Advances in Immunology Volume 81

ADVANCES IN Immunology VOLUME 81 This Page Intentionally Left Blank ADVANCES IN Immunology EDITED BY FREDERICK W...

Author: Frederick W. Alt

48 downloads 1042 Views 6MB Size Report

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

DOWNLOAD PDF

ADVANCES IN

Immunology VOLUME 81

This Page Intentionally Left Blank

ADVANCES IN

Immunology EDITED BY

FREDERICK W. ALT Howard Hughes Medical Institute Children’s Hospital, Boston, Massachusetts ASSOCIATE EDITORS K. Frank Austen Tadamitsu Kishimoto Fritz Melchers Jonathan W. Uhr Emil Raphael Unanue

VOLUME 81

Amsterdam Boston Heidelberg London New York Oxford Paris San Diego San Francisco Singapore Sydney Tokyo

This book is printed on acid-free paper. Copyright ß 2003, Elsevier Inc. (USA). All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the Publisher. The appearance of the code at the bottom of the first page of a chapter in this book indicates the Publisher’s consent that copies of the chapter may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay the stated per copy fee through the Copyright Clearance Center, Inc. (222 Rosewood Drive, Danvers, Massachusetts 01923), for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying, such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. Copy fees for pre-2003 chapters are as shown on the title pages. If no fee code appears on the title page, the copy fee is the same as for current chapters. 0065-2776/2003 $35.00 Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: ( þ 44) 1865 843830, fax: ( þ 44) 1865 853333, e-mail: [email protected]. You may also complete your request on-line via the Elsevier homepage (http://elsevier.com), by selecting ‘‘Customer Support’’ and then ‘‘Obtaining Permissions.’’

Academic Press An Elsevier Imprint. 525 B Street, Suite 1900, San Diego, California 92101-4495, USA http://www.academicpress.com

Academic Press 84 Theobald’s Road, London WC1X 8RR, UK http://www.academicpress.com

International Standard Book Number: 0-12-022481-X

PRINTED IN THE UNITED STATES OF AMERICA 03 04 05 06 07 08 9 8 7 6 5

4

3

2

1

CONTENTS

CONTRIBUTORS

ix

FOREWORD

xiii

Regulation of the Immune Response by the Interaction of Chemokines and Proteases

SOFIE STRUYF, PAUL PROOST, I. II. III. IV.

AND JO

VAN DAMME

Introduction to Chemokines Posttranslational Modification of CXC Chemokines CC Chemokines as Protease Substrates The CD26-Chemokine Connection References

1 13 20 23 32

Molecular Mechanisms of Host–Pathogen Interaction: Entry and Survival of Mycobacteria in Macrophages

JOHN GATFIELD I. II. III. IV. V. VI.

AND JEAN

PIETERS

Introduction Innate Recognition of Pathogens Phagocytosis and Intracellular Trafficking in Macrophages Intracellular Survival of Bacterial Pathogens Mycobacterial Infection Conclusions References

45 51 57 65 70 84 85

B Lymphoid Neoplasms of Mice: Characteristics of Naturally Occurring and Engineered Diseases and Relationships to Human Disorders

HERBERT C. MORSE III, TOM MCCARTY, CHEN-FENG QI, TED A. TORREY, ZOHREH NAGHASHFAR, SISIR K. CHATTOPADHYAY, TORGNY N. FREDRICKSON, AND JANET W. HARTLEY I. Introduction II. Classification of Mouse B Cell-lineage Lymphomas III. Pathogenesis of Spontaneous B Cell-lineage Neoplasms v

97 102 109

vi

CONTENTS

IV. Conclusions References

116 116

Prions and the Immune System: A Journey Through Gut, Spleen, and Nerves

ADRIANO AGUZZI I. II. III. IV. V. VI. VII. VIII. IX. X. XI.

Prion Biology: Some Basic Facts Peripheral Entry Sites of Prions: Complicity of Immune Cells Lymphocytes and Prion Pathogenesis Prion Hideouts in Lymphoid Organs A Spleen is not a Lymph Node: Idiosyncrasies in the Lymphotropism of Prions Sympathetic Nerves: A Neuroimmune Link? Innate Immunity and Antiprion Defense Adaptive Immunity and Pre-Exposure Prophylaxis Against Prions The Prion Doppelganger Prion Immunology: Quo Vadis? An Essential Glossary of Prion Jargon References

123 128 134 137 141 143 148 150 153 158 159 160

Roles of the Semaphorin Family in Immune Regulation

ATSUSHI KUMANOGOH I. II. III. IV. V. VI. VII.

AND

HITOSHI KIKUTANI

Overview Class IV Semaphorin (1): CD100/Sema4D Class IV Semaphorin (2): Sema4A Virus-Encoded Semaphorins and their Cellular Counterparts Class III Semaphorin: Sema3A in Immune Cell Migration Neuropilin-1 in initial T cell/DC Contacts Perspectives References

173 177 187 191 192 193 194 195

HLA-G Molecules: from Maternal–Fetal Tolerance to Tissue Acceptance

EDGARDO D. CAROSELLA, PHILIPPE MOREAU, JOE¨L LE MAOULT, MAGALI LE DISCORDE, JEAN DAUSSET, AND NATHALIE ROUAS-FREISS I. Introduction II. The HLA-G Gene and Polymorphism III. Regulation of HLA-G Gene Expression

199 201 206

CONTENTS

IV. V. VI. VII. VIII.

Processing and Transport of HLA-G Molecules Structural and Functional Properties of HLA-G Molecules Role of HLA-G in Normal and Pathological Pregnancies HLA-G in Organ Transplantation HLA-G in Malignancies References

vii 213 217 225 232 236 243

The Zebrafish as a Model Organism to Study Development of the Immune System

DAVID TRAVER, PHILIPPE HERBOMEL, E. ELIZABETH PATTON, RYAN D. MURPHEY, JEFFREY A. YODER, GARY W. LITMAN, ANDRE´ CATIC, CHRIS T. AMEMIYA, LEONARD I. ZON, AND NIKOLAUS S. TREDE I. II. III. IV. V. VI. VII. VIII. IX. X.

Introduction Innate Immunity of Teleosts Ontogeny of Adaptive Immunity from Fishes to Mammals Phenotypic Characterization of Zebrafish Hematolymphoid Cells The Zebrafish as a Vertebrate Model System for Forward Genetic Screens Zebrafish Screens for Lymphoid Mutants Reverse Genetic Approaches Gene Expression Screens Use of Genomics: Getting Started with the Zebrafish Concluding Remarks: Impact of Zebrafish on Immunology References

253 257 272 290 296 306 307 311 313 315 316

Control of Autoimmunity by Naturally Arising Regulatory CD4 þ T Cells

SHOHEI HORI, TAKESHI TAKAHASHI,

AND

SHIMON SAKAGUCHI

I. Introduction II. Key Roles of Naturally Arising CD4 þ Treg Cells in the Maintenance of Natural Self-Tolerance: Induction of Autoimmune Disease in Normal Animals by their Depletion III. The Phenotype of Naturally Arising CD4 þ Treg Cells IV. Functional Characteristics of CD25 þ CD4 þ Treg Ex Vivo and In Vitro V. Molecular Basis of Treg Functions VI. The Development, Specificity, and Homeostasis of CD4 þ Treg Cells VII. Concluding Remarks References

331

333 338 342 345 352 364 365

INDEX

373

CONTENTS OF RECENT VOLUMES

385


CONTRIBUTORS

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

Adriano Aguzzi (123), Institute of Neuropathology, Universita¨tsspital Zu¨rich, Schmelzbergstrasse 12, CH-8091 Zu¨rich, Switzerland Chris T. Amemiya (253), Benaroya Research Institute, Molecular Genetics Dept., 1201 Ninth Avenue, Seattle, WA 98101 Edgardo D. Carosella (199), Service de Recherches en He´matoImmunologie, Direction des Sciences du Vivant, De´partement de Recherche Me´dicale, Commissariat a` l’Energie Atomique, Institut Universitaire d’He´matologie, Hoˆpital Saint-Louis, 75010 Paris, France Andre´ Catic (253), Division of Hematology, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115 Sisir K. Chattopadhyay (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland Jean Dausset (199), Fondation Jean Dausset, CEPH, Paris, France Torgny N. Fredrickson (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland John Gatfield (45), Biozentrum der Universitaet Basel, Department of Biochemistry, Klingelbergstrasse 50-70, 4056 Basel, Switzerland Janet W. Hartley (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland Philippe Herbomel (253), Unite´ Macrophages et De´veloppement de l’Immunite´, Institut Pasteur, 25 rue du Dr Roux, 75724 Paris cedex 15, France Shohei Hori (329), Laboratory of Immunopathology, Research Center for Allergy and Immunology, The Institute for Physical and Chemical Research (RIKEN), Yokohama 230-0045, Japan Hitoshi Kikutani (173), Department of Molecular Immunology, Research Institute for Microbial Diseases, Osaka University, 3-1 Yamada-oka, Suita, Osaka 565-0871, Japan

ix

x

CONTRIBUTORS

Atsushi Kumanogoh (173), Department of Molecular Immunology, Research Institute for Microbial Diseases, Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-0871, Japan Magali Le Discorde (199), Service de Recherches en He´mato-Immunologie, Direction des Sciences du Vivant, De´partement de Recherche Me´dicale, Commissariat a` l’Energie Atomique, Institut Universitaire d’He´matologie, Hoˆpital Saint-Louis, 75010 Paris, France Joe¨l Le Maoult (199), Service de Recherches en He´mato-Immunologie, Direction des Sciences du Vivant, De´partement de Recherche Me´dicale, Commissariat a` l’Energie Atomique, Institut Universitaire d’He´matologie, Hoˆpital Saint-Louis, 75010 Paris, France Gary W. Litman (253), Immunology Program, H. Lee Moffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Tampa, FL 33612 and Department of Pediatrics, Children’s Research Institute, University of South Florida, 140 Seventh Avenue South, St. Petersburg, FL 33701 Tom McCarty (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland Philippe Moreau (199), Service de Recherches en He´mato-Immunologie, Direction des Sciences du Vivant, De´partement de Recherche Me´dicale, Commissariat a` l’Energie Atomique, Institut Universitaire d’He´matologie, Hoˆpital Saint-Louis, 75010 Paris, France Herbert C. Morse (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland Ryan D. Murphey (253), Division of Hematology, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115 Zohreh Naghashfar (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland E. Elizabeth Patton (253), Division of Hematology, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115 Jean Pieters (45), Biozentrum der Universitaet Basel, Department of Biochemistry, Klingelbergstrasse 50-70, 4056 Basel, Switzerland Paul Proost (1), Laboratory of Molecular Immunology, Rega Institute for Medical Research, Minderbroedersstraat 10, University of Leuven, B-3000 Leuven, Belgium Chen-Feng Qi (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland Nathalie Rouas-Freiss (199), Service de Recherches en He´matoImmunologie, Direction des Sciences du Vivant, De´partement de

CONTRIBUTORS

xi

Recherche Me´dicale, Commissariat a` l’Energie Atomique, Institut Universitaire d’He´matologie, Hoˆpital Saint-Louis, 75010 Paris, France Shimon Sakaguchi (329), Laboratory of Immunopathology, Research Center for Allergy and Immunology, The Institute for Physical and Chemical Research (RIKEN), Yokohama 230-0045, Japan and Department of Experimental Pathology, Institute for Frontier Medical Sciences, Kyoto University, Kyoto 606-8507, Japan Sofie Struyf (1), Laboratory of Molecular Immunology, Rega Institute for Medical Research, Minderbroedersstraat 10, University of Leuven, B-3000 Leuven, Belgium Takeshi Takahashi (329), Department of Experimental Pathology, Institute for Frontier Medical Sciences, Kyoto University, Kyoto 606-8507, Japan Ted A. Torrey (97), Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland David Traver (253), Division of Hematology, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115 Nikolaus S. Trede (253), Division of Hematology, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115 Jo Van Damme (1), Laboratory of Molecular Immunology, Rega Institute for Medical Research, Minderbroedersstraat 10, University of Leuven, B-3000 Leuven, Belgium Jeffrey A. Yoder (253), Department of Biology, University of South Florida, 4202 East Fowler Avenue, SCA110, Tampa, FL 33620; Immunology Program, H. Lee Moffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Tampa, FL 33612 and Department of Pediatrics, Children’s Research Institute, University of South Florida, 140 Seventh Avenue South, St. Petersburg, FL 33701 Leonard I. Zon (253), Division of Hematology, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115 and Howard Hughes Medical Institute, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115


FOREWORD

Advances in Immunology is one of the oldest and most widely read and cited of all review publications in the field of Immunology. This issue of Advances in Immunology marks a major transition in the history of the journal. Frank Dixon, who has served as Editor in Chief of Advances in Immunology for nearly 4 decades since 1964, has ‘‘folded up his tent’’ and passed the position on to new leadership. The current editorial board, which has remained intact with Frederick W. Alt moving to the position of Editor in Chief, served with Dr. Dixon for a decade or more and are well-poised to maintain this outstanding publication. However, they all will miss Dr. Dixon’s leadership and his devotion to the series. Under Dr. Dixon’s hand, Advances in Immunology rapidly rose to become one of the highest impact immunology publications of any type, a position it has maintained until the present time. Dr. Dixon’s own work sought to understand the underlying biology of a variety of human disorders that were related to immunologic processes. Dixon’s work, in particular his studies of immune complex diseases, along with that of Henry Kunkel, who co-edited Advances in Immunology with Dixon from 1967 (volume 7) to 1986 (volume 38), was instrumental in leading to the establishment of Immunopathology, a discipline that studies how protective immune responses may go awry to cause injury and disease. In its early stages, Advances in Immunology reflected this discipline and featured work on the pathobiology of the immune response. In particular, it emphasized a mechanistic approach to the elucidation of the immunologic basis of human disease. Immunology arguably represents the most definitive example of systems biology to which state-of-the-art technology has been sequentially introduced. xiii

xiv

FOREWORD

As the discipline of Immunology progressed, Advances in Immunology paralleled this evolution by covering in vitro systems, animal models, cellular mechanisms, structure, genetics, and molecular genetics. Thus, Advances in Immunology has reflected the progression of our understanding of immunologic processes in molecular and cellular terms. In this tradition, Advances in Immunology will continue to present comprehensive reviews in immunology and related fields with a focus on current developments. In addition, we will continue to seek reviews that describe advances in separate fields that provide technical or conceptual impact on our thinking about immune processes. Likewise, we will be alert to problems in other systems that, as has happened in the past, may be advanced by analogies to mechanisms elucidated in the context of immune processes. The overall goal of Advances in Immunology is to provide comprehensive and timely information that will help direct future research and stimulate new ideas, while serving as a repository of up-to-date information in a field that remains very rapidly evolving. Thus, chapters are meant to provide extremely thorough coverage of the subject with full and comprehensive bibliographies so that each can serve as the choice source on the topic for several years. We will maintain the many unique aspects of Advances in Immunology that have helped make it so successful. One of these has been the active involvement of an editorial board that represents different sub-disciplines and different geographical locations. This board will remain constantly alert to new advances and continue to strive to select authors who can present their topic in a timely, accessible and balanced style. We will also maintain one of the most distinctive qualities of Advances in Immunology—the freedom to review areas in depth from an author’s own perspective. In this context, we continue to emphasize the lack of unnecessary restrictions on style and length. That said, we will introduce a few changes to further enhance Advances in Immunology. An ongoing challenge has been speed of publication. We aim to further accelerate our publication time to approximately 6 months from submission. We will also continue to publish figures in color as required to represent the data. Finally, a new feature will be the occasional publication of thematic volumes, which have been successful in other Advances series. In this regard, the next volume of Advances in Immunology will be a thematic volume on T cell subsets and will be guest-edited by Laurie Glimcher and Harvey Canter. In conclusion, the Editors would like once again to thank Frank Dixon for establishing Advances in Immunology as one of the most unique and important publications in our field. We obviously will continue to cover all of the areas currently emphasized by Advances in Immunology. However, we note that systems biology, as so well represented by the immune system, will continue to provide a basis for the application of newly emerging technologies

FOREWORD

xv

such as genomics. Moreover, the immune system undoubtedly will provide a biological readout that is necessary for the ultimate success of interpreting the significance of modern methodologies with respect to impact on basic biology and disease. Thus, the current editorial board will all work hard to continue the tradition of bringing all of the newest advances in immunology rapidly to print. Frederick W. Alt K. Frank Austen Tadamitsu Kishimoto Jonathan W. Uhr Emil R. Unanue


ADVANCES IN IMMUNOLOGY, VOL. 81

Regulation of the Immune Response by the Interaction of Chemokines and Proteases SOFIE STRUYF, PAUL PROOST, AND JO VAN DAMME1 Laboratory of Molecular Immunology, Rega Institute for Medical Research, Minderbroedersstraat 10, University of Leuven, B-3000 Leuven, Belgium

I. Introduction to Chemokines

Controlled cell migration is an essential mechanism in organogenesis during foetal life and after birth for protection of the body against pathogens. Indeed, T lymphocytes migrate during maturation from the bone marrow to the thymus where the self-recognizing T cells are deleted. Other leukocytes accomplish their maturation in the bone marrow and subsequently travel in the blood stream to their final destination. Some leukocytes head for peripheral immune organs like lymph nodes, others leave the blood stream at non-immune organs to develop into highly specialized phagocytic cells, which are part of the first line of defense against pathogens. Some of these cells can also capture antigen and start an adaptive immune reaction that allows mounting a stronger and faster response when there is a second encounter with the same microorganism. To reach this goal, phagocytes migrate after antigen uptake to the draining lymph node and present processed antigen to naive lymphocytes that continuously patrol throughout the body. It is obvious that the traffic of all these different cell types needs to be tightly controlled. The orchestrating molecules are called chemokines, i.e., small, secreted proteins that contain conserved cysteines and selectively attract specific leukocyte subtypes (Rollins, 1997; Wuyts et al., 1999b; Zlotnik et al., 1999; Gerard and Rollins, 2001; Schaniel et al., 2001; Yoshie et al., 2001; Locati et al., 2002). The number of cysteine residues and their position in the primary structure allows for division of the family into four groups: CC, CXC, CX3C and C chemokines. In CC chemokines the first two cysteines are adjacent, whereas in CXC and CX3C chemokines they are separated by one and three amino acids, respectively. The single cysteine motif (SCM) C chemokines SCM-1 and SCM-1 possess only two instead of four cysteines, corresponding to the second and fourth cysteine residue in other chemokines. Fractalkine is the only CX3C chemokine and possesses the unique feature to exist as a membrane-bound molecule. Normally, chemokines are secreted to act in a Tel.: þ 32 16 337348; fax: þ 32 16 337340

1

1 Copyright ß 2003 by Elsevier (USA) All rights of reproduction in any form reserved. 0065-2776/03 $35.00

2

S. STRUYF ET AL.

paracrine way as most cytokines do. Their receptors on target cells are G protein-coupled receptors characterized by seven transmembrane domains. Since the number of known chemokines has dramatically increased due to the identification of new sequences by bioinformatics, a systematic nomenclature (Table I) has been introduced (Murphy et al., 2000; Zlotnik and Yoshie, 2000). The classification of the ligands (L) and their receptors (R) reflects the conserved chemokine structure (CC, CXC, CX3C and C) and the numbering follows the already assigned gene numbers (L) or has a historical base (R). The chemokine genes were designated scy (small cytokine), with SCYa, SCYb, SCYc and SCYd corresponding to the structural subfamilies CC, CXC, C and CX3C, respectively. For example, the gene scya1 codes for the CC chemokine CCL1 that binds to CC chemokine receptor 8 (CCR8). A relation exists between the structural division of the chemokine family and the target cells of each subgroup (Wuyts et al., 1999b; Zlotnik et al., 1999). TABLE I HUMAN CHEMOKINE NOMENCLATURE CC familya

CXC family CXCL1 CXCL2 CXCL3 CXCL4 CXCL5 CXCL6 CXCL7 CXCL8 CXCL9 CXCL10 CXCL11 CXCL12 CXCL13 CXCL14

XCL1 XCL2

CX3CL1

a

GRO GRO GRO PF-4 ENA-78 GCP-2 PBP/NAP-2 IL-8 Mig IP-10 I-TAC SDF-1/ BCA-1 BRAK/bolekine

C chemokines Lymphotactin/SCM-1 SCM-1

CX3C chemokine Fractalkine

CCL1 CCL2 CCL3 CCL3L1 CCL4 CCL5 CCL7 CCL8 CCL11 CCL13 CCL14 CCL15 CCL16 CCL17 CCL18 CCL19 CCL20 CCL21 CCL22 CCL23 CCL24 CCL25 CCL26 CCL27 CCL28

I-309 MCP-1 MIP-1/LD78 MIP-1/LD78 MIP-1 RANTES MCP-3 MCP-2 Eotaxin MCP-4 HCC-1 Lkn-1 LEC TARC DC-CK1/PARC MIP-3 /ELC MIP-3/LARC 6Ckine/SLC MDC MPIF-1/CK 8 MPIF-2/eotaxin-2 TECK Eotaxin-3 CTACK/ILC MEC

CCL numbers that are omitted, refer to CC chemokines of which only the mouse homologue is identified.

THE CD26-CHEMOKINE CONNECTION

3

CXC chemokines attract neutrophils if their primary structure contains the sequence Glu-Leu-Arg (ELR-motif ) just in front of the first cysteine; if they don’t possess this tripeptide motif then they induce migration to lymphocytes or NK cells. CC chemokines have a more divergent spectrum of action. They can activate monocytes, lymphocytes, eosinophils, basophils and/or dendritic cells. Lymphotactin (XCL1) attracts NK cells and lymphocytes, whereas fractalkine (CX3CL1) acts on T cells, NK cells and monocytes. Fractalkine is a unique chemokine as far as its structure is concerned. It is membrane-bound through a mucin-like stalk, although it can also be released from the cellular surface. The other chemokines are secreted and have a common tertiary structure consisting of an exposed NH2-terminal region connected by a loop to a central three-stranded anti-parallel -sheet, over which the COOH-terminal -helix following the last cysteine residue is orientated (Murphy et al., 2000). The NH2-terminus of the chemokine is involved in high-affinity receptor binding; the amino acids in the COOHterminal -helix form the site of low-affinity interaction with heparin or glycosaminoglycans (GAG). The latter part is the docking site for interaction with the extracellular matrix or heparan sulfates on endothelial cells and contributes to the formation of a chemokine gradient. Chemokines seem to act in a network, constituted by many overlapping interactions between ligands and their receptors. Multiple chemokines can recognize the same receptor and one chemokine can bind to several receptors. This complex organization guarantees a solid defense system that assures protection, even if one of the links in the web is affected. This robustness of the chemokine system, in its most prominent role, is evidenced by the apparent health of most chemokine or chemokine receptor knockout mice. However, when more specialized disease models are tested, a specific role for individual receptors and their ligands can be demonstrated. Exceptions are the mice of which the CXCR4 gene or the gene of its ligand, stromal cellderived factor-1 (SDF-1) (Nagasawa et al., 1994), is disrupted by gene targeting. These animals die perinatally of defects in embryonic neuronal cell migration, organ vascularization and hematopoiesis (Nagasawa et al., 1996; Ma et al., 1998). In zebrafish this ligand-receptor pair guides migratory germ cells toward their target tissues (David et al., 2002; Knaut et al., 2003). The general biological functions and the more unique roles of selected members of the chemokine family are discussed in the following subsections. A. INFLAMMATORY CHEMOKINES The connection between chemotaxis and members of the chemokine family, originally designated as the pro-inflammatory ‘‘intercrine’’ cytokine family (Oppenheim et al., 1991), was only revealed a decade after the identification of the first chemokine structure. In 1977, a procoagulant and angiostatic

4

S. STRUYF ET AL.

factor stored in platelet -granules, called platelet factor-4 (PF-4) was sequenced (Walz et al., 1977). Subsequently, structurally related cDNAs were isolated from cells stimulated with inflammatory mediators (Oppenheim et al., 1991). However, immunologists still had not solved the enigma of early and selective neutrophil influx at inflammatory sites. Until then, only the complement-derived peptide C5a and some other molecules attracting leukocytes, e.g., the arachidonic acid derivative leukotriene B4, without discriminating between subsets, were known. By the end of the eighties, several groups identified a selective neutrophil chemotactic protein, interleukin-8 (IL-8). Soon, the chemokine family was extended with structurally related peptides inducing selective migration of other leukocyte subtypes: the monocyte chemoattractant MCP-1 (monocyte chemotactic protein-1) and the eosinophil chemoattractant eotaxin. Known ‘‘intercrines’’, e.g., RANTES (‘‘regulated on activation normal T cell expressed and secreted’’) and MIP-1 (macrophage inflammatory protein-1), were demonstrated to attract lymphocytes. The group of chemokines identified on the basis of biological activity is now called the inflammatory chemokine group, since these proteins are produced under inflammatory conditions. Neutrophil-attracting chemokines, such as IL-8 are released early after pathogen encounter at the site of infection to recruit polymorphonuclear phagocytes. Later on, mononuclear leukocytes infiltrate the inflamed tissue in response to CC chemokines e.g., MCP-1 and MIP-1. Eosinophil influx in allergic airway inflammation is mediated by eotaxin or other CC chemokines. Cytokines and bacterial, viral or plant products turn on the expression of inflammatory chemokines and almost any cell type produces these proteins if appropriately stimulated. Especially IL-8 and MCP-1 are abundantly and ubiquitously expressed chemokines. The expression pattern of some other chemokines seems to be more restricted in terms of producer cells or inducing agents. To optimally induce chemokine production synergy between several cytokines is often required. For instance, MCP-2 could be isolated only in a minute amount from conditioned media, but if cell cultures are stimulated with interferon (IFN ) or IFN plus IL-1 , production levels that match that of MCP-1 could be reached (Van Damme et al., 1992, 1994; Struyf et al., 1998c). It is most probable that in vivo a number of chemokine inducers are simultaneously released during an inflammatory response to act in concert. Although certain mediators can selectively induce individual chemokines, most often pro-inflammatory cytokines like tumor necrosis factor- (TNF-), IL-1 or IFNs will elicit the release of several chemokines with overlapping activities. Tissue cells are not the sole source of inflammatory chemokines. Platelets store RANTES, PF-4 and cleavage products of platelet basic protein (PBP), which are released when platelets are activated. At the site of platelet


5

aggregation, these chemokines stick to the endothelium through their heparan sulfate-binding site, to prevent them from being washed away by the blood stream. Also chemokines produced by endothelial cells near inflammatory sites can be immobilized on the vessel wall. The function of captured chemokines is to activate circulating leukocytes so as to allow strong interaction between cellular adhesion molecules and subsequent diapedesis of the leukocyte through the vessel wall. Furthermore, chemokines will stimulate leukocytes to release extracellular matrix degrading enzymes that are necessary for passage through the extracellular space. Thus, in addition to guiding leukocytes through the tissue to sites of infection, inflammatory chemokines are important signals for leukocytes to leave the blood stream since they enable circulating cells to firmly adhere to the endothelium and transmigrate the vessel wall. B. LYMPHOID CHEMOKINES Recently (1996–2001), a number of new chemokine cDNAs were identified by screening databases of expressed sequence tags (ESTs) with known chemokine sequences. Production of these new chemokines seems to be constitutive and often limited to a particular lymphoid organ or leukocyte subtype. The rather restricted expression pattern and low expression levels compared to those of classical inflammatory chemokines are the reasons why these chemokines were not discovered by routine biological tests. Functional characterization of these new chemokines points to a different role in the immune system. As a result, the chemokine family can be functionally divided into inflammatory and constitutive chemokines. Alternative names for these ‘‘second generation’’ chemokines are ‘‘lymphoid’’ or ‘‘homeostatic’’ chemokines, because they are involved in the regulation of physiological lymphocyte trafficking, in T and B cell development and in structuring lymphoid tissues (Melchers et al., 1999; Rossi and Zlotnik, 2000). Migration of immunocompetent cells is crucial during their development and during the generation of an efficient immune response. Leukocytes are generated from pluripotent hematopoietic stem cells in the bone marrow. Whereas T cells undergo their maturation in the thymus before populating the T cell-rich areas of spleen and lymph nodes, B cells stay in the bone marrow until they express surface IgM. After negative selection, B cells leave the bone marrow and head toward the spleen, where they localize in the ‘‘peri-arteriolar lymphoid sheath’’ (PALS) and complete their maturation (expression of surface IgM and IgD). At this stage, B cells are ready to form primary follicles. Although all the signposts in this T and B cell traffic between the immunological organs are not yet identified, a role for SDF-1 and secondary lymphoid-tissue chemokine (SLC) and their respective receptors CXCR4 and

6

S. STRUYF ET AL.

CCR7 has been demonstrated (Melchers et al., 1999). Other chemokines were shown to regulate the internal organization of lymphoid organs and promote cell–cell contacts required for a helper T cell-dependent humoral immune response. Activated B cells and dendritic cells in T cell areas of lymph nodes release the CCR4 ligands macrophage-derived chemokine (MDC) and thymus and activation-regulated chemokine (TARC), to make contact with T cells (Schaniel et al., 1999; Tang and Cyster, 1999). Follicular dendritic cells establish interaction with B cells by producing B cell-attracting chemokine-1 (BCA-1) (Fo¨rster et al., 1996; Gunn et al., 1998). Within the thymus different chemokines regulate the migration of T cells during their development from the cortex to the medulla through the temporary expression of chemokine receptors (Campbell et al., 1999) (Fig. 1). Immature CD4 þ /CD8 þ thymocytes in the cortex are responsive to thymus-expressed chemokine (TECK) and SDF-1. Subsequently, as they enter the medulla after positive selection, they additionally express CCR4. Single-positive cells that have only CD4 or CD8 on their cellular surface migrate to MIP-3 and SLC. Finally, mature T cells that express the peripheral homing molecule L-selectin are not responsive anymore to the CCR4 ligands (TARC, MDC) and TECK. These examples constitute only a few steps in the complex cascade of cell interactions and cell movement leading to the elaboration of an effective immune system. Many more links of the network and steps in the cascade of reactions need to be unravelled, but it is clear that chemokines and their receptors play an important role in this process. Although the subdivision of chemokines based on their mode of expression seems attractive, some chemokines appear to have a dual expression pattern and can be classified both as constitutive and inflammatory. The CCR4 ligand MDC is constitutively expressed in the thymus medulla where it functions as a thymocyte attractant, but its expression can be regulated in dendritic cells by inflammatory stimuli (Vulcano et al., 2001). Furthermore, this chemokine was detected by immunohistochemistry in allergic inflammation of the skin (Vestergaard et al., 1999; Galli et al., 2000). For other ‘‘constitutive’’ chemokines, such as hemofiltrate CC chemokine-1 (HCC-1) the cellular source is not yet known, thus their classification is preliminary. Since NH2terminally processed HCC-1 is a potent ligand for CCR1 and CCR5 (Detheux et al., 2000), a rather inflammatory role can be predicted on the basis of its receptor usage. The most useful feature of this functional division of the chemokine family is that it clearly shows a fundamental role of its members in immunology that is more extended than attracting phagocytes to an inflammatory focus. Other recognized chemokine effects (Wang et al., 1998; Belperio et al., 2000; Hasegawa et al., 2000; Youn et al., 2000; Mu¨ller et al., 2001) on leukocytes (hematopoiesis), endothelial cells (angiogenesis) and tumor cells


7

FIG. 1. Chemokine receptor expression during T cell maturation in the thymus. Schematic representation of the maturation of T cells in the thymus. During this process, thymocytes gradually migrate from the outer cortex to the medulla, where they enter the blood stream as mature T cells. Along this route, the expression of cell surface markers [T cell receptor complex (TCR), L-selectin (L-sel), CD4 and CD8] on thymocytes changes and interactions take place with different stromal cells. First, thymocytes are positively selected for recognition of major histocompatibility molecules on cortical epithelial cells (EC). In the medulla antigen presenting cells (APC), macrophages (M) and dendritic cells (DC), induce deletion of self-recognizing T cells. The function of the Hassall’s corpuscles in the medulla, containing concentric layers of degenerating epithelial cells, is not yet known. At the different maturation stages, thymocytes selectively express different chemokine receptors. As a result, their responsiveness to chemokines changes. It is hypothesized that selectively produced chemokines, e.g., TECK in the cortex, and MDC, MIP-3 and SLC in the medulla, form a chemotactic gradient to guide the thymocytes through the thymus.

(metastasis) are at present not well understood and are therefore not discussed in detail. The recent discovery of regakine-1 as a constitutive plasma chemokine that synergizes with other CC and CXC chemokines, as well as with the bacterial chemotactic peptide formyl-methionyl-leucyl-phenylalanine (fMLP) adds another dimension of action to the chemokine system (Struyf et al., 2001b,c).

8

S. STRUYF ET AL.

C. CHEMOKINE RECEPTORS Chemokines and other leukocyte chemoattractants bind to G proteincoupled receptors (Murdoch and Finn, 2000; Murphy et al., 2000). These receptors contain approximately 350 amino acids of which 25–80% are conserved. Seven -helices of 20–25 hydrophobic residues each span the cell membrane and connect the internal COOH-terminal tail via six loops (three intra- and three extracellular) with the extracellular NH2-terminus. The primary sequence of the NH2-terminus is the most divergent as it constitutes a major region of interaction with the ligands. It may be sulfated on Tyr residues and contains N-linked glycosylation sites, as well as a Cys that is linked to another Cys in the third extracellular loop by a disulfide bond. This connection is critical for chemokine binding. Interaction of chemokines with their receptor occurs in two steps (Crump et al., 1997; Monteclaro and Charo, 1997) (Fig. 2). First, the ligand binds to the NH2-terminus of the receptor at high-affinity binding sites. Next, low-affinity interactions occur, the flexible NH2-terminal tail of the chemokine binds to central domains of the receptor and triggers receptor signaling (Murphy et al., 2000).

FIG. 2. A model for activation of chemokine receptors. The current hypothesis proposes that chemokine ligands interact (arrows in B and C) in two steps with two sites of their heptahelical receptor, first with the NH2-terminus and then with one or more extracellular domains (Crump et al., 1997; Monteclaro and Charo, 1997; Murphy et al., 2000). In the first step (B) high-affinity interaction between the NH2-terminus of the ligand and the NH2-terminus of the receptor triggers a conformational change in the receptor exposing the second binding domain. This latter binding cleft (formed by several residues in the extracellular domains of the receptor) interacts with additional residues in the flexible NH2-terminal region of the chemokine, triggering activation of a G protein (C). As a result GDP can be exchanged for GTP, which finally results in dissociation of the G protein in the G and G G subunits and the activation of secondary effectors.


9

Binding to G proteins occurs via the intracellular parts of the chemokine receptor. Specifically, the COOH-terminus contains conserved residues for signal transduction, including Ser and Thr residues that are phosphorylated to desensitize the receptor (Murdoch and Finn, 2000). A common feature of this type of receptors is that cellular responses to agonists are interrupted within seconds to minutes through rapid G-protein uncoupling. Different mechanisms contribute to attenuation of the signal: receptor desensitization by phosphorylation, receptor endocytosis and receptor down modulation (Bo¨hm et al., 1997). The latter process only occurs after prolonged exposure of the cells to ligand by depletion of the cellular receptor content through alterations in the rates of receptor degradation and synthesis. Two types of desensitization that are executed by different types of kinases can be distinguished. Homologous desensitization is mediated by G protein-coupled receptor kinases after agonist binding to the receptor. In contrast, heterologous desensitization does not require ligand binding. This latter process is mediated by a second messenger-dependent kinase, such as protein kinase A or protein kinase C, that has been activated by a different receptor (Bo¨hm et al., 1997). For instance, the bacterial chemotactic peptide fMLP can desensitize the calcium and chemotactic response of neutrophils to IL-8 although fMLP and IL-8 activate different receptors (Tomhave et al., 1994; Campbell et al., 1997). Chemokine receptor expression is tightly regulated and is dependent on the cell type, its maturation and activation status (Fig. 3). For example, some chemokine receptors are differentially expressed on TH1 and TH2 cells. This selective expression profile can be exploited to discriminate between these cell types (Annunziato et al., 1999). Immature Langerhans’ cells can be distinguished from other immature dendritic cells by the selective expression of CCR6 (Charbonnier et al., 1999). As already mentioned, thymocytes express different receptors for homeostatic chemokines as they migrate from the thymus cortex to the medulla (Campbell et al., 1999). Dendritic cells also switch chemokine receptor expression during maturation (Dieu-Nosjean et al., 1999). Immature dendritic cells express CCR1 and CCR5 for migration in response to the inflammatory chemokines MIP-1, MIP-1 and RANTES. After encounter of lipopolysaccharide (LPS) or another maturation stimulus at sites of inflammation, CCR1 and CCR5 are downregulated on the dendritic cell surface and CCR7 is newly synthesized and expressed to enable migration to the draining lymph node (Dieu et al., 1998; Sallusto et al., 1998; Sozzani et al., 1998; Yanagihara et al., 1998). Furthermore, activation of T cells by IL-2 induces the expression of CXCR3 (Loetscher et al., 1998). Downmodulation of CCR1, CCR2 and CCR5 on monocytes can be achieved by incubation with LPS (Sica et al., 1997). It was proposed that this mechanism would arrest and retain monocytes at sites of inflammation by rendering the cells insensitive for the established chemotactic gradient. Recently, it was shown that, in an

10

S. STRUYF ET AL.

FIG. 3. CC chemokine receptor expression by non-activated human leukocytes. Human monocytes (1), lymphocytes (2), dendritic cells (3), neutrophils (4) and eosinophils (5) express different CC chemokine receptors. The expression of CCR2 and CCR3 on dendritic cells and CCR3 on monocytes was demonstrated at the mRNA and/or protein level, but signal transduction through these receptors could not be detected. Receptor agonists are connected with their receptors by solid arrows, receptor antagonists by dashed arrows. Full-length eotaxin and MIP-1 have been demonstrated to be high affinity ligands for CCR2 and CCR1, respectively, but are only poor agonists and can function as antagonists in blocking calcium or chemotactic responses to more effective receptor ligands (Martinelli et al., 2001; Ogilvie et al., 2001; Chou et al., 2002).

inflammatory environment, IL-10 inhibits downregulation of chemokine receptors on monocytes by LPS and uncouples the receptors from their signaling pathway (D’Amico et al., 2000). Consequently, the chemokine receptors remain expressed at the cellular surface and sequester chemokines present without being triggered. Furthermore, IL-10 inhibits production of inflammatory chemokines, such as MCP-1 and MIP-1 by monocytes and dendritic cells. In this way, the anti-inflammatory cytokine IL-10 attenuates inflammation by blocking leukocyte recruitment and activation. In addition to leukocytes, endothelial cells also express chemokine receptors e.g., CXCR4 and CXCR3, as well as the chemokine receptor-like proteins Duffy antigen receptor (DARC), and D6 (Horuk, 1994; Nibbs et al., 1997). Since chemokine binding to DARC and D6 is not followed by signal transduction, DARC and D6 are thought to function as decoy receptors for


11

inflammatory chemokines. DARC expressed on erythrocytes is considered to function as a sink to rapidly capture chemokines in the blood circulation (Horuk, 1994). DARC expressed by vascular endothelial cells could also participate in transcytosis and surface presentation of inflammatory chemokines on the luminal side of the endothelial layer (Middleton et al., 1997). Analogously, D6 might be crucial for chemokine presentation on endothelial cells of the afferent lymphatics in lymph nodes (Nibbs et al., 2001). However, D6 could also act as a scavenger because inflammatory chemokines that are internalized after binding D6 are intracellularly degraded (Fra et al., 2003). D. CHEMOKINES

ARE INHIBITORS OF

HIV INFECTION

It is known for more than a decade that cells expressing CD4 are susceptible for infection by HIV-1. However, CD4 expression is not the only prerequisite to render cells permissive for fusion with HIV-1 virions, an additional molecule or ‘coreceptor’ is needed. Furthermore, the CD4 molecule cannot explain the HIV-1 strain tropism. Syncytium inducing strains only infect T cells (T-tropic HIV-1 strains); non-syncytium inducing strains can also enter macrophages (M-tropic HIV-1 strains) (Littman, 1998). The discovery that chemokine receptors are the missing coreceptors for HIV-1 entry in CD4 þ cells was a major breakthrough in 1996 (Balter, 1996) and has been extensively reviewed (Lee and Montaner, 1999; Doms and Trono, 2000; Kinter et al., 2000; O’Brien and Moore, 2000). First it was shown that some chemokines could block infection of peripheral blood mononuclear cells (PBMC) with HIV-1 (Cocchi et al., 1995). Later, the receptor recognizing this inhibitory chemokines (CCR5) was demonstrated to be essential for the virus to achieve membrane fusion with the host cell (Choe et al., 1996; Deng et al., 1996; Doranz et al., 1996; Dragic et al., 1996). It is now clear that CXCR4, the receptor for SDF-1 is the coreceptor for T-tropic or X4 HIV-1 strains (Bleul et al., 1996) and that CCR5, binding LD78, LD78 , MIP-1 , RANTES, MCP-2, MCP-3 and MCP-4 is used by M-tropic or R5 HIV-1 strains (Doms and Trono, 2000; Murphy et al., 2000). The cell tropism also correlates with disease progression (Scarlatti et al., 1997). Strains isolated from patients in the early, asymptomatic disease state are M-tropic. Sometime before the disease progresses to AIDS, dual-tropic strains recognizing both CCR5 and CXCR4 occur and afterwards more pathological T-tropic strains are found. The increased resistance to HIV-1 infection of persons carrying a deletion in their CCR5 gene, termed 32, indicates that the disease is primarily transmitted by R5 strains (Samson et al., 1996). Besides CCR5 and CXCR4, a multitude of alternative coreceptors have been identified that enable a restricted number of HIV strains to infect cells, but their in vivo relevance is still questionable (Doms and Trono, 2000). It is generally accepted that CCR5 and CXCR4 are the major HIV-1 coreceptors,

12

S. STRUYF ET AL.

because viral isolates that can use multiple coreceptors (in addition to CCR5 or CXCR4) are completely inhibited by CCR5- and/or CXCR4-specific antagonists. However, these less frequently used coreceptors could become more important as an escape route for the virus during treatment with R5 or X4 antagonists (Doms and Trono, 2000). The first contact between the target cell and HIV-1 virus can occur through many different molecules, although CD4 binding is a prerequisite for virus entry (Doms and Trono, 2000). The interaction of the viral envelope protein gp120 with CD4 induces a conformational change in gp120 that enables it to bind to the coreceptor. This contact with a second receptor leads to membrane fusion, a process in which the other viral envelope protein gp41 is implicated. Upon coreceptor triggering, the hydrophobic NH2-terminal fusion peptide of gp41 is exposed and interacts with the host cell membrane (Chan and Kim, 1998). Membrane fusion is a cooperative process, suggested to involve four to six CCR5 receptors, multiple CD4 and three to six gp41 molecules (Kuhmann et al., 2000). Interaction regions that determine the preferential chemokine receptor use are located in the V3 variable domain of the viral gp120 subunit (Cocchi et al., 1996; Xiao et al., 1998). Mutations arising at this site of gp120 accompany transition of R5 to X4 in patients with progressive HIV infection and allow escape of the virus from the antiviral control of CC chemokines. A second, highly conserved region in the bridging sheet of gp120 is involved in coreceptor binding and is possibly the target for antibodies that can neutralize HIV-1 strains that use different chemokine receptors (Rizzuto et al., 1998). This common domain becomes exposed after contact of gp120 with CD4, and allows interaction with the coreceptor. In some HIV strains that are capable to infect CD4/coreceptor þ cells, this conserved region is probably constitutively exposed (Hoffman et al., 1999; Reeves et al., 1999). As a consequence, no change in the configuration of gp120 is required to enable binding to the coreceptor and subsequent viral entry. Interaction between gp120 and a coreceptor is blocked by the corresponding chemokine ligands. In addition to the hindering of this binding, chemokines trigger internalization of their receptors. Besides, prolonged exposure of G protein-coupled receptors to their ligands causes downmodulation of the receptor. Consequently, chemokines and small molecule chemokine receptor antagonists were proposed as antiviral therapeutics. Alternatively, chemokine receptor expression can be influenced by the microenvironment of the cell. As mentioned above, cytokines, bacterial products and cell–cell interactions can modulate chemokine receptor expression. For example, stimulation of CD4 þ T cells with immobilized antibodies to CD3 and CD28 downregulates CCR5 expression (Carroll et al., 1997). Furthermore, cross talk between different G protein-coupled receptors


13

is possible and downmodulation of HIV coreceptors has been demonstrated through triggering of non-coreceptor G-protein-coupled receptors on monocytes (Deng et al., 1999; Lai et al., 2001). Finally, it was suggested that the CCR2-64I variant (Ile on position 64), but not wild type CCR2 (Val on position 64) is able to form heterodimers with CCR5 and/or CXCR4 sequestering these coreceptors in the endoplasmatic reticulum (Mellado et al., 1999). This might slow down the expression of the coreceptors on the cell surface and can explain the delay in progression to AIDS in individuals carrying the CCR2-64I allele (Smith et al., 1997). In conclusion, the discovery of the important role that chemokine receptors play in HIV infection opened a whole new field of drug development for AIDS therapy. II. Posttranslational Modification of CXC Chemokines

The most prominent and intensively studied function of chemokines is the attraction of leukocytes to inflammatory foci. To defend the organism against invading microbes, chemokines recruit phagocytes, often in synergy with other cytokines or microbial products (e.g., fMLP) (Struyf et al., 2001c). In addition, chemokines can activate the attracted phagocytes and trigger the release of proteases and the production of reactive oxygen species. Thus, chemokines can have destructive effects as well, and therefore their production and functioning are tightly regulated. This is also the case for other cytokines and most of the thoroughly studied mechanisms that guard cytokine activity can be superposed on chemokine biology. The different control levels that can be distinguished are first at the transcriptional and translational level of the cytokine gene, second at the protein level and finally at the receptor level. Here, we focus on the protein level, in particular, on chemokine processing by proteases. The original reports on the isolation and identification of a neutrophil chemotactic protein, IL-8, from leukocyte culture supernatant (Schro¨der et al., 1987; Walz et al., 1987; Yoshimura et al., 1987; Van Damme et al., 1988) indicated that the obtained NH2-terminal sequence did not match with the predicted mature protein sequence derived from the IL-8 cDNA (Schmid and Weissmann, 1987). Furthermore, a second chemokine structure described, -thromboglobulin ( -TG), was shown to be a posttranslationally modified form of PBP (Begg et al., 1978; Holt et al., 1986). Hence, evidence that CXC chemokines are natural protease substrates is present in the early chemokine literature. In the following years enzymes capable of clipping PBP and IL-8 were identified. Only a decade later, the reports describing naturally NH2-terminally processed CC chemokines first appeared (Noso et al., 1996). By now, a multitude of proteases acting on chemokines (Tables II and III) has been described and their effect on chemokine activity is far-reaching.

14

TABLE II HUMAN CXC CHEMOKINE VARIANTS

Chemokine

Isoforms

NH2-term. residue

Cellular source of chemokine variant

Degradation

NDb

NIb

GRO

(2–73) (3–73) (4–73) (5–73) (6–73) (3–73) (5–73) (5–73) Degradation (17–71)

Ser Val Ala Thr Glu Leu Thr Thr ND Ser

CTAP-III/ TGd

(8–78) (9–78) (3–77) (6–77) (9–77) Degradation

NAP-2d IL-8 IL-8 IL-8 IL-8

GRO

GRO c GRO PF-4 PF-4 ENA-78 ENA-78 GCP-2 GCP-2

Receptor recognitiona

Effect on biological activitya

NAb

Inactivation

Osteosarcoma cells

MMP-9, proteinase 3, elastase ND

ND

ND

Monocytes

ND

ND

" Neu activation

BM stroma

CD26/DPP IV?

" CXCR2

ND MMP-9 ND

ND NA Unknown

Leu Arg Val Val Glu ND

Monocytes NA Activated leukocytes/platelets Monocytes Monocytes Osteosarcoma cells Osteosarcoma cells

ND Cathepsin G CD26/DPP IV ND

ND ND ND ND

" Neu activation HSC mobilizationc " Neu activation Inactivation " Endothelial cell growth inhibition " Neu activation " Neu activation ¼ Neu activation ¼ Neu activation

NI

NA

Inactivation

(59–128) Degradation (6–77) (7–77)

Ala ND Ser Ala

Platelets/leukocytes NA Monocytes, lymphocytes Mononuclear leukocytes

Proteinase 3 elastase MMP-9 Cathepsin G Elastase, CD13 cathepsin G Plasmin thrombin MMP-9

" Neu activation Inactivation " Neu activation " Neu activation

(8–77)

Lys

Mononuclear leukocytes

Proteinase 3

" CXCR2 NA ND " CXCR1 " CXCR2 ND

c

" Neu activation

S. STRUYF ET AL.

GRO

Protease implicated

(9–77) (3–103)

Glu Val

Keratinocytes NI

ND CD26/DPP IV

ND # CXCR3

IP-10

Ser Arg Thr Leu

Osteosarcoma cells

ND

ND

IP-10

(4–77) (5–77) (6–77) (3–77)

NI

CD26/DPP IV

# CXCR3

I-TAC SDF-1

(3–73) (3–67)

Met Val

NI NI

CD26/DPP IV CD26/DPP IV

# CXCR3 # CXCR4

SDF-1

(4–67)

Ser

NI

Elastase

# CXCR4

SDF-1

(5–67)

Leu

NI

# CXCR4

SDF-1

(6–67)

Ser

NI

MMP-1, -2, -3, -9, -13 and -14 Cathepsin G

# CXCR4

ND # Ly chemotaxis ¼ angiogenesis inhibition ND

# Ly chemotaxis ¼ angiogenesis inhibition # Ly chemotaxis # Ly chemotaxis # CD34 þ chemotaxis # X4 HIV inhibition # Ly chemotaxis # X4 HIV inhibition # Ly chemotaxis # X4 HIV inhibition # Ly chemotaxis

a The effect of NH2-terminal truncation of the chemokine on its receptor recognition and its biological activity is indicated by " for increase, by # for decrease or by ¼ when unaffected. Leukocyte subtypes are indicated by Neu, neutrophils, Ly, lymphocytes or CD34 þ , CD34 þ progenitor cells from cord blood. b ND, not determined; NI, this isoform was not isolated from conditioned media but was generated by incubating recombinant chemokine with protease; NA, not applicable. c This GRO isoform is produced by bone marrow stromal cells (BM stroma) and is possibly generated by sequential CD26/DPP IV cleavage after the Pro and Ala residues at the 2nd and 4th NH2-terminal position. After subcutaneous injection of this chemokine variant (alone or in combination with granulocyte colony-stimulating factor) longterm repopulating hematopoietic stem cells (HSC) were mobilized into the peripheral blood. d For CXCL7, LDGF is considered as the full-length protein. The biological function of the CXCL7 isoforms PBP, CTAP-III and TG is not fully established, although it is suggested in some reports that these isoforms act as receptor antagonists (Brandt et al., 2000).


IL-8 Mig

15

16

TABLE III HUMAN NH2-TERMINAL CC CHEMOKINE VARIANTS GENERATED

Chemokine

MCP-4 Eotaxin Eotaxin HCC-1 HCC-1 PARC MDC MDC

NH2-term. residue

PROTEASES

OTHER THAN

Cellular source of chemokine variant

Protease implicated

CD26/DPP IV Effect on biological activitya

(5–70) (3–69) (5–76) (6–76) (5–76) (6–76) (5–76) (4–76) (5–76) (8–76) (4–74) Degradation (3–74) (4–74) (9–74)

Ala Met Ile Asn Val Ser Ile Ala Leu Pro Ser ND Thr Glu Gly

Monocytes Lymphocytes Mononuclear leukocytes Mononuclear leukocytes NIb Mononuclear leukocytes Fibroblasts NI

NDb ND, CD26/DPP IV? MMP-1, and -3 ND MMP-3 ND MMP-1, -2, -3, -13 and -14 MMP-1, MMP-3

ND " CCR1, # CA for # CA for # CA for # CA for # CA for # CA for

NI Fibroblasts NAb Plasma

MMP-1 ND Metalloprotease ND

# CA for Mo (binds CCR1, 2 and 3) ND Inactivation ND

Plasma

(3–69) (4–69) (7–69) (9–69)

Val Gly Met Asp

Mononuclear leukocytes and platelet rich plasma Dendritic cells CD8 þ T cells

Urokinase plasminogen activator, plasmin ND

" CA for Mo, Ly, Eo; inhibits R5 HIV-1 (" CCR1, 3 and 5) ND

ND ND

ND ND

2 inhibits R5 HIV-1 Mo (binds CCR2) Mo Mo (binds CCR1, 2 and 3) Mo Mo (binds CCR1, 2 and 3) Mo (binds CCR1, 2 and 3)

a The increase (") or decrease (#) in receptor binding or in chemotactic activity (CA) for the different leukocytic cells types (Mo, monocytes; Ly, lymphocytes; Eo, eosinophils) is indicated. b ND, not determined; NA, not applicable; NI, not isolated.

S. STRUYF ET AL.

LD78/ MIP-1 MCP-1 MCP-1 MCP-2 MCP-2 MCP-3 MCP-4

Isoforms

BY


17

Structurally, the CXC chemokine subfamily can be divided based on the presence or absence of a tripeptide (Glu-Leu-Arg, ELR-motif) in their NH2terminus. Depending on the presence of this ELR-motif, the chemokine will act primarily on neutrophils via CXCR1 and/or CXCR2 (ELR þ ; CXCL1 to CXCL8, except CXCL4; Table I) or lymphocytes (ELR, CXCL4 and CXCL9 to CXCL14; Table I). Furthermore, ELR þ CXC chemokines stimulate angiogenesis, i.e., the formation of new blood vessels, whereas ELR CXC chemokines tend to be inhibitors of angiogenesis and endothelial cell proliferation (Strieter et al., 1995). The following examples illustrate that the biological activities of CXC chemokines are differently influenced by limited truncation of their NH2-termini. A. ELR þ CXC CHEMOKINES It has been hypothesized that the ELR-motif is essential for selective receptor binding and activation (Clark-Lewis et al., 1994). Indeed, so far no natural isoforms of ELR þ CXC chemokines have been isolated which are trimmed beyond the tripeptide motif and all of these truncated variants are at least as active as the intact chemokine protein. Further proof for this theory is found in the structure and activity of the different CXCL7/PBP isoforms. In the longer forms of CXCL7, the extended NH2-terminus covers the ELR-motif, explaining why this chemokine has to be processed to become active on neutrophils (Malkowski et al., 1997). The longest CXCL7 form, leukocyte-derived growth factor (LDGF), is secreted by activated neutrophils, macrophages and T lymphocytes (Iida et al., 1996). LDGF is a mitogenic factor for fibroblasts and is the most recently described CXCL7 variant. Earlier on, the CXCL7 isoforms isolated from platelet supernatants were characterized: PBP (94 residues), connective tissue activating protein-III (CTAP-III, 85 residues) and -thromboglobulin ( -TG, 81 residues) (Begg et al., 1978; Holt et al., 1986). PBP and in particular CTAP-III are stored in the -granules of platelets and are released by thrombin stimulation. The simultaneous release of proteases during platelet aggregation accounts for the immediate conversion of PBP and CTAP-III to -TG (Holt et al., 1986). The shortest CXCL7 isoform, neutrophil-activating peptide-2 (NAP-2) can be generated from CTAP-III by cathepsin G (Brandt et al., 1991; Car et al., 1991; Cohen et al., 1992). NAP-2 has an intact ELR-motif and chemoattracts neutrophils by binding to CXCR2 (Wuyts et al., 1999b). The crystal structure of a synthetic, inactive NAP-2 isoform with five additional NH2-terminal residues is shown to be similar to that of NAP-2, but the NH2-terminal extension folds back, masking the ELR-region (Malkowski et al., 1997). Whereas NAP-2 is the only natural neutrophil-activating PBP isoform, the natural GRO, GRO and epithelial cell-derived neutrophil attractant-78 (ENA-78) variants all have neutrophil chemotactic properties, the shorter

18

S. STRUYF ET AL.

forms (with an intact ELR-motif) being more potent than the full-length proteins (Proost et al., 1993b; Nufer et al., 1999; Wuyts et al., 1999a). ENA-78 can be cleaved by cathepsin G into ENA-78(9–78) (Nufer et al., 1999). The proteases responsible for the processing of GRO and GRO are not yet identified. The truncated proteins are, however, more abundantly present in leukocyte-conditioned medium than in supernatants of tumor cells (Wuyts et al., 1999a). Similar results were obtained for IL-8. Fibroblasts (Van Damme et al., 1989a; Schro¨der et al., 1990) and endothelial cells (Gimbrone et al., 1989) predominantly produce IL-8 with a longer NH2-terminus, whereas in leukocyte-derived preparations (Yoshimura et al., 1987, 1989; Gregory et al., 1988; Lindley et al., 1988; Schro¨der et al., 1988; Van Damme et al., 1989b) mainly NH2-terminally truncated forms are found. Apparently, the leukocyteconditioned media contain more enzymes capable of converting CXC chemokines to more active forms. In this way a positive feedback loop is provided during inflammation. Indeed, IL-8 can induce granule release from neutrophils, liberating the matrix metalloprotease 9 (MMP-9) or gelatinase B, of which IL-8 is proven to be a substrate (Van den Steen et al., 2000; Opdenakker et al., 2001). Also other proteases are reported to process IL-8 (He´bert et al., 1990; Nakagawa et al., 1991; Padrines et al., 1994; Kanayama et al., 1995). MMP-9 converts the mature 77-residue IL-8 into IL-8(7–77), which is a more potent neutrophil activator (Van den Steen et al., 2000). It is shown that the affinity of IL-8(7–77) for both CXCR1 and CXCR2 is enhanced compared to IL-8(1–77). These data confirm previous in vitro experiments that characterized the shorter IL-8 variants to be more active than the full-length protein (Nourshargh et al., 1992). Enhancement of the neutrophil chemotactic activity by proteolytic processing is also observed for GRO (King et al., 2000). A specific isoform of this chemokine was isolated from bone marrow stromal cells stimulated with an immunomodulatory peptide. Surprisingly, a hematopoiesis synergistic factor in this conditioned medium was identified as GRO (5–73). The chemokine variant enhances the proliferation of hematopoietic progenitor cells stimulated by a suboptimal concentration of macrophage colonystimulating factor (King et al., 2000). Furthermore, GRO (5–73) mobilizes hematopoietic stem cells and neutrophils into the peripheral blood and synergizes in this process with granulocyte colony-stimulating factor (King et al., 2001). The truncated GRO isoform selectively binds to CXCR2 with higher affinity than intact GRO and is 10-fold more potent in neutrophil activation assays (King et al., 2000, 2001). In contrast to the previously discussed ELR þ CXC chemokines, different natural isoforms of human granulocyte chemotactic protein-2 (GCP-2), i.e., GCP-2(1–77), GCP-2(3–77), GCP-2(6–77) and GCP-2(9–77) isolated from


19

osteosarcoma cells, are equally active in neutrophil chemotaxis assays (Proost et al., 1993a). GCP-2 possesses a Pro residue in the penultimate NH2-terminal position and is thus a theoretical substrate of dipeptidyl-peptidase IV (DPP IV/CD26) (Proost et al., 1998c). It should be noticed that the truncated GCP-2 isoform, GCP-2(3–77), generated by incubation of GCP-2 with CD26/ DPP IV is present in osteosarcoma cell supernatants (Proost et al., 1993a,b). Additional, still unidentified proteases are probably secreted by the osteosarcoma cells and might work in concert with CD26/DPP IV to generate the shorter GCP-2 isoforms. B. ELR CXC CHEMOKINES Several of the ELR CXC chemokines, i.e., Mig, IP-10, I-TAC and SDF-1, possess a penultimate Pro and hence are theoretical substrates of CD26/DPP IV (Oravecz et al., 1997; Proost et al., 1998c, 2001). All of these ELR CXC chemokines are indeed processed by CD26/DPP IV and contrary to GCP-2, this conversion has impact on their biological activity. Incubation with CD26/ DPP IV generates SDF-1(3–67), that has severely reduced chemotactic activity for lymphocytes and cord blood CD34 þ progenitor cells compared to intact SDF-1 (Ohtsuki et al., 1998; Proost et al., 1998b; Christopherson et al., 2002). Likewise, SDF-1 activity diminishes after incubation with elastase, metalloproteases or cathepsin G that specifically remove three, four or five NH2-terminal residues, respectively (Delgado et al., 2001; McQuibban et al., 2001; Valenzuela-Fernandez et al., 2002). Synthetic SDF-1(3–67) interacts with CXCR4, but has a lower binding affinity compared to intact SDF-1 (Crump et al., 1997). This also explains the impaired capacity of SDF-1(3–67) to inhibit infection of mononuclear cells by X4 HIV-1 strains (Ohtsuki et al., 1998; Proost et al., 1998b). Remarkably, leukocyte-released elastase not only NH2-terminally cleaves SDF-1, but also its receptor CXCR4 (ValenzuelaFernandez et al., 2002). After removal of the NH2-terminal dipeptide, the CXCR3 ligands retain some affinity for their receptor (Proost et al., 2001). Accordingly, the CD26/ DPP IV-treated chemokines can antagonize the attraction of lymphocytes by their intact counterparts. Hence, CD26/DPP IV converts IP-10 and I-TAC into inhibitors of CXCR3-mediated T-lymphocyte chemotaxis. However, CD26/DPP IV-truncated IP-10 and Mig inhibit IL-8-induced angiogenesis equally well as intact IP-10 and Mig. This implicates that the angiostatic potential of IP-10 is either not dependent on CXCR3 or triggered by a different intracellular signal transduction pathway than required for induction of chemotaxis (Proost et al., 2001). Further research is needed to elucidate the precise mechanism mediating angiogenesis or endothelial cell growth inhibition and to prove that CXCR3 is indeed involved. Surprisingly, a preparation of natural IP-10 isolated from osteosarcoma cells did not contain

20

S. STRUYF ET AL.

the CD26/DPP IV-truncated isoform of IP-10, although from the same supernatant CD26/DPP IV-truncated RANTES was recovered (Struyf et al., 1998a). Osteosarcoma cells produced NH2-terminally intact IP-10 and IP-10 missing three, four or five NH2-terminal residues (Proost et al., 1993b). Angiostatic activity is also reported for the first identified chemokine, PF-4 (Maione et al., 1990). However, it is not clear yet whether this ELR CXC chemokine binds to a G protein-coupled receptor to exert its effect on endothelial cells. It is speculated that PF-4 prevents the formation of the ternary signal transduction complex between basic fibroblast growth factor (bFGF), the bFGF receptor and proteoglycans by interacting strongly with the sulfated glycosaminoglycan chains on the proteoglycans (Sato et al., 1990; Yayon et al., 1991; Brandt et al., 2000). It is demonstrated that IP-10 and PF-4 bind to shared specific heparan sulfate sites on the cellular surface of endothelial cells (Luster et al., 1995). The high affinity of chemokines for heparin and heparan sulfates is one of the hallmarks of the chemokine family and resides in their COOH-terminal tail. For PF-4 this COOH-terminal heparin-binding domain might be critical for influencing angiogenesis since synthetic COOH-terminal peptides of PF-4 retain the angiostatic activity (Maione et al., 1990). A PF-4 fragment, lacking 16 NH2-terminal residues, was isolated from leukocyte supernatants and has a 30 to 50-fold enhanced inhibitory activity on endothelial cell growth compared to mature PF-4 (Gupta et al., 1995). The protease generating this isoform is not yet identified. III. CC Chemokines as Protease Substrates

A considerable part of the literature on NH2-terminal variants of CC chemokines reports on the processing of chemokines by the specific serine protease CD26/DPP IV. Indeed, the primary structure of several CC chemokines is characterized by an NH2-terminal penultimate Pro. For other CC chemokine isoforms, the protease that is responsible for the truncation is not yet identified (Table III). Furthermore, altered CC chemokine NH2termini can also result from alternative splicing of the chemokine RNA (Youn et al., 1998; Baird et al., 1999; Tanaka et al., 1999). Finally, several CC chemokines seem to be rather resistant to proteolytic activity. MCPs (e.g., MCP-2) are protected from proteolytic degradation by CD26/ DPP IV through cyclization of the NH2-terminal Gln to a pyroglutamate residue (Van Coillie et al., 1998). A structural study of MCP-2 suggests that this residue stabilizes the MCP-2 dimer formation (Blaszczyk et al., 2000). In the MCP-2 monomer the NH2-terminus is exposed, whereas in the dimer the NH2-terminal extension is buried in the cleft between the two monomers and is not accessible for CD26/DPP IV. The inability of CD26/DPP IV to process recombinant MIP-1 in vitro, although its NH2-terminus theoretically


21

complies with the substrate qualifications of this enzyme (Proost et al., 2000), may also be related to the tendency of this molecule to form dimers. In vivo, cleavage of MIP-1 by CD26/DPP IV might occur since MIP-1 (3–69) was isolated from stimulated peripheral blood lymphocytes (Guan et al., 2002). The CC chemokines that are cleaved by incubation with CD26/DPP IV in vitro are LD78 , RANTES, eotaxin and MDC (Fig. 4). Unexpectedly, two dipeptides are removed consecutively by CD26/DPP IV from the MDC NH2terminus (Proost et al., 1999). The second cleavage after a Gly residue demonstrates that the substrate specificity of CD26/DPP IV (peptides with a penultimate Pro or Ala residue) is less restricted than anticipated and can be influenced by amino acids surrounding the scissile bond. The fact that several CC chemokine variants, which are generated in vitro after incubation with CD26/DPP IV, are isolated from supernatants of cell cultures provide evidence that these isoforms can occur in vivo (Noso et al., 1996, 1998; Pal et al., 1997; Mochizuki et al., 1998; Struyf et al., 1998a; Menten et al., 1999). To measure these isoforms in body fluids or biopsies, antibodies selectively recognizing the shortened variants or full-length proteins only, need to be

FIG. 4. Processing of human chemokines by CD26/DPP IV. Of the 25 known human CC chemokines, nine possess a penultimate Pro residue and are theoretical substrates of CD26/DPP IV. However, the NH2-terminus of MCP-1, MCP-2 and MCP-3 was not truncated after incubation with CD26/DPP IV. In contrast, CD26/DPP IV clipped off the NH2-terminal dipeptide of RANTES, LD78 , MDC, eotaxin, SDF-1, IP-10, Mig, I-TAC and GCP-2. As a consequence of this truncation, the biological activity of these chemokines increased ("), diminished (#) or remained unchanged ( ¼ ).

