a, Millar et al 2003) (Suzue & Young 1996a,b, Suzue et al 1997)

CC chemokines

human monocytes and monocyte-derived DC human DC human DC mouse DC

Recombinant Mycobacterium tuberculosis Hsp70–p24 fusion protein Recombinant DNA vaccine Hyperthermia Adenovirus Plasmodium falciparum Recombinant MSP1 fused to Hsc70

(Asea et al 2000b, Asea et al 2002) (Wang et al 2005)

cytokine release

human monocyte

References

Recombinant human and mouse

Effect

Target

Chaperokine activity of Hsp72

Source of Hsp70

TABLE 1

174 ASEA

beta chemokines protective immunity cross presentation

CD8 + CTL HIV human DC human DC human NK cells cancer

Non-human primates Non-human primates recombinant Hsp70–CCR5 fusion Human tumours

Human exosomes

Human tumours

(Gastpar et al 2005) (Srivastava & Udono 1994, Srivastava 2000, 2005, 2006)

(Srivastava et al 1994, Noessner et al 2002) (Bausero et al 2005, Gastpar et al 2005)

(Binder et al 2000a, Panjwani et al 2002) (Binder et al 2000b) (Li & Srivastava 1993, Udono & Srivastava 1993, Heike et al 1994) (Casey et al 2003, Wang et al 2003a,b) (Lehner et al 2000) (Bogers et al 2004a,b)

CCR, chemokine receptor; CTL, cytotoxic T lymphocytes; DC, dendritic cells; HIV, human immunodeficiency virus; MSP, Merozoite Surface Protein; NK, natural killer; PEC, peritoneal exudates cells.

cytokine release, costimulatory molecules, maturation signal migration tumor regression

cross-presentation

cross presentation protective immunity

mouse PEC CD8+ CTL melanoma

nitric oxide release, migration

mouse DC

Mouse and recombinant

Mouse liver and kidney

HSP70: A CHAPEROKINE 175

176

ASEA

2002); nitric oxide, a potent apotogenic mediator (Panjwani et al 2002); and chemokines including MIP1, MCP1 and RANTES (Lehner et al 2000, Panjwani et al 2002). This part of the immune response does not require peptide, since both peptide-bearing and non peptide-bearing eHsp72 is capable of inducing proinflammatory cytokine production by APCs (Asea et al 2000a). However, peptide is required for specific CD8 + cytotoxic T lymphocyte (CTL) responses (Srivastava et al 1994, Srivastava 2000, 2005). eHsp72 induces the dendritic cell (DC) maturation by augmenting the surface expression of CD40, CD83, CD86 and MHC class II molecules on DCs (Basu et al 2000, Singh-Jasuja et al 2000, Asea et al 2002, Noessner et al 2002) and migration of DCs (Binder et al 2000a) and NK cells (Gastpar et al 2005). Mechanisms of eHsp72 release There are currently two recognized mechanisms by which Hsp72 is released from cells: a passive release mechanism, including necrotic cell death (Calderwood 2005), severe blunt trauma (Pittet et al 2002), surgery (Dybdahl et al 2005) and following infection with lytic viruses (Moehler et al 2005); and an active release mechanism, which involves the non classical protein release pathway in which Hsp72 is released both as free Hsp72 and within highly immunologically potent exosomes (Bausero et al 2005, Gastpar et al 2005, Mambula & Calderwood 2006). Biological significance Whether Hsp72 enters the circulation via an active or passive release mechanism, what is its role in the circulation? The danger theory postulates that immune activation involves danger/non-danger molecular recognition schemas and suggests that innate immune cells are activated by danger signals that are derived from stressed or damaged self-proteins (Matzinger 1998, Gallucci & Matzinger 2001). It is now widely accepted that eHsp72 fit these criteria. The hypothesis is further reinforced by studies showing that circulating eHsp72 is increased and upregulated in diseased conditions including renal disease (Wright et al 2000), hypertension (Pockley et al 2002), atherosclerosis (Pockley et al 2003) and sickle cell disease (Adewoye et al 2005). However, intriguing questions still remain as to the role of increased circulating eHsp72 during psychological stress such as that which occurs when a Sprague Dawley rat is exposed to a cat (Fleshner et al 2004), or studies demonstrating that the human brain is able to release eHsp72 into the circulation in response to exercise (Lancaster et al 2004). In these situations, is it possible that circulating eHsp72 is priming the immune system to real or perceived danger? These are important questions, the answers to which will provide key insights into numerous psychological and pathophysiological conditions. However, extensive studies and a concerted

HSP70: A CHAPEROKINE

177

effort that brings together heat shock protein (HSP) researchers from such disparate fields as immunology, molecular biology, cell biology, neurophysiology and psychology, utilizing new breakthrough technologies including proteomics and the recently deciphered human genome, is required before conclusive answers can be given.

Acknowledgements This work was supported in part by the National Institute of Health grant RO1CA91889, Scott & White Clinic, the Texas A&M University System Health Science Center College of Medicine, the Central Texas Veterans Health Administration and an Endowment from the Cain Foundation.

References Adewoye AH, Klings ES, Farber HW et al 2005 Sickle cell vaso-occlusive crisis induces the release of circulating serum heat shock protein-70. Am J Hematol 78:240–242 Asea A 2003 Chaperokine-induced signal transduction pathways. Exerc Immunol Rev 9:25–33 Asea A 2007 Release of heat shock proteins: passive vs active release mechanisms. In: Asea A, DeMaio A (eds) Heat shock proteins: potent mediators of inflammation and immunity. Springer Publishers, Dordrecht, The Netherlands, p 3–20 Asea A, Kabingu E, Stevenson MA, Calderwood SK 2000a HSP70 peptide-bearing and peptide-negative preparations act as chaperokines. Cell Stress Chaperones 5:425–431 Asea A, Kraeft SK, Kurt-Jones EA et al 2000b HSP70 stimulates cytokine production through a CD14-dependant pathway, demonstrating its dual role as a chaperone and cytokine. Nat Med 6:435–442 Asea A, Rehli M, Kabingu E et al 2002 Novel signal transduction pathway utilized by extracellular HSP70: role of toll-like receptor (TLR) 2 and TLR4. J Biol Chem 277:15028– 15034 Basu S, Binder RJ, Suto R, Anderson KM, Srivastava PK 2000 Necrotic but not apoptotic cell death releases heat shock proteins, which deliver a partial maturation signal to dendritic cells and activate the NF-kappa B pathway. Int Immunol 12:1539–1546 Bausero MA, Gastpar R, Multhoff G, Asea A 2005 Alternative mechanism by which IFN- γ enhances tumor recognition: active release of heat shock protein 72. J Immunol 175: 2900–2912 Binder RJ, Anderson KM, Basu S, Srivastava PK 2000a Cutting edge: heat shock protein gp96 induces maturation and migration of CD11c+ cells in vivo. J Immunol 165:6029–6035 Binder RJ, Harris, ML, Menoret A, Srivastava PK 2000b Saturation, competition, and specificity in interaction of heat shock proteins (hsp) gp96, hsp90, and hsp70 with CD11b+ cells. J Immunol 165:2582–2587 Bogers WM, Bergmeier LA, Ma J et al 2004a A novel HIV-CCR5 receptor vaccine strategy in the control of mucosal SIV/HIV infection. AIDS 18:25–36 Bogers WM, Bergmeier LA, Oostermeijer H et al 2004b CCR5 targeted SIV vaccination strategy preventing or inhibiting SIV infection. Vaccine 22:2974–2984 Calderwood SK 2005 Chaperones and slow death–a recipe for tumor immunotherapy. Trends Biotechnol 23:57–59

178

ASEA

Casey DG, Lysaght J, James T, Bateman A, Melcher AA, Todryk SM 2003 Heat shock protein derived from a non-autologous tumour can be used as an anti-tumour vaccine. Immunology 110:105–111 Chen W, Lin Y, Liao C, Hsieh S 2000 Modulatory effects of the human heat shock protein 70 on DNA vaccination. J Biomed Sci 7:412–419 Dybdahl B, Slordahl SA, Waage A, Kierulf P, Espevik T, Sundan A 2005 Myocardial ischaemia and the inflammatory response: release of heat shock protein 70 after myocardial infarction. Heart 91:299–304 Fleshner M, Campisi J, Amiri L, Diamond DM 2004 Cat exposure induces both intra- and extracellular Hsp72: the role of adrenal hormones. Psychoneuroendocrinology 29:1142– 1152 Gallucci S, Matzinger P 2001 Danger signals: SOS to the immune system. Curr Opin Immunol 13:114–19 Gastpar R, Gehrmann M, Bausero MA et al 2005 Heat shock protein 70 surface-positive tumor exosomes stimulate migratory and cytolytic activity of natural killer cells. Cancer Res 65:5238–5247 Heike M, Blachere NE, Srivastava PK 1994 Protective cellular immunity against a spontaneous mammary carcinoma from ras transgenic mice. Immunobiology 190:411–423 Huang XF, Ren W, Rollins L et al 2003 A broadly applicable, personalized heat shock proteinmediated oncolytic tumor vaccine. Cancer Res 63:7321–7329 Kawabata Y, Udono H, Honma K et al 2002 Merozoite surface protein 1-specific immune response is protective against exoerythrocytic forms of Plasmodium yoelii. Infect Immun 70:6075–6082 Lancaster GI, Moller K, Nielsen B et al 2004 Exercise induces the release of heat shock protein 72 from the human brain in vivo. Cell Stress Chaperones 9:276–280 Lehner T, Bergmeier LA, Wang Y et al 2000 Heat shock proteins generate beta-chemokines which function as innate adjuvants enhancing adaptive immunity. Eur J Immunol 30:594–603 Li Z, Srivastava PK 1993 Tumor rejection antigen gp96/grp94 is an ATPase: implications for protein folding and antigen presentation. EMBO J 12:3143–3151 Mambula SS, Calderwood SK 2006 Heat shock protein 70 is secreted from tumor cells by a nonclassical pathway involving lysosomal endosomes. J Immunol 177:7849–7857 Matzinger P 1998 An innate sense of danger. Semin Immunol 10:399–415 Millar DG, Garza KM, Odermatt B et al 2003 Hsp70 promotes antigen-presenting cell function and converts T-cell tolerance to autoimmunity in vivo. Nat Med 9:1469–1476 Moehler MH, Zeidler M, Wilsberg V et al 2005 Parvovirus H-1-induced tumor cell death enhances human immune response in vitro via increased phagocytosis, maturation, and cross-presentation by dendritic cells. Hum Gene Ther 16:996–1005 Noessner E, Gastpar R, Milani V et al 2002 Tumor-derived heat shock protein 70 peptide complexes are cross-presented by human dendritic cells. J Immunol 169:5424– 5432 Oglesbee MJ, Pratt M, Carsillo T 2002 Role for heat shock proteins in the immune response to measles virus infection. Viral Immunol 15:399–416 Panjwani NN, Popova L, Srivastava PK 2002 Heat shock proteins gp96 and hsp70 activate the release of nitric oxide by APCs. J Immunol 168:2997–3003 Pittet JF, Lee H, Morabito D, Howard MB, Welch WJ, Mackersie RC 2002 Serum levels of Hsp 72 measured early after trauma correlate with survival. J Trauma 52:611–617; discussion 617 Pockley AG, De Faire U, Kiessling R, Lemne C, Thulin T, Frostegard J 2002 Circulating heat shock protein and heat shock protein antibody levels in established hypertension. J Hypertens 20:1815–1820

HSP70: A CHAPEROKINE

179

Pockley AG, Georgiades A, Thulin T, de Faire U, Frostegard J 2003 Serum heat shock protein 70 levels predict the development of atherosclerosis in subjects with established hypertension. Hypertension 42:235–238 Qazi KR, Wikman M, Vasconcelos NM, Berzins K, Stahl S, Fernandez C 2005 Enhancement of DNA vaccine potency by linkage of Plasmodium falciparum malarial antigen gene fused with a fragment of HSP70 gene. Vaccine 23:1114–1125 Singh-Jasuja H, Scherer HU, Hilf N et al 2000 The heat shock protein gp96 induces maturation of dendritic cells and down-regulation of its receptor. Eur J Immunol 30:2211–2215 Srivastava PK 2000 Heat shock protein-based novel immunotherapies. Drug News Perspect 13:517–522 Srivastava P 2002 Interaction of heat shock proteins with peptides and antigen presenting cells: chaperoning of the innate and adaptive immune responses. Annu Rev Immunol 20: 395–425 Srivastava PK 2005 Immunotherapy for human cancer using heat shock protein-peptide complexes. Curr Oncol Rep 7:104–108 Srivastava PK 2006 Therapeutic cancer vaccines. Curr Opin Immunol 18:201–5 Srivastava PK, Udono H 1994 Heat shock protein-peptide complexes in cancer immunotherapy. Curr Opin Immunol 6:728–732 Srivastava PK, Udono H, Blachere NE, Li Z 1994 Heat shock proteins transfer peptides during antigen processing and CTL priming. Immunogenetics 39:93–98 Suzue K, Young RA 1996a Adjuvant-free hsp70 fusion protein system elicits humoral and cellular immune responses to HIV-1 p24. J Immunol 156:873–879 Suzue K, Young RA 1996b Heat shock proteins as immunological carriers and vaccines. EXS 77:451–465 Suzue K, Zhou X, Eisen HN, Young RA 1997 Heat shock fusion proteins as vehicles for antigen delivery into the major histocompatibility complex class I presentation pathway. Proc Natl Acad Sci USA 94:13146–13151 Udono H, Srivastava PK 1993 Heat shock protein 70-associated peptides elicit specific cancer immunity. J Exp Med 178:1391–1396 Wang XY, Chen X, Manjili MH, Repasky E, Henderson R, Subjeck JR 2003a Targeted immunotherapy using reconstituted chaperone complexes of heat shock protein 110 and melanomaassociated antigen gp100. Cancer Res 63:2553–2560 Wang XY, Kazim L, Repasky EA, Subjeck JR 2003b Immunization with tumor-derived ER chaperone grp170 elicits tumor-specific CD8+ T-cell responses and reduces pulmonary metastatic disease. Int J Cancer 105:226–231 Wang Y, Whittall T, McGowan E et al 2005 Identification of stimulating and inhibitory epitopes within the heat shock protein 70 molecule that modulate cytokine production and maturation of dendritic cells. J Immunol 174:3306–3316 Wright BH, Corton JM, El-Nahas AM, Wood RF, Pockley AG 2000 Elevated levels of circulating heat shock protein 70 (Hsp70) in peripheral and renal vascular disease. Heart Vessels 15:18–22

DISCUSSION Panayi: I found what you said about exercise very interesting. There are various things you can measure as a result of exercise: lung capacity, oxygen utilization, changes in blood lipids and so on. The variation seen may be genetically determined. Thus, there is a strong familial aggregation in the response of blood

180

ASEA

lipids to exercise (Rice et al 2002). Although you showed nicely that certain things happened during exercise, and you couldn’t improve them more by giving them Hsp70, if there were any animals that didn’t show an improvement with the exercise, which would parallel the human condition, could you make them correct the deficit that nature gave them? Did you test enough animals to pick this up? Asea: We did have enough animals, and there were very few animals that didn’t show any improvement. However, we did not look any further at them. Panayi: Did they improve when you gave them exogenous Hsp70? Asea: Yes. Panayi: That’s an important point. Naturally they couldn’t do it, but they could do whatever it was you were measuring when they were given exogenous Hsp70. Asea: I would say that 80% would improve when they are injected with Hsp70. Macario: Were those rats inbred Asea: Yes, they are spontaneously hypertensive rats. Macario: Are the rats that are prevented from running on the wheel stressed? Asea: Yes, initially. These experiments take a long time because when you lock their wheel they are really upset. A study by Monika Fleshner (Fleshner et al 2004) addressed the question of whether the released HSPs could be due to tearing of the muscle or a similar event. In this study a rat was placed in a glass cage and a cat placed on top of it. The rat was scared, and when blood was drawn, the levels of Hsp72 were found to be significantly increased. These results could be reproduced using fox urine. Triantafilou: What do you feel about the disparity between pro-inflammatory and anti-inflammatory responses that have been observed with HSPs? Do you think it could be a situation where you have a concentration-dependent mechanism, where depending on the concentration of HSP you either have a pro- or antiinflammatory response? Asea: It should be concentration dependent, but also it could be cell-type dependent. For example, if there is a case where the host has a relatively large number of Treg cells in circulation, perhaps because of an infection that is now cleared. The host’s response to the HSP may be different than when there is a relatively small number of Tregs. On the other hand, the effect of HSPs on Th17 cells could make it go the other way. Nixon: You have identified no fewer than six receptors to which Hsp70 can actually bind. Do the differences in response, anti-inflammatory or pro-inflammatory, reflect the nature of the binding between Hsp70 and the actual receptor that is involved, and thus the signal transduction pathways that are initiated as a consequence of that?