22

S. STRUYF ET AL.

developed. Alternatively, dosage of NH2-terminally truncated chemokines via mass spectrometry is at present under investigation. The impact of NH2-terminal clipping on the activity of CC chemokines can, analogously to the CXC chemokine substrates, not be predicted (Fig. 4). The eosinophil chemotactic activity of truncated eotaxin is diminished due to reduced binding affinity for its receptor CCR3 (Struyf et al., 1999). Similarly, CD26/DPP IV processing results in decreased CCR3 binding affinity of RANTES and LD78 (Struyf et al., 1998a, 2001a). The latter CC chemokines also recognize CCR1 and CCR5. Whereas the receptor affinity of RANTES decreases for CCR1 and increases for CCR5, LD78 interacts stronger with both receptors after CD26/DPP IV processing (Oravecz et al., 1997; Proost et al., 2000; Struyf et al., 1998a, 2001a). This results in enhanced lymphocyte and monocyte chemotactic potency for LD78 and loss of monocyte chemotactic activity for RANTES, although lymphocytes still migrate toward truncated RANTES (Iwata et al., 1999b; Proost et al., 2000; Struyf et al., 1998a, 2001a). In contrast, CD26/DPP IV processing of MDC reduces its chemotactic potency for lymphocytes and dendritic cells and its interaction with CCR4, but does not affect its monocyte chemotactic activity (Struyf et al., 1998b; Proost et al., 1999). Based on these results, it can be deduced that RANTES and LD78 induce chemotaxis of monocytes and lymphocytes mainly through CCR1 and CCR5, respectively. Furthermore, lymphocytes and dendritic cells are attracted by MDC through CCR4 and monocytes probably express another, unknown MDC receptor (Struyf et al., 1998b). All the CC chemokine substrates of CD26/DPP IV have been investigated as inhibitors of HIV infection. The importance of CCR3 as a coreceptor is a subject for debate as is the capacity of MDC to prevent entry of HIV-1 in CD4 þ cells, but the role of CCR5 as a coreceptor for M-tropic HIV-1 strains is fully established. CD26/DPP IV processing of the CCR5 ligands, LD78 and RANTES, enhances their capacity to protect CD4 þ cells against HIV-1 infection, as could be expected from the resulting increase in binding affinity for CCR5 (Proost et al., 1998c; Schols et al., 1998; Xin et al., 1999; Struyf et al., 2001a). In addition to the CC chemokine isoforms that can be generated in vitro by CD26/DPP IV, other NH2-terminally modified CC chemokines [variants of MCP-1, LD78, LD78 , MCP-2, MCP-3, eotaxin, HCC-1, pulmonary and activation-regulated chemokine (PARC) and MDC] were identified in conditioned media (Table III). These natural truncated isoforms of eotaxin, PARC and MDC could not be purified to homogeneity and were never synthesized or expressed for biological characterization, neither was the protease identified that processed these chemokines (DeVico et al., 1998; Noso et al., 1998; Schutyser et al., 2001; Vulcano et al., 2001). Natural variants of MCP-1 and MCP-2, purified to homogeneity, have reduced agonistic


23

activity (Proost et al., 1998a). Furthermore, truncated MCP-2 and MCP-3 lack monocyte chemotactic activity but can block the function of the corresponding intact chemokines and other CC chemokines (Proost et al., 1998a; McQuibban et al., 2000). Whereas the protease trimming MCP-2 is still unknown, gelatinase A can clip MCP-3 and provide a natural mechanism to dampen inflammation. All four MCPs are substrates for matrix metalloproteases and cleavage results in the generation of CC chemokine receptor antagonists (McQuibban et al., 2002). Negative feedback on chemokine activity can also be achieved by full enzymatic degradation. Hookworms are shown to release metalloproteases that cause rapid proteolysis of eotaxin (Culley et al., 2000). As mentioned above for LD78 , the inflammatory activity of CC chemokines can also be enhanced by limited NH2-terminal truncation (Proost et al., 2000). The plasma chemokine HCC-1 was originally classified as a constitutive chemokine, without any obvious bioactivity (Schulz-Knappe et al., 1996; Richter et al., 2000). Recently a highly active isoform of this chemokine was isolated from human hemofiltrate (Detheux et al., 2000). In contrast to the full-length protein, HCC-1 missing eight NH2-terminal residues efficiently attracts eosinophils, monocytes and lymphocytes through activation of CCR1, CCR3 or CCR5. This HCC-1 variant can be generated by incubation of HCC-1 with urokinase plasminogen activator and plasmin (Vakili et al., 2001). The truncated HCC-1(9–74), with a proline residue in the second NH2-terminal position, is further processed by CD26/DPP IV resulting in decreased activity (Parmentier, personal communication). It can be proposed that contrary to the CXC chemokine subgroup, several CC chemokines are resistant to proteolytic activity. It must be noticed however, that the CC chemokine subgroup has expanded tremendously in recent years, by identification of several new members using bioinformatics. For several of these new chemokines only recombinant protein is available and the NH2-terminus was predicted based on the cDNA or gene sequence. Before these chemokines are purified from a natural source, it cannot be excluded that they exist in different NH2-terminal forms. IV. The CD26-Chemokine Connection

When reviewing the literature on posttranslational chemokine processing (Fig. 4, Tables II and III), one protease in particular catches the eye, i.e., CD26/DPP IV. In the recent years, our laboratory and other groups focused research on this serine protease and its interaction with chemokines. CD26/ DPP IV cleaves off the NH2-terminal dipeptide from peptides and rather small proteins starting with an X-Pro or X-Ala motif. The compliance of many chemokine structures with these substrate qualifications, the purification of natural, truncated chemokines missing the NH2-terminal dipeptide and the

24

S. STRUYF ET AL.

evidence that the NH2-terminus of a chemokine is a principal domain for receptor recognition were major triggers to investigate several chemokine substrates in detail and to determine the consequences for their biological activity. In the next subsection the biology of CD26/DPP IV is summarized, paying attention only to well-established functions of this protease, which are no longer subject of debate, as it is the case for several aspects of its biology. A. BIOLOGY

OF

CD26/DPP IV

Dipeptidyl-peptidase IV (EC 3.4.14.5) was first described in 1966 as an enzymatic activity from rat liver hydrolysing glycyl-prolyl- -naphthylamide (Hopsu-Havu and Glenner, 1966). It is a serine type peptidase that removes dipeptides from the NH2-terminal end of peptide chains if the penultimate residue is proline or alanine, but under certain conditions other amino acids may be accepted (Yaron and Naider, 1993; De Meester et al., 1999; Mentlein, 1999; Lambeir et al., 2001). The human CD26/DPP IV cDNA sequence was published in 1992 and encodes 766 amino acids (Darmoul et al., 1992). Its structure and functional domains were reviewed in detail (von Bonin et al., 1998; De Meester et al., 1999; Mentlein, 1999). Six NH2-terminal residues are located intracellularly and are followed by a hydrophobic transmembrane domain (22 residues) and a large multifunctional extracellular part that also includes the COOH-terminus. The extracellular part contains the catalytic domain and can be released from the cell surface. This soluble form of CD26/ DPP IV is enzymatically fully active and occurs at high levels in seminal fluid, lower amounts are detected in plasma, urine and cerebrospinal fluid (Yaron and Naider, 1993; De Meester et al., 1999). Cell bound CD26/DPP IV is rather ubiquitously expressed on blood cells, fibroblasts, epithelial and endothelial cells and can be detected in placenta, kidney, intestine, prostate, gall bladder, pancreas and liver (Yaron and Naider, 1993; Morimoto and Schlossman, 1998; von Bonin et al., 1998; De Meester et al., 1999; Mentlein, 1999). Furthermore, modulated expression of CD26/DPP IV on malignant hematological or solid tumor cells has been reported and may be of help in cytologic diagnosis (Iwata and Morimoto, 1999a). For instance, loss of CD26/ DPP IV expression occurs during melanoma progression and re-expression of CD26/DPP IV can reverse the malignant melanoma phenotype towards cells with characteristics of normal melanocytes (Wesley et al., 1999). In addition, CD26/DPP IV has been demonstrated on several leukocytes, e.g., dendritic cells, activated B cells, NK cells and T cells (Bu¨hling et al., 1994; De Meester et al., 1999). CD26 was originally characterized as a T-cell subset marker, present on 10–60% of resting T cells (Vanham et al., 1993). Systematically lower CD26/ DPP IV levels were detected on CD8 þ compared to CD4 þ cells (Vanham


25

et al., 1993). Expression of this membrane antigen is strongly upregulated after activation (Fox et al., 1984) and is therefore a suitable marker for activated T cells. T cells expressing high levels of CD26/DPP IV constitute a subpopulation of CD45RO þ memory T cells and produce IL-2 in response to mitogenic or alloantigenic stimulation (De Meester et al., 1999). This unique population of CD4 þ cells is the only one that responds to recall antigens, induces synthesis of IgG in B cells and activates MHC-restricted cytotoxic T cells (Morimoto and Schlossman, 1998). Furthermore, CD26/DPP IV may function in an alternative pathway of T cell activation (von Bonin et al., 1998). The molecular mechanisms underlying the costimulatory and immunomodulatory functions of CD26/DPP IV are not yet fully unraveled (De Meester et al., 1999). Furthermore, CD26/DPP IV fulfils other functions: (1) as adhesion molecule binding cellular matrix proteins like collagen and fibronectin, (2) by interacting with adenosine deaminase on the lymphocyte cell surface and (3) in cellular metabolism and amino acid transport by allowing full breakdown of proteolysis resistant peptides that possess a penultimate proline residue (Yaron and Naider, 1993; De Meester et al., 1999; Mentlein, 1999). CD26/DPP IV also influences the activity of several small peptide hormones, including the glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic peptide (GIP) that influence production, release and activity of insulin and glucagon (Mentlein, 1999). The contribution of CD26/DPP IV to the regulation of blood glucose levels was recently confirmed in mice lacking the CD26/DPP IV gene (Marguet et al., 2000). Therapeutic intervention by the development of CD26/DPP IV-resistant GLP-1 and GIP analogues is considered for treatment of type II diabetes (Mentlein, 1999). The physiological relevance of a most recently described function of this protease, chemokine processing, is still under investigation. These substrates and the protease itself are certainly co-expressed in several pathological conditions. Indeed, large numbers of CD26/DPP IV þ T cells have been detected in peripheral blood and/or inflamed tissues of patients with multiple sclerosis, Graves’ disease, tuberculoid leprosy and rheumatoid arthritis (Morimoto and Schlossman, 1998; Scheel-Toellner et al., 1995). In the latter disease, CD26/DPP IV expression is increased on T cells infiltrating the synovial cavity (Gerli et al., 1996), a body compartment where several chemokines are shown to be present as well (Kunkel et al., 1996). In addition, the chemokine receptors CCR5 and CXCR3 are present on almost all synovial fluid T lymphocytes (Qin et al., 1998), which are of the activated/memory subtype (characterized by the CD45RO marker, that is shown to be co-expressed with CD26) (Torimoto et al., 1991). It can thus be postulated that CCR5 or CXCR3 ligands attract CD26/DPP IV þ T cells to the synovial cavity. Several of these ligands, i.e., RANTES, LD78 , Mig, IP-10 and I-TAC

26

S. STRUYF ET AL.

are substrates of CD26/DPP IV. In this way a negative feedback loop on inflammatory chemokine activity is probably established in vivo. Studies in CD26/DPP IV knockout mice (Marguet et al., 2000) using experimental inflammatory models must confirm this issue. The rapid processing of a number of chemokines by CD26/DPP IV is an additional argument for a role for this protease on the regulation of chemokine activity in vivo (Lambeir et al., 2001). In the next subsection the possible impact of CD26/DPP IV processing on chemokine biology is discussed. B. INFLAMMATION

IS

MODULATED

BY

CD26/DPP IV

At first glance, constitutive chemokines, which are mostly CC chemokines, are apparently resistant to proteolytic activity. Hence, one could deduce that proteases would predominantly affect the inflammatory function of chemokines. This is also relevant, since most of the proteases reported to clip chemokines are only present in their active form in an inflammatory context, as they are released from activated neutrophils or during platelet aggregation (Brandt et al., 2000; Bank and Ansorge, 2001). Examples are given in Tables II and III and include elastase, cathepsin G, proteinase 3, gelatinase B, thrombin and plasmin. The effect of these proteases on chemokine activity is ambiguous even if only ELR þ CXC chemokines are considered. Of the latter group some are degraded, which is suggestive for a negative feedback regulation. Other ELR þ CXC chemokines, in contrast, are selectively and specifically clipped and gain potency after being processed. For instance, IL-8 is a substrate for several proteases and is in vitro a more potent neutrophil chemoattractant in its 72 residue-form compared to the 77 residue-protein (He´bert et al., 1990; Nourshargh et al., 1992). If these isoforms are injected intradermally, they elicited neutrophil influx with equivalent potency, probably because the longer isoform is rapidly converted in vivo (Nourshargh et al., 1992). CD26/DPP IV in particular, does not downregulate inflammation elicited by ELR þ CXC chemokines. The neutrophil attractant potency of GCP-2 is not affected by CD26/DPP IV processing and the other potential ELR þ CD26/DPP IV substrate, GRO , is in its truncated form a more efficient neutrophil activator (Proost et al., 1993a, 1998c; King et al., 2000). In contrast, the lymphocyte chemotactic activity of the ELR CXC chemokines mediated through CXCR3 is downregulated by CD26/DPP IV truncation (Proost et al., 2001). Moreover, after treatment with CD26/DPP IV, IP-10 and I-TAC could inhibit the chemotactic activity of their intact counterparts. This duality of pro-inflammatory versus contra-inflammatory characteristics of CD26/DPP IV is also obvious when considering its CC chemokine substrates (Fig. 4). Whereas, several CC chemokines lose their chemotactic potential for one or


27

more leukocyte subtypes, generating inhibitors of chemotaxis, other chemoattractants gain activity upon truncation. These effects are the result of changes in receptor recognition after chemokine processing, but are also determined by the expression pattern of chemokine receptors on the leukocyte subtype (Fig. 5). The latter is flexible and dependent on microenvironmental factors, like cytokines and LPS, in addition to the genetic background of the individual. For instance, it has been shown that CCR1 is highly expressed on eosinophils of certain individuals together with CCR3, the major chemokine receptor expressed on eosinophils (Sabroe et al., 1999). The CCR3 ligands that are processed by CD26/DPP IV still bind to this receptor when truncated but lose their chemotactic activity (Struyf et al., 1998a, 1999, 2001a). The potency of the CCR1 ligand LD78 however, is enhanced by CD26/DPP IV treatment (Proost et al., 2000). The impact of CD26/DPP IV on eosinophil chemotaxis will thus depend on whether CCR1 is highly expressed or not. When CCR1 is only marginally expressed, CD26/DPP IV would downregulate eosinophil influx and provide a

FIG. 5. The microenvironment influences the expression of CD26/DPP IV and chemokines and hence regulates the outcome of inflammation. Inflammation in response to infection or injury is commenced and down modulated by factors produced in the microenvironment. Proteases, such as CD26/DPP IV can influence the activity of released soluble mediators (e.g., chemokines) by posttranslational modification. The balance between chemokine production and chemokine (in)activation, e.g., by CD26/DPP IV, determines whether the inflammatory process is enhanced or dampened.

28

S. STRUYF ET AL.

negative feedback on allergic inflammation. This inhibitory effect would be counteracted in individuals expressing high levels of CCR1. Although it is difficult to make a general conclusion on the clinical outcome of chemokine processing by CD26/DPP IV, it is certain that the truncation has impact on the inflammatory properties of chemokines. Depending on the genetic background, the body compartment, the (pro-)inflammatory agents, the chemokines and leukocytes involved, CD26/DPP IV activity will enhance or downregulate inflammation (Fig. 5). C. THE ANTI-HIV-ACTIVITY CD26/DPP IV

OF

CHEMOKINES

IS

MODIFIED

BY

A direct relation between CD26/DPP IV and HIV infection has been incorrectly proposed in a context not linked to chemokines (Broder et al., 1994). It was suggested that CD26/DPP IV functions as a coreceptor for HIV1 infection and interacts with the viral Tat protein. However, CD26/DPP IV has impact on HIV-1 infection indirectly, through processing of chemokines and altering their receptor affinities (Fig. 6). It is generally accepted that chemokine receptors (CXCR4 and CCR5) are essential coreceptors for HIV-1 to enter CD4 þ cells (Kinter et al., 2000). After binding of a chemokine to its receptor the latter is internalized and unavailable for the virus. Whereas CD26/DPP IV-truncated LD78 and RANTES interact stronger with CCR5 (coreceptor for M-tropic or R5 HIV-1 strains) and are more potent as antiretroviral agent than the full-length proteins, processed SDF-1 (CXCR4 ligand, blocks T-tropic or X4 HIV-1 strains) is less effective in the protection of CD4 þ cells against HIV-1 infection (Ohtsuki et al., 1998; Proost et al., 1998b,c; Struyf et al., 2001a). Hence, CD26/DPP IV processing of these chemokines during early stages of HIV-1 infection, when R5 strains prevail, is beneficial for protection against HIV-1. When the disease progresses and X4 strains predominate, CD26/DPP IV enfeebles the anti-HIV-1 chemokine defense. The role of CD26/DPP IV in the control of HIV-1 infection is therefore dependent on the disease stage. Evidence is accumulating to show that chemokine processing by CD26/ DPP IV and the effect thereof on HIV infection is a relevant issue. Both CD26/DPP IV and CCR5 are considered as markers of CD4 þ TH1 cells (Willheim et al., 1997; Annunziato et al., 1998). Furthermore, co-expression of CCR5 and CD26/DPP IV has also been shown on activated memory T cells (Bleul et al., 1997). CXCR4 and CD26/DPP IV co-localize on the cellular surface and are internalized in the same vesicles after cellular activation with SDF-1 (Herrera et al., 2001). Thus, the HIV-1 coreceptors and the enzyme processing the coreceptor ligands are situated close to each other on the cell membrane. With regard to the rate of conversion, a half-life of less than 1 min


29

FIG. 6. The role of CD26/DPP IV in HIV infection through the processing of chemokines. The HIV-1 inhibitory capacity of the chemokines RANTES, LD78 and SDF-1 is influenced by the proteolytic activity of CD26/DPP IV. RANTES and LD78 become more efficient antiviral agents against M-tropic HIV-1 strains that are present during the asymptomatic disease phase, whereas the inhibitory activity of SDF-1 is reduced after processing by CD26/DPP IV. The latter chemokine is important as the disease progresses and T-tropic HIV-1 strains become the predominant viruses. In contrast to the M-tropic or NSI (non-syncytium-inducing) strains, T-tropic HIV-1 strains have a distinct phenotype, as they can induce formation of syncytia in infected cell cultures and are therefore also referred to as SI (syncytium-inducing) HIV-1 strains. The number of CD26/DPP IVhigh T cells in circulation is lower at the end than in the beginning of the infection. Hence, the disease-promoting role of CD26/DPP IV through inactivation of SDF-1 is expected to be rather limited and the protecting effect through RANTES and LD78 processing might be more important.

is expected for SDF-1 in the blood circulation (Lambeir et al., 2001). Processing of chemokines by cell-bound CD26/DPP IV is realistic as well. An increased rate of HIV entry in CD26/DPP IVbright CD4 þ /CXCR4 þ T cells has been observed (Callebaut et al., 1998; Shioda et al., 1998). Furthermore, SDF-1-induced kinase activation in T cells was prolonged when degradation of SDF-1 by CD26/DPP IV was blocked (Tilton et al., 2000) and CXCR4 was

30

S. STRUYF ET AL.

not downregulated on a CD26/DPP IV þ T cell line after incubation with SDF-1 (Shioda et al., 1998). It is reported that the percentage and absolute number of CD26/DPP IV þ cells in HIV patients decrease as the disease progresses (Blazquez et al., 1992; Vanham et al., 1993) and hence less CD26/DPP IV would be available for processing and inactivation of SDF-1. All these data indicate that not only the presence or absence of chemokines merits consideration while studying HIV-1, their biological activity is largely dependent on the presence of proteases like CD26/DPP IV. D. POSSIBLE ROLE OF CD26/DPP IV HEMATOPOIESIS

IN

ANGIOGENESIS

AND

As mentioned above, most constitutive chemokines were only described in recent years and some natural proteins are not yet characterized. It is hence difficult to conclude that these chemokines, regulating normal trafficking of blood cells between immunological compartments, are not enzymatically processed. For a conclusive evaluation of the impact of posttranslational modifications on the angiogenic and hematopoietic activities of chemokines we have to await more data. For angiogenesis, it can be speculated that CD26/DPP IV processing of chemokines does not alter the capacity of the ELR CXC chemokines (Mig, IP-10, I-TAC) to hinder development of new blood vessels. It has been demonstrated that the truncated isoforms generated by CD26/DPP IV still interact with CXCR3, albeit with reduced efficiency, and that IP-10(3–77) is as potent in inhibiting IL-8-induced angiogenesis as intact IP-10(1–77) (Proost et al., 2001). With regard to the ELR þ CXC chemokines that enhance angiogenesis through interaction with CXCR2 (Addison et al., 2000; Belperio et al., 2000), some members of this group are substrates of CD26/DPP IV (Table II). Truncation by CD26/DPP IV has no impact on GCP-2-induced neutrophil chemotaxis (Proost et al., 1993a, 1998c), whereas the neutrophilactivating properties and CXCR2-binding affinity of truncated GRO increases (King et al., 2001). It is not known whether CD26/DPP IV processing selectively affects the affinity of GCP-2 for CXCR2. Neither have these truncated CXC chemokines been evaluated in angiogenesis assays, although their angiogenic properties are expected to be unaffected or enhanced by CD26/DPP IV. Other proteases cleaving ELR þ CXC chemokines have been described (Table II) but these reports focused on the impact of processing on the neutrophil-activating properties of the chemokines. For IL-8 in particular, it can be expected that limited processing by gelatinase B enhances its angiogenic properties, since the CXCR2 binding affinity of IL-8(7–77) is about five-fold higher, compared to that of intact IL-8 (1–77) (Van den Steen et al., 2000).


31

Human CXCR2 might also be the dominant receptor for CXC chemokineinduced stem cell mobilization (King et al., 2001). Recently a truncated form of GRO was isolated, which only binds to CXCR2 and mobilizes hematopoietic stem cells into the circulation after subcutaneous injection. Similar activity has been demonstrated for IL-8 (Laterveer et al., 1995) and MIP-1 (Hunter et al., 1995). Furthermore, chemokines not only regulate the migration of immature and mature blood cells, they also influence the production of these cells at the level of hematopoietic stem/progenitor cells (Youn et al., 2000). More than 20 chemokines, belonging to the C, CC, CXC and CX3C subfamilies have been shown to possess suppressor activity on the proliferation of immature, multipotential progenitor cells. On more mature subsets of progenitors that responded to stimulation by a single growth factor, chemokines had rather stimulatory effects on colony formation. Finally, the chemokine MIP-1 is important as survival factor and an essential component of long-term bone marrow cultures (Verfaillie et al., 1994). This latter chemokine was also the first chemokine demonstrated to influence blood cell progenitor growth (Graham et al., 1990). After injection of MIP-1 in mice, suppression rather than stimulation of hematopoiesis is observed. This is expected to be a useful characteristic to protect bone marrow cells during chemotherapy and further evaluation of the application of a single chemokine or a combination of different chemokines for this purpose might be worthwhile (Broxmeyer and Kim, 1999b). Whereas it is clarified which chemokines might influence hematopoiesis, the receptors and signal transduction pathways involved are largely unknown. Nevertheless, some data with chemokine receptor knockouts are available. The murine IL-8 receptor is required for the murine GRO proteins to exert their inhibitory effect on progenitor proliferation in mice (Broxmeyer et al., 1996). The suppressive effect of MIP-1 is not dependent on CCR1, but this receptor mediates the stimulatory and mobilizing activity of MIP-1 (Broxmeyer et al., 1999a). Since the affinity of MIP-1/LD78 for CCR1 changes dramatically by CD26/DPP IV-mediated NH2-terminal truncation (Struyf et al., 2001a), chemokine-processing proteases can have an important impact on hematopoiesis. Whether posttranslational modification of chemokines is a regulatory mechanism in the formation and development of blood cells is still an unanswered question. For SDF-1 it is clear that its production and processing must be tightly controlled during fetal life, because NH2-terminal truncation by CD26/DPP IV reduces the affinity of this chemokine for its receptor CXCR4 as both the ligand and its receptor are indispensable during embryogenesis. Mice missing either SDF-1 or CXCR4 die perinatally (Nagasawa et al., 1996; Ma et al., 1998). It can thus be expected that constitutive chemokines are more resistant to NH2-terminal processing.

32

S. STRUYF ET AL.

ACKNOWLEDGMENTS This work was supported by the Fund for Scientific Research of Flanders (FWO-Vlaanderen), the Concerted Research Actions of the Regional Government of Flanders, the InterUniversity Attraction Pole initiative (IUAP) of the Belgian Federal Government and the Quality of Life Program of the European Community. S.S. and P.P. are senior research assistants of the FWOVlaanderen. The authors wish to thank Prof. A. Billiau, Prof. B. Goddeeris, Prof. G. Opdenakker, Prof. D. Schols and Prof. J. Vanderleyden from the University of Leuven and Prof. J. Van Snick (Ludwig Institute, Brussels) for critically reading this manuscript.

REFERENCES Addison, C. L., Daniel, T. O., Burdick, M. D., Liu, H., Ehlert, J. E., Xue, Y. Y., Buechi, L., Walz, A., Richmond, A., and Strieter, R. M. (2000). The CXC chemokine receptor 2, CXCR2, is the putative receptor for ELR þ CXC chemokine-induced angiogenic activity. J. Immunol. 165, 5269–5277. Annunziato, F., Cosmi, L., Galli, G., Beltrame, C., Romagnani, P., Manetti, R., Romagnani, S., and Maggi, E. (1999). Assessment of chemokine receptor expression by human Th1 and Th2 cells in vitro and in vivo. J. Leukoc. Biol. 65, 691–699. Annunziato, F., Galli, G., Cosmi, L., Romagnani, P., Manetti, R., Maggi, E., and Romagnani, S. (1998). Molecules associated with human Th1 or Th2 cells. Eur. Cytokine. Netw. 9(Suppl no. 3), 12–16. Baird, J. W., Nibbs, R. J., Komai-Koma, M., Connolly, J. A., Ottersbach, K., Clark-Lewis, I., Liew, F. Y., and Graham, G. J. (1999). ESkine, a novel beta-chemokine, is differentially spliced to produce secretable and nuclear targeted isoforms. J. Biol. Chem. 274, 33496–33503. Balter, M. (1996). New hope in HIV disease. Science 274, 1988–1989. Bank, U., and Ansorge, S. (2001). More than destructive: neutrophil-derived serine proteases in cytokine bioactivity control. J. Leukoc. Biol. 69, 197–206. Begg, G. S., Pepper, D. S., Chesterman, C. N., and Morgan, F. J. (1978). Complete covalent structure of human beta-thromboglobulin. Biochemistry 17, 1739–1744. Belperio, J. A., Keane, M. P., Arenberg, D. A., Addison, C. L., Ehlert, J. E., Burdick, M. D., and Strieter, R. M. (2000). CXC chemokines in angiogenesis. J. Leukoc. Biol. 68, 1–8. Blaszczyk, J., Van Coillie, E., Proost, P., Van Damme, J., Opdenakker, G., Bujacz, G. D., Wang, J. M., and Ji, X. (2000). Complete crystal structure of monocyte chemotactic protein-2, a CC chemokine that interacts with multiple receptors. Biochemistry 39, 14075–14081. Blazquez, M. V., Maduenõ, J. A., Gonzalez, R., Jurado, R., Bachovchin, W. W., Penã, J., and Munõz, E. (1992). Selective decrease of CD26 expression in T cells from HIV-1-infected individuals. J. Immunol. 149, 3073–3077. Bleul, C. C., Farzan, M., Choe, H., Parolin, C., Clark-Lewis, I., Sodroski, J., and Springer, T. A. (1996). The lymphocyte chemoattractant SDF-1 is a ligand for LESTR/fusin and blocks HIV-1 entry. Nature 382, 829–833. Bleul, C. C., Wu, L., Hoxie, J. A., Springer, T. A., and Mackay, C. R. (1997). The HIV coreceptors CXCR4 and CCR5 are differentially expressed and regulated on human T lymphocytes. Proc. Natl. Acad. Sci. USA 94, 1925–1930. Bo¨hm, S. K., Grady, E. F., and Bunnett, N. W. (1997). Regulatory mechanisms that modulate signalling by G-protein-coupled receptors. Biochem. J. 322, 1–18. Brandt, E., Van Damme, J., and Flad, H.-D. (1991). Neutrophils can generate their activator neutrophil-activating peptide 2 by proteolytic cleavage of platelet-derived connective tissueactivating peptide III. Cytokine 3, 311–321.


33

Brandt, E., Petersen, F., Ludwig, A., Ehlert, J. E., Bock, L., and Flad, H. D. (2000). The betathromboglobulins and platelet factor 4: blood platelet-derived CXC chemokines with divergent roles in early neutrophil regulation. J. Leukoc. Biol. 67, 471–478. Broder, C. C., Nussbaum, O., Gutheil, W. G., Bachovchin, W. W., Berger, E. A., Patience, C., McKnight, A., Clapham, P. R., Boyd, M. T., Weiss, R. A., Schulz, T. F., Camerini, D., Planelles, V., Chen, I. S., Alizon, M., Dragic, T., Callebaut, C., Jacotot, E., Krust, B., and Hovanessian, A. G. (1994). CD26 antigen and HIV fusion? Science 264, 1156–1165. Broxmeyer, H. E., Cooper, S., Cacalano, G., Hague, N. L., Bailish, E., and Moore, M. W. (1996). Involvement of Interleukin (IL) 8 receptor in negative regulation of myeloid progenitor cells in vivo: evidence from mice lacking the murine IL-8 receptor homologue. J. Exp. Med. 184, 1825–1832. Broxmeyer, H. E., Cooper, S., Hangoc, G., Gao, J.-L., and Murphy, P. M. (1999a). Dominant myelopoietic effector functions mediated by chemokine receptor CCR1. J. Exp. Med. 189, 1987–1992. Broxmeyer, H. E., and Kim, C. H. (1999b). Regulation of hematopoiesis in a sea of chemokine family members with a plethora of redundant activities. Exp. Hematol. 27, 1113–1123. Bu¨hling, F., Kunz, D., Reinhold, D., Ulmer, A. J., Ernst, M., Flad, H.-D., and Ansorge, S. (1994). Expression and functional role of dipeptidyl peptidase IV (CD26) on human natural killer cells. Nat. Immun. 13, 270–279. Callebaut, C., Jacotot, E., Blanco, J., Krust, B., and Hovanessian, A. G. (1998). Increased rate of HIV-1 entry and its cytopathic effect in CD4 þ /CXCR4 þ T cells expressing relatively high levels of CD26. Exp. Cell Res. 241, 352–362. Campbell, J. J., Foxman, E. F., and Butcher, E. C. (1997). Chemoattractant receptor cross talk as a regulatory mechanism in leukocyte adhesion and migration. Eur. J. Immunol. 27, 2571–2578. Campbell, J. J., Pan, J., and Butcher, E. C. (1999). Cutting edge: developmental switches in chemokine responses during T cell maturation. J. Immunol. 163, 2353–2357. Car, B. D., Baggiolini, M., and Walz, A. (1991). Formation of neutrophil-activating peptide 2 from platelet-derived connective-tissue-activating peptide III by different tissue proteinases. Biochem. J. 275, 4. Carroll, R. G., Riley, J. L., Levine, B. L., Feng, Y., Kaushal, S., Ritchey, D. W., Bernstein, W., Weislow, O. S., Brown, C. R., Berger, E. A., June, C. H., and St.Louis, D. C. (1997). Differential regulation of HIV-1 fusion cofactor expression by CD28 costimulation of CD4 þ T cells. Science 276, 273–276. Chan, D. C., and Kim, P. S. (1998). HIV entry and its inhibition. Cell 93, 681–684. Charbonnier, A.-S., Kohrgruber, N., Kriehuber, E., Stingl, G., Rot, A., and Maurer, D. (1999). Macrophage inflammatory protein 3alpha is involved in the constitutive trafficking of epidermal langerhans cells. J. Exp. Med. 190, 1755–1768. Choe, H., Farzan, M., Sun, Y., Sullivan, N., Rollins, B., Ponath, P. D., Wu, L., Mackay, C. R., LaRosa, G., Newman, W., Gerard, N., Gerard, C., and Sodroski, J. (1996). The -chemokine receptors CCR3 and CCR5 facilitate infection by primary HIV-1 isolates. Cell 85, 1135–1148. Chou, C.-C., Fine, J. S., Puglies-Sivo, C., Gonsiorek, W., Davies, L., Deno, G., Petro, M., Schwarz, M., Zavodny, P. J., and Hipkin, R. W. (2002). Pharmacological characterization of the chemokine receptor hCCR1 in a stable transfectant and differentiated HL-60 cells: antagonism of hCCR1 activation by MIP-1 . Br. J. Pharmacol. 137, 663–675. Christopherson, K. W. II, Hangoc, G., and Broxmeyer, H. E. (2002). Cell surface peptidase CD26/ dipeptidylpeptidase IV regulates CXCL12/stromal cell-derived factor-1-mediated chemotaxis of human cord blood CD34 þ progenitor cells. J. Immunol. 169, 7000–7008. Clark-Lewis, I., Dewald, B., Loetscher, M., Moser, B., and Baggiolini, M. (1994). Structural requirements for interleukin-8 function identified by design of analogs and CXC chemokine hybrids. J. Biol. Chem. 269, 16075–16081.

34

S. STRUYF ET AL.

Cocchi, F., DeVico, A. L., Garzino-Demo, A., Arya, S. K., Gallo, R. C., and Lusso, P. (1995). Identification of RANTES, MIP-1 alpha, and MIP-1 beta as the major HIV-suppressive factors produced by CD8 þ T cells. Science 270, 1811–1815. Cocchi, F., DeVico, A. L., Garzino-Demo, A., Cara, A., Gallo, R. C., and Lusso, P. (1996). The V3 domain of the HIV-1 gp120 envelope glycoprotein is critical for chemokine-mediated blockade of infection. Nat. Med. 2, 1244–1247. Cohen, A. B., Stevens, M. D., Miller, E. J., Atkinson, M. A., and Mullenbach, G. (1992). Generation of the neutrophil-activating peptide-2 by cathepsin G and cathepsin G-treated human platelets. Am. J. Physiol. 263, 249–256. Crump, M. P., Gong, J.-H., Loetscher, P., Rajarathnam, K., Amara, A., Arenzana-Seisdedos, F., Virelizier, J.-L., Baggiolini, M., Sykes, B. D., and Clark-Lewis, I. (1997). Solution structure and basis for functional activity of stromal cell-derived factor-1; dissociation of CXCR4 activation from binding and inhibition of HIV-1. EMBO J. 16, 6996–7007. Culley, F. J., Brown, A., Conroy, D. M., Sabroe, I., Pritchard, D. I., and Williams, T. J. (2000). Eotaxin is specifically cleaved by hookworm metalloproteases preventing its action in vitro and in vivo. J. Immunol. 165, 6447–6453. D’Amico, G., Frascaroli, G., Bianchi, G., Transidico, P., Doni, A., Vecchi, A., Sozzani, S., Allavena, P., and Mantovani, A. (2000). Uncoupling of inflammatory chemokine receptors by IL-10: generation of functional decoys. Nat. Immunol. 1, 387–391. Darmoul, D., Lacasa, M., Baricault, L., Marguet, D., Sapin, C., Trotot, P., Barbat, A., and Trugnan, G. (1992). Dipeptidyl peptidase IV (CD26) gene expression in enterocyte-like colon cancer cell lines HT-29 and Caco-2. Cloning of the complete human coding sequence and changes of dipeptidyl peptidase IV mRNA levels during cell differentiation. J. Biol. Chem. 267, 4824–4833. David, N. B., Sape`de, D., Saint-Etienne, L., Thisse, C., Thisse, B., Dambly-Chaudie`re, C., Rosa, F. M., and Ghysen, A. (2002). Molecular basis of cell migration in the fish lateral line: role of the chemokine receptor CXCR4 and of its ligand, SDF-1. Proc. Natl. Acad. Sci. USA 99, 16297–16302. De Meester, I., Korom, S., Van Damme, J., and Scharpe´, S. (1999). CD26, let it cut or cut it down. Immunol. Today 20, 367–375. Delgado, M. B., Clark-Lewis, I., Loetscher, P., Langen, H., Thelen, M., Baggiolini, M., and Wolf, M. (2001). Rapid inactivation of stromal cell-derived factor-1 by cathepsin G associated with lymphocytes. Eur. J. Immunol. 31, 699–707. Deng, H., Liu, R., Ellmeier, W., Choe, S., Unutmaz, D., Burkhart, M., Di Marzio, P., Marmon, S., Sutton, R. E., Hill, C. M., Davis, C. B., Peiper, S. C., Schall, T. J., Littman, D. R., and Landau, N. R. (1996). Identification of a major co-receptor for primary isolates of HIV-1. Nature 381, 661–666. Deng, X., Ueda, H., Su, S. B., Gong, W., Dunlop, N. M., Gao, J. L., Murphy, P. M., and Wang, J. M. (1999). A synthetic peptide derived from human immunodeficiency virus type 1 gp120 downregulates the expression and function of chemokine receptors CCR5 and CXCR4 in monocytes by activating the 7-transmembrane G-protein-coupled receptor FPRL1/LXA4R. Blood 94, 1165–1173. Detheux, M., Sta¨ndker, L., Vakili, J., Mu¨nch, J., Forssmann, U., Adermann, K., Po¨hlmann, S., Vassart, G., Kirchhoff, F., Parmentier, M., and Forssmann, W.-G. (2000). Natural proteolytic processing of hemofiltrate CC chemokine 1 generates a potent CC chemokine receptor (CCR)1 and CCR5 agonist with anti-HIV properties. J. Exp. Med. 192, 1501–1508. DeVico, A. L., Pal, R., and Markham, P. D. (1998). -chemokine MDC and HIV-1 infection. Science 281, 487. Dieu-Nosjean, M.-C., Vicari, A., Lebecque, S., and Caux, C. (1999). Regulation of dendritic cell trafficking: a process that involves the participation of selective chemokines. J. Leukoc. Biol. 66, 252–262.


35

Dieu, M.-C., Vanbervliet, B., Vicari, A., Bridon, J.-M., Oldham, E., Aı¨t-Yahia, S., Brie`re, F., Zlotnik, A., Lebecque, S., and Caux, C. (1998). Selective recruitment of immature and mature dendritic cells by distinct chemokines expressed in different anatomic sites. J. Exp. Med. 188, 373–386. Doms, R. W., and Trono, D. (2000). The plasma membrane as a combat zone in the HIV battlefield. Genes Dev. 14, 2677–2688. Doranz, B. J., Rucker, J., Yi, Y., Smyth, R. J., Samson, M., Peiper, S. C., Parmentier, M., Collman, R. G., and Doms, R. W. (1996). A dual-tropic primary HIV-1 isolate that uses fusin and the beta-chemokine receptors CKR-5, CKR-3, and CKR-2b as fusion cofactors. Cell 85, 1149–1158. Dragic, T., Litwin, V., Allaway, G. P., Martin, S. R., Huang, Y., Nagashima, K. A., Cayanan, C., Maddon, P. J., Koup, R. A., Moore, J. P., and Paxton, W. A. (1996). HIV-1 entry into CD4 þ cells is mediated by the chemokine receptor CC-CKR-5. Nature 381, 667–673. Fox, D. A., Hussey, R. E., Fitzgerald, K. A., Acuto, O., Poole, C., Palley, L., Daley, J. F., Schlossman, S. F., and Reinherz, E. L. (1984). Ta1, a novel 105 KD human T cell activation antigen defined by a monoclonal antibody. J. Immunol. 133, 1250–1256. Fo¨rster, R., Mattis, A. E., Kremmer, E., Wolf, E., Brem, G., and Lipp, M. (1996). A putative chemokine receptor, BLR1, directs B cell migration to defined lymphoid organs and specific anatomic compartments of the spleen. Cell 87, 1037–1047. Fra, A.M., Locati, M., Otero, K., Sironi, M., Signorelli, P., Massardi, M.L., Gobbi, M., Vecchi, A., Sozzani, S., and Mantovani, A. (2003). Scavenging of inflammatory CC chemokines by the promiscuous putatively silent chemokine receptor D6. J. Immunol. 170, 2279–2282. Galli, G., Chantry, D., Annunziato, F., Romagnani, P., Cosmi, L., Lazzeri, E., Manetti, R., Maggi, E., Gray, P. W., and Romagnani, S. (2000). Macrophage-derived chemokine production by activated human T cells in vitro and in vivo: preferential association with the production of type 2 cytokines. Eur. J. Immunol. 30, 204–210. Gerard, C., and Rollins, B. J. (2001). Chemokines and disease. Nat. Immunol. 2, 108–115. Gerli, R., Muscat, C., Bertotto, A., Bistoni, O., Agea, E., Tognellini, R., Fiorucci, G., Cesarotti, M., and Bombardieri, S. (1996). CD26 surface molecule involvement in T cell activation and lymphokine synthesis in rheumatoid and other inflammatory synovitis. Clin. Immunol. Immunopathol. 80, 31–37. Gimbrone, M. A., Obin, M. S., Brock, A. F., Luis, E. A., Hass, P. E., He´bert, C. A., Yip, Y. K., Leung, D. W., Lowe, D. G., Kohr, W. J., Darbonne, W. C., Bechtol, K. B., and Baker, J. B. (1989). Endothelial interleukin-8: a novel inhibitor of leukocyte-endothelial interactions. Science 246, 1601–1603. Graham, G. J., Wright, E. G., Hewick, R., Wolpe, S. D., Wilkie, N. M., Donaldson, D., Lorimore, S., and Pragnell, I. B. (1990). Identification and characterization of an inhibitor of haemopoietic stem cell proliferation. Nature 344, 442–444. Gregory, H., Young, J., Schro¨der, J.-M., Mrowietz, U., and Christophers, E. (1988). Structure determination of a human lymphocyte derived neutrophil activating peptide (LYNAP). Biochem. Biophys. Res. Commun. 151, 883–890. Guan, E., Wang, J., Roderiquez, G., and Norcross, M. A. (2002). Natural truncation of the chemokine MIP-1 /CCL4 affects receptor specificity but not anti-HIV-1 activity. J. Biol. Chem. 277, 32348–32352. Gunn, M. D., Ngo, V. N., Ansel, K. M., Ekland, E. H., Cyster, J. G., and Williams, L. T. (1998). A B-cell-homing chemokine made in lymphoid follicles activates Burkitt’s lymphoma receptor1. Nature 391, 799–803. Gupta, S. K., Hassel, T., and Singh, J. P. (1995). A potent inhibitor of endothelial cell proliferation is generated by proteolytic cleavage of the chemokine platelet factor 4. Proc. Natl. Acad. Sci. USA 92, 7799–7803.

36

S. STRUYF ET AL.

Hasegawa, H., Nomura, T., Kohno, M., Tateishi, N., Suzuki, Y., Maeda, N., Fujisawa, R., Yoshie, O., and Fujita, S. (2000). Increased chemokine receptor CCR7/EBI1 expression enhances the infiltration of lymphoid organs by adult T-cell leukemia cells. Blood 95, 30–38. Herrera, C., Morimoto, C., Blanco, J., Mallol, J., Arenzana, F., Lluis, C., and Franco, R. (2001). Comodulation of CXCR4 and CD26 in human lymphocytes. J. Biol. Chem. 276, 19532–19539. He´bert, C. A., Luscinskas, F. W., Kiely, J.-M., Luis, E. A., Darbonne, W. C., Bennett, G. L., Liu, C. C., Obin, M. S., Gimbrone, M. A., and Baker, J. B. (1990). Endothelial and leukocyte forms of IL-8. Conversion by thrombin and interactions with neutrophils. J. Immunol. 145, 3033–3040. Hoffman, T. L., LaBranche, C. C., Zhang, W., Canziani, G., Robinson, J., Chaiken, I., Hoxie, J. A., and Doms, R. W. (1999). Stable exposure of the coreceptor-binding site in a CD4-independent HIV-1 envelope protein. Proc. Natl. Acad. Sci. USA 96, 6359–6364. Holt, J. C., Harris, M. E., Holt, A. M., Lange, E., Henschen, A., and Niewiarowski, S. (1986). Characterization of human platelet basic protein, a precursor form of low-affinity platelet factor 4 and beta-thromboglobulin. Biochemistry 25, 1988–1996. Hopsu-Havu, V. K., and Glenner, G. G. (1966). A new dipeptide naphthylamidase hydrolyzing glycyl-prolyl- -naphthylamide. Histochemie 7, 197–201. Horuk, R. (1994). The Interleukin-8-receptor family: from chemokines to malaria. Immunol. Today 15, 169–174. Hunter, M. G., Bawden, L., Brotherton, D., Craig, S., Cribbes, S., Czaplewski, L. G., Dexter, T. M., Drummond, A. H., Gearing, A. H., Heyworth, C. M., Lord, B. I., McCourt, M., Varley, P. G., Wood, L. M., Edwards, R. M., and Lewis, P. J. (1995). BB-10010: an active variant of human macrophage inflammatory protein-1 with improved pharmaceutical properties. Blood 86, 4400–4408. Iida, N., Haisa, M., Igarashi, A., Pencev, D., and Grotendorst, G. R. (1996). Leukocyte-derived growth factor links the PDGF and CXC chemokine families of peptides. FASEB J. 10, 1336–1345. Iwata, S., and Morimoto, C. (1999a). CD26/dipeptidyl peptidase IV in context: the different roles of a multifunctional ectoenzyme in malignant transformation. J. Exp. Med. 190, 301–305. Iwata, S., Yamaguchi, N., Munakata, Y., Ikushima, H., Lee, J. F., Hosono, O., Schlossman, S. F., and Morimoto, C. (1999b). CD26/dipeptidyl peptidase IV differentially regulates the chemotaxis of T cells and monocytes toward RANTES: possible mechanism for the switch from innate to acquired immune response. Int. Immunol. 11, 417–426. Kanayama, N., Kajiwara, Y., Goto, J., El Maradny, E., Maehara, K., Andou, K., and Terao, T. (1995). Inactivation of interleukin-8 by aminopeptidase N (CD13). J. Leukoc. Biol. 57, 129–134. King, A. G., Johanson, K., Frey, C. L., DeMarsh, P. L., White, J. R., McDevitt, P., McNulty, D., Balcarek, J., Jonak, Z. L., Bhatnagar, P. K., and Pelus, L. M. (2000). Identification of unique truncated KC/GRO beta chemokines with potent hematopoietic and anti-infective activities. J. Immunol. 164, 3774–3782. King, A. G., Horowitz, D., Dillon, S. B., Levin, R., Farese, A. M., MacVittie, T. J., and Pelus, L. M. (2001). Rapid mobilization of murine hematopoietic stem cells with enhanced engraftment properties and evaluation of hematopoietic progenitor cell mobilization in rhesus monkeys by a single injection of SB-251353, a specific truncated form of the human CXC chemokine GRO . Blood 97, 1534–1542. Kinter, A., Arthos, J., Cicala, C., and Fauci, A. S. (2000). Chemokines, cytokines and HIV: a complex network of interactions that influence HIV pathogenesis. Immunol. Rev. 177, 88–98. Knaut, H., Werz, C., Geisler, R., The Tu¨bingen 2000 screen consortium, and Nu¨sslein-Volhard, C. (2003). A zebrafish homologue of the chemokine receptor CXCR4 is a germ-cell guidance receptor. Nature 421, 279–282. Kuhmann, S. E., Platt, E. J., Kozak, S. L., and Kabat, D. (2000). Cooperation of multiple CCR5 coreceptors is required for infections by human immunodeficiency virus type 1. J. Virol. 74, 7005–7015.


37

Kunkel, S. L., Lukacs, N., Kasama, T., and Strieter, R. M. (1996). The role of chemokines in inflammatory joint disease. J. Leukoc. Biol. 59, 6–12. Lai, J.-P., Ho, W.-Z., Zhan, G.-X., Yi, Y., Collman, R. G., and Douglas, S. D. (2001). Substance P antagonist (CP-96,345) inhibits HIV-1 replication in human mononuclear phagocytes. Proc. Natl. Acad. Sci. USA 98, 3970–3975. Lambeir, A.-M., Proost, P., Durinx, C., Bal, G., Senten, K., Augustyns, K., Scharpe´, S., Van Damme, J., and De Meester, I. (2001). Kinetic investigation of chemokine truncation by CD26/ dipeptidyl peptidase IV reveals a striking selectivity within the chemokine family. J. Biol. Chem. 276, 29839–29845. Laterveer, L., Lindley, I. J., Hamilton, M. S., Willemze, R., and Fibbe, W. E. (1995). Interleukin-8 induces rapid mobilization of hematopoietic stem cells with radioprotective capacity and longterm myelolymphoid repopulating ability. Blood 85, 2269–2275. Lee, B., and Montaner, L. J. (1999). Chemokine immunobiology in HIV-1 pathogenesis. J. Leukoc. Biol. 65, 552–565. Lindley, I., Aschauer, H., Seifert, J.-M., Lam, C., Brunowsky, W., Kownatzki, E., Thelen, M., Peveri, P., Dewald, B., von Tscharner, V., Walz, A., and Baggiolini, M. (1988). Synthesis and expression in Escherichia coli of the gene encoding monocyte-derived neutrophil-activating factor: biological equivalence between natural and recombinant neutrophil-activating factor. Proc. Natl. Acad. Sci. USA 85, 9199–9203. Littman, D. R. (1998). Chemokine receptors: keys to AIDS pathogenesis?. Cell 93, 677–680. Locati, M., Otero, K., Schioppa, T., Signorelli, P., Perrier, P., Baviera, S., Sozzani, S., and Mantovani, A. (2002). The chemokine system: tuning and shaping by regulation of receptor expression and coupling in polarized responses. Allergy 57, 972–982. Loetscher, M., Loetscher, P., Brass, N., Meese, E., and Moser, B. (1998). Lymphocyte-specific chemokine receptor CXCR3: regulation, chemokine binding and gene localization. Eur. J. Immunol. 28, 3696–3705. Luster, A. D., Greenberg, S. M., and Leder, P. (1995). The IP-10 chemokine binds to a specific cell surface heparan sulfate site shared with platelet factor 4 and inhibits endothelial cell proliferation. J. Exp. Med. 182, 219–231. Ma, Q., Jones, D., Borghesani, P. R., Segal, R. A., Nagasawa, T., Kishimoto, T., Bronson, R. T., and Springer, T. A. (1998). Impaired B-lymphopoiesis, myelopoiesis, and derailed cerebellar neuron migration in CXCR4- and SDF-1-deficient mice. Proc. Natl. Acad. Sci. USA 95, 9448–9453. Maione, T. E., Gray, G. S., Petro, J., Hunt, A. J., Donner, A. L., Bauer, S. I., Carson, H. F., and Sharpe, R. J. (1990). Inhibition of angiogenesis by recombinant human platelet factor-4 and related peptides. Science 247, 77–79. Malkowski, M. G., Lazar, J. B., Johnson, P. H., and Edwards, B. F. (1997). The amino-terminal residues in the crystal structure of connective tissue activating peptide-III (des10) block the ELR chemotactic sequence. J. Mol. Biol. 266, 367–380. Marguet, D., Baggio, L., Kobayashi, T., Bernard, A.-M., Pierres, M., Nielsen, P. F., Ribel, U., Watanabe, T., Drucker, D. J., and Wagtmann, N. (2000). Enhanced insulin secretion and improved glucose tolerance in mice lacking CD26. Proc. Natl. Acad. Sci. USA 97, 6874–6879. Martinelli, R., Sabroe, I., LaRosa, G., Williams, T. J., and Pease, J. E. (2001). The CC chemokine eotaxin (CCL11) is a partial agonist of CC chemokine receptor 2b. J. Biol. Chem. 276, 42957–42964. McQuibban, G. A., Gong, J.-H., Tam, E. M., McCulloch, C. A., Clark-Lewis, I., and Overall, C. M. (2000). Inflammation dampened by gelatinase A cleavage of monocyte chemoattractant protein-3. Science 289, 1202–1206. McQuibban, G. A., Butler, G. S., Gong, J.-H., Bendall, L., Power, C., Clark-Lewis, I., and Overall, C. M. (2001). Matrix metalloproteinase activity inactivates the CXC chemokine stromal cell-derived factor-1. J. Biol. Chem. 276, 43503–43508.

38

S. STRUYF ET AL.

McQuibban, G. A., Gong, J.-H., Wong, J. P., Wallace, J. L., Clark-Lewis, I., and Overall, C. M. (2002). Matrix metalloproteinase processing of monocyte chemoattractant proteins generates CC chemokine receptor antagonists with anti-inflammatory properties in vivo. Blood 100, 1160–1167. Melchers, F., Rolink, A. G., and Schaniel, C. (1999). The role of chemokines in regulating cell migration during humoral immune responses. Cell 99, 351–354. Mellado, M., Rodriguez-Frade, J. M., Vila-Coro, A. J., de Ana, A. M., and Martinez-A., C. (1999). Chemokine control of HIV-1 infection. Nature 400, 723–724. Menten, P., Struyf, S., Schutyser, E., Wuyts, A., De Clercq, E., Schols, D., Proost, P., and Van Damme, J. (1999). The LD78 isoform of MIP-1 is the most potent CCR5 agonist and HIV-1-inhibiting chemokine. J. Clin. Invest. 104, R1–R5. Mentlein, R. (1999). Dipeptidyl-peptidase IV (CD26)—role in the inactivation of regulatory peptides. Regul. Pept. 85, 9–24. Middleton, J., Neil, S., Wintle, J., Clark-Lewis, I., Moore, H., Lam, C., Auer, M., Hub, E., and Rot, A. (1997). Transcytosis and surface presentation of IL-8 by venular endothelial cells. Cell 91, 385–395. Mochizuki, M., Bartels, J., Mallet, A. I., Christophers, E., and Schro¨der, J.-M. (1998). IL-4 induces eotaxin: a possible mechanism of selective eosinophil recruitment in helminth infection and atopy. J. Immunol. 160, 60–68. Monteclaro, F. S., and Charo, I. F. (1997). The amino-terminal domain of CCR2 is both necessary and sufficient for high affinity binding of monocyte chemoattractant protein 1. Receptor activation by a pseudo-tethered ligand. J. Biol. Chem. 272, 23186–23190. Morimoto, C., and Schlossman, S. F. (1998). The structure and function of CD26 in the T-cell immune response. Immunol. Rev. 161, 55–70. Murdoch, C., and Finn, A. (2000). Chemokine receptors and their role in inflammation and infectious diseases. Blood 95, 3032–3043. Murphy, P. M., Baggiolini, M., Charo, I. F., He´bert, C. A., Horuk, R., Matsushima, K., Miller, L. H., Oppenheim, J. J., and Power, C. A. (2000). International union of pharmacology. XXII. Nomenclature for chemokine receptors. Pharmacol. Rev. 52, 145–176. Mu¨ller, A., Homey, B., Soto, H., Ge, N., Catron, D., Buchanan, M. E., McClanahan, T., Murphy, E., Yuan, W., Wagner, S. N., Barrera, J. L., Mohar, A., Vera´stegui, E., and Zlotnik, A. (2001). Involvement of chemokine receptors in breast cancer metastasis. Nature 410, 50–56. Nagasawa, T., Kikutani, H., and Kishimoto, T. (1994). Molecular cloning and structure of a preB-cell growth-stimulating factor. Proc. Natl. Acad. Sci. USA 91, 2305–2309. Nagasawa, T., Hirota, S., Tachibana, K., Takakura, N., Nishikawa, S., Kitamura, Y., Yoshida, N., Kikutani, H., and Kishimoto, T. (1996). Defects of B-cell lymphopoiesis and bone-marrow myelopoiesis in mice lacking the CXC chemokine PBSF/SDF-1. Nature 382, 635–638. Nakagawa, H., Hatakeyama, S., Ikesue, A., and Miyai, H. (1991). Generation of interleukin-8 by plasmin from AVLPR-interleukin-8, the human fibroblast-derived neutrophil chemotactic factor. FEBS Lett. 282, 412–414. Nibbs, R. J. B., Wylie, S. M., Yang, J., Landau, N. R., and Graham, G. J. (1997). Cloning and characterisation of a novel promiscuous human -chemokine receptor D6. J. Biol. Chem. 272, 32078–32083. Nibbs, R. J. B., Kriehuber, E., Ponath, P. D., Parent, D., Qin, S., Campbell, J. D. M., Henderson, A., Kerjaschki, D., Maurer, D., Graham, G. J., and Rot, A. (2001). The -chemokine receptor D6 is expressed by lymphatic endothelium and a subset of vascular tumors. Am. J. Pathol. 158, 867–877. Noso, N., Sticherling, M., Bartels, J., Mallet, A. I., Christophers, E., and Schro¨der, J.-M. (1996). Identification of an N-terminally truncated form of the chemokine RANTES and


39

granulocyte-macrophage colony-stimulating factor as major eosinophil attractants released by cytokine-stimulated dermal fibroblasts. J. Immunol. 156, 1946–1953. Noso, N., Bartels, J., Mallet, A. I., Mochizuki, M., Christophers, E., and Schro¨der, J.-M. (1998). Delayed production of biologically active O-glycosylated forms of human eotaxin by TNFstimulated dermal fibroblasts. Eur. J. Biochem. 253, 114–122. Nourshargh, S., Perkins, J. A., Showell, H. J., Matsushima, K., Williams, T. J., and Collins, P. D. (1992). A comparative study of the neutrophil stimulatory activity in vitro and pro-inflammatory properties in vivo of 72 amino acid and 77 amino acid IL-8. J. Immunol. 148, 106–111. Nufer, O., Corbett, M., and Walz, A. (1999). Amino-terminal processing of chemokine ENA-78 regulates biological activity. Biochemistry 38, 636–642. O’Brien, S. J., and Moore, J. P. (2000). The effect of genetic variation in chemokines and their receptors on HIV transmission and progression to AIDS. Immunol. Rev. 177, 99–111. Ogilvie, P., Bardi, G., Clark-Lewis, I., Baggiolini, M., and Uguccioni, M. (2001). Eotaxin is a natural antagonist for CCR2 and an agonist for CCR5. Blood 97, 1920–1924. Ohtsuki, T., Hosono, O., Kobayashi, H., Munakata, Y., Souta, A., Shioda, T., and Morimoto, C. (1998). Negative regulation of the anti-human immunodeficiency virus and chemotactic activity of human stromal cell-derived factor 1alpha by CD26/dipeptidyl peptidase IV. FEBS Lett. 431, 236–240. Opdenakker, G., Van den Steen, P. E., and Van Damme, J. (2001). Gelatinase B: a tuner and amplifier of immune functions. Trends Immunol. 22, 571–579. Oppenheim, J. J., Zachariae, C. O., Mukaida, N., and Matsushima, K. (1991). Properties of the novel proinflammatory supergene ‘‘intercrine’’ cytokine family. Annu. Rev. Immunol. 9, 617–648. Oravecz, T., Pall, M., Roderiquez, G., Gorrell, M. D., Ditto, M., Nguyen, N. Y., Boykins, R., Unsworth, E., and Norcross, M. A. (1997). Regulation of the receptor specificity and function of the chemokine RANTES (regulated on activation, normal T cell expressed and secreted) by dipeptidyl peptidase IV (CD26)-mediated cleavage. J. Exp. Med. 186, 1865–1872. Padrines, M., Wolf, M., Walz, A., and Baggiolini, M. (1994). Interleukin-8 processing by neutrophil elastase, cathepsin G and proteinase-3. FEBS Lett. 352, 231–235. Pal, R., Garzino-Demo, A., Markham, P. D., Burns, J., Brown, M., Gallo, R. C., and DeVico, A. L. (1997). Inhibition of HIV-1 infection by the beta-chemokine MDC. Science 278, 695–698. Proost, P., Wuyts, A., Conings, R., Lenaerts, J.-P., Billiau, A., Opdenakker, G., and Van Damme, J. (1993a). Human and bovine granulocyte chemotactic protein-2: complete amino acid sequence and functional characterization as chemokines. Biochemistry 32, 10170–10177. Proost, P., De Wolf-Peeters, C., Conings, R., Opdenakker, G., Billiau, A., and Van Damme, J. (1993b). Identification of a novel granulocyte chemotactic protein (GCP-2) from human tumor cells. In vitro and in vivo comparison with natural forms of GRO, IP-10 and IL-8. J. Immunol. 150, 1000–1010. Proost, P., Struyf, S., Couvreur, M., Lenaerts, J.-P., Conings, R., Menten, P., Verhaert, P., Wuyts, A., and Van Damme, J. (1998a). Posttranslational modifications affect the activity of the human monocyte chemotactic proteins MCP-1 and MCP-2: identification of MCP-2(6–76) as a natural chemokine inhibitor. J. Immunol. 160, 4034–4041. Proost, P., Struyf, S., Schols, D., Durinx, C., Wuyts, A., Lenaerts, J.-P., De Clercq, E., De Meester, I., and Van Damme, J. (1998b). Processing by CD26/dipeptidyl-peptidase IV reduces the chemotactic and anti-HIV-1 activity of stromal-cell-derived factor-1alpha. FEBS Lett. 432, 73–76. Proost, P., De Meester, I., Schols, D., Struyf, S., Lambeir, A.-M., Wuyts, A., Opdenakker, G., De Clercq, E., Scharpe´, S., and Van Damme, J. (1998c). Amino-terminal truncation of chemokines by CD26/dipeptidyl-peptidase IV. Conversion of RANTES into a potent inhibitor of monocyte chemotaxis and HIV-1-infection. J. Biol. Chem. 273, 7222–7227.

40

S. STRUYF ET AL.

Proost, P., Struyf, S., Schols, D., Opdenakker, G., Sozzani, S., Allavena, P., Mantovani, A., Augustyns, K., Bal, G., Haemers, A., Lambeir, A.-M., Scharpe´, S., Van Damme, J., and De Meester, I. (1999). Truncation of macrophage-derived chemokine by CD26/ dipeptidylpeptidase IV beyond its predicted cleavage site affects chemotactic activity and CC chemokine receptor 4 interaction. J. Biol. Chem. 274, 3988–3993. Proost, P., Menten, P., Struyf, S., Schutyser, E., De Meester, I., and Van Damme, J. (2000). Cleavage by CD26/dipeptidyl peptidase IV converts the chemokine LD78 into a most efficient monocyte attractant and CCR1 agonist. Blood 96, 1674–1680. Proost, P., Schutyser, E., Menten, P., Struyf, S., Wuyts, A., Opdenakker, G., Detheux, M., Parmentier, M., Durinx, C., Lambeir, A.-M., Neyts, J., Liekens, S., Maudgal, P., Billiau, A., and Van Damme, J. (2001). Aminoterminal truncation of CXCR3 agonists impairs receptor signaling and lymphocyte chemotaxis, while preserving anti-angiogenic properties. Blood 98, 3554–3561. Qin, S., Rottman, J. B., Myers, P., Kassam, N., Weinblatt, M., Loetscher, M., Koch, A. E., Moser, B., and Mackay, C. R. (1998). The chemokine receptors CXCR3 and CCR5 mark subsets of T cells associated with certain inflammatory reactions. J. Clin. Invest. 101, 746–754. Reeves, J. D., Hibbitts, S., Simmons, G., McKnight, A., Azevedo-Pereira, J. M., Moniz-Pereira, J., and Clapham, P. R. (1999). Primary human immunodeficiency virus type 2 (HIV-2) isolates infect CD4-negative cells via CCR5 and CXCR4: comparison with HIV-1 and simian immunodeficiency virus and relevance to cell tropism in vivo. J. Virol. 73, 7795–7804. Richter, R., Schulz-Knappe, P., John, H., and Forssmann, W.-G. (2000). Posttranslationally processed forms of the human chemokine HCC-1. Biochemistry 39, 10799–10805. Rizzuto, C. D., Wyatt, R., Hernańdez-Ramos, N., Sun, Y., Kwong, P. D., Hendrickson, W. A., and Sodroski, J. (1998). A conserved HIV gp120 glycoprotein structure involved in chemokine receptor binding. Science 280, 1949–1953. Rollins, B. J. (1997). Chemokines. Blood 90, 909–928. Rossi, D., and Zlotnik, A. (2000). The biology of chemokines and their receptors. Annu. Rev. Immunol. 18, 217–242. Sabroe, I., Hartnell, A., Jopling, L. A., Bel, S., Ponath, P. D., Pease, J. E., Collins, P. D., and Williams, T. J. (1999). Differential regulation of eosinophil chemokine signaling via CCR3 and non-CCR3 pathways. J. Immunol. 162, 2946–2955. Sallusto, F., Schaerli, P., Loetscher, P., Schaniel, C., Lenig, D., Mackay, C. R., Qin, S., and Lanzavecchia, A. (1998). Rapid and coordinated switch in chemokine receptor expression during dendritic cell maturation. Eur. J. Immunol. 28, 2760–2769. Samson, M., Libert, F., Doranz, B. J., Rucker, J., Liesnard, C., Farber, C.-M., Saragosti, S., Lapoume´roulie, C., Cognaux, J., Forceille, C., Muyldermans, G., Verhofstede, C., Burtonboy, G., Georges, M., Imai, T., Rana, S., Yi, Y., Smyth, R. J., Collman, R. G., Doms, R. W., Vassart, G., and Parmentier, M. (1996). Resistance to HIV-1 infection in caucasian individuals bearing mutant alleles of the CCR-5 chemokine receptor gene. Nature 382, 722–725. Sato, Y., Abe, M., and Takaki, R. (1990). Platelet factor 4 blocks the binding of basic fibroblast growth factor to the receptor and inhibits the spontaneous migration of vascular endothelial cells. Biochem. Biophys. Res. Commun. 172, 595–600. Scarlatti, G., Tresoldi, E., Bjo¨rndal, A., Fredriksson, R., Colognesi, C., Deng, H. K., Malnati, M. S., Plebani, A., Siccardi, A. G., Littman, D. R., Fenyo¨, E. M., and Lusso, P. (1997). In vivo evolution of HIV-1 co-receptor usage and sensitivity to chemokine-mediated suppression. Nat. Med. 3, 1259–1265. Schaniel, C., Sallusto, F., Ruedl, C., Sideras, P., Melchers, F., and Rolink, A. G. (1999). Three chemokines with potential functions in T lymphocyte-independent and -dependent B lymphocyte stimulation. Eur. J. Immunol. 29, 2934–2947. Schaniel, C., Rolink, A. G., and Melchers, F. (2001). Attractions and migrations of lymphoid cells in the organization of humoral immune responses. Adv. Immunol. 78, 111–168.


41

Scheel-Toellner, D., Richter, E., Toellner, K.-M., Reiling, N., Wacker, H.-H., Flad, H.-D., and Gerdes, J. (1995). CD26 expression in leprosy and other granulomatous diseases correlates with the production of interferon-gamma. Lab. Invest. 73, 685–690. Schmid, J., and Weissmann, C. (1987). Induction of mRNA for a serine protease and a beta-thromboglobulin-like protein in mitogen-stimulated human leukocytes. J. Immunol. 139, 250–256. Schols, D., Proost, P., Struyf, S., Wuyts, A., De Meester, I., Scharpe´, S., Van Damme, J., and De Clercq, E. (1998). CD26-processed RANTES(3–68), but not intact RANTES, has potent anti-HIV-1 activity. Antiviral Res. 39, 175–187. Schro¨der, J. M., Mrowietz, U., Morita, E., and Christophers, E. (1987). Purification and partial biochemical characterization of a human monocyte-derived, neutrophil-activating peptide that lacks interleukin 1 activity. J. Immunol. 139, 3474–3483. Schro¨der, J.-M., Mrowietz, U., and Christophers, E. (1988). Identification of different charged species of a human monocyte derived neutrophil activating peptide (MONAP). Biochem. Biophys. Res. Commun. 152, 277–284. Schro¨der, J. M., Sticherling, M., Henneicke, H. H., Preissner, W. C., and Christophers, E. (1990). IL-1 alpha or tumor necrosis factor-alpha stimulate release of three NAP-1/IL-8-related neutrophil chemotactic proteins in human dermal fibroblasts. J. Immunol. 144, 2223–2232. Schulz-Knappe, P., Ma¨gert, H.-J., Dewald, B., Meyer, M., Cetin, Y., Kubbies, M., Tomeczkowski, J., Kirchhoff, K., Raida, M., Adermann, K., Kist, A., Reinecke, M., Sillard, R., Pardigol, A., Uguccioni, M., Baggiolini, M., and Forssmann, W.-G. (1996). HCC-1, a novel chemokine from human plasma. J. Exp. Med. 183, 295–299. Schutyser, E., Struyf, S., Wuyts, A., Put, W., Geboes, K., Grillet, B., Opdenakker, G., and Van Damme, J. (2001). Selective induction of CCL18/PARC by staphylococcal enterotoxins in mononuclear cells and enhanced levels in septic and rheumatoid arthritis. Eur. J. Immunol. 31, 3755–3762. Shioda, T., Kato, H., Ohnishi, Y., Tashiro, K., Ikegawa, M., Nakayama, E. E., Hu, H., Kato, A., Sakai, Y., Liu, H., Honjo, T., Nomoto, A., Iwamoto, A., Morimoto, C., and Nagai, Y. (1998). Anti-HIV-1 and chemotactic activities of human stromal cell-derived factor 1alpha (SDF1alpha) and SDF-1beta are abolished by CD26/dipeptidyl peptidase IV-mediated cleavage. Proc. Natl. Acad. Sci. USA 95, 6331–6336. Sica, A., Saccani, A., Borsatti, A., Power, C. A., Wells, T. N., Luini, W., Polentarutti, N., Sozzani, S., and Mantovani, A. (1997). Bacterial lipopolysaccharide rapidly inhibits expression of C-C chemokine receptors in human monocytes. J. Exp. Med. 185, 969–974. Smith, M. W., Dean, M., Carrington, M., Winkler, C., Huttley, G. A., Lomb, D. A., Goedert, J. J., O’Brien, T. R., Jacobson, L. P., Kaslow, R., Buchbinder, S., Vittinghoff, E., Vlahov, D., Hoots, K., Hilgartner, M. W., and O’Brien, S. J. (1997). Contrasting genetic influence of CCR2 and CCR5 variants on HIV-1 infection and disease progression. Hemophilia Growth and Development Study (HGDS), Multicenter AIDS Cohort Study (MACS), Multicenter Hemophilia Cohort Study (MHCS), San Francisco City Cohort (SFCC), ALIVE Study. Science 277, 959–965. Sozzani, S., Allavena, P., D’Amico, G., Luini, W., Bianchi, G., Kataura, M., Imai, T., Yoshie, O., Bonecchi, R., and Mantovani, A. (1998). Differential regulation of chemokine receptors during dendritic cell maturation: a model for their trafficking properties. J. Immunol. 161, 1083–1086. Strieter, R. M., Polverini, P. J., Kunkel, S. L., Arenberg, D. A., Burdick, M. D., Kasper, J., Dzuiba, J., Van Damme, J., Walz, A., Marriott, D., Shan, S.-Y., Roczniak, S., and Shanafelt, A. B. (1995). The functional role of the ELR motif in CXC chemokine-mediated angiogenesis. J. Biol. Chem. 270, 27348–27357. Struyf, S., De Meester, I., Scharpe´, S., Lenaerts, J.-P., Menten, P., Wang, J. M., Proost, P., and Van Damme, J. (1998a). Natural truncation of RANTES abolishes signaling through the CC

42

S. STRUYF ET AL.

chemokine receptors CCR1 and CCR3, impairs its chemotactic potency and generates a CC chemokine inhibitor. Eur. J. Immunol. 28, 1262–1271. Struyf, S., Proost, P., Sozzani, S., Mantovani, A., Wuyts, A., De Clercq, E., Schols, D., and Van Damme, J. (1998b). Enhanced anti-HIV-1 activity and altered chemotactic potency of NH2terminally processed macrophage-derived chemokine (MDC) imply an additional MDC receptor. J. Immunol. 161, 2672–2675. Struyf, S., Van Collie, E., Paemen, L., Put, W., Lenaerts, J.-P., Proost, P., Opdenakker, G., and Van Damme, J. (1998c). Synergistic induction of MCP-1 and -2 by IL-1beta and interferons in fibroblasts and epithelial cells. J. Leukoc. Biol. 63, 364–372. Struyf, S., Proost, P., Schols, D., De Clercq, E., Opdenakker, G., Lenaerts, J.-P., Detheux, M., Parmentier, M., De Meester, I., Scharpe´, S., and Van Damme, J. (1999). CD26/dipeptidylpeptidase IV down-regulates the eosinophil chemotactic potency, but not the anti-HIV activity of human eotaxin by affecting its interaction with CC chemokine receptor 3. J. Immunol. 162, 4903–4909. Struyf, S., Menten, P., Lenaerts, J.-P., Put, W., D’Haese, A., De Clercq, E., Schols, D., Proost, P., and Van Damme, J. (2001a). Diverging binding capacities of natural LD78 isoforms of macrophage inflammatory protein-1 to the CC chemokine receptors 1, 3 and 5 affect their anti-HIV-1 activity and chemotactic potencies for neutrophils and eosinophils. Eur. J. Immunol. 31, 2170–2178. Struyf, S., Proost, P., Lenaerts, J.-P., Stoops, G., Wuyts, A., and Van Damme, J. (2001b). Identification of a blood-derived chemoattractant for neutrophils and lymphocytes as a novel CC chemokine, Regakine-1. Blood 97, 2197–2204. Struyf, S., Stoops, G., Van Coillie, E., Gouwy, M., Schutyser, E., Lenaerts, J.-P., Fiten, P., Van Aelst, I., Proost, P., Opdenakker, G., and Van Damme, J. (2001c). Gene cloning of a new plasma CC chemokine, activating and attracting myeloid cells in synergy with other chemoattractants. Biochemistry 40, 11715–11722. Tanaka, Y., Imai, T., Baba, M., Ishikawa, I., Uehira, M., Nomiyama, H., and Yoshie, O. (1999). Selective expression of liver and activation-regulated chemokine (LARC) in intestinal epithelium in mice and humans. Eur. J. Immunol. 29, 633–642. Tang, H. L., and Cyster, J. G. (1999). Chemokine Up-regulation and activated T cell attraction by maturing dendritic cells. Science 284, 819–822. Tilton, B., Ho, L., Oberlin, E., Loetscher, P., Baleux, F., Clark-Lewis, I., and Thelen, M. (2000). Signal transduction by CXC chemokine receptor 4. Stromal cell-derived factor 1 stimulates prolonged protein kinase B and extracellular signal-regulated kinase 2 activation in T lymphocytes. J. Exp. Med. 192, 313–324. Tomhave, E. D., Richardson, R. M., Didsbury, J. R., Menard, L., Snyderman, R., and Ali, H. (1994). Cross-desensitization of receptors for peptide chemoattractants. Characterization of a new form of leukocyte regulation. J. Immunol. 153, 3267–3275. Torimoto, Y., Dang, N. H., Vivier, E., Tanaka, T., Schlossman, S. F., and Morimoto, C. (1991). Coassociation of CD26 (dipeptidyl peptidase IV) with CD45 on the surface of human T lymphocytes. J. Immunol. 147, 2514–2517. Vakili, J., Sta¨ndker, L., Detheux, M., Vassart, G., Forssmann, W.-G., and Parmentier, M. (2001). Urokinase plasminogen activator and plasmin efficiently convert hemofiltrate CC chemokine 1 into its active [9–74] processed variant. J. Immunol. 167, 3406–3413. Valenzuela-Fernandez, A., Planchenault, T., Baleux, F., Staropoli, I., Le-Barillec, K., Leduc, D., Delaunay, T., Lazarini, F., Virelizier, J.-L., Chignard, M., Pidard, D., and ArenzanaSeisdedos, F. (2002). Leukocyte elastase negatively regulates stromal cell-derived factor-1 (SDF-1)/CXCR4 binding and functions by amino-terminal processing of SDF-1 and CXCR4. J. Biol. Chem. 277, 15677–15689. Van Coillie, E., Proost, P., Van Aelst, I., Struyf, S., Polfliet, M., De Meester, I., Harvey, D. J., Van Damme, J., and Opdenakker, G. (1998). Functional comparison of two human monocyte


43

chemotactic protein-2 isoforms, role of the amino-terminal pyroglutamic acid and processing by CD26/dipeptidyl peptidase IV. Biochemistry 37, 12672–12680. Van Damme, J., Van Beeumen, J., Opdenakker, G., and Billiau, A. (1988). A novel, NH2-terminal sequence-characterized human monokine possessing neutrophil chemotactic, skin-reactive, and granulocytosis-promoting activity. J. Exp. Med. 167, 1364–1376. Van Damme, J., Decock, B., Conings, R., Lenaerts, J.-P., Opdenakker, G., and Billiau, A. (1989a). The chemotactic activity for granulocytes produced by virally infected fibroblasts is identical to monocyte-derived interleukin 8. Eur. J. Immunol. 19, 1189–1194. Van Damme, J., Van Beeumen, J., Conings, R., Decock, B., and Billiau, A. (1989b). Purification of granulocyte chemotactic peptide/interleukin-8 reveals N-terminal sequence heterogeneity similar to that of beta-thromboglobulin. Eur. J. Biochem. 181, 337–344. Van Damme, J., Proost, P., Lenaerts, J.-P., and Opdenakker, G. (1992). Structural and functional identification of two human, tumor-derived monocyte chemotactic proteins (MCP-2 and MCP-3) belonging to the chemokine family. J. Exp. Med. 176, 59–65. Van Damme, J., Proost, P., Put, W., Arens, S., Lenaerts, J.-P., Conings, R., Opdenakker, G., Heremans, H., and Billiau, A. (1994). Induction of monocyte chemotactic proteins MCP-1 and MCP-2 in human fibroblasts and leukocytes by cytokines and cytokine inducers. Chemical synthesis of MCP-2 and development of a specific RIA. J. Immunol. 152, 5495–5502. Van den Steen, P. E., Proost, P., Wuyts, A., Van Damme, J., and Opdenakker, G. (2000). Neutrophil gelatinase B potentiates interleukin-8 tenfold by aminoterminal processing, whereas it degrades CTAP-III, PF-4, and GRO-alpha and leaves RANTES and MCP-2 intact. Blood 96, 2673–2681. Vanham, G., Kestens, L., De Meester, I., Vingerhoets, J., Penne, G., Vanhoof, G., Scharpe´, S., Heyligen, H., Bosmans, E., Ceuppens, J. L., and Gigase, P. (1993). Decreased expression of the memory marker CD26 on both CD4 þ and CD8 þ T lymphocytes of HIV-infected subjects. J. Acquir. Imm. Defic. Syndr. 6, 749–757. Verfaillie, C. M., Catanzarro, P. M., and Li, W.-N. (1994). Macrophage inflammatory protein 1 alpha, interleukin 3 and diffusible marrow stromal factors maintain human hematopoietic stem cells for at least eight weeks in vitro. J. Exp. Med. 179, 643–649. Vestergaard, C., Yoneyama, H., Murai, M., Nakamura, K., Tamaki, K., Terashima, Y., Imai, T., Yoshie, O., Irimura, T., Mizutani, H., and Matsushima, K. (1999). Overproduction of Th2specific chemokines in NC/Nga mice exhibiting atopic dermatitis-like lesions. J. Clin. Invest. 104, 1097–1105. von Bonin, A., Hu¨hn, J., and Fleischer, B. (1998). Dipeptidyl-peptidase IV/CD26 on T cells: analysis of an alternative T-cell activation pathway. Immunol. Rev. 161, 43–53. Vulcano, M., Albanesi, C., Stoppacciaro, A., Bagnati, R., D’Amico, G., Struyf, S., Transidico, P., Bonecchi, R., Del Prete, A., Allavena, P., Ruco, L. P., Chiabrando, C., Girolomoni, G., Mantovani, A., and Sozzani, S. (2001). Dendritic cells as a major source of macrophage-derived chemokine/CCL22 in vitro and in vivo. Eur. J. Immunol. 31, 812–822. Walz, A., Peveri, P., Aschauer, H., and Baggiolini, M. (1987). Purification and amino acid sequencing of NAF, a novel neutrophil-activating factor produced by monocytes. Biochem. Biophys. Res. Commun. 149, 755–761. Walz, D. A., Wu, V. Y., de Lamo, R., Dene, H., and McCoy, L. E. (1977). Primary structure of human platelet factor 4. Thromb. Res. 11, 893–898. Wang, J. M., Chertov, O., Proost, P., Li, J. J., Menten, P., Xu, L., Sozzani, S., Mantovani, A., Gong, W., Schirrmacher, V., Van Damme, J., and Oppenheim, J. J. (1998). Purification and identification of chemokines potentially involved in kidney-specific metastasis by a murine lymphoma variant: induction of migration and NFkappaB activation. Int. J. Cancer 75, 900–907. Wesley, U. V., Albino, A. P., Tiwari, S., and Houghton, A. N. (1999). A role for dipeptidyl peptidase IV in suppressing the malignant phenotype of melanocytic cells. J. Exp. Med. 190, 311–322.

44

S. STRUYF ET AL.

Willheim, M., Ebner, C., Baier, K., Kern, W., Schrattbauer, K., Thien, R., Kraft, D., Breiteneder, H., Reinisch, W., and Scheiner, O. (1997). Cell surface characterization of T lymphocytes and allergen-specific T cell clones: correlation of CD26 expression with T(H1) subsets. J. Allergy Clin. Immunol. 100, 348–355. Wuyts, A., Govaerts, C., Struyf, S., Lenaerts, J.-P., Put, W., Conings, R., Proost, P., and Van Damme, J. (1999a). Isolation of the CXC chemokines ENA-78, GRO alpha and GRO gamma from tumor cells and leukocytes reveals NH2-terminal heterogeneity. Functional comparison of different natural isoforms. Eur. J. Biochem. 260, 421–429. Wuyts, A., Struyf, S., Proost, P., and Van Damme, J. (1999b). Chemokines, in The`ze, J. (ed.) ‘‘The cytokine network and immune functions’’ Oxford, Oxford University Press, pp. 125–145. Xiao, L., Owen, S. M., Goldman, I., Lal, A. A., deJong, J. J., Goudsmit, J., and Lal, R. B. (1998). CCR5 coreceptor usage of non-syncytium-inducing primary HIV-1 is independent of phylogenetically distinct global HIV-1 isolates: delineation of consensus motif in the V3 domain that predicts CCR-5 usage. Virology 240, 83–92. Xin, X., Shioda, T., Kato, A., Liu, H., Sakai, Y., and Nagai, Y. (1999). Enhanced anti-HIV-1 activity of CC-chemokine LD78 , a non-allelic variant of MIP-1/LD78. FEBS Lett. 457, 219–222. Yanagihara, S., Komura, E., Nagafune, J., Watarai, H., and Yamaguchi, Y. (1998). EBI1/CCR7 is a new member of dendritic cell chemokine receptor that is up-regulated upon maturation. J. Immunol. 161, 3096–3102. Yaron, A., and Naider, F. (1993). Proline-dependent structural and biological properties of peptides and proteins. Crit. Rev. Biochem. Mol. Biol. 28, 31–81. Yayon, A., Klagsbrun, M., Esko, J. D., Leder, P., and Ornitz, D. M. (1991). Cell surface, heparinlike molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. Cell 64, 841–848. Yoshie, O., Imai, T., and Nomiyama, H. (2001). Chemokines in immunity. Adv. Immunol. 78, 57–110. Yoshimura, T., Matsushima, K., Tanaka, S., Robinson, E. A., Appella, E., Oppenheim, J. J., and Leonard, E. J. (1987). Purification of a human monocyte-derived neutrophil chemotactic factor that has peptide sequence similarity to other host defense cytokines. Proc. Natl. Acad. Sci. USA 84, 9233–9237. Yoshimura, T., Robinson, E. A., Appella, E., Matsushima, K., Showalter, S. D., Skeel, A., and Leonard, E. J. (1989). Three forms of monocyte-derived neutrophil chemotactic factor (MDNCF) distinguished by different lengths of the amino-terminal sequence. Mol. Immunol. 26, 87–93. Youn, B.-S., Zhang, S. M., Broxmeyer, H. E., Cooper, S., Antol, K., Fraser, M., and Kwon, B. S. (1998). Characterization of CKbeta8 and CKbeta8-1: two alternatively spliced forms of human beta-chemokine, chemoattractants for neutrophils, monocytes, and lymphocytes, and potent agonists at CC chemokine receptor 1. Blood 91, 3118–3126. Youn, B.-S., Mantel, C., and Broxmeyer, H. E. (2000). Chemokines, chemokine receptors and hematopoiesis. Immunol. Rev. 177, 150–174. Zlotnik, A., Morales, J., and Hedrick, J. A. (1999). Recent advances in chemokines and chemokine receptors. Crit. Rev. Immunol. 19, 1–47. Zlotnik, A., and Yoshie, O. (2000). Chemokines: a new classification system and their role in immunity. Immunity 12, 121–127.


Molecular Mechanisms of Host–Pathogen Interaction: Entry and Survival of Mycobacteria in Macrophages JOHN GATFIELD AND JEAN PIETERS1 Biozentrum der Universitaet Basel, Department of Biochemistry, Klingelbergstrasse 50-70, 4056 Basel, Switzerland

I. Introduction

Vertebrates are in constant danger of being colonized by microbes looking for favorable growing conditions. Despite the continuous contact with microbes, under normal circumstances, vertebrates do not develop infections, but are protected by their immune system. An efficient immune system against microbes should perform two tasks: microbe recognition and microbe elimination. Macrophages have a pivotal role in both of these processes during infection, primarily due to their ability to phagocytose and kill microorganisms. However, during the long period of host–pathogen co-evolution, microbes have developed most remarkable strategies to subvert their elimination by host cells. One of these carefully balanced pathogen–host relations is the interaction of Mycobacterium tuberculosis with its host cell, the macrophage. M. tuberculosis, the causative agent of tuberculosis, is a highly successful intracellular pathogen, which apparently has evolved to persist after phagocytosis within macrophages. Intracellular survival is based on the ability of mycobacteria to prevent the maturation of the phagosome they inhabit into a highly microbicidal phagolysosome. However, the mechanisms by which mycobacteria interfere with intracellular trafficking events and the host cell factors involved have remained largely unknown. Elucidation of the mechanisms operating during host–pathogen interaction may not only allow the design of novel pharmacological therapies to combat mycobacterial diseases, but will also give new insights into basic cell biological processes. A. INNATE IMMUNITY In the constant battle of hosts with microbes such as viruses, bacteria, fungi or eukaryotic parasites, hosts have developed a set of germ line encoded cellbound and soluble molecules (pattern recognition receptors, PRR) that recognize certain invariable and repetitive molecular patterns on the surface of pathogens (Janeway, 1989; Ezekowitz et al., 1990; Gordon, 1995; Silverstein, 1995; Medzhitov and Janeway, 2000). Recognition of pathogens 1

Tel.: 41-61-267-1494; fax: 41-61-267-2149; E-mail: [email protected] 45 Copyright ß 2003 by Elsevier (USA) All rights of reproduction in any form reserved. 0065-2776/03 $35.00

46

J. GATFIELD AND J. PIETERS

by PRRs leads to their engulfment into a vacuole, or phagosome. The cells that are highly efficient in phagocytosis are macrophages, neutrophils and dendritic cells. Intracellularly, the phagosome fuses with microbicidal lysosomes after which microbes are destroyed. Some microbes however survive phagocytosis and persist inside the cell. Pattern recognition and phagocytosis are central in eliminating bacteria and represent one part of the so-called innate immune system, a system that is active since the birth of an organism and functions independently of a previous encounter with the pathogen. Also soluble antimicrobial substances contribute to innate immunity. Examples are antimicrobial peptides such as defensins (Ganz and Lehrer, 1998) that perforate bacterial membranes, lytic enzymes such as lysozyme that digests the cell envelope of bacteria, and lactoferrin, an iron chelating protein that makes iron in body fluids unavailable to bacteria (Wilmott et al., 2000). Other mechanisms have evolved for the recognition and control of virally infected cells. Natural killer (NK) cells and some T-cell subsets survey the cells of the organism for reduced expression of defined surface molecules (MHC class I and MICA), which not only occurs during viral infection but also during tumorigenesis, and kill their target cell by secreting soluble and cell-bound molecules (Groh et al., 1998; Bauer et al., 1999; Colonna et al., 2000; Groh et al., 2001). Since the components of the innate immune system are omnipresent, without need for induction, they allow an immediate anti-microbial response. All organisms except vertebrates rely exclusively on the innate immune system for defense. In addition, vertebrates cope with most infections purely by using their innate immune system, emphasizing its central role. B. ADAPTIVE IMMUNITY The innate immune system cannot always control the pathogen after its entry into the organism. This is often due to the sophisticated microbial survival strategies. In this case, the initiation of an adaptive immune response supports the innate immune mechanisms. Lymphocytes, the key effector cells of the adaptive immune system, provide the host with additional pathogen detection strategies. Antigen recognition in the adaptive immune system is based on antigen receptors of the immunoglobulin super family that are generated with random specificities by somatic rearrangement of germ line encoded gene segments (Tonegawa, 1988). Cells expressing antigen receptors that can recognize antigen-specific molecules become activated upon pathogen recognition and start to proliferate. This necessity for proliferation causes a lag phase of several days before the pathogen-specific cells start to contribute to pathogen recognition and elimination. Besides antigen-specificity, the

MOLECULAR MECHANISMS OF HOST–PATHOGEN INTERACTION

47

adaptive immune system—in contrast to the innate immune system—displays a receptor diversity that principally allows responses to every chemical structure and an antigen-specific memory that gives survival advantage during the long life span and the low reproduction rate of vertebrates. Lymphocytes are subdivided into B- and T-lymphocytes. B cells secrete antigen-specific antibodies, the main effectors of the adaptive humoral response. Antibodies recognize antigens in their native three-dimensional state. Besides proteins, any other structure can serve as an antigenic epitope for B cells. This reflects the task of antibodies to recognize and neutralize extracellular pathogens and their products (Tonegawa, 1983; Lanzavecchia, 1987). T-lymphocytes are designed to recognize intracellular pathogens using T-cell receptors (TCR) present at the surface (Davis et al., 1998). Microbial persistence occurs at two topologically separated sites inside a cell, the cytosol and the membrane-enclosed intracellular compartments. To monitor both sites, T-lymphocytes are subdivided into two functionally distinct populations, the CD4 þ and the CD8 þ T cells (Germain, 1994). Viruses and tumor antigens present in the cytosolic compartment are recognized by the cytotoxic CD8 þ T cells that upon recognition kill their target cell thereby also eliminating the virus (Rammensee et al., 1993b). Bacteria taken up in the vesicular system are recognized by the CD4 þ T cells that upon activation do not kill the target cell, but produce cytokines that help macrophages to kill their intracellular parasite or help B cells to produce pathogen-specific antibodies. For this reason, CD4 þ cells are termed T-helper cells (Abbas et al., 1996). To trigger T-cell receptors, the intracellular antigens have to be transported to the cell surface. This transport and the surface presentation of the antigens relies on a family of highly polymorphic molecules encoded in the major histocompatibility complex (MHC) located on chromosome 6 in humans and chromosome 17 in the mouse (Rhodes and Trowsdale, 1999). 1. CD8 þ T cells Recognize Cytosolically-Derived Peptides on Major Histocompatibility Complex (MHC) Class I Molecules and Kill Their Target Cells Antigens derived from the cytosol including viral proteins, proteins of certain bacteria and tumor proteins are recognized by CD8 þ T cells. Before a protein antigen is transported to the cell surface, it is subjected to extensive intracellular processing (Lehner and Cresswell, 1996). Cytosolic antigen processing is mainly performed by the proteasome, the major cytosolic multiprotease complex that acts on polyubiquitinated protein substrates and processes these into shorter peptides (Monaco, 1995; Baumeister et al., 1998). However, other cytosolic proteases might also be involved in antigen processing such as the tricorn protease that can substitute for proteasome

48


activity in proteasome-deficient cells (Glas et al., 1998; Tamura et al., 1998; Geier et al., 1999). The proteolytic fragments are then translocated into the endoplasmic reticulum (ER) by the ATP-dependent TAP1/TAP2 heterodimer, a member of the family of ABC-transporters (Spies et al., 1990; Androlewicz et al., 1993; Neefjes et al., 1993). After further proteolytic processing by proteases in the ER-lumen the peptide is loaded into the peptide binding cleft of newly synthesized MHC class I molecules with the assistance of chaperone molecules (Sadasivan et al., 1996). The binding cleft of MHC molecules consists of an arrangement of two separated -helices placed on a -pleated sheet and holds the peptide in an extended conformation (Bjorkman et al., 1987; Stura et al., 1992). The MHC class I binding cleft imposes strong restrictions upon the structure of its peptide ligands, and therefore only a minority of the TAP-transported peptides are successful binders. However, multiple alleles for MHC proteins with different binding properties compensate for these restrictions. Once the peptide-MHC class I complex has formed, it travels along the exocytic pathway via the Golgi complex to the cell surface for recognition by CD8 þ T cells (Townsend and Bodmer, 1989). MHC class I molecules are present on all nucleated cells of an organism. By elution and sequencing of the bound peptides from purified MHC class I molecules, peptide lengths and origin have been determined (Rammensee et al., 1993a). MHC class I binds only octa-, nona-, and decapeptides due to a closed conformation of the peptide binding cleft and the majority of peptides is derived from cytosolic proteins. The sampling of peptides in the ER by MHC class I and their surface presentation occurs in every nucleated cell constitutively and randomly, i.e., without distinguishing between foreign and self-peptides. One cellular mechanism that might ensure a constant supply of the constitutively operating MHC class I pathway is the processing of defective ribosomal products (DRiPs) by the proteasome. DRiPs are polypeptides that never attain native structure owing to errors in translation or post-translational processes necessary for proper protein folding and can constitute up to 30% of the newly synthesized proteins (Schild and Rammensee, 2000; Schubert et al., 2000). The result of a cytotoxic CD8 þ T cells recognizing a viral, bacterial or an otherwise foreign peptide in the MHC class I complex is, as mentioned above, the killing of the target cell. Activated cytotoxic T cells kill the target cell by secretion (perforin/granzymes) or surface expression (Fas-ligand) of apoptosisinducing molecules (Berke, 1997). This process destroys the cytosolic pathogens (viruses), or releases them for uptake by macrophages or recognition by antibodies/PRRs (bacteria). Interestingly, in the case of mycobacterially infected target cells, mycobacteria get released in a viable state upon destruction of the host cell, but can be killed by the mycobactericidal action of granulysin (Stenger et al., 1998). This protein is co-secreted with perforin and


49

granzymes by CTLs specifically recognizing CD1 molecules loaded with mycobacterial lipids. CD1 molecules are related to MHC class I molecules and present lipid antigens to T cells (Sugita et al., 2000; Ulrichs and Porcelli, 2000). 2. CD4 þ T cells Recognize MHC Class II Molecules Containing Peptides Derived from the Endosomal/Lysosomal System Antigens present in the endocytic system of a cell are the source of peptides presented by MHC class II molecules that are recognized by CD4 þ T cells. Peptides can be derived from endogenous proteins residing in the endocytic or exocytic pathway or from material that has been taken up by the different forms of endocytosis including phagocytosis of microbes (Pieters, 1999). Following internalization, antigens are shuttled to degradative acidic lysosomal compartments where they are proteolytically fragmented mainly by a family of cystein proteases, the cathepsins (Chapman, 1998). The endocytic pathway that provides antigenic peptide fragments and the exocytic pathway by which MHC class II molecules egress from the trans-Golgi network (TGN) intersect in a vesicular compartment for loading of class II molecules (MHC class II compartment, also called MIIC) (Peters et al., 1991; Pieters et al., 1991; Amigorena et al., 1994; Tulp et al., 1994; West et al., 1994). Here, the surrogate peptide CLIP (class II associated invariant chain peptide), that prevents premature peptide loading in the ER, is exchanged for a high-affinity antigenic peptide (Mellins et al., 1990; Kelly et al., 1991; Denzin and Cresswell, 1995; Sloan et al., 1995; Kropshofer et al., 1996; Ferrari et al., 1997). The MHC class II peptide binding cleft has an open conformation and as a consequence MHC class II binding peptides follow less strict rules concerning length and sequence, compared to those bound to MHC class I molecules (Falk et al., 1994; Fremont et al., 1996). Once loaded with peptide, the MHC class II molecules become transported to the cell surface where they are recognized by the TCRs of CD4 þ T-helper cells (Fig. 1). MHC class II molecules are mainly expressed on professional antigen presenting cells (macrophages, dendritic cells, neutrophils and B cells) but can be induced on many cell types by interferon- that stimulates the synthesis of the MHC class II transactivator (CIITA), the transcriptional regulator of many molecules involved in the MHC class II pathway (Steimle et al., 1994; Chang and Flavell, 1995; Kern et al., 1995). Recognition of foreign peptides on MHC class II molecules of phagocytes causes CD4 þ T cells to produce cytokines. Two major subclasses of CD4 þ cells exist according to their secreted cytokine pattern. The T-helper 1 subpopulation of CD4 þ cells secretes cytokines that drastically augment the antimicrobial capacities of macrophages and neutrophils. The T-helper 2 subpopulation secretes cytokines that have essential roles in initiating and regulating antibody responses (Asnagli and Murphy, 2001).

50


FIG. 1. Activation of the immune system by MHC class I and MHC class II molecules. MHC class I molecules present peptides derived from cytosolic proteins at the cell surface to cytotoxic T-lymphocytes. Activation of cytotoxic T-lymphocytes leads to rapid lysis of the presenting cell (left side). Antigens that reside extracellularly and are taken up through phagocytosis are degraded in the endosomal/lysosomal system and presented by MHC class II molecules (right side). MHC class II-peptide complexes are recognized by T helper cells that can either activate the presenting cell (T helper 1 cells) or stimulate the production of antibodies by B lymphocytes (T helper 2 cells). (From Pieters (1999), by permission of JAI Press/Elsevier Science, Stamford, CT.)

C. COOPERATION

OF

ADAPTIVE

AND INNATE IMMUNE

RESPONSE

The innate and adaptive immune systems are not isolated from each other, but cooperate. During intracellular bacterial infections, T-helper 1 cells play an essential role in helping phagocytes to eliminate intracellularly residing microbes. Infected phagocytes indicate the presence of microbes in their endocytic system by presenting pathogen-derived peptides in MHC class II molecules, and are recognized by T-helper 1 cells which, in turn, synthesize the cytokines interferon- and tumor necrosis factor-. The combination of these two cytokines enhances the anti-microbial mechanisms of the infected macrophage, a phenomenon termed macrophage activation (Adams and Hamilton, 1992; Le Page et al., 2000). Activated macrophages produce highly microbicidal oxygen and nitrogen radicals by the action of two enyzme complexes, the inducible nitric oxidesynthase (iNOS) and nicotinamide adenine dinucleotide phosphate-oxidase (NADPH-oxidase) that are recruited to the bacterial phagosome (Bogdan et al., 2000). The generation of oxygen radicals is termed oxidative burst. In addition, phagosome–lysosome fusion is enhanced by unknown mechanisms, also contributing to microbicidal activity (Kaufmann and Flesch, 1988; Kagaya et al., 1989; Ishibashi and Arai, 1990). How crucial T-helper 1 cells and their


51

cytokines are to gain full antimicrobial capacity in macrophages is demonstrated by the fact that AIDS patients who have low numbers of CD4 þ cells suffer from intracellular infections. Similarly, mice and humans with defects in interferon- signaling are highly susceptible to some intracellular pathogens (Cooper et al., 1993; Dalton et al., 1993; Huang et al., 1993; Dai et al., 1997; de Jong et al., 1998). Interferon- , however not only leads to enhanced killing of the intracellular parasite, but also increases the antigen-presenting capacity of the macrophage which, in turn, ensures activation of more microbe-specific T cells (Collins et al., 1984; Hart et al., 1993; Steimle et al., 1994). To efficiently eliminate an intracellular infection, the presence of T helper 1, and not T helper 2 cells is required. Whether a naive CD4 þ T cell develops into a T helper 1 or a T helper 2 cell is influenced by the cytokines present during its first activation by an antigen-presenting cell. Early in an infection, these cytokines are produced by cells of the innate immune system such as natural killer cells, dendritic cells and macrophages (Fearon and Locksley, 1996). Bacterially infected macrophages and dendritic cells produce interleukin-12 that directs polarization into T helper 1 cells, precisely the subgroup needed for an effective response against intracellular bacteria (Heufler et al., 1996). Consequently, impairment in interleukin-12 production leads to similar susceptibilities toward intracellular infections as deficiency in interferon- signaling (Altare et al., 1998). Thus, the innate immune system organizes its own help by directing CD4 þ T-cell differentiation towards the interferon- secreting T helper 1 subclass. II. Innate Recognition of Pathogens

Innate immunity secures immediate recognition of pathogens after their entry into the host organism. This is in contrast to the adaptive immune system that is induced over a period of days. As mentioned, the innate immune system relies on a fixed and limited set of germ line encoded molecules (PRRs) that recognize a broad spectrum of evolutionary conserved patterns or repetitive antigenic structures (pathogen associated molecular patterns, PAMPs) that are present on microbes but absent on higher eukaryotes (Janeway, 1989). Since PRRs recognize general patterns, microbes usually cannot escape recognition despite their high mutation rate. Patterns recognized include repetitive mannosylated surface antigens (Ezekowitz et al., 1990), lipopolysaccharide (LPS) on Gram-negative bacteria (Yang et al., 1998b), lipoteichoic acids on Gram-positive bacteria or unmethylated GC-rich DNA (Hemmi et al., 2000). Receptors recognizing such patterns are for example the complement receptors, scavenger receptors, the mannose receptor and the LPS-receptor CD14.

52


Many of these recognition molecules are located on phagocytes and directly trigger phagocytosis and intracellular degradation of the ligand. In addition, soluble recognition molecules (complement proteins, soluble defense collagens) are present in serum and tissue fluids, that opsonize (mark) pathogens for recognition by phagocytes via opsonin-specific receptors. A. MECHANISMS

OF

COMPLEMENT ACTIVATION

A central role of pathogen elimination in innate immunity is played by the complement system. It represents a hierarchic system of serum proteins that become sequentially activated in a proteolytic cascade (Carroll, 1998). Upon triggering by various stimuli, the activation of the cascade culminates in the deposition of the so-called C3 convertase complex on the pathogen cell surface. The C3 convertase proteolytically activates C3 molecules into highly reactive, thioester-containing C3b molecules that covalently opsonize the surrounding bacterial surface. One convertase complex can catalyze the covalent deposition of over a thousand C3b molecules leading to massively opsonized bacteria, which are then recognized and eliminated by phagocytes. Besides being phagocytosed, opsonized bacteria can also be directly lysed by the complement components C5 to C9 that form a membrane attack complex (MAC) perforating the bacterial cell wall. Deficiencies in the early complement components in humans is associated with high susceptibility to infection with pyogenic bacteria and Neisseria ssp. whereas the constituents of the MAC may be of only limited importance in immunity (Ross and Densen, 1984). The surface deposition of the C3 convertase can occur via three mechanisms. In a pathway known as the alternative pathway of complement activation, opsonization is initiated by a coincidental reaction of spontaneously formed C3b with a microbial surface molecule. C3b assembles with another serum factor to form the C3 convertase complex and this assembly leads to opsonization. This process, that is maintained by a positive feedback loop, is largely specific for microbes as the reaction of host cell surfaces with C3b is abortive due to the presence of membrane bound and soluble complement regulatory proteins that protect host cells. Secondly, the complement cascade can be initiated by the mannose-binding lectin, a pattern recognition protein which is a multimer of carbohydrate recognition domains that assemble with each other via C-terminal -helical collagen-like domains (Gadjeva et al., 2001). Due to the spacing of the 3–6 carbohydrate recognition domains, high avidity recognition is restricted to terminally mannosylated structures only if they are highly repetitive (Lee et al., 1992). These patterns occur on microbes but rarely in higher eukaryotes, where most carbohydrate modifications terminate with sialic acid. Interestingly, carbohydrate patterns can change during oncogenic transformation of cells


53

suggesting a potential role of mannose binding lectin also in tumor control (Ma et al., 1999). Mannose binding lectin is a member of the family of soluble defense collagens (also including C1q, surfactant proteins A and D and conglutinin) and can, in addition, directly trigger phagocytosis by macrophages or neutrophils (Kuhlman et al., 1989; Tenner et al., 1995). A third mechanism of complement activation, termed the classical pathway, is initiated by the adaptive immune system. The Fc-parts of antibodies bound to their target structure are recognized by the C1q protein that starts the complement cascade. Here, the innate complement cascade is used by the adaptive immune system for signal amplification (Cooper, 1985). Regardless of the precise mode of complement initiation, opsonized pathogen surfaces are recognized by a set of complement receptors (CR types 1–4) present on professional antigen presenting cells which ensures the phagocytosis of these pathogens. Interestingly, complement receptor type 3 also recognizes certain pathogen surfaces directly via its carbohydrate recognition domain in the absence of complement factors (Cywes et al., 1997; Forsyth et al., 1998; Ehlers, 2000). B. SOLUBLE DEFENSE COLLAGENS Besides the mannose binding lectin, other collagen-like molecules recognize pathogen surfaces. These include the soluble lung surfactant proteins-A and -D, conglutinin and the complement component C1q. The surfactant proteins-A and -D opsonize bacteria via carbohydrate dependent binding for phagocytosis (Eggleton and Reid, 1999). The importance of surfactant proteins for bacterial clearance from the airways is illustrated by increased susceptibilities to bacterial infections in knockout mice lacking SP-A (LeVine et al., 1997). C1q, that does not contain a lectin domain, also recognizes certain bacterial and viral products such as HIV gp41, lipid A in LPS and some bacterial cell walls. Phagocytosis of pathogens opsonized by surfactant proteins-A and -D or mediated by one common receptor on macrophages and neutrophils are termed C1qRp. In addition, one SP-A specific receptor (SPR210) that enhances phagocytosis and cellular cytokine production has been found. However, the precise contribution of surfactant proteins and C1q as opsonins in innate immunity remains to be investigated (Tenner, 1999). C. LECTINS: THE MANNOSE RECEPTOR Macrophages express several membrane-bound lectins for binding and endocytosis of host-derived and microbial ligands. Two of these, the mannose receptor and the complement receptor type 3 are involved in microbial recognition and phagocytosis of many different microbes, as well as in the

54


induction of effector mechanisms (Aderem and Underhill, 1999; Linehan et al., 2000). The mannose receptor is expressed by most macrophages and dendritic cells and contains eight carbohydrate recognition domains that cooperatively bind branched -linked high-mannose containing saccharide structures terminating in mannose, fucose or N-acetylglucosamine moieties (Kery et al., 1992; Engering et al., 1997). It efficiently internalizes soluble and particulate ligands through the endocytic and phagocytic pathways such as highly mannosylated lysosomal enzymes that have been accidentally secreted (Stahl et al., 1978). Recognition of glycoconjugates on the surface of microorganisms by the mannose receptor leads to the phagocytosis of bacteria, fungi, protozoa and viruses such as Candida albicans, Klebsiella pneumoniae, Leishmania donovani, Pneumocystis carinii, Saccharomyces cerevisiae, M. tuberculosis and HIV-1. Involvement of the mannose receptor in endocytosis/phagocytosis has been demonstrated by inhibition of uptake with mono- or polysaccharides such as mannan, but interpretation of the results is compromised by the low specificity of these inhibitors (Linehan et al., 2000). However, using transfected COS cells expressing exclusively the mannose receptor it could be demonstrated that only the transfected cells could recognize and phagocytose C. albicans, K. pneumoniae and P. carinii indicating involvement of the mannose receptor for microbial binding and phagocytosis (Ezekowitz et al., 1990, 1991; Kabha et al., 1995). An important role for the mannose receptor in host defense is further indicated by the fact that the inflammatory cytokine interferon- upregulates mannose receptor-mediated phagocytosis (Marodi et al., 1993). The induction of microbicidal effector mechanisms by mannose receptor engagement has been shown to be complex and depend on multiple parameters such as type, differentiation and activation state of the phagocyte, pathogen and host species as well as involvement of other phagocytic receptors (Lefkowitz et al., 1997; Shibata et al., 1997; Yamamoto et al., 1997; Astarie-Dequeker et al., 1999). A mannose receptor deficient mouse might bring more insight into the role of the mannose receptor in host defense. D. LECTINS: THE COMPLEMENT RECEPTOR TYPE 3 Beside the mannose receptor, the second well characterized phagocyte lectin is complement receptor type 3, a 2-integrin (Ehlers, 2000). This heterodimer (CD11b/CD18) is expressed exclusively on leukocytes and mediates not only recognition and phagocytosis of particles opsonized with the complement component iC3b, but also displays multiple binding sites for a broad spectrum of other endogenous and microbial ligands. Complement receptor type 3 (CR3) is involved in cell spreading, epithelial transmigration and phagocytosis as shown in experiments with CR3/ mice and using blocking antibodies


55

(Plow and Zhang, 1997; Andrew et al., 1998; Balsam et al., 1998; Shang and Issekutz, 1998) and has been shown to bind for example LPS, lipophosphoglycan (LPG) of Leishmania, acylpolygalactosides of Klebsiella and polysaccharides of mycobacteria (Cywes et al., 1997; Forsyth et al., 1998). This receptor plays multiple roles in innate immunity, and as such is known to bind several endogenous and pathogen-derived ligands. Macrophage activation by ligation with endogenous ligands would be detrimental in a nonpathogenic scenario. For that reason, CR3 mediates macrophage activation only in conjunction with additional signals that indicate microbial presence such as co-ligation of the CR3 lectin site or of Fc receptors, or G-protein coupled signals induced by microbial surface molecules and chemoattractants (Jones et al., 1998a). E. TOLL-LIKE RECEPTORS The members of the mammalian family of Toll-like receptors (TLRs) were named after their homology to the Drosophila membrane protein Toll. Drosophila Toll contains a leucine-rich extracellular recognition domain and was identified as playing a role in dorso-ventral patterning during embryogenesis (Schupbach and Wieschaus, 1986). Its ligand, called Spaetzle, is a proteolytic fragment generated by an extracellular cascade of serine proteases (Schneider et al., 1994). The Toll signaling domain is homologous to the interleukin-1 receptor signaling domain. In fact, the overall Toll signaling pathway resembles the mammalian NF-kB pathway induced during inflammatory responses. This led to the identification of a role for Toll also in insect immunity. Triggering of the Toll receptor family members in Drosophila induces the generation of highly fungicidal or bactericidal peptides that are essential effectors of the Drosophila innate immune system (Lemaitre et al., 1996; Hoffmann et al., 1999). Meanwhile, also mammalian TLRs of which 10 are known to date, have been studied in the context of the innate immune system (Krutzik et al., 2001). TLRs are mainly expressed on phagocytic cells and epithelium, and they are involved in the recognition of PAMPs. Following activation, a strong proinflammatory cytokine response is induced as well as other microbicidal effector mechanisms. The first TLR to be implicated in recognition of pathogen-associated molecular patterns was TLR 2 that mediates response to LPS in transfected cell lines. The response to LPS is dependent on the LPS binding protein (LBP) and is enhanced by the GPI-linked LPS receptor CD14 (Yang et al., 1998a) that by itself cannot signal into the cytosol due to the lack of a transmembrane domain. Genetic evidence from LPS-hyporesponsive mice has meanwhile revealed that the physiological responses to LPS are mediated via TLR 4 (Poltorak et al., 1998).

56


Different TLRs respond to different pathogen-derived structures. Lipoproteins, including the 19 kDa lipoprotein of M. tuberculosis, are known to induce strong proinflammatory responses, and do this by triggering TLR 2 (Brightbill et al., 1999). Gram-positive and Gram-negative microbes trigger TLRs 2 and 4 differentially. In the respective knockout mice (Takeuchi et al., 1999), peptidoglycans, present in Gram-positive bacteria, induce innate immune responses in a TLR 2 dependent manner. LPS, present in Gramnegative bacteria, signals via TLR 4. A fine tuning of recognition may be obtained by combination of several TLRs (Ozinsky et al., 2000; Hajjar et al., 2001). Recently, TLR 9 has been identified as mediating the signaling in proinflammatory responses to unmethylated GC-rich bacterial DNA (Hemmi et al., 2000). Despite all the information on TLR involvement in pathogen recognition, the nature of the ligand is unknown. It is still elusive whether TLRs directly bind to the microbial products or whether they are receptors for host-derived proteins that are generated upon pathogen recognition by proper PRRs. The latter is the case in the Drosophila immune response where the Toll–ligand Spaetzle is proteolytically activated upon microbe encounter and then ligates Toll (Levashina et al., 1999). Thus, whether TLRs can be envisioned as true PRRs remains to be determined. F. SCAVENGER RECEPTORS The PRRs on macrophages with the probably broadest specificity are the scavenger receptors (Gough and Gordon, 2000). Most of the known scavenger receptor ligands are characterized by their polyanionic nature (including quartz dust), but not all polyanions are ligands indicating additional structural requirements. Scavenger receptors are multidomained glycoproteins and defined by their ability to bind oxidized or otherwise chemically modified low density lipoprotein thereby being critically involved in the formation of foam cells in artheroschlerotic lesions (Brown and Goldstein, 1990; Suzuki et al., 1997). More recently, scavenger receptors have been implicated in recognizing microbes and their products. In vitro experiments demonstrated, especially for the group of trimeric collagen-like type A scavenger receptors, binding and phagocytosis of lipoteichoic acid, LPS and whole Gram-positive and Gram-negative bacteria including mycobacteria (Hampton et al., 1991; Dunne et al., 1994; Zimmerli et al., 1996). In vivo studies with knockout mice demonstrated a role for type A scavenger receptors in clearance of LPS (Haworth et al., 1997) and of Listeria infections (Suzuki et al., 1997). Nothing is yet known about scavenger receptor-coupling to intracellular effector mechanisms. Macrophages express all of the described receptor molecules designed for recognition of opsonized pathogens or for direct binding of pathogen surface


57

structures. Recognition of the pathogen is normally followed by phagocytosis and intracellular killing. The mechanisms of phagocytosis and the sequential transformation of the phagosome into a microbicidal phagolysosome are described in the next sections. III. Phagocytosis and Intracellular Trafficking in Macrophages

Cells take up extracellular material via different mechanisms. Receptormediated endocytosis (clathrin-coated vesicles with a 100–150 nm diameter) and micropinocytosis (random fluid phase endocytosis in clathrin-coated vesicles) occur in most cell types. In contrast, macropinocytosis (membrane ruffling generating 0.5–2 m vesicles) and phagocytosis (uptake of particulate material which can be as large as the cell itself) only occur efficiently in specialized cells, macrophages, neutrophils and dendritic cells, the so-called professional phagocytes (Fig. 2) (Rabinovitch, 1995). Phagocytosis is a mechanism that arose early in evolution. After the development of multicellularity, organisms had to find means to eliminate and recycle their own cells that became superfluous during ontogeny (tissue

FIG. 2. Possible pathways for internalization of extracellular material. Particulate material such as bacteria is taken up via phagocytosis. In addition, most cells can take up fluid via fluid phase endocytosis, whereas a more specific way to internalize molecules occurs via receptor-mediated endocytosis. After internalization, material is transported along the endosomal/lysosomal pathways where it can be degraded. Some pathogenic microbes subvert this mechanism of degradation. (From Pieters (1999), by permission of JAI Press/Elsevier Science, Stamford, CT.)

58


remodeling) or damaged later on in their life (senescence). Another necessity for the survival of multicellular organisms was the ability to combat parasites. Phagocytic capacity exists already in evolutionary ancient organisms such as sponges (Bergquist and Glasgow, 1986), and more highly developed organisms have retained these functions in a subset of specialized cells until today. Following particle uptake, phagocytosed material has to be transported to degradative organelles, the lysosomes, and hydrolysed completely by protein-, polysaccharide-, lipid- and nucleic acid-processing enzymes. Most research on phagocytosis and the intracellular trafficking of cargo along the endocytic pathway has been done with antibody-, complement-coated or plain inert particles. However, the interaction of a living microorganism with a phagocyte can substantially differ from the pathways delineated for inert particles. A. MECHANISMS

OF

PHAGOCYTOSIS 1. The Role of Actin

Phagocytosis is an actin-dependent process in which ligation of phagocytic receptors induces F-actin polymerization at the site of particle uptake. The mechanistics of actin assembly are now relatively well understood. Central regulators of the actin cytoskeleton are Rac, RhoA and Cdc42, members of the Rho family of GTPases (Hall, 1998; Chimini and Chavrier, 2000). In a series of pioneering experiments Hall and colleagues demonstrated that these proteins were essential in consecutive actin-dependent stages of fibroblast locomotion, namely filopodium formation (Cdc42), lamellipodium formation (Rac1) and stress fiber and focal adhesion assembly (RhoA) (Ridley and Hall, 1992; Ridley et al., 1992; Nobes and Hall, 1995). Filopodia are finger-like extensions, lamellipodia are broad fronted extensions of the plasma membrane, and stress fibers with focal adhesion secure the attachment to a new substratum during locomotion. Using toxin B from Clostridium difficile, a specific inhibitor of all Rho family members, it could be shown that in mouse macrophages these proteins control actin assembly during cell motility and phagocytosis where they accumulate in the activated GTP binding form at the nascent phagosome (Caron and Hall, 1998; Massol et al., 1998). In FcR-mediated phagocytosis, Cdc42 and Rac, but not Rho, control actin assembly by recruiting members of the Wiskott Aldrich Syndrome Protein (WASP) family which, in turn, bind and stimulate the actin nucleation activity of the Arp2/3 complex, a multiprotein complex containing the actin-related proteins 2 and 3 (Machesky and Insall, 1998; Massol et al., 1998; May et al., 2000). Arp2/3 binds to sites of pre-existing actin filaments and nucleates new filaments thereby creating a branched actin network that is believed to push forward the leading edge of membranes. Indeed, it was demonstrated that Arp2/3 and WASP


59

are recruited to the site of phagosome formation. Engagement of FcRs during pseudopod extension would continuously trigger membrane propulsion via recruitment of activated Rho family GTPases and subsequent actin polymerization. The local control of actin polymerization by ligandreceptor interaction along the particle surface is the hallmark of the ‘‘zippering’’ model of phagocytosis (Swanson and Daer, 1995). Using dominant negative forms of Rho family members the contributions of Rac1 and Cdc42 in phagocytosis could be separated, revealing a crucial role for Rac1 in the pseudopod extension process, whereas Cdc42 is essential for the final stages of phagosome formation, namely distal contraction and closure of the membrane protrusions. Dominant negative RhoA had no effect on FcR-mediated phagocytosis (Cox et al., 1997; Caron and Hall, 1998; Massol et al., 1998). Closure is also sensitive to inhibitors of phosphatidyl-inositol3-kinase (PI3K) and myosins (Swanson et al., 1999). Indeed, several myosins are found on the nascent phagosome as well as the serine/threonine kinase PAK-1, a regulator of the actomyosin cytoskeleton that is activated by Rho family members (Dharmawardhane et al., 1999; van Leeuwen et al., 1999). How Rho family GTPases are recruited and activated during phagocytosis is not known. Generally, an exchange of GDP for GTP by guanine nucleotide exchange factors is the crucial step in GTPase activation. Consequently, a Rho specific guanine nucleotide exchange factor must lie in the transduction pathway between initial receptor triggering and recruitment of the Rho proteins. The missing link might be provided by the proto-oncogene Vav that is a guanine nucleotide exchange factor, which activates multiple Rho family members (Bustelo, 2000). 2. Signaling in FcR-Mediated Phagocytosis The signaling pathways involved in the various stages of phagocytosis are only partially understood. In the case of Fc R-mediated phagocytosis, receptor cross-linking by IgG opsonized particles initiates tyrosine phosphorylation of so-called immunoreceptor tyrosine-based activatory motifs (ITAMs) that are present in the cytoplasmic signaling parts of the receptors by a member of the Src kinase family (Greenberg, 1995). For further signaling, recruitment and phosphorylation of the non-receptor tyrosine kinase Syk is essential. This kinase associates with the ITAMs of the Fc R via its SH2 domain and triggers not only actin-based phagocytosis, but also gene transcription, cytokine release and initiation of oxidative burst in macrophages (Aderem and Underhill, 1999). The link between this initial ligation induced Syk-recruitment and the Rac1- and Cdc42-regulated mechanistics of actin polymerization is not precisely known. The lipid kinase PI3K, however seems to play a role early in

60


signal transduction. Phosphatidyl-inositol-3-kinase inhibitors or deficiency in either of the three molecules Syk, amphyphysin II or dynamin 2, result in a failure in actin-dependent pseudopod extension and phagosome closure during FcR-mediated phagocytosis (Araki et al., 1996; Crowley et al., 1997; Toker, 1998; Gold et al., 1999, 2000). Amphyphysin II and dynamin 2 are recruited to the nascent phagosome by activated phosphatidyl-inositol3-kinase. However, Syk, PI3K, amphyphysin II and dynamin 2 are not required for actin assembly below the site of particle binding indicating the existence of still unknown pathways signaling initial actin polymerization. In this context it is notable that PI3K is activated by many receptor linked-tyrosine kinases and that its reaction products strongly accumulate at the nascent phagosomal membrane. The GTPase dynamin 2 is also involved in vesicle scission in the TGN and in pinching off vesicles in clathrinmediated endocytosis (Takei et al., 1995; Jones et al., 1998b; Schmid et al., 1998). As mentioned before, Vav could serve as a guanine nucleotide exchange factor for the Rho proteins regulating actin assembly at the nascent phagosome. For activation, Vav has to be tyrosine-phosphorylated and Vav is also stimulated by the products of PI3K that bind to the pleckstrin homology domain of Vav (Crespo et al., 1997; Han et al., 1998). Indeed, during FcRmediated phagocytosis, Vav is tyrosine-phosphorylated (Kiener et al., 1993) and therefore could link Syk/PI3K activity and activation of the Rho family proteins. 3. Complement Receptor Type 3-Mediated Phagocytosis Electron microscopic comparison of FcR-mediated and complement receptor type 3-mediated particle ingestion has revealed morphologic differences that might reflect the differential use of Rho family members in actin assembly. FcR engagement is accompanied by the formation of membrane protrusions that extend along the IgG-coated particle and fuse on their distal end, involving, as mentioned, the regulators Rac and Cdc42. In contrast, complement receptor type 3-mediated uptake was found to occur in the absence of membrane protrusions. Instead, complement opsonized particles sink into the cell (Kaplan, 1977). Complement receptor type 3, actin and its nucleation complex Arp2/3 and other actin-binding proteins are located in focal contact points with the internalized particle distributed along the phagosomal membrane and resembling integrin-mediated focal adhesions (Allen and Aderem, 1996). The different phenotype of complement receptor type 3-mediated phagocytosis is not regulated by Rac, but by Rho which is known to regulate integrin-based sites of focal adhesions (Caron and Hall, 1998; Schoenwaelder and Burridge, 1999). The signaling events following complement receptor type 3 ligation that lead to the described distinct phenotype of the phagocytosis process are far


61

less clear than those in FcR-mediated phagocytosis. The integrin complement receptor type 3 does not contain any ITAMs and a direct coupling to kinase activity is not known. Tyrosine kinase inhibitors do not interfere with complement receptor type 3-mediated phagocytosis, inhibitors of protein kinase C, however do (Allen and Aderem, 1996). This indicates different initial signaling pathways for complement receptor type 3- and FcR-mediated phagocytosis. The requirement for protein kinase C signaling in complement receptor type 3-mediated phagocytosis was proposed to be due to a protein kinase C-dependent cytoplasmic phosphorylation of complement receptor type 3 that makes the integrin competent for phagocytosis. Such complement receptor type 3 activating stimuli can be provided by tumor necrosis factor-, phorbolesters, fibronectin-coated surfaces inducing PI3K activity or by G-protein-coupled mechanisms induced by chemoattractants independently of PI3K (Jones et al., 1998a). This integrin activation from inside the cell is a common phenomenon observed for members of this protein family and termed inside-out signaling. In addition, PI3K involvement in complement receptor type 3-mediated uptake has been deduced from the fact that complement receptor type 3 triggering induced phosphatidylinositol (3,4,5)triphosphate generation, a product of PI3K. In contrast to the FcR, complement receptor type 3 operates in many different processes such as cell adhesion, phagocytosis and transmigration (Ehlers, 2000), and thus its signaling depends probably on the precise physiological situation. The fact that complement receptor type 3 is not a pure defense receptor might be beneficial for microbes in allowing a less detrimental entry into phagocytes. Indeed, phagocytosis mediated by complement receptor type 3 alone does not generate an oxidative burst (Wright and Silverstein, 1983; Berton et al., 1992; Zhou and Brown, 1994) and complement receptor type 3 ligation has been shown to suppress interleukin12 secretion (Marth and Kelsall, 1997). B. DEGRADATION

IN THE

ENDOSOMAL-LYSOSOMAL PATHWAY

After phagocytosis, the microbe resides within a phagosome, enclosed by a host cell surface-derived membrane. Under normal circumstances, the phagosome undergoes transformation into a phagolysosome by membrane fusion and fission events that enrich the phagosome with degradative lysosomal enzymes while recycling plasma membrane components to the cell surface. All transport between the membrane-enclosed compartments and organelles within a cell is performed by different types of vesicles. In the exocytic pathway, newly synthesized luminal and transmembrane proteins are transported from the site of biosynthesis, the rough ER, through the stacks of the Golgi complex to the trans-Golgi network. Here, sorting into vesicles

62


destined for the plasma membrane, for intracellular storage or for the endosomal system occurs (Griffiths and Simons, 1986). In the endocytic pathway, internalized material such as nutrients, hormones or ingested particles, is transported to a sorting station, the early endosome, where separation of recycling plasma membrane components such as receptors from their ligands is achieved (Mellman et al., 1986). After the recycling step, the remaining endocytosed material is shuttled to the lysosomes and hydrolyzed (Fig. 2). 1. Principles and Components of Vesicular Traffic Vesicular transport is based on two principles that have emerged over the past 25 years. First, the correct cargo has to be sorted into the appropriate vesicle, and second, the vesicle formed has to be specifically targeted to the correct acceptor compartment. Biochemical and genetic approaches revealed that membranes carry specific protein complexes defining the vesicle type (Trimble et al., 1988; Baumert et al., 1989; Bennett et al., 1992; Sollner et al., 1993). Central to these complexes are the SNARE and the Rab-GTPase proteins. SNAREs come in two flavors, the target (t-) SNAREs that are present on the acceptor membrane and the vesicle (v-) SNAREs on the donor, i.e., vesicle membrane. Preceding fusion of membranes, a macromolecular protein complex is generated by coiled-coil interaction of specific v-SNAREs and t-SNAREs (Hanson et al., 1997; Sutton et al., 1998). This complex also includes organelle specific Rab proteins. Different cellular membrane and vesicle populations have different associated Rab GTPases (Chavrier et al., 1990), which are thought to increase the fidelity of membrane fusion events (Pfeffer, 1999). Tight docking via SNARE interactions is then followed by ATP-dependent membrane fusion in which the vacuolar type proton ATPase might be involved (Peters et al., 2001). The second requirement in vesicle transport is selectivity of vesicles for their cargo. This is accomplished by soluble cytoplasmic coat proteins that assemble in a lattice-like structure at the donor membrane and are thought to generate the forces for the membrane budding process. Cargo selection is performed by the assembling coat that possesses binding affinities for the cytoplasmic tails of its cargo, thereby allowing its active enrichment within the budding vesicle. After the budding event that is regulated by small GTPases, the coat dissociates from the vesicle membrane liberating the vesicle-specific v-SNAREs for docking and fusion with the target membrane. Three main vesicle coats based on: (1) clathrin, (2) COPI, and (3) COPIIoperating on different intracellular routes have been identified so far (Schekman and Orci, 1996; Kirchhausen, 2000; Pelham and Rothman, 2000). The first vesicle type to be described and biochemically characterized was the clathrin-coated vesicle, the only known peripheral coat involved in


63

endocytosis up to now (Heuser and Reese, 1973; Pearse, 1975). Originally described in oocytes and neurons, the ‘‘vesicles in baskets’’ (Kanaseki and Kadota, 1969) function in the receptor-mediated uptake of soluble molecules such as hormones (insulin), nutrients (LDL) and iron (transferrin). Clathrin forms pentagonal–hexagonal arrays on the cytoplasmic side of budding vesicles with fixed diameters between 100 and 150 nm. Mainly based on analysis of low density lipoprotein receptor internalization and recycling (Goldstein et al., 1979), three principles in clathrin-mediated endocytosis have become clear (Riezman et al., 1997; Marsh and McMahon, 1999). First, the cargo-specific receptors contain a motif in their cytoplasmic domain that is specifically recognized by a heterotetrameric clathrin adaptor protein (AP-2) that connects receptors with the clathrin coat (Robinson, 1989). Two motifs have been delineated and follow the consensus YXX or are dileucine-based (Rapoport et al., 1998; Bonifacino and Dell’Angelica, 1999). Second, the receptors are re-utilized several times by recycling to the cell surface in coated vesicles and third, the dissociation of receptor and ligand required for receptor re-utilization is a consequence of a pH-sensitive receptor ligand interaction. The slightly acidic (about pH 6.5) lumen of the early endosome represents the sorting organelle in the endocytic pathway in which receptors are uncoupled from their ligands and sorted into clathrin vesicles that recycle to the cell surface. The separated ligands are then shuttled via vesicular transport to late endosomes and lysosomes for degradation. The final scission event in clathrin-mediated endocytosis is dependent on the GTPase dynamin (van der Bliek and Meyerowitz, 1991; Takei et al., 1995). Clathrin-coated vesicles also have roles in the exocytic pathway where they are involved in the transfer of endosomally/lysosomally resident proteins from the Golgi complex to the respective organelles. Lysosomal hydrolases, for example, are sorted in the TGN into coated vesicles destined for fusion with lysosomes. Different clathrin adaptor proteins function in different transport steps. As mentioned, endocytosis from the plasma membrane exclusively makes use of AP-2 whereas in vesicular budding at the level of the early endosome and the TGN AP-1, AP-3 and AP-4 are in operation (Hirst and Robinson, 1998; Dell’Angelica et al., 1999; Hirst et al., 1999). The initial stages of the exocytic pathway are governed by other coat complexes, COPI and COPII. COPII coats assemble at the cytoplasmic side of the ER and operate in ER to cis Golgi transport (Bednarek et al., 1995). Whether COPII cargo that consists of all secreted, endosomal and membrane proteins has sorting determinants is not clear yet, but certainly pre-budding sorting mechanisms exist based on ER resident chaperones and glycosyltransferases that retain malfolded cargo from being packed into budding vesicles. In contrast to COPII vesicles, COPI vesicles, that mainly

64


mediate retrograde transport of cargo from the Golgi to the ER or transport within the Golgi, do recognize defined sorting determinants (Letourneur et al., 1994; Cosson and Letourneur, 1997). RRXX and KKXX motifs that are present in the cytoplasmic tails of many ER resident proteins have affinity for COPI coats and allow recycling of proteins that have escaped the ER. Endoplasmac reticulum luminal proteins contain KDEL motifs that are similarly recycled by a transmembrane KDEL receptor (Lewis and Pelham, 1992) that, in turn, has affinity to COPI coats. The assembly of COP coats on donor membranes is regulated by small GTPases (Arf). These GTPases become activated by membrane-bound guanine nucleotide exchange factors and serve together with the cytoplasmic tails of cargo proteins as docking sites for the appropriate coat proteins (Springer et al., 1999). 2. Phagosome Maturation The described mechanisms of vesicle budding and fusion are likely to operate also during the stepwise maturation of the phagosome. In recent years, a number of markers specific to different compartments within the endocytic pathway have been used to characterize the changes that occur during phagosome maturation in more detail. As the phagosome matures, proteins that initially are part of the newly formed compartment disappear and are replaced, first by proteins present in early endosomes, such as Rab5, the transferrin and mannose receptors and LDLs, and at a later stage by markers of late endosomes such as Rab7, Rab9, the vacuolar type proton ATPase and lysosomal membrane glycoproteins (LAMPs) (Pitt et al., 1992; Rabinowitz et al., 1992; Desjardins et al., 1994; Beron et al., 1995; Oh and Straubinger, 1996). The precise function of LAMPs is still unclear, but it has been suggested that they may protect lysosomal membranes from degradation by lumenal enzymes (von Figura and Hasilik, 1986). Finally, after further enrichment with active hydrolases such as cathepsins, lysosomal acid phosphatases and -glucuronidase, the phagosome resembles a degradative lysosome. The rate of phagosome maturation can vary considerably depending on the ingested particle. During FcR-mediated phagocytosis of antibody-coated S. aureus particles, the phagosome acquires lysosomal markers after 15 min. In contrast, phagosomes containing latex beads can remain non-fused with lysosomes for hours (Pitt et al., 1992; de Chastellier and Thilo, 1997). It has been demonstrated using beads with varying surface properties that high surface hydrophobicity of the ingested particle decreases the rate of phagosome maturation (de Chastellier and Thilo, 1997). As mentioned, specificity in membrane fusion is accomplished by the action of v- and t-SNAREs and the Rab family of low molecular weight GTPases.


65

During phagosome maturation, the small GTPases Rab5 (early endosomal marker), Rab7 (late endosomal marker) and Rap1A/B (lysosomal markers) associate sequentially with the phagosomal membrane (Desjardins et al., 1994; Pizon et al., 1994). Since homotypic fusion of early endosomes is dependent on the presence of Rab5 on both organelles (Gorvel et al., 1991), it is likely that Rab5 also facilitates fusion of the early phagosome with early endosomal compartments. Loss of Rab5 is paralleled by acquisition of Rab7 which is then exchanged by Rap1. In analogy to Rab5, the appearance of Rab7 and then Rap1 on the maturing phagosome may indicate their fusogenic capacity with late endosomes and lysosomes, respectively (Feng et al., 1995). The consequence of Rab-controlled fusion is that the membrane composition and contents of maturing phagosomes reflect the composition in the endocytic pathway as described above for several markers. The presence and absence of the various described endocytic markers in vacuoles containing replicating pathogens has also given insight into how microbes like M. tuberculosis, Legionella pneumophila, Chlamydia trachomatis, or Salmonella typhimurium survive in the hostile environment of a macrophage.

IV. Intracellular Survival of Bacterial Pathogens

Once inert particles and microbes have been internalized into a phagosome by macrophages, this organelle matures into an acidic phagolysosome rich in lysosomal hydrolases. To escape degradation, pathogenic bacteria have evolved several strategies. First, some pathogens avoid phagocytosis by professional phagocytes and rather enter other, less microbicidal cells. Second, other microbes, once within the macrophage phagosome, divert the phagosome from the maturation pathway and convert it into a hospitable vacuole. A third way of surviving is the escape from the endosomal/lysosomal system by lysing the phagosomal membrane and replicating in the cytoplasm of the host cell (Finlay and Cossart, 1997; Meresse et al., 1999). A. AVOIDING

THE

MACROPHAGE

To avoid the macrophage, the cell type that is most efficient in phagocytosis, some bacteria actively induce their uptake by non-professional phagocytes. This process, termed invasion, is dependent on receptor ligation and remodeling of the host cell actin cytoskeleton. Two main types of entry have been observed, the ‘‘zipper’’ mechanism, where bacterial ligands and host cell receptors interact along the whole surface of the pathogen forming a closely fitting phagosome (Yersinia), and the ‘‘trigger’’ mechanism where the microbe induces host cell macropinocytosis that ultimately leads to passive enclosure of

66


the pathogen into a spacious vacuole (Shigella, Salmonella) (Swanson and Daer, 1995). Yersinia enterocolitica, causing severe enterocolitis with fever, pain and diarrhea, replicates extracellularly in micro-colonies in the Peyers patches, which are the lymphatic organs underlying the intestine (Carter and Collins, 1974). To reach this place it induces its uptake by the intestinal M cells, which it then transmigrates. Yersinia expresses the surface molecule invasin that binds to a subset of 1-integrins and mediates uptake via a ‘‘zipper’’ mechanism (Isberg and Leong, 1990). Phagocytosis by professional phagocytes is lethal for Yersinia and is in fact prevented by the microbe using a contact-induced type III secretion system, a needle-like complex of a single oligomerized 6 kDa protein (Hoiczyk and Blobel, 2001). This complex is thought to span the bacterial inner and outer membranes and upon contact with phagocytes injects several Yop proteins into the cytoplasm of the host cell. YopE paralyses the host cell’s actin cytoskeleton by unknown mechanisms (Rosqvist et al., 1990, 1994; Cornelis and Van Gijsegem, 2000). YopH (Black and Bliska, 1997), a broad-spectrum tyrosine phosphatase, dephosphorylates host signaling molecules thereby abrogating phagocytosis signals, and YopO is a putative serine/threonine kinase of unknown function. YopJ binds to family members of the mitogenactivated protein kinase kinase (MAPKK) superfamily and prevents their phosphorylation, thereby inhibiting the NF-kB pathway for secretion of proinflammatory cytokines by macrophages (Palmer et al., 1998; Schesser et al., 1998). In addition YopJ, which belongs to the family of ubiquitin-like protein cystein proteases, induces apoptosis in macrophages via its proteolytic activity (Mills et al., 1997; Monack et al., 1997; Denecker et al., 2001). Salmonella typhimurium and Shigella flexneri enter intestinal epithelial cells by a ‘trigger’ mechanism which involves induction of extensive membrane ruffling by the contact-induced injection of bacterial factors (Francis et al., 1993; Alpuche-Aranda et al., 1994; Dehio et al., 1995). More than 20 components of the needle-like type III secretion apparatuses and most of the secreted effector molecules are encoded within the Salmonella pathogenicity island I (SPI-I) or the large Shigella virulence plasmid, respectively (Menard et al., 1996; Ochman et al., 1996; Kubori et al., 1998; Blocker et al., 1999). Salmonella enters epithelial cells from the apical side via an unknown receptor. Upon host cell binding, Salmonella injects SopE, a Cdc42- and Racspecific GTPase activating protein into the host cell thereby inducing the extensive actin remodeling necessary for membrane ruffling (Hardt et al., 1998). After internalization, Salmonella remains in the vesicular system. The phagosome in epithelial cells in which Salmonella replicates after a lag phase of few hours does not interact with the early endosomal system.


67

The Salmonella-phagosome acquires the vacuolar type proton ATPase and lysosomal glycoproteins, but excludes the mannose-6-phosphate receptor and the bulk of lysosomal hydrolases (Garcia-del Portillo and Finlay, 1995; SteeleMortimer et al., 1999). During onset of replication Salmonella induces and interconnects with a tubular network of membranes positive for lysosomal glycoproteins such as LAMP-2 (Garcia-del Portillo et al., 1993). The function of this network is unclear to date. Salmonella also triggers its uptake by macrophages in which it can survive with the help of an additional type III secretion system encoded on SPI-2 (Hensel et al., 1998) whose mode of action is still under examination. The receptor for Shigella is the 5 1integrin (Dehio et al., 1995). As this integrin is only expressed on the basolateral membrane of epithelia, Shigella has to transmigrate through M cells before it can invade the epithelium. During phagocytosis Src kinase is activated since inhibitors of tyrosine kinases abolish uptake. Unlike Salmonella, Shigella requires Rho for internalization (Adam et al., 1996). For entry, Shigella injects the virulence factors Ipa A, B, C and D into the host cell (Nhieu and Sansonetti, 1999). Ipa B and C assemble into a complex that forms a pore in the host membrane to allow entry of Ipa A that in turn binds to vinculin, an actin-binding protein and component of the nascent phagosome, and in doing so depolymerizes F-actin (Bourdet-Sicard et al., 1999). At a later stage, Ipa B also lyses the phagosomal membrane allowing Shigella to escape from the degradative lysosomal pathway into the cytosol (High et al., 1992). A well-studied Gram-positive bacterium entering a wide range of host cells is the food-borne pathogen Listeria monocytogenes (Hurme and Cossart, 1999; Cossart and Bierne, 2001). Two independently functioning bacterial surface molecules can mediate entry into host cells. The surface molecule internalin A is the ligand for E-cadherin, a transmembrane molecule involved in homotypic interaction of epithelia. Internalin A is sufficient to promote entry of Listeria into epithelial cells (Mengaud et al., 1996). The second bacterial protein mediating entry into a broader variety of cells (Parida et al., 1998) is internalin B for which two host receptors are known, the peripheral associated membrane protein gC1qR (Braun et al., 2000) and the receptor tyrosine kinase Met (Shen et al., 2000). In all cases, internalization occurs by a ‘‘zipper’’ mechanism and involves PI3K activation (Ireton et al., 1996). Like Shigella (and Rickettsia), Listeria also lyses its phagosomal membrane after phagocytosis and escapes into the cytosol (Gouin et al., 1999). To that end, Listeria secretes the pore-forming toxin listeriolysin O and phospholipases that rupture the phagosomal membrane. Interestingly, listeriolysin O is only active at mildly acidic pH, and therefore Listeria requires limited acidification of the phagosome to escape (Geoffroy et al., 1987). All three phagosomerupturing pathogens apply the same propelling mechanism in the host cytosol

68


by polymerizing actin at their rear side that can be observed in vivo and in vitro by fluorescence microscopy as an actin comet (Tilney and Portnoy, 1989). The barbed ends of the actin comet are directed toward the bacterium. At the interface of bacterium and comet a single bacterial protein is sufficient to induce prolongation of the comet by addition of actin monomers (Dramsi and Cossart, 1998). In the case of Listeria a protein called ActA recruits the host actin nucleation and branching complex Arp2/3. Interestingly, ActA is the bacterial equivalent to the cellular WASP-proteins that normally recruit and activate the Arp2/3 complex (Welch et al., 1998; Loisel et al., 1999). ActA also interacts with the vasodilatator-stimulated phosphoprotein (VASP) that has been shown to induce formation of actin bundles at focal adhesion points (Drees et al., 2000). This second type of actin polymerization probably plays a role in Listeria infection as well, because mutation of the VASP binding site on ActA leads to strong reduction in motility (Geese et al., 2000). The intracellular actin-based motility is also the basis for cell-to-cell spreading which as a result avoids contact with extracellular antibodies or complement. Listeria and Shigella infect neighboring epithelial cells by propelling forward against the plasma membrane forming invaginations into the next cell that are then phagocytosed. This results in a double membrane enclosed vesicle in the next cell that is subsequently lysed by the bacteria (Gedde et al., 2000). Together, these examples illustrate to what extent microbes go to prevent phagocytosis and destruction by macrophages. B. SURVIVING

IN THE

MACROPHAGE

Other microbes do not avoid phagocytosis by macrophages, but, once internalized, reside in remodeled vacuoles that neither fuse with lysosomes nor acidify. This strategy is employed by several species of Mycobacterium, by Legionella pneumophila, C. trachomatis, Brucella abortus and Toxoplasma gondii. Alternatively, some microbes are capable of surviving in acidic phagolysosomes (Coxiella burnetti, Leishmania spp.). Employing a third type of mechanism, Listeria, as in epithelial cells, escapes the macrophage phagosome by listeriolysin O-mediated lysis. Interestingly, listeriolysin O-defective mutants that remain within their phagosome retard phagosomal maturation, which is characterized by prolonged retention of Rab5a (AlvarezDominguez et al., 1996). This retardation of phagosome maturation might be of importance for Listeria in allowing more time for phagosome lysis. Toxoplasma gondii converts its phagosome into a vacuole that is essentially devoid of host cell proteins and thus, in the absence of SNAREs and Rab proteins, truly non-fusogenic (Suss-Toby et al., 1996). This status is maintained even if the parasite is killed after its entry. However, FcRmediated phagocytosis of antibody-opsonized parasites leads to rapid phagosome–lysosome fusion, corroborating the idea that receptor involvement


69

in phagocytosis can decide on subsequent phagosome trafficking (Joiner et al., 1990). T. gondii modifies its vacuolar membrane further by exocytosis. The vacuole associates with the host ER and mitochondria, and it acquires pores to allow for import of small host molecules such as ATP (Schwab et al., 1994). Legionella- and Chlamydia-containing vacuoles are associated with the biosynthetic pathway. The major outer membrane protein (MOMP) of Legionella efficiently fixes a complement factor leading to a complement receptor type 3-mediated phagocytosis (Bellinger-Kawahara and Horwitz, 1990). Legionella induces a unique phenotype of phagocytosis in which membrane protrusions wrap around the microbe (coiling phagocytosis) (Horwitz, 1984), however the induction mechanisms are unknown. The phagosome only slightly acidifies (pH 6.1) (Horwitz, 1983; Horwitz and Maxfield, 1984), associates with the ER and mitochondria and ultimately becomes ribosome studded (Swanson and Isberg, 1995). Furthermore, the vacuole lacks the transferrin receptor and does not interact with late endosomes/lysosomes (Vogel and Isberg, 1999). The Legionella icm/dot virulence genes that form a putative transmembrane type IV secretion system are required for vacuole biogenesis. Phagosomes containing Legionella with mutations in the transmembrane protein DotA (a putative ABC transporter) rapidly become Rab7 and LAMP positive and do not allow Legionella survival (Roy and Isberg, 1997; Roy et al., 1998). Chlamydia trachomatis resides in a vacuole that interacts with Golgiderived vesicles that provide sphingomyelin and nutrients (Scidmore et al., 1996). All host-derived proteins, however, are excluded from the membrane that acquires chlamydial proteins instead. How Chlamydia locates its vacuole in the biosynthetic pathway is not known, but several chlamydial proteins in the vacuolar membrane and a recently identified putative type III secretion system may be involved in this process (Hsia et al., 1997). The protozoan parasite L. donovani represents a microbe that resides in different vacuoles depending on its life cycle stage. The promastigote form occupies a phagosome with poor fusogenic capacity towards endosomes. This has been attributed to the selective exclusion of Rab7 from its phagosome (Scianimanico et al., 1999). The envelope molecule LPG has been shown to, at least in part, reduce fusogenicity with lysosomes (Desjardins and Descoteaux, 1997). Upon intraphagosomal transformation of the promastigote to the amastigote form, Leishmania downregulates LPG and phagosome–lysosome fusion is restored, however without affecting the viability of the parasite (Russell et al., 1992). Another parasite that can withstand the hostile environment of a phagolysosome is Coxiella burnetii. The replication of this obligate intracellular Gram-negative bacterium is dependent on transfer to lysosomes that are believed to provide the necessary nutrients (Heinzen et al., 1996).

70


The acidic Coxiella phagosome, in fact, does not display any restrictions toward fusion with endosomal organelles as it even fuses with mycobacterial phagosomes that normally remain segregated from acidic compartments (Gomes et al., 1999b). Interestingly, the low pH of 5.2 in the acidic vacuole protects Coxiella from antibiotics that lose their activity at low pH (Maurin et al., 1992). V. Mycobacterial Infection

In terms of number of infected people, the most successful pathogenic bacteria are members of the M. tuberculosis complex (Bloom, 1992; Murray and Salomon, 1998). Important members of this complex are M. tuberculosis itself, and the lesser virulent species Mycobacterium bovis, Mycobacterium avium and Mycobacterium intracellulare. According to estimates by the World Health Organization, one third of the global population is infected with M. tuberculosis and in the year 2002, approximately 3 million people, mainly in developing countries, will succumb to this disease, making it the leading cause of death from bacterial infectious diseases (Kaufmann, 2000) (see http:// www.stoptb.org/tuberculosis/#facts.html). Tuberculosis is transmitted via inhalation of droplets coming from a coughing person with the active disease. In a newly infected immunologically competent host, macrophages and macrophage-activating T cells quickly control the infection by containing it in a so-called granuloma which is a histologically defined aggregation of activated and infected macrophages surrounded by T cells, mainly of the CD4 þ subtype (Kindler et al., 1989). Contained in a granuloma, mycobacteria do not proliferate but can remain viable in a dormant state for decades and can become reactivated to proliferate upon loss of host immune competency as occurs in HIV infections, after organ transplantations or under conditions of malnutrition (Chan et al., 1996; Lim et al., 1999). Loss of immune control leads to granuloma liquification and dissemination of the bacteria within the lung and to other organs such as liver and kidney ultimately causing multi-organ-failure and death (Davis et al., 1990). Pathogenic mycobacteria successfully persist in their host because they can survive phagocytosis by macrophages. In fact, macrophages, instead of killing mycobacteria, become their hospitable host cells. The intracellular survival strategy of mycobacteria is thought to be based on their ability to prevent fusion of the phagosome they inhabit with macrophage lysosomes (Armstrong and Hart, 1971). Many studies have shown that the phagosome in which mycobacteria survive and replicate actually displays features of an early endosomal compartment (Russell, 1995; Hasan et al., 1997; Via et al., 1997). The control of infection that occurs in an immuno-competent individual is


71

dependent on T helper 1 cytokines such as interferon- that activate macrophage bactericidal mechanisms (Flesch and Kaufmann, 1987; Schaible et al., 1998). In accordance, mice and humans with genetic defects in interferon- signaling are highly susceptible to mycobacterial diseases (Huang et al., 1993; Jouanguy et al., 1997). The discovery that mycobacteria inhibit phagosome–lysosome fusion was made in the 1970s. However, the actual mechanism of inhibition of phagosome–lysosome fusion has remained unknown. A. THE MYCOBACTERIAL PHAGOSOME Armstrong and Hart (1971) discovered using electron microscopy that phagosomes containing live M. tuberculosis do not fuse with lysosomes. Mycobacterial phagosomes resist lysosomal fusion, but by no means are inert and incapable of membrane fusion. Mycobacterial phagosomes continuously remain accessible to extracellular fluid phase markers such as horseradish peroxidase (HRP) (Frehel et al., 1986) and plasma membrane molecules (Clemens and Horwitz, 1996). Exogenously added transferrin is within minutes transported into the phagosome and can be chased out equally fast. Also the glycosphingolipid GM1 steadily cycles from the plasma membrane through the mycobacterial phagosome (Russell et al., 1996). Furthermore, mycobacterial phagosomes may contain large amounts of MHC I and MHC II molecules that are removed from these organelles within days (Clemens and Horwitz, 1995). Thus, the mycobacterial phagosome is in constant exchange with the plasma membrane and the early endocytic organelles. The mycobacterial phagosome is not devoid of, but low in lysosomal markers such as LAMP-1 and -2, CD63 and cathepsins B, L and D, the latter being present in their immature intermediate forms (Sturgill-Koszycki et al., 1996). The mannose-6-phosphate receptor, which transports lysosomal hydrolases to late endosomal compartments, is absent, indicating an alternative delivery pathway for the cathepsins present in phagosomes, possibly via the plasma membrane (Xu et al., 1994). Together, these data indicate that the mycobacterial phagosome is positioned early in the endocytic pathway. In contrast to the mycobacterial phagosome, latex bead-containing phagosomes steadily acquire lysosomal markers and cathepsins in their mature active forms reflecting the rapid transfer through the endocytic default pathway to lysosomes (Clemens and Horwitz, 1995). As mentioned earlier, membrane fusion events along the endocytic pathway are regulated by small GTPases of the Rab family. Mycobacterial phagosomes containing M. bovis BCG are Rab5 positive, but Rab7 negative, which is in agreement with a phagosomal maturation arrest at the early endosomal stage (Via et al., 1997).