HSP70: A CHAPEROKINE

181

Asea: I’d like to think so. This would make it much easier to understand. Then you’d also be able to understand why different cells would have different concentrations of receptors on their surface. Knowing the type of cells that are exposed to, and respond to HSPs as they come into the inflammatory foci would be useful. We and others are also investigating which tissue or organ the HSP is released from. Work from the Febbraio lab (Lancaster et al 2004) using internal jugular venous to arterial difference and global cerebral blood flow, before and after exercise, demonstrated that Hsp70 is released from the brain after exercise. In a separate study from the Fleshner lab, Johnson et al (2005) reported that Hsp72 increases in the circulation via an α1-adrenergic receptor-mediated signalling pathway. They showed that activation of α1-adrenoceptors results in a rapid increase in circulating Hsp72, and blockade of α1-adrenoceptors prevents the stress-induced rise in circulating Hsp72. The mechanism is slowly coming together, but it is hard to put everything together yet. Csermely: How do you rationalize the change in the centenarians with decreased HSP concentrations? Should there be a change in the lifespan, or do people with high levels die out by this age? Asea: In situations in which there is a high concentration of HSP in the blood, we hypothesize that our body is telling us that there is some danger. As one ages, arthritis and other inflammatory diseases set in. Why the centenarians should have such low levels of HSPs is unclear. We are now trying to look at the composition of lipids on the surface of their cells. It seems that this group of people exhibit lower levels of cholesterol than others. We have recently demonstrated that extremely low levels of cholesterol inhibit the ability of cells to release free Hsp72 and Hsp72-containing exosomes (Bausero et al 2005). Gaston: Centenarians have done well to get to 100, but they will almost certainly have lots of arthritis and degenerative disease, as well as substantial plaques. Asea: We were looking at a special group, the Ashkenazi Jewish descents. They don’t have much debilitating arthritis or other degenerative diseases. Perhaps this is a highly selective group. Csermely: I don’t know whether it is that selective. We are participating in a European study, which includes a general cohort of centenarians, and they look similar to yours. Henderson: I’d like to return to the Hsp70 receptor. Which Hsp70 is your human one? Asea: We are using recombinant A1a. Henderson: You have shown that there are multiple receptors, as has Stuart Calderwood. The Guy’s group led by Tom Lehner working on mycobacterial proteins have shown even more. Have any of these groups tried to block individual

182

ASEA

receptors, to see whether blocking CD40 blocks the whole effect? How many receptors are needed to get the effect you are seeing? Asea: I think most people just focus on their HSP receptor. I think all the receptors are probably involved in triggering the cell response. Calderwood: It is not all that clear. When we did confocal studies it looked like we were getting large aggregates on the cell surface. I think these probably contain multiple receptors and downstream signalling modules. Henderson: Kathy Triantafi lou is finding rather similar activation clusters. Does this have a resonance with the lipopolysaccharide (LPS) effect you are seeing? Triantafilou: This is what I was thinking, especially with Stuart’s work with the scavengers, which have been shown to associate with TLR2. I am not sure whether they directly bind to the TLRs. I think they are binding to another accessory type of molecule which is probably nearby, resulting in a cluster. It could be that depending on what is clustering, the responses are different. Asea: We are just learning what happens when a ligand binds to its receptor. There are lots of things that happen on the surface of the cell, such as rearrangement of proteins and migration of other proteins and adaptor proteins. Nixon: So there is evidence for the movement of the receptors. Asea: Yes, there is movement. Habich: Our studies to characterize the receptor structures for human Hsp60 on innate immune cells have shown that TLR4 and CD14, both involved in LPS binding and signalling, are necessary for the Hsp60-mediated pro-inflammatory response. In contrast, we found that TLR4 and CD14 are not binding receptors for Hsp60. Based on these findings, we also think that a receptor complex for Hsp60 exists on innate immune cells and that the interaction of Hsp60 with these cells represents a highly complex process, including the contact of Hsp60 with cell surface structures involved in binding and in the initiation of pro-inflammatory responses. Henderson: If you add oxidised LDL to your cells, do you get the same Ca2+ transient response? If you say that Hsp70 is acting through LOX1, the natural ligand for LOX1 is oxidized LDL, you should expect that oxidized LDL would give you a Ca2+ transient. Is this known? Asea: I don’t know whether it is known, but one of the first real signals of ligand–receptor binding is intracellular Ca2+ flux. Calderwood: With oxidized LDL there may be something from the atherosclerosis field. It will induce NF-κ B signalling. I can’t remember if it induces Ca2+ signals. Triantafilou: We have tried oxidised LDL to see whether we can sequester LPS or LTA. We have had mixed results. Our readout is cytokine secretion. With LTA, it sequesters the LTA so we don’t get cytokine secretion. In the case of the LPS we see it augmenting the response.

HSP70: A CHAPEROKINE

183

References Bausero MA, Gastpar R, Multhoff G, Asea A 2005 Alternative mechanism by which IFNgamma enhances tumor recognition: active release of heat shock protein 72. J Immunol 175:2900–2912 Fleshner M, Campisi J, Amiri L, Diamond DM 2004 Cat exposure induces both intra- and extracellular Hsp72: the role of adrenal hormones. Psychoneuroendocrinology 29: 1142–1152 Johnson JD, Campisi J, Sharkey CM, Kennedy SL, Nickerson M, Fleshner M 2005 Adrenergic receptors mediate stress-induced elevations in extracellular Hsp72. J Appl Physiol 99:1789–1795 Lancaster GI, Moller K, Nielsen B, Secher NH, Febbraio MA, Nybo L 2004 Exercise induces the release of heat shock protein 72 from the human brain in vivo. Cell Stress Chaperones 9:276–280 Rice T, Després JP, Pérusse L et al 2002 Familial aggregation of blood lipid response to exercise training in the health, risk factors, exercise training, and genetics (HERITAGE) Family Study. Circulation 105:1904–1908

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Extracellular functions of thioredoxin Hajime Nakamura Thioredoxin Project, Department of Experimental Therapeutics, Translational Research Center, Kyoto University Hospital, 54 Shogoin Kawaharacho Sakyo, Kyoto 606-8507, Japan

Abstract. Thioredoxin (TRX) is a small multifunctional protein with a redox-active dithiol/disulfide in the conserved active site. Human TRX was originally identified as a cytokine-like factor in virus-transformed cells. The TRX family of proteins share the active site sequence: -Cys-Xxx-Xxx-Cys-. Intracellularly TRX scavenges reactive oxygen species and regulates various signal transduction pathways in cellular activation and apoptosis. TRX is induced and released from cells in response to oxidative stress. TRX levels in plasma or serum are good markers for oxidative stress. Extracellularly, TRX shows anti-inflammatory effects. Circulating TRX in plasma inhibits neutrophil extravasation into the inflammatory sites. More recently we found that TRX inhibits the expression and release of macrophage migration inhibitory factor (MIF), which is a classical proinflammatory cytokine and a member of TRX family. Administration of recombinant TRX protein may become a novel therapeutic strategy for acute inflammatory disorders. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 184–195

Thioredoxin was originally discovered in Escherichia coli as an electron donor to ribonucleotide reductase, which is an essential enzyme for DNA synthesis (Holmgren 1985, Laurent et al 1964). Human thioredoxin (TRX), which means cytosolic thioredoxin 1 here, was accidentally cloned as adult T cell leukaemia (ATL)-derived factor from the culture supernatant of ATL2 cells (Tagaya et al 1989, Yodoi & Tursz 1991). TRX consists of 105 amino acids and acts as a reductase via the dithiol/disulfide exchange reaction between two cysteines in the conserved active site (Nakamura 2005, Nakamura et al 1997, 2006). Reduced TRX scavenges hydrogen peroxide together with peroxiredoxin and reduces a disulfide in an oxidized protein. Oxidized TRX is reversibly reduced by NADPH and thioredoxin reductase (Fig. 1). TRX is induced by oxidative stress and acts as an anti-oxidant. However, the intracellular level of TRX is in the micromolar range, whereas intracellular glutathione is present at millimolar levels. Therefore, glutathione is a major intracellular antioxidant. However, TRX knockout mice are embryonic lethal and therefore TRX1 is an essential protein in mammals, because TRX plays a crucial 184

THIOREDOXIN

185 S S

Oxidative stress

S

S

Ligand S

Lipid raft

S

SH

Cell membrane

SH

Receptor oxidized TRX S

S

H 2O 2

NADPH Peroxiredoxin TRX reductase NADP

H 2O

SH SH SH SH

reduced protein S

reduced TRX

S

oxidized TRX S

oxidized protein

S

reduced TRX SH SH

ASK1 Transcription factors

ASK1 ASK1

NF-kB p53

p38MAPK

Nuclear membrane

AP-1 SH SH SH SH

SP-1

CRE

ORE ARE XRE

FIG. 1.

SH SH

TRX

TBP-2/VDUP-1

TRX Ref-1

TRX

TRX gene Apoptosis

Transcription

Redox regulation by thioredoxin.

role in signal transduction. For instance, TRX binds to apoptosis signal regulating kinase 1 (ASK1). When TRX is oxidized, ASK1 is dissociated from TRX and becomes a homodimer to transduce the apoptosis signal. DNA binding activities of transcription factors such as nuclear factor-κ B (NF-κ B), activator protein 1 (AP-1) and p53 are regulated by the reduction of the cysteine residues by TRX and/or redox factor 1 (Ref1). Moreover, TRX is released from cells in response to oxidative stress. Proteins that share a particular active site (-Cys-Xxx-Xxx-Cys-) and a similar three-dimensional structure are known as the thioredoxin family (Fig. 2). As mentioned above, cytosolic thioredoxin 1 is the focus of the present study and is abbreviated as TRX. Thioredoxin 2 is a mitochondrial-specific member of the family. Many members such as protein disulfide isomerase are in the endoplasmic reticulum and are involved in protein folding. Because of this, thioredoxin family members belong to the category of chaperone proteins. Recently, macrophage migration inhibitory factor (MIF), a classical proinflammatory cytokine, was also proved to be a member of the thioredoxin family (Kleemann et al 1998, Thiele & Bernhagen 2005). MIF is also released from cells and shows extracellular func-

186

NAKAMURA

:share –C-X1-X2-C- motif Cytosol

TRX1 Truncated TRX1 (Trx80)

TRX1

MIF (Macrophage migration inhibitory factor)

TRP14 TRP32/Txl-1 Txl-2 GRX1

TRX2 GRX2 Mitochondria

Nucleoredoxin GRX2

PDI CaBP1 ERp72 TMX ERdj5/JPDI

Tissue Specific SpTRX1 SpTRX2 SpTRX3 PC-TRP

Nucleus Endoplasmic Reticulum

:Involved in protein-folding

FIG. 2. Thioredoxin family.

tions. Since human TRX was originally identified as a cytokine-like factor, TRX is released from cells. Truncated TRX, which is composed of the 1–80 or 1–84 N-terminal polypeptide of TRX, is called Trx80 and also released from cells. Trx80 was originally identified as eosinophil cytotoxicity enhancing factor and shows more cytokine-like effects (Pekkari et al 2000, Pekkari & Holmgren 2004, Silberstein et al 1993). Trx80 is more mitogenic for lymphocytes, although it has no reducing activity. In contrast, wild-type TRX shows a growth-promoting effect in the presence of a reducing agent such as 2-mercaptoethanol (Nakamura et al 2006). TRX transgenic mice, in which human TRX is overexpressed under the β -actin promoter systemically, are more resistant to various oxidative stresses, including cerebral infarction (Takagi et al 1999), and survive longer than control mice (Mitsui et al 2002). TRX transgenic embryos are more resistant to high oxygen pressure (25% oxygen) and have less severe abnormalities when cultivated from E7.5 for 48 h (Kobayashi-Miura et al 2002, 2007). The generation of hydroxyl radicals was more attenuated in TRX-transgenic mice than in control mice after intratracheal instillation of diesel exhaust particles (Ahsan et al 2005). Collectively, the overexpression of TRX introduces the resistance to oxidative stress.

THIOREDOXIN

187

Previous reports have shown that TRX is involved in the growth of cancer (Arner & Holmgren 2006, Powis & Montfort 2001). It is debatable whether TRX is a friend or a foe for human diseases (Burke-Gaffney et al 2005). Intracellular TRX is induced in malignant tissues and associated with the resistance to anticancer agents. Therefore, some TRX inhibitors are under investigation as anticancer agents. In contrast, extracellular TRX does not show a growth promoting effect without a reducing agent (Nakamura et al 2006). Moreover, TRX becomes oxidized or dimeric in circulation, and the administration of TRX protein does not promote the growth of cancer tissue transplanted in nude mice nor inhibit the effect of anticancer agents (Ueda et al 2006). TRX is released from cells in response to oxidative stress. Transfected flagtagged TRX is released from cells in response to a sublethal amount of hydrogen peroxide, which is inhibited by an excess amount of exogenous TRX (Kondo et al 2004). Plasma/serum levels of TRX are a good marker for oxidative stress. In HIV infection, plasma levels of TRX are elevated and patients with higher plasma TRX die sooner than patients with normal plasma levels of TRX (Nakamura et al 2001a). In hepatitis C infection, serum levels of TRX and ferritin are good markers for the efficacy of interferon therapy (Sumida et al 2000). Patients with higher serum levels of TRX and ferritin suffer from more oxidative stress, caused by iron deposition in the liver, and are more resistant to interferon therapy. After phlebotomy, serum levels of TRX and ferritin are decreased and interferon therapy becomes more efficient (Nakashima et al 2005). In patients with acute respiratory distress syndrome, TRX levels are elevated in plasma and bronchoalveolar lavage fluids (Callister et al 2006). Plasma/serum levels of TRX are 10–30 ng/ml in healthy volunteers, rising to 40–140 ng/ml in patients suffering from oxidative stress. Tissue levels of TRX are 0.1–10 µg/ml. Lymphocytes contain rather a small amount (∼0.1 µg/mg protein) TRX, whereas placenta and liver contain rather high amounts (1∼10 µg/mg protein) TRX. TRX itself has a chemotactic activity. When more than 1 µg/ml TRX is administered intravenously, neutrophils are inhibited from extravasating into the inflammatory site (Nakamura et al 2001b). TRX suppresses p38 MAPK in neutrophils, L-selectin (CD62L) shedding from neutrophil cell surface and the adhesion of neutrophils on the endothelial cells. L-selectin (CD62L) shedding is mediated by a protease such as ADAM17 (Smalley & Ley 2005). Circulating TRX suppresses L-selectin shedding on circulating neutrophils (Nakamura et al 2007). When TRX is administered intraperitoneally, bleomycin-induced or inflammatory cytokine (IL2 + IL18)-induced interstitial pneumonia and pulmonary fibrosis is attenuated (Hoshino et al 2003). In a lipopolysaccharide (LPS)-induced acute lung injury model in rats, the continuous infusion of 8 mg/kg/day recombinant human TRX (rhTRX) suppressed neutrophil infiltration in the pulmonary interstitial space (Ueda et al 2006). In the sustained rat acute lung injury model caused by the

188

NAKAMURA

intratracheal instillation of bleomycin, the continuous infusion from day 3 for 96 hours suppressed the neutrophil recruitment in the airway (Nakamura et al 2007). Circulating TRX becomes oxidized and still shows the anti-inflammatory effects. This is in line with the previous in vitro study showing that extracellular TRX inhibits LPS-induced IL1β expression on macrophages (Billiet et al 2005). Based on these results, we started a translational research project in Kyoto University Hospital. The goal is to treat patients with acute lung injury, which is associated with neutrophil infi ltration in the lung tissues, by intravenous infusion of rhTRX protein. Preparing rhTRX at good manufacturing practice (GMP) grade and testing drug safety in rat and monkey using good laboratory practice (GLP) is expensive, so this work was supported by a venture company. We will soon start a phase I clinical trial to confirm the safety of this drug in healthy volunteers and then will start phase IIa to examine the efficacy for the treatment of acute lung injury. So far, no apparent toxicity has been observed in rat and monkey. To expand the clinical application, we have also studied the effect of rhTRX on acute pancreatitis, inflammatory bowel disease and a cigarette smoke-induced lung injury model. In inflammatory bowel disease, serum levels of TRX and MIF are elevated. We found that TRX inhibited the production and release of MIF (Tamaki et al 2006). MIF production was induced by 1 mg/ml LPS and 10 ng/ml IFNγ in THP1 cells, and was suppressed by rhTRX in a dose-dependent manner. We identified five plasma proteins that interact with TRX (Table 1). One of them is complement factor H (CFH). Choroidal neovascularization (CNV) is associated with age-related macular degeneration (AMD) which is a major reason for sight loss. Laser-induced CNV in mice was suppressed in TRX1-transgenic mice and by administration of rhTRX. Complement C3b deposition in the retina was also suppressed in TRX-transgenic mice and by administration of rhTRX. Moreover, rhTRX suppressed the infiltration of neutrophils and macrophages in the retina in the laser-induced CNV model. CFH plays a crucial role in the activation of the complement system and recently the risk of AMD was reported to be higher in CFH gene variants. In an in vitro assay, rhTRX and CFH additively sup-

TABLE 1

Plasma proteins interacting with TRX

Spot

kDa

Protein

1 2 3 4 5

30 40 50 71 150

Apolipoprotein A-I Scavenger receptor cysteine rich domain Fibrinogen, gamma polypeptide Albumin Complement factor H

THIOREDOXIN

189

MIF SH SH S

C35 S m u t ant TRX SH O H

S

S

SH SH

S

SH SH

T RX

Cell m em bran e

SH SH

SH SH

SH SH

Lipid raft

SH O H

FIG. 3. Possible interaction between extracellular thioredoxin and membranous target molecules including thioredoxin on the cell membrane.

pressed C3a production. Taken together, TRX is a good candidate for the prevention of and therapy for AMD. Recently we found that the C35S mutant can bind on the cell surface of ATL2 cells and enter into the cells, and this is suppressed by a lipid raft formation inhibitor methyl- β -cyclodextrin or excess amounts of exogenous wild-type TRX (Kondo et al 2007). These results suggest that extracellular TRX may interfere with the entrance of other proinflammatory signalling molecules through TRX and/or TRX family members on the cellular membrane, including the lipid rafts fraction (Fig. 3). In summary, extracellular or circulating TRX shows anti-inflammatory effects. The possible mechanisms may be explained by: (1) its anti-chemotactic effect; (2) anti-MIF function; (3) protease inhibition; (4) inhibition of adhesion through TRX in lipid rafts; and (5) regulation of complement activation (Fig. 4). The mechanisms by which extracellular TRX transduces the intracellular signals are not yet clarified. Shortly after this symposium, it was reported that CD30 is one of the target molecule for extracellular TRX (Schwertassek et al 2007). Studies are ongoing to clarify the extracellular functions of TRX.

190

NAKAMURA

GIF

MIF

TRX

Cys

Neutrophil

suppressive signaling

TRX protease inhibition ?

p 3 8 M A PK A c t i v a t i o n

Complement inactivation CD62L shedding Adhesion inhibition

TRX Complement factor H

Complement activation lipid raft

Endothelial cell

Adhe sio n o n endo thelial cells Transm igration

Inflammatory site FIG. 4. Possible mechanisms of anti-inflammatory effects of circulating thioredoxin.

References Ahsan M, Nakamura H, Tanito M, Yamada K, Utsumi H, Yodoi J 2005 Thioredoxin-1 suppresses lung injury and apoptosis induced by diesel exhaust particles (DEP) by scavenging reactive oxygen species and by inhibiting DEP-induced downregulation of Akt. Free Radic Biol Med 39:1549–1559 Arner ES, Holmgren A 2006 The thioredoxin system in cancer. Semin Cancer Biol 16:420– 426 Billiet L, Furman C, Larigauderie G et al 2005 Extracellular human thioredoxin-1 inhibits lipopolysaccharide-induced interleukin-1beta expression in human monocyte-derived macrophages. J Biol Chem 280:40310–40318 Burke-Gaffney A, Callister ME, Nakamura H 2005 Thioredoxin: friend or foe in human disease? Trends Pharmacol Sci 26:398–404 Callister ME, Burke-Gaffney A, Quinlan GJ et al 2006 Extracellular thioredoxin levels are increased in patients with acute lung injury. Thorax 61:521–527 Holmgren A 1985 Thioredoxin. Annu Rev Biochem 54:237–271 Hoshino T, Nakamura H, Okamoto M et al 2003 Redox-active protein thioredoxin prevents proinflammatory cytokine- or bleomycin-induced lung injury. Am J Respir Crit Care Med 168:1075–1083 Kleemann R, Kapurniotu A, Frank RW et al 1998 Disulfide analysis reveals a role for macrophage migration inhibitory factor (MIF) as thiol-protein oxidoreductase. J Mol Biol 280:85–102 Kobayashi-Miura M, Nakamura H, Yodoi J, Shiota K 2002 Thioredoxin, an anti-oxidant protein, protects mouse embryos from oxidative stress-induced developmental anomalies. Free Radic Res 36:949–956

THIOREDOXIN

191

Kobayashi-Miura M, Shioji K, Hoshino Y, Masutani H, Nakamura H, Yodoi J 2007 Oxygen sensing and redox signaling: the role of thioredoxin in embryonic development and cardiac diseases. Am J Physiol Heart Circ Physiol 292:H2040–H2050 Kondo N, Ishii Y, Kwon YW et al 2004 Redox-sensing release of human thioredoxin from T lymphocytes with negative feedback loops. J Immunol 172:442–448 Kondo N, Ishii Y, Kwon YW et al 2007 Lipid-raft-mediated uptake of cysteine-modified thioredoxin-1 enhances cisplatin-induced apoptosis by inhibiting the endogenous thioredoxin-1. Antioxid Redox Signal 9:1439–1448 Laurent TC, Moore EC, Reichard P 1964 Enzymatic synthesis of deoxyribonucleotides IV. Isolation and characterization of thioredoxin, the hydrogen donor from Echerichia coli B. J Biol Chem 239:3436–3444 Mitsui A, Hamuro J, Nakamura H et al 2002 Overexpression of human thioredoxin in transgenic mice controls oxidative stress and life span. Antioxid Redox Signal 4:693–696 Nakamura H 2005 Thioredoxin and its related molecules: update 2005. Antioxid Redox Signal 7:823–828 Nakamura H, Nakamura K, Yodoi J 1997 Redox regulation of cellular activation. Annu Rev Immunol 15:351–369 Nakamura H, De Rosa SC, Yodoi J, Holmgren A, Ghezzi P, Herzenberg LA 2001a Chronic elevation of plasma thioredoxin: inhibition of chemotaxis and curtailment of life expectancy in AIDS. Proc Natl Acad Sci USA 98:2688–2693 Nakamura H, Herzenberg LA, Bai J et al 2001b Circulating thioredoxin suppresses lipopolysaccharide-induced neutrophil chemotaxis. Proc Natl Acad Sci USA 98:15143–15148 Nakamura H, Masutani H, Yodoi J 2006 Extracellular thioredoxin and thioredoxin-binding protein 2 in control of cancer. Semin Cancer Biol 16:444–451 Nakamura T, Hoshino Y, Yamada A et al 2007 Recombinant human thioredoxin-1 becomes oxidized in circulation and suppresses bleomycin-induced neutrophil recruitment in the rat airway. Free Radic Res 41:1089–1098 Nakashima T, Sumida Y, Furutani M et al 2005 Elevation of serum thioredoxin levels in patients with nonalcoholic steatohepatitis. Hepatol Res 33:135–137 Pekkari K, Holmgren A 2004 Truncated thioredoxin: physiological functions and mechanism. Antioxid Redox Signal 6:53–61 Pekkari K, Gurunath R, Arner ES, Holmgren A 2000 Truncated thioredoxin is a mitogenic cytokine for resting human peripheral blood mononuclear cells and is present in human plasma. J Biol Chem 275:37474–37480 Powis G, Montfort WR 2001 Properties and biological activities of thioredoxins. Annu Rev Biophys Biomol Struct 30:421–455 Schwertassek U, Balmer Y, Gutscher M et al 2007 Selective redox regulation of cytokine receptor signaling by extracellular thioredoxin-1. EMBO J 26:3086–3097 Silberstein DS, McDonough S, Minkoff MS, Balcewicz-Sablinska MK 1993 Human eosinophil cytotoxicity-enhancing factor. Eosinophil-stimulating and dithiol reductase activities of biosynthetic (recombinant) species with COOH-terminal deletions. J Biol Chem 268:9138– 9142 Smalley DM, Ley K 2005 L-selectin: mechanisms and physiological significance of ectodomain cleavage. J Cell Mol Med 9:255–266 Sumida Y, Nakashima T, Yoh T et al 2000 Serum thioredoxin levels as an indicator of oxidative stress in patients with hepatitis C virus infection. J Hepatol 33:616–622 Tagaya Y, Maeda Y, Mitsui A et al 1989 ATL-derived factor (ADF), an IL-2 receptor/Tac inducer homologous to thioredoxin; possible involvement of dithiol-reduction in the IL-2 receptor induction. EMBO J 8:757–764 Takagi Y, Mitsui A, Nishiyama A et al 1999 Overexpression of thioredoxin in transgenic mice attenuates focal ischemic brain damage. Proc Natl Acad Sci USA 96:4131–4136

192

NAKAMURA

Tamaki H, Nakamura H, Nishio A et al 2006 Human thioredoxin-1 ameliorates experimental murine colitis in association with suppressed macrophage inhibitory factor production. Gastroenterology 131:1110–1121 Thiele M, Bernhagen J 2005 Link between macrophage migration inhibitory factor and cellular redox regulation. Antioxid Redox Signal 7:1234–1248 Ueda S, Nakamura T, Yamada A et al 2006 Recombinant human thioredoxin suppresses lipopolysaccharide-induced bronchoalveolar neutrophil infi ltration in rat. Life Sci 79: 1170–1177 Yodoi J, Tursz T 1991 ADF, a growth-promoting factor derived from adult T cell leukemia and homologous to thioredoxin: involvement in lymphocyte immortalization by HTLV-I and EBV. Adv Cancer Res 57:381–411

DISCUSSION Multhoff: Do you have any idea which lipid compound in the rafts is interacting with your TRX? Nakamura: We are not sure which component of the rafts is involved. When we treat the cells with NCD, which is a cholesterol-depleting agent, the binding is affected. Triantafilou: You showed that TRX can block extravasation and complement. It disables the whole innate immune system, in effect. Is this common for all chaperones, or is this unique to TRX? Nakamura: We think this depends on the concentration of TRX. TRX itself has some chemotactic activity in in vitro assays. When TRX levels in circulation are higher than in tissues, TRX can block extravasation of neutrophils. However, in the normal condition, the level of TRX in tissues is higher than in the circulation. Triantafilou: So you think this is concentration dependent? Nakamura: Yes. Recently we have found several other extracellular functions of TRX. Cheetham: The association with complement factor H is interesting. This is a major genetic risk factor for age-related macular degeneration (AMD). The major environmental risk factor we know is smoking. Also, oxidative stress has for a long time been thought to be a contributory factor in the disease process. Can you correlate any of your TRX levels in the circulation with the risk of developing AMD? Interestingly, the other major genetic risk factor for AMD is a polymorphism in the promoter of a gene called HTRA1, which has been referred to as a chaperone in some papers. Nakamura: Recently we have been looking at plasma levels of TRX in AMD patients. They show a higher level compared with controls. This may be predictive. Cheetham: So high TRX is bad in the patients, whereas in your mouse model it was good.

THIOREDOXIN

193

Nakamura: It is also dependent on the concentration of TRX. In general, leukocytes are drawn to higher concentrations of TRX. TRX transgenic mice express TRX systemically. Even the leukocytes contain higher levels of TRX in TRX transgenic mice. This is why it doesn’t go out into the inflammatory site. In normal mice or humans, when the TRX level in plasma is higher (up to 100 ng/ml), this indicates that they are suffering from chronic oxidative stress. However, it is not enough to suppress the inflammation. To suppress the inflammation, 10-fold higher amounts of TRX (>1 µg/ml) are necessary in circulation. Kregel: There are several forms of TRX in the cell. Can you match up what type of TRX you typically see circulating? Nakamura: Yes, it is TRX1. The mitochondrial type TRX2 might also be secreted into plasma. Kregel: You mentioned that a lot of this secretion is presumably associated with oxidative stress. Is this also associated with necrosis of cells? Is this the action for its secretion, or is it in cells that are just stressed but not necessarily necrotic? Nakamura: When we treated cells with a low (sublethal) level of hydrogen peroxide, the viability of the cells is good and they are not necrotic. But endogenous TRX is released. Intact cells can therefore release TRX in response to exogenous oxidative stress. Csermely: I have a question about the clinical trial. In human patients will you need continuous infusion of TRX as occurs in the rat experiments? Nakamura: Yes. When we start clinical trials, we will have to use continuous infusion, at least in the beginning. Feder: For TRX or the other circulating chaperones, what regulates their clearance from the circulation? Are there any specific mechanisms? Nakamura: In the case of TRX1, it is excreted into the urine. In the glomerulus it is excreted into the urine, but some TRX is reabsorbed by the proximal tubes. When the plasma level of TRX is too high, most of it is excreted. Feder: Is it picked up by a specific transporter in the convoluted tubule? Nakamura: We don’t know yet whether it is specific, but the process may be ATP-dependent. Feder: What about Hsp60 and 70 in the circulation? Henderson: We don’t know. Nakamura: The antigenicity of TRX is intact in the urine, so we can measure it. Gupta: How different are the mitochondrial and the cytosolic TRX? Are they encoded by the same genes? Nakamura: No. they are encoded by different genes. Mitochondrial thioredoxin is more similar to that of Escherichia coli. Cytosolic TRX1 has five cysteines; two in the active site and three additional ones. When it is oxidized, it makes a disulfide bond between the two cysteines in the active site and then another disulfide bond

194

NAKAMURA

between two other cysteines. Finally it becomes a homodimer by making a disulfide bond at the C-terminal cysteine. Mitochondrial TRX only has two cysteines in the active site and is thus more resistant to oxidation. Gupta: You mentioned that when you give exogenous TRX, it prevents secretion of intracellular TRX. What is the molecular mechanism underlying this? Nakamura: We don’t know. TRX can scavenge hydrogen peroxide extracellularly. I have a tentative hypothesis: TRX also exists on the cell membrane. This membrane TRX could play a role in the intracellular signal transduction when the cell suffers from exogenous oxidative stress. Perhaps membrane TRX is rapidly oxidized by extracellular oxidative stress. On the cell surface it may transduce a signal for endogenous TRX release. But when we add exogenous TRX in excess, the membrane TRX may be reduced. The redox state of the membrane-type TRX may play a role in the intracellular signal transduction. Lund: When you have these high levels of TRX in the blood, what does this do to the redox potential? Nakamura: Usually, in the circulation most of the TRX is already in the oxidized form. Even when we add more than 1 µg/ml exogenously, it is oxidized quickly. Panayi: You showed the effects of cigarette smoke on mice with COPD. In that experiment you administered TRX at the same time as you were giving smoke exposure, whereas most of our patients unfortunately come to us after 20 years of cigarette smoking. Can you show a therapeutic benefit in a mouse that has already developed COPD? Nakamura: That’s a good question. We are now studying whether various compounds can induce TRX when ingested. We don’t yet have a therapeutic model in an animal. Calderwood: Do you think these extracellular functions of TRX are related to its redox function? Nakamura: Recently we found that almost all thioredoxin in the circulation was in the oxidized form, so it may not be dependent on its reducing activity. The oxidized form itself showed several kinds of anti-inflammatory functions. Henderson: There are a number of reports from Holmgren’s group that a natural truncation of TRX, TRX80, has no redox function (Pekkari & Holmgren 2004). It also has a bizarrely different profi le of activation of monocytes. When you add large amounts of recombinant TRX do you generate this truncation? Nakamura: TRX80 is a truncated form, as you say, containing amino acids 1–80 of the 105 amino acids of TRX. We only see this when the macrophages are activated by PMA. They can release this. Henderson: Very early on in the work on IL1 secretion it was suggested that TRX and IL1 were secreted by the same pathway. Now we know what the IL1 pathway is, involving lysosomal endosomes. Is TRX secreted by this same pathway?

THIOREDOXIN

195

Calderwood: ABC transporters are involved in importing IL1 into the endosomes, which then move to the membrane, fuse and release IL1. This is stimulated by extracellular ATP. Nakamura: We don’t know the real mechanism for TRX secretion. Cheetham: You showed the redox and anti-inflammatory actions of TRX. Is there potential for direct cytoprotective functions of TRX? For example, Jose Sahel’s group have identified a protein that they call rod cell-derived cone viability factor (Leveillard et al 2004). This is a protein released by rod photoreceptors that leads to a preservation of cone cells. It has a TRX fold. They don’t know the mechanism of action for this protein but they assume it is acting as a direct cytoprotective factor. Is it possible TRX could be having a direct pro-viability effect on target cells rather than immune cells? Nakamura: I am not sure. More work needs to be done to clarify the relationship between the cytoprotective roles and reducing activity of TRX. References Leveillard T, Mohand-Said S, Lorentz O et al 2004 Identification and characterization of rodderived cone viability factor. Nat Genet 36:755–759 Pekkari K, Holmgren A 2004 Truncated thioredoxin: physiological functions and mechanism. Antioxid Redox Signal 6:53–61

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

HSP27: an anti-inflammatory and immunomodulatory stress protein acting to dampen immune function Carol L. Miller-Graziano, Asit De, Krzysztof Laudanski, Tara Herrmann and Sanjukta Bandyopadhyay University of Rochester Medical Center, Department of Surgery, 601 Elmwood Avenue—SURG, Rochester, NY 14642, USA

Abstract. The effects of HSP27 on human monocytes (MO) are predominantly antiinflammatory through preferential interleukin (IL)10 induction and by alteration of MO to immature dendritic cells (iDCs) or MO to macrophage (Mac) differentiation. Initial HSP27 inclusion in IL4+GM-CSF MO to iDC induction cultures allows Mac differentiation (CD14 ++ , CD16 + ), decreases iDC (CD1a + ) differentiation, and depresses DC induction of allogeneic T lymphocyte proliferation (MLR). HSP27 increased MO IL10 and M-CSF production but subsequent increased Mac differentiation isn’t responsible for depressed MO to iDC differentiation and function. Mac function after IL10 induced MO to Mac differentiation is also altered by HSP27 inclusion so that Mac phagocytic activity and scavenger receptor expression (CD163) are depressed. HSP27, in addition to immature DCs, doesn’t increase Mac differentiation but instead generates inhibitory DCs, which depress rather than stimulate T cell proliferation even during anti CD3+CD28 induction. Upon maturation, these HSP27-altered inhibitory DCs have increased production of the T cell and DC suppressive mediator, thrombospondin 1. HSP27’s anti-inflammatory and immunodepressive effects include deranging MO differentiation to both Mac and DCs, altering their receptor expression, and inducing production of inhibitory mediators such as thrombospondin-1 as well as IL10. These data suggest HSP27 belongs to a new group of ‘anti-danger signals’. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 196–211

Heat shock proteins (HSPs) are now recognized as having expanded physiological activities beyond their well described roles as molecular chaperones. Demonstrations that endogenous large molecular weight heat shock proteins could enhance monocyte innate responses to danger signals was predictable based on their strong pro-inflammatory cytokine induction of monocytes (MO)/macrophage (Mac) and their activation and maturation of dendritic cells (DCs) (Davies et al 196