72


Another characteristic of the mycobacterial phagosome is its limited acidification. Using pH-sensitive dyes several groups showed that M. avium or M. tuberculosis containing phagosomes do not acidify below pH 6–6.5 (Crowle et al., 1991; Sturgill-Koszycki et al., 1994) whereas latex beads are shuttled to compartments with an approximate pH of 5. The compromised acidification has been attributed to the exclusion of the vacuolar type proton ATPase from the mycobacterial phagosome (Sturgill-Koszycki et al., 1994; Xu et al., 1994). B. PHAGOCYTIC RECEPTORS

FOR

MYCOBACTERIA

Mycobacterium tuberculosis is recognized by at least nine receptors on phagocytes including complement receptors of the types 1, 3, and 4, the mannose receptor, CD14, scavenger receptor type A and surfactant protein receptors (Ernst, 1998). M. tuberculosis uses at least two mechanisms for becoming opsonized by C3b/C3bi. On one hand it activates the alternative complement pathway (Schlesinger et al., 1990). On the other hand pathogenic mycobacteria have developed a unique method to trigger their complement opsonization. To that end, the microbes capture the complement component C2b from serum, combine it with a factor from their own bacterial surface and as a result a unique C3 convertase is formed that subsequently opsonizes the microbe (Schorey et al., 1997). In addition to achieving C3bi opsonization, mycobacteria use two distinct docking sites on the complement receptor type 3, the C3bi binding site and the lectin site that binds capsular polysaccharide without prior opsonization (Cywes et al., 1997). Employing CHO transfectants expressing complement receptor type 3, different preferences of different mycobacterial strains for non-opsonic binding or complement-mediated binding were observed, however the consequences of using one or the other binding site on complement receptor type 3 for entry into macrophages are not known (Cywes et al., 1997). In this context it should be noted that simultaneous ligation of both binding sites on complement receptor type 3 can lead to macrophage activation (Xia et al., 1999). Preferential usage of complement receptors by mycobacteria for entry might be one way to manipulate the host response as has been found for other pathogens (Mosser and Edelson, 1987; Da Silva et al., 1989). Other receptors used for mycobacterial entry include the mannose receptor that can recognize lipoarabinomannan. Lipoarabinomannan is one of the abundant surface molecules on M. tuberculosis that contain terminal mannose residues (Schlesinger et al., 1990, 1996). The expression of the mannose receptor is downregulated in activated macrophages and thus might play a role mainly in the initial stages of mycobacterial infection (Schreiber et al., 1993). It has been shown that the mannose receptor transports lipoarabinomannan to the endosomal system for loading of CD1b molecules and thus might


73

contribute to recognition of infected macrophages by T-lymphocytes (Prigozy et al., 1997). As mycobacteria are transmitted via aerosol, the primary site of mycobacterial infection is the lung. For this reason, the role of the opsonizing lung surfactant proteins in mycobacterial entry into macrophages has been analyzed. Surfactant protein-A can function as an opsonin, however, the contribution to the anti-mycobacterial response is unknown (Downing et al., 1995; Gaynor et al., 1995). To date it cannot be excluded that different receptors lead to a different fate of the ingested mycobacteria. A study selectively blocking complement receptors type 1, 3 and 4, the mannose receptor and scavenger receptor type A on human macrophages did not reveal any apparent differences in intracellular survival and growth of the Erdman strain of M. tuberculosis (Zimmerli et al., 1996). However, receptor usage might influence the outcome of an infection more indirectly by differentially inducing cytokine secretion or by resulting in different intracellular trafficking and antigen processing/ presentation events. In summary, the available evidence to date would suggest that the different entry pathways do not have any effect on intracellular survival and growth of mycobacteria, and for mycobacterial success it might be only important—via one of the many receptors—to get into the cell from where the mycobacteria can control their fate by mechanisms not related to receptor usage. C. HOST MECHANISMS CONTROLLING INTRACELLULAR MYCOBACTERIA The control of mycobacterial infections requires both the action of CD8 þ and CD4 þ T cells as has been demonstrated using knockout mice (Flory et al., 1992; Flynn et al., 1992; Caruso et al., 1999). CD4 T cells of the T helper 1 subset activate infected macrophages by secreting cytokines, especially interferon- . Several studies published in the last 10 years have highlighted the roles of interferon- and also other proinflammatory cytokines such as interleukin-1, granulocyte-macrophage-colony-stimulating factor (GM-CSF) and tumor necrosis factor- (TNF-) in combating mycobacterial infections by contributing to the antimycobacterial potential of macrophages (Cooper et al., 1993; Dalton et al., 1993; Flynn et al., 1993, 1995). A well-studied interferon- -dependent anti-mycobacterial effector mechanism in mouse macrophages is the induction of nitric oxide-synthase (iNOS) that generates mycobactericidal nitric oxide (NO) (Flesch and Kaufmann, 1991; Chan et al., 1995; MacMicking et al., 1997; Flynn et al., 1998). In addition to interferon- , a second signal provided by the NF-kB pathway (interleukin-1, TNF- or mycobacterial products) is required for iNOS gene induction. Whereas from studies with iNOS-deficient mice the necessity of

74


NO generation for M. tuberculosis killing is clear (MacMicking et al., 1997), this mycobactericidal mechanism might not be as prominent in human macrophages that do not readily produce NO upon interferon- stimulation, but can kill mycobacteria by NO-independent mechanisms (James et al., 1990; Murray and Teitelbaum, 1992; Schneemann et al., 1993). One such mechanism is triggered by the 19 kDa lipoprotein of M. tuberculosis and probably involves TLR 2 (Thoma-Uszynski et al., 2001). NO-independent, interferon- -induced killing mechanisms must also exist in mice because control of M. avium infections requires interferon- but occurs also in iNOS-deficient mice (Gomes et al., 1999a). The precise contribution of NO in controlling human mycobacterial infection remains to be defined. One of the few polymorphic loci connected to relative natural resistance of mice against several bacterial including mycobacterial infections is the NRAMP-1 locus (natural resistance-associated macrophage protein 1). This transmembrane protein that is recruited to late endosomes/lysosomes most probably functions as an endosome-to-cytosol transporter for divalent cations. At the endosomal membrane, NRAMP-1 could serve to deplete the vesicle of divalent cations essential for microbial survival thereby supporting other microbicidal mechanisms in the macrophage (Vidal et al., 1995). Interestingly, the mycobacterial genome encodes for a homologue to NRAMP-1, which might compete with the host protein for divalent cations (Agranoff et al., 1999). Another major mechanism contributing to mycobacterial killing in interferon- -activated macrophages is thought to be the induction of phagosome–lysosome fusion. The interferon- -induced intracellular killing process in activated murine macrophages is preceded by a shift of the phagosome from an early endosomal stage to a late endosomal stage (Schaible et al., 1998; Via et al., 1998). This shift is characterized by the acquisition of the vacuolar type proton ATPase (and consequently vacuolar acidification), overlap of mycobacteria with lysosomal dyes and the non-accessibility of the matured vacuole for transferrin. The signaling mechanisms leading to cytokine-induced phagosomal maturation are unknown. Phagosome maturation is thought to be important in mycobacterial killing also because the low pH in the phagolysosome increases the toxicity of many microbicidal mechanisms operating in activated macrophages. A number of antimicrobial peptides have their optimal activity at the low pH of the lysosome (Miyakawa et al., 1996). In addition, nitric oxide gains highest toxicity at low pH and after reaction with products of the oxidative burst that takes place at the matured phagosome (Stuehr and Nathan, 1989; Halliwell and Gutteridge, 1992; Ehrt et al., 1997; St John et al., 2001). Furthermore, the activity of NRAMP-1, that functions to deplete the phagosome from divalent cations, is dependent on an acidic phagosomal pH (Jabado et al., 2000). Thus, phagosome maturation may


75

well synergize with other microbicidal mechanisms induced after macrophage activation. D. MYCOBACTERIAL VIRULENCE GENES The mycobacterial cell envelope differs from Gram-negative and Grampositive bacteria. In addition to cell membrane and peptidoglycan layers found in Gram-positive bacteria, the mycobacterial envelope contains a thick layer of mycolic acids (long chained -hydroxy fatty acids) that are covalently and non-covalently linked to the outer surface of the peptidoglycan, in the latter case in the form of trehalose dimycolate (TDM) (Brennan and Nikaido, 1995; Daffe and Draper, 1998). A broad spectrum of other non-covalently linked lipids and glycolipids contribute to the extreme hydrophobicity of the mycobacterial cell envelope. Many studies have highlighted bioactivity of individual envelope lipid components. The abundant lipoarabinomannan (LAM) can inhibit interferon- -mediated activation of macrophages in vitro and scavenge free oxygen radicals (Sibley et al., 1988; Chan et al., 1991). TDM by itself induces granulomatous inflammation after injection into mice (Ozeki et al., 1997) and in vitro inhibits Ca2 þ -dependent fusion of phospholipid vesicles (Spargo et al., 1991). The phthiocerol dimycocerosate family of envelope lipids can suppress lymphocyte responses (Prasad et al., 1987; Vachula et al., 1989) and facilitate Schwann cell invasion by Mycobacterium leprae by binding to peripheral laminin (Ng et al., 2000). Large parts of the mycobacterial genome encode for enzymes involved in biosynthetic pathways for lipids that are incorporated into the complex lipophilic mycobacterial cell envelope. With the help of newly developed screening and mutagenesis techniques, specific pathways for lipid biosynthesis have been shown to be essential for virulence or persistence. Analysis of mutants displaying small alterations in the structures of distinct cell envelope lipids emphasizes that individual lipid components of the bacterial envelope contribute to distinct virulence/persistence phenotypes. Therefore, the mycobacterial envelope cannot be seen merely as a hydrophobic inert capsule but probably represents a collection of immuno-modulatory molecules that mediate specific aspects of mycobacterial virulence (Camacho et al., 1999; Cox et al., 1999; Dubnau et al., 2000). The sequencing of the M. tuberculosis H37Rv circular chromosome has revealed the presence of large protein families with virulence potential (Cole et al., 1998). One tenth of the genome encodes for proteins containing proline-glutamate or proline-proline-glutamate residues at their predicted N-terminus and share conserved N-terminal domains of 110–180 amino acid residues (PE and PPE protein families). A search for Mycobacterium marinum promoters active in host frog granulomata identified two members of the PE family to be selectively upregulated intracellularly and essential

76


for replication in macrophages after phagocytosis (Ramakrishnan et al., 2000), although their function remains elusive. Another large proportion of the mycobacterial genome seems to be dedicated to lipid breakdown. Many predicted mycobacterial proteins show homology to the E. coli Fad system of fatty acid -oxidation and are present in multiple copies (e.g., acyl-CoA-synthetase is represented by 36 homologues (Cole et al., 1998)). As mycobacteria do mainly metabolize fatty acids when growing in mouse tissues (Wheeler and Ratledge, 1988), the strong putative -oxidation apparatus might reflect the intracellular specialization of mycobacteria toward this energy and nutritional source. Two other gene products are thought to allow mycobacterial adaptation to the intracellular milieu of macrophages. First, the catalase-peroxidase KatG contributes to virulence by inactivating reactive oxygen intermediates inside the phagosome (Li et al., 1998). Second, the enzyme isocitrate lyase (Icl) is upregulated by mycobacteria when residing in activated macrophages and is essential for persistence in mice after onset of specific immunity which occurs after about two weeks. Icl is the key enzyme of the glyoxylate shunt pathway for gluconeogenesis and allows mycobacteria to grow on acetyl-CoA as the sole carbon source, probably a requirement for persistence in immune competent hosts (Honer Zu Bentrup et al., 1999). Crucial for mycobacterial survival is the mycobacterial production of mycobactins, siderophores which function in the acquisition of iron in the restrictive environment of the macrophage (De Voss et al., 2000). These compounds are synthesized from salicylate, serine and threonine through the action of a cluster of genes that include both nonribosomal peptide synthase and polyketide synthase (Quadri et al., 1998). Thus, partially because of the possibility now to successfully manipulate mycobacteria on the genomic level, mycobacterial genes involved in the production of lipid virulence factors and involved in the metabolic adaptation to the macrophage intracellular milieu have been identified. E. TACO, A COAT PROTEIN

AT THE

MYCOBACTERIAL PHAGOSOME

In a recent proteomic approach to identify host proteins potentially involved in the inhibition of phagosomal maturation, a 50 kDa protein was described which coats mycobacterial phagosomes containing viable bacilli and is absent from phagosomes containing heat-killed bacilli (Ferrari et al., 1999). The 461 amino acid protein (Fig. 3) was termed TACO (tryptophan aspartatecontaining coat protein) and belongs to the family of tryptophan-aspartate 40 (WD40) repeat proteins, a protein family whose members have thus far only been found to be present in eukaryotes (Neer et al., 1994). Its members have diverse functions in processes such as signal transduction, cell motility (King et al., 1995), vesicle budding ( -COP of COPI coats) (Duden et al.,


77

FIG. 3. Tryptophan aspartate containing coat protein (TACO). The 461 residue peripheral membrane protein belongs to the family of WD repeat proteins and contains five WD40 repeat motifs located in the N-terminal part of the sequence. The 30 amino acids preceding the C-terminus are predicted to form a coiled-coil structure.

1991; Pryer et al., 1993) and cytoskeletal organization (Li and Suprenant, 1994; Wilkerson et al., 1995). The precise role of the WD40-domains is unknown, but they may provide contact surfaces within protein complexes (Wall et al., 1995; Sondek et al., 1996). In addition to its WD40 repeat sequences, TACO contains a C-terminal stretch of approximately 30 amino acids predicted to form coiled-coils. TACO was found to be expressed in all leukocytes examined which is in agreement with its presence in the lymphatic organs, the brain (microglia) and the lung (alveolar macrophages). Immunofluorescence analysis of J774 macrophages infected with M. bovis BCG expressing the green fluorescent protein (GFP) demonstrated that viable mycobacteria recruit TACO from its cortical sites at the plasma membrane to the site of uptake and actively retain the protein at the mycobacterial phagosome. In contrast, TACO is released from phagosomes containing dead mycobacteria, followed by lysosomal delivery of the microbes indicating the requirement for metabolically active bacteria in the phagosomal retention of TACO (Fig. 4). Transfection of TACO cDNA into a TACO-negative phagocytic cell line followed by infection and analysis of intracellular bacterial trafficking and killing demonstrated that TACO was instrumental in preventing lysosomal delivery of mycobacteria. Cells containing TACO displayed a block in lysosomal delivery of the mycobacteria and supported mycobacterial survival in contrast to mock transfected control cells in which survival was about 10-fold decreased 15 h after infection. Additional evidence for a role of TACO in mycobacterial survival was provided by the observation that Kupffer cells, the resident macrophages of the liver and the only TACO-negative macrophages known to date, kill mycobacteria very efficiently (North, 1974; Wardle, 1987). These experiments demonstrated a causative role for TACO in maintaining the non-fusogenic status of the mycobacterial phagosome. F. ROLE OF PLASMA MEMBRANE CHOLESTEROL INFECTION

IN

MYCOBACTERIAL

TACO, like other known coat proteins, is not a transmembrane protein and does not display any ER-targeting sequence despite its association with the

78


FIG. 4. Trafficking of pathogenic mycobacteria in macrophages. Material internalized by receptor mediated endocytosis or fluid phase endocytosis as well as dead mycobacteria are transported to lysosomes for degradation. Live mycobacteria however remain segregated from lysosomes and reside within the mycobacterial phagosome. Recruitment and retention of TACO by viable mycobacteria to the phagosome prevents their lysosomal delivery. In contrast, killed mycobacteria cannot retain TACO at the phagosomal membrane and are consequently transported to lysosomes. (From Pieters, 1999, by permission of JAI Press/Elsevier Science, Stamford, CT.)

phagosomal and the plasma membrane. Biochemical analysis of membraneassociated TACO revealed its TX-100 insolubility at low temperatures and sensitivity to the cholesterol-sequestering agent digitonin indicating that TACO was associated to membranes via cholesterol (Gatfield and Pieters, 2000). Cholesterol is a major constituent of the animal plasma membrane and an essential component of membrane microdomains that have been shown to be involved in membrane sorting and signal transduction (Simons and Ikonen, 1997). The finding that cholesterol is essential for the attachment of TACO to the phagosomal and plasma membrane, led to investigating the function of this lipid during mycobacterial infection of macrophages. Visualization of cellular cholesterol using the fluorescent cholesterol-binding dye filipin (Drabikowski et al., 1973; Bornig and Geyer, 1974) revealed an accumulation of the sterol at the site of mycobacterial uptake, suggesting a role for cholesterol in mycobacterial trafficking within macrophages (Gatfield and Pieters, 2000) (Fig. 5). Evidence for such a role came from experiments using cholesterol depleted cells aiming at disrupting TACO association with the plasma


79

FIG. 5. Distribution of cholesterol during mycobacterial uptake. Macrophages were incubated with M. bovis BCG expressing the green fluorescent protein, washed and fixed. Cells were stained with the cholesterol specific dye filipin (blue). Scale bar ¼ 10 m. Reproduced, with permission, from Gatfield, J. and Pieters, J. (2000) Science 288, 1647–1650 copyright 2000, AAAS (http:// www.sciencemag.org).

membrane and analyzing the effect on intracellular mycobacterial trafficking. To reduce cholesterol levels in cultured cells, cholesterol biosynthesis was inhibited using lovastatin and plasma membrane cholesterol was extracted with the selective cholesterol-extracting compound methyl- -cyclodextrin (Klein et al., 1995; Simons et al., 1998). Instead of resulting in lysosomal transfer because of lack of TACO retention, in cholesterol depleted macrophages the rate of entry was reduced by up to 85% (Fig. 6) (Gatfield and Pieters, 2000). The role of cholesterol during mycobacterial entry into their host cells was specific for mycobacteria because uptake of E. coli, Yersinia pseudotuberculosis, S. typhimurium and Lactobacillus casei was not affected by depletion of cholesterol (Fig. 6) (Gatfield and Pieters, 2000). One possibility to explain the role of cholesterol in mycobacterial entry is the need for cholesterol in receptor functioning during mycobacterial entry. When antibodies towards complement receptor type 3 were included during infection, the degree of uptake was strongly reduced both in J774 and in bone marrow derived macrophages (BMM) demonstrating that mycobacteria enter these two cell types mainly via complement receptor type 3 (Fig. 7a) (Gatfield and Pieters, 2000). Entry into monocyte-derived human dendritic cells was reduced by 60% suggesting an additional receptor being involved in the uptake of BCG by human dendritic cells (Fig. 7a). In summary, mycobacterial entry in all three cell types examined was largely mediated by complement receptor type 3. Experiments addressing the role of cholesterol in complement receptor mediated uptake using serum-opsonized L. casei, indicated that, first, uptake

80


FIG. 6. Effect of cholesterol depletion on entry of different species of bacteria. Control and cholesterol depleted macrophages were allowed to internalize the different fluorescent or radiolabeled microorganisms and uptake was analyzed by flow cytometry or liquid scintillation counting. Reproduced, with permission, from Gatfield, J. and Pieters, J. (2000) Science 288, 1647–1650 copyright 2000, AAAS (http://www.sciencemag.org).

FIG. 7. Receptor usage during mycobacterial entry into phagocytes and the role of cholesterol. a. Effect of anti-CR3 antibodies on mycobacterial uptake by J774 macrophages, bone marrow derived macrophages (BMM) and human dendritic cells (hDC). b. Analysis of CR3-mediated uptake of complement-opsonized lactobacilli by control and cholesterol depleted J774 macrophages. Reproduced, with permission, from Gatfield, J. and Pieters, J. (2000) Science 288, 1647–1650 copyright 2000, AAAS (http://www.sciencemag.org).


81

was mediated via complement receptor type 3, and second, that cholesterol was not required for uptake of lactobacilli via this receptor. Thus, cholesterol is not needed for complement receptor function in general, but is specific for mycobacterial entry (Fig. 7b) (Gatfield and Pieters, 2000). At least two possibilities exist for the cholesterol involvement in complement receptor type 3-mediated mycobacterial entry. First, plasma membrane cholesterol could serve as a direct adhesion site for the mycobacterial envelope (Brennan and Nikaido, 1995). The envelope is known to be very hydrophobic and to display a high capacity for binding cholesterol in vitro (Gatfield and Pieters, 2000). However, as cholesterol is largely embedded in the core of the plasma membrane between the fatty acyl chains, a direct interaction between the bacterial envelope and the membrane sterol might be difficult. In addition, the glycocalix limits the accessibility of the lipid surface of the membrane. Second, cholesterol could play a role during signaling processes required for complement receptor type 3-mediated mycobacterial entry. Complement receptor type 3 is known, like many other integrins, to require, besides its ligation, a second signal to initiate high affinity binding and phagocytosis (Jones et al., 1998a), a phenomenon termed integrin activation. Cholesterol, possibly as a component of cholesterol-rich microdomains, might be essential for the generation or transduction of the second signal needed for complement receptor type 3-mediated uptake of non-opsonized mycobacteria. Time-lapse microscopy of macrophage–Mycobacterium interaction revealed that depleted cells have lost the ability to tightly adhere to mycobacteria whereas their chemotactic responses remain unaltered (Gatfield and Pieters, 2000). The lack of affinity between host cell and pathogen in the absence of cholesterol might be precisely based on failure in integrin activation. In contrast, complement-opsonized lactobacilli would provide a cholesterolindependent integrin activation signal and be taken up also by cholesterol depleted cells. A recent study on the role of cholesterol in complement receptor type 3-mediated uptake of Mycobacterium kansasii by neutrophils (Peyron et al., 2000), supports the notion that signaling events involving cholesterol are essential for mycobacterial entry. This group provided evidence that complement receptor type 3 itself has to be positioned in cholesterol-rich microdomains to allow mycobacterial uptake and that this positioning is a result of this receptor interacting with GPI-linked proteins. Taken together, the data suggest that mycobacterial uptake is dependent on a signaling event in which the phagocytic receptor and the signaling apparatus for its activation are brought together via cholesterol. From the point of view of the pathogen, an entry into host cells via cholesterol-enriched domains might be of great advantage for subsequent survival. Mycobacteria would, during entry, recruit TACO that is peripherally

82


attached to cholesterol-rich microdomains from the cytoplasmic side. In so doing, the mycobacteria would ensure sequestration into TACO-coated phagosomes (Fig. 8). Interestingly, as mentioned, mycobacteria display a preference for complement receptor type 3-mediated uptake. Not only do mycobacteria possess a site for non-opsonic binding, but they also induce their own opsonization by complement, either by activating the alternative pathway or by assembling a unique C3 convertase on their surface (Schorey et al., 1997; Ernst, 1998). Therefore, it is quite possible that complement receptormediated entry has significant advantage for mycobacteria. Information on whether or not cholesterol and TACO also play a role in mycobacterial entry via other receptors, would help to assess if complement receptor type 3 is a specifically hospitable port of entry. Interestingly, when dendritic cells, which express high levels of active mannose receptor, are infected with M. bovis BCG, the pathogens still enter via complement receptor type 3 avoiding the mannose receptor (unpublished results). Mannose receptor-mediated phagocytosis of mycobacteria induces strong anti-microbial responses in mouse macrophages, and avoiding the mannose receptor might be part of a mycobacterial survival strategy (Ezekowitz et al., 1990). In summary, the TACO protein represents a host protein that is used by a pathogenic microbe to its own benefit in allowing its survival and replication within the host cell. TACO allows mycobacteria to survive by preventing phagosome maturation. When mycobacteria enter macrophages via complement receptor type 3 they recruit TACO, cholesterol and possibly a set of signaling molecules to the site of uptake. The combined presence of complement receptor type 3 and cholesterol is necessary for uptake and might have evolved as a strategy of the Mycobacterium to gain entry into TACO-coated phagosomes. Immune competent hosts have mechanisms counteracting the blockage in phagosome–lysosome fusion, namely interferon -induced macrophage activation that possibly leads to phagosome–lysosome fusion and subsequent mycobacterial killing (Schaible et al., 1998; Via et al., 1998). How cytokines regulate this fusion and whether they do so by acting on TACO function is unknown and could provide more insight into how the adaptive immune system of immune-competent hosts controls mycobacterial infections. The normal cellular function of TACO is not known. TACO is an abundant protein representing up to 0.5% of the total protein in macrophages (unpublished results). Therefore it might play a structural role in coating the inner leaflet of the plasma membrane and early endosomes/phagosomes. This coating might define the plasma membrane status and thus prevent inappropriate docking of late endosomal and lysosomal vesicles (Ferrari et al., 1999; Pieters, 2001). Work on the putative TACO ancestor coronin (35% homology) in Dictyostelium discoideum, a slime mould living off soil bacteria,


83

FIG. 8. Model of cholesterol-dependent mycobacterial entry into macrophages. During entry of mycobacteria into macrophages, plasma membrane cholesterol accumulates at the site of uptake. Cholesterol plays a two-fold role in, on the one hand, mediating attachment of the tryptophan aspartate containing coat protein (TACO) to the plasma membrane, on the other hand being crucial for the functioning and the proper localization of complement receptor type 3 in membrane microdomains during mycobacterial entry. By entering via complement receptor type 3 at cholesterol-rich microdomains, mycobacteria might ensure their subsequent sequestration in TACO-coated phagosomes. TACO-coated phagosomes do not fuse with lysosomes thereby allowing mycobacteria to survive inside the macrophage.

84


revealed similar recruitment of coronin to nascent phagosomes and dissociation after internalization in this unicellular organism (de Hostos et al., 1991; Maniak et al., 1995). Knockout studies in Dictyostelium demonstrated a role for coronin in phagocytosis, cytokinesis and motility without being absolutely essential for these processes (de Hostos et al., 1993). In how far TACO is required for these processes in mouse leukocytes remains to be analyzed. Like other known coat proteins, TACO is likely to exist in a soluble cytosolic and a membrane-bound form. These two TACO-pools are probably interconvertible and regulated by cellular (and bacterial) signaling pathways that induce TACO assembly and disassembly at the plasma membrane and early endosomes. Revealing these pathways should allow a better understanding of how mycobacteria recruit and retain TACO and whether or how TACO is connected to components of the membrane fusion machinery and the cytoskeleton. VI. Conclusions

The components of the innate and adaptive immune systems and their functions during an immune response against pathogens are largely known today. Depending on the given microbe the contributions of the individual host cell types and molecules to an effective immune response vary. Most microbes are recognized, killed and degraded by first line innate immune mechanisms, without causing any symptoms of disease. Only the few longer persisting pathogens, which survive innate immunity, trigger adaptive immune responses that are based on molecularly well understood pathways of immunoglobulin receptor rearrangement, antigen processing and presentation and T- or B- cell activation. Innate immunity as a first line defense system should act against all invading pathogens and therefore has made available a diverse repertoire of pathogen recognition molecules and mechanisms of elimination. The central innate immune cell, the macrophage, is equipped with a variety of recognition receptors, large phagocytic capacity, and a highly efficient killing and degradation machinery and is therefore a main mediator of anti-bacterial immune responses. Surviving first line innate immunity is essential for microbes to proliferate as successful pathogens, and thus many pathogenic bacteria have developed strategies to subvert or escape the elimination by macrophages at many different levels. The microbial tactics for macrophage subversion are often based on offensive interference with host cell physiology, such as protein injection to avoid phagocytosis, induction of apoptosis, rupturing the phagosome and host-actin based intracellular motility, but can also, in the


85

case of seemingly passive, slow growing pathogens, be tailored to allow a long term survival in the presence of the host immune system. As a result, some of the latter microbes can remain in largely symptom-free symbiosis with their hosts. The analysis of these different mechanisms of host–pathogen interaction is contributing greatly to our understanding of disease processes as well as of basic cell biological mechanisms. A promising model system for revealing principles governing vesicular transport and vesicle fusion is the interaction of macrophages with pathogenic mycobacteria. Mycobacteria establish a long-term balance with the host’s defense systems by preventing lysosomal transfer within macrophages. Microbiological and cell biological research is starting to explain the mechanisms of mycobacterial adaptation, and as a result we hope to learn also about mechanisms governing phagosome lysosome fusion or vesicular trafficking in general. The recent identification of a phagosomal coat protein termed TACO gives a first explanation of how phagosome–lysosome fusion might be regulated in infected and non-infected cells. One of the open questions is still how mycobacteria signal to macrophages to retain TACO at the phagosome to inhibit fusion with lysosomes and how precisely the presence of TACO at the phagosomal membrane prevents fusion. Furthermore, how macrophage activation can overcome the mycobacterially induced inhibition of fusion remains a challenging question. Despite being among the first bacteria to be discovered over a century ago, mycobacteria still remain among the least understood pathogens at the molecular level. The multidisciplinary approach, however, followed over the past few years contributes to the decoding of the mechanisms operating during the persistence of mycobacteria within their host cells. As a result, this work might not only help to develop more specific therapies to combat tuberculosis but also offer new insights into basic cell biological processes. ACKNOWLEDGMENTS Research of our laboratory described in this chapter was in part performed at the Basel Institute for Immunology which was founded and supported by Hoffmann-LaRoche Ltd., Switzerland. Current support is provided by grants from the Swiss National Science Foundation, the World Health Organization and the Olga Mayenfisch Stiftung to J.P.

REFERENCES Abbas, A. K., Murphy, K. M., and Sher, A. (1996). Nature 383, 787. Adam, T., Giry, M., Boquet, P., and Sansonetti, P. (1996). EMBO J. 15, 3315. Adams, D. O., and Hamilton, T. A. (1992). In Lewis, C. E., and McGee, J. O. D. (eds.), ‘‘The Macrophage’’, Oxford-New York-Tokyo, Oxford University Press, p. 75. Aderem, A., and Underhill, D. M. (1999). Annu. Rev. Immunol. 17, 593.

86


Agranoff, D., Monahan, I. M., Mangan, J. A., Butcher, P. D., and Krishna, S. (1999). J. Exp. Med. 190, 717. Allen, L. A., and Aderem, A. (1996). J. Exp. Med. 184, 627. Alpuche-Aranda, C. M., Racoosin, E. L., Swanson, J. A., and Miller, S. I. (1994). J. Exp. Med. 179, 601. Altare, F., Durandy, A., Lammas, D., Emile, J. F., Lamhamedi, S., Le Deist, F., Drysdale, P., Jouanguy, E., Doffinger, R., Bernaudin, F., Jeppsson, O., Gollob, J. A., Meinl, E., Segal, A. W., Fischer, A., Kumararatne, D., and Casanova, J. L. (1998). Science 280, 1432. Alvarez-Dominguez, C., Barbieri, A. M., Beron, W., Wandinger-Ness, A., and Stahl, P. D. (1996). J. Biol. Chem. 271, 13834. Amigorena, S., Drake, J. R., Webster, P., and Mellman, I. (1994). Nature 369, 113. Andrew, D. P., Spellberg, J. P., Takimoto, H., Schmits, R., Mak, T. W., and Zukowski, M. M. (1998). Eur. J. Immunol. 28, 1959. Androlewicz, M. J., Anderson, K. S., and Cresswell, P. (1993). Proc. Natl. Acad. Sci. USA 90, 9130. Araki, N., Johnson, M. T., and Swanson, J. A. (1996). J. Cell Biol. 135, 1249. Armstrong, J. A., and Hart, P. D. A. (1971). J. Exp. Med. 134, 713. Asnagli, H., and Murphy, K. M. (2001). Curr. Opin. Immunol. 13, 242. Astarie-Dequeker, C., N’Diaye, E. N., Le Cabec, V., Rittig, M. G., Prandi, J., and MaridonneauParini, I. (1999). Infect. Immun. 67, 469. Balsam, L. B., Liang, T. W., and Parkos, C. A. (1998). J. Immunol. 160, 5058. Bauer, S., Groh, V., Wu, J., Steinle, A., Phillips, J. H., Lanier, L. L., and Spies, T. (1999). Science 285, 727. Baumeister, W., Walz, J., Zuhl, F., and Seemuller, E. (1998). Cell 92, 367. Baumert, M., Maycox, P. R., Navone, F., De Camilli, P., and Jahn, R. (1989). EMBO J. 8, 379. Bednarek, S. Y., Ravazzola, M., Hosobuchi, M., Amherdt, M., Perrelet, A., Schekman, R., and Orci, L. (1995). Cell 83, 1183. Bellinger-Kawahara, C., and Horwitz, M. A. (1990). J. Exp. Med. 172, 1201. Bennett, M. K., Calakos, N., and Scheller, R. H. (1992). Science 257, 255. Bergquist, P. R., and Glasgow, K. (1986). Exp. Biol. 45, 111. Berke, G. (1997). Curr. Opin. Hematol. 4, 32. Beron, W., Alvarez-Dominguez, C., and Stahl, P. D. (1995). Trends Cell Biol. 5, 100. Berton, G., Laudanna, C., Sorio, C., and Rossi, F. (1992). J. Cell Biol. 116, 1007. Bjorkman, P. J., Saper, M. A., Samraoui, B., Bennett, W. S., Strominger, J. L., and Wiley, D. C. (1987). Nature 329, 506. Black, D. S., and Bliska, J. B. (1997). EMBO J. 16, 2730. Blocker, A., Gounon, P., Larquet, E., Niebuhr, K., Cabiaux, V., Parsot, C., and Sansonetti, P. (1999). J. Cell Biol. 147, 683. Bloom, B. R. (1992). Nature 358, 538. Bogdan, C., Rollinghoff, M., and Diefenbach, A. (2000). Curr. Opin. Immunol. 12, 64. Bonifacino, J. S., and Dell’Angelica, E. C. (1999). J. Cell Biol. 145, 923. Bornig, H., and Geyer, G. (1974). Acta Histochem. 50, 110. Bourdet-Sicard, R., Rudiger, M., Jockusch, B. M., Gounon, P., Sansonetti, P. J., and Nhieu, G. T. (1999). EMBO J. 18, 5853. Braun, L., Ghebrehiwet, B., and Cossart, P. (2000). EMBO J. 19, 1458. Brennan, P. J., and Nikaido, H. (1995). Annu. Rev. Biochem. 64, 29. Brightbill, H. D., Libraty, D. H., Krutzik, S. R., Yang, R. B., Belisle, J. T., Bleharski, J. R., Maitland, M., Norgard, M. V., Plevy, S. E., Smale, S. T., Brennan, P. J., Bloom, B. R., Godowski, P. J., Modlin, R. L. (1999). Science 285, 732. Brown, M. S., and Goldstein, J. L. (1990). Nature 343, 508. Bustelo, X. R. (2000). Mol. Cell Biol. 20, 1461.


87

Camacho, L. R., Ensergueix, D., Perez, E., Gicquel, B., and Guilhot, C. (1999). Mol. Microbiol. 34, 257. Caron, E., and Hall, A. (1998). Science 282, 1717. Carroll, M. C. (1998). Annu. Rev. Immunol. 16, 545. Carter, P. B., and Collins, F. M. (1974). Infect. Immun. 9, 851. Caruso, A. M., Serbina, N., Klein, E., Triebold, K., Bloom, B. R., and Flynn, J. L. (1999). J. Immunol. 162, 5407. Chan, J., Fan, X. D., Hunter, S. W., Brennan, P. J., and Bloom, B. R. (1991). Infect. Immun. 59, 1755. Chan, J., Tanaka, K., Carroll, D., Flynn, J., and Bloom, B. R. (1995). Infect. Immun. 63, 736. Chan, J., Tian, Y., Tanaka, K. E., Tsang, M. S., Yu, K., Salgame, P., Carroll, D., Kress, Y., Teitelbaum, R., and Bloom, B. R. (1996). Proc. Natl. Acad. Sci. USA 93, 14857. Chang, C. H., and Flavell, R. A. (1995). J. Exp. Med. 181, 765. Chapman, H. A. (1998). Curr. Opin. Immunol. 10, 93. Chavrier, P., Parton, R. G., Hauri, H. P., Simons, K., and Zerial, M. (1990). Cell 62, 317. Chimini, G., and Chavrier, P. (2000). Nat. Cell Biol. 2, E191. Clemens, D. L., and Horwitz, M. A. (1995). J. Exp. Med. 181, 257. Clemens, D. L., and Horwitz, M. A. (1996). J. Exp. Med. 184, 1349. Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., Gordon, S. V., Eiglmeier, K., Gas, S., Barry, C. E. R., Tekaia, F., Badcock, K., Basham, D., Brown, D., Chillingworth, T., Connor, R., Davies, R., Devlin, K., Feltwell, T., Gentles, S., Hamlin, N., Holroyd, S., Hornsby, T., Jagels, K., and Barrell, B. G. (1998). Nature 393, 537. Collins, T., Korman, A. J., Wake, C. T., Boss, J. M., Kappes, D. J., Fiers, W., Ault, K. A., Gimbrone, M. A., Jr., Strominger, J. L., and Pober, J. S. (1984). Proc. Natl. Acad. Sci. USA 81, 4917. Colonna, M., Moretta, A., Vely, F., and Vivier, E. (2000). Immunol. Today 21, 428. Cooper, A. M., Dalton, D. K., Stewart, T. A., Griffin, J. P., Russell, D. G., and Orme, I. M. (1993). J. Exp. Med. 178, 2243. Cooper, N. R. (1985). Adv. Immunol. 37, 151. Cornelis, G. R., and Van Gijsegem, F. (2000). Annu. Rev. Microbiol. 54, 735. Cossart, P., and Bierne, H. (2001). Curr. Opin. Immunol. 13, 96. Cosson, P., and Letourneur, F. (1997). Curr. Opin. Cell Biol. 9, 484. Cox, D., Chang, P., Zhang, Q., Reddy, P. G., Bokoch, G. M., and Greenberg, S. (1997). J. Exp. Med. 186, 1487. Cox, J. S., Chen, B., McNeil, M., and Jacobs, W. R., Jr. (1999). Nature 402, 79. Crespo, P., Schuebel, K. E., Ostrom, A. A., Gutkind, J. S., and Bustelo, X. R. (1997). Nature 385, 169. Crowle, A. J., Dahl, R., Ross, E., and May, M. H. (1991). Infect. Immun. 59, 1823. Crowley, M. T., Costello, P. S., Fitzer-Attas, C. J., Turner, M., Meng, F., Lowell, C., Tybulewicz, V. L., and DeFranco, A. L. (1997). J. Exp. Med. 186, 1027. Cywes, C., Hoppe, H. C., Daffe, M., and Ehlers, M. R. (1997). Infect. Immun. 65, 4258. Da Silva, R. P., Hall, B. F., Joiner, K. A., and Sacks, D. L. (1989). J. Immunol. 143, 617. Daffe, M., and Draper, P. (1998). Adv. Microb. Physiol. 39, 131. Dai, W. J., Bartens, W., Kohler, G., Hufnagel, M., Kopf, M., and Brombacher, F. (1997). J. Immunol. 158, 5297. Dalton, D. K., Pitts-Meek, S., Keshav, S., Figari, I. S., Bradley, A., and Stewart, T. A. (1993). Science 259, 1739. Davis, B. D., Dulbecco, R., Eisen, H. N., and Ginsberg, H. S. (1990). ‘‘Microbiology’’, 4th edn., Philadelphia, J.B. Lippincott Company. Davis, M. M., Boniface, J. J., Reich, Z., Lyons, D., Hampl, J., Arden, B., and Chien, Y. (1998). Annu. Rev. Immunol. 16, 523.

88


de Chastellier, C., and Thilo, L. (1997). Eur. J. Cell Biol. 74, 49. de Hostos, E. L., Bradtke, B., Lottspeich, F., Guggenheim, R., and Gerisch, G. (1991). EMBO J. 10, 4097. de Hostos, E. L., Rehfuess, C., Bradtke, B., Waddell, D. R., Albrecht, R., Murphy, J., and Gerisch, G. (1993). J. Cell Biol. 120, 163. de Jong, R., Altare, F., Haagen, I. A., Elferink, D. G., Boer, T., van Breda Vriesman, P. J., Kabel, P. J., Draaisma, J. M., van Dissel, J. T., Kroon, F. P., Casanova, J. L., and Ottenhoff, T. H. (1998). Science 280, 1435. De Voss, J. J., Rutter, K., Schroeder, B. G., Su, H., Zhu, Y., and Barry, C. E. (2000). Proc. Natl. Acad. Sci. USA 97, 1252. Dehio, C., Prevost, M. C., and Sansonetti, P. J. (1995). EMBO J. 14, 2471. Dell’Angelica, E. C., Mullins, C., and Bonifacino, J. S. (1999). J. Biol. Chem. 274, 7278. Denecker, G., Declercq, W., Geuijen, C. A., Benabdillah, R., van Gurp, M., Sory, M. P., Vandenabeele, P., and Cornelis, G. R. (2001). J. Biol. Chem. 14, 14. Denzin, L. K., and Cresswell, P. (1995). Cell 82, 155. Desjardins, M., and Descoteaux, A. (1997). J. Exp. Med. 185, 2061. Desjardins, M., Huber, L. A., Parton, R. G., and Griffiths, G. (1994). J. Cell Biol. 124, 677. Dharmawardhane, S., Brownson, D., Lennartz, M., and Bokoch, G. M. (1999). J. Leukoc. Biol. 66, 521. Downing, J. F., Pasula, R., Wright, J. R., Twigg, H. L., and Martin, W. J. (1995). Proc. Natl. Acad. Sci. USA 92, 4848. Drabikowski, W., Lagwinska, E., and Sarzala, M. G. (1973). Biochim. Biophys. Acta 291, 61. Dramsi, S., and Cossart, P. (1998). Annu. Rev. Cell Dev. Biol. 14, 137. Drees, B., Friederich, E., Fradelizi, J., Louvard, D., Beckerle, M. C., and Golsteyn, R. M. (2000). J. Biol. Chem. 275, 22503. Dubnau, E., Chan, J., Raynaud, C., Mohan, V. P., Laneelle, M. A., Yu, K., Quemard, A., Smith, I., and Daffe, M. (2000). Mol. Microbiol. 36, 630. Duden, R., Griffiths, G., Frank, R., Argos, P., and Kreis, T. E. (1991). Cell 64, 649. Dunne, D. W., Resnick, D., Greenberg, J., Krieger, M., and Joiner, K. A. (1994). Proc. Natl. Acad. Sci. USA 91, 1863. Eggleton, P., and Reid, K. B. (1999). Curr. Opin. Immunol. 11, 28. Ehlers, M. R. (2000). Microbes. Infect. 2, 289. Ehrt, S., Shiloh, M. U., Ruan, J., Choi, M., Gunzburg, S., Nathan, C., Xie, Q., and Riley, L. W. (1997). J. Exp. Med. 186, 1885. Engering, A. J., Cella, M., Fluitsma, D. M., Hoefsmit, E. C. M., Lanzavecchia, A., and Pieters, J. (1997). Eur. J. Immunol. 27, 2417. Ernst, J. D. (1998). Infect. Immun. 66, 1277. Ezekowitz, R. A., Sastry, K., Bailly, P., and Warner, A. (1990). J. Exp. Med. 172, 1785. Ezekowitz, R. A., Williams, D. J., Koziel, H., Armstrong, M. Y., Warner, A., Richards, F. F., and Rose, R. M. (1991). Nature 351, 155. Falk, K., Rotzschke, O., Stevanovic, S., Jung, G., and Rammensee, H. G. (1994). Immunogenetics 39, 230. Fearon, D. T., and Locksley, R. M. (1996). Science 272, 50. Feng, Y., Press, B., and Wandinger-Ness, A. (1995). J. Cell Biol. 131, 1435. Ferrari, G., Knight, A. M., Watts, C., and Pieters, J. (1997). J. Cell Biol. 139, 1433. Ferrari, G., Naito, M., Langen, H., and Pieters, J. (1999). Cell 97, 435. Finlay, B. B., and Cossart, P. (1997). Science 276, 718. Flesch, I., and Kaufmann, S. H. (1987). J. Immunol. 138, 4408. Flesch, I. E., and Kaufmann, S. H. (1991). Infect. Immun. 59, 3213. Flory, C. M., Hubbard, R. D., and Collins, F. M. (1992). J. Leukoc. Biol. 51, 225.


89

Flynn, J. L., Chan, J., Triebold, K. J., Dalton, D. K., Stewart, T. A., and Bloom, B. R. (1993). J. Exp. Med. 178, 2249. Flynn, J. L., Goldstein, M. M., Chan, J., Triebold, K. J., Pfeffer, K., Lowenstein, C. J., Schreiber, R., Mak, T. W., and Bloom, B. R. (1995). Immunity 2, 561. Flynn, J. L., Goldstein, M. M., Triebold, K. J., Koller, B., and Bloom, B. R. (1992). Proc. Natl. Acad. Sci. USA 89, 12013. Flynn, J. L., Scanga, C. A., Tanaka, K. E., and Chan, J. (1998). J. Immunol. 160, 1796. Forsyth, C. B., Plow, E. F., and Zhang, L. (1998). J. Immunol. 161, 6198. Francis, C. L., Ryan, T. A., Jones, B. D., Smith, S. J., and Falkow, S. (1993). Nature 364, 639. Frehel, C., Ryter, A., Rastogi, N., and David, H. (1986). Ann. Inst. Pasteur. Microbiol. 137B, 239. Fremont, D. H., Hendrickson, W. A., Marrack, P., and Kappler, J. (1996). Science 272, 1001. Gadjeva, M., Thiel, S., and Jensenius, J. C. (2001). Curr. Opin. Immunol. 13, 74. Ganz, T., and Lehrer, R. I. (1998). Curr. Opin. Immunol. 10, 41. Garcia-del Portillo, F., and Finlay, B. B. (1995). J. Cell Biol. 129, 81. Garcia-del Portillo, F., Zwick, M. B., Leung, K. Y., and Finlay, B. B. (1993). Proc. Natl. Acad. Sci. USA 90, 10544. Gatfield, J., and Pieters, J. (2000). Science 288, 1647. Gaynor, C. D., McCormack, F. X., Voelker, D. R., McGowan, S. E., and Schlesinger, L. S. (1995). J. Immunol. 155, 5343. Gedde, M. M., Higgins, D. E., Tilney, L. G., and Portnoy, D. A. (2000). Infect. Immun. 68, 999. Geese, M., Schluter, K., Rothkegel, M., Jockusch, B. M., Wehland, J., and Sechi, A. S. (2000). J. Cell Sci. 113, 1415. Geier, E., Pfeifer, G., Wilm, M., Lucchiari-Hartz, M., Baumeister, W., Eichmann, K., and Niedermann, G. (1999). Science 283, 978. Geoffroy, C., Gaillard, J. L., Alouf, J. E., and Berche, P. (1987). Infect. Immun. 55, 1641. Germain, R. N. (1994). Cell 76, 287. Glas, R., Bogyo, M., McMaster, J. S., Gaczynska, M., and Ploegh, H. L. (1998). Nature 392, 618. Gold, E. S., Morrissette, N. S., Underhill, D. M., Guo, J., Bassetti, M., and Aderem, A. (2000). Immunity 12, 285. Gold, E. S., Underhill, D. M., Morrissette, N. S., Guo, J., McNiven, M. A., and Aderem, A. (1999). J. Exp. Med. 190, 1849. Goldstein, J. L., Anderson, R. G., and Brown, M. S. (1979). Nature 279, 679. Gomes, M. S., Florido, M., Pais, T. F., and Appelberg, R. (1999a). J. Immunol. 162, 6734. Gomes, M. S., Paul, S., Moreira, A. L., Appelberg, R., Rabinovitch, M., and Kaplan, G. (1999b). Infect. Immun. 67, 3199. Gordon, S. (1995). Bioessays 17, 977. Gorvel, J.-P., Chavrier, P., Zerial, M., and Gruenberg, J. (1991). Cell 64, 915. Gough, P. J., and Gordon, S. (2000). Microbes. Infect. 2, 305. Gouin, E., Gantelet, H., Egile, C., Lasa, I., Ohayon, H., Villiers, V., Gounon, P., Sansonetti, P. J., and Cossart, P. (1999). J. Cell Sci. 112, 1697. Greenberg, S. (1995). Trends Cell Biol. 5, 93. Griffiths, G., and Simons, K. (1986). Science 234, 438. Groh, V., Rhinehart, R., Randolph-Habecker, J., Topp, M. S., Riddell, S. R., and Spies, T. (2001). Nat. Immunol. 2, 255. Groh, V., Steinle, A., Bauer, S., and Spies, T. (1998). Science 279, 1737. Hajjar, A. M., O’Mahony, D. S., Ozinsky, A., Underhill, D. M., Aderem, A., Klebanoff, S. J., and Wilson, C. B. (2001). J. Immunol. 166, 15. Hall, A. (1998). Science 279, 509. Halliwell, B., and Gutteridge, J. M. (1992). FEBS Lett. 307, 108.

90


Hampton, R. Y., Golenbock, D. T., Penman, M., Krieger, M., and Raetz, C. R. (1991). Nature 352, 342. Han, J., Luby-Phelps, K., Das, B., Shu, X., Xia, Y., Mosteller, R. D., Krishna, U. M., Falck, J. R., White, M. A., and Broek, D. (1998). Science 279, 558. Hanson, P. I., Roth, R., Morisaki, H., Jahn, R., and Heuser, J. E. (1997). Cell 90, 523. Hardt, W. D., Chen, L. M., Schuebel, K. E., Bustelo, X. R., and Galan, J. E. (1998). Cell 93, 815. Hart, D. N., Starling, G. C., Calder, V. L., and Fernando, N. S. (1993). Immunology 79, 616. Hasan, Z., Schlax, C., Kuhn, L., Lefkovits, I., Young, D., Thole, J., and Pieters, J. (1997). Mol. Microbiol. 24, 545. Haworth, R., Platt, N., Keshav, S., Hughes, D., Darley, E., Suzuki, H., Kurihara, Y., Kodama, T., and Gordon, S. (1997). J. Exp. Med. 186, 1431. Heinzen, R. A., Scidmore, M. A., Rockey, D. D., and Hackstadt, T. (1996). Infect. Immun. 64, 796. Hemmi, H., Takeuchi, O., Kawai, T., Kaisho, T., Sato, S., Sanjo, H., Matsumoto, M., Hoshino, K., Wagner, H., Takeda, K., and Akira, S. (2000). Nature 408, 740. Hensel, M., Shea, J. E., Waterman, S. R., Mundy, R., Nikolaus, T., Banks, G., Vazquez-Torres, A., Gleeson, C., Fang, F. C., and Holden, D. W. (1998). Mol. Microbiol. 30, 163. Heufler, C., Koch, F., Stanzl, U., Topar, G., Wysocka, M., Trinchieri, G., Enk, A., Steinman, R. M., Romani, N., and Schuler, G. (1996). Eur. J. Immunol. 26, 659. Heuser, J. E., and Reese, T. S. (1973). J. Cell Biol. 57, 315. High, N., Mounier, J., Prevost, M. C., and Sansonetti, P. J. (1992). EMBO J. 11, 1991. Hirst, J., Bright, N. A., Rous, B., and Robinson, M. S. (1999). Mol. Biol. Cell 10, 2787. Hirst, J., and Robinson, M. S. (1998). Biochim. Biophys. Acta 1404, 173. Hoffmann, J. A., Kafatos, F. C., Janeway, C. A., and Ezekowitz, R. A. (1999). Science 284, 1313. Hoiczyk, E., and Blobel, G. (2001). Proc. Natl. Acad. Sci. USA 98, 4669. Honer Zu Bentrup, K., Miczak, A., Swenson, D. L., and Russell, D. G. (1999). J. Bacteriol. 181, 7161. Horwitz, M. A. (1983). J. Exp. Med. 158, 2108. Horwitz, M. A. (1984). Cell 36, 27. Horwitz, M. A., and Maxfield, F. R. (1984). J. Cell Biol. 99, 1936. Hsia, R. C., Pannekoek, Y., Ingerowski, E., and Bavoil, P. M. (1997). Mol. Microbiol. 25, 351. Huang, S., Hendriks, W., Althage, A., Hemmi, S., Bluethmann, H., Kamijo, R., Vilcek, J., Zinkernagel, R. M., and Aguet, M. (1993). Science 259, 1742. Hurme, R., and Cossart, P. (1999). In Gordon, S. (ed.), ‘‘Phagocytosis: Microbial invasion’’, Stamford, Connecticut, JAI Press Inc, p. 59. Ireton, K., Payrastre, B., Chap, H., Ogawa, W., Sakaue, H., Kasuga, M., and Cossart, P. (1996). Science 274, 780. Isberg, R. R., and Leong, J. M. (1990). Cell 60, 861. Ishibashi, Y., and Arai, T. (1990). FEMS Microbiol. Immunol. 2, 75. Jabado, N., Jankowski, A., Dougaparsad, S., Picard, V., Grinstein, S., and Gros, P. (2000). J. Exp. Med. 192, 1237. James, S. L., Cook, K. W., and Lazdins, J. K. (1990). J. Immunol. 145, 2686. Janeway, C. A., Jr. (1989). Cold Spring Harb. Symp. Quant. Biol. 54 Pt 1, 1. Joiner, K. A., Fuhrman, S. A., Miettinen, H. M., Kasper, L. H., and Mellman, I. (1990). Science 249, 641. Jones, S. L., Knaus, U. G., Bokoch, G. M., and Brown, E. J. (1998a). J. Biol. Chem. 273, 10556. Jones, S. M., Howell, K. E., Henley, J. R., Cao, H., and McNiven, M. A. (1998b). Science 279, 573. Jouanguy, E., Lamhamedi-Cherradi, S., Altare, F., Fondaneche, M. C., Tuerlinckx, D., Blanche, S., Emile, J. F., Gaillard, J. L., Schreiber, R., Levin, M., Fischer, A., Hivroz, C., and Casanova, J. L. (1997). J. Clin. Invest. 100, 2658.


91

Kabha, K., Nissimov, L., Athamna, A., Keisari, Y., Parolis, H., Parolis, L. A., Grue, R. M., Schlepper-Schafer, J., Ezekowitz, A. R., Ohman, D. E., et al. (1995). Infect. Immun. 63, 847. Kagaya, K., Watanabe, K., and Fukazawa, Y. (1989). Infect. Immun. 57, 609. Kanaseki, T., and Kadota, K. (1969). J. Cell Biol. 42, 202. Kaplan, G. (1977). Scand. J. Immunol. 6, 797. Kaufmann, S. H. (2000). Nat. Med. 6, 955. Kaufmann, S. H., and Flesch, I. E. (1988). Springer. Semin. Immunopathol. 10, 337. Kelly, A. P., Monaco, J. J., Cho, S. G., and Trowsdale, J. (1991). Nature 353, 571. Kern, I., Steimle, V., Siegrist, C. A., and Mach, B. (1995). Int. Immunol. 7, 1295. Kery, V., Krepinsky, J. J., Warren, C. D., Capek, P., and Stahl, P. D. (1992). Arch. Biochem. Biophys. 298, 49. Kiener, P. A., Rankin, B. M., Burkhardt, A. L., Schieven, G. L., Gilliland, L. K., Rowley, R. B., Bolen, J. B., and Ledbetter, J. A. (1993). J. Biol. Chem. 268, 24442. Kindler, V., Sappino, A. P., Grau, G. E., Piguet, P. F., and Vassalli, P. (1989). Cell 56, 731. King, S. M., Patel-King, R. S., Wilkerson, C. G., and Witman, G. B. (1995). J. Cell Biol. 131, 399. Kirchhausen, T. (2000). Nat. Rev. Mol. Cell Biol. 1, 187. Klein, U., Gimpl, G., and Fahrenholz, F. (1995). Biochemistry 34, 13784. Kropshofer, H., Vogt, A. B., Moldenhauer, G., Hammer, J., Blum, J. S., and Hammerling, G. J. (1996). EMBO J. 15, 6144. Krutzik, S. R., Sieling, P. A., and Modlin, R. L. (2001). Curr. Opin. Immunol. 13, 104. Kubori, T., Matsushima, Y., Nakamura, D., Uralil, J., Lara-Tejero, M., Sukhan, A., Galan, J. E., and Aizawa, S. I. (1998). Science 280, 602. Kuhlman, M., Joiner, K., and Ezekowitz, R. A. (1989). J. Exp. Med. 169, 1733. Lanzavecchia, A. (1987). Immunol. Rev. 99, 39. Le Page, C., Genin, P., Baines, M. G., and Hiscott, J. (2000). Rev. Immunogenet. 2, 374. Lee, R. T., Ichikawa, Y., Kawasaki, T., Drickamer, K., and Lee, Y. C. (1992). Arch. Biochem. Biophys. 299, 129. Lefkowitz, D. L., Lincoln, J. A., Lefkowitz, S. S., Bollen, A., and Moguilevsky, N. (1997). Immunol. Cell Biol. 75, 136. Lehner, P. J., and Cresswell, P. (1996). Curr. Opin. Immunol. 8, 59. Lemaitre, B., Nicolas, E., Michaut, L., Reichhart, J. M., and Hoffmann, J. A. (1996). Cell 86, 973. Letourneur, F., Gaynor, E. C., Hennecke, S., Demolliere, C., Duden, R., Emr, S. D., Riezman, H., and Cosson, P. (1994). Cell 79, 1199. Levashina, E. A., Langley, E., Green, C., Gubb, D., Ashburner, M., Hoffmann, J. A., and Reichhart, J. M. (1999). Science 285, 1917. LeVine, A. M., Bruno, M. D., Huelsman, K. M., Ross, G. F., Whitsett, J. A., and Korfhagen, T. R. (1997). J. Immunol. 158, 4336. Lewis, M. J., and Pelham, H. R. (1992). Cell 68, 353. Li, Q., and Suprenant, K. A. (1994). J. Biol. Chem. 269, 31777. Li, Z., Kelley, C., Collins, F., Rouse, D., and Morris, S. (1998). J. Infect. Dis. 177, 1030. Lim, A., Eleuterio, M., Hutter, B., Murugasu-Oei, B., and Dick, T. (1999). J. Bacteriol. 181, 2252. Linehan, S. A., Martinez-Pomares, L., and Gordon, S. (2000). Microbes. Infect. 2, 279. Loisel, T. P., Boujemaa, R., Pantaloni, D., and Carlier, M. F. (1999). Nature 401, 613. Ma, Y., Uemura, K., Oka, S., Kozutsumi, Y., Kawasaki, N., and Kawasaki, T. (1999). Proc. Natl. Acad. Sci. USA 96, 371. Machesky, L. M., and Insall, R. H. (1998). Curr. Biol. 8, 1347. MacMicking, J. D., North, R. J., LaCourse, R., Mudgett, J. S., Shah, S. K., and Nathan, C. F. (1997). Proc. Natl. Acad. Sci. USA 94, 5243. Maniak, M., Rauchenberger, R., Albrecht, R., Murphy, J., and Gerisch, G. (1995). Cell 83, 915.

92


Marodi, L., Schreiber, S., Anderson, D. C., MacDermott, R. P., Korchak, H. M., and Johnston, R. B. (1993). J. Clin. Invest. 91, 2596. Marsh, M., and McMahon, H. T. (1999). Science 285, 215. Marth, T., and Kelsall, B. L. (1997). J. Exp. Med. 185, 1987. Massol, P., Montcourrier, P., Guillemot, J. C., and Chavrier, P. (1998). EMBO J. 17, 6219. Maurin, M., Benoliel, A. M., Bongrand, P., and Raoult, D. (1992). J. Infect. Dis. 166, 1097. May, R. C., Caron, E., Hall, A., and Machesky, L. M. (2000). Nat. Cell Biol. 2, 246. Medzhitov, R., and Janeway, C. (2000). Immunol. Rev. 173, 89. Mellins, E., Smith, L., Arp, B., Cotner, T., Celis, E., and Pious, D. (1990). Nature 343, 71. Mellman, I., Fuchs, R., and Helenius, A. (1986). Annu. Rev. Biochem. 55, 663. Menard, R., Prevost, M. C., Gounon, P., Sansonetti, P., and Dehio, C. (1996). Proc. Natl. Acad. Sci. USA 93, 1254. Mengaud, J., Ohayon, H., Gounon, P., Mege, R. M., and Cossart, P. (1996). Cell 84, 923. Meresse, S., Steele-Mortimer, O., Moreno, E., Desjardins, M., Finlay, B., and Gorvel, J. P. (1999). Nat. Cell Biol. 1, E183. Mills, S. D., Boland, A., Sory, M. P., van der Smissen, P., Kerbourch, C., Finlay, B. B., and Cornelis, G. R. (1997). Proc. Natl. Acad. Sci. USA 94, 12638. Miyakawa, Y., Ratnakar, P., Rao, A. G., Costello, M. L., Mathieu-Costello, O., Lehrer, R. I., and Catanzaro, A. (1996). Infect. Immun. 64, 926. Monack, D. M., Mecsas, J., Ghori, N., and Falkow, S. (1997). Proc. Natl. Acad. Sci. USA 94, 10385. Monaco, J. J. (1995). J. Leukoc. Biol. 57, 543. Mosser, D. M., and Edelson, P. J. (1987). Nature 327, 329. Murray, C. J., and Salomon, J. A. (1998). Proc. Natl. Acad. Sci. USA 95, 13881. Murray, H. W., and Teitelbaum, R. F. (1992). J. Infect. Dis. 165, 513. Neefjes, J. J., Momburg, F., and Hammerling, G. J. (1993). Science 261, 769. Neer, E. J., Schmidt, C. J., Nambudripad, R., and Smith, T. F. (1994). Nature 371, 297. Ng, V., Zanazzi, G., Timpl, R., Talts, J. F., Salzer, J. L., Brennan, P. J., and Rambukkana, A. (2000). Cell 103, 511. Nhieu, G. T., and Sansonetti, P. J. (1999). Curr. Opin. Microbiol. 2, 51. Nobes, C. D., and Hall, A. (1995). Cell 81, 53. North, R. J. (1974). Infect. Immun. 10, 66. Ochman, H., Soncini, F. C., Solomon, F., and Groisman, E. A. (1996). Proc. Natl. Acad. Sci. USA 93, 7800. Oh, Y. K., and Straubinger, R. M. (1996). Infect. Immun. 64, 319. Ozeki, Y., Kaneda, K., Fujiwara, N., Morimoto, M., Oka, S., and Yano, I. (1997). Infect. Immun. 65, 1793. Ozinsky, A., Underhill, D. M., Fontenot, J. D., Hajjar, A. M., Smith, K. D., Wilson, C. B., Schroeder, L., and Aderem, A. (2000). Proc. Natl. Acad. Sci. USA 97, 13766. Palmer, L. E., Hobbie, S., Galan, J. E., and Bliska, J. B. (1998). Mol. Microbiol. 27, 953. Parida, S. K., Domann, E., Rohde, M., Muller, S., Darji, A., Hain, T., Wehland, J., and Chakraborty, T. (1998). Mol. Microbiol. 28, 81. Pearse, B. M. (1975). J. Mol. Biol. 97, 93. Pelham, H. R., and Rothman, J. E. (2000). Cell 102, 713. Peters, C., Bayer, M. J., Buhler, S., Andersen, J. S., Mann, M., and Mayer, A. (2001). Nature 409, 581. Peters, P. J., Neefjes, J. J., Oorschot, V., Ploegh, H. L., and Geuze, H. J. (1991). Nature 349, 669. Peyron, P., Bordier, C., N’Diaye, E. N., and Maridonneau-Parini, I. (2000). J. Immunol. 165, 5186. Pfeffer, S. R. (1999). Nat. Cell Biol. 1, E17.


93

Pieters, J. (1999). In Gordon, S. (ed.), ‘‘Phagocytosis: The Host’’, Stamford, Connecticut, JAI Press Inc, p. 379. Pieters, J. (2001). Curr. Opin. Immunol. 13, 37. Pieters, J., Horstmann, H., Bakke, O., Griffiths, G., and Lipp, J. (1991). J. Cell Biol. 115, 1213. Pitt, A., Mayorga, L. S., Stahl, P. D., and Schwartz, A. L. (1992). J. Clin. Invest. 90, 1978. Pizon, V., Desjardins, M., Bucci, C., Parton, R. G., and Zerial, M. (1994). J. Cell Sci. 107, 1661. Plow, E. F., and Zhang, L. (1997). J. Clin. Invest. 99, 1145. Poltorak, A., He, X., Smirnova, I., Liu, M. Y., Huffel, C. V., Du, X., Birdwell, D., Alejos, E., Silva, M., Galanos, C., Freudenberg, M., Ricciardi-Castagnoli, P., Layton, B., and Beutler, B. (1998). Science 282, 2085. Prasad, H. K., Mishra, R. S., and Nath, I. (1987). J. Exp. Med. 165, 239. Prigozy, T. I., Sieling, P. A., Clemens, D., Stewart, P. L., Behar, S. M., Porcelli, S. A., Brenner, M. B., Modlin, R. L., and Kronenberg, M. (1997). Immunity 6, 187. Pryer, N. K., Salama, N. R., Schekman, R., and Kaiser, C. A. (1993). J. Cell Biol. 120, 865. Quadri, L. E., Sello, J., Keating, T. A., Weinreb, P. H., and Walsh, C. T. (1998). Chem. Biol. 5, 631. Rabinovitch, M. (1995). Trends Cell Biol. 5, 85. Rabinowitz, S., Horstmann, H., Gordon, S., and Griffiths, G. (1992). J. Cell Biol. 116, 95. Ramakrishnan, L., Federspiel, N. A., and Falkow, S. (2000). Science 288, 1436. Rammensee, H. G., Falk, K., and Rotzschke, O. (1993). Annu. Rev. Immunol. 11, 213. Rammensee, H. G., Rotzschke, O., and Falk, K. (1993). Chem. Immunol. 57, 113. Rapoport, I., Chen, Y. C., Cupers, P., Shoelson, S. E., and Kirchhausen, T. (1998). EMBO J. 17, 2148. Rhodes, D. A., and Trowsdale, J. (1999). Rev. Immunogenet. 1, 21. Ridley, A. J., and Hall, A. (1992). Cell 70, 389. Ridley, A. J., Paterson, H. F., Johnston, C. L., Diekmann, D., and Hall, A. (1992). Cell 70, 401. Riezman, H., Woodman, P. G., van Meer, G., and Marsh, M. (1997). Cell 91, 731. Robinson, M. S. (1989). J. Cell Biol. 108, 833. Rosqvist, R., Forsberg, A., Rimpilainen, M., Bergman, T., and Wolf-Watz, H. (1990). Mol. Microbiol. 4, 657. Rosqvist, R., Magnusson, K. E., and Wolf-Watz, H. (1994). EMBO J. 13, 964. Ross, S. C., and Densen, P. (1984). Medicine (Baltimore) 63, 243. Roy, C. R., Berger, K. H., and Isberg, R. R. (1998). Mol. Microbiol. 28, 663. Roy, C. R., and Isberg, R. R. (1997). Infect. Immun. 65, 571. Russell, D. G. (1995). Trends Cell Biol. 5, 125. Russell, D. G., Dant, J., and Sturgillkoszycki, S. (1996). J. Immunol. 156, 4764. Russell, D. G., Xu, S., and Chakraborty, P. (1992). J. Cell Sci. 103, 1193. Sadasivan, B., Lehner, P. J., Ortmann, B., Spies, T., and Cresswell, P. (1996). Immunity 5, 103. Schaible, U. E., Sturgill-Koszycki, S., Schlesinger, P. H., and Russell, D. G. (1998). J. Immunol. 160, 1290. Schekman, R., and Orci, L. (1996). Science 271, 1526. Schesser, K., Spiik, A. K., Dukuzumuremyi, J. M., Neurath, M. F., Pettersson, S., and WolfWatz, H. (1998). Mol. Microbiol. 28, 1067. Schild, H., and Rammensee, H. G. (2000). Nature 404, 709. Schlesinger, L. S., Bellinger-Kawahara, C. G., Payne, N. R., and Horwitz, M. A. (1990). J. Immunol. 144, 2771. Schlesinger, L. S., Kaufman, T. M., Iyer, S., Hull, S. R., and Marchiando, L. K. (1996). J. Immunol. 157, 4568. Schmid, S. L., McNiven, M. A., and De Camilli, P. (1998). Curr. Opin. Cell Biol. 10, 504.

94


Schneemann, M., Schoedon, G., Hofer, S., Blau, N., Guerrero, L., and Schaffner, A. (1993). J. Infect. Dis. 167, 1358. Schneider, D. S., Jin, Y., Morisato, D., and Anderson, K. V. (1994). Development 120, 1243. Schoenwaelder, S. M., and Burridge, K. (1999). Curr. Opin. Cell Biol. 11, 274. Schorey, J. S., Carroll, M. C., and Brown, E. J. (1997). Science 277, 1091. Schreiber, S., Perkins, S. L., Teitelbaum, S. L., Chappel, J., Stahl, P. D., and Blum, J. S. (1993). J. Immunol. 151, 4973. Schubert, U., Anton, L. C., Gibbs, J., Norbury, C. C., Yewdell, J. W., and Bennink, J. R. (2000). Nature 404, 770. Schupbach, T., and Wieschaus, E. (1986). Dev. Biol. 113, 443. Schwab, J. C., Beckers, C. J., and Joiner, K. A. (1994). Proc. Natl. Acad. Sci. USA 91, 509. Scianimanico, S., Desrosiers, M., Dermine, J. F., Meresse, S., Descoteaux, A., and Desjardins, M. (1999). Cell Microbiol. 1, 19. Scidmore, M. A., Fischer, E. R., and Hackstadt, T. (1996). J. Cell Biol. 134, 363. Shang, X. Z., and Issekutz, A. C. (1998). Eur. J. Immunol. 28, 1970. Shen, Y., Naujokas, M., Park, M., and Ireton, K. (2000). Cell 103, 501. Shibata, Y., Metzger, W. J., and Myrvik, Q. N. (1997). J. Immunol. 159, 2462. Sibley, L. D., Hunter, S. W., Brennan, P. J., and Krahenbuhl, J. L. (1988). Infect. Immun. 56, 1232. Silverstein, S. C. (1995). Trends Cell Biol. 5, 141. Simons, K., and Ikonen, E. (1997). Nature 387, 569. Simons, M., Keller, P., De Strooper, B., Beyreuther, K., Dotti, C. G., and Simons, K. (1998). Proc. Natl. Acad. Sci. USA 95, 6460. Sloan, V. S., Cameron, P., Porter, G., Gammon, M., Amaya, M., Mellins, E., and Zaller, D. M. (1995). Nature 375, 802. Sollner, T., Bennett, M. K., Whiteheart, S. W., Scheller, R. H., and Rothman, J. E. (1993). Cell 75, 409. Sondek, J., Bohm, A., Lambright, D. G., Hamm, H. E., and Sigler, P. B. (1996). Nature 379, 369. Spargo, B. J., Crowe, L. M., Ioneda, T., Beaman, B. L., and Crowe, J. H. (1991). Proc. Natl. Acad. Sci. USA 88, 737. Spies, T., Bresnahan, M., Bahram, S., Arnold, D., Blanck, G., Mellins, E., Pious, D., and DeMars, R. (1990). Nature 348, 744. Springer, S., Spang, A., and Schekman, R. (1999). Cell 97, 145. St John, G., Brot, N., Ruan, J., Erdjument-Bromage, H., Tempst, P., Weissbach, H., and Nathan, C. (2001). Proc. Natl. Acad. Sci. USA 98, 9901. Stahl, P. D., Rodman, J. S., Miller, M. J., and Schlesinger, P. H. (1978). Proc. Natl. Acad. Sci. USA 75, 1399. Steele-Mortimer, O., Meresse, S., Gorvel, J. P., Toh, B. H., and Finlay, B. B. (1999). Cell Microbiol. 1, 33. Steimle, V., Siegrist, C. A., Mottet, A., Lisowska-Grospierre, B., and Mach, B. (1994). Science 265, 106. Stenger, S., Hanson, D. A., Teitelbaum, R., Dewan, P., Niazi, K. R., Froelich, C. J., Ganz, T., Thoma-Uszynski, S., Melian, A., Bogdan, C., Porcelli, S. A., Bloom, B. R., Krensky, A. M., and Modlin, R. L. (1998). Science 282, 121. Stuehr, D. J., and Nathan, C. F. (1989). J. Exp. Med. 169, 1543. Stura, E. A., Matsumura, M., Fremont, D. H., Saito, Y., Peterson, P. A., and Wilson, I. A. (1992). J. Mol. Biol. 228, 975. Sturgill-Koszycki, S., Schaible, U. E., and Russell, D. G. (1996). EMBO J. 15, 6960. Sturgill-Koszycki, S., Schlesinger, P. H., Chakraborty, P., Haddix, P. L., Collins, H. L., Fok, A. K., Allen, R. D., Gluck, S. L., Heuser, J., and Russell, D. G. (1994). Science 263, 678.