HSP27, NO DANGER SIGNAL

197

2006, Johnson & Fleshner 2006, Osterloh et al 2007, Schmitt et al 2007). The direct ability of released HSPs to extracellularly signal MO and other cell types independent of lipopolysaccharide (LPS) contaminants is now clear (Johnson & Fleshner 2006, Kajino et al 2007). However, activation of innate and adaptive immune function must be balanced with damping of inflammation and leukocyte return to immune quiescence to prevent autoimmunity and/or hyper inflammatory pathology of normal tissue (Braun et al 2006, Schmitt et al 2007, Zanin-Zhorov et al 2006). Down-regulation of lymphocyte activation and proinflammatory cytokine production can result from MO/Mac production of inhibitory mediators such as IL10 and TGFβ (Wing & Sakaguchi 2006, Zanin-Zhorov et al 2006). We have previously published that the small HSP, HSP27, preferentially induces IL10 over proinflammatory cytokines suggesting a role for HSP27 to act as an ‘anti-danger signal’ controlling inflammation, a newly described concept (De et al 2000, Rosenblum et al 2006). Such an anti-inflammatory role for HSP27 has been suggested by others (Hashiguchi et al 2001, Njemini et al 2006). However, premature or excessive depression of inflammatory and adaptive immune responses can also augment susceptibility to infections and contribute to tumour cell progression, implicating a possible detrimental effect of HSP27 on immunity. Increased cancer cell expression of HSP27 mediates resistance to apoptosis and tumour progression (Lee et al 2005, Schmitt et al 2007). Thus, the negative prognosis of increased intracellular tumour cell levels of HSP27 was initially attributed primarily to its anti-apoptotic properties (Schmitt et al 2007). However, increased levels of circulating antibodies to HSP27 and decreased circulating HSP27 levels were associated with improved survival in breast cancer patients suggesting that extracellular HSP27 released either by cell apoptosis, necrosis, and/or stimulated induction could modulate immune responses favouring cancer progression (Conroy et al 1998, Niwa et al 2006, Njemini et al 2006). Myeloid DCs are pivotal cells not only in the induction of adaptive immune responses, but also in maintaining peripheral self-tolerance, and suppressing T cell responses compromising immunity in cancer and chronic infections (Reis e Sousa 2006, Smits et al 2005, Tarbell et al 2006). MO-derived DCs are the primary source of differentiating human DCs and Mac (Shortman & Naik 2007, Szekanecz & Koch 2007). Human MO can be differentiated to immature DCs by culture with IL4+GM-CSF, then matured to full antigen presenting potential by further culture with cytokines such as TNFα or by bidirectional interactions during culture with T lymphocytes (Reis e Sousa 2006, Tan & O’Neill 2005, Tarbell et al 2006). Conversely, MO can be differentiated to Mac by culture with M-CSF (Bakri et al 2005). We have previously shown that HSP27 can alter human MO to DC differentiation (Laudanski et al 2005). We have now assessed the immunomodulatory effects of HSP27 on MO–DC and MO–Mac differentiation characterizing changes in phagocytic function, receptor expression, and DC

198

MILLER-GRAZIANO ET AL

FIG. 1. Monocyte (MO) HSP27 treatment inhibits subsequent IL4+GM-CSF-induced MO to dendritic cell (DC) differentiation. Monocytes were cultured 5 days in the presence of IL4 (500 IU/ml) + 500 IU/ml GM-CSF (n = 19) with or without co-stimulation by 100 ng/ml LPS (n = 7) or 100–250 ng/ml HSP27 (n = 12). Expression levels of CD1a (iDC marker), CD14 (retained MO/Mac marker) and the DC-induced levels of T cell proliferation in MLR were compared for MO differentiated under the three DC differentiation conditions. *P < 0.01 **P < 0.05.

ability to stimulate T cells. When IL10 is added to MO cultures it induces their differentiation to Mac similar to the Mac differentiation induced by M-CSF (Bakri et al 2005). IL10 addition to MO simultaneously with IL4+GM-CSF prevents DC differentiation and drives Mac differentiation (Cohen et al 2006). Consequently, HSP27 with its potent MO IL10 induction capacity might also increase Mac differentiation and prevent DC development when HSP27 was simultaneously added to IL4+GM-CSF cultures of MO differentiating to DC. As we previously showed, addition of HSP27+IL4+GM-CSF to MO cultures depressed their differentiation to iDC as evidenced by decreased CD1a levels (Fig. 1) (Laudanski et al 2005). As expected, simultaneous HSP27 addition to IL4+GM-CSF MO differentiation cultures significantly increased IL10 levels (36.5 ± 6.6 to 241.7 ± 36.7 pg/ml, n = 8). HSP27 addition also increased the M-CSF levels in these MO–iDC differentiation cultures (127.4 ± 44.7 to 3383.2 ± 429.7 pg/ml n = 8). Increases in these Mac differentiating cytokines (IL10 and

HSP27, NO DANGER SIGNAL

199

M-CSF) might be partially mediating the HSP27 inhibition of MO–iDC differentiation by preferentially inducing macrophage differentiation. Initial LPS addition to IL4+GM-CSF cultures also inhibits MO–iDC differentiation by an IL10 dependent mechanism (Fig. 1) (Kwan et al 2007). However, our HSP27 preparation had been treated with polymyxin B-bound agarose beads and all HSP27-containing cultures also contained polymyxin B negating the possibility that an LPS contamination mediates this HSP27 effect. In addition, the HSP27 inhibition of MO–iDC differentiation could be prevented by preabsorption of the HSP27 preparation with anti-human HSP27 but not by anti-CD14. The reduction in the HSP27-treated iDC population’s ability to stimulate T cell proliferation in a mixed lymphocyte reaction (MLR) was paralleled by retention of CD14 and appearance of increased macrophage markers CD16 and CD163 (Fig. 1a,b). Finally, the up-regulation of CD86 and CD1a expression typical of MO differentiated to iDC was reduced in the HSP27-pretreated group (Fig. 1a,b). The data showing increased macrophage markers and increases in Mac differentiating cytokines like IL10 and M-CSF in populations of HSP27+IL4+GM-CSF-differentiated iDC suggests that IL10 and/or M-CSF-mediated increased Mac differentiation was totally responsible for the reduced T cell stimulatory effect of iDC populations differentiated in the presence of HSP27. We therefore assessed the ability of HSP27 to differentiate MO to Mac or to augment IL10 or M-CSF induced MO to Mac differentiation. Surprisingly, three-day culture with HSP27 (100– 500 ng/ml) did not induce MO to Mac differentiation (CD16 and CD163 increased) as compared to culture with M-CSF (50 IU/ml) or IL10 (100 IU/ml) (Table 1).

TABLE 1 HSP27 treatment neither induces nor augments MO-Mac differentiation during 3 day Mac differentiation % positive cells MO culture treatment a

CD16b

CD163c

None HSP27 M-CSF M-CSF + HSP27 IL10 IL10 + HSP27

36.6 ± 5.6 37.6 ± 6.3 62.6 ± 5.6*d 58.1 ± 10.3* 53.9 ± 11.4* 54.1 ± 12.7*

30.2 ± 8.8 25.7 ± 6.1 62.1 ± 5.5* 23.6 ± 5.9 46.2 ± 3.4* 58.2 ± 2.6*

a

MO cultured for 3 days in the presence of HSP27, M-CSF, M-CSF + HSP27, IL10, or IL10 + HSP27 to differentiated Mac as indicated by increased Mac markers (n = 6). b Up-regulation of CD16 (FcgRIII) as an indication of Mac differentiation detected by Flow cytometry. c Up-regulation of scavenger receptor, CD163, as an indication of MO to Mac differentiation. d * indicates P < 0.5 as compared to undifferentiated MO (MO cultured 3 days no stimuli) by ANOVA controlled for multiple variables.

200

MILLER-GRAZIANO ET AL

HSP27, NO DANGER SIGNAL

201

Interestingly, addition of HSP27 to the M-CSF or IL10 MO–Mac differentiation cultures actually inhibited M-CSF induced up-regulation of CD163 while failing to augment the IL10 mediated MO–Mac differentiation (Table 1). More strikingly, HSP27 inclusion depressed the increased phagocytosis exhibited by the Mac differentiated in the presence of IL10 or M-CSF alone (Fig. 2). These data suggest that HSP27 can affect MO differentiation to Mac as well as MO to iDC differentiation by altering receptor expression independent of its induced IL10’s or MCSF’s macrophage effect. Depressed CD163 expression by Mac depresses their ability to clear haemoglobin after injury or endotoxin shock (Schaer et al 2007). Thus, HSP27 depression of differentiating MO-derived Mac could increase the toxicity in injury or endotoxaemia by depressing the CD163 mediated clearance of haemoglobin. This HSP27 effect is in distinct contrast to the IL10 MO–Mac differentiating effect and indicates an IL10-independent activity of HSP27. IL10 addition to semi-differentiated iDCs does not induce Mac development but alters the DCs’ subsequent differentiation to potent antigen presenting cells by up-regulating inhibitory co-receptors, thereby depressing the DCs’ T cell activation capacity (Cohen et al 2006, Okazaki & Honjo 2006). Inhibitory DCs with up-regulated co-inhibitory receptors are critical for self-tolerance and resolution of inflammation (Okazaki & Honjo 2006, Sharpe et al 2007). One of the most potent co-inhibitory receptors, program death receptor ligand 1 (PDL1), belongs to the B-7 family of co-stimulatory/co-inhibitory receptors (Okazaki & Honjo 2006, Sharpe et al 2007). Up-regulated PDL1 and/or PD1 concomitant to depressed CD86 (co-stimulatory receptor) expression by DC results in T cell tolerance induction instead of activation by DCs (Okazaki & Honjo 2006, Sharpe et al 2007). Simultaneous addition of HSP27 and IL4+GM-CSF during iDC differentiation also increased MO expression of PDL1 simultaneously with down-regulation of CD86 (Fig. 3). These data imply that HSP27 release during inflammatory activation might also have DC tolerizing effects as well as DC differentiation inhibitory effects both independently and in concert with its induced IL10 effects. To assess the DC tolerizing effects of HSP27 and to compare these effects with the DC tolerizing effects of IL10, we employed a new experimental design which parallels that of experiments tolerizing DCs with IL10 (Okazaki & Honjo 2006, Sharpe et al 2007, Vlad et al 2005). HSP27 or IL10 were added at 48 h to cultures of MO and IL4+GM-CSF then the iDC differentiation cultures were continued for an additional 3 days. IL10 addition to such partially differentiated iDC cultures 䉳 FIG. 2. HSP27 inhibits phagocytic capacity of MO-derived macrophage (Mac). Ac LDL Alexa 488 uptake was compared by flow cytometry in MO differentiated to Mac by 3 days of culture with IL10 (100 IU/ml), or M-CSF (50 IU/ml) in the presence or absence of 100 ng/ml HSP27. One representative experiment of six is shown.

202

MILLER-GRAZIANO ET AL

FIG. 3. MO HSP27 treatment inhibits the subsequent IL4+GM-CSF-induced DC differentiation induced expression of CD86 but stimulates increased PDL1 expression. MO were differentiated to iDCs by IL4+GM-CSF (classic DCs as in Fig. 1) in the presence or absence of 250 ng HSP27. Solid bars represent the Mean Fluorescent Intensity (flow cytometry) of HSP27 untreated classic DCs. Crosshatched bars represent HSP27 pretreated DCs. P < 0.0.3 as compared to IL4+GM-CSF group for n = 8.

had previously been shown to increase DC expression of the co-inhibitory receptors PDL1 as well as ILT3 and ILT4 while decreasing their co-stimulatory receptor expression (Okazaki & Honjo 2006, Sharpe et al 2007, Vlad et al 2005). This increase in iDC inhibitory receptors concomitant with decreased expression of co-stimulatory receptors induces not T cell activation but T cell anergy and/or regulatory T cells (Sharpe et al 2007). The addition of HSP27 to semi-differentiated iDCs depressed their subsequent T cell activation capacity (Fig. 4) to a similar degree, as did IL10 treatment (Table 2). However, the effect of the IL10 vs. HSP27 modulation of differentiating DC receptor expression varied dramatically between IL10-treated and HSP27-treated iDC populations. As expected, IL10 addition depressed co-stimulatory receptor expression (CD86) and increased co-inhibitory receptor expression (ILT4+PDL1) (Table 3). The resultant imbalance between co-inhibitory and co-stimulatory receptors has been repeatedly demonstrated as the mechanism behind IL10-treated regulatory DC activity, implicated in generation of regulatory and anergic T cells (Sharpe et al 2007, Vlad et al 2005). In contrast to the results when HSP27 was added to MO before their differentiation to DC, PDL1 was not increased nor was CD86 decreased when HSP27 was added to iDCs. The depressed activation of T cell proliferation by HSP27-treated iDCs could not, however, be attributed to an IL10-like increase in ILT4 nor to an increased percentage of CD14 + Mac (Table 3). Nevertheless, the inhibitory effect

HSP27, NO DANGER SIGNAL

203

FIG. 4. HSP27 treatment of partially differentiated iDCs alters their subsequent induction effect on T lymphocyte proliferation from activation to inhibition. MO were differentiated to iDCs with IL4+GM-CSF for 24–30 h, then HSP27 (250 ng) or IL10 was added and culture continued for 3 additional days. The DCs were harvested and added to purified allogenic T cells in the presence of anti CD3+CD28. Proliferation of the T cells after 4 day culture measured in CPM by tritiated thymidine. Solid dark bars are T cell proliferation in the presence of IL4+GM-CSF differentiated DCs. Hatched bars are T cell proliferation when the iDCs were IL10 treated. Solid light bars are proliferation when the DC were HSP27 treated. Two representative experiments out of six are shown.

TABLE 2 Decreased T lymphocyte proliferation in cultures with allogeneic iDCs previously differentiated with HSP27 or IL10 (CPM ¥ 103 ) Donor T cells

T cells witha aCD3 + CD28

T cells withb Classic iDCs

T cells with c,d IL10 Tx iDCs

T cells withd,e HSP27 Tx iDCs

A B C D E F

38.5 21.3 36.7 – – –

48.6 29.9 42.9 34.2 33.0 27.5

34.8 21.7 30.8 21.7 15.0 20.1

39.6 15.8 26.7 23.2 12.6 17.7

a

Control 1 (Cnt1) T cells stimulated with anti-CD3 + CD28. Cont1 T cells co-cultured with IL4 + GM-CSF differentiated allogeneic Cnt2 DCs (classic DCs) and stimulated with anti-CD3 + CD28. c Donor T cells co-cultured with an allogeneic second donor’s IL10-treated (Tx) DCs as a control for induction of inhibitory DCs (IL10 added after 2 days of IL4 + GM-CSF differentiation). d Donor T cells co-cultured with an allogeneic second donor’s HSP27 Tx iDCs (500 ng/ml HSP27 added after 2 days of IL4 + GM-CSF differentiation). e Significant proliferation inhibition versus classic iDCs. P = <0.03 by both Wilcoxon-Signed Rank Test and ANOVA. b

204 TABLE 3

MILLER-GRAZIANO ET AL Phenotype of MO-derived DCs after delayed IL10 or HSP27 addition Classic iDCs a

HSP27 Tx iDCs c

Mean ± SEM

DC marker CD1a %pos CD14 %pos CD86 MFI CD40 MFI ILT4 MFI PDL1% HLA-DR MFI

IL10 Tx iDCs b

76.2 ± 2.0 ± 77.4 ± 10.3 ± 64.9 ± 8.8 ± 97.7 ±

5.3 0.4 19.2 1.5 7.8 1.1 16.6

71.8 ± 5.6 ± 21.9 ± 12.2 ± 95.9 ± 21.4 ± 44.5 ±

5.8 1.9d 4.4 e 1.8 13.2d 3.9e 5.5e

74.7 ± 1.6 ± 74.0 ± 8.6 ± 60.9 ± 7.2 ± 85.5 ±

5.5 0.3f 16.8 1.3 7.2f 1.1g 13.7g

MO were differentiated to iDC by 5 day co-culture with IL4 + GM-CSF (Classic iDC). IL10 added to partially differentiated iDCs (MO culture 2 days with IL4 + GM-CSF). n = 9. HSP27 added to iDCs (MO differentiated to iDCs for 2 days with IL4 + GM-CSF, then 250 µg HSP27 was added). n = 9. d P < 0.05 in comparison to classic DCs (ANOVA). e P < 0.02 in comparison to classic DCs. f P < 0.05 in comparison to IL10 TX iDCs. g P < 0.02 in comparison to IL10 TX iDCs. a

b c

of HSP27-treated iDCs on T cell proliferation could still be due to increase of a multitude of other untested inhibitory iDC receptors described in the literature which act independently of IL10 (Braun et al 2006, Rosenblum et al 2006). Alternatively, or in addition, the HSP27-triggered iDC production of immunosuppressive mediators such as indoleamine-2,3-dioxygenase (IDO) and/or thrombospondin 1 (TSP1) could also inhibit T cell activation (Krispin et al 2006, Masli et al 2006, Orabona et al 2006). Heat shock is reported to increase TSP1 levels and TSP1 is emerging as a major mediator of inhibitory DC activation and function (Doyen et al 2003, Grimbert et al 2006, Kang et al 2006, Krispin et al 2006, Masli et al 2006). Consequently, we assessed whether treating iDCs with HSP27 could alter their TSP1 production levels. HSP27 was added to cultures of partially differentiated iDCs (2 days of IL4+GM-CSF). After 3 additional days of differentiation, the semi-mature DCs were either isolated and assessed for TSP1 (TSP1 ELISA, Chemicon) or continued in fresh culture media for an additional 2 days with TNFα as a maturation stimulus and then assessed for TSP1 levels. Increases in TSP1 levels over 5 days in IL4+GMCSF-cultured DCs (classic DCs) were detected. However, TNFα maturation of the HSP27-differentiated iDCs resulted in a significant increase in TSP1 release into the culture supernates versus the TSP1 levels detected in supernates of TNFα matured classic DCs (Fig. 5). Both HSP27-treated and classic mature DC populations had similar increases in CD83 expression (the marker for mature DC). These

HSP27, NO DANGER SIGNAL

205

FIG. 5. Production of thrombospondin 1 by HSP27-treated DCs either matured by adding exogenous TNF or without added maturation stimuli. MO were differentiated to DC for 30 h, then HSP27 or medium added. Culture was continued for three additional days. On day 5 TNFα or medium was added for an additional 24–30 h of culture. Culture supernatants were collected at the end of 5 days of DC differentiation and at the end of maturation period and assessed for TSP1 (ELISA). #P < 0.02 for an n of 4.

data suggest that the HSP27-treated iDCs matured by stimulated T cells in our DC plus T cell experiments could also be inhibiting T cell activity by increasing their production of TSP1. Other possibilities still need to be explored which could also contribute to the deviation of DC function from stimulatory to inhibitory for T cells. Our data suggest that like other HSPs, HSP27 can stimulate through the TLR4 receptor ( Johnson & Fleshner 2006, Zanin-Zhorov et al 2006). Antibodies to TLR4 but not TLR2 or CD14 can partially reverse the HSP27-increased p38 activation in MO and ameliorate the HSP27-induced DC differentiation defect. Stimulation of semi-mature iDCs through TLR4 by LPS is reported to result in exhausted DCs which cannot stimulate T cells (Kajino et al 2007, Kwan et al 2007). Consequently, a similar exhausted DC phenotype might be stimulated by HSP27 triggering of TLR4. This exhaustion mechanism is dependent on IL10 and M-CSF sequential interactions with iDCs (Kajino et al 2007, Kwan et al 2007). Both of these cytokines are triggered by HSP27 treatment. Therefore, in HSP27 cultures, the combined effect of IL10 and M-CSF would be acting on differentiating DCs versus solely an IL10 effect. Dissecting a single mechanism by which HSP27 can mediate T cell suppression is complicated by the bidirectional interplay of iDCs and T cells in mediating

206

MILLER-GRAZIANO ET AL

reciprocal functional alterations. Regulatory T cells could be stimulated in cocultures of HSP27-treated iDCs. TSP1 released by iDCs could stimulate T cell CD47 and IL10 triggering of regulatory T cells (Grimbert et al 2006). These induced regulatory T cells could then inhibit DC function on the remaining T cell population (Braun et al 2006, Grimbert et al 2006). Our HSP27 experiments suggest multiple mechanistic possibilities for HSP27-mediated down-regulation or alteration of inflammatory and specific immune responses, which remain to be explored. In conclusion, we have demonstrated that HSP27 can inhibit MO to iDCs and MO to Mac differentiation both by an IL10-dependent and an IL10independent pathway. The HSP27-mediated increase in MO PDL1 expression and decrease in CD86 levels will depress MO activation of NK and T lymphocytes, even beyond its depression of T cell activation by preventing MO to DC differentiation. The slight increase in Mac differentiation that results from HSP27 treatment of MO before iDC differentiation is most likely a default event rather than a HSP27-directed increase in Mac differentiation. HSP27treated MO also had altered IL10 or M-CSF-induced MO–Mac differentiation. The depressed Mac phagocytic function mediated by HSP27 during MO–Mac differentiation occurred even when CD163 was not decreased (IL10-mediated MO–Mac differentiation). This decreased Mac phagocytic capacity will increase susceptibility to bacterial infection. Meanwhile, the HSP27-decreased CD163 expression in M-CSF-differentiated Mac will depress haemoglobin clearance and result in increased toxicity after injury or endotoxaemia (Schaer et al 2007). The modulation of the ability of DCs to activate T cells resulting when partially matured iDCs are triggered by HSP27 could mediate beneficially decreased T cell activation in the case of massive injury and release of self proteins. Alternatively, this same HSP27-induced decrease in T cell activation could augment pathology by damping an immune response in the face of malignancy or chronic infection. HSP27 induction of DCs to produce the newly described immunosuppressive mediator, TSP1, may represent an additional mechanism for its down modulation of highly activated immune responses. HSP27 seems to have a unique anti-inflammatory, negative danger signal activity in host defence. This down-regulation effect is related to the HSP27 modulation of MO receptor expression, alteration of MO–DC differentiation and maturation, its derangement of MO–Mac differentiation, and its capacity to induce unique inhibitory mediator production.

Acknowledgements Supported by NIH NIGMS Grant R01-GM036214 and American Cancer Society Grant RSG-06-266-01-LIB.

HSP27, NO DANGER SIGNAL

207

References Bakri Y, Sarrazin S, Mayer UP et al 2005 Balance of MafB and PU.1 specifies alternative macrophage or dendritic cell fate. Blood 105:2707–2716 Braun D, Galibert L, Nakajima T et al 2006 Semimature stage: a checkpoint in a dendritic cell maturation program that allows for functional reversion after signal-regulatory protein- α ligation and maturation signals. J Immunol 177:8550–8559 Cohen N, Mouly E, Hamdi H et al 2006 GILZ expression in human dendritic cells redirects their maturation and prevents antigen-specific T lymphocyte response. Blood 107:2037– 2044 Conroy SE, Sasieni PD, Amin V et al 1998 Antibodies to heat-shock protein 27 are associated with improved survival in patients with breast cancer. Br J Cancer 77:1875–1879 Davies EL, Bacelar MM, Marshall MJ et al 2006 Heat shock proteins form part of a danger signal cascade in response to lipopolysaccharide and GroEL. Clin Exp Immunol 145: 183–189 De AK, Kodys KM, Yeh BS, Miller-Graziano C 2000 Exaggerated human monocyte IL-10 concomitant to minimal TNF- α induction by heat-shock protein 27 (Hsp27) suggests Hsp27 is primarily an antiinflammatory stimulus. J Immunol 165:3951–3958 Doyen V, Rubio M, Braun D et al 2003 Thrombospondin 1 is an autocrine negative regulator of human dendritic cell activation. J Exp Med 198:1277–1283 Grimbert P, Bouguermouh S, Baba N et al 2006 Thrombospondin/CD47 interaction: a pathway to generate regulatory T cells from human CD4+CD25- T cells in response to inflammation. J Immunol 177:3534–3541 Hashiguchi N, Ogura H, Tanaka H et al 2001 Enhanced expression of heat shock proteins in activated polymorphonuclear leukocytes in patients with sepsis. J Trauma 51:1104–1109 Johnson JD, Fleshner M 2006 Releasing signals, secretory pathways, and immune function of endogenous extracellular heat shock protein 72. J Leukoc Biol 79:425–434 Kajino K, Nakamura I, Bamba H, Sawai T, Ogasawara K 2007 Involvement of IL-10 in exhaustion of myeloid dendritic cells and rescue by CD40 stimulation. Immunology 120:28– 37 Kang JH, Kim SA, Hong KJ 2006 Induction of TSP1 gene expression by heat shock is mediated via an increase in mRNA stability. FEBS Lett 580:510–516 Krispin A, Bledi Y, Atallah M et al 2006 Apoptotic cell thrombospondin-1 and heparin binding domain lead to dendritic cell phagocytic and tolerizing states. Blood 108:3580–3589 Kwan WH, Boix C, Gougelet N, Fridman WH, Mueller CG 2007 LPS induces rapid IL-10 release by M-CSF-conditioned, tolerogenic dendritic cell precursors. J Leukoc Biol 82: 133–141 Laudanski K, De AK, Miller-Graziano CL 2005 HSP-27 as an anti-inflammatory protein. In: Henderson B, Pockley G (eds) Molecular chaperones and cell signaling, Cambridge University Press, p 220–233 Lee SA, Ndisang D, Patel C et al 2005 Expression of the Brn-3b transcription factor correlates with expression of HSP-27 in breast cancer biopsies and is required for maximal activation of the HSP-27 promoter. Cancer Res 65:3072–3080 Masli S, Turpie B, Streilein JW 2006 Thrombospondin orchestrates the tolerance-promoting properties of TGFβ -treated antigen-presenting cells. Int Immunol 18:689–699 Niwa M, Hotta K, Hara A et al 2006 TNF-alpha decreases hsp 27 in human blood mononuclear cells: involvement of protein kinase c. Life Sci 80:181–186 Njemini R, Lambert M, Demanet C, Mets T 2006 The effect of aging and inflammation on heat shock protein 27 in human monocytes and lymphocytes. Exp Gerontol 41:312–319 Okazaki T, Honjo T 2006 The PD-1-PD-L pathway in immunological tolerance. Trends Immunol 27:195–201

208

MILLER-GRAZIANO ET AL

Orabona C, Puccetti P, Vacca C et al 2006 Towards the identification of a tolerogenic signature in IDO-competent dendritic cells. Blood 107:2846–2854 Osterloh A, Kalinke U, Weiss S, Fleischer B, Breloer M 2007 Synergistic and differential modulation of immune responses by Hsp60 and lipopolysaccharide. J Biol Chem 282:4669– 4680 Reis e Sousa C 2006 Dendritic cells in a mature age. Nat Rev Immunol 6:476–483 Rosenblum MD, Yancey KB, Olasz EB, Truitt RL 2006 CD200, a “no danger” signal for hair follicles. J Dermatol Sci 41:165–174 Schaer CA, Vallelian F, Imhof A, Schoedon G, Schaer DJ 2007 CD163-expressing monocytes constitute an endotoxin-sensitive Hb clearance compartment within the vascular system. J Leukoc Biol 82:106–110 Schmitt E, Gehrmann M, Brunet M, Multhoff G, Garrido C 2007 Intracellular and extracellular functions of heat shock proteins: repercussions in cancer therapy. J Leukoc Biol 81:15–27 Sharpe AH, Wherry EJ, Ahmed R, Freeman GJ 2007 The function of programmed cell death 1 and its ligands in regulating autoimmunity and infection. Nat Immunol 8:239–245 Shortman K, Naik SH 2007 Steady-state and inflammatory dendritic-cell development. Nat Rev Immunol 7:19–30 Smits HH, de Jong EC, Wierenga EA, Kapsenberg ML 2005 Different faces of regulatory DCs in homeostasis and immunity. Trends Immunol 26:123–129 Szekanecz Z, Koch AE 2007 Macrophages and their products in rheumatoid arthritis. Curr Opin Rheumatol 19:289–295 Tan JK, O’Neill HC 2005 Maturation requirements for dendritic cells in T cell stimulation leading to tolerance versus immunity. J Leukoc Biol 78:319–324 Tarbell KV, Yamazaki S, Steinman RM 2006 The interactions of dendritic cells with antigenspecific, regulatory T cells that suppress autoimmunity. Semin Immunol 18:93–102 Vlad G, Cortesini R, Suciu-Foca N 2005 License to heal: bidirectional interaction of antigenspecific regulatory T cells and tolerogenic APC. J Immunol 174:5907–5914 Wing K, Sakaguchi S 2006 Regulatory T cells as potential immunotherapy in allergy. Curr Opin Allergy Clin Immunol 6:482–488 Zanin-Zhorov A, Cahalon L, Tal G, Margalit R, Lider O, Cohen IR 2006 Heat shock protein 60 enhances CD4+ CD25+ regulatory T cell function via innate TLR2 signaling. J Clin Invest 116:2022–2032

DISCUSSION Gaston: What is the concentration of HSP27 needed for these effects? Miller-Graziano: We are using 100–250 ng/ml. Gaston: Have you tried using your inhibitory DCs in Transwell experiments? Are cell–cell contacts needed between those cells and T cells to turn them off? Miller-Graziano: Yes. Cell–cell contact is needed, which is why I am emphasizing the receptor interaction. Remember, thrombospondin is bound on to the CD47 receptor. It is presented on a plate, and this is a concentration-dependent mechanism. It is like the old macrophage story, but in this case it is the DCs coming up to the T cells and giving them the kiss of death, as opposed to the kiss of life. Conway de Macario: What was your rationale for choosing HSP27, as opposed to other small chaperones?

HSP27, NO DANGER SIGNAL

209

Miller-Graziano: We chose HSP27 for historical reasons. We thought by giving HSP27 to patients’ dysfunctional macrophages, we could reverse the blockade or deficit activity that we see in the macrophage. HSP27 has kinase activity downstream of p38, and our original work was looking at the signal transduction to see whether it was activated. Simultaneously with our findings, data were published showing that breast cancer patients who had elevated circulating levels of HSP27 were much more immunosuppressed. Then it was shown that the cancer patients with better prognoses were those who developed antibodies against HSP27. This made me think HSP27 had some sort of immunomodulatory effect. When I saw our data on peripheral monocyte IL10 induction by HSP27, I thought this might be an important immunosuppressive mechanism. At the time, IL10 was thought to be the critical factor in immunosuppression in trauma patients. So I rushed off to work with HSP27. Conway de Macario: Have you used another HSP as a control to show specificity for HSP27? For example, HSP60 and HSP10 have been reported to have some sort of role in carcinogenesis and related pathological pathways (Capello et al 2007) Miller-Graziano: We have done a lot of antibody work with HSP27 to show that it is specific, and it isn’t blocked by antibody to an irrelevant murine Hsp25. We haven’t tried another human HSP, though. Based on the literature, I suspect that Hsp10 may have the same immunoinhibitory effect, though. The main reason I didn’t try the large HSPs is because all the literature data stated that they were pro-inflammatory. Quintana: TGFβ can stay as a membrane-bound form in something called LAP. A recent paper (Hyytiainen et al 2004) describes that there is a subpopulation of DCs that express LAP on the membrane. One of the known activators of this latent form of TGFβ is thrombospondin. This might be a link to your mechanism. Miller-Graziano: I agree. But the Montreal group’s data are convincing that there is a direct effect of thrombospondin also. IL10 can induce TSP1, which is also induced by TGFβ. TGFβ is activated from the cell-bound LAP form. It’s an amplifying loop. Quintana: What surprises me is that contact dependency is still needed. In your system the T cells are activated by the plate-bound CD28/CD3-specific antibodies. The DCs are secreting something and it should work even in a Transwell system. Why do you think that cell–cell interaction is still needed? Is there an unknown molecule playing a role? Miller-Graziano: I think both IL10 and TGFβ may be membrane bound to the DC. We have some data on this but I don’t think it’s convincing enough to show. We are making things hard for ourselves. The supernatants are immunosuppressive if we use a CD3 plus CD4 T cell stimulation instead of CD3 plus CD28. The reason I picked an anti CD3+CD28 stimulation is that this breaks anergy. In order

210

MILLER-GRAZIANO ET AL

to define direct DC inhibition we need a powerful inhibitory system. The reason for the contact requirement may be to get the cell–cell interaction’s augmentation of suppression necessary to counteract the strength of the anti CD3 plus 28 signalling. We wanted to demonstrate a real DC inhibitory action rather than a less potent DC stimulation. Corrigall: Have you looked at the up-regulation of indolamine deaminase? Miller-Graziano: The graduate student whose project this was became ill. We are trying to look at this, but I don’t really want to speculate yet. Corrigall: What about intracellular CTLA4? Miller-Graziano: This is up-regulated on the T cells along with PD1. The problem here is because of this T cell CTLA4–DC interaction, we don’t know who is upregulating who. Macario: How many genes are there for HSP27 in humans? Miller-Graziano: Supposedly only two. I am being careful here: I have heard so much new work I’ve not heard of before at this meeting I don’t want to get it wrong! Supposedly one of them is the mitochondrial form. Csermely: What is the phosphorylation status of the HSP27? Miller-Graziano: It is phosphorylated. Pockley: You showed that the HSP27-treated DCs inhibited proliferation of your responding T cell populations. Have you looked at the cytokine profi les of these T cells? Miller-Graziano: Yes. When we collected the T cells’ iDC supernates, there is a high level of IL10 and M-CSF, and there is supposedly some kind of sequential interaction that takes place between M-CSF and IL10 which then affects both DCs and T cells. The IL2 levels are slightly lower, but not significantly so. If in fact we are making adaptive Tregs, we would expect that early on. Pockley: You need to do some restimulation experiments. Miller-Graziano: We haven’t done that. You don’t get an expansion of Tregs if there is no IL2. I shouldn’t have been so surprised that there wasn’t any IL2 down-regulation. There is very little IL12, and IFNγ production is reduced. We don’t know whether the failure of the cells to proliferate is a real inhibition, or just that they didn’t get going. The real ongoing experiment is the restimulation experiment. Pockley: Thymidine uptake gives you a global picture of what is happening, but have you looked to see whether the DCs have preferential effects on different T cell subsets? You may have a reduction in thymidine uptake as an example, but you might have quite a highly proliferating sub-population with a suppressed subpopulation within that. Miller-Graziano: We are trying to look with CSFE. We are doing these experiments by FACS to see whether there is a difference in the subsets. Remember, these are human cells and there are never enough of them. We can’t re-bleed someone for 8 weeks after we have taken a draw from them.