95

Sugita, M., Peters, P. J., and Brenner, M. B. (2000). Traffic 1, 295. Suss-Toby, E., Zimmerberg, J., and Ward, G. E. (1996). Proc. Natl. Acad. Sci. USA 93, 8413. Sutton, R. B., Fasshauer, D., Jahn, R., and Brunger, A. T. (1998). Nature 395, 347. Suzuki, H., Kurihara, Y., Takeya, M., Kamada, N., Kataoka, M., Jishage, K., Ueda, O., Sakaguchi, H., Higashi, T., Suzuki, T., Takashima, Y., Kawabe, Y., Cynshi, O., Wada, Y., Honda, M., Kurihara, H., Aburatani, H., Doi, T., Matsumoto, A., Azuma, S., Noda, T., Toyoda, Y., Itakura, H., Yazaki, Y., and Kodama, T. (1997). Nature 386, 292. Swanson, J. A., and Daer, S. C. (1995). Trends Cell Biol. 5, 89. Swanson, J. A., Johnson, M. T., Beningo, K., Post, P., Mooseker, M., and Araki, N. (1999). J. Cell Sci. 112, 307. Swanson, M. S., and Isberg, R. R. (1995). Infect. Immun. 63, 3609. Takei, K., McPherson, P. S., Schmid, S. L., and De Camilli, P. (1995). Nature 374, 186. Takeuchi, O., Hoshino, K., Kawai, T., Sanjo, H., Takada, H., Ogawa, T., Takeda, K., and Akira, S. (1999). Immunity 11, 443. Tamura, N., Lottspeich, F., Baumeister, W., and Tamura, T. (1998). Cell 95, 637. Tenner, A. J. (1999). Curr. Opin. Immunol. 11, 34. Tenner, A. J., Robinson, S. L., and Ezekowitz, R. A. (1995). Immunity 3, 485. Thoma-Uszynski, S., Stenger, S., Takeuchi, O., Ochoa, M. T., Engele, M., Sieling, P. A., Barnes, P. F., Rollinghoff, M., Bolcskei, P. L., Wagner, M., Akira, S., Norgard, M. V., Belisle, J. T., Godowski, P. J., Bloom, B. R., and Modlin, R. L. (2001). Science 291, 1544. Tilney, L. G., and Portnoy, D. A. (1989). J. Cell Biol. 109, 1597. Toker, A. (1998). Curr. Opin. Cell Biol. 10, 254. Tonegawa, S. (1983). Nature 302, 575. Tonegawa, S. (1988). Biosci. Rep. 8, 3. Townsend, A., and Bodmer, H. (1989). Ann. Rev. Immunol. 7, 601. Trimble, W. S., Cowan, D. M., and Scheller, R. H. (1988). Proc. Natl. Acad. Sci. USA 85, 4538. Tulp, A., Verwoerd, D., Dobberstein, B., Ploegh, H. L., and Pieters, J. (1994). Nature 369, 120. Ulrichs, T., and Porcelli, S. A. (2000). Rev. Immunogenet. 2, 416. Vachula, M., Holzer, T. J., and Andersen, B. R. (1989). J. Immunol. 142, 1696. van der Bliek, A. M., and Meyerowitz, E. M. (1991). Nature 351, 411. van Leeuwen, F. N., van Delft, S., Kain, H. E., van der Kammen, R. A., and Collard, J. G. (1999). Nat. Cell Biol. 1, 242. Via, L. E., Deretic, D., Ulmer, R. J., Hibler, N. S., Huber, L. A., and Deretic, V. (1997). J. Biol. Chem. 272, 13326. Via, L. E., Fratti, R. A., McFalone, M., Pagan-Ramos, E., Deretic, D., and Deretic, V. (1998). J. Cell Sci. 111, 897. Vidal, S., Gros, P., and Skamene, E. (1995). J. Leukoc. Biol. 58, 382. Vogel, J. P., and Isberg, R. R. (1999). Curr. Opin. Microbiol. 2, 30. von Figura, K., and Hasilik, A. (1986). Ann. Rev. Biochem. 55, 167. Wall, M. A., Coleman, D. E., Lee, E., Iniguez-Lluhi, J. A., Posner, B. A., Gilman, A. G., and Sprang, S. R. (1995). Cell 83, 1047. Wardle, E. N. (1987). Liver 7, 63. Welch, M. D., Rosenblatt, J., Skoble, J., Portnoy, D. A., and Mitchison, T. J. (1998). Science 281, 105. West, M. A., Lucocq, J. M., and Watts, C. (1994). Nature 369, 147. Wheeler, P. R., and Ratledge, C. (1988). J. Gen. Microbiol. 134, 2111. Wilkerson, C. G., King, S. M., Koutoulis, A., Pazour, G. J., and Witman, G. B. (1995). J. Cell Biol. 129, 169. Wilmott, R. W., Khurana-Hershey, G., and Stark, J. M. (2000). Curr. Opin. Pediatr. 12, 187.

96


Wright, S. D., and Silverstein, S. C. (1983). J. Exp. Med. 158, 2016. Xia, Y., Vetvicka, V., Yan, J., Hanikyrova, M., Mayadas, T., and Ross, G. D. (1999). J. Immunol. 162, 2281. Xu, S., Cooper, A., Sturgill-Koszycki, S., van Heyningen, T., Chatterjee, D., Orme, I., Allen, P., and Russell, D. G. (1994). J. Immunol. 153, 2568. Yamamoto, Y., Klein, T. W., and Friedman, H. (1997). Infect. Immun. 65, 1077. Yang, R. B., Mark, M. R., Gray, A., Huang, A., Xie, M. H., Zhang, M., Goddard, A., Wood, W. I., Gurney, A. L., and Godowski, P. J. (1998a). Nature 395, 284. Yang, R. B., Mark, M. R., Gray, A., Huang, A., Xie, M. H., Zhang, M., Goddard, A., Wood, W. I., Gurney, A. L., and Godowski, P. J. (1998b). Nature 395, 284. Zhou, M. J., and Brown, E. J. (1994). J. Cell Biol. 125, 1407. Zimmerli, S., Edwards, S., and Ernst, J. D. (1996). Am. J. Respir. Cell Mol. Biol. 15, 760.


B Lymphoid Neoplasms of Mice: Characteristics of Naturally Occurring and Engineered Diseases and Relationships to Human Disorders HERBERT C. MORSE III, TOM McCARTY, CHEN-FENG QI, TED A. TORREY, ZOHREH NAGHASHFAR, SISIR K. CHATTOPADHYAY, TORGNY N. FREDRICKSON, AND JANET W. HARTLEY Laboratory of Immunopathology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland, USA

I. Introduction

A. A HISTORICAL PERSPECTIVE The origins of the mouse as a model for human neoplastic diseases dates to the first part of the 20th century, when Clarence Cook Little, E. C. MacDowell, Leonell Strong, Leo Loeb, Ernest Tyzzer, and others postulated from studies of mice that cancers were genetically determined (Morse, 1981). With remarkable foresight, they decided that genetically homogeneous inbred strains would be essential in unraveling host determinants of neoplastic diseases and set about generating the ancestral stocks of most inbred strains in use today. Using partially inbred strains, they demonstrated clear genetic links to the development of tumors and the occurrence of specific types of solid as well as lymphoid neoplasms. Leukemias were also known to occur at low frequency in some strains of mice, but animals diagnosed with the disease were usually moribund or too old to breed for purposes of developing inbred strains; however, lymphoid diseases were found to occur at greatly increased incidence in mice treated with ionizing radiation (Krebs et al., 1930), estrogens (Lacassagne, 1937), and coal tar derivatives (Brues and Marble, 1939). During this time frame, MacDowell was successful in developing the highly leukemic inbred strain, C58 (MacDowell and Richter, 1935; Richter and MacDowell, 1935) and reported the first transplantable cell line. Unfortunately, he did not make the mice accessible to others. A strain with a reproducibly high incidence of leukemia and a relatively short latency became generally available for study only with the development of the strain AKR by Jacob Furth (Cole and Furth, 1941). The demonstration that thymectomy of young mice prevented the development of disease and that thymus transplants restored susceptibility tied the disorder to a tissue recognized until then simply as another lymphoid organ, long before a role for the thymus in T cell development was established. 97 0065-2776/03 $35.00

98

HERBERT C. MORSE ET AL.

A causal role for viruses in mouse mammary tumor development was established in the 1930s for mammary tumor viruses (MMTV) (Staff et al., 1933; Korteweg, 1936). Although viruses were known to cause lymphoid leukemia and lymphomas in chickens and were thus suspected to have a role in mouse leukemia as well, repeated attempts to infect adult mice with filterable agents failed to induce disease. Then, in 1950, Ludwik Gross showed that cell-free extracts of AKR thymic tumors could induce the same disease in C3H mice inoculated as newborns (Gross, 1951). These studies were critical in uncovering the contributions of what came to be known as murine leukemia viruses (MuLV) to thymic lymphomas (Hartley et al., 1977; Rosenberg and Jolicoeur, 1997) and the much more recent unraveling of molecular mechanisms in the pathogenesis of T- as well as B-cell malignancies. Transplantable tumors obtained from some of these mice provided the first testing ground for chemotherapeutic agents (Law, 1951). Further passage study of tumor extracts uncovered MuLV that were oncogenic in vivo—either acutely or with latencies of less than a year—and sometimes in vitro. Molecular analyses of replication-defective acutely transforming MuLV, such as v-abl (Goff et al., 1980), revealed that they had captured oncogenic forms of cellular proto-oncogenes as presaged by earlier studies of acutely transforming chicken retroviruses. Viruses that induced specific tumors with longer latencies, such as Moloney MuLV and thymic lymphomas, were found to be full length and replication competent (Shinnick et al., 1981). By the early 1980s, then, the mouse had provided a variety of insights into cancer biology—the importance of genetic factors and contributions of infectious agents and chemicals being at the top of the list. What was missing, however, was the ability to link phenotype with genotype and then ascertain how specific genes contributed to tumor pathogenesis. All this changed in 1984 and 1985 with the groundbreaking experiments of several laboratories that yielded tumor-prone mice through the use of transgenes to over express viral or cellular oncogenes in specific tissues (Brinster et al., 1984; Stewart et al., 1984; Adams et al., 1985; Hanahan, 1985). The E-myc mouse of Adams et al. was remarkably informative in showing that tissue-specific regulatory sequences linked to a proto-oncogene in a manner equivalent to that found in translocations of human Burkitt lymphoma (BL) and mouse plasmacytoma (PCT) were lymphomagenic for B cells. Because the lymphomas in these mice were mostly pre-B cell in origin, they failed to accurately model human BL. Nonetheless, the studies provided the conceptual framework for the contention that it should be possible to model in mice any human neoplasm with an established molecular mechanism. The use of gene targeting technology in mouse embryonic stem (ES) cells to inactivate the prototype tumor suppressor genes Rb (Jacks et al., 1992) and p53 (Donehower et al., 1992) forcefully extended this concept while showcasing

B LYMPHOID NEOPLASMS OF MICE

99

the powers of targeted recombination. The most recent technologic breakthroughs in this field allow the engineering of tissue- and stage of differentiation-specific mutations. Over the last several years, a number of publications have served to demonstrate the power of gene expression profiling in understanding the biology of a variety of human tumors including non-Hodgkin lymphomas (Alizadeh et al., 2000). This approach made possible molecular characterization of previously unexpected subtypes of cancers (Alizadeh et al., 2000; Clark et al., 2000) and uncovered the existence of critical signaling pathways that distinguish these subsets (Davis et al., 2001; Shipp et al., 2002), thereby identifying new rational targets for intervention. Applications of microarray technology to studies of mouse tumors are much less advanced but should provide a powerful means for validating spontaneous or induced models of human lymphoma and leukemia. B. OF MICE

AND

MEN: MICE ARE

NOT

SMALL HUMANS

The underlying premise for all this work is the understanding of the pathogenetic mechanisms involved in mouse lymphomas and leukemia that will permit the discovery of new cancer-related genes, unappreciated signaling pathways that are affected in common with related human neoplasms and, as a result, provide opportunities for developing new therapeutic approaches. Hope along these lines is greatest for diseases for which the molecular pathogenesis has been determined in humans and that are thus potentially amenable to recreation in the mouse by genetic manipulation. It is essential, however, to understand that fundamental differences between mice and men may preclude the exact modeling of any human disease. These distinctions extend beyond fur, tails, weight (3000 : 1), growth rate after birth (100 : 1), and rate of aging after maturity (30 : 1). Several disparities are of obvious importance for thinking about lymphoid neoplasms in mice. First is the persistence of the thymus into adult life, providing a population of highly proliferative cells that may be particularly susceptible to transformation. This may explain the frequency with which thymic T-cell-lymphomas, in particular, but also B-cell-lineage lymphoma appear in a substantial number of genetically engineered mice (GEM) in which the manipulation would not have been expected to have lymphoma as a readout. These can be considered ‘‘accidental tumors’’ in many instances. For example, mice deficient in the DNA mismatch repair gene associated with hereditary non-polyposis colon cancer, MSH2, develop a high indigence of both T- and B-cell-lineage lymphomas (Dewind et al., 1995; Reitmair et al., 1995) even though this gene has not been found to be mutated in human lymphoid malignancies; however, lymphomas do occur in patients with mutations in another mismatch repair gene, MLH1 (Ricciardone et al., 1999).

100


Second, while lymph nodes are the dominant secondary lymphoid compartment in humans and the source of most B-cell-lineage neoplasms, the spleen plays these roles in mice. The lifelong extramedullary hematopoiesis that occurs in mouse but not human spleens also provides a markedly different histologic context for determining lymphoma types. Recent studies have also shown species-specific differences in the nature of the human and rodent splenic marginal zones (Steiniger et al., 2001), and splenic marginal zone lymphomas of mice and humans seem to be quite different entities (Morse et al., 2001a). Finally, mice are not susceptible to infection with Epstein-Barr virus (EBV) or HTLV-1, which are lymphomagenic in humans. Patterns of differentiation associated with expression of specific markers or the effects of mutation in specific signaling pathways are likewise sometimes divergent for mice and humans. For example, human germinal center B cells and plasma cells are CD38-positive, while those of the mouse are CD38-dull or -negative (Oliver et al., 1997). Most humans with mutations of the BTK gene have no B cells in blood, spleen, or nodes and have profound hypogammaglobulinemia, while mutant mice exhibit only modestly reduced numbers of B cells and levels of immunoglobulin (Kerner et al., 1995; Khan et al., 1995). Furthermore, IL7R-deficient mice have no B cells (Peschon et al., 1994), while similarly affected humans have normal numbers (Puel et al., 1998). How well spontaneous neoplasms or those in GEM can be expected to model human lymphoid neoplasm may critically hinge on species-specific differences in the number of genetic changes required for tumor induction, as repeated observations suggest that fewer are needed in the mouse. These differences appear, in turn, to reflect divergent barriers to immortalization and responses to introduced oncogenes. Immortalization is essential to the development of a tumor cell, and studies of human cells have shown that immortalization can be achieved by overcoming two lines of defense— replicative senescence and cellular crisis. This can be effected by disabling the Rb and p53 tumor-suppressor pathways and activating telomerase (reviewed by Hahn and Weinberg, 2002). In contrast, repression of ARF-p53 signaling pathway is all that is required to immortalize many mouse cells (Kamijo et al., 1998). The fact that expression of telomerase is constitutive in most somatic tissues of inbred mice but repressed in human cells explains part of this difference; however, the basis for the distinctive dependence on the two suppressor pathways—p53 in the mouse and p53 plus Rb in human cells—remains without explanation. The responses of mouse and human cells to the combination of Myc or Myc plus Ras were also revealing. Full transformation of mouse hematopoietic cells by Myc could be achieved using cells from p53-deficient mice (Metz et al., 1995). In addition, transformation of embryonic fibroblasts by Myc plus Ras


101

was observed in cells lacking either p53 or ARF (Lin and Lowe, 2001). In contrast, human embryonic kidney cells or mammary epithelial cells could be transformed only when both the p53 and Rb suppressor pathways were disarmed by the presence of SV40 early region sequences (Hahn and Weinberg, 2002). These observations indicate that there are likely to be some human diseases that cannot be modeled at all in the mouse and others that may be only partially successful due to intrinsic species differences. The latter view is strongly buttressed by studies of mice with germline mutations that reproduce those seen in humans with inborn errors that predispose them to cancer. The spectrum of cancers seen in these mice is often at great variance with the range of tumors seen in affected humans (summarized in Hahn and Weinberg, 2002). The influence of environmental factors on the incidence and time course of mouse lymphoid neoplasms is substantial but poorly understood. For example, BALB/c mice raised under conventional conditions develop a high incidence of plasmacytomas (PCT) following intraperitoneal injections of pristane, while mice of the same strain raised specific pathogen-free (SPF) are almost completely resistant (Byrd et al., 1991). Similar effects of environment on the incidence of PCT were reported for mice bearing an E-v-abl transgene (Symons et al., 2002). A suggestion that inhibition of prostaglandin production may be partly responsible for this effect was suggested by the finding that treatment with indomethacin blocks PCT induction by pristane or other agents (Potter et al., 1985). More recent studies have shown that the time course for lymphoma development in NFS.V þ mice raised SPF is delayed by approximately 5 months as compared with animals raised under conventional conditions (Hartley et al., unpublished observations). The reasons for this shift are almost certainly different from those responsible for the differences in PCT induction between SPF and conventional mice, because treatment of conventional NFS.V þ mice with indomethacin had no effect on the time course of disease (Hartley et al., unpublished observations). Susceptibility to neoplasia in mice and humans is influenced by well established factors such as tumor immunity but also by genetic modifiers of tumor susceptibility and resistance phenotypes. Striking strain-dependent differences in susceptibility to PCT development following treatment with pristane are determined by at least five loci, one of which was shown to be an ‘‘efficiency’’ allele of the p16INK4a gene (Zhang et al., 2001). Loci modifying strain- as well as sex-dependent differences in PCT-related mortality in E-vabl transgenic mice have also been described (Symons et al., 2002). Last, it must be recognized that most of our detailed knowledge of mouse lymphoid neoplasms derives primarily from studies of spontaneous disease in a relatively limited number of inbred strains, particularly those that express high levels of MuLV. This is changing dramatically with the proliferation of strains

102


of GEM. Indeed, diseases have been seen in GEM that have never been found to occur spontaneously. How these unique disorders can be used to understand human conditions will be governed in part by the purposes for which they are employed. Directed expression of a fusion gene that reproduces a human protein determined by translocation, such as BCR/ABL tyrosine kinase, could be used to study the effects of inhibitors on the oncogene or on pathways activated downstream by the oncogene that cannot be reproduced in cultured cells; however, the farther one moves from the ‘‘inciting event,’’ the more difficult it might become to relate effects to human biology.

II. Classification of Mouse B Cell-lineage Lymphomas

A. BACKGROUND Classifications of human hematopoietic neoplasms have gone through an extensive series of iterations that recently culminated in the publication of the consensus World Health Organization blue book on Tumors of Hematopoietic and Lymphoid Tissues (Jaffe et al., 2001). The approach to classification detailed therein emphasizes the use of all available sources of information to define a disease entity—clinical presentation, morphology, immunophenotype, and genetic features. For each of the lymphoid diseases, a presumed cell of origin is postulated to reflect the state of differentiation exhibited by the tumor population (Fig. 1). While precursor-B-cell-lymphoblastic (pre-B LBL) lymphoma/leukemia derives from bone marrow pre-B cells and mantle cell lymphoma from resting non-mutated B-cells of the mantle zone, most human B-cell-lineage lymphomas are of germinal center (GC) or post-GC origin (Jaffe et al., 2001). The GC is the site for ordered alterations in the DNA of B cells— somatic hypermutation and Ig isotype switching. The normal GC reaction is characterized histologically by dark and light zones. The dark zone is populated by ‘‘B blasts’’ (lymphoblasts), the precursors of cells in the light zone. Centroblasts accumulate at the juncture of the zones and migrate to the light zone comprised of immunoblasts (cells on their way to being plasma cells), centroblasts, and centrocytes. Burkitt, follicular, and diffuse large-cell lymphomas bear morphologic, phenotypic, and molecular marks of the GC reaction. All CLL have the phenotype of memory, post-GC B cells, although some appear to be pre-GC in having germline IgV region sequences (Klein et al., 2001; Rosenwald et al., 2001). Splenic marginal zone lymphomas segregate for those with mutated and those with non-mutated IgV regions, thus appearing to arise from both naive and post-GC B cells (Algara et al., 2002). The relative weight given to clinical history, morphology, immunophenotype, and molecular characteristics in making a diagnosis varies from one


103

FIG. 1. Relationships of human B-cell-lineage lymphomas to normal B-cell differentiation. Normal B-cell-specific mechanisms of gene rearrangement and diversification through somatic mutation and class switch recombination are shown at the top along with the cell types in which they occur. The germinal center reaction with its rich mixture of B cells and dendritic cells is emphasized, because most human B-cell-lineage derive from cells in this anatomic compartment or their progeny. Lymphoma types of precursor B-cell lymphoblastic (pre-B LBL), mantle cell (MCL), Burkitt (BL), follicular (FL), diffuse large B-cell (DLBCL), chronic lymphotic leukemia/small lymphocytic lymphoma (CLL/SLL), multiple myeloma (MM), and plasmacytoma (PCT) are shown in bold in relation to their normal cell counterparts. Marginal zone B cells that surround primary follicles and their transformed counterpart, marginal zone lymphoma (MZL), are not shown.

entity to another. While morphology and immunophenotype may dominate the process of classification for one disorder, the detection of a specific molecular feature may be a prerequisite for another. For diseases of mice, Thelma Dunn was the first to develop a classification of mouse hematopoietic neoplasms (Dunn, 1954), with a series of other schemes appearing through the 1990s. Most, representing the work of one or a few individuals, adopted terminologies from previously published human classifications in an attempt to relate mouse neoplasms to human disorders (Pattengale and Taylor, 1983; Frith et al., 1993; Fredrickson et al., 1994). Unfortunately, these taxonomies were used irregularly or not at all used in publications describing hematopoietic tumors. This limited the opportunities for the scientific community to understand if a disease was new or resembled the one described in the past. These considerations prompted an international

104


panel of experts in mouse and human pathology and molecular biologists involved in model building to develop a consensus list of neoplasms with descriptions and criteria for diagnosis. It was reasoned that such a uniform classification would ‘‘enhance the value of the scientific literature and will make it possible to meaningfully compare and contrast murine lesions with human lesions (Dunn, 1954).’’ Using the WHO classification as a model, the resulting ‘‘Bethesda proposals’’ provide the first consensus classification of mouse lymphoma/leukemia (Kogan et al., 2002; Morse et al., 2002). The lymphomas evaluated to assemble this nomenclature include spontaneous neoplasms, lymphomas occurring in mice infected with acutely transforming MuLV, mice reconstituted with retrovirus oncogene-transduced bone marrow cells, and transgenic mice as well as knockout and knockin mice. The resulting classifications of mouse B-cell-lineage neoplasms in the Bethesda proposals and their closest human counterparts from the WHO classification are listed in Table I. In reviewing these comparisons, it must be recognized that each of the human disorders has been classified using the full range of morphologic, phenotypic, clinical, and molecular parameters, while the mouse categories have been defined primarily on the basis of histopathology, with variable contributions from analyses of immunophenotype, clonality of immune receptor gene rearrangements, and expression of a limited number of genes determined by RT-PCR. With this caveat in mind, the range of well-defined neoplasms in humans is broader than that for the mouse. Nonetheless, there are examples of neoplasms from either species not seen in the other. Most of the mouse diagnostic categories have been observed to occur spontaneously, but two—Burkitt and primary mediastinal (thymic) diffuse large-B-cell lymphoma (DLBCL)—have been seen only in transgenic mice (Kovalchuk et al., 2000) and mice infected with a replication-defective MuLV, BM5def (Morse, unpublished observations), respectively. Several of the named categories for mouse diseases are identical to those for humans, while others are similar but distinct. With one exception, splenic marginal zone lymphoma (MZL), those with identical designations are felt to exhibit enough parallels to suggest they represent the same disease but in different species. Mouse MZL was given that designation because of clear evidence for its origins from splenic marginal zone B cells (Fredrickson et al., 1999). Human MZL, in contrast, may actually derive from another B-cell type and thus be misnamed (Morse et al., 2001b). For those with similar but different names, mouse small-B-cell lymphoma (SBL) was labeled as distinct from human small lymphocytic lymphoma/CLL because when leukemia occurs, it appears to be secondary ‘‘spillover’’ from a pre-existing lymphoma rather than the other way around. Mouse follicular B-cell-lymphoma (FBL) appears to be decomposed of the same cell types as


HUMAN Human classification

AND

105

TABLE I MOUSE B CELL-LINEAGE LYMPHOMAS/LEUKEMIAS Mouse classification

Immature B cell neoplasms Precursor B-cell-lymphoblastic Precursor B-cell-lymphoblastic lymphoma/leukemia lymphoma/leukemia Mature B cell neoplasms Chronic lymphocytic leukemia/ Small-B-cell lymphoma small lymphocytic lymphoma B cell prolymphocytic leukemia Lymphoplasmacytic lymphoma/ Waldenstrom’s macroglobulinemia Splenic marginal zone (MZ) lymphoma Splenic MZ lymphoma Hairy cell leukemia Plasma cell neoplasms Monoclonal gammopathy of undetermined significance Plasmacytoma Solitary plasmacytoma of bone Extraosseus plasmacytoma Extraosseus plasmacytoma Plasmacytoma Anaplastic plasmacytoma Primary amyloidosis Heavy chain disease Extranodal MZB cell lymphoma of mucosa-associated lymphoid tissue Follicular lymphoma Follicular B cell lymphoma Mantle cell lymphoma Diffuse large B cell lymphoma (DLBCL) DLBCL Morphologic variants Centroblastic Centroblastic Immunoblastic Immunoblastic T-cell-rich/histiocyte-rich Histiocyte-associated Subtype Mediastinal (thymic) large B cell lymphoma Primary mediastinal (thymic) DLBCL Intravascular large B cell lymphoma Primary effusion lymphoma Burkitt lymphoma Burkitt lymphoma Burkitt-like lymphoma B-natural killer cell lymphoma

human follicular lymphoma, but the characteristic follicular feature of the human disease is lacking in mice. Finally, the histiocytic component of mouse histiocyte-associated DLBCL rarely exceed 70%, while 90–95% of the cells are histiocytes in human histiocyte-rich DLBCL. A mouse category never seen in humans is B-natural killer cell lymphoma (Lu and Hiai, 1999), while human disorders never seen in the mouse include

106


prolymphocytic, lymphoplasmacytic, hairy cell, several plasma cell, extranodal MZL, mantle cell and intravascular lymphomas and leukemia, as well as monoclonal gammopathy of undetermined significance (MGUS). Two large series of lymphomas have been studied morphologically and molecularly for the occurrence of lymphomas of different types: NFS.V þ congenics (Hartley et al., 2000) and AKXD RI strains (Morse et al., 2001b). The two groups of mice differed significantly for the frequency of specific nonT lineage lymphomas. For example, SBL comprised 13.1% of the congenic but only 1.1% of the RI B-lineage neoplasms. Even more striking was the finding that not one DLBCL of CB or IB type was found among the RI cases, while over 21% of the congenic cases were of these types. Conversely, cases of DLBCL (HA) are very rare among congenic lymphomas but represented 15.2% of the B-cell-lineage lymphomas in the RI series. Nonetheless, the histologic features of specific B-cell-lineage subsets seen in either group were indistinguishable for histologic and cytologic features. The features of each subset are detailed elsewhere (Fredrickson and Harris, 2000; Taddesse-Heath and Morse, 2000; Morse et al., 2002). B. SPONTANEOUS MOUSE B CELL NEOPLASMS The following is a brief description of the occurrence and some features of spontaneous mouse B cell neoplasms. 1. Precursor-B Cell Lymphoblastic Leukemia/Lymphoma This lymphoma type with features of pre-B cells occurs with regularity in SL /Kh mice, where it is well characterized (Yamada et al., 1994) and also seems to occur at an appreciable frequency in the AKXD RI strains (Mucenski et al., 1986; Morse et al., 2001b). The latter presumption is based on molecular criteria only—lymphomas rearranged for IgH but not IgL. No flow cytometric or other studies to confirm the state of differentiation have been performed. 2. Small B Cell Lymphoma This lymphoma occurs primarily in old mice of several strains. Less than a third of cases have an associated leukemia, which is always secondary to the lymphoma. This contrasts with human SLL/CLL, where the leukemia is primary. 3. Splenic Marginal Zone Lymphoma First documented to occur in NZB mice, this lymphoma exhibits a striking progression from low to high grade (Fredrickson et al., 1999). Simple expansion of the splenic marginal zone, usually only one to two cell layers thick, is the earliest sign of disease. The most advanced form, arising with


107

continued expansion of the marginal zone, has cytologic features often indistinguishable from those of DLBCL or CB type. High-grade SMZL can lead to marked compression of the white pulp and invasion of the red pulp. Spread to the splenic node, and more rarely, the liver is uncommon. Human SMZL IgV region sequences can be germline or mutated similar to the sequences seen with normal marginal zone B cells. There is no comparable information on mouse SMZL. In addition, all SMZL were negative for BCL6 by immunohistochemistry and immunoblotting (Qi et al., 2000). 4. Follicular B Cell Lymphoma The normal germinal center reaction is characterized histologically by dark and light zones. The dark zone is populated by B blasts, the precursors of cells in the light zone. With an ovoid nucleus, stippled chromatin, and inconspicuous nucleoli, these resemble the cells of lymphoblastic lymphomas. Centroblasts accumulate at the juncture of these zones and migrate to the light zone comprised of immunoblasts (cells on their way to being plasma cells), centroblasts, and centrocytes. Follicular lymphomas in humans and mice are a mixture of transformed centroblasts and centrocytes that no longer exhibit the spatial relationships they would in a germinal center. In both species, they occupy the follicles, but only in humans do they exhibit true follicular structures. Analyses of a limited number of IgH V region sequences (D. Zhu, T. Torrey, F. Stevenson, H. Morse, unpublished observations) indicate that they are not mutated, while those of humans are; however, a FBL from a TCL1 transgenic mouse was found to have mutated sequences (Hoyer et al., 2002), indicating that further work needs to be done before any firm conclusions can be made. Most cases studied were BCL6 negative by immunohistochemistry and immunoblotting, also suggesting that they differ from human follicular lymphoma (Qi et al., 2000). 5. Diffuse Large B Cell Lymphomas The histologic subtypes of centroblastic and immunoblastic DLBCL are optional for use in the classification of human lymphoma. Centroblastic lymphomas comprise uniform populations of the same centroblastic cells as seen in FBL, while immunoblastic lymphomas share many features of the GC immunoblasts. Mouse CB lymphomas are histologically quite consistent, a feature supported by the unity of this lymphoma type seen with gene expression profiling (see below). Sheets of immunoblasts are very uncommon, and in IB, there is an admixture of immunoblasts with centroblasts and sometimes plasma cells. Most CB and IB are BCL6 positive by immunohistochemistry and immunoblotting, indicating they are of GC or post-GC origin (Qi et al., 2000).

108


An unusual subset of DLBCL arising in the thymus was observed in mice infected helper-free with a replication defective virus, BM5def (H. Morse et al., unpublished observations). When inoculated together with helper viruses, BM5def induces a retrovirus-induced immunodeficiency syndrome termed MAIDS (reviewed in Rosenberg and Jolicoeur, 1997). The origins and histologic features exhibited parallels with human primary mediastinal (thymic) DLBCL. Future studies for BCL6 and gene expression profiling should shed more light on this unusual lymphoma type. 6. Burkitt Lymphoma A lymphoma type not seen to occur spontaneously in mice was observed in all mice carrying a human MYC transgene regulated by human Ig 30 regulatory sequences (-MYC mice; Kovalchuk et al., 2000). The cytology was essentially indistinguishable from human BL and different from that seen in E-myc transgenic mice (Adams et al., 1985). Human BL almost always exhibit hypermutation of IgV region sequences and appears to be of post-GC origin, although this is somewhat controversial. Recently, our collaborators and we have found that the lymphomas of -MYC mice have germline IgH V regions and that the disease develops with equal kinetics in Bcl6 knockout mice that are unable to form germinal centers (T. Torrey, G. Catoretti, D. Zhu, F. Stevenson, R. Dalla-Favera, H. Morse, manuscript in preparation). Finally, the mice express cell surface antigens found on normal pre-B cells even though they are IgM positive, but they are BCL6 negative by immunohistochemistry. These results indicate that the lymphomas of -MYC mice are pre-GC and may be most like tumors of transitional or immature B cells. The nomenclature committee will have to reconsider this classification, but it is likely it will be changed. 7. Burkitt-Like Lymphoma This terminology was adopted for lymphomas with morphologic features of lymphoblasts that are sIg þ . Other lymphomas that are indistinguishable histologically from these derive from precursor B cells or T lymphocytes. These neoplasms would appear to be of post-GC origin, as all are BCL6 positive by immunohistochemistry and immunoblotting and were found to have rearrangements of Bcl6 genomic structure as for human DLBCL (Qi et al., 2000). No sequence analyses of lymphomas of this class from NFS congenic mice have been performed to determine whether IgV regions are mutated, but those of TCL1 transgenic mice are (M. Teitell, H. Morse, unpublished observations). A change in the nomenclature for BL would clearly necessitate a change in terminology for this neoplasm. While the Bethesda proposals were in draft form, thought was given for including these tumors as a subset of DLBCL. Further study may provide increasingly strong support to this view.


109

8. Neoplasms of Plasma Cells Studies of mouse plasma cell neoplasms originated with now-classic studies by Michael Potter, and his deep involvement in this work continues to the present (reviewed in part in Potter and Wiener, 1992). His studies are based on analyses of mice injected intraperitoneally with a mineral oil, pristane, or silicon that develop plasmacytomas (PCT). The tumors develop as implants on the omentum. Spontaneous plasma cell neoplasms occur in old SJL mice, mice mutant for Fas or Fas ligand (Davidson et al., 1998), and occasionally in NFS.V þ mice. In addition, some NFS congenic mice develop anaplastic PCT (T. Fredrickson, J. Hartley, unpublished observations). Recently, it was found that BALB/c mice bearing an IL-6 transgene exhibit a very high frequency of PCT involving the spleen and lymph nodes (Kovalchuk et al., 2002). Essentially all pristane-induced and IL-6-associated PCT have T(12;15) that activate the Myc locus (Potter and Wiener, 1992), while the Fas and Fas ligand neoplasms do not (Davidson et al., 1998). The nomenclature for mouse lymphoid neoplasms is a work in progress. The findings that suggest that BL and BLL may have to be renamed exemplifies the type of evolution to be expected as we come to learn more about the genetic basis of disease pathogenesis. The structure will also change as new models of human neoplasms appear and are studied in depth. III. Pathogenesis of Spontaneous B Cell lineage Neoplasms

A. PROVIRAL INSERTIONAL MUTAGENESIS Most of the well understood genetic errors associated with human B celllineage neoplasms are occasioned by recurrent chromosomal translocations or inversions. These result in the generation of fusion genes encoding chimeric oncoproteins or, more frequently, aberrant expression of cellular protooncogenes. These are deemed to be the primary events in transformation, which are later complemented by silencing of tumor suppressors by mutation, deletion, or promoter methylation. Similar primary contributions of translocations to mouse lymphoid neoplasms have been documented as a regular finding only for plasmacytomas induced by treatment with pristane or in mice bearing an IL-6 transgene (Kovalchuk et al., 2002). Here, MYC is activated by translocations involving Ig loci. Some cases of spontaneous DLBCL have also been shown to harbor translocations affecting the BCL6 proto-oncogene (Qi et al., 2000). The dominant mode of proto-oncogene activation in spontaneous mouse hematopoietic neoplasms is quite distinct, however, involving insertional mutagenesis by MuLV activated from endogenous sequences. This mechanism, first revealed in studies of avian leukosis virus-induced bursal

110


lymphomas, can entail a large variety of genes involved in growth and differentiation, many of which were first identified as common targets of proviral insertion (Rosenberg and Jolicoeur, 1997). The altered expression of a gene controlling growth can give the affected cell a proliferative advantage, leading to the eventual development of a clonal tumor with a provirus integrated at the same site in every cell. The nature of the gene activated by proviral insertion can be determined by cloning the virus and its flanking sequences or by the approach of inverse PCR (Silver and Keerikatte, 1989). These studies have uncovered several modes of proto-oncogene activation: (i) by enhancer sequences within the LTR. This occurs with viruses integrated upstream or downstream from the proto-oncogene regardless of the transcriptional orientation of the virus; (ii) by production of readthrough transcripts; and (iii) by altering noncoding sequences that affect expression of the proto-oncogene; 30 truncations of sequences determining message stability are one example. The identification of loci that are targets of integration in more than one tumor defines a common integration site (CIS) that is likely to be associated with a disease-causing gene. Recent studies have indicated that a CIS can be defined by two integrations occurring within 26 kb of a gene or three integrations within 300 kb (Li et al., 1999; Lund et al., 2002; Suzuki et al., 2002). While it might be assumed that genes targeted by specific translocations associated with a particular human tumor type might be those affected by proviral insertional mutagenesis in histologically related spontaneous mouse B-cell-lymphomas, this possibility is yet to be adequately tested. The reasons for this are several. First, few of the primary tumors examined for proviral insertions have been adequately characterized histologically. In a study of CIS in AKXD and NFS.V þ mice in which the tumors were classified (Suzuki et al., 2002), only eight sites were characterized in follicular B-cell-lymphomas (FBL), for example, and none of these scored for BCL2 or BCL6, the genes most frequently affected by translocations in human follicular lymphoma (FL). It must be recognized, however, that only 10% of human FL cases have the t(14;18) as the sole chromosomal anomaly. One study of human FL reported a mean of six additional breaks involving most frequently chromosomes 1, 2, 4, 5, 13, and 17 and sometimes additions of X, 12, or 18 (Tilly et al., 1994). Notably, four of the FBL integrations localized to regions syntenic with portions of human chromosomes X, 2, and 17 (Suzuki et al., 2002). CIS may thus lead to genes involved in secondary as well as primary disease-causing mutations. Second, there may be strain-specific differences in mechanisms involved in tumor development. Support for this view comes from the finding that some CIS were identified only in AKXD neoplasms, while others were restricted to those of NFS.V þ mice (Suzuki et al., 2002). If more strains were available for


111

testing, some might more closely approximate the genetic defects associated with specific disease categories of humans. Finally, the possibility must be entertained that while CIS identified in mouse lymphoma and leukemia provide powerful tags for cancer gene discovery, the exact pathways engaged in the transformation of B-lineage cells may differ from those active in parallel human diseases. It has been suggested that all cancer phenotypes can be ascribed to six fundamental changes in cell physiology: self-sufficiency in growth signals, insensitivity to anti-growth signals, evasion of apoptosis, limitless replicative potential, sustained angiogenesis and tissue invasion, and metastasis (Hanahan and Weinberg, 2000). Differences in the mechanistic strategies used to obtain these capabilities may be powerfully influenced by species-specific differences. Nonetheless, the power of proviral tagging for identification of cancer-related genes is forcefully demonstrated by the fact that 36 of the 153 CIS characterized by Suzuki et al. (2002) identify known or suspected human cancer genes or their homologs. The possibility that the other CIS might identify genes involved in human leukemia and lymphoma is evidenced by the demonstration that the gene associated with the Evi9 CIS, a zinc-finger protein that interacts with BLC6 (Nakamura et al., 2000), has been identified as the target of a rare t(2;14)(p13;q32) translocation in human NHL now designated BCL11A (Satterwhite et al., 2001). This translocation occurs in a range of human NHL from B cell ALL to multiple myeloma (MM). Even more recent data indicate that Evi9/BCL11A is only the second example of a cell non-autonomous tumor suppressor gene (N. Copeland, personal communication). Just as the dissection of rare but recurrent Ig translocation in aggressive DLBCL continues to define new proto-oncogenes and pathogenic mechanisms, we can expect that the mining of CIS will prove equally rewarding. B. GENE EXPRESSION PROFILING A series of experiments demonstrated the power of gene expression profiling for uncovering previously unrecognized subsets of human B-celllymphomas while providing insights into genetic events involved in pathogenesis (Alizadeh et al., 2000). It is anticipated that these observations will evolve into new therapeutic strategies directed specifically at alterations characteristic of a particular lymphoma class (Davis et al., 2001; Barrans et al., 2002; Shipp et al., 2002). Efforts to utilize this technology for understanding the biology of mouse lymphomas are currently in progress in several laboratories. In the Laboratory of Immunopathology, we have been using spotted oligonucleotide arrays (Operon) comprising approximately 7400 genes to assess the gene expression patterns of primary lymphomas of categories including small B cell, marginal zone, follicular B cell, centroblastic, immunoblastic, Burkitt, and Burkitt-like,

112


together with cell lines derived from pristane-induced plasmacytomas (T. McCarty et al., manuscript submitted for publication). We also examined unmanipulated nude spleen cells as a source of in vivo resting B cells as well as LPS-stimulated nude spleen cells as a source of activated B cells. The data sets were first examined using a hierarchic clustering algorithm to relate the expression patterns of over 90 tumors and cell lines for these 7400 genes (Fig. 2). There were two major branches of the tree. The branch on the right side contains all the low-grade lymphoma types including small B cell, follicular, and marginal zone lymphomas. The only other class represented is immunoblastic lymphomas. In these lymphomas, the proportion of immunoblasts is often quite low, such that the gene readout may be more of normal passenger cells rather than transformed immunoblasts. Another possibility is that the immunoblasts in these cases are the progeny of clonally related centrocytes and centroblast that form the dominant population in the test sample. The small B lymphoma cases are very tightly clustered, indicative of a remarkably close readout by histologic and gene expression criteria for this lymphoma subset. The familial relationships on this tree among lymphomas diagnosed histologically as follicular B cell and splenic marginal zone were less well conserved. Nonetheless, even for these classifications, there were significant number of cases that grouped together, suggesting that studies of additional cases might lead to the resolution of distinct subsets as was demonstrated for human DLBCL (Alizadeh et al., 2000) and multiple myeloma (Zhan et al., 2002). The branch to the left conversely contains all the high-grade tumors and the plasmacytoma cell lines along with a few follicular and immunoblastic tumors. The distinct sub-branch occupied by the Burkitt and plasmacytoma samples could represent a Myc-driven signature, but further studies are required to examine this possibility. Evaluation of primary plasmacytomas will also help exclude the possibility that the plasmacytoma signature is markedly influenced by their adaptation to culture and long passage history. As one step toward understanding the utility of these molecular phenotypes for classification and potential identification of class-specific gene targets for therapy, we determined for each histologic category the genes that best distinguished that subset from all the others. Table II lists, as one example, some of the genes over- or under-expressed by plasmacytomas relative to all other lymphoma types. Stated otherwise, these are the genes that are the best discriminators for diagnosing a lymphoma as PCT. Several genes were previously characterized as over expressed in mouse PCT as compared to normal plasma cells—Pet2, Xlr3a, and Xlr3b (Bergsagel et al., 1994). Others were recently identified as being highly unregulated in human MM—Tra1, Irf4, and Grp58 (Claudio et al., 2002). These can be seen as providing little


GENE TRANSCRIPTS Increased Pik3cg Pet2 Xlr3a Xlr3b Tra1 Xbp1 Irf4 Vegf Sdc1 Grp58 Rps6ka2

113

TABLE II DISTINGUISH PLASMACYTOMA

THAT

Up in MM a

þ þ þ (BLIMP regulated) þ þ (BLIMP regulated)

Decreased

Down in MM a

ICSBP MHC class I CD19 Fcer1g Hcph Lrmp Edr2 Brdg1

(BLIMP repressed) þ þ (BSAP regulated)

(Low in PC)

a From data by Claudio et al. (2002) and Shaffer et al. (2002). þ indicates finding made with myeloma cells or cell lines.

insight into pathogenesis, simply marking a state of differentiation or poorly understood features of transformed plasma cells common to mature, malignant B cells of both species; however, the findings that VEGF and Pik3cg chain are unregulated may provide important clues to pathogenesis that are becoming apparent in the study of human multiple myeloma. For example, VEGF signaling, both autocrine and paracrine—stimulating angiogenesis and release of IL-6 from bone marrow stromal cells—has been shown to contribute to signaling pathways in myeloma cells that regulate cell proliferation and migration (Podar et al., 2001). The observation that heightened expression of Pik3cg is the second most telling discriminator of PCT may also indicate the utility of comparing mouse to human lymphoid neoplasms. PI3K is an important intermediate in signaling by IL-6 and IGF1, two molecules implicated in mouse and human PCT/MM. One recent study showed that PI3K was constitutively activated in purified CD38 þ bone marrow cells from one of four patients examined (Tu et al., 2000). Others have demonstrated that AKT, a downstream target of PI3K, is constitutively active in primary MM cells (Hsu et al., 2001). Treatment of PCT and MM cells lines with inhibitors of PI3K or AKT induces growth retardation or death (Tu et al., 2000; Hsu et al., 2001; Mitsiades et al., 2002). C. MODELING HUMAN LYMPHOMAS In contrast to the successes that have been achieved in modeling different human non-lymphoid leukemia, the record for B-cell-lineage neoplasms

114 HERBERT C. MORSE ET AL.

FIG. 2. Hierarchic clustering of mouse B-lineage lymphomas based on gene expression data from oligonucleotide microarrays. This algorithm was used to group genes based on their similarity of their expression patterns against all the histologically defined categories (not shown). The same clustering method was used to group the cell lines and primary lymphoma samples as determined by similarities in their expression of these genes.


115

is relatively sparse. Mouse models of human pre-B LL/BL associated with BCR/ABL translocations have been developed using p190 and p210 transgene constructs and by reconstituting lethally irradiated recipients with bone marrow cells infected by BCR/ABL-encoding retroviral constructs. These models have been used to understand signaling pathways and for pre-clinical drug testing. Transduction of mouse bone marrow cells by a TEL/AML1containing retrovirus followed by transplantation to an irradiated recipient resulted in one B-lineage lymphoma (Bernardin et al., 2002). Further work is clearly needed to know whether this finding is reproducible and useful for understanding pathogenesis or treatment options. TCL1 is expressed normally in B-lineage cells until downregulation occurs with B cell maturation within the germinal center. Many B-lineage lymphomas ranging from pre-B to mature B and post-GC AIDS-related lymphomas express TCL1 (Teitell et al., 1999; Pekarsky et al., 2001). Recent studies of mice bearing a human TCL1 transgene controlled by an Ig VH promoter and E enhancer were recently described as a model for human CLL/SLL (Bichi et al., 2002). Notably, the mice develop lymphomas before becoming leukemic, similar to what is seen for spontaneous mouse SBL. It will be of interest to compare the gene expression profiles for spontaneous and transgene-associated neoplasms. Also of interest, studies of mice with TCL1 under the control of the E enhancer but the B29 promoter developed a distinct spectrum of more mature B-lineage neoplasms (Hoyer et al., 2002). PAX5 is affected by translocations in human lymphoplasmacytic lymphomas (Iida et al., 1996, 1999). Studies by John Kehrl and collaborators have shown that mice bearing a Pax5 transgene develop a spectrum of mature B-celllymphomas with none truly similar to the human disease (J. Kehrl, H. Morse, unpublished observations). IgH translocations affecting the BCL2 locus are almost invariantly associated with the development of follicular lymphoma. Efforts to model this in mice have been reported to have modest success, with some animals carrying different transgenes exhibiting a low incidence of B-lineage lymphomas of different types (McDonnell and Korsmeyer, 1991; Strasser et al., 1993). Later studies of the same mice in other laboratories have not confirmed these initial observations. The clinical evidence is overwhelming that BCL6 is involved in human DLBCL as a result of translocations into the locus by Ig and non-Ig genes (Kerckaert et al., 1993; Ye et al., 1993; Miki et al., 1994). Multiple efforts from several excellent laboratories to model this effect in mice through different transgene constructs, retrovirus transduction/transplantation, and most recently, a knockin have, to date, been unsuccessful. This may be one of the circumstances where species-specific differences in gene effects make it impossible to induce a human-like disease in mice.

116


IV. Conclusions

From the 1950s to the present, investigators have sought to understand the pathogenesis of hematopoietic neoplasms in mice as models for human disease. The spontaneous diseases of mice form an essential foundation of knowledge about efforts to molecularly model specific diseases, because it will be critically important to distinguish induced from ‘‘background’’ cases. Efforts such as those by Haines et al. (2001) to carefully define the range of tumor types seen in aging mice of different strains is a critical starting point. Gene expression profiling of histologically defined spontaneous neoplasms provides additional richness to these analyses. One of the most striking findings to come from the latter studies is the intimate relationship between morphologic and ‘‘genetic’’ diagnoses. This clearly indicates that the pathologist—in screening the hundreds of ‘‘gene readouts’’ that determine cell size, cytoplasmic volume, nuclear shape, chromatin pattern, and nucleolar number, size, and position—provides a valuable foil to play against the molecular fingerprints portrayed by the data from microarrays. The synergy provided by these approaches in understanding the disease holds tremendous promise for developing better treatments for human hematologic neoplasms.

REFERENCES Adams, J. M., Harris, A. W., Pinkert, C. A., Corcoran, L. M., Alexander, W. S., Cory, S., Palmiter, R. D., and Brinster, R. L. (1985). The c-myc oncogene driven by immunoglobulin enhancers induces lymphoid malignancy in transgenic mice. Nature 318, 533–538. Algara, P., Mateo, M. S., Sanchez-Beato, M., Mollejo, M., Navas, I. C., Romero, L., Sole, F., Salido, M., Florensa, L., Martinez, P., Campo, E., and Piris, M. A. (2002). Analysis of the Igv(H) somatic mutations in splenic marginal zone lymphoma defines a group of unmutated cases with frequent 7q deletion and adverse clinical course. Blood 99, 1299–1304. Alizadeh, A. A., Eisen, M. B., Davis, R. E., Ma, C., Lossos, I. S., Rosenwald, A., Boldrick, J. G., Sabet, H., Tran, T., Yu, X., Powell, J. I., Yang, L. M., Marti, G. E., Moore, T., Hudson, J., Lu, L. S., Lewis, D. B., Tibshirani, R., Sherlock, G., Chan, W. C., Greiner, T. C., Weisenburger, D. D., Armitage, J. O., Warnke, R., Levy, R., Wilson, W., Grever, M. R., Byrd, J. C., Botstein, D., Brown, P. O., and Staudt, L. M. (2000). Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403, 503–511. Barrans, S. L., Carter, I., Owen, R. G., Davies, F. E., Patmore, R. D., Haynes, A. P., Morgan, G. J., and Jack, A. S. (2002). Germinal center phenotype and Bcl-2 expression combined with the International Prognostic Index improves patient risk stratification in diffuse large B-cell lymphoma. Blood 99, 1136–1143. Bergsagel, P. L., Timblin, C. R., Kozak, C. A., and Kuehl, W. M. (1994). Sequence and expression of murine cDNAs encoding Xlr3a and Xlr3b, defining a new X-linked lymphocyte-regulated Xlr gene subfamily. Gene 150, 345–350. Bernardin, F., Yang, Y. D., Cleaves, R., Zahurak, M., Cheng, L. Z., Civin, C. I., and Friedman, A. D. (2002). Tel-Aml1, expressed from T(12;21) in human acute lymphocytic leukemia, induces acute leukemia in mice. Cancer Res. 62, 3904–3908.


117

Bichi, R., Shinton, S. A., Martin, E. S., Koval, A., Calin, G. A., Cesari, R., Russo, G., Hardy, R. R., and Croce, C. M. (2002). Human chronic lymphocytic leukemia modeled in mouse by targeted Tcl1 expression. Proc. Natl. Acad. Sci. USA 99, 6955–6960. Brinster, R. L., Chen, H. Y., Messing, A., Vandyke, T., Levine, A. J., and Palmiter, R. D. (1984). Transgenic mice harboring Sv40 T-antigen genes develop characteristic brain tumors. Cell 37, 367–379. Brues, A. M., and Marble, B. B. (1939). Lymphoblastoma in mice following administration of carcinogenic tar. Am. J. Cancer 37, 45–53. Byrd, L. G., Mcdonald, A. H., Gold, L. G., and Potter, M. (1991). Specific pathogen-free BALB/c mice are refractory to plasmacytoma induction by pristane. J. Immunol. 147, 3632–3637. Clark, E. A., Golub, T. R., Lander, E. S., and Hynes, R. O. (2000). Genomic analysis of metastasis reveals an essential role for Rhoc. Nature 406, 532–535. Claudio, J. O., Masih-Khan, E., Tang, H., Goncalves, J., Voralia, M., Li, Z. H., Nadeem, V., Cukerman, E., Francisco-Pabalan, O., Liew, C. C., Woodgett, J. R., and Stewart, A. K. (2002). A molecular compendium of genes expressed in multiple myeloma. Blood 100, 2175–2186. Cole, R. K., and Furth, J. (1941). Experimental studies on the genetics of spontaneous leukemia in mice. Cancer Res. 1, 957–965. Davidson, W. F., Giese, T., and Fredrickson, T. N. (1998). Spontaneous development of plasmacytoid tumors in mice with defective Fas-Fas ligand interactions. J. Exp. Med. 187, 1825–1838. Davis, P. E., Brown, K. D., Siebenlist, U., and Staudt, L. M. (2001). Constitutive nuclear factor B activity is required for survival of activated B cell-like diffuse large B cell lymphoma cells. J. Exp. Med. 194, 1861–1874. Dewind, N., Dekker, M., Berns, A., Radman, M., and Riele, H. T. (1995). Inactivation of the mouse Msh2 gene results in mismatch repair deficiency, methylation tolerance, hyperrecombination, and predisposition to cancer. Cell 82, 321–330. Donehower, L. A., Harvey, M., Slagle, B. L., Mcarthur, M. J., Montgomery, C. A., Butel, J. S., and Bradley, A. (1992). Mice deficient for P53 are developmentally normal but susceptible to spontaneous tumors. Nature 356, 215–221. Dunn, T. B. (1954). Normal and pathologic anatomy of the reticular tissue in laboratory mice. J. Natl. Cancer Inst. 14, 1281–1433. Fredrickson, T. N., and Harris, A. W. (2000). ‘‘Atlas of Mouse Hematopathology.’’ Amsterdam: Harwood Academic Publishers. Fredrickson, T. N., Hartley, J. W., Morse, H.C. III, Chattopadhyay, S. K., and Lennert, K. (1994). Classification of mouse lymphomas. Curr. Topics Microbiol. Immunol. 194, 109–116. Fredrickson, T. N., Lennert, K., Chattopadhyay, S. K., Morse, H. C., and Hartley, J. W. (1999). Splenic marginal zone lymphomas of mice. Am. J. Pathol. 154, 805–812. Frith, C. H., Ward, J. M., and Chandra, M. (1993). The morphology, immunohistochemistry, and incidence of hematopoietic neoplasms in mice and rats. Toxicol. Pathol. 21, 206–218. Goff, S., Gilboa, E., Witte, O., and Baltimore, D. (1980). Structure of the Abelson murine leukemia virus genome and the homologous cellular gene: studies with cloned viral DNA. Cell 22, 7777–7785. Gross, L. (1951). Spontaneous leukemia developing in C3H mice following inoculation, in infancy, with AK-leukemia extracts, or AK-embryos. Proc. Soc. Exp. Biol. Med. 76, 27–32. Hahn, W. C., and Weinberg, R. A. (2002). Modeling the molecular circuitry of cancer. Nature Rev. Cancer 2, 331–341. Haines, D. C., Chattopadhyay, S., and Ward, J. M. (2001). Pathology of aging B6;129 mice. Toxicol. Pathol. 29, 653–661. Hanahan, D. (1985). Heritable formation of pancreatic -cell tumors in transgenic mice expressing recombinant insulin simian virus-40 oncogenes. Nature 315, 115–122.

118


Hanahan, D., and Weinberg, R. A. (2000). The hallmarks of cancer. Cell 100, 57–70. Hartley, J. W., Chattopadhyay, S. K., Lander, M. R., Taddesse-Heath, L., Naghashfar, Z., Morse, H.C. III, and Fredrickson, T. N. (2000). Accelerated appearance of multiple B cell lymphoma types in NFS/N mice congenic for ecotropic murine leukemia viruses. Lab. Invest. 80, 159–169. Hartley, J. W., Wolford, N. K., Old, L. J., and Rowe, W. P. (1977). A new class of murine leukemia virus associated with development of spontaneous lymphomas. Proc. Natl. Acad. Sci. USA 74, 789–792. Hoyer, K. K., French, S. W., Turner, D. E., Nguyen, M. T. N., Renard, M., Malone, C. S., Knoetig, S., Qi, C.-F., Su, T. T., Cheroutre, H., Wall, R., Rawlings, D. J., Morse, H. C. III, and Teitell, M. A. (2002) Dysregulated TCL1 promotes multiple classes of mature B cell lymphoma. Proc. Natl. Acad. Sci. USA 99, 14392–14397. Hsu, J. H., Shi, Y. J., Krajewski, S., Renner, S., Fisher, M., Reed, J. C., Franke, T. F., and Lichtenstein, A. (2001). The Akt kinase is activated in multiple myeloma tumor cells. Blood 98, 2853–2855. Iida, S., Rao, P. H., Nallasivam, P., Hibshoosh, H., Butler, M., Louie, D. C., Dyomin, V., Ohno, H, Chaganti, R. S., and Dalla-Favera, R. (1996). The t(9;14)(p13;q32) chromosomal translocation associated with lymphoplasmacytoid lymphoma involves the PAX-5 gene. Blood 88, 4110–4117. Iida, S., Rao, P. H., Ueda, R., Chaganti, R. S. K., and Dalla-Favera, R. (1999). Jun chromosomal rearrangement of the Pax-5 locus in lymphoplasmacytic lymphoma with T(9;14)(P13;Q32). Leuk. Lymph. 34, 25–33. Jacks, T., Fazeli, A., Schmitt, E. M., Bronson, R. T., Goodell, M. A., and Weinberg, R. A. (1992). Effects of an Rb mutation in the mouse. Nature 359, 295–300. Jaffe, E. S., Harris N. L., Stein, H., and Vardiman, J. W. (2001). World Health Organisation series on classification of tumors: Pathology and Genetics. ‘‘Tumours of Haematopoietic and Lymphoid Tissues.’’ Monograph. Lyon: IARC Press. Kamijo, T., Weber, J. D., Zambetti, G., Zindy, F., Roussel, M. F., and Sherr, C. J. (1998). Functional and physical interactions of the ARF tumor suppressor with p53 and MDM2. Proc. Natl. Acad. Sci. USA 95, 8292–8297. Kerckaert, J.-P., Deweindt, C., Tilly, H., Quief, S., Lecocq, G., and Bastard, C. (1993). LAZ3, a novel zinc-finger encoding gene, is disrupted by recurring chromosome 3q27 translocations in human lymphomas. Nature Genet. 5, 66–70. Kerner, J. D., Appleby, M. W., Mohr, R. N., Chien, S., Rawlings, D. J., Maliszewski, C. R., Witte, O. N., and Perlmutter, R. M. (1995). Impaired expansion of mouse B-cell progenitors lacking Btk. Immunity 3, 301–312. Khan, W. N., Alt, F. W., Gerstein, R. M., Malynn, B. A., Larsson, I., Rathbun, G., Davidson, L., Muller, S., Kantor, A. B., Herzenberg, L. A., Rosen, F. S., and Sideras, P. (1995). Defective Bcell development and function in Btk-deficient mice. Immunity 3, 283–299. Klein, U., Tu, Y. H., Stolovitzky, G. A., Mattioli, M., Cattoretti, G., Husson, H., Freedman, A., Inghirami, G., Cro, L., Baldini, L., Neri, A. N., Califano, A., and Dalla-Favera, R. (2001). Gene expression profiling of B cell chronic lymphocytic leukemia reveals a homogeneous phenotype related to memory B cells. J. Exp. Med. 194, 1625–1638. Kogan, S. C., Ward, J. M., Anver, M. R., Berman, J. J., Brayton, C., Cardiff, R. D., Carter, J. S., De Coronado, S., Downing, J. R., Fredrickson, T. N., Haines, D. C., Harris, A. W., Harris, N. L., Hiai, H., Jaffe, E. S., Maclennan, I. C. M., Pandolfi, P. P., Pattengale, P. K., Perkins, A. S., Simpson, R. M., Tuttle, M. S., Wong, J. F., and Morse, H. C. (2002). Bethesda proposals for classification of nonlymphoid hematopoietic neoplasms in mice. Blood 100, 238–245. Korteweg, R. (1936). On the manner to which the disposition to carcinoma of the mammary gland is inherited in mice. Genetica 18, 350–371.


119

Kovalchuk, A. L., Kim, J. S., Park, S. S., Coleman, A. E., Ward, J. M., Morse, H. C., Kishimoto, T., Potter, M., and Janz, S. (2002). IL-6 transgenic mouse model for extraosseous plasmacytoma. Proc. Natl. Acad. Sci. USA 99, 1509–1514. Kovalchuk, A. L., Qi, C. F., Torrey, T. A., Taddesse-Heath, L., Feigenbaum, L., Park, S. S., Gerbitz, A., Klobeck, G., Hoertnagel, K., Polack, A., Bornkamm, G. W., Janz, S., and Morse, H. C. (2000). Burkitt lymphoma in the mouse. J. Exp. Med. 192, 1183–1190. Krebs, C., Rask-Nielsen, H. C., and Wagner, A. C. (1930). The origins of lymphosarcomatosis and its relation to other forms of leukosis in white mice. Acta Radiol. 10, 1–53. Lacassagne, A. (1937). Sarcomes lymphoides apparus chez des souris longuement traitees pas des hormones oestrogenes. Compt. Rend. Sco. Biol. 126, 193. Law, L. W. (1951). Observations on properties of leukemic cells resistant to folic acid antagonists. J. Natl. Cancer Inst. 11, 849–865. Li, J., Shen, H., Himmel, K. L., Dupuy, A. J., Largespada, D. A., Nakamura, T., Shaughnessy, J. D., Jr., Jenkins, N. A., and Copeland, N. G. (1999). Leukemia disease genes: large-scale cloning and pathway predictions. Nature Genet. 23, 348–353. Lin, A. W., and Lowe, S. W. (2001). Oncogenic Ras activates the Arf-P53 pathway to suppress epithelial cell transformation. Proc. Natl. Acad. Sci. USA 98, 5025–5030. Lu, L.-M., and Hiai, H. (1999). Mixed phenotype lymphomas in thymectomized (SL/Kh x AKR/ Ms)F1 mice. Jpn. J. Cancer Res. 90, 1218–1223. Lund, A. H., Turner, G., Trubetskoy, A., Verhoeven, E., Wientjens, E., Hulsman, D., Russell, R., Depinho, R. A., Lenz, J., and Van Lohuizen, M. (2002). Genome-wide retroviral insertional tagging of genes involved in cancer in Cdkn2a-deficient mice. Nature Genet. 32, 160–165. MacDowell, E. C., and Richter, M. N. (1935). Mouse leukemia. IX. The role of heredity in spontaneous cases. Arch. Pathol. 20, 709–724. McDonnell, T. J., and Korsmeyer, S. J. (1991). Progression from lymphoid hyperplasia to highgrade malignant lymphoma in mice transgenic for the T(14, 18). Nature 349, 254–256. Metz, T., Harris, A. W., and Adams, J. M. (1995). Absence of P53 allows direct immortalization of hematopoietic cells by the Myc and Raf oncogenes. Cell 82, 29–36. Miki, T., Kawamata, N., Hirosawa, S., and Aoki, N. (1994). Gene involved in the 3q27 translocation associated with B-cell lymphoma, BCL5, encodes a Kruppel-like zinc-finger protein. Blood 83, 26–32. Mitsiades, C. S., Mitsiades, N., Poulaki, V., Schlossman, R., Akiyama, M., Chauhan, D., Hideshima, T., Treon, S. P., Munshi, N. C., Richardson, P. G., and Anderson, K. C. (2002). Activation of NF-B and upregulation of intracellular anti-apoptotic proteins via the Igf-1/Akt signaling in human multiple myeloma cells: therapeutic implications. Oncogene 21, 5673–5683. Morse, H. C. (1981). The laboratory mouse—a historical perspective. In: ‘‘The Mouse in Biomedical Research,’’ New York: Academic Press, pp. 1–16. Morse, H.C. III, Kearney, J. F., Isaacson, P. G., Carroll, M., Fredrickson, T. N., and Jaffe, E. S. (2001a). Cells of the marginal zone—origins, function and neoplasia. Leuk. Res. 25, 169–178. Morse, H. C., Qi, C. F., Chattopadhyay, S. K., Hori, M., Taddesse-Heath, L., Ozato, K., Hartley, J. W., Taylor, B. A., Ward, J. M., Jenkins, N. A., Copeland, N. G., and Fredrickson, T. N. (2001b). Combined histiologic and molecular features reveal previously unappreciated subsets of lymphoma in AKXD recombinant inbred mice. Leuk. Res. 25, 719–733. Morse, H. C., Anver, M. R., Fredrickson, T. N., Haines, D. C., Harris, A. W., Harris, N. L., Jaffe, E. S., Kogan, S. C., Maclennan, I. C. M., Pattengale, P. K., and Ward, J. M. (2002). Bethesda proposals for classification of lymphoid neoplasms in mice. Blood 100, 246–258. Mucenski, M. L., Taylor, B. A., Jenkins, N. A., and Copeland, N. G. (1986). AKXD recombinant inbred strains: models for studying the molecular genetic basis of murine lymphomas. Mol. Cell. Biol. 6, 4236–4243.

120


Nakamura, T., Yamazaki, Y., Saiki, Y., Moriyama, M., Largaespada, D. A., Jenkins, N. A., and Copeland, N. G. (2000). Evi9 Encodes a novel zinc finger protein that physically interacts with Bcl6, a known human B-cell proto-oncogene product. Mol. Cell. Biol. 20, 3178–3186. Oliver, A. M., Martin, F., and Kearney, J. F. (1997). Mouse Cd38 is down-regulated on germinal center B cells and mature plasma cells. J. Immunol. 158, 1108–1115. Pattengale, P. K., and Taylor, C. R. (1983). Experimental models of lymphoproliferative disease. The mouse as a model for human non-Hodgkin’s lymphomas and related leukemias. Am. J. Pathol. 113, 237–265. Pekarsky, Y., Hallas, C., and Croce, C. M. (2001). The role of Tcl1 in human T-cell leukemia. Oncogene 20, 5638–5643. Peschon, J. J., Morrissey, P. J., Grabstein, K. H., Ramsdell, F. J., Maraskovsky, E., Gliniak, B. C., Park, L. S., Ziegler, S. F., Williams, D. E., Ware, C. B., Meyer, J. D., and Davison, B. L. (1994). Early lymphocyte expansion is severely impaired in interleukin-7 receptor-deficient mice. J. Exp. Med. 180, 1955–1960. Podar, K., Tai, Y. T., Davies, F. E., Lentzsch, S., Sattler, M., Hideshima, T., Lin, B. K., Gupta, D., Shima, Y., Chauhan, D., Mitsiades, C., Raje, N., Richardson, P., and Anderson, K. C. (2001). Vascular endothelial growth factor triggers signaling cascades mediating multiple myeloma cell growth and migration. Blood 98, 428–435. Potter, M., Wax, J. S., Anderson, A. O., and Nordan, R. P. (1985). Inhibition of plasmacytoma development in BALB/c mice by indomethacin. J. Exp. Med. 161, 996–1012. Potter, M., and Wiener, F. (1992). Plasmacytomagenesis in mice: model of neoplastic development dependent on chromosomal translocations. Carcinogenesis 13, 1681–1697. Puel, A., Ziegler, S. F., Buckley, R. H., and Leonard, W. J. (1998). Defective IL7r expression in T-B þ NK þ severe combined immunodeficiency. Nature Genet. 20, 394–397. Qi, C. F., Hori, M., Coleman, A. E., Torrey, T. A., Taddesse-Heath, L., Ye, B. H., Chattopadhyay, S. K., Hartley, J. W., and Morse, H. C. (2000). Genomic organisation and expression of Bcl6 in murine B-cell lymphomas. Leuk. Res. 24, 719–732. Reitmair, A. H., Schmits, R., Ewel, A., Bapat, B., Redston, M., Mitri, A., Waterhouse, P., Mittrucker, H. W., Wakeham, A., Liu, B., Thomason, A., Griesser, H., Gallinger, S., Ballhausen, W. G., Fishel, R., and Mak, T. W. (1995). Msh2 deficient mice are viable and susceptible to lymphoid tumors. Nature Genet. 11, 64–70. Ricciardone, M. D., Ozcelik, T., Cevher, B., Ozdag, H., Tuncer, M., Gurgey, A., Uzunalimoglu, O., Cetinkaya, H., Tanyeli, A., Erken, E., and Ozturk, M. (1999). Human Mlh1 deficiency predisposes to hematological malignancy and neurofibromatosis type 1. Cancer Res. 59, 290–293. Richter, M. N., and MacDowell, E. C. (1935). Experiments with mammalian leukemia. Physiol. Rev. 15, 509–524. Rosenberg, N., and Jolicoeur, P. (1997). Retroviral pathogenesis. In: ‘‘Retroviruses,’’ Plaiview, NY: Cold Spring Harbor Laboratory Press, pp. 475–486. Rosenwald, A., Alizadeh, A. A., Widhopf, G., Simon, R., Davis, R. E., Yu, X., Yang, L. M., Pickeral, O. K., Rassenti, L. Z., Powell, J., Botstein, D., Byrd, J. C., Grever, M. R., Cheson, B. D., Chiorazzi, N., Wilson, W. H., Kipps, T. J., Brown, P. O., and Staudt, L. M. (2001). Relation of gene expression phenotype to immunoglobulin mutation genotype in B cell chronic lymphocytic leukemia. J. Exp. Med. 194, 1639–1647. Satterwhite, E., Sonoki, T., Willis, T. G., Harder, L., Nowak, R., Arriola, E. L., Liu, H., Price, H. P., Gesk, S., Steinemann, D., Schlegelberger, B., Oscier, D. G., Siebert, R., Tucker, P. W., and Dyer, M. J. S. (2001). The Bcl11 gene family: involvement of Bcl11a in lymphoid malignancies. Blood 98, 3413–3420. Shaffer, A. L., Lin, K. I., Kuo, T. C., Yu, X., Hurt, E. M., Rosenwald, A., Giltnane, J. M., Yang, L. M., Zhao, H., Calame, K., and Staudt, L. M. (2002). Blimp-1 orchestrates plasma cell differentiation by extinguishing the mature B cell gene expression program. Immunity 17, 51–62.


121

Shinnick, T. M., Lerner, R. A., and Sutcliffe, J. G. (1981). Nucleotide sequence of Moloney murine leukemia virus. Nature 293, 543–548. Shipp, M. A., Ross, K. N., Tamayo, P., Weng, A. P., Kutok, J. L., Aguiar, R. C. T., Gaasenbeek, M., Angelo, M., Reich, M., Pinkus, G. S., Ray, T. S., Koval, M. A., Last, K. W., Norton, A., Lister, T. A., Mesirov, J., Neuberg, D. S., Lander, E. S., Aster, J. C., and Golub, T. R. (2002). Diffuse large B-cell lymphoma outcome prediction by gene expression profiling and supervised machine learning. Nature Med. 8, 68–74. Silver, J., and Keerikatte, V. (1989). Novel use of polymerase chain reaction to amplify cellular DNA adjacent to an integrated provirus. J. Virol. 63, 1924–1928. Staff, Roscoe B. Jackson Memorial Laboratory. (1933). The existence of nonchromosomal influence in the incidence of mammary tumors in mice. Science 78, 465–466. Steiniger, B., Barth, P., and Hellinger, A. (2001). The perifollicular and marginal zones of the human splenic white pulp—do fibroblasts guide lymphocyte immigration? Am. J. Pathol. 159, 501–512. Stewart, T. A., Pattengale, P. K., and Leder, P. (1984). Spontaneous mammary adenocarcinomas in transgenic mice that carry and express Mtv/Myc fusion genes. Cell 38, 627–637. Strasser, A., Harris, A. W., and Cory, S. (1993). E-bcl-2 transgene facilitates spontaneous transformation of early pre-B and immunoglobulin-secreting cells but not T cells. Oncogene 8, 1–9. Suzuki, T., Shen, H. F., Akagi, K., Morse, H. C., Malley, J. D., Naiman, D. Q., Jenkins, N. A., and Copeland, N. G. (2002). New genes involved in cancer identified by retroviral tagging. Nature Genet. 32, 166–174. Symons, R. C. A., Daly, M. J., Fridlyand, J., Speed, T. P., Cook, W. D., Gerondakis, S., Harris, A. W., and Foote, S. J. (2002). Multiple genetic loci modify susceptibility to plasmacytoma-related morbidity in E-V-Abl transgenic mice. Proc. Natl. Acad. Sci. USA 99, 11299–11304. Taddesse-Heath, L., and Morse, H.C. III. (2000). Lymphoma in genetically engineered mice. In: ‘‘Pathology of Genetically Engineered Mice,’’ Ames, IA: Iowa State University Press, pp. 365–381. Teitell, M., Damore, M. A., Sulur, G. G., Stern, M. H., Said, J. W., Denny, C. T., and Wall, R. (1999). Tcl1 oncogene expression in AIDS-related lymphomas and lymphoid tissues. Proc. Natl. Acad. Sci. USA 96, 9809–9814. Tilly, H., Rossi, A., Stamatoullas, A., Lenormand, B., Bigorgne, C., Kunlin, A., Monconduit, M., and Bastard, C. (1994). Prognostic value of chromosomal abnormalities in follicular lymphoma. Blood 84, 1043–1049. Tu, Y. P., Gardner, A., and Lichtenstein, A. (2000). The phosphatidylinositol 3-kinase/Akt kinase pathway in multiple myeloma plasma cells: roles in cytokine-dependent survival and proliferative responses. Cancer Res. 60, 6763–6770. Yamada, Y., Matsushiro, H., Ogawa, M. S., Okamoto, K., Nakakuki, Y., Toyokuni, S., Fukumoto, M., and Hiai, H. (1994). Genetic predisposition to pre-B lymphomas in Sl Kh strain mice. Cancer Res. 54, 403–407. Ye, B. H., Rao, P. H., Chaganti, R. S. K., and Dalla-Favera, R. (1993). Cloning of bcl-6, the locus involved in chromosome translocations affecting band 3q27 in B-cell lymphoma. Cancer Res. 53, 2732–2735. Zhan, F. H., Hardin, J., Kordsmeier, B., Bumm, K., Zheng, M. Z., Tian, E. M., Sanderson, R., Yang, Y., Wilson, C., Zangari, M., Anaissie, E., Morris, C., Muwalla, F., Van Rhee, F., Fassas, A., Crowley, J., Tricot, G., Barlogie, B., and Shaughnessy, J. (2002). Global gene expression profiling of multiple myeloma, monoclonal gammopathy of undetermined significance, and normal bone marrow plasma cells. Blood 99, 1745–1757. Zhang, S. L., Dubois, W., Ramsay, E. S., Bliskovski, V., Morse, H. C., Taddesse-Heath, L., Vass, W. C., Depinho, R. A., and Mock, B. A. (2001). Efficiency alleles of the Pctr1 modifier locus for plasmacytoma susceptibility. Mol. Cell. Biol. 21, 310–318.