HSP27, NO DANGER SIGNAL

211

Calderwood: Do you have any idea about the receptor signalling system for binding Hsp27? Miller-Graziano: The only data we have are that we can block with antibodies to TLR4. We did FRET, and it didn’t show a distinctive TLR4–HSP27 association. I think we are looking at a complex here. Antibody against CD14 has no effect whatsoever. There is a very different modulation of the receptor expression. I don’t know what receptors I should be looking for. CD47 is up-regulated, but I don’t think this is binding HSP27. I think it is part of a whole HSP27 modulation of the cell surface. Lund: You said that you required intact whole protein to see the effects you are seeing. Do you know anything about the oligomeric state of the HSP? Small HSPs have very complex equilibria between monomers, dimers, 8-mers and so on. In the higher oligomeric forms they will bind in the capped state of the proteins. When you are putting two things together, is it conceivable that you are finishing with a lot of one protein inside a HSP27 cage? Miller-Graziano: I don’t know: it’s not something I like to think about. Let’s turn that around: how would I approach making sure this isn’t the case? Csermely: If you phosphorylate small HSPs do they change their complex formation? Cheetham: It depends on the particular small HSP you are looking at. Lund: One of the roles of small HSP is to act as ‘safari parks’ where you put the endangered proteins so they are reactivated for folding by the other chaperones inside the cell. Things that become partially denatured will get soaked up by the small HSPs. In your case there may be a reaction taking place in your exogenously added proteins between the small HSP and the other cytokines, which may change the property of the cytokines. Miller-Graziano: I haven’t thought about this. Quintana: Does it also work in mice? Miller-Graziano: HSP27 is species specific. Hsp25 seems in vivo to have some effects (Masubuchi et al 2003, Park et al 2002). References Cappello F, Czarnecka AM, La Rocca G, Di Stefano A, Zummo G, Macario AJL 2007 Hsp60 and Hsp10 as antitumor molecular agents. Cancer Biol Ther 6:487–489 Hyytiainen M, Penttinen C, Keski-Oja J 2004 Latent TGF-beta binding proteins: extracellular matrix association and roles in TGF-beta activation. Crit Rev Clin Lab Sci 41:233–264 Masubuchi Y, Bourdi M, Reilly TP, Graf ML, George JW, Pohl LR 2003 Role of interleukin-6 in hepatic heat shock protein expression and protection against acetaminophen-induced liver disease. Biochem Biophys Res Commun 304:207–212 Park KM, Kramers C, Vayssier-Taussat M, Chen A, Bonventre JV 2002 Prevention of kidney ischemia/reperfusion-induced functional injury, MAPK and MAPK kinase activation, and inflammation by remote transient ureteral obstruction. J Biol Chem 277:2040–2049

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

BiP, an anti-inflammatory ER protein, is a potential new therapy for the treatment of rheumatoid arthritis Gabriel S. Panayi and Valerie M. Corrigall Academic Department of Rheumatolog y, Division of Immunolog y, Infection and Infl ammatory Disease, King’s College London School of Medicine at Guy’s, King’s College and St. Thomas’ Hospitals, King’s College London, Guy’s Hospital, London SE1 9RT, UK

Abstract. The endoplasmic reticulum chaperone and stress protein BiP has hitherto been considered as having only crucial intracellular cell protective functions. However, we have shown that BiP can be present in the extracellular environment and that it binds to a putative but as yet uncloned cell surface receptor. It will stimulate human monocytes via this receptor to express a gene profi le that is anti-inflammatory. It will generate T cells with a regulatory function from human peripheral blood most likely by altering dendritic cell development. Intravenous BiP will both prevent and treat ongoing collagen induced arthritis in the DBA/1 mouse. Part of the suppression of arthritis is linked to interleukin (IL)4 as BiP-specific lymph node and spleen cells from these mice secrete IL4, and BiP has no suppressive effect on collagen induced arthritis in IL4 knockout mice. Lymph node and spleen cells isolated from mice administered intravenous BiP will suppress arthritis when transferred intravenously into recipient arthritic mice without any further BiP having to be given. Thus, both in vitro work with human peripheral blood mononuclear cells and in vivo work in the collagen arthritis model lead to the conclusion that BiP induces regulatory cells. Finally, intravenous BiP will ablate the inflammatory cell infi ltrate and inflammatory cytokine expression in rheumatoid synovial membrane tissue transplanted subcutaneously into SCID mice. The conclusion from all this experimental work is that BiP may be a novel therapy for the treatment of patients with rheumatoid arthritis. 2008 The biolog y of extracellular molecular chaperones. Wiley, Chichester (Novartis Foundation Symposium 291) p 212–220

BiP has been recognized as having an important intracellular role in protecting cells from stress. Our work has shown that secreted BiP also has potent extracellular anti-inflammatory and immunomodulatory properties. One could hypothesise that if cellular stress reached a certain level at which intracellular protective mechanisms were no longer adequate, then extracellular protective mechanisms could be turned on to provide additional cellular and tissue protection (Panayi et al 2004). 212

IMMUNOMODULATION BY BIP

213

Discovery of BiP BiP was originally discovered using a proteomics approach when we were looking for novel rheumatoid arthritis (RA) autoantigens. However, there was a feature of BiP that suggested that all was not as it seemed. T cell proliferation, and then only to a low level, was observed predominantly in rheumatoid synovial fluid cells but not in the paired peripheral blood. The cytokine profi le of the supernatants was unusual in that they contained no interleukin (IL)2 and little or no interferon (IFN) γ. BiP did not induce arthritis in several strains of mice or rats. At this point, we reasoned that the relative inability of BiP to stimulate T cell responses may be because it has regulatory properties. We then showed that intravenous BiP could prevent collagen-induced arthritis (CIA) in DBA/1 mice and was able to divert the anti-collagen antibody response from Th1 to Th2 (Corrigall et al 2001). Hence, we postulated that BiP could stimulate a Th2 T cell response that thereby suppressed the Th1 response driving CIA. Human studies with BiP T cells In parallel with these studies in CIA, we studied the effect of BiP on human T cell responses. BiP-specific T cell clones were established from peripheral blood but with some difficulty perhaps related to the fact that BiP poorly stimulates T cell proliferation. The first surprise, in view of the difficulty usually encountered in establishing human CD8 + clones, was that the majority of the clones were CD8 + (Bodman-Smith et al 2003). They produced a very interesting pattern of cytokines of interleukin IL4, -5 and -10 in various combinations. These are the characteristics of the Th2 T cell subset. However, we have not investigated whether these clones are able to suppress or down modulate Th1 cells. However, we have shown that BiP will suppress T cell proliferation in response to the recall antigen, tuberculin purified protein derivative (PPD) (Corrigall et al 2004). Monocytes Dr Valerie Corrigall has extensively studied the effects of BiP on human peripheral blood monocytes. She showed that BiP suppresses tumour necrosis factor (TNF) α secretion but potently stimulates the production of IL10 (Corrigall et al 2005). Hence, BiP can stimulate IL10 secretion by two pathways: from BiP-specific T cells and from monocytes. The secretion of IL10 will lead to the de-activation of monocytes from cells able to produce pro-inflammatory cytokines, such as TNFα , that have been called ‘alternatively activated’ macrophages (Gordon 2003).

214

PANAYI & CORRIGALL

However BiP itself, independently of the stimulated production of IL10, may directly stimulate monocytes down this pathway as neutralization of IL10 does not lead to abrogation of all of the effects of BiP (Corrigall et al 2004). Analysis of gene transcription at 24 h after BiP stimulation shows decreased transcription of pro-inflammatory cytokine and chemokine genes, reduced transcription of co-stimulatory molecules such as HLA-DR and CD86, and increased transcription of anti-inflammatory genes such as TNF receptor II and IL1 receptor antagonist (Corrigall et al 2005). These changes collectively support the concept that BiP induces ‘alternative activation’ of human monocytes thereby leading to anti-inflammatory effects. Perhaps these alterations of monocyte gene transcription could explain some of the beneficial effects of BiP on CIA (see above). How does BiP stimulate monocytes? Corrigall showed that 95% of monocytes, 10–30% of B cells, and up to 10% of T cells in human peripheral blood, and 80% of primary cultures of RA fibroblast-like synoviocytes could bind fluoresceinconjugated BiP (Corrigall et al 2003). Apart from monocytes, there is no published work on the effect of directly stimulating cells via their constitutively expressed BiP receptor. Equally, there are no data on the structure of the receptor. Such knowledge would greatly facilitate our analysis of the mode of action of BiP by shedding light on intracellular signalling pathways, although we have shown in preliminary studies that BiP may be acting partially via the MAPK p38 pathway (Corrigall et al 2004). Dendritic cells Monocytes are able to develop into different end-stage cells, such as antigen presenting dendritic cells (DCS) or osteoclasts, depending on the cytokine milieu in which they are cultured. We have shown that BiP inhibits the development of immature dendritic cells (iDC) from monocyte precursors when cultured with IL4 and granulocyte macrophage-colony-stimulating factor (GM-CSF) (manuscript in preparation). These BiP-induced iDCs have decreased expression of HLA-DR, CD86 and persisting expression of CD14, a monocyte marker, as compared to ‘physiological’ iDCs. On maturation with lipopolysaccharide these DC have increased numbers of indolamine-2,3-dioxygenase (IDO)-containing cells and do not up-regulate CD83. Not only is allogeneic T cell proliferation reduced after contact with BiP-treated immature and mature DCs but also T cell contact with BiP-induced mDCs generates T cells that have immune regulating properties (manuscript in preparation). The decreased T cell proliferation should remind us that BiP also reduces T cell proliferation in vitro and that BiP poorly stimulates RA synovial T cell proliferation (see above). These findings suggest that BiP may be able to affect T cell development but whether this results in cells able to secrete

IMMUNOMODULATION BY BIP

215

Th2 cytokines, or whether it induces regulatory cells either of already known function and phenotype, or induces a regulatory cell with a novel phenotype and mode of action, is presently unknown.

How does BiP work? As usual, answers to such a question may be obtained more quickly and perhaps more easily by experiments in animals, although results obtained from animal models must be confirmed by human studies. In previous work we have shown that intravenous BiP could prevent CIA (see above). In further work we showed that either intravenous or subcutaneous BiP could treat ongoing CIA with intravenous BiP being about 20 times more potent than subcutaneous (Brownlie et al 2006). When lymph node and spleen cells, removed from animals with active CIA or CIA inhibited by either intravenous or subcutaneous BiP, were cultured with various concentrations of BiP, all three cell populations produced similar amounts of interferon IFNγ but only the cells from the animals given either intravenous or subcutaneous BiP secreted IL4, IL5 and IL10. This finding had a resonance with the cytokine profi le of human BiP-specific CD8 + T cell clones (see above). The suggestion that IL4 produc-tion may be a critical feature of the immunomodulating activity of BiP was confirmed by the finding that BiP could not inhibit CIA in IL4 knockout mice (Brownlie et al 2006). Furthermore, lymph node and spleen cells taken from DBA/1 mice given intravenous BiP, when transferred intravenously into mice with CIA, could suppress CIA without any BiP having to be administered to the recipient mice (Brownlie et al 2006). This exciting finding was the first observation that BiP may be acting via the induction of a regulatory cell(s). The nature of the BiP-induced regulatory cell is being intensively investigated. One obvious question is whether IL4 is needed for its induction or for its effector function or both.

Will BiP suppress rheumatoid synovitis? Short of a clinical trial of BiP in patients with active RA, the only way to answer this question is by the use of SCID/RA synovial membranes. This experimental system has been used to investigate other treatment modalities in RA such as an antibody directed against the soluble IL6 receptor (Matsuno et al 1998). In collaborative experiments with our Japanese colleagues we have shown that intravenous BiP will suppress inflammatory cell infi ltrate, expression of co-stimulatory molecules such as HLA-DR and CD86, and inflammatory cytokines such as TNFα and IL6 (Yoshida, Corrigall, Panayi and Matsuno; unpublished).

216

PANAYI & CORRIGALL

Summary BiP induces regulatory cells possibly via the alteration of DC development and function as well as by inducing the development of anti-inflammatory cells from monocytes. These are novel functions for this stress protein and suggest that BiP may be an effective therapy for chronic, immune-mediated diseases such as RA. References Bodman-Smith M, Corrigall VM, Kemeny DM, Panayi GS 2003 BiP, a putative autoantigen in rheumatoid arthritis, stimulates IL-10-producing CD8 positive T cells from normal individuals. Rheumatology (Oxford) 42:637–644 Brownlie RJ, Myers LK, Wooley PH et al 2006 Treatment of murine collagen-induced arthritis by the stress protein BiP via interleukin-4-producing regulatory T cells: a novel function for an ancient protein. Arthritis Rheum 54:854–863 Corrigall VM, Bodman-Smith MD, Fife MS et al 2001 The human endoplasmic reticulum molecular chaperone BiP is an autoantigen for rheumatoid arthritis and prevents the induction of experimental arthritis. J Immunol 166:1492–1498 Corrigall VM, Bodman-Smith MD, Panayi GS 2003 The stress protein, immunoglobulin binding protein (BiP), has a specific cell surface receptor. Rheumatology (Oxford) 42 Suppl:43 Corrigall VM, Bodman-Smith MD, Brunst M, Cornell H, Panayi GS 2004 Inhibition of antigen-presenting cell function and stimulation of human peripheral blood mononuclear cells to express an antiinflammatory cytokine profi le by the stress protein BiP: relevance to the treatment of inflammatory arthritis. Arthritis Rheum 50:1164–1171 Corrigall VM, Bodman-Smith M, Panayi GS 2005 BiP stimulation induces an anti-inflammatory gene activation profi le in monocytes. Rheumatology (Oxford) 44 Suppl:38 Gordon S 2003 Alternative activation of macrophages. Nat Rev Immunol 3:23–35 Matsuno H, Sawai T, Nezuka T et al 1998 Treatment of rheumatoid synovitis with anti-reshaping human interleukin-6 receptor monoclonal antibody. Arthritis Rheum 41:2014–2021 Panayi GS, Corrigall VM, Henderson B 2004 Stress cytokines: pivotal proteins in immune regulatory networks; Opinion. Curr Opin Immunol 16:531–534

DISCUSSION Lund: What are the levels of BiP circulating in the blood in normal human populations? Corrigall: We have screened quite a few normal subjects and rheumatoid arthritis patients. The maximum level we have found is 10 µg/ml, with the minimum undetectable. It appears that the RA patients have slightly lower levels than the normal subjects. Habich: Are there any data available regarding the role of BiP in the pathogenesis of type 1 diabetes? Panayi: No. I’m not aware that anyone has really worked on this. BiP is expressed in tumours and it has been suggested that this is because they are hypoxic. Interestingly, the rheumatoid joint is hypoxic as well.

IMMUNOMODULATION BY BIP

217

Nixon: The in vivo work is very exciting. Do you have any information on the stability of BiP once it is injected? Panayi: I have no idea. We’re hoping to get a company to do pharmacokinetics, because we’d like these sorts of data before we embark on human studies. This has to be done under GLP (good laboratory practice) conditions, which we are not licensed for. In the collagen-induced arthritis model as little as 1 µg of BiP given intravenously at the onset at arthritis suppresses it almost completely. This goes on for another 30 days when we have to kill the animal. Plus, if we take lymph node and spleen cells from animals given BiP intravenously, wash them and then inject them into animals with arthritis, it prevents or suppresses the arthritis, depending on when we give it. Provided you give enough BiP to induce the cellular effect, I don’t think the survival of BiP in the circulation will be critical. In other words the pharmacodynamic effect of BiP will not be related to its pharmacokinetic profi le. Hightower: It is wonderful to be discussing an extracellular chaperone for which we know the mechanism by which it leaves cells. At least in cultured rat cells, the ER to Golgi inhibitors almost completely block its release. Gaston: In the CIA model there is an IL4-driven effect of BiP, and in monocytes BiP’s effect is driven by IL10. Panayi: Because of the work showing that BiP stimulates IL10 release from human monocytes, we went for IL10. Obviously, because we know that the BiP T cell clones in humans and the BiP-responsive T cells in mice make this pattern of Th2 cytokines, we looked at IL4 and IL6 as well as IL10. It is interesting that in the circulation we could only measure IL6: we know that the rheumatoid synovium makes a lot of IL6. The histology of IL10 shows that of all the cytokines that we have looked at, IL10 is not abolished by giving these mice BiP. Gaston: Have you tried to block with anti-IL10 or anti-IL4? Panayi: One of the things that happens with these synovial pieces is that when BiP is applied, the weight of the pieces goes down, because all this cellular infi ltrate has been removed. But if anti-IL10 is applied, the weight stays the same as controls. IL10 may be an important part of this process. Gupta: I assume this effect of BiP is quite different from its chaperone activity in the ER, because it is not carrying out the same function. Panayi: I am not an expert on intracellular chaperones, but we know that the extracellular BiP probably binds to a receptor. Gupta: In your studies you are using intact, full-length BiP. Have you tried using fragments? Panayi: No. Gupta: This could be important. It would further illustrate that the activity you see is different from the chaperone activity. If you want to take it to the next clinical level, having a defined fragment will be helpful.

218

PANAYI & CORRIGALL

Panayi: I agree, but we haven’t looked at fragments yet. Gupta: Does the BiP that you find in the serum contain the retention peptide KDEL? Panayi: It doesn’t. Gupta: Then it is not full length BiP. So how much is deleted? Is the deletion specifically of KDEL? Panayi: Yes, as far as we can tell. Macario: BiP plays a key role in the ER stress response (Macario & Conway de Macario 2007). Perhaps what you are doing is instructing some cells to mount an ER stress response, if the protein is internalized somehow, or a signal is delivered. This could be what prolongs the effect. Panayi: I don’t have any data as to whether exogenous BiP is able to stimulate an ER BiP response. One of the ways we are looking at this is that we have a certain class of molecular chaperones, and when the cell is getting to a point where it is finding it difficult to cope, it says to itself let’s export this outside and get other cells to come and help us. Calderwood: It has been shown that GRP78 BiP can form a complex with CD91 in the membrane. Does this have anything to do with the signalling? Corrigall: We haven’t found BiP binding to CD91. We have tried blocking experiments with antibodies and also with α 2 macroglobulin, and they don’t have any effect on BiP binding to the cell. This is quite different to Hsp70 binding. Calderwood: We can’t find any effect of deleting CD91 on Hsp70 binding. Hightower: To extend Alberto’s point about stress to the ER, I don’t know enough about RA to know whether this means anything, but stress to the ER, while it involves the accumulation of abnormal proteins, has another feature that is distinct from nuclear/cytoplasmic stress. Stresses to the ER are reductive stresses because this is normally an oxidative environment. If this is specific to inducing ER chaperones and not nuclear/cytoplasmic, then you would look for conditions of reductive stress rather than oxidative stress. Panayi: The joint has low oxygen, reactive oxygen species, and is low glucose. BiP is a glucose-regulated protein. Immunohistology of the synovial membrane from patients with RA shows that there is marked up-regulation of the ER BiP. Hightower: This would make sense. Glucose deprivation was one of the first inducers of BiP discovered. Henderson: Can we go back to the beginning of the whole process? You found BiP because you were looking for the autoantigen in RA. This was a proteomic study using rheumatoid sera. BiP must then be an important immunogen in RA patients. Panayi: We don’t know what the affinities of these antibodies are. Throughout this meeting there has been a lot of discussion about how there exist antibodies to most of these stress proteins in humans, and whether or not they are functional.

IMMUNOMODULATION BY BIP

219

Paul Woolley looked at mice with CIA. They have antibodies to BiP, but nevertheless BiP could still prevent and clear up arthritis. There are situations with biological therapies where the presence of antibodies has not always led to loss of efficacy or toxicity. It will be a question of looking to see what happens. Quintana: What happens with the T cell responses to BiP in correlation with disease amelioration? Is there a cytokine shift and then they become more Th2? Panayi: The best evidence for that was the very first prevention study, in which the anti-collagen antibodies showed a shift from a Th1 type pattern to a Th2 pattern. In other studies, where lymph node and spleen cells have been removed and the cells stimulated with type 2 collagen, then they produce IL4 and IL10, plus IFNγ, whereas the lymph node and spleen cells from mice that weren’t given BiP but which were given control protein make IFNγ. In the mice it looks as if IFNγ is a constant, but they don’t make IL4 or IL10. For a variety of reasons, in these mice we have not yet measured the anti-collagen antibodies to see whether there is a Th1/Th2 shift as we showed in the first study. Quintana: What happens in humans? You have found T cell responses to BiP in humans. How does the T cell response to BiP in patients with arthritis change through the disease course? Panayi: We haven’t done the longitudinal study needed to answer that question. Cheetham: With regard to autoantibodies, can you reduce complications by local administration, or do you need a systemic effect to affect the immune system globally? Panayi: This is a multi-joint disease. It also has systemic manifestations. We know from studies with anti-TNF biologics that just injecting them into a joint isn’t a viable therapy. I think we have to go down the systemic administration route. Triantafilou: You say that you are getting CD8 + T cell clones. Why do you think that is? In the past we have found BiP associated on the cell membrane with MHC class I. Panayi: That is one thought: for some reason, extracellular BiP is taken up and is put in the MHC class I pathway. A concomitant explanation might be that the effects on things such as DC development might preferentially select for CD8 cells. Most of us are obsessed with CD4 regulatory cells, but we mustn’t forget that there are CD8 regulatory cells. I’m afraid I can’t give you any mechanistic explanation for the finding because we have not looked at it. Quintana: Are the BiP reactive T cells γδ or αβ ? Panayi: They are all αβ. We haven’t looked specifically at γδs. Henderson: BiP is inducing a Th2 response, so would it have any efficacy in diseases caused by parasites?

220

PANAYI & CORRIGALL

Panayi: I am not a parasitologist. It could do. I am not so sure that I subscribe to the view that Th1 cells do one thing and Th2 cells do another. It may be more complicated than this in real life. Miller-Graziano: The cells you are getting are essentially CD8 clones, and they are certainly not displaying typical Th1 or Th2 cytokine patterns as you would see in a mouse. Reference Macario AJL, Conway de Macario E 2007 Viral virulence and stress. In: Fink G (ed) The encyclopedia of stress second edition. Academic Press, Oxford, Vol 3, p 842–849

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Final discussion Lund: What many of the papers at this meeting have shown is that evolution is good at co-opting proteins to take on different functions. Evolution will grab whatever raw material is around, and work in a way that is not necessarily ‘intelligent’. Because I come from a chaperone background, when someone begins to talk about BiP or Hsp27, for example, I immediately think about them in terms of what they may be doing as chaperones. This may be a red herring in terms of understanding what they are doing, because they could be working as signalling molecules. If we are to make sense of this, we need more meetings such as this. Henderson: One of the key elements we need to grasp hold of is the relationship between the protein-folding capacity of molecular chaperones and their potential ‘moonlighting’ additional functions. What concentrations should we be looking at for folding and activity as opposed to lower levels where a chaperone will no longer be acting in this way? Lund: That’s a hard question to answer. To borrow from Popper, in our infi nite ignorance we are all equal! Henderson: A guideline like this would be quite useful. Cheetham: If you are thinking in terms of a folding activity, you should consider the clients, not the chaperone. The concentration of chaperone needed would depend on how much client you need to fold. Henderson: Some of these chaperones are fairly promiscuous, though. Cheetham: I think we are more likely to exclude a folding function by showing that critical residues are not important, certain domains are irrelevant, or that there are no co-chaperones present. As we have heard, many chaperones exist in the extracellular space and circulation: I’d be intrigued to know which chaperones are not present, especially in terms of what has been looked for and not found. Csermely: I can picture chaperones as proteins with a relatively open surface that does contain a lot of segments that are promiscuously binding to many things. They might continuously screen the environment for other proteins that don’t have their original conformational states or original partners. Calderwood: There are two important things we need to work out over the next few years. What are the mechanisms of secretion? And we need to sort out the problem of the receptors. There are a large number of postulated receptors, but in our hands only a few survive rigorous inspection. If we work out the receptors we can start to look at signalling. 221

222

FINAL DISCUSSION

Macario: Seeing the wide range of functions and locations of these molecules— and I am not sure that we can still call all of them ‘molecular chaperones’—one can predict that deficient chaperones would have a tremendous impact on health. There are many diseases that probably involve deficient chaperones in one way or another, namely the chaperonopathies (Macario & Conway de Macario 2005, 2007). This is a young field, as is the use of chaperones for prevention and treatment of chaperonopathies (i.e. chaperonotheraphy), and both are likely to become a big story in the next few years. Gupta: It is clear that chaperones have another important dimension, with multiple functions not related to protein-folding activities. In our studies on these proteins, it is necessary that we characterize these proteins a little more. When we discuss Hsp70, for example, we need to know which particular form of Hsp70 it is. This is important if we want to address mechanisms. We also need to understand that when these proteins are present in the membrane many of them don’t have any membrane-associating domain and they are not typical membrane proteins. How do they associate with the membrane and what is their tertiary-quaternary structure when they are there? These are important questions, which have not yet been looked at carefully. Gaston: This meeting has made me realize how much we need to concentrate on upstream events. We can gather a lot of data about what is happening downstream, but without knowing the receptors we are hassled by the rest of the world who don’t want to grapple with extracellular chaperones. I just wonder whether the receptor function is going to be different from what we are usually thinking about. We often think of heat shock proteins acting at quite high concentrations, certainly as compared with cytokines. If there was an obvious receptor for some of these things, surely it would have popped out of many studies earlier on. Perhaps something different is going on here. There may be ways of addressing this by mutating HSPs and looking at how this affects their extracellular function. Even if we don’t know what membrane events are occurring, there is more scope for looking at the immediate signalling pathways, such as Ca2+ flux. Multhoff: We were in the age of DNA and RNA; then we were in the age of proteins, and we are still in the age of proteomics. But for me the most exciting prospect is to combine proteomics and lipidomics! This might be a way forward for this field. Pockley: I think we have reached a watershed. We are no longer having to convince people that these proteins are released, but we still need to explain why they are being released and the consequences of this release. The challenge will be to develop an integrated approach, putting all this valid scientific information into a unified explanation. We focus on the immunology, but we can’t ignore the other aspects of extracellular stress response proteins, such

FINAL DISCUSSION

223

as cytoprotective effects. Conceptually, when people talk about extracellular molecular chaperones or stress proteins, the immediate thought is that we are looking at the presence of these proteins in plasma as a soluble protein. But they can be integrated into the plasma membrane. Why is that? Are they fixed there? Is there a dynamic relationship between extracellular and intracellular stress proteins? Panayi: The thought has struck me that we started off with the nervous system, with Charles Sherrington talking about its integrated action, and then along came the immune system. These two systems hung around more or less separately. Then people began to realize that the hypothalamic–pituitary–adrenal (HPA) axis can regulate inflammation, for example—the nervous system affecting the immune system. At the same time the immune system can impinge on the CNS through impulses via the vagus nerve. There is a ‘third world’ system, the system we have been discussing. We have a system that probably evolved just to protect cells from stress. Then this moved outside the cell with the realization that the outside environment needs some regulation. We have heard how this third world system of molecular chaperones is affecting the immune system. Can the chaperone system have an effect also on the CNS? To my knowledge, no one has given heat shock protein (HSP) to a mouse and then looked at the brain. This a great integrative system, and I can’t believe that it is not integrated with the other great integrative systems of the body. Csermely: I think in the field of extracellular chaperones we often know where we would like to go to, but sometimes we feel that we don’t know where we are. I think this meeting has helped us see more clearly where we are. I’d just like to add two of the take home messages that haven’t already been covered by others. • We know an increasing amount about how extracellular chaperones are leaving the cell. There is a lot to do in this area, but there is promise. • Without a cargo, HSPs activate natural killer (NK) cells, but with a cargo they have more diverse immune effects, activating many types of T cells. In general terms, many of these dual actions, such as pro-inflammatory and antiinflammatory, might be rationalized by saying that there are many receptors that are interacting with the same HSP in the circulation: one could be activating and the other inhibiting, or they could be acting antagonistically. It could also involve co-chaperones. This could explain some of the more controversial results. This meeting clearly showed us that we are at the beginning of an understanding of the field of extracellular chaperones, and the realization of its complexity. I would like to suggest to the participants and to their younger colleagues in their

224

FINAL DISCUSSION

labs to organise many more meetings like this with a lot of discussion time to meet the interested scientists who will certainly join the field in the near future. References Macario AJL, Conway de Macario E 2005 Sick chaperones, cellular stress and disease. New Engl J Med 353:1489–1501 Macario AJL, Conway de Macario E 2007 Chaperonopathies by defect, excess, or mistake. Ann N Y Acad Sci 1113:178–191

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Contributor Index

Non-participating co-authors are indicated by asterisk. Entries in bold indicate papers; other entries refer to discussion contributions. A

F

Asea, A. 54, 69, 70, 81, 113, 132, 139, 173, 180, 181, 182

Feder, M. E. 36, 43, 71, 82, 83, 96, 131, 132, 155, 158, 172, 193

B

G

*Bandyopadhyay, S. 196 *Bowes. T. 59

Gaston, J. S. H. 40, 41, 79, 111, 112, 135, 169, 181, 208, 217, 222 *Gong, J. 115 *Gray Jnr, P. J. 115 Gupta, R. S. 38, 39, 41, 57, 59, 69, 70, 71, 72, 84, 96, 134, 139, 155, 170, 171, 193, 194, 217, 218, 222

C Calderwood, S. K. 37, 43, 55, 82, 115, 132, 133, 134, 135, 139, 170, 182, 194, 195, 211, 218, 221 *Cehovin, A. 160 Cheetham, M. 17, 18, 41, 98, 99, 133, 134, 192, 195, 211, 219, 221 Coates, A. R. M. 41, 42, 70, 112, 133, 154, 155, 160, 169, 170, 171, 172 *Cohen, I. R. 101 Conway de Macario, E. 99, 138, 208, 209 Corrigall, V. M. 97, 138, 171, 210, 212, 216, 218 Csermely, P. 1, 16, 42, 45, 54, 55, 56, 57, 58, 69, 80, 97, 98, 155, 158, 171, 181, 193, 210, 211, 221, 223 D *De, A. 196

H *Haak, J. 3 Habich, C. 157, 182, 216 Henderson, B. 18, 21, 56, 79, 84, 96, 98, 99, 133, 137, 139, 141, 154, 155, 156, 157, 158, 181, 182, 193, 194, 218, 219, 221 *Herrmann, T. 196 Hightower, L. E. 17, 43, 55, 58, 82, 83, 96, 133, 137, 217, 218 *Hu, Y. 160 K *Korcsmáros, T. 45 *Kovács, I. A. 45 Kregel, K. C. 3, 20, 21, 193 L

E Ellis, R. J. 16, 17, 18, 19, 23, 36, 37, 38, 39, 40, 41, 42, 43, 72, 83, 137, 154, 155, 156, 169

*Laudanski, K. 196 Lund, P. A. 16, 20, 23, 39, 43, 57, 58, 83, 84, 96, 99, 135, 138, 154, 171, 194, 211, 216, 221 225

226

CONTRIBUTOR INDEX

M

Q

Macario, A. J. L. 15, 38, 42, 43, 54, 71, 80, 96, 97, 112, 132, 155, 169, 171, 180, 210, 218, 222 *Mambula, S. S. 115 Miller-Graziano, C. L. 80, 81, 83, 84, 98, 112, 134, 196, 208, 209, 210, 211, 220 Multhoff, G. 19, 42, 43, 57, 70, 86, 97, 132, 135, 171, 172, 192, 222

Quintana, F. J. 80, 101, 111, 112, 113, 114, 138, 156, 209, 211, 219

N

*Sawyer, D. 74 *Singh, B. 59 *So˝ti, C. 45 *Szalay, M. S. 45

*Nakamura, H. 184, 192, 193, 194, 195 Nixon, B. 18, 19, 39, 40, 55, 70, 98, 133, 138, 157, 158, 180, 182, 217 *Nor, A. 74

R *Ramachandra, N. B. 59 S

T P Panayi, G. S. 69, 79, 82, 84, 99, 135, 137, 138, 139, 155, 156, 157, 158, 169, 172, 179, 180, 194, 212, 216, 217, 218, 219, 220, 223 Pockley, A. G. 20, 43, 81, 86, 97, 98, 99, 113, 114, 133, 134, 135, 137, 139, 157, 210, 222

*Theriault, J. R. 115 Triantafi lou, K. 42, 74, 79, 80, 84, 171, 180, 182, 192, 219 *Triantafi lou, M. 74 V *Vakakis, E. 74

The Biology of Extracellular Molecular Chaperones: Novartis Foundation Symposium 291, Volume 291. Edited by Derek J. Chadwick and Jamie Goode Copyright  Novartis Foundation 2008. ISBN: 978-0-470-72333-3

Subject Index A acr2 147 actin 17, 41 activator protein 1 (AP-1) 185 acute respiratory distress syndrome 187 ADAM protease 157 ADAM17 187 adjuvant arthritis (AA) 102, 103, 164 adult T cell leukaemia (ATL) 184 ageing 13 heat shock response 20–21 Hsp60 21, 98 age-related macular degeneration (AMD) 188, 192 aggregation 41–42 AIDS 161 alcohol acetaldehyde dehydrogenase 145 αβ domain proteins 16 αβ class 39 αβ -crystallin 50 Alzheimer’s disease 12, 138 amylase, Hsp60 and 63 anti-ergotypic cells 105 antigen cross-presentation 123 antigen processing cells, Hsp70 family in 116–118 Antlion 151 apoptosis signal regulating kinase 1 (ASK1) 185 Arabidopsis thaliana 48 arthritis 103 juvenile chronic 103 see also rheumatoid arthritis asthma 164 atherosclerosis 21, 90, 102, 134, 137, 176 autoimmune diabetes 102, 104–105, 157, 164 B B7 119 B7.1 121 B7.2 121

bacterial heat shock protein 147–151 as host cell modulating agonists 148–149 stress of infection 147–148 symbiosis 151 as therapeutic targets 151 as virulence factors 149–151 BAG 38 Bechet’s disease 102 BiP 10, 83, 89, 138, 146, 212–216, 216–222 anti-inflammatory effects 92 dendritic cells, effects on 214–215 discovery 213 human studies 213–215 monocytes, effects on 213–214 suppression of rheumatoid synovitis 215 T cells, effects on 213 BiP receptor 158 brefeldin A 88, 89 C Caco-2 88 Caenorhabditis elegans 20, 57 calnexin 27 calreticulin 27, 120, 124 Campylobacter jejuni 145 CCR5 146 CCT complex 31, 48 CD1a 199 CD14 75, 120, 146, 211 Hsp60 and 182 CD28 83, 84 CD40 119, 120, 121, 146 CD47 211 CD62L 187 CD83 204 CD86 112, 199, 206 CD91 120, 123, 146, 218 CD94 132 CD163 199, 201, 206 cell stress protein see heat shock protein cellular stress response 3–4, 15–21 classical view 5–8 227

228 historical overview 4–5 immune responses 11 inhibition of apoptotic pathways 9–10 modern view 8–11 paradoxical inflammatory actions 10– 11 stress protein dysfunction, potential ramifications 11–13 c-ERB-B2/neu 132 chaotrope 34 chaperokines 10 see also Hsp72, exogenous chaperome 38 chaperomics 38 chaperone networks 48–50 chaperone overload 47 chaperone-rich lysates 124 chaperones categories 27–32 defi nition 26 historical overview 4–5 myths 32–4 origin of concept 25 chaperonins 31, 34 chaperonopathies 12–13, 71–72, 222 acquired 12 chemical chaperone 34 CHIP 38, 43 Chlamydia 145 Chlamydia pneumoniae 162 choroidal neovascularization (CNV) 188 chronic obstructive pulmonary disease (COPD) 194 Chs7p 27 cis ring 32 c-Jun N-terminal kinase ( JNK) 10 CLEC1 120 CLEC2 120 CLIPS chaperones 48 ClpB 39 ClpP 151 co-chaperones 26, 38 protein folding 30–32 colchicine 87 compensatory anti-inflammatory response syndrome (CARS) 83 competitive inhibition of chaperones 47 complement factor H (CFH) 188 Con A 105 CP20961 103 Cpn10 see Hsp10