Prions and the Immune System: A Journey Through Gut, Spleen, and Nerves ADRIANO AGUZZI1 Institute of Neuropathology, Universita¨tsspital Zu¨rich, Schmelzbergstrasse 12, CH-8091 Zu¨rich, Switzerland

For more than two decades it has been contended that prion infection does not elicit immune responses: transmissible spongiform encephalopathies do not go along with conspicuous inflammatory infiltrates, and antibodies to the prion protein are typically undetectable. Why is it, then, that prions accumulate in lymphoid organs, and that various states of immune deficiency prevent peripheral prion infection? This review revisits the current evidence of the involvement of the immune system in prion diseases, while attempting to trace the elaborate mechanisms by which peripherally administered prions invade the brain and ultimately cause damage. The investigation of these questions leads to unexpected detours, including the neurophysiology of lymphoid organs, and even the function of a prion protein homolog in male fertility.

I. Prion Biology: Some Basic Facts

Prion diseases are inevitably fatal neurodegenerative conditions which affect humans and a wide variety of animals (Aguzzi et al., 2001c). Although prion diseases may be present with certain morphological and pathophysiological parallels to other progressive encephalopathies, such as Alzheimer’s and Parkinson’s disease (Aguzzi and Raeber, 1998), they are unique in that they are transmissible. Homogenization of brain tissue from affected individuals and intracerebral inoculation into another individual of the same species will typically reproduce the disease. This important fact was recognized more than half a century ago in the case of scrapie (Cuille and Chelle, 1939), a prototypic prion disease that affects sheep and goats. In the 1960s and 1970s, the work of Gajdusek on Kuru and CreutzfeldtJakob disease (Gajdusek et al., 1966; Gibbs et al., 1968), partly inspired by a suggestion of Hadlow (Hadlow, 1959) on the similarity of scrapie and Kuru, established that these diseases are also transmissible to primates, to mice, and in some unfortunate iatrogenic instances, also to other humans (Brown et al., 2000). Therefore, prion diseases are also called transmissible spongiform encephalopathies (TSEs), a term that emphasizes their infectious character. Tel.: þ 41 1 255 2107; fax: þ 41 1 255 4402; E-mail: [email protected]

1


124

A. AGUZZI

While only less than 1% of all reported cases of Creutzfeldt-Jakob disease (CJD) can be traced to a defined infectious source, the identification of bovine spongiform encephalopathy (BSE) (Wells et al., 1987) and its subsequent epizootic spread has highlighted prion-contaminated meat-and-bone meal as a tremendously efficient vector for bovine prion diseases (Weissmann and Aguzzi, 1997), which does not completely lose its infectious potential even after extensive autoclaving (Taylor, 2000). BSE is most likely transmissible to humans, too, and strong circumstantial evidence (Aguzzi, 1996; Collinge et al., 1996; Aguzzi and Weissmann, 1996b; Bruce et al., 1997; Hill et al., 1997a) suggests that BSE is the cause of variant Creutzfeldt-Jakob disease (vCJD) which has claimed more than 130 lives in the United Kingdom (Will et al., 1996; http://www.doh.gov.uk/cjd/stats/aug02.htm, 2002), and to a much smaller extent, in some other countries (Chazot et al., 1996). When transmitted to primates, BSE produces a pathology strikingly similar to that of vCJD (Aguzzi and Weissmann, 1996b; Lasmezas et al., 1996b). While there is no direct indication yet that vCJD has been transmitted from one human to another, this is certainly a very worrying scenario (Aguzzi, 2000). Not only surgical instruments may represent potential vectors of transmission, but possibly also transfusions. Significantly, sheep infected with BSE can efficiently transmit the agent to other sheep via blood transfusion (Houston et al., 2000; Hunter et al., 2002). Several aspects of CJD epidemiology continue to be enigmatic. For example, CJD incidence in Switzerland has risen two-fold in 2001, and appears to be increasing even further in the year 2002 (Glatzel et al., 2002). A screen for recognized or hypothetical risk factors for CJD has, to date, not exposed any causal factors. Several scenarios may account for the increase in incidence, including improved reporting, iatrogenic transmission, and transmission of a prion zoonosis. Prion diseases typically exhibit a very long latency period between the time of infection and the clinical manifestation: this is the reason why these diseases were originally thought to be caused by ‘‘slow viruses.’’ From the viewpoint of interventional approaches, this peculiarity may be exploitable, since it opens a possible window of intervention after infection has occurred, but before brain damage is being initiated. Prions spend much of this latency time executing neuroinvasion, which is the process of reaching the central nervous system (CNS) after entering the body from peripheral sites (Aguzzi, 1997a; Nicotera, 2001). During this process, little or no damage occurs to the brain, and one might hope that its interruption may prevent neurodegeneration. A. STANLEY PRUSINER’S PROTEIN-ONLY HYPOTHESIS The most widely accepted hypothesis on the nature of the infectious agent causing TSEs (which was termed prion by Stanley B. Prusiner) (Prusiner,

PRIONS AND THE IMMUNE SYSTEM

125

1982) predicates that it consists essentially of PrPSc, an abnormally folded, protease-resistant, beta-sheet rich isoform of a normal cellular protein termed PrPC. According to this fascinating theory, the prion does not contain any informational nucleic acids, and its infectivity propagates simply by recruitment and ‘‘autocatalytic’’ conformational conversion of cellular prion protein into disease-associated PrPSc (Aguzzi and Weissmann, 1997). A large body of experimental and epidemiological evidence is compatible with the protein-only hypothesis, and very stringently designed experiments have failed to disprove it. It would go well beyond the scope of this chapter to review all efforts that have been undertaken to this effect. Perhaps most impressively, knockout mice carrying a homozygous deletion of the Prnp gene that encodes PrPC, fail to develop disease upon inoculation with infectious brain homogenate (Buëler et al., 1993), nor does their brain carry prion infectivity (Sailer et al., 1994). Reintroduction of Prnp by transgenesis—even in a shortened, redacted form—restores infectibility and prion replication in Prnpo/o mice (Fischer et al., 1996; Shmerling et al., 1998; Supattapone et al., 1999; Flechsig et al., 2000). In addition, all familial cases of human TSEs are characterized by Prnp mutations (Aguzzi and Weissmann, 1996a; Prusiner et al., 1998). Informational nucleic acids of >50 nucleotides in length do not participate in prion infectivity (Kellings et al., 1993; Riesner et al., 1993), but shorter non-coding oligonucleotides have not been formally excluded—a fact that may have some relevance in view of the surprising discoveries related to RNA-mediated gene silencing. Prnp exhibits a long reading frame on its noncoding strand (Moser et al., 1993) which is conserved among mammals (Aguzzi, 1997b; Rother et al., 1997), suggesting that it may be transcribed— maybe only in specific pathological situations. One might wonder, therefore, whether this peculiar property of Prnp might result in the production of double-stranded transcripts with silencing properties. However, bona fide antisense transcription of the Prnp locus has never been demonstrated in vivo, nor in cultured cells.

B. SOME MAJOR OPEN QUESTIONS

IN

PRION BIOLOGY

In recent years, prion research has progressed at a faster pace than many of us would have thought possible. As a consequence, many enigmas surrounding prion diseases have now been solved. However, the areas that are still obscure do not relate only to the details: some of these concern the core of the prion concept (Chesebro, 1998). In my opinion, there are four large groups of questions regarding the basic science of prion replication and of

126

A. AGUZZI

development of TSE diseases that deserve to be addressed with a vigorous research effort:

Which are the molecular mechanisms of prion replication? How does the disease-associated prion protein, PrPSc, achieve the conversion of its cellular sibling, PrPC, into a likeness of itself ? Which other proteins assist this process? Can we inhibit this process? If so, how? What is the essence of prion strains, which are operationally defined as variants of the infectious agent capable of retaining stable phenotypic traits upon serial passage in syngeneic hosts? The existence of strains is very well known in virology, but it was not predicted to exist in the case of an agent that propagates epigenetically. How do prions reach the brain after having entered the body? Which molecules and which cell types are involved in this process of neuroinvasion? Which inhibitory strategies are likely to succeed? The mechanisms of neurodegeneration in spongiform encephalopathies are not understood. Which are the pathogenetic cascades that are activated upon accumulation of disease-associated prion protein, and ultimately lead to brain damage? What is the physiological function of the highly conserved, normal prion protein, PrPC? The Prnp gene encoding PrPC was identified in 1985 (Oesch et al., 1985; Basler et al., 1986), Prnp knockout mice were described in 1992 (Buëler et al., 1992), and some PrPC-interacting proteins have been identified (Oesch et al., 1990; Rieger et al., 1997; Yehiely et al., 2002; Zanata et al., 2002). Yet the function of PrPC remains unknown.

This chapter will discuss some of the progress recently achieved in some of the areas delineated above, with special reference to the topics that have directly interested my laboratory. C. BRAIN DAMAGE

IN

PRION DISEASES

An interesting question regards the molecular mechanism underlying neuropathological changes, in particular cell death, resulting from prion disease. Depletion of PrPC is an unlikely cause; in view of the finding that abrogation of PrP does not cause scrapie-like neuropathological changes (Buëler et al., 1992), even when elicited postnatally (Mallucci et al., 2002). More likely, toxicity of PrPSc or some PrPC-dependent process is responsible. To address the question of neurotoxicity, brain tissue of Prnpo/o mice was exposed to a continuous source of PrPSc. For this purpose, telencephalic tissue from transgenic mice overexpressing PrP (Fischer et al., 1996) was transplanted into the forebrain of Prnpo/o mice and the ‘‘pseudochimeric’’ brains were inoculated with scrapie prions. All grafted and scrapie-inoculated


127

mice remained free of scrapie symptoms for at least 70 weeks; this exceeded at least seven-fold the survival time of scrapie-infected donor mice (Brandner et al., 1996a). Therefore, the presence of a continuous source of PrPSc and of scrapie prions does not exert any clinically detectable adverse effects on a mouse devoid of PrPC. On the other hand, the grafts developed characteristic histopathological features of scrapie after inoculation. The course of the disease in the graft was very similar to that observed in the brain of scrapie-inoculated wild-type mice (Brandner et al., 1998). Importantly, grafts had extensive contact with the recipient brain, and prions could navigate between the two compartments, as shown by the fact that inoculation of wild-type animals engrafted with PrP-expressing neuroectodermal tissue resulted in scrapie pathology in both graft and host tissue. Nonetheless, histopathological changes never extended into host tissue, even at the latest stages (>450 days), although PrPSc was detected in both grafts and recipient brain, and immunohistochemistry revealed PrP deposits in the hippocampus, and occasionally in the parietal cortex, of all animals (Brandner et al., 1996a). Thus, prions moved from the grafts to some regions of the PrPdeficient host brain without causing pathological changes or clinical disease. The distribution of PrPSc in the white matter tracts of the host brain suggests diffusion within the extra-cellular space (Jeffrey et al., 1994) rather than axonal transport. These findings suggest that the expression of PrPC by an infected cell, rather than the extra-cellular deposition of PrPSc, is the critical prerequisite for the development of scrapie pathology. Perhaps PrPSc is inherently nontoxic and PrPSc plaques found in spongiform encephalopathies are an epiphenomenon rather than a cause of neuronal damage. This hypothesis appears to be supported by the recent data (Ma and Lindquist, 2002; Ma et al., 2002) indicating that exaggerated retrograde transport of the prion protein from the endoplasmic reticulum into the cytosol, or functionally equivalent inhibition of proteasome function, might induce a self-propagating, extremely cytotoxic cellular form, which may ultimately be responsible for neuronal damage. These results are very exciting, and it will be important to validate them by investigating whether the aggregated self-propagating material can also be transmitted between individual animals in a classic transmission experiment. One may therefore propose that availability of PrPC for some intracellular process elicited by the infectious agent, perhaps the formation of a toxic form of PrP (PrP* ; Weissmann, 1991) other than PrPSc is responsible for spongiosis, gliosis, and neuronal death. This would be in agreement with the fact that in several instances, and especially in fatal familial insomnia, spongiform pathology is detectable although very little PrPSc is present (Aguzzi and Weissmann, 1997).

128

A. AGUZZI

II. Peripheral Entry Sites of Prions: Complicity of Immune Cells

The fastest and most efficient method for inducing spongiform encephalopathy in the laboratory is intracerebral inoculation of brain homogenate. Inoculation of 1,000,000 infectious units (defined as the amount of infectivity that will induce TSE with 50% likelihood in a given host) will yield disease in approximately half year; a remarkably strict inverse relationship can be observed between the logarithm of the inoculated dose and the incubation time (Prusiner et al., 1982) (Fig. 1). However, the above situation does not correspond to typically what happens in the field. There, acquisition of prion infectivity through any of several peripheral routes is the rule. However, prion diseases can also be initiated by feeding (Wells et al., 1987; Kimberlin and Wilesmith, 1994; Anderson et al., 1996), by intravenous and intraperitoneal injection (Kimberlin and Walker, 1978) as well as from the eye by conjunctival instillation (Scott et al., 1993), corneal grafts (Duffy et al., 1974) and intraocular injection (Fraser, 1982).

FIG. 1. Prion bioassay by the incubation time method. This figure is reproduced from an important study by Prusiner and colleagues that demonstrates an inverse logarithmic relationship between size of infectious inoculum and latency period of scrapie in experimental animals (Prusiner et al., 1982). The relationship is so precise that incubation time can actually be used to back-calculate the prion titer of a test article, e.g., for inactivation studies. Measurement of prion titers by the incubation time method has become a standard procedure which can be used for both wild-type and transgenic animals.


129

Two routes of infection have suggested for a long time that immune cells might be of importance for this phase of prion pathogenesis: oral challenge and administration by scarification. A. THE PATHWAY

OF

ORALLY ADMINISTERED PRIONS

Upon oral challenge, an early rise in prion infectivity can be observed in the distal ileum of infected organisms: this applies to several species but was most extensively investigated in the sheep (Wells et al., 1994; Vankeulen et al., 1996). There, Peyer’s patches acquire strong immunopositivity for the prion protein. Immunohistochemical stains with antibodies to the prion protein typically reveal a robust signal in primary B cell follicles and germinal centers, which roughly colocalizes with the complement receptor, CD35, in a wide variety of secondary lymphoid organs including appendix and tonsils (Hill et al., 1997b). Although conventional light microscopy does not allow differentiating between PrPC and PrPSc, Western blot analysis has not left any doubt about the fact that Peyer’s patches do accumulate the disease-associated form of the prion protein. The latter is also true in the mouse model of scrapie, which is being used as a convenient experimental paradigm by many laboratories including ours. Administration of mouse-adapted scrapie prions (Rocky Mountain Laboratory or RML strain, originally derived from the Chandler sheep scrapie isolate) induces a surge in intestinal prion infectivity as early as a few days after inoculation (Marco Prinz, Gerhard Huber and Adriano Aguzzi, unpublished results). All of the above evidence conjures the suggestion that Peyer’s patches may represent a portal of entry for orally administered prions on their journey from the luminal aspect of the gastroenteric tube to the CNS. However, the question as to whether the same applies for BSE-affected cattle has been answered less unambiguously. In a monumental study of BSE pathogenesis in cattle carried out at the UK Veterinary Laboratory Agency, cows of various ages were fed with 100 g, 10 g, 1 g, or 100 mg of brain homogenate derived from BSE-sick cows (Bradley, 2000). A large variety of tissues was taken at various points in time, homogenized, and transmitted intracerebrally to indicator organisms in order to assess their prion content. This study was designed to be performed over a time frame of more than a decade and is still underway at the time of writing: it has uncovered a transient surge in infectivity in the distal ileum of cows at approximately 6 months post infection. Infectivity then subsides, but it appears to return to the terminal ileum at the end stages of disease, maybe by means of some sort of retrograde transport (Wells et al., 1998). Although this was not formally confirmed, it appears likely that Peyer’s patches are the sites of prion accumulation in the gastrointestinal tract of cattle challenged orally with prions.

130

A. AGUZZI

B. ORAL PRION SUSCEPTIBILITY CORRELATES STRUCTURE OF PEYER’S PATCHES

WITH

NUMBER

BUT NOT

Recruitment of activated B lymphocytes to Peyer’s patches requires 4 7 integrin as an essential homing receptor: Peyer’s patches of mice that lack 7 integrin are normal in number, but are atrophic and almost entirely devoid of B cells (Wagner et al., 1996). Therefore, it seemed interesting to investigate the susceptibility to orally administered prions of 7/ mice. Surprisingly, we found that minimal infectious dose and disease incubation after oral exposure to logarithmic dilutions of prion inoculum were similar in 7/ and wild-type mice (Prinz et al., 2003). Despite their atrophy, Peyer’s patches of both 7/ and wild-type mice contained 3–4 logLD50/g prion infectivity at 125 or more days after challenge. Why does reduced mucosal lymphocyte trafficking not impair, as expected, the susceptibility to orally initiated prion disease? One possible reason may relate to the fact that, despite marked reduction of B cells, M cells were still present in 7/ mice. In contrast, mice deficient in both tumor necrosis factor and lymphotoxin- (TNF/ LT/) or in lymphocytes (RAG-1/, MT), in which Peyer’s patches are reduced in number, were highly resistant to oral challenge, and their intestines were virtually devoid of prion infectivity at all times after challenge. Therefore, lymphoreticular requirements for enteric and for intraperitoneal uptake of prions differ from each other and that susceptibility to prion infection following oral challenge correlates with the number of Peyer’s patches, but is independent of the number of intestinal mucosa-associated lymphocytes (Prinz et al., 2003). C. DIAGNOSIS OF PRION DISEASE: IDENTIFICATION INFECTIOUS AGENT

OF THE

As explained in Section XI, the term ‘‘prion’’ is used here in an operational meaning: it is not meant to be necessarily identical with PrPSc, but it simply denotes the infectious agent, whatever this agent exactly consists of (Aguzzi and Weissmann, 1997). If one takes this viewpoint, PrPSc defined originally as the protease-resistant moiety associated with prion disease would be a mere surrogate marker for the prion. Even if this position may seem all too negativistic in view of the wealth of compelling evidence in favor of the protein-only hypothesis, it is still advisable to maintain that we do not understand any of the molecular details of prion replication. In other words, we still need to be open to the possibility that infectivity might be brought about by a conformer of PrPC that is not identical with what was originally defined as PrPSc (see Section XI). This position appears to gain increasing acceptance, as documented by the recent description of ‘‘protease sensitive PrPSc’’ by the group of Stanley Prusiner (Safar et al., 1998, 2000).

131


Because protease resistance and infectivity of PrPSc may not necessarily be congruent, it is desirable to expand the range of reagents that will interact specifically with disease-associated prion protein, but not (or to a lesser degree) with its counterpart, PrPC. One IgM monoclonal antibody termed 15B3 has been described to fulfill the above conditions (Korth et al., 1997), but this antibody has not become generally available, and no follow-up studies have been published. Disappointingly, expression of the complementaritydetermining regions of 15B3 in the context of a phage, as well as expression of its heavy chain variable region in transgenic mice, failed to reproduce the specificity of PrPSc binding (Heppner et al., 2001b). We have described the puzzling phenomenon that a prevalent constituent of blood serum, plasminogen, captures efficiently PrPSc but not PrPC when immobilized onto the surface of magnetic beads (Fischer et al., 2000; Maissen et al., 2001). The significance of this phenomenon in vivo has yet to be addressed, and it has been claimed on the basis of experiments that binding occurs only in the presence of specific detergents (Shaked et al., 2002). While the latter study was performed only in vitro, and related only to the binding of a mixture of serum proteins rather than purified plasminogen, this does not detract from the usefulness of the binding phenomenon for diagnostic purposes: plasminogen binds to PrPSc from a variety of species and genotypes (Maissen et al., 2001), and the captured PrPSc retains the glycotype profile characteristic of the prion strain from which it was isolated (M. Maissen and Adriano Aguzzi, unpublished data). D. TRANSEPITHELIAL ENTERIC PASSAGE

OF

PRIONS: A ROLE

FOR

M CELLS?

We have set out to investigate some of the preconditions of transepithelial passage of prions. Membranous epithelial cells (M cells) are key sites of antigen sampling for the mucosal-associated lymphoid system (MALT) and have been recognized as major ports of entry for enteric pathogens in the gut via transepithelial transport (Neutra et al., 1996). Interestingly, maturation of M cells is dependent on signals transmitted by intraepithelial B cells. The group of Jean-Pierre Kraehenbuhl (Lausanne) has developed efficient in vitro systems, in which epithelial cells can be instructed to undergo differentiation to cells that resemble M cells by morphological and functional-physiological criteria. Therefore, we investigated whether M cells are a plausible site of prion entry in a co-culture model (Kerneis et al., 1997). Colon carcinoma cells (line Caco-2) were seeded on the upper face of transwell filters and cultured until confluency was reached. Next, B-lymphoblastoid cells Raji B cells were added onto the lower side of the filters. Lymphoid cells migrated through the pores of the filter and settled within the epithelial monolayer (Fig. 2), inducing differentiation of some Caco-2 cells into M cells. Successful conversion was monitored by measuring transport of fluorescein-conjugated latex beads in

132

A. AGUZZI

FIG. 2. Transepithelial transport of prions via M cells. Filter membranes with 3 m pores were overlaid with epithelial cells (colon carcinoma, line Caco-2). Human B-lymphoblastoid Raji cells were then cultured on the back side of the membrane. (a) morphology of co-cultures on filters, visualized by hematoxylin/eosin stains (HE). (b) Immunohistochemical stain for cytokeratins (pan-CK) visualizing human Caco-2 epithelial cells and (c) for the B cell marker CD20, displaying B cells that have successfully migrated through the pores of the filter (arrow). (d) Flow cytometric analysis demonstrating active (i.e., temperature-dependent) transport of FITCconjugated latex beads from the apical to the basolateral chamber compartment upon integration of B cells within pouches of the epithelial Caco-2 cell layer, and differentiation of M cells. Intactness of co-cultures was assessed by measuring transepithelial resistance at 37 C, which was typically > 200 cm2. The presence of additional peaks corresponds to aggregates of several latex beads. (e–f) Recovery of prion infectivity upon challenge with different prion inocula (5 (black) or 3 (grey) logLD50 input infectious units) was only visible in co-cultures containing M cells (triangles, lane 3–4). No prion transport was observed in Caco-2 cultures (circles, lane 1–2) without M cells, except in one case in which traces of infectivity were present. Controls: mock inoculum (lane 1), Caco-2 culture (Ca) after slight mechanical manipulation resulting in a TER < 50 cm2 (lane 2). Prion infectivity dilutions as indicated before (positive control, lane 3–4) and after incubating with either Caco-2 (lane 5) or M cell containing (M) cultures (lane 6). Each symbol represents the mean incubation time in days (ordinate) of 4 tga20 indicator mice until terminal disease.


133

cocultures that exhibited a high transepithelial resistance and were therefore tight. Active transepithelial transport of beads, but not passive leakage, was blocked at 4 C. Scrapie prions were administered to the apical compartment of co-cultures that combined integrity and active transport of beads. After 24 h, infectivity was determined within the basolateral compartment by bioassay with tga20 mice, which overexpress a Prnp transgene and develop scrapie rapidly after infection (Fischer et al., 1996). Upon challenge with 5 logLD50 scrapie prions, we consistently recovered prions in the basolateral compartment of cocultures containing M-cells, suggesting transepithelial prion transport (Fig. 2). Even at low prion doses (3 logLD50), we found infectivity in at least one M cell-containing co-culture. In contrast, there was hardly any prion transport in Caco-2 cultures without M cells (Heppner et al., 2001a). These findings indicate that M-cell differentiation is necessary and sufficient for active transepithelial prion transport in vitro. M cell-dependent uptake of foreign antigens or particles is known to be followed by rapid transcytosis directly to the intraepithelial pocket, where key players of the immune system, e.g., macrophages, dendritic cells and lymphocytes (Neutra et al., 1996), are located. The remainder of this review article will discuss the fact that some of these immune cells are crucially involved in the process of neuroinvasion (Aguzzi and Heppner, 2000). Therefore, prions may exploit M cell-dependent transcytosis to gain access to the immune system. While these findings suggest that M cells are a plausible candidates for the mucosal portal of prion infection, it still remains to be established whether the pathway delineated above does indeed represent the first portal of entry of orally administered prions into the body. This will necessitate in vivo experimentation, for example by ablation of M cells through suicide transgenetic strategies, or by M-cell specific expression of Prnp transgenes. E. UPTAKE

OF

PRION THROUGH

THE

SKIN

Even less well-understood, yet possibly much more efficient than oral administration of prions is challenged by scarification. Removal of the most superficial layers of the skin, and subsequent administration of prions, has been known for a long time to be a highly efficacious method of inducing prion disease (Carp, 1982; Taylor et al., 1996). It is thinkable that dendritic cells in the skin may become loaded with the infectious agent by this method, and in fact recent work has implicated dendritic cells as potential vectors of prions in oral (Huang et al., 2002) and in hematogenous spread (Aucouturier et al., 2001) of the agent. It is equally possible (and in my view more probable), however, that scarification induces direct neural entry of prions into skin nerve terminals. The latter hypothesis has not yet been studied in much detail, but it would help explain the remarkable speed with which CNS pathogenesis follows inoculation by this route: dermal inoculation of scarified mice yields

134

A. AGUZZI

typical latency periods of the disease that are similar to those obtained by intracerebral inoculation. The possibility of direct neural spread of the agent has been brought up also by a series of elegant experiments in the laboratories of Oldstone and Chesebro: transgenic mice that lack the endogenous prion gene but express a Prnp transgene exclusively in nervous tissue (under transcriptional control of the neuron specific enolase regulatory elements) can be efficiently infected by the oral route despite lack of prion protein expression in lymphoid organs (Race et al., 1995). Of course, these experiments do not exclude that dendritic or other mobile cells may participate to neuroinvasion even if they do not express endogenous PrPC.

III. Lymphocytes and Prion Pathogenesis

It was Neil Cashman, more than a decade ago, who discovered that the normal prion protein is consistently expressed, albeit at moderate levels, in circulating lymphocytes (Cashman et al., 1990). Subsequently it was clearly shown in a wealth of experimental paradigms that innate or acquired deficiency of lymphocytes would impair peripheral prion pathogenesis, whereas no aspects of pathogenesis were affected by the presence or absence of lymphocytes upon direct transmission of prions to the CNS (Kitamoto et al., 1991; Lasmezas et al., 1996a). Klein and colleagues were then able to pinpoint the lymphocyte requirement to B cells (Aguzzi, 1997a; Klein et al., 1997): at first blush this was very surprising, since there had been no suggestions that any aspect of humoral immunity would be involved in prion diseases. In the same study, it was shown that T-cell deficiency brought about by ablation of the T-cell receptor (TCR) chain did not affect prion pathogenesis (Table I). TCR deficient mice, however, still contain TCR T-lymphocytes. Although the latter represents a subpopulation of T cells, the experiments described did not allow excluding a role for TCR T-lymphocytes. Therefore, we challenged also TCR TCR / deficient mice with prions. Incubation times after intracerebral and intraperitoneal inoculation of limiting or saturating doses of prions, however, elicited disease in these mice with the same kinetics as in wild-type mice (Michael Klein and Adriano Aguzzi, unpublished data). Also, accumulation of PrPSc and development of histopathological changes in the brain were indistinguishable in these two strains of mice. We therefore conclude that the complete absence of T cells has no measurable impact on prion diseases. Therefore, it is unlikely that the T cell infiltrates which have been reported to occur in the CNS during the course of prion infections (Betmouni et al., 1996; Betmouni and Perry, 1999; Perry et al., 2002) represent more than an epiphenomenon.

135


TABLE I SUSCEPTIBILITY OF VARIOUS STRAINS OF IMMUNODEFICIENT MICE TO INTRACEREBRALLY OR INTRAPERITONEALLY ADMINISTERED PRIONS (Klein et al., 1997) Intracerebral route Defect

CD-4o/o b CD-8o/o b 2-o/o b Perforin/

T T T T T T T T

Genotype

and and and and

B B B B

B IgG

SCID b RAG 2o/o RAG 1o/o AGRo/o d IgMo/o

b b

b b

t11umt

b

Intraperitoneal route

Scrapie

Time to terminal disease

Scrapie

Time to terminal disease

7/7 6/6 8/8 3/4a

159 11 157 15 162 11 171 2

8/8 6/6 7/7 4/4

191 1 202 5 211 6 204 3

7/8a 7/7 3/3 6/6

160 11 167 2 175 2 184 10

6/8e 0/7 0/5 0/7

226 15 healthy (> 504) healthy (> 258) healthy (> 425)

8/8

181 6

0/8

healthy (> 510)

5/5

170 3

4/4

223 2

FDC

TNFR1

7/7

165 3

9/9

216 4

Controls

129Sv C57BL/6

4/4 4/4

167 9 166 2

4/4 4/4

193 3 206 2

o/o c

Note: We measured the latency of scrapie (days from inoculation to terminal disease) upon delivery of a standard prion inoculum. All mice developed spongiform encephalopathy after i.c. inoculation. In contrast, all mice that carried a defect in B-cell differentiation stayed healthy after i.p. inoculation of RML scrapie prions. Intriguingly, in any of these several subsequent study, TNF receptor I deficient mice developed disease with the same kinetics as wild-type mice, although there are no morphologically detectable follicular dendritic cells in the spleens of TNFRI/ mice. a One perforin-deficient and one SCID mouse suffered from intercurrent death 135 and 141 days after inoculation, respectively. b Genetic background was C57BL/6. c Genetic background was 129sv. d Genetic background was C57BL/6 129sv. e Two SCID mice remained healthy and were sacrificed 303 and 323 days after inoculation.

In 1996–1997, when the results of the studies described above were being collected, we had no precise idea of what the mechanistic role of B cells might be in prion pathogenesis. It had become rather clear, on the other hand, that lymphocytes alone could not account for the entirety of prion pathogenesis, and an additional sessile compartment had to be involved: adoptive transfer of Prnp þ / þ bone marrow to Prnpo/o recipient mice did not suffice to restore infectibility of Prnp-expressing brain grafts, indicating that neuroinvasion was still defective (Bla¨ttler et al., 1997). In an immediately subsequent series of bone marrow transfer experiments, it emerged that peripheral prion pathogenesis required the physical presence

136

A. AGUZZI

of B cells, yet intraperitoneal infection occurred efficiently even in B-cell deficient hosts that had been transferred with B cell from Prnp knockout mice (Klein et al., 1998). Therefore, presence of B cells—but not expression of the cellular prion protein by these cells—is indispensable for pathogenesis upon intraperitoneal infection in the mouse scrapie model (Aguzzi et al., 2001a). The above results have been reproduced and confirmed several times over the years by many laboratories in various experimental paradigms: the requirement for B cells in particular appears to be very stringent in most instances investigated. However, it has emerged that not all strains of prions induce identical patterns of peripheral pathogenesis, even when propagated in the same, isogenic strain of host organism. One quite intriguing exception to the B-cell requirement appears to have been recently identified: Manuelidis and colleagues have reported efficient peripheral infection of B cell deficient mice with a mouse-adapted strain of human CJD prions (Shlomchik et al., 2001). Despite a somewhat polemical undertone and selective interpretation of previously published reports, this study is interesting because it identifies the unique situation of a prion strain which induces a highly anomalous peripheral pathogenesis. A. STRAINING

THE

LYMPHOCYTES

Another interesting discrepancy that remains to be addressed concerns the actual nature of the cells that replicate and accumulate prions in lymphoid organs. In the RML paradigm, four series of rigorously controlled experiments over five years (Aguzzi et al., 2000; Kaeser et al., 2001; Klein et al., 2001; Prinz et al., 2002) unambiguously reproduced the original observation by Thomas Bla¨ttler and colleagues that transfer of wild-type bone marrow cells (or fetal liver cells) to Prnp deficient mice restored accumulation and replication of prions in spleen (Bla¨ttler et al., 1997). In contrast, Brown et al. (1999) reported a diametrically opposite outcome of similar experiments when mice were inoculated with prions of the ME7 strain. Since trivial experimental artifacts appear to have been excluded, it is enticing to speculate that this discrepancy may identify yet another significant difference in the cellular tropism of different prion strains. It is important to note that the Bla¨ttler results do not necessarily indicate that lymphocytes are the primary splenic repository of prions: in fact, other experiments suggest that this is quite unlikely. Instead, bone marrow transplantation may (1) transfer an ill-defined population with the capability to replenish splenic stroma and to replicate prions, or—less probably—(2) donor-derived PrPC expressing hematopoietic cells may confer prion replication capability to recipient stroma by virtue of ‘‘GPI painting,’’ i.e., the post-translational cell-to-cell transfer of glycophosphoinositol linked extracellular membrane proteins (Kooyman et al., 1998). Some evidence might be accrued for either possibility: stromal splenic


137

follicular dendritic cells have been described by some authors to possibly derive from hematopoietic precursors, particularly when donors and recipients were young (Szakal et al., 1995; Kapasi et al., 1998). Conversely, instances have been described in which transfer of GPI-linked proteins occurs in vivo with suprisingly high efficiency (Kooyman et al., 1995). Most recently, GPI painting has been described specifically for the cellular prion protein (Liu et al., 2002). While it has been known for a long time that specific strains of prions may preferentially affect specific subsets of neurons, the Bla¨ttler/Brown paradox may uncover an analogous phenomenon in peripheral prion pathogenesis. The latter question may be much more important than it may appear at face value, and its clarification may warrant the investment of significant resources, since the molecular and cellular basis of peripheral tropism of prion strains is likely to be directly linked to the potential danger of BSE in sheep (Glatzel and Aguzzi, 2001; Bruce et al., 2002; Kao et al., 2002), as well as the potential presence of vCJD prions in human blood (Aguzzi, 2000). IV. Prion Hideouts in Lymphoid Organs

As mentioned above, prion infectivity rises very rapidly (in a matter of days) in the spleen of intraperitoneally infected mice. Although B-lymphocytes are crucial for neuroinvasion, a series of prion titrations in Prnp expressing and Prnp ablated mice, as well as reciprocal bone marrow chimaeras thereof, has virtually ruled out that the bulk of infectivity might be contained in lymphocytes. Instead, all evidence point to the fact that most splenic prion infectivity resides in a ‘‘stromal’’ fraction. Therefore, lymphocytes may be important for trafficking prions within lymphoid organs, but they do not appear to represent a major hideout for the infectious agent. Follicular dendritic cells are a prime candidate prion reservoir, since they express large amounts of cellular prion protein, and PrP accumulations tend to colocalize with FDCs in light and electron microscopical analyses of prion-infected spleens. More recently, elegant immuno electron microscopic studies have evidenced that prion immunoreactivity is situated in the immediate neighborhood of iccosomes (Jeffrey et al., 2000). A definitive assessment of the contribution of FDCs to prion pathogenesis continues to be problematic since the histogenesis and the molecular characteristics of these cells are ill-defined. In particular, there is a dearth of molecular markers for FDCs. FDCs express S-100 proteins, as well as the complement receptors 2 (CD35) and 4 (identical to the marker FDC-M2). All of these markers, however, are also expressed by additional cell types, even within lymphoid organs (Bofill et al., 2000). The FDC-M1 marker recognized by hybridoma clone 4C11 appears to be more specific, but it also stains

138

A. AGUZZI

tingible body macrophages, which are most likely of hematopoietic origin. In addition, the antigen recognized by 4C11 has not been molecularly characterized. Identification of FDC-specific genes would be extremely important and useful to this effect, since it would allow for transgenic studies of FDC histogenesis using fluorescent marker proteins. Other studies on the immediate wish list that would depend on FDC-specific transcription might include (1) transgenetic expression of Prnp restricted to FDCs, and (2) lineage ablation experiments by expressing conditional suicide genes— such as diphtheria toxin receptor or herpes simplex virus thymidine kinase—in FDCs. Gene deletion experiments in mice have shown that signaling by both TNF and lymphotoxins is required for FDC development (Fu et al., 1997; Koni et al., 1997; Endres et al., 1999). Membrane-bound lymphotoxin-/ (LT-/ ) heterotrimers signal through the LT- receptor (LT- R) (Ware et al., 1995) thereby activating a signaling pathway required for the development and maintenance of secondary lymphoid organs (Mackay et al., 1997; Matsumoto et al., 1997). Membrane LT-/ heterotrimers are mainly expressed by activated lymphocytes (Browning et al., 1993, 1995). Maintenance of pre-existing FDCs in a differentiated state requires continuous interaction with B-lymphocytes expressing surface LT-/ (Gonzalez et al., 1998). Inhibition of the LT-/ pathway in mice by treatment with LT Rimmunoglobulin fusion protein (LT R-Ig) (Crowe et al., 1994) leads to the disappearance of mature, functional FDCs (defined as cells that express markers such as FDC-M1, FDC-M2 or CD35) within one day, both in spleen and in lymph nodes (Mackay et al., 1997; Mackay and Browning, 1998). Prolonged administration of the LT R-Ig protein leads to disruption of B-cell follicles. All of the above prompted us (as a collaboration with Fabio Montrasio, Charles Weissmann, and Fabienne McKay) to study the effect of selective ablation of functional FDCs on the pathogenesis of scrapie in mice. FDCdepletion was maintained by weekly administration of the LT R-Ig fusion protein for eight weeks. Histological examination of spleen sections revealed that FDC networks had disappeared one week after treatment (Fig. 3), as expected. In mice, following peripheral inoculation, infectivity in the spleen rises within days and reaches a plateau after a few weeks (Bruce, 1985; Rubenstein et al., 1991; Buëler et al., 1993). Even after intracerebral inoculation, spleens of C57BL/6 animals already contain infectivity 4 days post infection (p.i.), (Buëler et al., 1993). However, Western blot analysis (Fig. 4) revealed that eight weeks after inoculation spleens of control mice showed strong bands of protease-resistant PrP, whereas mice injected weekly with LT R-Ig, starting


139

FIG. 3. Depletion of follicular dendritic cells by pharmacological inhibition of lymphotoxin signaling. Time course of FDC depletion in spleen of LT R-Ig treated mice. Frozen sections of treated (left) and control mice (right) immunostained with follicular dendritic cell specific antibody FDC-M1 at different times points after injection of LT R-Ig (original magnifications: upper row 8 25; lower row 8 63). Germinal centers FDCs networks were depleted already one week after treatment as described (Mackay and Browning, 1998; Montrasio et al., 2000). Some FDC-M1 positive cells, which may represent residual FDCs or tingible body macrophages, were still detectable in the spleens of treated mice.

either one week before or one week after inoculation, showed no detectable signal (less than 1/50th of the controls). Prion infectivity in three spleens for each time point was assayed by intracerebral inoculation into indicator mice (Fischer et al., 1996). In spleens of mice treated with LT R-Ig one week before intraperitoneal inoculation no infectivity could be detected after three or eight weeks (50-fold).

to control mice. Whenever FDC depletion was initiated one week before inoculation, the effect on incubation time was even more pronounced: Two out of three mice developed scrapie symptoms 60 days later than huIgGtreated controls and one mouse survived >340 days. Double-color immunofluorescence analysis of spleen sections of terminally sick animals revealed that the FDC networks were reconstituted after interruption of LT R-Ig administration at eight weeks and that PrPC and/or PrPSc co-localized with FDCs. Immunoblot analysis showed that PrPSc accumulation was restored upon reappearance of FDCs in the spleens of terminal sick LT R-Ig treated mice. These results show that follicular dendritic cells are essential for the deposition of PrPSc and generation of infectivity in the spleen and suggest that they participate in the process of neuroinvasion. Therefore, the requirement


141

for B cells for these processes might be explained by their essential role in the maturation of FDCs. PrP knockout mice expressing PrP transgenes only in B cells do not sustain prion replication in the spleen or elsewhere, suggesting that prions associated with splenic B cells (Raeber et al., 1999a) may be acquired from FDCs. Finally, these results suggest that strategies aimed at depleting FDCs might be envisaged for post-exposure prophylaxis (Aguzzi and Collinge, 1997) of prion infections initiated at extracerebral sites.

V. A Spleen is not a Lymph Node: Idiosyncrasies in the Lymphotropism of Prions

As discussed above, inhibition of the LT signaling pathway with a soluble receptor depletes FDCs (Mackay and Browning, 1998) and abolishes prion replication in spleens, thereby prolonging the latency of scrapie after intraperitoneal challenge (Montrasio et al., 2000). This suggests that B cell deficient MT mice (Kitamura et al., 1991) may be resistant to prions intraperitoneal (Klein et al., 1997) because of impaired FDC maturation (Klein et al., 1998; Montrasio et al., 2000). However, additional experimentation appears to indicate that PrPC-expressing hematopoietic cells are required in addition to FDCs for efficient lymphoreticular prion propagation (Bla¨ttler et al., 1997; Kaeser et al., 2001). This apparent discrepancy called for additional studies of the molecular requirements for prion replication competence in lymphoid stroma. We therefore set out to study peripheral prion pathogenesis in mice lacking TNF, LT/ , or their receptors. After intracerebral inoculation, all treated mice developed clinical symptoms of scrapie with incubation times, attack rates, and histopathological characteristics similar to those of wild-type mice, indicating that TNF/LT signaling is not relevant to cerebral prion pathogenesis. Upon intraperitoneal prion challenge, mice defective in LT signaling (LT/, LT /, LT R/, or LTTNF/) proved virtually non-infectible with 5 logLD50 scrapie infectivity, and establishment of subclinical disease (Frigg et al., 1999) was prevented. In contrast, TNFR1/ mice were almost fully susceptible to all inoculum sizes, and TNF/ mice showed dose-dependent susceptibility. TNFR2/ mice had intact FDCs and germinal centers, and were fully susceptible to scrapie. Unexpectedly, all examined lymph nodes (Fig. 5) of TNFR1/ and TNF/ mice had consistently high infectivity titers. Even inguinal lymph nodes, which are distant from the injection site and do not drain the peritoneum, contained infectivity titers equal to all other lymph nodes. Therefore, TNF deficiency prevents lymphoreticular prion accumulation in spleen but not in lymph nodes.

142

A. AGUZZI

FIG. 5. Determination of prion infectivity titers in spleens and lymph nodes of scrapiechallenged TNF- and LT-deficient mice. The diagrams represent concentration of prions at various time points after inoculation. Titers were determined in spleens (black circles, d), inguinal (grey circles, d), mesenteric (grey crosses, x) and cervical (grey triangles, ,) lymph nodes at various time points as indicated. Mice were inoculated intraperitoneally with 6 logLD50 of scrapie prions when not otherwise indicated. To avoid excessive cluttering of the diagrams, we opted to draw standard deviations within groups only when these exceeded 0.75 logLD50. Symbols on the abscissa indicate prion titers below detection limit (none of the 4 indicator mice developed scrapie). If one or more indicator mice survived >180 dpi, or the mean incubation time was over 120 days, titer was regarded as close to the detection threshold of the bioassay, and symbols were drawn on the threshold line.

Why is susceptibility to peripheral prion challenge preserved in the absence of TNFR1 or TNF, while deletion of LT signaling components confers high resistance to peripheral prion infection? After all, each of these defects (except TNFR2/) abolishes FDCs. For one thing, prion pathogenesis in the lymphoreticular system appears to be compartmentalized, with lymph nodes (rather than spleen) being important reservoirs of prion infectivity during disease. Second, prion replication appears to take place in lymph nodes even in the absence of mature FDCs. We generated chimeric mice in order to determine whether hematopoietic components are involved in prion propagation in TNFR1/ lymph nodes. In Prnpo/o mice grafted with TNFR1/ hematopoietic cells, high infectivity loads were detectable in lymph nodes but not spleens, indicating that TNFR1


143

deficient, PrPC expressing hematopoietic cells may support prion propagation within lymph nodes. These findings are in line with previous studies, which showed that chimeras of PrP-deficient hosts with PrP-expressing hematopoietic cells can accumulate prions chronically (Bla¨ttler et al., 1997; Kaeser et al., 2001). The PrP signal colocalized with a subset of macrophages in TNFR1/ lymph nodes. Since marginal zone macrophages are in close contact to FDCs and also interact with marginal zone B cells, this cell type is certainly a candidate supporter in prion uptake and replication. On the other hand, it has been reported that in short-time infection experiments depletion of macrophages appears to enhance the amount of recoverable infectivity, implying that macrophages may degrade prions rather than transport them (Beringue et al., 2000). The finding that a cell type other than mature FDCs is involved in prion replication and accumulation within lymph nodes may be relevant to the development of post-exposure prophylaxis strategies. VI. Sympathetic Nerves: A Neuroimmune Link?

In the last several years, a model has emerged that predicts prion neuroinvasion consists of two distinct phases (Aguzzi et al., 2001c). The details of the first phase are discussed above: widespread colonization of lymphoreticular organs is achieved by mechanisms that depend on B-lymphocytes (Klein et al., 1997, 1998), follicular dendritic cells (Montrasio et al., 2000), and complement factors (Klein et al., 2001). The second phase has long been suspected to involve peripheral nerves, and possibly the autonomic nervous system, and may depend on expression of PrPC by these nerves (Glatzel and Aguzzi, 2000). We have been attempting to test the requirement for expression of PrPC in peripheral nerves for several years, and have developed a gene transfer protocol to spinal ganglia aimed at resolving this question (Glatzel et al., 2000); however, we were never able to recover infectivity in spinal cords of Prnpo/o mice whose spinal nerves had been transduced by Prnp-expressing adenoviruses (Glatzel and Aguzzi, unpublished results). Also, fast axonal transport does not appear to be involved in prion neuroinvasion, since mice that are severely impaired in this transport mechanism experience prion pathogenesis with kinetics similar to that of wild-type mice (Ku¨nzi et al., 2002). There is substantial evidence suggesting that prion transfer from the lymphoid system to the CNS occurs along peripheral nerves in a PrPC dependent fashion (Bla¨ttler et al., 1997; Glatzel and Aguzzi, 2000; Race et al., 2000). Studies focusing on the temporal and spatial dynamics of neuroinvasion have suggested that the autonomic nervous system might be responsible for transport from lymphoid organs to the CNS (Clarke and Kimberlin, 1984; Cole and Kimberlin, 1985; Beekes et al., 1998; McBride and Beekes, 1999).

144

A. AGUZZI

FIG. 6. The sympathetic innervation of the spleen can be ablated by 6-hydroxydopamine. The density of sympathetic innervation correlates with the concentration of tyrosine hydroxylase, which was measured by Western blot analysis in spleens of adult 6-OHDA treated mice, K14NGF transgenic mice and controls. The two main tyrosine hydroxylase bands are visible at ca. 55 kDa. While, there is a certain increase in sympathetic innervation with age, it is obvious that 6-OHDA treatment depletes, and the K14NGF transgene increases sympathetic innervation. Lower panel: quantification of tyrosine hydroxylase content by chemiluminescence scanning. The bars represent the ratio between signal intensity of the tyrosine hydroxylase band and that of the corresponding beta-actin bands (arbitrary units).

The innervation pattern of lymphoid organs is mainly sympathetic (Felten and Felten, 1988). Markus Glatzel and colleagues have shown that denervation by injection of the drug 6-hydroxydopamine (6-OHDA), as well as ‘‘immunosympathectomy’’ by injection of antibodies against nerve growth factor (NFG) leads to a rather dramatic decrease in the density of sympathetic innervation of lymphoid organs (Fig. 6), and significantly delayed the development of scrapie (Glatzel et al., 2001) (Fig. 7). Sympathectomy appears to delay the transport of prions from lymphatic organs to the thoracic spinal cord, which is the entry site of sympathetic nerves to the CNS. Transgenic mice overexpressing NGF under control of the K14 promoter, whose spleens are hyperinnervated, developed scrapie significantly earlier than nontransgenic control mice. No alteration in lymphocyte subpopulations was detected in spleens at any time of investigation. In particular, we did not detect any significant differences in the content of FDCs of treated and


145

untreated mice, which negates the possibility that the observed protection may be due to modulation of FDC microanatomy. While the sympathetic nervous system may represent a major component of the second phase of prion neuroinvasion, many details remain to be elucidated. It is not known whether prions can be transferred directly from FDCs to sympathetic endings, or whether additional cell types are involved. The latter possibility is particularly enticing, since FDCs have not been shown to entertain physical contact with sympathetic nervous system terminals. Moreover, it is unclear how prions are actually transported within peripheral nerves. Axonal and non-axonal transport mechanisms may be involved, and non-neuronal cells (such as Schwann cells) may play a role. Within the framework of the protein-only hypothesis, one may hypothesize a ‘‘domino’’ mechanism, by which incoming PrPSc converts resident PrPC on the axolemmal surface, thereby propagating spatially the infection. While speculative, this model is attractive since it may accommodate the finding that the velocity of neural prion spread is extremely slow (Kimberlin et al., 1983) and may not follow the canonical mechanisms of fast axonal transport. Indeed, recent studies may favor a non-axonal transport mechanism that results in periaxonal deposition of PrPSc (Hainfellner and Budka, 1999; Glatzel and Aguzzi, 2000). The fact that denervated mice eventually developed scrapie may be due to (i) an alternative, low-efficiency route of entry that may become uncovered by the absence of sympathetic fibers, or (ii) because of incomplete sympathectomy. Entry through the vagal nerve has been proposed in studies of the dynamics of vacuolation following oral and intraperitoneal challenge with prions (Baldauf et al., 1997; Beekes et al., 1998). There was no evidence for transport along the vagal nerve in sympathectomized mice, since vagal nuclei were affected similarly to other regions of the brain stem unrelated to the vagal system. This supports the hypothesis that delayed neuroinvasion in denervated mice may occur because of residual innervation of lymphoreticular organs. The surprising finding that infectious titers in hyperinnervated spleens are at least two logs higher and show enhanced PrPSc accumulations compared to control mice suggests that sympathetic nerves, besides being involved in the transport of prions, may also accumulate and replicate prions in lymphatic organs (Clarke and Kimberlin, 1984). Obviously this finding has implications related to the permanence and possibly eradication of prions in subclinically infected hosts. A. SPREAD

OF

PRIONS WITHIN

THE

CENTRAL NERVOUS SYSTEM

Ocular administration of prions has proved particularly useful to study neural spread of the agent, since the retina is a part of the CNS and

146

A. AGUZZI

FIG. 7. Modulation of sympathetic nerve density in lymphoid organs affects the efficiency of peripheral prion pathogenesis. Survival plots displaying the incubation time (days) until development of terminal scrapie in C57Bl/6 mice inoculated i.p. with prions (A–C). Mice injected with 6-OHDA in their adulthood (A), as well as mice injected neonatally with anti-NGF


147

intraocular injection does not produce direct physical trauma to the brain, which may disrupt the blood-brain barrier and impair other aspects of brain physiology. The assumption that spread of prions occurs axonally rests mainly on the demonstration of diachronic spongiform changes along the retinal pathway following intraocular infection (Fraser, 1982). To investigate whether spread of prions within the CNS is dependent on PrPC expression in the visual pathway, PrP-producing neural grafts were used as sensitive indicators of the presence of prion infectivity in the brain of an otherwise PrP-less host. Following inoculation with prions into the eye of grafted Prnpo/o mice, none of the grafts showed signs of scrapie. Therefore, it was concluded that infectivity administered to the eye of PrP-deficient hosts cannot induce scrapie in a PrP-expressing brain graft (Brandner et al., 1996b). Engraftment of Prnpo/o mice with PrPC-producing tissue might lead to an immune response to PrP which, in turn, was shown to be in principle capable of neutralizing infectivity (Heppner et al., 2001b). In order to definitively rule out the possibility that prion transport was disabled by a neutralizing immune response, Prnpo/o mice were rendered tolerant by expressing PrPC under the control of the lck promoter. These mice overexpress PrP on T-lymphocytes, but are resistant to scrapie and do not replicate prions in brain, spleen and thymus after intraperitoneal inoculation with scrapie prions (Raeber et al., 1999b). Engraftment of these mice with PrP-overexpressing neuroectoderm did not lead to the development of antibodies to PrP after intracerebral or intraocular inoculation, presumably due to clonal deletion of PrP-immunoreactive T-lymphocytes. As before, intraocular inoculation with prions did not provoke scrapie in the graft, supporting the conclusion that lack of PrPC, rather than immune response to PrP, prevented prion spread (Brandner et al., 1996b). Therefore, PrPC appears to be necessary for the spread of prions along the retinal projections and within the intact CNS. These results indicate that intracerebral spread of prions is based on a PrPC-paved chain of cells, perhaps because they are capable of supporting prion replication. When such a chain is interrupted by interposed cells that lack PrPC, as in the case described here, no propagation of prions to the target tissue can occur. Perhaps prions require PrPC for propagation across synapses: PrPC is present in the synaptic region (Fournier et al., 1995) and certain synaptic properties are altered in Prnpo/o mice (Collinge et al., 1994; Whittington et al., 1995). Perhaps transport of prions within (or on the surface

antibodies or with 6-OHDA (B) developed terminal scrapie significantly later than their respective controls. Instead K14NGFmice, which have hyperinnervated spleens, developed terminal scrapie earlier than matched wild-type siblings of the same genetic background (C). The size of inoculum is expressed in infectious units (logLD50).

148

A. AGUZZI

of) neuronal processes is PrPC-dependent. Within the framework of the protein-only hypothesis (Griffith, 1967; Prusiner, 1989), these findings may be accommodated by a ‘‘domino-stone’’ model in which spreading of scrapie prions in the CNS occurs per continuitatem through conversion of PrPC by adjacent PrPSc (Aguzzi, 1997a).

VII. Innate Immunity and Antiprion Defense

A. MACROPHAGES

AND

TOLL-LIKE RECEPTORS

Cells of the monocyte/macrophage lineage typically represent the first line of defense against an extremely broad variety of pathogens. In the case of prions, it might be conceivable that macrophages protect against prions. However, it would be equally conceivable that macrophages, by virtue of their phagocytic properties and of their intrinsic mobility, may function as Trojan horses that transport prion infectivity between sites of replication within the body. This interesting question has not yet been fully resolved. In a short-term prion infection paradigm, Beringue and colleagues administered dichloromethylene disphosphonate encapsulated into liposomes to mice: this eliminates all spleen macrophages temporarily. Accumulation of newly synthesized PrPSc was accelerated, suggesting that macrophages participate in the clearance of prions, rather than being involved in PrPSc synthesis. On the basis of the results presented above, Beringue and colleagues have suggested that activation or targeting of macrophages may represent a therapeutic pathway to explore in TSE infection. This suggestion was taken up by Sethi and Kretzschmar, who recently reported that activation of Toll-like receptors (TLRs), which function as general stimulators of innate immunity by driving expression of various sets of the immune regulatory molecules, can effect postexposure prophylaxis in an experimental model of intraperitoneal scrapie infection (Sethi et al., 2002). In this experimental paradigm administration of prions intraperitoneally elicited disease after approximately 180 days, whereas the administration of CpG oligodeoxynucleotides 7 h after prion inoculation and daily for 20 days led to disease-free intervals of ‘‘more than 330 days’’—although it appears that all inoculated mice died of scrapie shortly thereafter (communicated by H. Kretzschmar at the TSE conference in Edinburgh, September 2002). This finding is very surprising, since most available evidence indicates that general activation of the immune system would typically sensitize mice to prions, rather than protect them. The mechanism by which activation of TLR can result in post-exposure prophylaxis are wholly unclear at present, particularly in view of the fact that mice lacking Myd88 (Adachi et al., 1998),


149

which is an essential mediator of TLR signaling, develop prion disease with exactly the same sensitivity and kinetics as wild-type mice (M. Prinz, M. Heikenwa¨lder, and A. Aguzzi, unpublished results). B. COMPLEMENTING SCRAPIE Another prominent component at the crossroad between innate and adaptive immunity is represented by the complement system. Opsonization by complement system components also appears to be relevant to prion pathogenesis: mice genetically engineered to lack complement factors (Klein et al., 2001), or mice depleted of the C3 complement component by administration of cobra venom (Mabbott et al., 2001), exhibit a remarkable resistance to peripheral prion inoculation. This phenomenon may, once again, be related to the pathophysiology of FDCs, which typically function as antigen traps. Trapping mechanisms essentially consist of capture of immune complexes by Fc receptors, and binding of opsonized antigens (linked covalently to C3d and C4b complement adducts) to the CD21/CD35 complement receptors. Capture mediated by Fc receptors does not appear to be very important in prion disease: for one thing, knockout mice lacking Fc receptors (Takai et al., 1994; Hazenbos et al., 1996; Takai et al., 1996; Park et al., 1998) are just as susceptible to intraperitoneally administered scrapie as wild-type mice. Further, introduction into MT mice of a generic immunoglobulin chain fully restored prion neuroinvasion irrespective of whether this heavy chain allowed for secretion of immunoglobulins, or only for production of membrane-bound immunoglobulins. We therefore conclude that circulating immunoglobulins are certainly not crucial to prion replication in lymphoid organs and to neuroinvasion. A second mechanism exploited by FDCs for antigen trapping involves covalent linking of proteolytic fragments of the complement components C3 and C4 (Szakal and Hanna, 1968; Carroll, 1998). The CD21/CD35 complement receptors on FDCs bind C3b, iC3b, C3d, and C4b through short consensus repeats in their extracellular domain. Ablation of C3, or of its receptor CD21/CD35, as well as C1q (alone or combined with BF/C2/), delayed neuroinvasion significantly after intraperitoneal inoculation when a limiting dose of prions was administered. These effects suggest that opsonization of the infectious agent may enhance its accessibility to germinal centers by facilitating docking to FDCs. Very large prion inocula ( 106 infectious units) appear to over-ride the requirement for a functional complement receptor in prion pathogenesis. This is similar to systemic viral infections and coreceptor-dependent retention within the follicular compartment, whose necessity can be overridden by very high affinity antigens (Fischer et al., 1998) or adjuvants (Wu et al., 2000).

150

A. AGUZZI

Additional retention mechanisms for prions may therefore exist in FDCs, which are not complement-dependent, or depend on hitherto unidentified complement receptors. VIII. Adaptive Immunity and Pre-Exposure Prophylaxis Against Prions

For many conventional viral agents, vaccination is the most effective method of infection control. But is it possible to induce protective immunity in vivo against prions at all? Prions are extremely sturdy and their resistance against sterilization is proverbial. Pre-incubation with anti-PrP antisera was reported to reduce the prion titer of infectious hamster brain homogenates by up to 2 log units (Gabizon et al., 1988) and an anti-PrP antibody was found to inhibit formation of PrPSc in a cell-free system (Horiuchi and Caughey, 1999). Also, antibodies (Klein et al., 2001) and F(ab) fragments raised against certain domains of PrP (Peretz et al., 2001) can suppress prion replication in cultured cells. However, it is difficult to induce humoral immune responses against PrPC and PrPSc. This is most likely due to tolerance of the mammalian immune system to PrPC, which is an endogenous protein expressed rather ubiquitously. Ablation of the Prnp gene (Buëler et al., 1992), which encodes PrPC, renders mice highly susceptible to immunization with prions (Brandner et al., 1996b), and many of the best available monoclonal antibodies to the prion protein have been generated in Prnpo/o mice (Prusiner et al., 1993). However, Prnpo/o mice are unsuitable for testing vaccination regimens since they do not support prion pathogenesis (Buëler et al., 1993). We have therefore asked whether genes encoding high-affinity anti-PrP antibodies (originally generated in Prnpo/o mice) may be utilized to reprogram B-cell responses of prion-susceptible mice that express PrPC. Indeed, introduction of the epitope-interacting region the heavy chain of 6H4, a highaffinity anti-PrP monoclonal antibody (Korth et al., 1997) into the germ line of mice sufficed to produce high-titer anti-PrPC immunity. The 6H4 heavy chain transgene induced similar anti-PrPC titers in Prnpo/o, Prnpo/ þ and Prnp þ / þ mice, indicating that deletion of autoreactive B cells does not prevent anti-PrP immunity. The buildup of anti-PrPC titers, however, was more sluggish in the presence of endogenous PrPC, suggesting that some clonal deletion is actually occurring. How can these observations be interpreted? The total anti-PrPC titer results from pairing of one transgenic heavy chain with a large repertoire of endogenous and chains: some pairings may lead to reactive moieties, while others may be anergic (Fig. 8b). Maybe the B cell clones with the highest affinity to PrPC are being eliminated by the immune tolerization machinery, and only clones with medium affinity are retained (Fig. 8b). This would explain the delay in titer buildup in the presence of PrPC, and would be in


151

FIG. 8. Affinity and avidity of antibodies against the prion protein. (a) In order to gain some indication on the relative avidity of the immune response against PrPC in 6H4 transgenic mice, we compared the binding of serum from transgenic mice to a dilution series of the 6H4 original monoclonal antibody. When normalized against the immunoglobulin concentration, taking into account that the anti-PrPC antibodies 6H4 transgenic mice are pentameric, we found that the total avidity of the transgenic serum is approximately 2 log units lower than that of the monoclonal antibody from which the transgene was derived. A possible explanation for this phenomenon is depicted in (b). While the heavy chain is kept constant in the transgene, it may pair with a large repertoire of endogenous light chains. Some of the pairs will yield very high-affinity antibodies, others will have low affinity, but the majority may have no affinity at all for the prion protein. It is possible that some degree of clonal deletion occurs, and that combinations with the highest affinity are eliminated.

152

A. AGUZZI

agreement with our affinity measurements—which indicate that the total molar avidity of 6H4m serum is approximately 100-fold lower than that of the original 6H4 antibody from which the transgene was derived (Fig. 8a). But most intriguingly, expression of the 6H4 heavy chain sufficed to block peripheral prion pathogenesis upon intraperitoneal inoculation of the prion agent (Heppner et al., 2001a). PrPC is a normal protein expressed by most tissues of the body. Therefore an anti-PrP immune response may conceivably induce an autoimmune disease, and defeat any realistic prospect for prion vaccination. We did not observe any blatant autoimmune disease as a consequence of anti-prion immunization—unless PrPC was artificially transgenically expressed at nonphysiological, extremely high levels. The strategy outlined above delivers proof-of-principle that a protective humoral response against prions can be mounted by the mammalian immune system, and suggests that B cells are not intrinsically tolerant to PrPC. If the latter is generally true, lack of immunity to prions may be due to T-helper tolerance. The latter problem is not trivial, but may perhaps be overcome by presenting PrPC to the immune system along with highly active adjuvants. These findings, therefore, encourage a reassessment of the possible value of active and passive immunization (Westaway and Carlson, 2002), and perhaps of reprogramming B-cell repertoires by chain transfer, in prophylaxis or in therapy of prion diseases. A. PRION IMMUNIZATION

AND ITS

REDUCTION

TO

PRACTICE

Approximately one year later, a first example of a reduction to practice of the approach proposed by Heppner was demonstrated by the laboratory of Thomas Wisnieswki (Sigurdsson et al., 2002). The authors have explored whether immunization with recombinant prion protein might be protective against prion diseases. Two paradigms were chosen: prophylactic immunization, and rescue after infection (post-exposure prophylaxis). The first remarkable surprise of the study was that it was indeed possible to induce antibody responses in wild-type mice—although the actual titers were not determined. Immunization was simply achieved by injecting 50 g of recombinant prion protein emulsified in Freund’s adjuvant. This procedure had been utilized extensively in the Zurich laboratory (F. L. Heppner, M. Polymenidou, E. Pellicioli, M. Bachmann, and A. Aguzzi, unpublished observations), but had never produced any reasonable titers in wild-type mice. Upon inoculation with a high dose or with a lower dose of prion inoculum, vaccinated mice exhibited a modest delay in development of prion disease. Although the success of the study, from the viewpoint of survival of the mice, might be regarded as limited, the study in my view is very promising as it


153

shows that vaccination approaches can be translated into models that are closer to the real-life situation than immunoglobulin-transgenic mice. The possibility of producing protective immunity against prions has captured the imagination of a considerable number of scientists, and additional reports are appearing for original ways to break tolerance and induce an immune response in animals that express the normal prion protein, including mixing of the immunogenic moiety with bacterial chaperons, among others (Koller et al., 2002).

IX. The Prion Doppelganger

Despite considerable efforts, the physiological function of PrPC is still unclear. PrPC was reported to have weak superoxide dismutase activity in vitro (a report that awaits independent confirmation) and evidence for a physiological relevance is scant (Brown et al., 2001). PrPC can bind copper (Brown et al., 1997; Wadsworth et al., 1999), but reports of increased copper content of neurons lacking PrPC have been questioned by contradictory findings (Waggoner et al., 2000). Similarly, a proposed role for PrPC in synaptic function has been questioned by other investigators (Collinge et al., 1994; Lledo et al., 1996) and a possible regulation of circadian rhythm by PrPC has remained more-or-less anecdotal (Tobler et al., 1996). It is fair to say that if it was not for its role in prion diseases, the significance of PrPC would be obscure and it might be viewed as just another one among many poorly understood GPI-linked proteins. After the advent of large scale sequencing efforts and genome projects, it was realized that there is an open reading frame (ORF) directly adjacent of Prnp that encodes a protein sharing significant homology with PrPC (Moore et al., 1999). The novel gene, Prnd, is located 16 kb downstream Prnp in the mouse genome, and encodes a protein of 179 residues, which was termed Dpl (‘‘downstream of the Prnp locus’’ or ‘‘doppel,’’ German for ‘‘double’’) (Moore et al., 1999; Weissmann and Aguzzi, 1999; Behrens and Aguzzi, 2002). The Prnd gene is evolutionarily conserved from humans to sheep and cattle, and shows roughly 25% identity with the carboxy-proximal two thirds of PrPC. Structural studies indicate that Dpl contains three alpha helices like PrPC and two disulphide bridges between the second and third helix (Lu et al., 2000; Silverman et al., 2000; Mo et al., 2001) Dpl mRNA is expressed at high levels in testis, less in other peripheral organs and, notably, at very low levels in brain of adult wild-type mice. However, significant Prnd mRNA transcripts were detected during embryogenesis and in the brains of newborn mice, arguing for a possible function of Dpl in brain development (Li et al., 2000).

154

A. AGUZZI

To understand the function of Dpl better, we have inactivated the Prnd gene in embryonic stem (ES) cells (Behrens et al., 2001). Similarly mice lacking PrPC, mice devoid of Dpl survive to adulthood and do not show obvious phenotypical alterations, suggesting that Dpl is dispensable for embryogenesis and postnatal development (Behrens et al., 2002). The similarities between PrPC and Dpl in primary amino acid sequence, structure and subcellular localization suggest related biological functions. Therefore, a possible role of PrPC and Dpl during development may be masked by functional redundancy. To address this question it will be necessary to generate mice lacking Prnp as well as Prnd and to study whether the lack of both PrPC and Dpl will result in an exacerbation of the mutant phenotypes. These double-mutant mice may finally reveal the true physiological function of PrPC and Dpl, and pave the way for the long-awaited understanding of these proteins. The Dpl protein resembles an N-terminally truncated PrPC protein lacking the octamer repeats. But the latter version of PrPC is actually capable of supporting PrPSc propagation (Flechsig et al., 2000), suggesting that the Dpl protein may in principle be susceptible to conversion into ‘‘DplSc.’’ However, presently there is no evidence whether upon scrapie inoculation Dpl can be converted into a misfolded beta-strand-rich, protease-resistant conformation. Embryonic stem (ES) cells carrying a homozygous null mutation of the Prnd locus, and a normal Prnp locus, were found to be capable of giving rise to all neural cell lineages when transplanted into host brains. After inoculation with scrapie prions, Dpl-deficient neural grafts showed spongiosis, gliosis and unimpaired accumulation of PrPSc and infectivity similar to wildtype neuroectodermal grafts (Behrens et al., 2001). Therefore, in neural grafts Prnd deficiency does not prevent prion pathogenesis. It is important to note, however, that this experimental approach does not rule out a role for Dpl in peripheral prion pathogenesis and in PrPSc transport to the brain. The latter possibility is worth studying, because Prnd is expressed in the spleen, a major peripheral reservoir of PrPSc and prion infectivity (Li et al., 2000). Four polymorphisms in human PRND were detected, but no strong association was found between any of these polymorphisms and human prion diseases (Mead et al., 2000; Peoc’h et al., 2000). These findings further argue against an important function of Dpl in neurons during prion disease, at least in genetically determined forms of these diseases. A. PHENOTYPES

OF

Prnp DEFICIENT MICE: A PARADOX RESOLVED

Although this was not realized for quite some time, a Dpl-associated phenotype had been accidentally produced in knockout mice lacking PrPC.

155


Several mouse lines with targeted disruptions of Prnp were independently generated. All mutant mouse lines lacked significant portions of the Prnp ORF and did not produce PrPC protein, but showed two strikingly different phenotypes. Zrch Prnpo/o and Edbg Prnp/ (termed after the city of origin) showed only minor defects (Buëler et al., 1992; Collinge and Palmer, 1994; Manson et al., 1994; Tobler et al., 1996) whereas Ngsk Prnp/, Zu¨rich II and Rcm0 mice develop cerebellar Purkinje cell degeneration causing ataxia with advancing age (Sakaguchi et al., 1996; Moore et al., 1999; Rossi et al., 2001). This conundrum was solved when David Westaway and colleagues realized that in the brain of ataxic, but not of healthy Prnp-mutant mice, Prnd mRNA was upregulated (Moore et al., 1999; Li et al., 2000). An intergenic splicing event places the Dpl locus under the control of the Prnp promoter, probably due to the deletion of the Prnp intron 2 sequence including its splicing acceptor (Moore et al., 1999). This intergenic splicing event could be detected at very low levels also in wild-type mice, but was greatly enhanced by the absence of the intron 2 splice acceptor (Moore et al., 1999). Whereas the Prnp promoter is strongly expressed in neuronal cells, the Prnd promoter is not (Moore et al., 1999; Li et al., 2000; Rossi et al., 2001) and therefore Prnd expression from the Prnp promoter results in overproduction of Dpl in the brain (Moore et al., 1999; Li et al., 2000; Rossi et al., 2001). Further experiments have demonstrated an inverse correlation between the mRNA levels of Prnd and the onset of ataxia. Disease progression was accelerated by increasing Prnd levels, supporting the notion that ectopic Dpl expression, but not functional loss of PrPC, may be responsible for neuronal degeneration in ataxic Prnp-deficient mice (Rossi et al., 2001). B. CONSEQUENCES PATHOLOGY

OF

DOPPEL DEFICIENCY: A DETOUR

TO

REPRODUCTIVE

We then inactivated the Prnd gene. However, we have not been able to generate any progeny from intercrosses of Prnd/ mice. Female Prnd/ mice, when crossed to Prnd/ or Prnd þ / mice, yielded litter sizes similar to those of wild-type. In contrast, male Prnd/ mice were infertile. Their sexual activity was similar to that of controls, as shown by a normal number of copulation plugs. However, the number of spermatozoa in the cauda epididymis of Prnd/ males was reduced, and motility of mutant sperm was decreased. Therefore, sterility of Dpl-mutant males is not due to behavioral abnormalities, but may be due to a spermatogenesis defect. Indeed, Prnd/ sperm heads were severely malformed and lacked a discernible welldeveloped acrosome (Fig. 9). As the acrosome is essential for sperm-egg interaction, this defect could explain the sterility of Prnd deficient males. In vitro fertilization (IVF) experiments confirmed that spermatozoa isolated from Prnd/ males were unable to fertilize wild-type oocytes. Spermatozoa of

156

A. AGUZZI

FIG. 9. Morphology of sperms in mice lacking the Doppel gene. Spermatozoa from Dpldeficient mice are often heavily malformed. (A, B) Photographs of bright field images of spermatozoa isolated from wild-type (A) and Prnd/ mice (B). (C–D) Spermatozoa isolated from Prnd þ / þ (C) and Prnd/ mice (D) were stained with mitotracker to detect mitochondria (in green) and with the DNA stain Hoechst (in blue). (E–H) Photographs of bright field images of sperm heads from Prnd þ / þ (E) and Prnd/ mice (F–H).