SUBJECT INDEX Cpn60 see Hsp60 CTLA4 112 cyclophosphamide 107 cyclophosphamide-accelerated diabetes (CAD) 104 cytochrome c 66 cytochrome c/caspase cascade 10 D DC-SIGN 120 Dectin1 120 Dengue virus 145 15-deoxyspergualin 116 detergent-resistant microdomains (DRMs) 88 diabetes 103, 114 autoimmune 102, 104–105, 107, 164 cyclophosphamide-accelerated (CAD) 104 Hsp60 107, 157 dimethylsulfoxide 34 DnaJ see Hsp40 DnaK see Hsp70 Drosophila 20 Hsp60 96 Hsp70 17, 43 Drosophila melanogaster 4 E early pregnancy factor (EPF) 64, 143 EF-Tu 16 endoplasmin see Gp96 endotoxaemic shock 82 Enterobacter aerogenes 151 Escherichia coli 48, 145, 146, 150, 162, 193 chaperone-networks 49 DnaJ 38 DnaK system 39, 49, 55 GroEL 16, 38–39 Hsp70 39, 49, 55 exosomes 42–43 experimental autoimmune encephalitis (EAE) 107 F FEEL1 120, 123 ferritin 187 fibroblast growth factors 1 and 2 87

SUBJECT INDEX fluorescent resonance energy transfer (FRET) 76, 79 fumarase 66 G β -galactosidase

124 gamma-interferon 148 giant squid axon 17, 18, 87, 89 glutamic acid decarboxylase (GAD) 102 glycerol 34 glycodelin A 157 Gp96 10, 51, 81, 90, 98, 120, 123, 124, 145 anti-inflammatory effects 92–93 antigen presentation 11 immunological properties 133 gene families 115 groups 117 granuloma activation–suppression (GAS) cycle 164–166, 167 granzyme B 91 GroE system 16 GroEK 55 GroEL proteins 31, 32, 71, 89, 151, 156 in E. coli 16, 38, 39 folding 40 in mycoplasmas 49 system 16 in Mycobacterium smegmatis 96 GroEL/S team 54 GroES proteins 31–32 growth hormone granules (GHGs) 63, 64 Grp78 see BiP Grp94 (endoplasmin) see Gp96 Grp170 117, 124 gene families 115 GrpE 15, 54 H heat shock elements (HSE) 7 heat shock protein (Hsp) see entries under Hsp heat shock transcription factor 7 Helicobacter pylori 151, 162 hepatitis C 187 histone 2B 61, 72 HIV transmembrane glycoprotein gp41 61 Hop 51 HP0175 151 Hsc70 17, 43, 62, 84, 86–89, 118 interaction with phosphatidylserine 88

229 Hsp bacterial 147–151 in biological fluids 138 circulating host 146–147 innate functions 106 inflammation biomarkers 106–107 interaction with immune cells 118–120 location and function 5, 7–8 modulators of bacteria–host interactions 141–152, 154–158 receptors for pathogens 145 targets of adaptive immunity 102–106 target of pathogenic T and B cell responses 102–103 target of regulatory responses 103–105 Hsp families 5 Hsp receptors 121 Hsp10 (Cpn10) 32, 89, 155 assay 146 binding to MARCO 134 innate immunity 77, 79, 80 LPS responses 77, 79, 80 mitochondrial 64, 69–72 as moonlighting CSP 143 rheumatoid arthritis 147 Hsp20 gene 147 Hsp25 211 Hsp27 55, 89, 134, 196–206, 221 alterations by TLR3 80 anti-inflammatory effects 92 IL10 10–11, 197–206, 208–211 inhibition of TLR4 receptor responses 80 MO-Mac differentiation 199 Tregs 81 Hsp40 (DnaJ) 15, 26, 27, 37, 38, 39, 40, 96–97 association with Hsp70 97 families 30 protein folding 30 team 54 Hsp60 10, 41, 32, 51, 101–108, 111–114, 120 activation by TLR2 81, 106 activation of TLR4 106, 182 ageing 21, 98 assays 97, 146 atherosclerosis 137 autoimmune diabetes 104–105 B and T cell responses 102 diabetes 107, 157 Drosophila 96 ELISA-based studies 98

230 endosymbiont (symbionin) 151 extracellular 145 granuloma activation–suppression (GAS) cycle 164–166, 167 groups 27 immunostimulatory effects 163–164 immunosuppressive actions 164 innate functions 106 macrophage activation 160–167, 169–172 mitochondrial 60–64, 69–72 mitochondrial targeting sequence (MTS) 60 mouse sperm 18, 19 multiple 96 potential receptors 162 pro-inflammatory molecules 90, 91 psychological stress 98 reproductive tract 138 rheumatoid arthritis 147 target of anti-ergotypic regulatory T cells 105–106 Hsp60.1, Mycobacterium tuberculosis (MtCpn60.1) 161, 164, 166–167, 171, 172 Hsp60.2, Mycobacterium tuberculosis (MtCpn60.2) 160, 161, 162, 163, 164, 166, 171, 172 Hsp65 102, 103, 104, 105, 114 Hsp70 (DnaK) 5, 8, 9–10, 37, 86, 120, 124, 125, 222 ageing 98 antigen presentation 11 antigen processing and cross presentation 116–118 assays 97 association with Hsp40 97 atherosclerosis 21 binding of dendritic cells 11 binding sites 82–83 binding to CCR5 123 binding to CD40 84 binding to CD94 132 binding to NKG2D 133 binding to SLC receptor 134 candidate receptors 120 co-chaperone functions 26 diabetes 107 domains of life 15–16 Drosophila 17, 43 E. coli 39, 49, 55 effect of endogenous Hsp70 99

SUBJECT INDEX ELISA-based studies 98 extracellular, role in antitumour immunity 115–116 giant squid axon 17, 18 glial axon transfer protein 17, 18 granzyme B uptake 91 heat activation 99 human vs Mycobacterium tuberculosis 84 inhibition of LPS-induced inflammatory responses 75–76, 79 innate immunity 75–77, 79, 80 interaction with phosphatidylserine 88 interaction with TLR4 75, 76–77, 106 lipid rafts 55 mitochondrial (mHsp70; mortalin) 65, 66, 69–72 multiple 96 normal levels 99 pro-inflammatory molecules 90, 91, 92 protection of heat-stressed cells from cytotoxic damage 89 protein folding 30 protein–lipid interactions 57 psychological stress 98 release from Caco-2 88 role 31, 32 specificity 97 sperm surface 19 SSA 48 SSB 48 TLR4 and TLR2 activation 106 tumours 11, 87–88, 89, 90, 92, 126, 131–132 Hsp70 family 5, 8, 30, 31, 115 Hsp70 group 27 Hsp70 receptors 120 Hsp70(DnaK) team 54 Hsp70/Bag-4 88 Hsp70/Hsp40 88 Hsp70B′ 96 Hsp70-based vaccine design 124–126 Hsp70–PC internalization by APCs 123–124 Hsp72, exogenous (eHsp72) 173–177, 179–182 biological significance 176–177 chaperokine activity 173–176 mechanisms of release 176 Hsp78 49–50 Hsp90 10, 11, 26, 117, 120, 124, 125 cancer cells 11, 19

SUBJECT INDEX co-chaperone functions 26 Drosophila melanogaster 47 immunostimulation 80 lipid rafts 55 role in antigen presentation 11 yeast 38 Hsp90 chaperone complex 51 Hsp90 gene families 115 Hsp90 group 27 Hsp90 α 145 Hsp90b1 mouse sperm 18, 19 reproductive tract 138 Hsp100 27, 31, 87 Hsp104 16–17, 37, 48 Hsp110 86, 120, 124 gene families 115 Hsp-induced APC maturation and pro-inflammatory signalling 121– 123 HSP-OC uptake by APCs 121 HTR A1 192 Hu3 103, 105 Huntington’s disease 12

231 interferon gamma (IFNgamma) 88, 103, 105, 106 interleukin see under IL J JlpA 145 juvenile chronic arthritis 103 K KDEL 218 kosmotrope 34 L LAMP1 123 LAMP2 123 LAMP2a 118 LAP 209 Legionella pneumophila 150 lipid rafts 55, 57, 79, 88, 189, 192 lipidome 57 Listeria monocytogenes 145 LOX1 120, 123, 132, 146, 182 LOX1 receptor 133 L-selectin (CD62L) 187 M

I IgG1 104 IgG2b 104 IL1 10, 87, 146, 194–195 Hsp27 10–11, 197–206, 208–211 IL1β 120, 122 IL2 91 IL2 receptor 105 IL4 98 IL6 10, 120, 122, 126 IL10 10, 11, 90, 103, 106, 171–172 IL12 10, 120, 210 IL15 10 ILT receptors 134 ILT3 202 ILT4 202 indolamine deaminase 210 indoleamine-2,3-dioxygenase (IDO) 204, 214 inflammatory bowel disease 188 insulin, Hsp60 and 63 insulin secretory granules (ISGs) 63 interferon 138, 187

α2

macroglobulin 218 MARCO 134 MD2 75 melanoma, malignant 126 2-mercaptoethanol 186 methyl- β -cyclodextrin 88, 189 migration inhibitory factor (MIF) 185 Mip protein 150 misassembly 41 mitochondrial aspartate aminotransferase 65 molecular chaperone function, defi nition 26 molecular chaperones see entries under chaperone monensin 87, 88 moonlighting 143–147 mortalin (mitochondrial Hsp70; mHsp70) 65, 66, 69–72 MUC1 125 multiple sclerosis (MS) 107, 137–138 Mycobacterium smegmatis 96, 170 Mycobacterium tuberculosis 102, 133, 146, 147 life cycle 160–161 Hsp70 84

232 MtCpn60.1 161, 164, 166–167, 171, 172 MtCpn60.2 160, 161, 162, 163, 164, 166, 171, 172 Mycoplasma, chaperone-networks in 49 N nascent chain-associated complex (NAC) 27, 30, 31 Neisseria gonorrheae 147 NF-κ B 121, 122, 123, 185 NKG2A 120, 123, 132 NKG2C 132 NKG2D 120, 123 nucleoplasmin 25 O OmpA 120, 145 ovalbumin 124 OX6 105 OX40L 121 P P13K 106 ras p21 61, 72 p23 51 p38 106 p53 185 p277 102, 104, 106 Pam3Cys-SK4 91 pancreatitis, acute 188 Parkinson’s disease 12 pathogen-associated molecular patterns (PAMPs) 75 peptide binding protein (PBP74) 65 peptidyl prolyl isomerases (PPIs) 150, 151 phosphatidylserine (PS) 88 phosphoglucoisomerase 143 phospholipase C inhibition 89 pHSP60 103, 104, 105 pHSP60 vaccination 104 pHsp65 103, 104 pHsp65 vaccination 104 placental protein 14 157 podophyllotoxin 60 polyribosomes 24 Porphyromonas gingivalis 147, 162 prefolin 16 prion protein (Prpc ) 61 pristine 103

SUBJECT INDEX program death receptor ligand 1 (PDL1) 201 protein aggregation 24–25, 37, 38 protein degradation 38 protein disulfide isomerase 26, 185 protein folding 27–32, 36–37, 115 protein kinase A 61 protein refolding 25 protein self-assembly 23–24 assisted 24 spontaneous 24 R Rab4 70 Rab5a 123 RbcX 40 reactive oxygen species (ROS) generation 20 rheumatoid arthritis 103, 155, 156, 157 BiP 212–216, 216–222 Hsp10 147 rheumatoid synovitis 215 Rhizobium leguminosarum 162 RNA chaperones 42 rpoH 147 Rubisco 40 S self-Hsp60 103 Shr3p 27 sHSP groups 27 sickle cell disease 176 SOCS3 164 SPARC 51–52 SREC1 120, 123, 132 SSE1 see Hsp104 Staphylococcus aureus 145 Streptococcus hawaiiensis 151 Streptococcus mutans 150 symbionin 151 system biology 45–62, 54–58 cellular network approach 45–46 chaperone networks 48–50, 52 extracellular chaperones and networks 51–52 special roles of chaperones 46–48 systemic inflammatory responses syndrome (SIRS) 82, 83 systemic lupus erythematosus 102

SUBJECT INDEX T TAT system 72 taxol 70 TGFβ 90, 209 TGFβ1 103, 105, 106 thermosome 31 thermotolerance 8, 11 thioredoxin (TRX) 147 anti-inflammatory mechanisms 189 extracellular functions 184–190, 192– 195 thioredoxin 1 184, 185, 183 thioredoxin 2 185, 193 thrombospondin 1 (TSP1) 204, 205, 206, 209 TLR2 90, 106, 113, 114, 120, 182 activation of Hsp60 81, 106 TLR2/4 146 TLR3 alteration of Hsp27 80 modulation of HSP receptors 80–81 TLR4 12, 75, 120 Hsp60 106, 182 inhibition by Hsp27 80 interaction with Hsp70 75 Tregs 81 TLRs 75, 90, 132 TNFR:Fc fusion protein 82 Toll-like receptors see TLRs

233 transferrin 123 transporters associated with antigen processing (TAP) 117 TAP1 117 TAP2 117 Tregs (T-regulatory cells) 98–99, 106, 112, 113, 135, 180 Hsp27 81 TLR4 81 TRiC complex 31, 48 trigger factor (TF) 30, 31, 32, 150 trimentylamine N-oxide 34 Trx80 186, 194 tuberculosis 133, 160, 169, 170 vaccine 133 tubulin 41, 61, 62, 69, 70, 71, 72, 82 tumour necrosis factor α (TNFα ) 10, 11, 120, 122, 204 V vincristine 70 ViP 43 Y Yersinia enterocolitica 164 Z zymogen granules (ZGs) 63, 64

The Biology of Extracellular Molecular Chaperones (Novartis Foundation Symposium 291)

The Limits of Reductionism in Biology (Novartis Foundation Symposium 213)

Mammalian TRP Channels as Molecular Targets (Novartis Foundation Symposium 258)

Mitochondrial Biology: New Perspectives (Novartis Foundation Symposium 287)

Vascular Development (Novartis Foundation Symposium 283)

Cancer and Inflammation (Novartis Foundation Symposium 256)

The Genetics of Autoimmunity (Novartis Foundation Symposium 267)

Decoding the Genomic Control of Immune Reactions: Novartis Foundation Symposium

Anaphylaxis (Novartis Foundation Symposia)

Evolutionary Developmental Biology of the Cerebral Cortex (Novartis Foundation Symposia)

Evolutionary Developmental Biology of the Cerebral Cortex (Novartis Foundation Symposia)

Evolutionary Developmental Biology of the Cerebral Cortex (Novartis Foundation Symposia)

Neuronal and Cognitive Effects of Oestrogens (Novartis Foundation Symposium 230)

Neuronal and Cognitive Effects of Oestrogens (Novartis Foundation Symposium 230)

Tissue Engineering of Cartilage and Bone (Novartis Foundation Symposium 249)

Endocrine Facets of Ageing (Novartis Foundation Symposium 242)

Molecular Chaperones: Methods and Protocols (Methods in Molecular Biology, v787)

Molecular Chaperones: Methods and Protocols (Methods in Molecular Biology 787)

Molecular Chaperones and Cell Signalling

Molecular Chaperones and Cell Signalling

Nuclear Organization in Development and Disease (Novartis Foundation Symposium 264)

Mucus Hypersecretion in Respiratory Disease (Novartis Foundation Symposium 248)

Autism : Neural Basis and Treatment Possibilities (Novartis Foundation Symposium 251)

Growth Factors and Psychiatric Disorders (Novartis Foundation Symposium 289)

Molecular Mechanisms Influencing Aggressive Behaviours (Novartis Foundation Symposia)

Extracellular Nucleic Acids (Nucleic Acids and Molecular Biology)

New Foundation of Biology

New Foundation of Biology

Foundation of Mathematical Biology

Matrix Metalloproteinases (Biology of Extracellular Matrix)

Empathy and Fairness (Novartis Foundation Symposia)

The Biology of Extracellular Molecular Chaperones (Novartis Foundation Symposium 291)

The Limits of Reductionism in Biology (Novartis Foundation Symposium 213)

Mammalian TRP Channels as Molecular Targets (Novartis Foundation Symposium 258)

Mitochondrial Biology: New Perspectives (Novartis Foundation Symposium 287)

Vascular Development (Novartis Foundation Symposium 283)

Cancer and Inflammation (Novartis Foundation Symposium 256)

The Genetics of Autoimmunity (Novartis Foundation Symposium 267)

Decoding the Genomic Control of Immune Reactions: Novartis Foundation Symposium

Anaphylaxis (Novartis Foundation Symposia)

Evolutionary Developmental Biology of the Cerebral Cortex (Novartis Foundation Symposia)

Evolutionary Developmental Biology of the Cerebral Cortex (Novartis Foundation Symposia)

Evolutionary Developmental Biology of the Cerebral Cortex (Novartis Foundation Symposia)

Neuronal and Cognitive Effects of Oestrogens (Novartis Foundation Symposium 230)

Neuronal and Cognitive Effects of Oestrogens (Novartis Foundation Symposium 230)

Tissue Engineering of Cartilage and Bone (Novartis Foundation Symposium 249)

Endocrine Facets of Ageing (Novartis Foundation Symposium 242)

Molecular Chaperones: Methods and Protocols (Methods in Molecular Biology, v787)

Molecular Chaperones: Methods and Protocols (Methods in Molecular Biology 787)

Molecular Chaperones and Cell Signalling

Molecular Chaperones and Cell Signalling

Nuclear Organization in Development and Disease (Novartis Foundation Symposium 264)

Mucus Hypersecretion in Respiratory Disease (Novartis Foundation Symposium 248)

Autism : Neural Basis and Treatment Possibilities (Novartis Foundation Symposium 251)

Growth Factors and Psychiatric Disorders (Novartis Foundation Symposium 289)

Molecular Mechanisms Influencing Aggressive Behaviours (Novartis Foundation Symposia)

Extracellular Nucleic Acids (Nucleic Acids and Molecular Biology)

New Foundation of Biology

New Foundation of Biology

Foundation of Mathematical Biology

Matrix Metalloproteinases (Biology of Extracellular Matrix)

Empathy and Fairness (Novartis Foundation Symposia)

Recommend Documents