Prnd/ males never penetrated the zona pellucida. However, if the zona pellucida was partially dissected and IVF was performed with sperm suspension from Prnd/ males, fertility was partially rescued. These data indicate that Prnd/ spermatozoa are capable of oocyte fertilization, albeit at a lower frequency than controls, but that they cannot overcome the barrier imposed by the zona pellucida (Fig. 10). A significant amount of PrPC is expressed in mature spermatozoa. The PrP protein found in testes was truncated in its C terminus in the vicinity of residue 200 (Shaked et al., 1999). A protective role for PrP against copper toxicity has been proposed: sperm cells originating from Prnp ablated mice were more susceptible to high copper concentrations than wild-type sperm. However, male Prnpo/o knockout mice are not sterile and produce normal litter sizes. PrP expressed in testes is clearly not capable of compensating for the absence of Dpl, suggesting non-redundant functions for the two proteins. However, Dpl may mask a minor function of PrPC in testes development or spermiogenesis: therefore it will be very interesting to explore whether males


157

FIG. 10. A possible model for the function of Dpl in fertilization. (A) Male sterility in mice that lack the Doppel protein appears to come about because Doppel deficient sperms are incapable of fertilizing eggs. The process that appears to be disturbed relates to the acrosomal reaction and to the penetration of the zona pellucida. In fact, mechanical dissection of the zona pellucida restores, at least in part, fertility. (B) Interestingly, haploid spermatozoa lacking the Doppel gene (Prnd) are perfectly fertile when generated in the context of a heterozygous Prnd þ / mouse. Instead, Prnd sperms are infertile when generated in a Prnd/ mouse. This may be because sperms spend much of the maturation time in the form of syncytia, with maturing cells connected to each other by cytoplasmic bridges. This may allow sufficient amounts of Doppel protein to be transferred from Prnd þ to Prnd spermatids, and rescue fertility in trans.

158

A. AGUZZI

lacking both PrPC and Dpl might display a more severe defect than Prnd/ single mutant mice. At present the molecular mechanism of Dpl-regulated acrosome development is unclear. Dpl may be present on the acrosomic vesicles through its GPI-anchor, and possibly participate in acrosome morphogenesis. Alternatively, Dpl may regulate acrosome function in a more indirect way. We have also observed that sperm isolated from Prnd/ mice is greatly impaired in sperm-egg interaction and that Prnd/ spermatozoa fail to trigger the acrosomal reaction. Oligosaccharides have been implicated in sperm binding and signaling for the acrosome reaction, but the composition and structure of the essential carbohydrate moieties remain controversial (Wassarman et al., 2001). Dpl is a highly glycosylated protein located at the outside of the plasma membrane (Moore et al., 1999; Silverman et al., 2000). It is possible that Dpl present on the sperm plasma membrane is directly involved in sperm-egg interaction. In this context, it is interesting to mention that the laboratory of David Melton in Edinburgh has reported a slightly different phenotype of Dpldeficient mice (communicated by Derek Paisley and David Melton at the international TSE conference in Edinburgh, September 2002). In contrast to the mice generated in Zurich, Doppel deficient homozygous males in Edinburgh appear to be able to fertilize eggs. In a series of in vitro fertilizations, the Melton laboratory has reported that progression to the early cleavage divisions occurred, but thereafter was soon followed by death of the embryos at the preimplantation stage. Instead, there were no obvious malformations of sperms. Whether this discrepancy is related to a divergent genetic background of the mice utilized, whether it points to slightly different targeting strategies, or finally whether it might uncover yet another surprising phenomenon in the genetics of prion-related genes, is—at the time of writing—wholly unclear. With regard to the keen interest in the development of new methods of contraception (in particular those targeting the male) the phenotype of Prnd/ mutant mice suggests that inhibition of Dpl function may provide a novel target for contraceptive intervention. X. Prion Immunology: Quo Vadis?

Until approximately 10 years ago, it was thought that prion diseases are completely independent of the immune system. The lack of overt inflammation in spongiform lesions lent support to this viewpoint. It is fascinating to note that this view has been completely turned upside-down, and it is now obvious that the relationships between the infectious agent and various components of the immune system are intimate, complex, and


159

multifaceted. While several crucial questions are still open, notably regarding the true portal of entry of BSE prions into the human body, and the mechanism by which they cause new vJCD, a fascinating multitude of unexpected findings has been reported, some of them in extraordinary detail. In the long run, one would hope that this accrued knowledge will be put into useful practice. To that effect, it is encouraging to note that a sizeable number of the steps in prion transport, which have been discussed above, appears to be rate-limiting. Because of that, these steps lend themselves as a target for interventions, which may be therapeutic or prophylactic (Aguzzi et al., 2001b). A number of substances appear capable of influencing the outcome of a contact of mammalian organisms with prions: a non-exhaustive list includes compounds as diverse as Congo red (Caughey and Race, 1992), amphotericin B (Pocchiari et al., 1987), anthracyclin derivatives (Tagliavini et al., 1997), sulfated polyanions (Caughey and Raymond, 1993), pentosan polysulphate (Farquhar et al., 1999), soluble lymphotoxin- receptors (Montrasio et al., 2000), porphyrins (Priola et al., 2000), branched polyamines (Supattapone et al., 2001), and beta-sheet breaker peptides (Soto et al., 2000). However, it is sobering that none of the substances have yet made it to any validated clinical use: quinacrine appears to represent the most recent unfulfilled promise (Collins et al., 2002). On the other hand, the tremendous interest in this field has attracted researchers from various neighboring disciplines, including immunology, genetics, and pharmacology, and therefore it is to hope that rational and efficient methods for managing prion infections will be developed in the future.

XI. An Essential Glossary of Prion Jargon

Prion: Agent of transmissible spongiform encephalopathy (TSE), with unconventional properties. The term does not have structural implications other than that a protein is an essential component. ‘‘Protein-only’’ hypothesis: Maintains that the prion is devoid of informational nucleic acid, and that the essential pathogenic component is protein (or glycoprotein). Genetic evidence indicates that the protein is an abnormal form of PrP (perhaps identical with PrPSc). The association with other ‘‘non-informational’’ molecules (such as lipids, glycosamino glycans, or maybe even short nucleic acids) is not excluded. PrPC: The naturally occurring form of the mature Prnp gene product. Its presence in a given cell type is necessary, but not sufficient, for replication of the prion.

160

A. AGUZZI

PrPSc: An ‘‘abnormal’’ form of the mature Prnp gene product found in tissue of TSE sufferers, defined as being partly resistant to digestion by proteinase K under standardized conditions. It is believed to differ from PrPC only (or mainly) conformationally, and is often considered to be the transmissible agent or prion. These definitions (Aguzzi and Weissmann, 1997) describe our terminology which is, however, not agreed on by convention and is not necessarily used by others. ACKNOWLEDGMENTS This chapter is primarily dedicated to the present and past members of my laboratory, without whose enthusiastic dedication the work described here would not have been performed. Special thanks to Petra Schwarz for maintaining our prion-infected mouse colony in an impeccable shape, to Rolf Zinkernagel for critical reading of the manuscript, and to Susanne Tiefenthaler for excellent secretarial assistance. The work of my lab is supported by the Canton of Zu¨rich, the Faculty of Medicine at the University of Zu¨rich, the Swiss Federal Offices of Education and Science, of Health, and of Animal Health, the Swiss National Foundation, the National Center for Competence in Research on neural plasticity and repair, the Migros foundation, the Coop foundation, the U.K. Department for Environment, Food, & Rural Affairs, and the Stammbach foundation. Individual members of the Aguzzi lab have been directly supported by the EMBO fellowship program (Isabelle Arrighi), HFSP and Bonizzi-Theler foundation (Frank Heppner), DFG (Michael Klein, Marco Prinz, Kirsten Mertz), the Koetser Foundation (Nicolas Genoud), the FAN Society for the Support of Young Academic Scientists (Mathias Heikenwa¨lder, Christoph Huber), the SBF foundation (Christoph Huber), the Catello family (Mathias Heikenwa¨lder), the Federal Office of Health (Manuela Maissen), the Center of Neuroscience Zu¨rich (Magda Polymenidou), the Roche Foundation (Mark Zabel), the UK Biotechnology and Biological Sciences Research Council (Gino Miele), and the Career Development Awards of the University of Zu¨rich (Markus Glatzel, Erich Brunner).

REFERENCES Adachi, O., Kawai, T., Takeda, K., Matsumoto, M., Tsutsui, H., Sakagami, M., Nakanishi, K., and Akira, S. (1998). Targeted disruption of the MyD88 gene results in loss of IL-1- and IL-18mediated function. Immunity 9, 143–150. Aguzzi, A. (1996). Between cows and monkeys. Nature 381, 734. Aguzzi, A. (1997a). Neuro-immune connection in spread of prions in the body? The Lancet 349, 742–743. Aguzzi, A. (1997b). Prions and antiprions. Biol. Chem. 378, 1393–1395. Aguzzi, A. (2000). Prion diseases, blood and the immune system: concerns and reality. Haematologica 85, 3–10. Aguzzi, A., Brandner, S., Fischer, M. B., Furukawa, H., Glatzel, M., Hawkins, C., Heppner, F. L., Montrasio, F., Navarro, B., Parizek, P., Pekarik, V., Prinz, M., Raeber, A. J., Rockl, C., and Klein, M. A. (2001a). Spongiform encephalopathies: insights from transgenic models. Adv. Virus Res. 56, 313–352. Aguzzi, A., and Collinge, J. (1997). Post-exposure prophylaxis after accidental prion inoculation. Lancet 350, 1519–1520. Aguzzi, A., Glatzel, M., Montrasio, F., Prinz, M., and Heppner, F. L. (2001b). Interventional strategies against prion diseases. Nat. Rev. Neurosci. 2, 745–749.


161

Aguzzi, A., and Heppner, F. L. (2000). Pathogenesis of prion diseases: a progress report. Cell Death Differ. 7, 889–902. Aguzzi, A., Klein, M. A., Montrasio, F., Pekarik, V., Brandner, S., Furukawa, H., Kaser, P., Rockl, C., and Glatzel, M. (2000). Prions: pathogenesis and reverse genetics. Ann. N.Y. Acad. Sci. 920, 140–157. Aguzzi, A., Montrasio, F., and Kaeser, P. S. (2001c). Prions: health scare and biological challenge. Nat. Rev. Mol. Cell Biol. 2, 118–126. Aguzzi, A., and Raeber, A. J. (1998). Transgenic models of neurodegeneration. Neurodegeneration: of (transgenic) mice and men. Brain Pathol. 8, 695–697. Aguzzi, A., and Weissmann, C. (1996a). Sleepless in Bologna: transmission of fatal familial insomnia. Trends Microbiol. 4, 129–131. Aguzzi, A., and Weissmann, C. (1996b). Spongiform encephalopathies: a suspicious signature. Nature 383, 666–667. Aguzzi, A., and Weissmann, C. (1997). Prion research: the next frontiers. Nature 389, 795–798. Anderson, R. M., Donnelly, C. A., Ferguson, N. M., Woolhouse, M. E., Watt, C. J., Udy, H. J., MaWhinney, S., Dunstan, S. P., Southwood, T. R., Wilesmith, J. W., Ryan, J. B., Hoinville, L. J., Hillerton, J. E., Austin, A. R., and Wells, G. A. (1996). Transmission dynamics and epidemiology of BSE in British cattle. Nature 382, 779–788. Aucouturier, P., Geissmann, F., Damotte, D., Saborio, G. P., Meeker, H. C., Kascsak, R., Carp, R. I., and Wisniewski, T. (2001). Infected splenic dendritic cells are sufficient for prion transmission to the CNS in mouse scrapie. J. Clin. Invest. 108, 703–708. Baldauf, E., Beekes, M., and Diringer, H. (1997). Evidence for an alternative direct route of access for the scrapie agent to the brain bypassing the spinal cord. J. Gen. Virol. 78, 1187–1197. Basler, K., Oesch, B., Scott, M., Westaway, D., Walchli, M., Groth, D. F., McKinley, M. P., Prusiner, S. B., and Weissmann, C. (1986). Scrapie and cellular PrP isoforms are encoded by the same chromosomal gene. Cell 46, 417–428. Beekes, M., McBride, P. A., and Baldauf, E. (1998). Cerebral targeting indicates vagal spread of infection in hamsters fed with scrapie. J. Gen. Virol. 79 Part 3, 601–607. Behrens, A., and Aguzzi, A. (2002). Small is not beautiful: antagonizing functions for the prion protein PrP(C) and its homologue Dpl. Trends Neurosci. 25, 150–154. Behrens, A., Brandner, S., Genoud, N., and Aguzzi, A. (2001). Normal neurogenesis and scrapie pathogenesis in neural grafts lacking the prion protein homologue Doppel. EMBO Rep. 2, 347–352. Behrens, A., Genoud, N., Naumann, H., Rulicke, T., Janett, F., Heppner, F. L., Ledermann, B., and Aguzzi, A. (2002). Absence of the prion protein homologue Doppel causes male sterility. EMBO J. 21, 3652–3658. Beringue, V., Demoy, M., Lasmezas, C. I., Gouritin, B., Weingarten, C., Deslys, J. P., Andreux, J. P., Couvreur, P., and Dormont, D. (2000). Role of spleen macrophages in the clearance of scrapie agent early in pathogenesis. J. Pathol. 190, 495–502. Betmouni, S., and Perry, V. H. (1999). The acute inflammatory response in CNS following injection of prion brain homogenate or normal brain homogenate [In Process Citation]. Neuropathol. Appl. Neurobiol. 25, 20–28. Betmouni, S., Perry, V. H., and Gordon, J. L. (1996). Evidence for an early inflammatory response in the central nervous system of mice with scrapie. Neuroscience 74, 1–5. Bla¨ttler, T., Brandner, S., Raeber, A. J., Klein, M. A., Voigtla¨nder, T., Weissmann, C., and Aguzzi, A. (1997). PrP-expressing tissue required for transfer of scrapie infectivity from spleen to brain. Nature 389, 69–73. Bofill, M., Akbar, A. N., and Amlot, P. L. (2000). Follicular dendritic cells share a membranebound protein with fibroblasts. J. Pathol. 191, 217–226.

162

A. AGUZZI

Bradley, R. (2000). Veterinary research at the Central Veterinary Laboratory, Weybridge, with special reference to scrapie and bovine spongiform encephalopathy. Rev. Sci. Tech. 19, 819–830. Brandner, S., Isenmann, S., Kuhne, G., and Aguzzi, A. (1998). Identification of the end stage of scrapie using infected neural grafts. Brain Pathol. 8, 19–27. Brandner, S., Isenmann, S., Raeber, A., Fischer, M., Sailer, A., Kobayashi, Y., Marino, S., Weissmann, C., and Aguzzi, A. (1996a). Normal host prion protein necessary for scrapieinduced neurotoxicity. Nature 379, 339–343. Brandner, S., Raeber, A., Sailer, A., Blattler, T., Fischer, M., Weissmann, C., and Aguzzi, A. (1996b). Normal host prion protein (PrPC) is required for scrapie spread within the central nervous system. Proc. Natl. Acad. Sci. USA 93, 13148–13151. Brown, D. R., Clive, C., and Haswell, S. J. (2001). Antioxidant activity related to copper binding of native prion protein. J. Neurochem. 76, 69–76. Brown, D. R., Qin, K., Herms, J. W., Madlung, A., Manson, J., Strome, R., Fraser, P. E., Kruck, T., von Bohlen, A., Schulz-Schaeffer, W., Giese, A., Westaway, D., and Kretzschmar, H. (1997). The cellular prion protein binds copper in vivo. Nature 390, 684–687. Brown, K. L., Stewart, K., Ritchie, D. L., Mabbott, N. A., Williams, A., Fraser, H., Morrison, W. I., and Bruce, M. E. (1999). Scrapie replication in lymphoid tissues depends on prion proteinexpressing follicular dendritic cells. Nat. Med. 5, 1308–1312. Brown, P. M., Brandel, J. P., Sato, T., McShane, L., Zerr, I., Fletcher, A., Will, R. G., Pocchiari, M., Cashman, N. R., Aignaux, J. H., Cervenakova, L., Fradkin, J., Schonberger, L. B., and Collins, S. J. (2000). Iatrogenic Creutzfeldt-Jakob disease at the millennium. Neurology 55, 1075–1081. Browning, J. L., Dougas, I., Ngam-ek, A., Bourdon, P. R., Ehrenfels, B. N., Miatkowski, K., Zafari, M., Yampaglia, A. M., Lawton, P., Meier, W., et al. (1995). Characterization of surface lymphotoxin forms. Use of specific monoclonal antibodies and soluble receptors. J. Immunol. 154, 33–46. Browning, J. L., Ngam-ek, A., Lawton, P., DeMarinis, J., Tizard, R., Chow, E. P., Hession, C., O’Brine-Greco, B., Foley, S. F., and Ware, C. F. (1993). Lymphotoxin beta, a novel member of the TNF family that forms a heteromeric complex with lymphotoxin on the cell surface. Cell 72, 847–856. Bruce, M. E. (1985). Agent replication dynamics in a long incubation period model of mouse scrapie. J. Gen. Virol. 66, 2517–2522. Bruce, M. E., Boyle, A., Cousens, S., McConnell, I., Foster, J., Goldmann, W., and Fraser, H. (2002). Strain characterization of natural sheep scrapie and comparison with BSE. J. Gen. Virol. 83, 695–704. Bruce, M. E., Will, R. G., Ironside, J. W., McConnell, I., Drummond, D., Suttie, A., McCardle, L., Chree, A., Hope, J., Birkett, C., Cousens, S., Fraser, H., and Bostock, C. J. (1997). Transmissions to mice indicate that ‘‘new variant’’ CJD is caused by the BSE agent [see comments]. Nature 389, 498–501. Buëler, H. R., Aguzzi, A., Sailer, A., Greiner, R. A., Autenried, P., Aguet, M., and Weissmann, C. (1993). Mice devoid of PrP are resistant to scrapie. Cell 73, 1339–1347. Buëler, H. R., Fischer, M., Lang, Y., Bluethmann, H., Lipp, H. P., DeArmond, S. J., Prusiner, S. B., Aguet, M., and Weissmann, C. (1992). Normal development and behaviour of mice lacking the neuronal cell-surface PrP protein. Nature 356, 577–582. Carp, R. I. (1982). Transmission of scrapie by oral route: effect of gingival scarification [letter]. Lancet 1, 170–171. Carroll, M. C. (1998). CD21/CD35 in B cell activation. Semin. Immunol. 10, 279–286. Cashman, N. R., Loertscher, R., Nalbantoglu, J., Shaw, I., Kascsak, R. J., Bolton, D. C., and Bendheim, P. E. (1990). Cellular isoform of the scrapie agent protein participates in lymphocyte activation. Cell 61, 185–192.


163

Caughey, B., and Race, R. E. (1992). Potent inhibition of scrapie-associated PrP accumulation by congo red. J. Neurochem. 59, 768–771. Caughey, B., and Raymond, G. J. (1993). Sulfated polyanion inhibition of scrapie-associated PrP accumulation in cultured cells. J. Virol. 67, 643–650. Chazot, G., Broussolle, E., Lapras, C., Bla¨ttler, T., Aguzzi, A., and Kopp, N. (1996). New variant of Creutzfeldt-Jakob disease in a 26-year-old French man [letter]. Lancet 347, 1181. Chesebro, B. (1998). BSE and prions: uncertainties about the agent. Science 279, 42–43. Clarke, M. C., and Kimberlin, R. H. (1984). Pathogenesis of mouse scrapie: distribution of agent in the pulp and stroma of infected spleens. Vet. Microbiol. 9, 215–225. Cole, S., and Kimberlin, R. H. (1985). Pathogenesis of mouse scrapie: dynamics of vacuolation in brain and spinal cord after intraperitoneal infection. Neuropathol. Appl. Neurobiol. 11, 213–227. Collinge, J., and Palmer, M. S. (1994). Human prion diseases. Baillie`res. Clin. Neurol. 3, 241–247. Collinge, J., Sidle, K. C., Meads, J., Ironside, J., and Hill, A. F. (1996). Molecular analysis of prion strain variation and the aetiology of ‘‘new variant’’ CJD. Nature 383, 685–690. Collinge, J., Whittington, M. A., Sidle, K. C., Smith, C. J., Palmer, M. S., Clarke, A. R., and Jefferys, J. G. (1994). Prion protein is necessary for normal synaptic function. Nature 370, 295–297. Collins, S. J., Lewis, V., Brazier, M., Hill, A. F., Fletcher, A., and Masters, C. L. (2002). Quinacrine does not prolong survival in a murine Creutzfeldt-Jakob disease model. Ann. Neurol. 52, 503–506. Crowe, P. D., VanArsdale, T. L., Walter, B. N., Ware, C. F., Hession, C., Ehrenfels, B., Browning, J. L., Din, W. S., Goodwin, R. G., and Smith, C. A. (1994). A lymphotoxin-betaspecific receptor. Science 264, 707–710. Cuille, J., and Chelle, P. L. (1939). Experimental transmission of trembling to the goat. C. R. Seances Acad. Sci. 208, 1058–1160. Duffy, P., Wolf, J., Collins, G., DeVoe, A. G., Streeten, B., and Cowen, D. (1974). Possible personto-person transmission of Creutzfeldt-Jakob disease. N. Engl. J. Med. 290, 692–693. Endres, R., Alimzhanov, M. B., Plitz, T., Futterer, A., Kosco-Vilbois, M. H., Nedospasov, S. A., Rajewsky, K., and Pfeffer, K. (1999). Mature follicular dendritic cell networks depend on expression of lymphotoxin beta receptor by radioresistant stromal cells and of lymphotoxin beta and tumor necrosis factor by B cells. J. Exp. Med. 189, 159–168. Farquhar, C., Dickinson, A., and Bruce, M. (1999). Prophylactic potential of pentosan polysulphate in transmissible spongiform encephalopathies [letter]. Lancet 353, 117. Farquhar, C. F., Dornan, J., Somerville, R. A., Tunstall, A. M., and Hope, J. (1994). Effect of Sinc genotype, agent isolate and route of infection on the accumulation of protease-resistant PrP in non-central nervous system tissues during the development of murine scrapie. J. Gen. Virol. 75, 495–504. Felten, D. L., and Felten, S. Y. (1988). Sympathetic noradrenergic innervation of immune organs. Brain. Behav. Immun. 2, 293–300. Fischer, M., Ru¨licke, T., Raeber, A., Sailer, A., Moser, M., Oesch, B., Brandner, S., Aguzzi, A., and Weissmann, C. (1996). Prion protein (PrP) with amino-proximal deletions restoring susceptibility of PrP knockout mice to scrapie. EMBO J. 15, 1255–1264. Fischer, M. B., Goerg, S., Shen, L., Prodeus, A. P., Goodnow, C. C., Kelsoe, G., and Carroll, M. C. (1998). Dependence of germinal center B cells on expression of CD21/CD35 for survival. Science 280, 582–585. Fischer, M. B., Roeckl, C., Parizek, P., Schwarz, H. P., and Aguzzi, A. (2000). Binding of diseaseassociated prion protein to plasminogen. Nature 408, 479–483. Flechsig, E., Shmerling, D., Hegyi, I., Raeber, A. J., Fischer, M., Cozzio, A., von Mering, C., Aguzzi, A., and Weissmann, C. (2000). Prion protein devoid of the octapeptide repeat region restores susceptibility to scrapie in PrP knockout mice. Neuron 27, 399–408.

164

A. AGUZZI

Fournier, J. G., Escaig Haye, F., Billette de Villemeur, T., and Robain, O. (1995). Ultrastructural localization of cellular prion protein (PrPc) in synaptic boutons of normal hamster, hippocampus. C. R. Acad. Sci. III. 318, 339–344. Fraser, H. (1982). Neuronal spread of scrapie agent and targeting of lesions within the retinotectal pathway. Nature 295, 149–150. Frigg, R., Klein, M. A., Hegyi, I., Zinkernagel, R. M., and Aguzzi, A. (1999). Scrapie pathogenesis in subclinically infected B-cell-deficient mice. J. Virol. 73, 9584–9588. Fu, Y. X., Huang, G., Matsumoto, M., Molina, H., and Chaplin, D. D. (1997). Independent signals regulate development of primary and secondary follicle structure in spleen and mesenteric lymph node. Proc. Natl. Acad. Sci. USA 94, 5739–5743. Gabizon, R., McKinley, M. P., Groth, D., and Prusiner, S. B. (1988). Immunoaffinity purification and neutralization of scrapie prion infectivity. Proc. Natl. Acad. Sci. USA 85, 6617–6621. Gajdusek, D. C., Gibbs, C. J., and Alpers, M. (1966). Experimental transmission of a Kuru-like syndrome to chimpanzees. Nature 209, 794–796. Gibbs, C. J., Jr., Gajdusek, D. C., Asher, D. M., Alpers, M. P., Beck, E., Daniel, P. M., and Matthews, W. B. (1968). Creutzfeldt-Jakob disease (spongiform encephalopathy): transmission to the chimpanzee. Science 161, 388–389. Glatzel, M., and Aguzzi, A. (2000). PrP(C) expression in the peripheral nervous system is a determinant of prion neuroinvasion. J. Gen. Virol. 81, 2813–2821. Glatzel, M., and Aguzzi, A. (2001). The shifting biology of prions. Brain Res. Rev. 36, 241–248. Glatzel, M., Flechsig, E., Navarro, B., Klein, M. A., Paterna, J. C., Bueler, H., and Aguzzi, A. (2000). Adenoviral and adeno-associated viral transfer of genes to the peripheral nervous system. Proc. Natl. Acad. Sci. USA 97, 442–447. Glatzel, M., Heppner, F. L., Albers, K. M., and Aguzzi, A. (2001). Sympathetic innervation of lymphoreticular organs is rate limiting for prion neuroinvasion. Neuron 31, 25–34. Glatzel, M., Rogivue, C., Ghani, A., Streffer, J., Amsler, L., and Aguzzi, A. (2002). Incidence of Creutzfeldt-Jakob disease in Switzerland. Lancet 360, 139–141. Gonzalez, M., Mackay, F., Browning, J. L., Kosco-Vilbois, M. H., and Noelle, R. J. (1998). The sequential role of lymphotoxin and B cells in the development of splenic follicles. J. Exp. Med. 187, 997–1007. Griffith, J. S. (1967). Self-replication and scrapie. Nature 215, 1043–1044. Hadlow, W. J. (1959). Scrapie and kuru. Lancet 2, 289–290. Hainfellner, J. A., and Budka, H. (1999). Disease associated prion protein may deposit in the peripheral nervous system in human transmissible spongiform encephalopathies. Acta Neuropathol. (Berl.) 98, 458–460. Hazenbos, W. L., Gessner, J. E., Hofhuis, F. M., Kuipers, H., Meyer, D., Heijnen, I. A., Schmidt, R. E., Sandor, M., Capel, P. J., Daeron, M., van de Winkel, J. G., and Verbeek, J. S. (1996). Impaired IgG-dependent anaphylaxis and Arthus reaction in Fc gamma RIII (CD16) deficient mice. Immunity 5, 181–188. Heppner, F. L., Christ, A. D., Klein, M. A., Prinz, M., Fried, M., Kraehenbuhl, J. P., and Aguzzi, A. (2001a). Transepithelial prion transport by M cells. Nat. Med. 7, 976–977. Heppner, F. L., Musahl, C., Arrighi, I., Klein, M. A., Rulicke, T., Oesch, B., Zinkernagel, R. M., Kalinke, U., and Aguzzi, A. (2001b). Prevention of scrapie pathogenesis by transgenic expression of anti-prion protein antibodies. Science 294, 178–182. Hill, A. F., Desbruslais, M., Joiner, S., Sidle, K. C., Gowland, I., Collinge, J., Doey, L. J., and Lantos, P. (1997a). The same prion strain causes vCJD and BSE [letter] [see comments]. Nature 389, 448–450. Hill, A. F., Zeidler, M., Ironside, J., and Collinge, J. (1997b). Diagnosis of new variant CreutzfeldtJakob disease by tonsil biopsy. Lancet 349, 99.


165

Horiuchi, M., and Caughey, B. (1999). Specific binding of normal prion protein to the scrapie form via a localized domain initiates its conversion to the protease-resistant state [In Process Citation]. EMBO J. 18, 3193–3203. Houston, F., Foster, J. D., Chong, A., Hunter, N., and Bostock, C. J. (2000). Transmission of BSE by blood transfusion in sheep. Lancet 356, 999–1000. http://www.doh.gov.uk/cjd/stats/aug02.htm (2002) Monthly Creutzfeldt-Jakob disease statistics. In. Department of Health. Huang, F. P., Farquhar, C. F., Mabbott, N. A., Bruce, M. E., and MacPherson, G. G. (2002). Migrating intestinal dendritic cells transport PrP(Sc) from the gut. J. Gen. Virol. 83, 267–271. Hunter, N., Foster, J., Chong, A., McCutcheon, S., Parnham, D., Eaton, S., MacKenzie, C., and Houston, F. (2002). Transmission of prion diseases by blood transfusion. J. Gen. Virol. 83, 2897–2905. Jeffrey, M., Goodsir, C. M., Bruce, M., McBride, P. A., Scott, J. R., and Halliday, W. G. (1994). Correlative light and electron microscopy studies of PrP localisation in 87V scrapie. Brain Res. 656, 329–343. Jeffrey, M., McGovern, G., Goodsir, C. M., Brown, K. L., and Bruce, M. E. (2000) Sites of prion protein accumulation in scrapie-infected mouse spleen revealed by immuno-electron microscopy [In Process Citation]. J. Pathol. 191, 323–332. Kaeser, P. S., Klein, M. A., Schwarz, P., and Aguzzi, A. (2001). Efficient lymphoreticular prion propagation requires prp(c) in stromal and hematopoietic cells. J. Virol. 75, 7097–7106. Kao, R. R., Gravenor, M. B., Baylis, M., Bostock, C. J., Chihota, C. M., Evans, J. C., Goldmann, W., Smith, A. J., and McLean, A. R. (2002). The potential size and duration of an epidemic of bovine spongiform encephalopathy in British sheep. Science 295, 332–335. Kapasi, Z. F., Qin, D., Kerr, W. G., Kosco-Vilbois, M. H., Shultz, L. D., Tew, J. G., and Szakal, A. K. (1998). Follicular dendritic cell (FDC) precursors in primary lymphoid tissues. J. Immunol. 160, 1078–1084. Kellings, K., Meyer, N., Mirenda, C., Prusiner, S. B., and Riesner, D. (1993). Analysis of nucleic acids in purified scrapie prion preparations. Arch. Virol. Suppl. 7, 215–225. Kerneis, S., Bogdanova, A., Kraehenbuhl, J. P., and Pringault, E. (1997). Conversion by Peyer’s patch lymphocytes of human enterocytes into M cells that transport bacteria. Science 277, 949–952. Kimberlin, R. H., Hall, S. M., and Walker, C. A. (1983). Pathogenesis of mouse scrapie. Evidence for direct neural spread of infection to the CNS after injection of sciatic nerve. J. Neurol. Sci. 61, 315–325. Kimberlin, R. H., and Walker, C. A. (1978). Pathogenesis of mouse scrapie: effect of route of inoculation on infectivity titres and dose-response curves. J. Comp. Pathol. 88, 39–47. Kimberlin, R. H., and Wilesmith, J. W. (1994). Bovine spongiform encephalopathy. Epidemiology, low dose exposure and risks. Ann. N. Y. Acad. Sci. 724, 210–220. Kitamoto, T., Muramoto, T., Mohri, S., Doh ura, K., and Tateishi, J. (1991). Abnormal isoform of prion protein accumulates in follicular dendritic cells in mice with Creutzfeldt-Jakob disease. J. Virol. 65, 6292–6295. Kitamura, D., Roes, J., Kuhn, R., and Rajewsky, K. (1991). A B cell-deficient mouse by targeted disruption of the membrane exon of the immunoglobulin mu chain gene. Nature 350, 423–426. Klein, M. A., Frigg, R., Flechsig, E., Raeber, A. J., Kalinke, U., Bluethmann, H., Bootz, F., Suter, M., Zinkernagel, R. M., and Aguzzi, A. (1997). A crucial role for B cells in neuroinvasive scrapie. Nature 390, 687–690. Klein, M. A., Frigg, R., Raeber, A. J., Flechsig, E., Hegyi, I., Zinkernagel, R. M., Weissmann, C., and Aguzzi, A. (1998). PrP expression in B lymphocytes is not required for prion neuroinvasion. Nat. Med. 4, 1429–1433.

166

A. AGUZZI

Klein, M. A., Kaeser, P. S., Schwarz, P., Weyd, H., Xenarios, I., Zinkernagel, R. M., Carroll, M. C., Verbeek, J. S., Botto, M., Walport, M. J., Molina, H., Kalinke, U., Acha-Orbea, H., and Aguzzi, A. (2001). Complement facilitates early prion pathogenesis. Nat. Med. 7, 488–492. Koller, M. F., Grau, T., and Christen, P. (2002). Induction of antibodies against murine full-length prion protein in wild-type mice. J. Neuroimmunol. 132, 113–116. Koni, P. A., Sacca, R., Lawton, P., Browning, J. L., Ruddle, N. H., and Flavell, R. A. (1997). Distinct roles in lymphoid organogenesis for lymphotoxins alpha and beta revealed in lymphotoxin beta-deficient mice. Immunity 6, 491–500. Kooyman, D. L., Byrne, G. W., and Logan, J. S. (1998). Glycosyl phosphatidylinositol anchor. Exp. Nephrol. 6, 148–151. Kooyman, D. L., Byrne, G. W., McClellan, S., Nielsen, D., Tone, M., Waldmann, H., Coffman, T. M., McCurry, K. R., Platt, J. L., and Logan, J. S. (1995). In vivo transfer of GPIlinked complement restriction factors from erythrocytes to the endothelium. Science 269, 89–92. Korth, C., Stierli, B., Streit, P., Moser, M., Schaller, O., Fischer, R., Schulz-Schaeffer, W., Kretzschmar, H., Raeber, A., Braun, U., Ehrensperger, F., Hornemann, S., Glockshuber, R., Riek, R., Billeter, M., Wuthrich, K., and Oesch, B. (1997). Prion (PrPSc)-specific epitope defined by a monoclonal antibody. Nature 390, 74–77. Ku¨nzi, V., Glatzel, M., Nakano, M.Y., Greber, U.F., Van Leuven, F., and Aguzzi, A. (2002) Unhampered prion neuroinvasion despite impaired fast axonal transport in transgenic mice overexpressing four-repeat tau. J. Neurosci. 22, 7471–7477. Lasmezas, C. I., Cesbron, J. Y., Deslys, J. P., Demaimay, R., Adjou, K. T., Rioux, R., Lemaire, C., Locht, C., and Dormont, D. (1996a). Immune system-dependent and -independent replication of the scrapie agent. J. Virol. 70, 1292–1295. Lasmezas, C. I., Deslys, J. P., Demaimay, R., Adjou, K. T., Lamoury, F., Dormont, D., Robain, O., Ironside, J., and Hauw, J. J. (1996b). Bse Transmission to Macaques. Nature 381, 743–744. Li, A., Sakaguchi, S., Shigematsu, K., Atarashi, R., Roy, B. C., Nakaoke, R., Arima, K., Okimura, N., Kopacek, J., and Katamine, S. (2000). Physiological expression of the gene for PrP-like protein, PrPLP/Dpl, by brain endothelial cells and its ectopic expression in neurons of PrP- deficient mice ataxic due to purkinje cell degeneration [In Process Citation]. Am. J. Pathol. 157, 1447–1452. Liu, T., Li, R., Pan, T., Liu, D., Petersen, R. B., Wong, B. S., Gambetti, P., and Sy, M. S. (2002) Intercellular transfer of the cellular prion protein. J. Biol. Chem. Lledo, P. M., Tremblay, P., Dearmond, S. J., Prusiner, S. B., and Nicoll, R. A. (1996). Mice deficient for prion protein exhibit normal neuronal excitability and synaptic transmission in the hippocampus. Proc. Natl. Acad. Sci. USA 93, 2403–2407. Lu, K., Wang, W., Xie, Z., Wong, B. S., Li, R., Petersen, R. B., Sy, M. S., and Chen, S. G. (2000). Expression and structural characterization of the recombinant human doppel protein(,) [In Process Citation]. Biochemistry 39, 13575–13583. Ma, J., and Lindquist, S. (2002) Conversion of PrP to a Self-Perpetuating PrPSc-like Conformation in the Cytosol. Science. Ma, J., Wollmann, R. and Lindquist, S. (2002) Neurotoxicity and Neurodegeneration. When PrP Accumulates in the Cytosol. Science 298, 1781–1785. Mabbott, N. A., Bruce, M. E., Botto, M., Walport, M. J., and Pepys, M. B. (2001). Temporary depletion of complement component C3 or genetic deficiency of C1q significantly delays onset of scrapie. Nat. Med. 7, 485–487. Mackay, F., and Browning, J. L. (1998). Turning off follicular dendritic cells. Nature 395, 26–27. Mackay, F., Majeau, G. R., Lawton, P., Hochman, P. S., and Browning, J. L. (1997). Lymphotoxin but not tumor necrosis factor functions to maintain splenic architecture and humoral responsiveness in adult mice. Eur. J. Immunol. 27, 2033–2042.


167

Maissen, M., Roeckl, C., Glatzel, M., Goldmann, W., and Aguzzi, A. (2001). Plasminogen binds to disease-associated prion protein of multiple species. Lancet 357, 2026–2028. Mallucci, G. R., Ratte, S., Asante, E. A., Linehan, J., Gowland, I., Jefferys, J. G., and Collinge, J. (2002). Post-natal knockout of prion protein alters hippocampal CA1 properties, but does not result in neurodegeneration. EMBO J. 21, 202–210. Manson, J. C., Clarke, A. R., Hooper, M. L., Aitchison, L., McConnell, I., and Hope, J. (1994). 129/Ola mice carrying a null mutation in PrP that abolishes mRNA production are developmentally normal. Mol. Neurobiol. 8, 121–127. Matsumoto, M., Fu, Y. X., Molina, H., Huang, G., Kim, J., Thomas, D. A., Nahm, M. H., and Chaplin, D. D. (1997). Distinct roles of lymphotoxin alpha and the type I tumor necrosis factor (TNF) receptor in the establishment of follicular dendritic cells from non-bone marrow-derived cells. J. Exp. Med. 186, 1997–2004. McBride, P. A., and Beekes, M. (1999). Pathological PrP is abundant in sympathetic and sensory ganglia of hamsters fed with scrapie. Neurosci. Lett. 265, 135–138. Mead, S., Beck, J., Dickinson, A., Fisher, E. M., and Collinge, J. (2000). Examination of the human prion protein-like gene doppel for genetic susceptibility to sporadic and variant Creutzfeldt-Jakob disease. Neurosci. Lett. 290, 117–120. Mo, H., Moore, R. C., Cohen, F. E., Westaway, D., Prusiner, S. B., Wright, P. E., and Dyson, H. J. (2001). Two different neurodegenerative diseases caused by proteins with similar structures. Proc. Natl. Acad. Sci. USA 98, 2352–2357. Montrasio, F., Frigg, R., Glatzel, M., Klein, M. A., Mackay, F., Aguzzi, A., and Weissmann, C. (2000). Impaired prion replication in spleens of mice lacking functional follicular dendritic cells. Science 288, 1257–1259. Moore, R. C., Lee, I. Y., Silverman, G. L., Harrison, P. M., Strome, R., Heinrich, C., Karunaratne, A., Pasternak, S. H., Chishti, M. A., Liang, Y., Mastrangelo, P., Wang, K., Smit, A. F., Katamine, S., Carlson, G. A., Cohen, F. E., Prusiner, S. B., Melton, D. W., Tremblay, P., Hood, L. E., and Westaway, D. (1999). Ataxia in prion protein (PrP)-deficient mice is associated with upregulation of the novel PrP-like protein doppel [In Process Citation]. J. Mol. Biol. 292, 797–817. Moser, M., Oesch, B., and Bueler, H. (1993). An anti-prion protein? Nature 362, 213–214. Neutra, M. R., Frey, A., and Kraehenbuhl, J. P. (1996). Epithelial M cells: gateways for mucosal infection and immunization. Cell 86, 345–348. Nicotera, P. (2001). A route for prion neuroinvasion. Neuron 31, 345–348. Oesch, B., Teplow, D. B., Stahl, N., Serban, D., Hood, L. E., and Prusiner, S. B. (1990). Identification of cellular proteins binding to the scrapie prion protein. Biochemistry 29, 5848–5855. Oesch, B., Westaway, D., Walchli, M., McKinley, M. P., Kent, S. B., Aebersold, R., Barry, R. A., Tempst, P., Teplow, D. B., Hood, L. E., and Weissmann, C. (1985). A cellular gene encodes scrapie PrP 27-30 protein. Cell 40, 735–746. Park, S. Y., Ueda, S., Ohno, H., Hamano, Y., Tanaka, M., Shiratori, T., Yamazaki, T., Arase, H., Arase, N., Karasawa, A., Sato, S., Ledermann, B., Kondo, Y., Okumura, K., Ra, C., and Saito, T. (1998). Resistance of Fc receptor-deficient mice to fatal glomerulonephritis. J. Clin. Invest. 102, 1229–1238. Peoc’h, K., Guerin, C., Brandel, J. P., Launay, J. M., and Laplanche, J. L. (2000). First report of polymorphisms in the prion-like protein gene (PRND): implications for human prion diseases. Neurosci. Lett. 286, 144–148. Peretz, D., Williamson, R. A., Kaneko, K., Vergara, J., Leclerc, E., Schmitt-Ulms, G., Mehlhorn, I. R., Legname, G., Wormald, M. R., Rudd, P. M., Dwek, R. A., Burton, D. R., and Prusiner, S. B. (2001). Antibodies inhibit prion propagation and clear cell cultures of prion infectivity. Nature 412, 739–743.

168

A. AGUZZI

Perry, V. H., Cunningham, C., and Boche, D. (2002). Atypical inflammation in the central nervous system in prion disease. Curr. Opin. Neurol. 15, 349–354. Pocchiari, M., Schmittinger, S., and Masullo, C. (1987). Amphotericin B delays the incubation period of scrapie in intracerebrally inoculated hamsters. J. Gen. Virol. 68, 219–223. Prinz, M., Huber, G., Macpherson, A.J.S., Heppner, F.L., Glatzel, M., Eugster, H., Wagner, N., and Aguzzi, A. (2003) Oral prion infection requires normal numbers of Peyer’s patches but not of enteric lymphocytes. Am. J. Pathohol. 162, 1103–1111. Prinz, M., Montrasio, F., Klein, M. A., Schwarz, P., Priller, J., Odermatt, B., Pfeffer, K., and Aguzzi, A. (2002). Lymph nodal prion replication and neuroinvasion in mice devoid of follicular dendritic cells. Proc. Natl. Acad. Sci. USA 99, 919–924. Priola, S. A., Raines, A., and Caughey, W. S. (2000). Porphyrin and phthalocyanine antiscrapie compounds [see comments]. Science 287, 1503–1506. Prusiner, S. B. (1982). Novel proteinaceous infectious particles cause scrapie. Science 216, 136–144. Prusiner, S. B. (1989). Scrapie prions. Annu. Rev. Microbiol. 43, 345–374. Prusiner, S. B., Cochran, S. P., Groth, D. F., Downey, D. E., Bowman, K. A., and Martinez, H. M. (1982). Measurement of the scrapie agent using an incubation time interval assay. Ann. Neurol. 11, 353–358. Prusiner, S. B., Groth, D., Serban, A., Koehler, R., Foster, D., Torchia, M., Burton, D., Yang, S. L., and DeArmond, S. J. (1993). Ablation of the prion protein (PrP) gene in mice prevents scrapie and facilitates production of anti-PrP antibodies. Proc. Natl. Acad. Sci. USA 90, 10608– 10612. Prusiner, S. B., Scott, M. R., DeArmond, S. J., and Cohen, F. E. (1998). Prion protein biology. Cell 93, 337–348. Race, R., Oldstone, M., and Chesebro, B. (2000). Entry versus blockade of brain infection following oral or intraperitoneal scrapie administration: role of prion protein expression in peripheral nerves and spleen. J. Virol. 74, 828–833. Race, R. E., Priola, S. A., Bessen, R. A., Ernst, D., Dockter, J., Rall, G. F., Mucke, L., Chesebro, B., and Oldstone, M. B. (1995). Neuron-specific expression of a hamster prion protein minigene in transgenic mice induces susceptibility to hamster scrapie agent. Neuron 15, 1183–1191. Raeber, A. J., Klein, M. A., Frigg, R., Flechsig, E., Aguzzi, A., and Weissmann, C. (1999a). PrPdependent association of prions with splenic but not circulating lymphocytes of scrapie-infected mice. EMBO J. 18, 2702–2706. Raeber, A.J., Sailer, A., Hegyi, I., Klein, M.A., Rulicke, T., Fischer, M., Brandner, S., Aguzzi, A., and Weissmann, C. (1999b) Ectopic expression of prion protein (PrP) in T lymphocytes or hepatocytes of PrP knockout mice is insufficient to sustain prion replication. Proc. Natl. Acad. Sci. USA 96, 3987–3992. Rieger, R., Edenhofer, F., Lasmezas, C. I., and Weiss, S. (1997). The human 37-kDa laminin receptor precursor interacts with the prion protein in eukaryotic cells [see comments]. Nat. Med. 3, 1383–1388. Riesner, D., Kellings, K., Wiese, U., Wulfert, M., Mirenda, C., and Prusiner, S. B. (1993). Prions and nucleic acids: search for ‘‘residual’’ nucleic acids and screening for mutations in the PrPgene. Dev. Biol. Stand. 80, 173–181. Rossi, D., Cozzio, A., Flechsig, E., Klein, M. A., Aguzzi, A., and Weissmann, C. (2001). Onset of ataxia and Purkinje cell loss in PrP null mice inversely correlated with Dpl level in brain. EMBO J. 20, 1–9. Rother, K. I., Clay, O. K., Bourquin, J. P., Silke, J., and Schaffner, W. (1997). Long non-stop reading frames on the antisense strand of heat shock protein 70 genes and prion protein (PrP) genes are conserved between species [see comments]. Biol. Chem. 378, 1521–1530.


169

Rubenstein, R., Merz, P. A., Kascsak, R. J., Scalici, C. L., Papini, M. C., Carp, R. I., and Kimberlin, R. H. (1991). Scrapie-infected spleens: analysis of infectivity, scrapie-associated fibrils, and protease-resistant proteins. J. Infect. Dis. 164, 29–35. Safar, J., Cohen, F. E., and Prusiner, S. B. (2000). Quantitative traits of prion strains are enciphered in the conformation of the prion protein. Arch. Virol. Suppl. 16, 227–235. Safar, J., Wille, H., Itri, V., Groth, D., Serban, H., Torchia, M., Cohen, F. E., and Prusiner, S. B. (1998). Eight prion strains have PrP(Sc) molecules with different conformations [see comments]. Nat. Med. 4, 1157–1165. Sailer, A., Buëler, H., Fischer, M., Aguzzi, A., and Weissmann, C. (1994). No propagation of prions in mice devoid of PrP. Cell 77, 967–968. Sakaguchi, S., Katamine, S., Nishida, N., Moriuchi, R., Shigematsu, K., Sugimoto, T., Nakatani, A., Kataoka, Y., Houtani, T., Shirabe, S., Okada, H., Hasegawa, S., Miyamoto, T., and Noda, T. (1996). Loss of cerebellar purkinje cells in aged mice homozygous for a disrupted Prp gene. Nature 380, 528–531. Scott, J. R., Foster, J. D., and Fraser, H. (1993). Conjunctival instillation of scrapie in mice can produce disease. Vet. Microbiol. 34, 305–309. Sethi, S., Lipford, G., Wagner, H., and Kretzschmar, H. (2002). Postexposure prophylaxis against prion disease with a stimulator of innate immunity. Lancet 360, 229–230. Shaked, Y., Engelstein, R., and Gabizon, R. (2002). The binding of prion proteins to serum components is affected by detergent extraction conditions. J. Neurochem. 82, 1–5. Shaked, Y., Rosenmann, H., Talmor, G., and Gabizon, R. (1999). A C-terminal-truncated PrP isoform is present in mature sperm. J. Biol. Chem. 274, 32153–32158. Shlomchik, M. J., Radebold, K., Duclos, N., and Manuelidis, L. (2001). Neuroinvasion by a Creutzfeldt-Jakob disease agent in the absence of B cells and follicular dendritic cells. Proc. Natl. Acad. Sci. USA 98, 9289–9294. Shmerling, D., Hegyi, I., Fischer, M., Blattler, T., Brandner, S., Gotz, J., Rulicke, T., Flechsig, E., Cozzio, A., von Mering, C., Hangartner, C., Aguzzi, A., and Weissmann, C. (1998). Expression of amino-terminally truncated PrP in the mouse leading to ataxia and specific cerebellar lesions. Cell 93, 203–214. Sigurdsson, E. M., Brown, D. R., Daniels, M., Kascsak, R. J., Kascsak, R., Carp, R., Meeker, H. C., Frangione, B., and Wisniewski, T. (2002). Immunization delays the onset of prion disease in mice. Am. J. Pathol. 161, 13–17. Silverman, G. L., Qin, K., Moore, R. C., Yang, Y., Mastrangelo, P., Tremblay, P., Prusiner, S. B., Cohen, F. E., and Westaway, D. (2000). Doppel is an N-glycosylated, glycosylphosphatidylinositol-anchored protein. Expression in testis and ectopic production in the brains of Prnp(0/0) mice predisposed to Purkinje cell loss. J. Biol. Chem. 275, 26834–26841. Soto, C., Kascsak, R. J., Saborio, G. P., Aucouturier, P., Wisniewski, T., Prelli, F., Kascsak, R., Mendez, E., Harris, D. A., Ironside, J., Tagliavini, F., Carp, R. I., and Frangione, B. (2000). Reversion of prion protein conformational changes by synthetic beta-sheet breaker peptides. Lancet 355, 192–197. Supattapone, S., Bosque, P., Muramoto, T., Wille, H., Aagaard, C., Peretz, D., Nguyen, H. O., Heinrich, C., Torchia, M., Safar, J., Cohen, F. E., DeArmond, S. J., Prusiner, S. B., and Scott, M. (1999). Prion protein of 106 residues creates an artifical transmission barrier for prion replication in transgenic mice. Cell 96, 869–878. Supattapone, S., Wille, H., Uyechi, L., Safar, J., Tremblay, P., Szoka, F. C., Cohen, F. E., Prusiner, S. B., and Scott, M. R. (2001). Branched polyamines cure prion-infected neuroblastoma cells. J. Virol. 75, 3453–3461. Szakal, A. K., and Hanna, M. G., Jr. (1968). The ultrastructure of antigen localization and viruslike particles in mouse spleen germinal centers. Exp. Mol. Pathol. 8, 75–89.

170

A. AGUZZI

Szakal, A. K., Kapasi, Z. F., Haley, S. T., and Tew, J. G. (1995). Multiple lines of evidence favoring a bone marrow derivation of follicular dendritic cells (FDCs). Adv. Exp. Med. Biol. 378, 267–272. Tagliavini, F., McArthur, R. A., Canciani, B., Giaccone, G., Porro, M., Bugiani, M., Lievens, P. M., Bugiani, O., Peri, E., Dall’Ara, P., Rocchi, M., Poli, G., Forloni, G., Bandiera, T., Varasi, M., Suarato, A., Cassutti, P., Cervini, M. A., Lansen, J., Salmona, M., and Post, C. (1997). Effectiveness of anthracycline against experimental prion disease in Syrian hamsters. Science 276, 1119–1122. Takai, T., Li, M., Sylvestre, D., Clynes, R., and Ravetch, J. V. (1994). FcR gamma chain deletion results in pleiotrophic effector cell defects. Cell 76, 519–529. Takai, T., Ono, M., Hikida, M., Ohmori, H., and Ravetch, J. V. (1996). Augmented humoral and anaphylactic responses in Fc gamma RII-deficient mice. Nature 379, 346–349. Taylor, D. M. (2000). Inactivation of transmissible degenerative encephalopathy agents: A review [see comments]. Vet. J. 159, 10–17. Taylor, D. M., McConnell, I., and Fraser, H. (1996). Scrapie infection can be established readily through skin scarification in immunocompetent but not immunodeficient mice. J. Gen. Virol. 77, 1595–1599. Tobler, I., Gaus, S. E., Deboer, T., Achermann, P., Fischer, M., Ru¨licke, T., Moser, M., Oesch, B., McBride, P. A., and Manson, J. C. (1996). Altered circadian activity rhythms and sleep in mice devoid of prion protein. Nature 380, 639–642. Vankeulen, L. J. M., Schreuder, B. E. C., Meloen, R. H., Mooijharkes, G., Vromans, M. E. W., and Langeveld, J. P. M. (1996). Immunohistochemical detection of prion protein in lymphoid tissues of sheep with natural scrapie. ‘‘Journal of Clinical Microbiology’’ 34, 1228–1231. Wadsworth, J. D., Hill, A. F., Joiner, S., Jackson, G. S., Clarke, A. R., and Collinge, J. (1999). Strain-specific prion-protein conformation determined by metal ions [see comments]. Nat. Cell. Biol. 1, 55–59. Waggoner, D. J., Drisaldi, B., Bartnikas, T. B., Casareno, R. L., Prohaska, J. R., Gitlin, J. D., and Harris, D. A. (2000). Brain copper content and cuproenzyme activity do not vary with prion protein expression level. J. Biol. Chem. 275, 7455–7458. Wagner, N., Lohler, J., Kunkel, E. J., Ley, K., Leung, E., Krissansen, G., Rajewsky, K., and Muller, W. (1996). Critical role for beta7 integrins in formation of the gut-associated lymphoid tissue. Nature 382, 366–370. Ware, C. F., VanArsdale, T. L., Crowe, P. D., and Browning, J. L. (1995). The ligands and receptors of the lymphotoxin system. Curr. Top. Microbiol. Immunol. 198, 175–218. Wassarman, P. M., Jovine, L., and Litscher, E. S. (2001). A profile of fertilization in mammals. Nat. Cell. Biol. 3, E59–64. Weissmann, C. (1991). Spongiform encephalopathies. The prion’s progress. Nature 349, 569–571. Weissmann, C., and Aguzzi, A. (1997). Bovine spongiform encephalopathy and early onset variant Creutzfeldt-Jakob disease. Curr. Opin. Neurobiol. 7, 695–700. Weissmann, C., and Aguzzi, A. (1999). Perspectives: neurobiology. PrP’s double causes trouble. Science 286, 914–915. Wells, G. A., Dawson, M., Hawkins, S. A., Green, R. B., Dexter, I., Francis, M. E., Simmons, M. M., Austin, A. R., and Horigan, M. W. (1994). Infectivity in the ileum of cattle challenged orally with bovine spongiform encephalopathy. Vet. Rec. 135, 40–41. Wells, G. A., Hawkins, S. A., Green, R. B., Austin, A. R., Dexter, I., Spencer, Y. I., Chaplin, M. J., Stack, M. J., and Dawson, M. (1998). Preliminary observations on the pathogenesis of experimental bovine spongiform encephalopathy (BSE): an update. Vet. Rec. 142, 103–106. Wells, G. A., Scott, A. C., Johnson, C. T., Gunning, R. F., Hancock, R. D., Jeffrey, M., Dawson, M., and Bradley, R. (1987). A novel progressive spongiform encephalopathy in cattle. Vet. Rec. 121, 419–420.


171

Westaway, D., and Carlson, G. A. (2002). Mammalian prion proteins: enigma, variation and vaccination. Trends. Biochem. Sci. 27, 301–307. Whittington, M. A., Sidle, K. C., Gowland, I., Meads, J., Hill, A. F., Palmer, M. S., Jefferys, J. G., and Collinge, J. (1995). Rescue of neurophysiological phenotype seen in PrP null mice by transgene encoding human prion protein. Nat. Genet. 9, 197–201. Will, R. G., Ironside, J. W., Zeidler, M., Cousens, S. N., Estibeiro, K., Alperovitch, A., Poser, S., Pocchiari, M., Hofman, A., and Smith P. G. (1996). A new variant of Creutzfeldt-Jakob disease in the UK. Lancet 347, 921–925. Wu, X., Jiang, N., Fang, Y. F., Xu, C., Mao, D., Singh, J., Fu, Y. X., and Molina, H. (2000). Impaired affinity maturation in Cr2/ mice is rescued by adjuvants without improvement in germinal center development. J. Immunol. 165, 3119–3127. Yehiely, F., Bamborough, P., Costa, M. D., Perry, B. J., Thinakaran, G., Cohen, F. E., Carlson, G. A., and Prusiner, S. B. (2002). Identification of Candidate Proteins Binding to Prion [Protein. Volume 3, Number 4 (1997), pp. 339–355. Erratum in Neurobiol. Dis. 10, 67–68]. Zanata, S. M., Lopes, M. H., Mercadante, A. F., Hajj, G. N., Chiarini, L. B., Nomizo, R., Freitas, A. R., Cabral, A. L., Lee, K. S., Juliano, M. A., De Oliveira, E., Jachieri, S. G., Burlingame, A., Huang, L., Linden, R., Brentani, R. R., and Martins, V. R. (2002). Stressinducible protein 1 is a cell surface ligand for cellular prion that triggers neuroprotection. EMBO J. 21, 3307–3316.



Roles of the Semaphorin Family in Immune Regulation ATSUSHI KUMANOGOH AND HITOSHI KIKUTANI1 Department of Molecular Immunology, Research Institute for Microbial Diseases, Osaka University, 3-1 Yamada-oka, Suita, Osaka 565-0871, Japan

The immune system and the nervous system have distinct roles in maintaining physiological homeostasis. These independent systems, however, influence each other while sharing common resources, including the cytokines and members of the immunoglobulin superfamily. Semaphorins are one of these shared molecular families that are biologically active in both systems. Although semaphorins were originally identified as axon guidance factors functioning in the nervous system, recent studies have uncovered additional immunological functions. For example, ligand-receptor systems distinct from those characterized in the nervous system govern class IV semaphorin, CD100/Sema4D and Sema4A, activity in immune responses. This review provides an overview of the currently emerging immunoregulatory functions of ‘‘Immuno-semaphorins.’’

I. Overview

A. HISTORY The guidance of axons to appropriate targets is essential for proper assembly of the nervous system. Neurons require specific cellular environments to extend new axons. To respond appropriately, the task of perceiving and integrating the myriad of signals present in the immediate vicinity falls to the growth cone, a sensory and motor apparatus located at the distal tip of a developing axon (Giger and Kolodkin, 2001; Steward, 2002). In the 1980s, in vitro examination of growth cone behavior and axonal navigation suggested the existence of molecules guiding axons by repelling growth cones. In 1993, Luo et al. cloned a gene encoding an avian protein named collapsin that induces the collapse of sensory growth cones (Luo et al., 1993). The identified amino acid sequence of collapsin was closely related to fasciclin IV, a protein implicated in the guidance of peripheral sensory axons in the embryonic grasshopper nervous system (Kolodkin et al., 1992, 1993). The structural conservation between collapsin and fasciclin IV suggested the existence of a novel family of proteins contributing to the guidance of developing axons. The identification of more than 30 members of this gene family, now termed semaphorins, from invertebrates to humans (Table I) also Tel.: þ 81-6-6879-8363; fax: þ 81-6-6875-4465; E-mail: [email protected]

1


174

A. KUMANOGOH AND H. KIKUTANI

TABLE I SEMAPHORIN MEMBERS Class Invertebrates 1 2 5 Rodents/Other Vertebrates 3

Name (original name) Sema-1a (G-Sema I, D-Sema I, T-Sema I, Ce-Sema-I) Sema-1b (Sema 1b) Sema-2a (D-Sema II, Ce-Sema II, gSema II) Sema-5c (D-Sema 5c) Sema3A/SEMA3A (C-Collapsin (Coll-1), H-Sema III, M-SemD, R-Sema III, Sema-Z1a) Sema3B/SEMA3B (M-SemA, H-SemaA, H-Sema V) Sema3C/SEMA3C (M-SemE, C-Coll-3, H-Sema E) Sema3D/SEMA3D (C-Coll-2, Sema-Z2) Sema3E/SEMA3E (C-Coll-5, M-Sema H) Sema3F/SEMA3F (H-Sema IV, M-Sema IV, H-SEMA3F

4

Sema4A/SEMA4A (M-SemB) Sema4B/SEMA4B (M-SemC) Sema4C/SEMA4C (M-Sema F) Sema4D/SEMA4D (CD100, M-Sema G, C-Coll-4) Sema4E/SEMA4E (Sema-Z7) Sema4F/SEMA4F (M-Sema W, R-Sema W, H-Sema W) Sema4G/SEMA4G

5

Sema5A/SEMA5A (M-SemF) Sema5B/SEMA5B (M-SemG)

6

Sema6A/SEMA6A (M-Sema Via) Sema6B/SEMA6B (M-SemaVIb, R-Sema Z) Sema6C/SEMA6C (M-Sema Y, R-Sema Y) Sema6D/SEMA6D

7

Sema7A/SEMA7A (CD108, H-Sema K1, H-Sema L, M-Sema L, M-Sema K1) SEMAVA (A39R, Vaccinia sema, Variola sema) SEMAVB (AHV sema)

Viral

supported this possibility (Semaphorin Nomenclature Committee, 1999). In addition, virus-encoded semaphorins have been found in the genomes of non-neurotropic DNA viruses (Kolodkin et al., 1993). B. STRUCTURE The semaphorin family is comprised of a large number of phylogenetically conserved proteins. Both secreted and transmembrane proteins possess both a sema domain (of approximately 500 amino acids) within the extracellular region and a class-specific C-terminus containing additional sequence motifs, such as immunoglobulin domains, thrombospondin domains, or glycosylphosphatidylinositol (GPI) linkage sites. Based on these structural features,

ROLES OF THE SEMAPHORIN FAMILY IN IMMUNE REGULATION

175

FIG. 1. Semaphorins and their receptors. A schematic representation of semaphorin–receptor interactions (arrows). The semaphorin family contains a large number of phylogenetically conserved secreted and transmembrane proteins. Based on structural features, the members have been divided into eight classes, which includes a unique viral class. The members of the semaphorin family share a common sema domain. The plexin family also possesses a sema domain within the extracellular region. Neuropilins and plexins differ, however, in the structural motifs found in their extracellular domains. Within the immune system, class IV semaphorins, CD100/Sema4D and Sema4A, bind to CD72 and Tim-2, respectively.

176


semaphorins have been divided into eight classes (Fig. 1) (Tessier-Lavigne and Goodman, 1996; Semaphorin Nomenclature Committee, 1999; Yu and Kolodkin, 1999). Class I and II semaphorins are found in invertebrates. The class III–VII semaphorins are vertebrate proteins, whereas class VIII semaphorins are encoded by viruses. Class IV, V, and VI semaphorins are transmembrane proteins, while class III semaphorins are secreted. The class VII semaphorins resemble class III semaphorins, but are bound to the cell surface by a GPI anchor. The structural conservation of the sema domain suggests that this region plays an essential role in the biological signaling activity of semaphorins. Truncated secreted-type class III semaphorins containing only the sema domain are biologically active (Koppel and Raper, 1998). Moreover, the biological specificity of semaphorins is determined by a relatively short stretch of amino acids within the sema domain, as shown for avian semaphorins (collapsins) (Koppel et al., 1997). The carboxyl-terminal domain containing basic amino acids adjacent to an immunoglobulin domain enhances the binding of class III semaphorins to their receptors (Raper, 2000). It remains to be determined, however, if these findings are generally applicable to other semaphorins. C. RECEPTORS Two classes of semaphorin receptors have been identified in the nervous system (Fig. 1): the neuropilins, consisting of Neuropilins-1 and -2, and the plexins, which are subdivided into the four subfamilies of plexin-A1-4, plexinB1-3, plexin-C1, and plexin-D1 (Yu and Kolodkin, 1999). Expression cloning, utilizing a chimera of Sema3A fused to alkaline phosphatase, identified neuropilin-1 as the high-affinity binding receptor for class III semaphorins (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997). While Sema3A binds only to neuropilin-1, Sema3B, Sema3C, and Sema3F bind to both neuropilin-1 and -2. A structure–function analysis revealed that the amino-terminal CUB (complement-binding protein homology) domain of neuropilins functions as a semaphorin binding site (Grunwald and Klein, 2002). Neuropilins, however, have short cytoplasmic tails with no obvious signaling motifs; these domains are dispensable for repulsive semaphorin guidance, suggesting that neuropilins aid in the assembly of a receptor complex including additional transmembrane signaling molecules (Yu and Kolodkin, 1999). Subsequently, a population of plexins were identified as receptors for semaphorins (Comeau et al., 1998; Winberg et al., 1998; Takahashi et al., 1999; Tamagnone et al., 1999; Tamagnone and Comoglio, 2000). Although plexins themselves have a sema domain in their extracellular regions (sharing 17% amino acid identity with semaphorin sema domains), this family is phylogenetically distinct from the semaphorins. Although class III semaphorins do not directly bind plexins, their signals are mediated through a


177

neuropilin–plexin-A1 complex (Takahashi et al., 1999). Sema3A initiates cell collapse upon coexpression of neuropilin-1 and plexin-A1 (Takahashi and Strittmatter, 2001). A portion of non-class III semaphorins, however, binds plexins directly. Drosophila Sema 1a and Sema 1b bind to plexin-A (Winberg et al., 1998; Tamagnone and Comoglio, 2000), while the viral semaphorins A39R/SEMAVA and AHVsema/SEMAVB bind the cell surface virus-encoded semaphorin protein receptor (VESPR) (Comeau et al., 1998), also designated plexin-C1 or CD232. CD108/Sema7A, a human class VII semaphorin homologous to AHVsema, specifically binds plexin-C1/VESPR/CD232 (Tamagnone et al., 1999). A class IV semaphorin, CD100/Sema4D, binds specifically to plexin-B1 (Tamagnone et al., 1999). Of importance in immune responses, the class IV semaphorins CD100 and Sema4A utilize CD72 and Tim-2, respectively, as their functional lymphocyte receptors (Kumanogoh et al., 2000; Kumanogoh et al., 2002a). D. ACTIVITIES

OF

SEMAPHORINS

Several members of the semaphorin family induce axonal growth cone collapse and axon repulsion. Many subsets of neuronal cells respond to semaphorins, including sympathetic, motor, cerebellar, hippocampal, olfactory, corticospinal and dorsal root ganglion neurons (Tessier-Lavigne and Goodman, 1996; Raper, 2000). In addition to the activity as a chemorepulsive axonal guidance factors, semaphorins also function in organogenesis, vascularization, angiogenesis, and cancer progression (Kitsukawa et al., 1995; Behar et al., 1996; Sekido et al., 1996). The molecular cloning of CD100/ Sema4D, a T-cell surface antigen, as a member of the class IV semaphorin subfamily also suggested a role for semaphorins in the immune system (Hall et al., 1996). Several semaphorins, particularly the transmembrane and virusencoded semaphorins, are also implicated in immune responses (Kumanogoh and Kikutani, 2001). While the detailed immunological activities of semaphorins remain to be elucidated, unique functional roles for these molecules in the immune system are emerging slowly. II. Class IV Semaphorin (1): CD100/Sema4D

A. DISCOVERY CD100 was first identified as a 150-kDa surface antigen on the surface of human T-cells (Bougeret et al., 1992). In 1996, molecular cloning revealed CD100 as a transmembrane-type semaphorin family member, the first found to be expressed in the immune system (Hall et al., 1996). Mouse CD100 was independently isolated from mouse embryonic cDNA by PCR cloning using degenerate oligonucleotide primers based on conserved sequences present in

178


the semaphorin family (Furuyama et al., 1996). Mouse CD100 cDNA was also isolated from a mouse B cell line by subtractive cDNA cloning as a CD40inducible gene (Kumanogoh et al., 2000). B. STRUCTURE CD100 contains an amino-terminal signal sequence followed by a sema domain, an Ig-like domain, a lysine-rich stretch, a hydrophobic transmembrane region, and a cytoplasmic tail (Fig. 2) (Furuyama et al., 1996; Hall et al., 1996; Kumanogoh and Kikutani, 2001). According to its structure, CD100, also known as Sema4D, was determined to be a class IV semaphorin (Semaphorin Nomenclature Committee, 1999). Approximately 81% amino acid identity exists between the human and mouse molecules. Expressed at the cell surface as a homodimer, the CD100 extracellular region contains several conserved cysteine residues within the sema domain. Mutational analysis of the conserved cysteine residues of human CD100 has demonstrated that the cysteine at position 679 in the sema domain is required for homodimerization, essential for CD100 biological function (inhibition of immune cell migration) (Delaire et al., 2001). Sequence analysis also reveals several consensus N-glycosylation sites; N-glycosidase treatment reduces the molecular weight of mouse CD100, confirming the physiological presence of these sugar modifications (Kumanogoh et al., 2000). Although there is no catalytic domain within the cytoplasmic domain of CD100, there are consensus sites for tyrosine and serine phosphorylation (Delaire et al., 1998; Kumanogoh and Kikutani, 2001). C. EXPRESSION CD100 mRNA is highly expressed in a broad range of human tissues, including embryonic and adult brain, kidney, and heart (Hall et al., 1996). In mice, CD100 mRNA is detectable throughout the embryonic nervous system, with intense expression in the cortical plate and dorsal root ganglia (Furuyama et al., 1996). The embryonic thymus also exhibits marked expression, while embryonic lung and kidney demonstrate moderate expression. CD100 protein expression is detected on the majority of hematopoietic cells, with the exception of immature bone marrow cells, red blood cells, and platelets (Delaire et al., 1998). In particular, CD100 is abundantly expressed on resting T cells but only weakly on resting B cells and antigen-presenting cells (APCs), including dendritic cells (DCs) (Kumanogoh and Kikutani, 2001; Kumanogoh et al., 2000). Various stimuli induce CD100 upregulation. In T cells, CD100 expression is upregulated after treatment with anti-CD3, ConA, or PHA. Stimulation with anti-CD40 or LPS upregulates CD100 expression on B cells and APCs (Fig. 2).


179

FIG. 2. The structure and expression of CD100. CD100 mRNA is expressed at high levels in embryonic and adult brain, kidney and heart. CD100 exhibits different expression levels among hematopoietic cells. CD100 is abundantly expressed on T cells, but demonstrates low levels on B cells and DCs. Following APC activation, CD100 expression is upregulated. CD100 is a class IV transmembrane-type semaphorin. This protein contains a sema domain, an Ig-like domain, a transmembrane region and a cytoplasmic tail. CD100 has several conserved cysteines within the sema domain and potential N-linked glycosylation sites. There are multiple consensus sequences for serine phosphorylation in the cytoplasmic region. A 120-kDa soluble form of CD100 can also be released from the cell surface by proteolytic cleavage. The generation of the soluble form may be dependent on metalloprotease activity and regulated by serine kinases, which associate with the cytoplasmic region of CD100.

D. BIOLOGICAL ACTIVITIES

OF

CD100

AS A

LIGAND

Cumulative evidence suggests that CD100 functions as a signaling ligand. Freeman and colleagues reported that CD100-expressing transfectants promote the in vitro aggregation and survival of B cells (Hall et al., 1996). In addition, human CD100 stimulation induces shedding of CD23 from the cell surface of B cells. In mice, the exogenous expression of CD100 or the addition of soluble recombinant mouse CD100 can enhance CD40-induced B cell proliferation and immunoglobulin production (Kumanogoh et al., 2000). CD100 also plays a role in activation and maturation of DCs (Kumanogoh et al., 2002b); soluble recombinant mouse CD100 enhances CD40-induced

180


DC maturation, as measured by the upregulation of CD40 and CD80 and the enhanced production of IL-12. Boumsell and colleagues reported that soluble human CD100 inhibits both spontaneous and chemokine-induced migration of human monocytes and monocytic and B cell lineage cell lines, suggesting a role for CD100 in immune cell migration (Delaire et al., 2001). These findings strongly suggest that CD100 performs a critical function in the immune system. E. BIOLOGICAL ACTIVITIES

OF

CD100

AS A

RECEPTOR

Antibody-crosslinking of human CD100 enhances T-cell proliferation in the presence of submitogenic doses of anti-CD3 or anti-CD2 mAb (Bougeret et al., 1992; Herold et al., 1995). Thus, CD100 likely transduces extracellular signals through its cytoplasmic domain, In addition, human T-cell CD100 is associated with serine/threonine kinase (Elhabazi et al., 1997) and protein tyrosine phosphatase (PTP), including CD45 activity (Herold et al., 1996). A switch in the type of PTP activity associated with human CD100 at the terminal stage of B cell differentiation has also been reported (Billard et al., 2000). These findings suggest that CD100 may function as a receptor transmitting activation signal for lymphocytes. The idea that CD100 is involved in physiological or pathological immune responses as a signaltransducing receptor, however, is still controversial. F. BIOLOGICALLY ACTIVE SOLUBLE FORM

OF

CD100 (sCD100)

Although CD100 is a transmembrane-type semaphorin, a 120-kDa soluble form is released from the cell surface by proteolysis in both humans and mice (Fig. 2) (Herold et al., 1995, 1996; Elhabazi et al., 2001; Wang et al., 2001). Mouse sCD100 either affinity-purified from the supernatants of activated T-cell cultures or produced recombinantly retains the ability to stimulate B cells (Wang et al., 2001). In addition, a significant amount of sCD100 is detectable in the sera of mice immunized with T-cell dependent (TD) antigens, as well as in the sera of autoimmunity-prone MRL/lpr mice. In these systems, the levels of sCD100 correlate well with antigen-specific antibody titers, although sCD100 is undetectable in the sera of unimmunized normal mice (Wang et al., 2001). These findings suggest that sCD100 is involved in both physiological and pathological immune responses in vivo. The generation of sCD100 is well regulated; its release from primary T and B cells is dependent on cellular activation (Wang et al., 2001). The cleavage of membrane CD100 on the cell surface relies on an enzymatic process, as it is inhibited by either azide or incubation at 4 C (Elhabazi et al., 2001). In addition, the light metal chelators, EDTA and EGTA inhibit up to 50% of CD100 shedding, demonstrating that metalloprotease activity may facilitate sCD100 release. Several metalloprotease inhibitors, including a zinc chelator,


181

1,10-phenanthroline, and the MMP inhibitor GM 6001, cannot inhibit this process. Thus, the mechanism governing CD100 cleavage is likely to be different from that employed for other surface molecules. The proteolytic cleavage of surface molecules often possesses structural requirements and requires post-translational changes such as phosphorylation and/or modification of glycosylation state. Recent evidence suggests that the cytoplasmic region of CD100 may be involved in the control of sCD100 shedding (Elhabazi et al., 2001). Staurosporine, a cell-permeable, broad range inhibitor of serine/threonine kinases, enhances the release of sCD100. This result suggests that serine/threonine phosphorylation regulates CD100 cleavage. The cytoplasmic region of CD100, which is associated with serine kinases, has multiple consensus sites for serine phosphorylation. Large quantities of sCD100 are detectable in the sera of transgenic mice expressing a truncated form of CD100 that lacks the cytoplasmic region, despite weak cell surface expression of the transgene product (Watanabe et al., 2001). These data support the idea that the CD100 cytoplasmic tail regulates the generation of sCD100. G. TWO TYPES

OF

CD100 RECEPTORS

CD100 utilizes two types of receptors located in different tissues with distinct binding affinities: plexin-B1 and CD72 (Fig. 3) (Tamagnone and Comoglio, 2000; Kumanogoh and Kikutani, 2001). 1. Plexin-B1 Plexin-B1, a receptor widely expressed in the fetal brain and kidney (Maestrini et al., 1996), demonstrates a high affinity (Kd ¼ 1 109 M) for CD100 (Tamagnone et al., 1999). Human CD100 stimulation enhances the interaction of the cytoplasmic region of plexin-B1 with the small GTPase Rac1 in a GTP-dependent manner (Vikis et al., 2000). In addition, activation of plexin-B1 by CD100 activates the Rho guanine nucleotide exchange factors, PDZ-RhoGEF/LARG, leading to RhoA activation (Aurandt et al., 2002; Perrot et al., 2002). Little is known, however, about the physiological relevance of CD100-plexin B1 interaction, as nor plexin-B1 or PDZ-RhoGEF/LARG is expressed in primary immune cells, such as T and B cells (Kumanogoh et al., 2000; Perrot et al., 2002). 2. CD72 Binding studies using soluble mouse CD100 revealed that there are at least two binding sites for CD100 on interacting cells (Kumanogoh et al., 2000). A mouse kidney-derived cell line, in which plexin-B1 mRNA is abundantly expressed, possesses a high affinity binding partner for mouse CD100 (Kd ¼ 1 109 M). A low affinity binding interaction (Kd ¼ 3 107 M)

182


FIG. 3. Two types of receptors for CD100. CD100 utilizes two receptors, located within distinct tissues; plexin-B1 is a receptor for CD100 in the nervous system, while CD72 is a receptor in the immune system.

is observed with mouse B cells lines, in which plexin-B1 mRNA is not detectable. Expression cloning allowed the identification of a B cell-related antigen, CD72, as a receptor for CD100 on lymphocytes (Kumanogoh et al., 2000). The binding affinity of CD100 for CD72 (Kd ¼ 3 107 M), measured by surface plasmon resonance, is in good agreement with that of CD100 to B cells. This interaction is comparable to the interactions between other costimulatory or adhesion partners such as CD28–CD80 and CD58–CD2, which are crucial to immune cell function (van der Merwe et al., 1997). CD72 is a 45-kDa type II transmembrane protein that belongs to the C-type lectin family (Nakayama et al., 1989; Von Hoegen et al., 1990). Antibodycrosslinking of CD72 promotes B cell survival and proliferation, blocks BCRinduced cell death, alters the expression of MHC class II and CD23, and activates signaling components such as phospholipase C- , CD19, Lyn, Blk and Btk (Yakura et al., 1982; Subbarao and Mosier, 1984; Yakura et al., 1986; Subbarao et al., 1988; Gordon et al., 1991; Gordon, 1994; Venkataraman et al., 1998a,b). Therefore, CD72 may mediate positive signals in B cells. The potential role of CD72 as a negative regulator in B cell responses, however, has also been reported. Tsubata and colleagues demonstrated that the cytoplasmic domain of CD72 contains an immunoreceptor tyrosine-based


183

inhibitory motifs (ITIMs), to which SHP-1, a tyrosine phosphatase, is recruited (Adachi et al., 1998). SHP-1 associates through this ITIM domain with many inhibitory receptors, such as CD22 and killer inhibitory receptors (Doody et al., 1995; Burshtyn et al., 1996), leading to the inhibition of immune cell function. Thus, the stimulation of multiple immune cell subsets, including B cells and NK cells, is inhibited by inducing tyrosine-dephosphorylation and inactivation of signaling proteins. B cells from CD72-deficient mice are hyper-proliferative in response to various stimuli (Pan et al., 1999). H. REGULATION

OF

CD72

BY

CD100

It seems paradoxical that the binding of CD100 to a ITIM-containing negative regulatory molecule generates positive B cell responses. The most likely explanation proposes that CD100 turns off the negative CD72-mediated signals, in turn enhancing B cell responses (Fig. 4) (Kumanogoh and Kikutani, 2001). The addition of recombinant soluble CD100 or anti-CD72 blocks anti- induced tyrosine phosphorylation and SHP-1 association with CD72 in B cells (Kumanogoh et al., 2000). Negative signaling through CD72 appears to be abnormally active in B cells derived from CD100-deficient mice (Shi et al., 2000). In these mice, CD72 is constitutively tyrosine-phosphorylated and associated with SHP-1. As a consequence of abnormal negative signaling through CD72, B cells from CD100-deficient mice respond poorly to stimuli, such as anti-CD40 and LPS. The phenotype of CD100-deficient mice is opposed to that of CD72-deficient mice (Pan et al., 1999), described below. In general, positive outputs are generated from stimulatory receptors while negative outputs originate from inhibitory receptors. The CD100–CD72 interaction is a unique ligand–receptor pair interaction, as this example of ligand binding to an inhibitory receptor creates a positive output. It remains unclear, however, whether CD72 alone is responsible for CD100-mediated signaling or other unidentified molecules that associate with CD72 are involved in the CD100-mediated regulation of immune responses. I. CD100

IN

B CELL DEVELOPMENT

The phenotypes of CD100-deficient and CD100-overexpressing transgenic (CD100-TG) mice have highlighted the in vivo roles of CD100 in immunity (Shi et al., 2000; Watanabe et al., 2001). The numbers of CD5 þ B220 þ B-1 cells are considerably reduced in both the spleen and the peritoneal cavity of CD100-deficient mice. As mutant mice carrying defects in SHP-1 signaling, such as SHP-1- and CD22-deficient mice, exhibit an expansion of B-1 cells, SHP-1 is thought to negatively regulate either the development or maintenance of B-1 cells. Thus, CD100 may positively regulate the development and autonomous growth of B-1 cells by inactivating SHP-1 signals mediated by CD72. The numbers of CD23 þ B220 þ B cells are increased in

184


FIG. 4. CD100 regulate B cell responses through CD72. Signals from BCR and CD40 and TLR4 are homeostatically regulated by CD100–CD72 interactions. In the absence of CD100, SHP-1 is associated with the ITIMs of CD72. SHP-1 induces tyrosine dephosphorylation and the inactivation of several signaling proteins including syk and lyn. Binding of CD100 to CD72 induces the dephosphorylation of CD72 ITIMs, resulting in the dissociation of SHP-1 from CD72.

CD100-deficient mice, consistent with the idea that CD100 stimulation induces shedding of CD23 from the cell surface of B cells (Hall et al., 1996). There remains the possibility, however, that the impaired development of B cell subsets expressing low levels of CD23, such as marginal zone B cells, contributes to this phenotype. J. CD100

IN

HUMORAL IMMUNITY

As expected from in vitro ability of CD100 to enhance B cell responses, the proliferative responses and Ig production of CD100-deficient B cells are impaired (Shi et al., 2000). Antibody responses against the TD antigen,


185

NP-CGG, are also reduced. The production of high-affinity antibodies is most profoundly affected. For T-cell independent (TI) responses, although B cell responses to LPS are affected in vitro, CD100-deficient mice do not demonstrate any apparent defects in type 1 T-cell independent (TI-1) responses. CD100-deficient mice, however, display enhanced type 2 T-cell independent (TI-2) antibody responses for particular isotypes such as IgG1. TI-2 antigens, the responses to which are related to the strength of BCR signaling, have biphasic effects on B cells; heavy crosslinking of the BCR renders the B cell unresponsive, whereas moderate BCR crosslinking stimulates the B cell. It thus appears that CD100 is crucially relevant to BCRmediated signals via CD72. Indeed, BCR-induced apoptosis and the tyrosinephosphorylation of several signaling molecules, including Syk and lyn, are also impaired in CD100-deficient mice (Kumanogoh et al., unpublished data). By contrast, the immunological phenotypes of CD100-Tg mice, in which serum levels of soluble CD100 are elevated, are opposite to those of CD100-deficient mice (Watanabe et al., 2001). These mice display in vitro B cell hyperresponsiveness to various stimuli, enhanced TD antibody responses, and reduced TI-2 antibody responses of certain isotypes. Transgenic expression of a truncated CD100, however, can rescue impaired immune responses of CD100-deficient mice, suggesting that the functions of the membrane-bound CD100 can be compensated for by overexpression of soluble CD100. The comparison of CD100-deficient and CD100-Tg mice demonstrates the crucial roles of CD100 in antibody responses directed against TD and TI-2 antigens. Furthermore, the phenotype of CD72-deficient mice (Pan et al., 1999), is opposite to that of CD100-deficient mice (Shi et al., 2000) and are similar to CD100-Tg mice (Watanabe et al., 2001), substantiating the regulation of CD72-mediated negative signals by CD100. K. CD100

IN

CELLULAR IMMUNITY

CD100 is abundantly expressed in T cells, while CD72 is expressed in both B cells and APCs (Kumanogoh et al., 2002b). Therefore, CD100 is likely to play an important role in T-cell-professional APC interactions (Fig. 5). 1. Involvement of CD100 in Activation and Maturation of DCs Recombinant sCD100 can enhance CD40-induced DC activation, leading to the upregulation of costimulatory molecules and IL-12 production. Agonistic anti-CD72 stimulation mimics the effects of sCD100 on DCs (Kumanogoh et al., 2002b). Thus, CD100 is involved in the activation and maturation of DCs. Furthermore, the activities of CD100 against DCs may have potential utility in DC-based vaccinations. As DCs have been used as an adjuvant to enforce immunity against both infections and tumors, stimulation of antigen-pulsed DCs with both recombinant mouse CD100 and anti-CD40

186


FIG. 5. The involvement of class IV semaphorins, CD100/Sema4D and Sema4A in several phases of immune reactions. (A) CD100 and CD40L are concomitantly expressed on the surface of activated T cells, while CD72 and CD40 are expressed on the surface of B cells and APCs. In T cell–B cell interactions, CD100 functions through CD72 during the same phase as CD40–CD40L interactions, which contribute to B cell activation resulting in enhanced proliferation and antibody production. (B) In T-APC interactions, CD100 enhances CD40-induced maturation of APCs, including DCs. Sema4A, expressed on APCs but not on T cells, enhances T cell activation through Tim-2 ligation. This reciprocal stimulation between T cells and APCs acting through CD100 and Sema4A contributes to cellular immunity. (C) The expression of CD100 is also upregulated on the surface of B cells. Thus, CD100 derived from B cells may also be involved in regulation of B cell and activation, contributing to both TI responses but also the survival of GC B cells. (D) Both CD100 and CD72 are expressed on the surface of APCs, including DCs. CD100 derived from APCs might also play a role in autocrine activation of APCs, which might be involved in innate immune responses.

mAb may significantly augment in vivo immunogenicity, resulting in enhanced generation of antigen-specific and memory T cells. 2. Essential Requirement for T-Cell-Derived CD100 in T Cell Differentiation Although T-cells express abundant CD100, the expression of CD100 is only upregulated on DCs and other APCs upon stimulation (Kumanogoh and


187

Kikutani, 2001). Thus, during periods of inflammation, CD100 can be supplied by either the T cells or APCs. DCs from CD100-deficient mice display poor allostimulatory activity in MLRs. Following stimulation with anti-CD40, CD100-deficient DCs demonstrate defects in costimulatory molecule (such as CD80 and CD40) and MHC class II expression, as well as in IL-12 production (Kumanogoh et al., 2002b). These findings indicate that DC-derived CD100 contributes to DC activation and maturation. DC-derived CD100, however, appears to be insufficient for the generation of functional effector T cells. Naive T cells from TCR-transgenic mice differentiate into cytokine-producing effector cells following incubation with normal APCs in vitro. In this experimental system, CD100-deficient TCR-Tg T cells fail to differentiate into effector cells even after culture with specific antigens in the presence of CD100-expressing APCs. In contrast, CD100-expressing TCR-Tg T cells differentiate into effector cells in the presence of CD100-deficient APCs (Kumanogoh et al, 2002b). Thus, in T cell/APC interactions, T cells are the primary, essential source of CD100. 3. Involvement of CD100 in the In Vivo Generation of Antigen-Specific T Cells Although the T cells of CD100-deficient mice develop normally, in vivo generation of antigen-specific T cells is severely impaired (Shi et al., 2000). Defective antigen-specific T cell generation in these mice can be rescued by administration of recombinant soluble mouse CD100 protein. Of note, CD100-deficient mice are resistant to the induction of experimental autoimmune encephalomyelitis (EAE) via myelin oligodendrocyte glycoprotein (MOG)-peptide administration, due to impaired generation of MOG-specific T cells (Kumanogoh et al., 2002b). These observations demonstrate a critical role for CD100 in the generation of antigen-specific T cell responses in vivo (Fig. 5).

III. Class IV Semaphorin (2): Sema4A

A. STRUCTURE

AND

EXPRESSION

The expression of Sema4A, originally identified as semB (Puschel et al., 1995), gradually increases during embryonic development. Its precise function and expression profile, however, has not yet been elucidated. A cDNA fragment of Sema4A was isolated in a search for semaphorins expressed in DCs through PCR cloning using degenerated oligonucleotide primers derived from motifs conserved among semaphorin family members (Kumanogoh et al., 2002a). Sema4A belongs to the class IV subfamily, possessing a structure

188


FIG. 6. Structure and expression of Sema4A. Sema4A is a class IV transmembrane-type semaphorin member. It contains a sema domain, an Ig-like domain, a transmembrane region, and a cytoplasmic region, which has several conserved cysteines in its sema domain and several potential N-linked glycosylation sites. Sema4A is preferentially expressed on the surface of DCs and B cells, not resting T cells.

similar to that of CD100 (Fig. 6) (Semaphorin Nomenclature Committee, 1999). All of the cysteine residues within the sema domain are conserved between CD100 and Sema4A. Sema4A mRNA is detectable in brain, spleen, lung, kidney, and the testis in adults, in addition to embryonic expression. Flow cytometric analysis using an anti-Sema4A mAb demonstrated that Sema4A is expressed on bone marrowderived and splenic DCs as well as B cells, but not resting T cells (Kumanogoh et al., 2002a). No differences were found in Sema4A expression between CD8 þ and CD8 DCs. When B cells are stimulated with anti-CD40 mAb, Sema4A is upregulated. The expression of Sema4A becomes detectable upon the stimulation of T cells with anti-CD3 mAb. The expression pattern of Sema4A contrasts that of CD100, in which expression is preferentially seen on T cells.


B. ACTIVITIES

OF

189

SEM4A IN VITRO

Although recombinant soluble Sema4A (sSema4A) does not have a stimulatory effect on B cells and DCs, Sema4A possesses the activity to provide T cell costimulation. The addition of recombinant sSema4A enhances the T cell proliferation and IL-2 production following stimulation by antiCD3. sSema4A significantly enhances the induction of either Th1 cells producing IFN- or Th2 cells producing IL-4 under the appropriate respective culture conditions. In addition, sSema4A enhances MLRs between allogeneic T cells and DC. Anti-Sema4A can block these MLRs, indicating that Sema4A functions in T cell activation by mediating stimulatory interactions between T cells and DCs (Kumanogoh et al., 2002a). C. ACTIVITIES

OF

SEMA4A IN VIVO

The role of Sema4A in in vivo immune responses has been clarified using sSema4A and anti-Sema4A mAb (Kumanogoh et al., 2002a). The administration of sSema4A significantly enhances the generation of antigen-specific T cells following subcutaneous immunization with keyhole limpet hemocyanin (KLH) in complete Freund’s adjuvant (CFA). In contrast, administration of anti-Sema4A blocks the priming of antigen-specific T cells. The treatment of mice with anti-Sema4A inhibits the development of EAE induced by MOG-peptide. Delayed administration of sSema4A or anti-Sema4A does not affect antigen-specific T cell generation, suggesting that Sema4A acts at an early phase of in vivo T cell activation. D. TIM-2

AS A

SEMA4A RECEPTOR

The staining of various cells with biotinylated Sema4A-Fc demonstrates that the Sema4A receptor is expressed on the surface of activated T cells, not on B cells and DCs. Sema4A-binding partners, however, are observed on the surface of a subset of T cell lines, such as EL-4 cells. Expression cloning using Sema4A-Fc isolated a full-length cDNA encoding T cell, immunoglobulin domain and mucin domain protein (TIM)-2, belonging to the Tim protein family (Kumanogoh et al., 2002a; McIntire et al., 2001). Tim-2 is a transmembrane protein consisting of 305 amino acids (Fig. 7). Tim-2 is a functional receptor for Sema4A; Sema4A-Fc specifically binds COS7 cells transfected with Tim-2 cDNA but not to COS7 cells transfected with either Tim-3 cDNA or vector alone. The binding of Sema4A to Tim-2-expressing COS7 cells is specifically blocked by excess Tim-2-Fc. Surface plasmon resonance estimates the affinity constant (Kd) of Sema4A for Tim-2 at 7.0 108 M (kon ¼ 7.0 104 M1 s1, koff ¼ 4.9 103 s1). Sema4Astimulation also induces the tyrosine-phosphorylation of Tim-2, not Tim-3, in COS7 cells.

190


FIG. 7. Structure of Tim-2. Tim-2, a member of the Tim-protein family, possesses an Ig-like domain, a mucin domain, a transmembrane region, and an intracellular tail containing consensus tyrosine-phosphorylation sites. The mucin domain has multiple putative sites for O-linked glycosylation, while the Ig domain has several sites for putative N-linked glycosylation.

E. TIM PROTEIN FAMILY The Tim protein family, expressed on T cells, is characterized by conserved immunoglobulin and mucin domains. McIntire et al. identified a locus conferring susceptibility to mouse allergen-induced airway hypersensitivity, dubbed T cell and airway phenotype regulator (Tapr) (McIntire et al., 2001). Subsequent positional cloning revealed a new family of genes, designated as Tims. All three Tim family members encode cell surface glycoproteins with common structural motifs, including a signal peptide, immunoglobulin domain, mucin domain, a transmembrane region, and an intracellular tail containing consensus tyrosine-phosphorylation sites. The mucin domain has multiple putative sites for O-linked glycosylation and the Ig domain contains several sites of putative N-linked glycosylation. Kuchroo and colleagues independently identified Tim-3 through a screening for Th1-reactive mAbs as a new Th1-specific cell surface protein (Monney et al., 2002). The administration of anti-Tim-3 mAbs promoted the development of EAE, a Th1-dependent autoimmune disease. This result suggests the involvement of Tim-3 in the interaction between Th1 cells and macrophages, resulting in the expansion and activation of macrophages. Mouse Tim-1 possesses 78% identity with rat KIM-1 and 42% identity with human Hepatitis A virus cellular receptor (hHAVcr-1), its orthologs. There are remarkable numbers of mouse Tim-1 sequence polymorphisms in Tapr as well as in human Tim-1 (hHAVcr-1) (observed in the human genome and EST


191

databases). Tim-2 exhibits 64% identity to mouse Tim-1, 60% identity to rat KIM-1 and 32% identity to hHAVcr-1. Polymorphisms within the Tim-2 sequence have not been reported. The human homologue of Tim-2 has not been determined, although there is also substantial identity between Tim-2 and hHAVcr-1, raising the possibility that variants of hHAVcr-1 are homologs of Tim-2. Alternatively, a Tim-related gene located near the human hHAVcr-1 locus, recently identified by a database search (Kumanogoh et al., unpublished), may function as the human homolog of Tim-2. Although the natural ligands of Tim-1 and Tim-3 have not been identified, these Tim protein ligands, including Sema4A, will likely regulate the activation and differentiation of T cells.

IV. Virus-Encoded Semaphorins and their Cellular Counterparts

Viruses encode proteins that can function as immune modulators. Such proteins interfere with the host immune system to facilitate the infectious process or support viral transmission. Many viral immune modulators have significant homology with host proteins, probably deriving from viral scavenging during the co-evolution of virus and host (Spriggs, 1999). In recent years, the studies of these viral immune modulators have significantly advanced our understanding of the pathogenesis associated with these viral infections. Several viruses also encode semaphorins, suggesting a function for these molecules as immune modulators in an infected host. A. A39R/SEMAVA

AND

AHVSEMA/SEMAVB

An open reading frame (ORF), A39R, related to the semaphorin family was discovered in the genome of vaccinia virus, a member of the poxvirus family (Kolodkin et al., 1993). A39R, the smallest semaphorin, consists of only a truncated extracellular sema domain. The mammalian homolog still remains to be identified. In contrast to other semaphorins, A39R is thought to act as a monomer (Spriggs, 1999). Infection with vaccinia virus is associated with skin lesions at the site of inoculation followed by secondary viral replication in lymphoid organs. A39R induces robust responses of cell aggregation in human monocytes, the induction of proinflammatory cytokines (e.g., IL-6 and IL-8), and the upregulation of the monocyte cell surface marker CD54 (ICAM-1) (Comeau et al., 1998; Spriggs, 1999), suggesting the involvement of A39R in the immune responses of vaccinia-infected hosts. Ensser and Fleckenstein have reported another semaphorin, AHVsema, encoded by the alcelaphine herpes virus (AHV). Unlike other herpes viruses, AHV is not neurotropic, but causes wildebeest-associated malignant catarrhal fever, a lymphoproliferative syndrome occurring in ungulate species other

192


than the natural host. Both mouse and human homologs of AHVsema have been identified. Spriggs and colleagues have identified that both of these viral semaphorins, A39R and AHVsema, bind to the cellular receptor, VESPR/CD232/plexin-C1. A39R-induced induction of inflammatory cytokines is abrogated by a blocking antibody against VESPR/CD232/plexin-C1 (Comeau et al., 1998). These findings suggest a pathogenic role for viral semaphorins and evoke great interest in discovering the physiological roles of their mammalian counterparts. B. CD108/SEMA7A/SEMA-K1 Two independent groups identified the class VII semaphorin Sema7A (Ensser and Fleckenstein, 1995; Xu et al., 1998; Yamada et al., 1999), also known as CD108 or Sema-K1, a molecule homologous to AVHsema (Ensser and Fleckenstein, 1995). CD108, originally discovered as the John-MiltonHagen human blood group antigen (Mudad et al., 1995), is a GPI-anchored cell surface glycoprotein preferentially expressed on activated lymphocytes and thymocytes (Yamada et al., 1999). The molecular structure of CD108 revealed its membership in the semaphorin family. A search for vertebrate homologs to AHVsema using the BLAST algorithm identified Sema-K1 (Xu et al., 1998), later confirmed as CD108. CD108/Sema7A/Sema-K1 specifically binds to VESPR/CD232/plexin-C1 (Tamagnone et al., 1999), confirming this semaphorin is a cellular counterpart of AHVsema. Recombinant soluble CD108/Sema7A/Sema-K1 protein exhibits similar activities against monocytes as AHVsema; inducing the production of inflammatory cytokines, such as TNF-, IL-6 and IL-8. In addition, CD108 demonstrates chemotactic attraction to monocytes (Holmes et al., 2002), substantiating a role for CD108/Sema7A/Sema-K1 in inflammatory responses. Thus, the viral semaphorins may mediate an immunomodulatory effect by mimicking CD108/Sema7A/Sema-K1. It is therefore important to determine the physiological importance of interactions between CD108/Sema7A/ Sema-K1 and VESPR/CD232/plexin-C1. V. Class III Semaphorin: Sema3A in Immune Cell Migration

Sema3A/H-SemaIII is the human homolog of collapsin-1, the first identified vertebrate semaphorin (Kolodkin et al., 1993). This molecule is involved in repulsive growth cone guidance during neuronal development (Tessier-Lavigne and Goodman, 1996). The function of Sema3A/H-SemaIII, the representative semaphorin member, as an axonal guidance factor has been extensively studied in neurobiology. Sema3A/H-SemaIII requires the neuropilin-1 receptor to exhibit its chemorepulsive activity on neurons.


193

Boumsell and colleagues reported that Sema3A/H-SemaIII also plays a role in regulating monocyte migration (Delaire et al., 2001). Sema3A/H-SemaIII inhibits spontaneous monocytic-cell migration in a transwell assay. The soluble form of CD100 also inhibits immune-cell migration. As Sema3A/H-SemaIII can be recognized by some anti-human CD100 mAbs, the ability of Sema3A/ H-SemaIII to inhibit monocytic cell migration can be abolished by preincubation with these mAbs. These experiments propose the possibility of a homologous domain common to Sema3A/H-SemaIII and human CD100 that is important in cell migration. Neuropilin-1, the known receptor of Sema3A/H-SemaIII, is not expressed on the target immune cells, suggesting the inhibitory effect of Sema3A/H-SemaIII on immune-cell migration is mediated by different receptors than those previously known to bind within the nervous system (Delaire et al., 2001). Neuropilin-1, however, is expressed on the surface of human T cells and DCs (Tordjman et al., 2002). Immune cell migration, hence, may be mediated by similar mechanisms as those identified in the nervous system. Migration, a basic feature of many cell types throughout a broad range of species, occurs in both the nervous and immune systems. Thus, it will be critical to determine the significance of semaphorins in immune cell migration. VI. Neuropilin-1 in initial T cell/DC Contacts

Neuropilin-1 was the first semaphorin receptor demonstrated to bind class III semaphorins (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997). VEGF has been shown to bind neuropilin-1 (Soker et al., 1998), suggesting the involvement of neuropilin-1 in both axonal guidance and angiogenesis in a ligand-dependent manner. In contrast, neuropilin-1 forms homo-oligomers; thus, neuropilin-1–neuropilin-1 interactions may occur concurrently in trans (Chen et al., 1998). No function, however, had yet been reported for this homophilic interaction. Recently, Tordjman et al. have determined that neuropilin-1 is expressed by DCs and resting T cells (Tordjman et al., 2002). They demonstrated that neuropilin-1 is critically involved in the initiation of a primary immune response originating in the interaction of DCs and resting T cells through a homophilic interaction of neuropilin-1. In the initial contact between DCs and resting T cells, a variety of molecules function in the formation of the immunological synapse, including the DC-specific nonintegrin (DC-SIGN) with ICAM-3, CD58 (LFA-3) with CD2 or LFA-1 with either ICAM-1 or ICAM-2 (Banchereau and Steinman, 1998). Polarization of neuropilin-1 localization is observed on resting T cells following interaction with DCs. Double immunofluorescence staining with anti-CD3 demonstrated that neuropilin-1 and CD3 colocalize. Thus, following the interaction with a

194


professional APC, neuropilin-1 is positioned at the T cell pole of the immunological synapse. In addition, soluble neuropilin-1 binds both DCs and resting T cells. Anti-neuropilin-1 blocked the induction of T cell proliferation by allogeneic DCs (Tordjman et al., 2002). Collectively, these findings indicate a role for neuropilin-1 in the initiation of the primary immune responses; these studies, however, do not exclude the possibility that unidentified transmembrane proteins, including these semaphorins, are involved in stabilizing the initial interaction between DCs and resting T cells via neuropilin-1.

VII. Perspectives

The term Semaphorins is derived from ‘‘Semaphore,’’ a primitive but effective communication tool. In a similar manner, semaphorins act as communication tools during neuronal development. The current findings indicate that several semaphorins also play crucial roles in the immune system. In addition to the semaphorins described here, we recently found additional semaphorin molecules that appear to function in immune regulation. These molecules represent a new family of immunoregulatory molecules. CD100, the most extensively studied immunoregulatory semaphorin, utilizes CD72 as a receptor within the immune system. It must be clarified if other immunoregulatory semaphorins also use molecules distinct from neuropilins and plexins as their receptors. Indeed, the class IV semaphorin Sema4A utilizes Tim-2 as a receptor in T cells, although this evidence does not exclude the use of another member of the plexin family as a Sema4A receptor. Further studies on additional semaphorins expressed in the immune system may provide clues aiding in our understanding of how these molecules exert different biological activities in the immune and nervous systems. As studies with ‘‘Immuno-semaphorins’’ suggest that these molecules function in immune responses including autoimmune diseases and viral infections, semaphorins may prove to be targets for therapy of immunological and infectious diseases. Future studies should help to clarify the biological function of semaphorins in the immune system. Such information should help to establish a new paradigm of immune cell communication acting through an immunoregulatory semaphorin network. ACKNOWLEDGMENTS We would like to thank K. Kubota for excellent secretarial assistance. We also thank Dr. C. Watanabe and N. Takegahara for outstanding figure illustrations. A. K. and H. K were supported by research grants from the Ministry of Education, Culture, Science and Technology of Japan.


195

REFERENCES Adachi, T., Flaswinkel, H., Yakura, H., Reth, M., and Tsubata, T. (1998). The B cell surface protein CD72 recruits the tyrosine phosphatase SHP-1 upon tyrosine phosphorylation. J. Immunol. 160, 4662–4665. Aurandt, J., Vikis, H. G., Gutkind, J. S., Ahn, N., and Guan, K. L. (2002). The semaphorin receptor plexin-B1 signals through a direct interaction with the Rho-specific nucleotide exchange factor, LARG. Proc. Natl. Acad. Sci. USA 99, 12085–12090. Banchereau, J., and Steinman, R. M. (1998). Dendritic cells and the control of immunity. Nature 392, 245–252. Behar, O., Golden, J. A., Mashimo, H., Schoen, F. J., and Fishman, M. C. (1996). Semaphorin III is needed for normal patterning and growth of nerves, bones and heart. Nature 383, 525–528. Billard, C., Delaire, S., Raffoux, E., Bensussan, A., and Boumsell, L. (2000). Switch in the protein tyrosine phosphatase associated with human CD100 semaphorin at terminal B-cell differentiation stage. Blood 95, 965–972. Bougeret, C., Mansur, I. G., Dastot, H., Schmid, M., Mahouy, G., Bensussan, A., and Boumsell, L. (1992). Increased surface expression of a newly identified 150-kDa dimer early after human T lymphocyte activation. J. Immunol. 148, 318–323. Burshtyn, D. N., Scharenberg, A. M., Wagtmann, N., Rajagopalan, S., Berrada, K., Yi, T., Kinet, J.-P., and Long, E. O. (1996). Recruitment of tyrosine phosphatase HCP by the killer cell inhibitor receptor. Immunity 4, 77–85. Chen, H., He, Z., Bagri, A., and Tessier-Lavigne, M. (1998). Semaphorin-neuropilin interactions underlying sympathetic axon responses to class III semaphorins. Neuron 21, 1283–1290. Comeau, M. R., Johnson, R., DuBose, R. F., Petersen, M., Gearing, P., VandenBos, T., Park, L., Farrah, T., Buller, R. M., Cohen, J. I. et al. (1998). A poxvirus-encoded semaphorin induces cytokine production from monocytes and binds to a novel cellular semaphorin receptor, VESPR. Immunity 8, 473–482. Delaire, S., Billard, C., Tordjman, R., Chedotal, A., Elhabazi, A., Bensussan, A., and Boumsell, L. (2001). Biological activity of soluble CD100. II. Soluble CD100, similarly to H-SemaIII, inhibits immune cell migration. J. Immunol. 166, 4348–4354. Delaire, S., Elhabazi, A., Bensussan, A., and Boumsell, L. (1998). CD100 is a leukocyte semaphorin. Cell Mol. Life Sci. 54, 1265–1276. Doody, G. M., Justement, L. B., Delibrias, C. C., Matthews, R. J., Lin, J., Thomas, M. L., and Fearon, D. T. (1995). A role in B cell activation for CD22 and the protein tyrosine phosphatase SHP. Science 269, 242–244. Elhabazi, A., Delaire, S., Bensussan, A., Boumsell, L., and Bismuth, G. (2001). Biological activity of soluble CD100. I. The extracellular region of CD100 is released from the surface of T lymphocytes by regulated proteolysis. J. Immunol. 166, 4341–4347. Elhabazi, A., Lang, V., Herold, C., Freeman, G. J., Bensussan, A., Boumsell, L., and Bismuth, G. (1997). The human semaphorin-like leukocyte cell surface molecule CD100 associates with a serine kinase activity. J. Biol. Chem. 272, 23515–23520. Ensser, A., and Fleckenstein, B. (1995). Alcelaphine herpesvirus type 1 has a semaphorin-like gene. J. Gen. Virol. 76, 1063–1067. Furuyama, T., Inagaki, S., Kosugi, A., Noda, S., Saitoh, S., Ogata, M., Iwahashi, Y., Miyazaki, N., Hamaoka, T., and Tohyama, M. (1996). Identification of a novel transmembrane semaphorin expressed on lymphocytes. J. Biol. Chem. 271, 33376–33381. Giger, R. J., and Kolodkin, A. L. (2001). Silencing the siren: guidance cue hierarchies at the CNS midline. Cell 105, 1–4. Gordon, J. (1994). B-cell signalling via the C-type lectins CD23 and CD72. Immunol. Today 15, 411–417.

196


Gordon, J., Katira, A., Strain, A. J., and Gillis, S. (1991). Inhibition of interleukin 4-promoted CD23 production in human B lymphocytes by transforming growth factor-beta, interferons or anti-CD19 antibody is overridden on engaging CD40. Eur. J. Immunol. 21, 1917–1922. Grunwald, I. C., and Klein, R. (2002). Axon guidance: receptor complexes and signaling mechanisms. Curr. Opin. Neurobiol. 12, 250–259. Hall, K. T., Boumsell, L., Schultze, J. L., Boussiotis, V. A., Dorfman, D. M., Cardoso, A. A., Bensussan, A., Nadler, L. M., and Freeman, G. J. (1996). Human CD100, a novel leukocyte semaphorin that promotes B-cell aggregation and differentiation. Proc. Natl. Acad. Sci. USA A93, 11780–11785. He, Z., and Tessier-Lavigne, M. (1997). Neuropilin is a receptor for the axonal chemorepellent Semaphorin III. Cell 90, 739–751. Herold, C., Bismuth, G., Bensussan, A., and Boumsell, L. (1995). Activation signals are delivered through two distinct epitopes of CD100, a unique 150 kDa human lymphocyte surface structure previously defined by BB18 mAb. Int. Immunol. 7, 1–8. Herold, C., Elhabazi, A., Bismuth, G., Bensussan, A., and Boumsell, L. (1996). CD100 is associated with CD45 at the surface of human T lymphocytes. Role in T cell homotypic adhesion. J. Immunol. 157, 5262–5268. Holmes, S., Downs, A., Fosberry, A., Hayes, P., Michalovich, D., Murdoch, P., Moores, K., Fox, J., Deen, K., Pettman, G., Wattam, T., and Lewis, C. (2002). Sema7A is a Potent Monocyte Stimulator. Scand. J. Immunol. 56, 270–275. Kitsukawa, T., Shimono, A., Kawakami, A., Kondoh, H., and Fujisawa, H. (1995). Overexpression of a membrane protein, neuropilin, in chimeric mice causes anomalies in the cardiovascular system, nervous system and limbs. Development 121, 4309–4318. Kolodkin, A. L., Levengood, D. V., Rowe, E. G., Tai, Y. T., Giger, R. J., and Ginty, D. D. (1997). Neuropilin is a semaphorin III receptor. Cell 90, 753–762. Kolodkin, A. L., Matthes, D. J., and Goodman, C. S. (1993). The semaphorin genes encode a family of transmembrane and secreted growth cone guidance molecules. Cell 75, 1389–1399. Kolodkin, A. L., Matthes, D. J., O’Connor, T. P., Patel, N. H., Bentley, D., and Goodman, C. S. (1992). Fascilin IV: sequence, expression, and function during growth cone guidance in the grasshopper embryo. Neuron 9, 831–845. Koppel, A. M., Feiner, L., Kobayashi, H., and Raper, J. A. (1997). A 70 amino acid region within the semaphorin domain activates specific cellular response of semaphorin family members. Neuron 19, 531–537. Koppel, A. M., and Raper, J. A. (1998). Collapsin-1 covalently dimerizes, and dimerization is necessary for collapsing activity. J. Biol. Chem. 273, 15708–15713. Kumanogoh, A., and Kikutani, H. (2001). The CD100-CD72 interaction: a novel mechanism of immune regulation. Trends Immunol. 22, 670–676. Kumanogoh, A., Marukawa, S., Suzuki, K., Takegahara, N., Watanabe, C., Ch’ng, E., Ishida, I., Fujimura, H., Sakoda, S., Yoshida, K., and Kikutani, H. (2002a). Class IV semaphorin Sema4A enhances T-cell activation and interacts with Tim-2. Nature 419, 629–633. Kumanogoh, A., Suzuki, K., Ch’ng, E., Watanabe, C., Marukawa, S., Takegahara, N., Ishida, I., Sato, T., Habu, S., Yoshida, K., and Kikutani, H. (2002b). Requirement for the lymphocyte semaphorin, CD100, in the induction of antigen-specific T cells and the maturation of dendritic cells. J. Immunol. 169, 1175–1181. Kumanogoh, A., Watanabe, C., Lee, I., Wang, X., Shi, W., Araki, H., Hirata, H., Iwahori, K., Uchida, J., Yasui, T., Matsumoto, M., Yoshida, K., Yakura, H., Pan, C., Parnes, J. R., and Kikutani, H. (2000). Identification of CD72 as a lymphocyte receptor for the class IV semaphorin CD100: a novel mechanism for regulating B cell signaling. Immunity 13, 621–631. Luo, Y., Raible, D., and Raper, J. A. (1993). Collapsin: a protein in brain that induces the collapse and paralysis of neuronal growth cones. Cell 75, 217–227.


197

Maestrini, E., Tamagnone, L., Longati, P., Cremona, O., Bione, S., Tamanini, F., Neel, B. G., Toniolo, D., and Comoglio, P. M. (1996). A family of transmembrane proteins with homology to the MET-hepatocyte growth factor receptor. Proc. Natl. Acad. Sci. USA 93, 674–678. McIntire, J. J., Umetsu, S. E., Akbari, O., Potter, M., Kuchroo, V. K., Barsh, G. S., Freeman, G. J., Umetsu, D. T., and DeKruyff, R. H. (2001). Identification of Tapr (an airway hyperreactivity regulatory locus) and the linked Tim gene family. Nat. Immunol. 2, 1109–1116. Monney, L., Sabatos, C. A., Gaglia, J. L., Ryu, A., Waldner, H., Chernova, T., Manning, S., Greenfield, E. A., Coyle, A. J., Sobel, R. A. et al. (2002). Th1-specific cell surface protein Tim-3 regulates macrophage activation and severity of an autoimmune disease. Nature 415, 536–541. Mudad, R., Rao, N., Angelisova, P., Horejsi, V., and Telen, M. J. (1995). Evidence that CDw108 membrane protein bears the JMH blood group antigen. Transfusion 35, 566–570. Nakayama, E., von Hoegen, I., and Parnes, J. R. (1989). Sequence of the Lyb-2 B-cell differentiation antigen defines a gene superfamily of receptors with inverted membrane orientation. Proc. Natl. Acad. Sci. USA 86, 1352–1356. Pan, C., Baumgarth, N., and Parnes, J. R. (1999). CD72-deficient mice reveal nonredundant roles of CD72 in B cell development and activation. Immunity 11, 495–506. Perrot, V., Vazquez-Prado, J., and Gutkind, J. S. (2002). Plexin B regulates Rho through the guanine nucleotide exchange factors Leukemia-associated RhoGEF (LARG) and PDZRhoGEF. J. Biol. Chem 277, 43115–43120. Puschel, A. W., Adams, R. H., and Betz, H. (1995). Murine semaphorin D/collapsin is a member of a diverse gene family and creates domains inhibitory for axonal extension. Neuron 14, 941–948. Raper, J. A. (2000). Semaphorins and their receptors in vertebrates and invertebrates. Curr. Opin. Neurobiol. 10, 88–94. Sekido, Y., Bader, S., Latif, F., Chen, J. Y., Duh, F. M., Wei, M. H., Albanesi, J. P., Lee, C. C., Lerman, M. I., and Minna, J. D. (1996). Human semaphorins A(V) and IV reside in the 3p21.3 small cell lung cancer deletion region and demonstrate distinct expression patterns. Proc. Natl. Acad. Sci. USA 93, 4120–4125. Semaphorin Nomenclature Committee (1999). Unified nomenclature for the semaphorins/ collapsins. Cell 97, 551–552. Shi, W., Kumanogoh, A., Watanabe, C., Uchida, J., Wang, X., Yasui, T., Yukawa, K., Ikawa, M., Okabe, M., Parnes, J. R., Yoshida, K., and Kikutani, H. (2000). The class IV semaphorin CD100 plays nonredundant roles in the immune system: defective B and T cell activation in CD100deficient mice. Immunity 13, 633–642. Soker, S., Takashima, S., Miao, H. Q., Neufeld, G., and Klagsbrun, M. (1998). Neuropilin-1 is expressed by endothelial and tumor cells as an isoform-specific receptor for vascular endothelial growth factor. Cell 92, 735–745. Spriggs, M. K. (1999). Shared resources between the neural and immune systems: semaphorins join the ranks. Curr. Opin. Immunol. 11, 387–391. Steward, O. (2002). Translating axon guidance cues. Cell 110, 537. Subbarao, B., Morris, J., and Baluyut, A. R. (1988). Properties of anti-Lyb-2-mediated B-cell activation and the relationship between Lyb-2 molecules and receptors for B-cell stimulatory factor-1 on murine B lymphocytes. Cell Immunol. 112, 329–342. Subbarao, B., and Mosier, D. E. (1984). Activation of B lymphocytes by monovalent anit-Lyb-2 antibodies. J. Exp. Med. 159, 1796–1801. Takahashi, T., Fournier, A., Nakamura, F., Wang, L. H., Murakami, Y., Kalb, R. G., Fujisawa, H., and Strittmatter, S. M. (1999). Plexin-neuropilin-1 complexes form functional semaphorin-3A receptors. Cell 99, 59–69. Takahashi, T., and Strittmatter, S. M. (2001). Plexina1 autoinhibition by the plexin sema domain. Neuron 29, 429–439.

198


Tamagnone, L., Artigiani, S., Chen, H., He, Z., Ming, G. I., Song, H., Chedotal, A., Winberg, M. L., Goodman, C. S., Poo, M., Tessier-Lavigne, M., and Comoglio, P. M. (1999). Plexins are a large family of receptors for transmembrane, secreted, and GPI-anchored semaphorins in vertebrates. Cell 99, 71–80. Tamagnone, L., and Comoglio, P. M. (2000). Signalling by semaphorin receptors: cell guidance and beyond. Trends Cell Biol. 10, 377–383. Tessier-Lavigne, M., and Goodman, S. C. (1996). The molecular biology of axon guidance. Science 274, 1123–1133. Tordjman, R., Lepelletier, Y., Lemarchandel, V., Cambot, M., Gaulard, P., Hermine, O., and Romeo, P. H. (2002). A neuronal receptor, neuropilin-1, is essential for the initiation of the primary immune response. Nat. Immunol. 3, 477–482. van der Merwe, P. A., Bodian, D. L., Daenke, S., Linsley, P., and Davis, S. J. (1997). CD80 (B7-1) binds to both CD28 and CTLA-4 with a low affinity and very fast kinetics. J. Exp. Med. 185, 393–403. Venkataraman, C., Lu, P. J., Buhl, A. M., Chen, C. S., Cambier, J. C., and Bondada, S. (1998a). CD72-mediated B cell activation involves recruitment of CD19 and activation of phosphatidylinositol 3-kinase. Eur. J. Immunol. 28, 3003–3016. Venkataraman, C., Muthusamy, N., Muthukkumar, S., and Bondada, S. (1998b). Activation of lyn, blk, and btk but not syk in CD72-stimulated B lymphocytes. J. Immunol. 160, 3322–3329. Vikis, H. G., Li, W., He, Z., and Guan, K. L. (2000). The semaphorin receptor plexin-B1 specifically interacts with active Rac in a ligand-dependent manner. Proc. Natl. Acad. Sci. USA 97, 12457–12462. Von Hoegen, I., Nakayama, E., and Parnes, J. R. (1990). Identification of a human protein homologous to the mouse Lyb-2 B cell differentiation antigen and sequence of the corresponding cDNA. J. Immunol. 144, 4870–4877. Wang, X., Kumanogoh, A., Watanabe, C., Shi, W., Yoshida, K., and Kikutani, H. (2001). Functional soluble CD100/Sema4D released from activated lymphocytes: possible role in normal and pathologic immune responses. Blood 97, 3498–3504. Watanabe, C., Kumanogoh, A., Shi, W., Suzuki, K., Yamada, S., Okabe, M., Yoshida, K., and Kikutani, H. (2001). Enhanced immune responses in transgenic mice expressing a truncated form of the lymphocyte semaphorin CD100. J. Immunol. 167, 4321–4328. Winberg, M. L., Noordermeer, J. N., Tamagnone, L., Comoglio, P. M., Spriggs, M. K., TessierLavigne, M., and Goodman, C. S. (1998). Plexin A is a neuronal semaphorin receptor that controls axon guidance. Cell 95, 903–916. Xu, X., Ng, S., Wu, Z. L., Nguyen, D., Homburger, S., Seidel-Dugan, C., Ebens, A., and Luo, Y. (1998). Human semaphorin K1 is glycosylphosphatidylinositol-linked and defines a new subfamily of viral-related semaphorins. J. Biol. Chem. 273, 22428–22434. Yakura, H., Kawabata, I., Ashida, T., Shen, F. W., and Katagiri, M. (1986). A role for Lyb-2 in B cell activation mediated by a B cell stimulatory factor. J. Immunol. 136, 1475–1481. Yakura, H., Shen, F. W., Bourcet, E., and Boyse, E. A. (1982). Evidence that Lyb-2 is critical to specific activation of B cells before they become responsive to T cell and other signals. J. Exp. Med. 155, 1309–1316. Yamada, A., Kubo, K., Takeshita, T., Harashima, N., Kawano, K., Mine, T., Sagawa, K., Sugamura, K., and Itoh, K. (1999). Molecular cloning of a glycosylphosphatidylinositolanchored molecule CDw108. J. Immunol. 162, 4094–4100. Yu, H. H., and Kolodkin, A. L. (1999). Semaphorin signaling: a little less per-plexin. Neuron 22, 11–14.


HLA-G Molecules: from Maternal–Fetal Tolerance to Tissue Acceptance ¨ L LE MAOULT,* EDGARDO D. CAROSELLA,*,1 PHILIPPE MOREAU,* JOE MAGALI LE DISCORDE,* JEAN DAUSSET,y AND NATHALIE ROUAS-FREISS*

*Service de Recherches en He´mato-Immunologie, Direction des Sciences du Vivant, De´partement de Recherche Me´dicale, CEA Commissariat a` l’Energie Atomique, Institut Universitaire d’He´matologie, Hoˆpital Saint-Louis, 75010 Paris, France; and yFondation Jean Dausset, CEPH, Paris, France

Over the past few years, HLA-G, the non-classical HLA class I molecule, has been the center of investigations that have led to the description of its specific structural and functional properties. Although located in the HLA class I region of chromosome six, the HLA-G gene may be distinguished from other HLA class I genes by its low polymorphism and alternative splicing that generates seven HLA-G proteins, whose tissue-distribution is restricted to normal fetal and adult tissues that display a tolerogeneic function toward both innate and acquired immune cells. We review these points, with special emphasis on the role of HLA-G in human pathologies, such as cancer, viral infection, and inflammatory diseases, as well as in organ transplantation.

I. Introduction

Transcription of the non-classical human leukocyte antigen (HLA) HLA-G gene (Geraghty et al., 1987) and expression of the HLA-G protein (Kovats et al., 1990; McMaster et al., 1995) was initially described as restricted to the fetal–maternal interface on the extravillous cytotrophoblast. The presence of HLA-G at this immunologically privileged site was first proposed as a mechanism used by the fetal semi-allograft to avoid rejection by the mother’s immune system (Rouas-Freiss et al., 1997a, 1999). In this chapter, we review evidence supporting the ability of HLA-G to suppress immune cell functions, such as natural killer (NK) cell- and cytotoxic T lymphocyte (CTL)-mediated cytolysis, the T cell proliferative response, NK transendothelial migration, and dendritic cell (DC) maturation. HLA-G has been also shown to induce NK and T cell apoptosis, as well as a shift toward the Th2 cytokine profile (Carosella et al., 2000; Contini et al., 2003). These multiple suppressive effects 1 Service de Recherche en He´mato-Immunologie, CEA-DSV-DRM, Hoˆpital Saint Louis, Institut Universitaire d’He´matologie, 1, avenue Claude Vellefaux, 75475 Paris cedex 10, France. Tel.: þ 33(0) 1 53 72 22 27; fax: þ 33 (0) 1 48 03 19 60; E-mail: [email protected]


200

E. D. CAROSELLA ET AL.

FIG. 1. Schematic Representation of alternative splicing of HLA-G transcripts and protein isoforms. Exon 1 (E1) encodes leader peptide, exons 2, 3, and 4 (E2, E3, and E4) encode the 1, 2, and 3 extracellular domains, respectively, exon 5 (E5) encodes the transmembrane region, and exon 6 (E6) encodes the reduced cytoplasmic domain of the HLA-G protein. Translation of HLA-G1, -G2, and -G4 transcripts give rise to membrane-bound forms of HLA-G proteins. Intron 4 is retained in both HLA-G5 and HLA-G6 transcripts, and intron 2 is retained in HLA-G7, thus generating soluble forms of HLA-G proteins. In these introns, open reading frames yield to a 21amino acid-specific tail for both HLA-G5, and HLA-G6 proteins, and a 2-amino acid-specific tail for the HLA-G7 isoform.

in both innate and adaptive immunity are mediated by the ligation of HLA-G to several receptors, including ILT2, ILT4, KIR2DL4/p49, and CD8. The gene structure of HLA-G is highly homologous when compared to that of the other HLA class I genes (Ellis et al., 1986), analysis of its transcription has led to the identification of specific mechanisms of alternative mRNA splicing. The HLA-G primary transcript has been shown to generate seven alternative mRNAs able to encode four membrane-bound (HLA-G1, G2, G3, and G4) and three soluble (HLA-G5, G6, and G7) protein isoforms (Fig. 1) (Ishitani, 1992; Kirszenbaum et al., 1994; Paul et al., 2000a). We also review evidence that over the past few years, the tissue distribution of these HLA-G molecules has been found to be broader than originally thought, particularly in pathological processes. While the HLA-G gene is

HLA-G MOLECULES

201

transcribed in almost all tissues, HLA-G transcripts are translated into proteins in only a few non-pathological tissues. Indeed, HLA-G molecules are detected in oocytes and pre-implantation embryos (Jurisicova et al., 1996b; Menicucci et al., 1999; Fuzzi et al., 2002) and in amniotic cells and fluid (Houlihan et al., 1995; Hammer et al., 1997a; McMaster et al., 1998). In adults, thymus is the only non-pathological tissue in which HLA-G has been detected (Crisa et al., 1997; Mallet et al., 1999a). In contrast, HLA-G mRNA and protein upregulation have been described under pathological conditions in peripheral tissues, such as grafted organs, skin and muscle fiber inflammation, after viral infection, and in malignancies. We discuss ectopic HLA-G expression both with respect to its ‘‘positive’’ role in allowing better acceptance of organ grafts, analogous to its role in protecting the fetal semiallograft from maternal immune recognition, and to its ‘‘negative’’ role, when expressed by malignant or virus-infected cells, whose escape from the host’s immunosurveillance it favors. In inflammatory diseases, HLA-G may constitute a tissue-protective molecule against inflammatory aggression (Carosella et al., 2001). We also discuss the fact that besides the immunosuppressive properties exhibited by HLA-G, another possible function is acting as an antigen peptide-presenting molecule capable of being recognized by antigen-specific T cells.

II. The HLA-G Gene and Polymorphism

The HLA-G gene is located within the class I MHC locus, on the short arm of chromosome 6, approximately 230 kb telomeric from HLA-A and 100 kb centromeric to HLA-F. The overall HLA-G gene structure, previously designed HLA-6.0, was described in 1987 (Geraghty et al., 1987) and found to be highly homologous to that of classical HLA-class I genes. It presents 86% similarity with the consensus sequence of the HLA-A, -B, and -C genes and consists of eight exons and seven introns: Exon 1 encodes the signal peptide, exons 2, 3, and 4 respectively encode the 1, 2 and 3 extracellular domains, and exon 5 encodes the transmembrane domain. Exon 6 has a stop codon in its second codon, which results in a shorter cytoplasmic tail region, compared to classical HLA-class I gene products. Evolutionary trees obtained using complete genomic DNA sequences show that HLA-A, G, H, and J form a group of HLA-A-related loci that are readily distinguishable from HLA-B, C, E, and F (Messer et al., 1992). Recent data have also accumulated that show that the HLA-G gene is not confined to the human species and that the HLA-G gene and HLA-G protein present limited polymorphism, which however, may be relevant in the regulation of HLA-G function.

202


A. HLA-G GENE HOMOLOGS

IN

NONHUMAN PRIMATES

No HLA-G gene homolog has been found in mice. Nevertheless, nonhuman primates that are phylogenetically close to humans express homologs of HLA-G. It has been shown that the cotton-tamarin (Saguinus oedipus—Saoe), which diverged from human lineage approximately 40 million years ago, expressed an HLA-G-‘‘like’’ gene (Watkins et al., 1990). More recently, data have shown that both New World and Old World monkeys have MHC-G genes (Corell et al., 1994; Arnaiz-Villena et al., 1997, 1999; Slukvin et al., 2000; Arnaiz-Villena et al., 2001). In monkeys, the degree of MHC-G gene polymorphism varies according to the species, being high in Saoe and low in Pongidae (Arnaiz-Villena et al., 2001). The MHC-G molecule in monkeys may exhibit a truncated cytoplasmic domain (Arnaiz-Villena et al., 2001), and alternative splicing of the primary transcript of MHC-G genes has also been reported (Castro et al., 2000b; Langat et al., 2002). Notably, a transcript retaining intron 4 and encoding a soluble Mamu-AG isoform (Manu-AG5) has been described in the rhesus monkey placenta, thereby confirming a homology between HLA-G in the human and Mamu-AG in the rhesus (Ryan et al., 2002). B. HLA-G GENE POLYMORPHISM Studies have converged on the existence of a low level of HLA-G allelic polymorphism compared with classical HLA-class I genes, and the absence of HLA-G gene imprinting (Kirszenbaum et al., 1997; Hiby et al., 1999a). To date, 15 HLA-G alleles have been assigned in the WHO nomenclature (Table I), including a null allele, known as HLA-G* 0105N. Among the ‘‘normal’’ alleles, the first allele that was sequenced, G* 01011 (recently renamed G* 010101), predominates in almost all populations studied to date (Asian, European, and African). Its frequency varies from 32% in the German/Croatian population (van der Ven et al., 1998) to 83% in the African Ghanaian population (Matte et al., 2000). The G* 01011 (G* 010101) and G* 01012 (G* 010102) alleles have been found with comparable frequencies among the German/Croatian (32% and 36%, respectively) (van der Ven et al., 1998) and Portuguese (37% and 31%, respectively) (Alvarez et al., 1999) populations. Nevertheless, the high frequency of the G* 01012 (G* 010102) allele seems to be confined to Caucasian populations and was found to be lower than that of the G* 01041 (G* 010401) allele, in both the Japanese (38%) and African (9.5–20.4%) populations (Alvarez et al., 1999). Some HLA-G alleles are found frequently only in certain specific populations. For example, G* 01013 (G* 010103) and G* 01018 (G* 010108) are found at frequencies of 17% in the Portuguese and 14.4% in the Shona population of Africa, respectively (Alvarez et al., 1999; Matte et al., 2000). Other alleles, such as

HLA-G ALLELES Alleles

Exon 2

Exon 3

30 UT

Exon 4

31

54

57

69

93

107

110

130

188

241

258

14 bp

ACG(T) ACG ACG ACG ACG ACG ACG ACG ACG TCG(S) ACG ACG ACG ACG ACG

CAG(Q) CAG CAG CAG CAG CAG CAG CAG CGG(R) CAG CAG CAG CAG CAG CAG

CCG(P) CCA CCA CCG CCG CCG CCA CCA CCG CCG CCA CCC CCG CCA CCA

GCC(A) GCC GCC GCT GCC GCC GCC GCC GCC GCC GCC GCC GCC GCC GCC

CAC(H) CAT CAC CAC CAC CAC CAT CAC CAC CAC CAC CAC CAC CAT CAT

GGA(G) GGA GGT GGA GGT GGA GGT GGA GGA GGA GGA GGA GGA GGA GGA

CTC(L) CTC CTC CTC CTC CTC CTC CTC CTC CTC ATC(I) ATC(I) ATC(I) CTC CTC

CTG(L) CTG CTG CTG CTG CTG CTG CTG CTG CTG CTG CTG CTG .TG(f) CTG

CAC(H) CAC CAC n.d. CAC CAT CAC CAC CAC n.d. CAC CAT CAC

TTC(F) TTC TCC(S) n.d. TTC TTC TTC TTC TTC n.d. TTC TTC TTC

ACG(T) ACG ACG n.d. ACG ACG ACG ACG ACG n.d. ACG ACG ACG

þ þ

CAC

TTC

ATG(M)

þ

HLA-G MOLECULES

G* 010101 G* 010102 G* 010103 G* 010104 G* 010105 G* 010106 G* 010107 G* 010108 G* 0102 G* 0103 G* 010401 G* 010402 G* 010403 G* 0105N G* 0106

TABLE I PROTEIN POLYMORPHISM

AND

þ þ

A, Ala; F, Phe; G, Gly; H, His; I, Ile; L, Leu; M, Met; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; f, frameshift.

203

204


G* 01015 (G* 010105), G* 01016 (G* 010106), G* 01017 (G* 010107), G* 0102, and G* 01043 (G* 010403) are very rare. Furthermore, the newly described G* 0106 allele (Hviid et al., 2001), which has not yet been extensively studied, was found at a frequency of 4% among Caucasians. Nevertheless, the existence of this allele was not confirmed by the studies of Geraghty’s group presented during the thirteenth Histocompatibility Workshop (Victoria, BC, Canada; May, 2002), and thus remains to be clarified. The G* 0105N null allele (Arnaiz-Villena et al., 1997) has been found at frequencies of 11% in African Shona, 7.4–8.3% among African-Americans, 7.3% in Nigerian and 5.3–6.3% in Cameroon populations, 4.8% among Ghanaians, 3% in Spaniards (in association with the HLA-A30-B13 haplotype), 2.9% among Mexican-Americans, 2.5% in Sardinians, 2.3% in mixed German-Croatian populations, 0.6% in Denmark, and absent in UK/USEuropean, Portuguese, and Japanese populations (Hviid et al., 1997; Suarez et al., 1997; Ober et al., 1998; van der Ven et al., 1998; Ishitani et al., 1999; Matte et al., 2000; Aldrich et al., 2002). This allele may have arisen in Africa and spread to Europe after being introduced in Spain on an HLA-A30 haplotype by Moorish conquerors during the eighth century, AD (Suarez et al., 1997). C. HLA-G PROTEIN POLYMORPHISM Although HLA-G alleles are essentially defined by non-synonymous substitutions in comparison with G* 01011 (G* 010101), single amino acid changes characterize six HLA-G proteins (Table I). They are localized in exon 2 at codon 31 (Thr ! Ser) and codon 54 (Gln ! Arg), in exon 3 at codon 110 (Leu ! Ile), and in exon 4 at codon 241 (Phe ! Ser) and codon 258 (Thr ! Met). Another HLA-G protein variant may be encoded by the G* 0105N allele, designated the ‘‘HLA-G null allele.’’ This allele has the same DNA sequence as G* 01012 (G* 010102), except for a cytosine deletion (1597C) at the third position of codon 129 or the first position of codon 130, causing a frameshift mutation. Consequently, all amino acids encoded by the second half of exon 3 are different, and a stop codon generated at the beginning of exon 4 gives rise to a truncated HLA-G protein (Ober et al., 1998; Castro et al., 2000c). The reason for the unusually high frequency of the 1597C mutation in African populations is unknown, but it has been suggested that it is a result of natural selection pressure (Aldrich et al., 2002). One hypothesis is that intrauterine pathogens may be the selective agents in populations from areas with a history of high pathogen load. Aldrich et al. hypothesized that reduced HLA-G1 expression in heterozygous placentas may result in an overall increase in the number of T cells available in the uterus to fight intrauterine infections. On the other hand, the existence of healthy individuals who are

205

HLA-G MOLECULES

homozygous for the G* 0105N allele raised doubts about the role of HLA-G1 in the maintenance of pregnancy (Ober et al., 1998). Although debated, it has been suggested that the HLA-G* 0105N allele can still produce the HLA-G2, G3, G6, and G7 isoforms that could substitute for HLA-G1 in fetal–maternal tolerance (Riteau et al., 2001c). Accordingly, the corresponding transcripts are produced in homozygous G* 0105N individuals, and HLA-G2 and G3 protein isoforms have been shown to inhibit NK function in vitro (Castro et al., 2000c; Riteau et al., 2001c). D. POLYMORPHISM

IN THE

NON-CODING REGION

OF THE

HLA-G GENE

Recent work has focused on HLA-G gene polymorphism in non-coding regions. First, a 14-bp deletion/insertion polymorphism has been localized in the 30 UT region of the HLA-G gene at respective frequencies of 58% for deletion and 42% for insertion (Harrison et al., 1993). HLA-G alleles containing a 14-bp insertion may undergo further alternative splicing, such that 92 bases of the 30 UT region are spliced out of a part of the transcripts (Hiby et al., 1999a). Analysis of this polymorphism in several other primates has shown that 14-bp polymorphism must be very recent, since the deletion is present in some human MHC-G alleles but not in other primate species. This suggests that the ‘‘long’’ MHC-G alleles are the oldest (Castro et al., 2000a). Such polymorphism may have functional consequences. One hypothesis is that the 92-base-pair deletion might influence the stability of the HLA-G transcripts. On the other hand, it is observed that the 14-bp polymorphism, in association with the T mutation at the third base of codon 93 (exon 3), may be associated with lower HLA-G transcript levels and be involved in the variation in alternative splicing profiles (for example, lack of HLA-G3 in pre-eclampsia) (O’Brien et al., 2001). Two studies have reported low polymorphism in the 1.4-kb promoter region of the HLA-G gene. One of the interesting polymorphisms is demonstrated in the Danish samples, located near the B2 site at position –201 from the gene, changing a G in the G* 01011 (G* 010101) allele to an A in the G* 01012 (G* 010102), G* 01013 (G* 010103), G* 0104, and G* 0105N alleles (Hviid et al., 1999). The other one is a nucleotide substitution (C ! T) at position –56 found in indigenous East-African populations and located in the putative binding site for polyomavirus enhancer-binding protein 2 (PEBP2) (Matte et al., 2002). The authors inferred that these polymorphisms might have an effect on the regulation of expression of some HLA-G alleles, a theory that requires further investigation. In conclusion, there is convergent data showing that in humans, the HLA-G gene exhibits little polymorphism, a property that contributes to supporting the HLA-G function during pregnancy. In some cases, the inheritance of specific HLA-G alleles, as well as the appearance of somatic mutations in the

206


HLA-G gene, may have an impact on the modulation of HLA-G expression, consequently, on its function. Indeed, Rebmann et al. demonstrated the association of HLA-G alleles with soluble HLA-G levels in the plasma of 94 unrelated healthy individuals (Rebmann et al., 2001). High levels of soluble HLA-G were found to be associated with the presence of the HLA-G* 01041 (G* 010401) allele. Mean soluble HLA-G levels in individuals with the most common HLA-G alleles G* 01011 (G* 010101) and G* 01012 (G* 010102) were very similar and were approximately 63% of the levels found in HLAG* 01041 (G* 010401) individuals. In contrast, significantly reduced soluble HLA-G levels were observed in individual bearing G* 01013 (G* 010103) or G* 0105N alleles. In addition, recent data reveal that HLA-G polymorphism may influence some pathologies of pregnancy. Hviid et al. did not find significant differences in the distribution of HLA-G alleles between controls and Recurrent Spontaneous Abortion (RSA) couples. However, 15% of the RSA women studied carried the HLA-G* 0106 allele, compared to 2% of the control women (Hviid et al., 2002). Carreiras et al. recently showed that the presence of the HLA-G* 0104 and DRB1* 07/06 alleles, HCMV sequences, and the maternal inheritance of G* 0104 should be considered conditioning factors for the development of pre-eclampsia (Carreiras et al., 2002). Moreover Aikhionbare et al. (2001) demonstrated that mother-to-child discordance in exon 2 of HLA-G is associated with a reduced risk of perinatal HIV-1 transmission. Considering that HLA-G may be induced in some tumors (Cabestre et al., 1999) as well as in certain allogenic situations (Lila et al., 2001), identification of the HLA-G allelic forms involved should be taken into account. III. Regulation of HLA-G Gene Expression

HLA-G protein expression in non-pathological situations is very restricted. HLA-G is secreted by blastocysts (Fuzzi et al., 2002) and displays local cell surface expression in extravillous cytotrophoblasts, amnionic epithelial cells, chorionic villi endothelial cells, and adult thymic epithelial cells (McMaster et al., 1995; Blaschitz et al., 1997; Crisa et al., 1997; Hammer et al., 1997b). HLA-G is also activated in Human Cytomegalovirus (HCMV)- and HIV-1infected cells (Onno et al., 2000b; Lozano et al., 2002), tumors (Paul et al., 1999; Ugurel et al., 2001; Urosevic et al., 2001; Ibrahim et al., 2001b; Lefebvre et al., 2002; Ugurel et al., 2002; Urosevic et al., 2002a,b; Wiendl et al., 2002), inflammatory pathologies (Aractingi et al., 2001; Khosrotehrani et al., 2001), transplanted heart, and during the mixed lymphocyte reaction (Lila et al., 2000, 2001, 2002). It is usually reported that HLA-G protein expression correlates with high transcriptional activity, whereas HLA-G

207

HLA-G MOLECULES

transcript levels are generally low or absent in situations in which the protein is absent (Onno et al., 1994). Notably, using RT-PCR analysis, the absence of HLA-G transcripts has been demonstrated in numerous cell-lines (Frumento et al., 1999, 2000), in CD34 þ and NK hematopoietic cells, and in firsttrimester fetal liver (Kirszenbaum et al., 1994, 1995; Amiot et al., 1996; Onno et al., 1997; Moreau et al., 1998). In spite of the fact that HLA-G mRNA accumulates during trophoblast cell differentiation prior to protein synthesis, suggesting tight post-transcriptional control (Copeman et al., 2000), apparently an important part of the regulation of HLA-G expression takes place at the transcriptional level. A. REGULATORY SEQUENCES IN THE PROMOTERS HLA-CLASS I AND HLA-G GENES

OF

CLASSICAL

In contrast to the promoter of classical HLA-class I genes, the HLA-G gene promoter is atypical since almost all conserved cis-regulatory elements are disrupted (van den Elsen et al., 1998a). They comprise two modules located 220 bp from the gene initiation codon (Fig. 2). 1. The Upstream Module of Classical HLA-Class I Gene Promoters The upstream module of classical HLA-class I gene promoters contains (i) the enhancer A, consisting of the B2 and B1 sites, which bind the NFB/rel and Sp1 DNA-binding proteins; and (ii) the interferon-stimulated response element (ISRE), which is a binding site for interferon-induced transcription factors such as interferon regulatory factor (IRF)-1 and interferon-stimulated gene factor (ISGF)-3, as well as repressors of HLA class I transcription, such as IRF-2 and interferon consensus sequence-binding protein (ICSBP) (Gobin et al., 1998a, 1999b). In the HLA-G promoter, enhancer A is unresponsive to NF-B, and the two B sites display binding affinity only for the P50 subunit of NF-B, as well as the Sp1 factor for B2. ISRE is partially deleted in the HLA-G promoter, thus has no binding affinity for proteins of the IRF1 family, and cannot mediate interferon-induced HLA-G expression (Gobin et al., 1998a, 1999b; Gobin and van den Elsen, 2000). 2. The Downstream SXY Module of Classical HLA-Class I Gene Promoters The downstream SXY module of the classical HLA-class I gene promoters has been identified by sequence comparison of HLA class I and HLA-class II promoters and the reduced level of MHC class I expression observed, in addition to the lack of HLA class II expression in the type III Bare Lymphocyte Syndrome (BLS) (van den Elsen et al., 1998b; Masternak et al., 2000). There is increasing evidence that this module is critical for constitutive and inducible expression of both HLA class I and HLA class II genes

208


FIG. 2. Schematic representation of the HLA-G gene promoter. Numbers indicate location of regulatory boxes relative to ATG (bp). LCR?: putative locus control region; CRE: three functional CRE/TRE sites bound by CREB1, ATF1 and c-Jun. ISRE: interferon sequence responsive element. Note that only one ISRE site is functional; HSE: heat shock element that bind Heat Shock Factor 1; GAS: nonfunctional Interferon- activated site. B2 and B1 are referred to enhancer A within classical HLA class I promoter. They are disrupted within the HLA-G promoter and display affinity for P50 a subunit of NF-B; The conserved X1 half of X box associates to RFX and Sp1 in vitro; ‘‘?’’ indicates that RFX member factor is not yet identified in vivo. X2 and Y boxes are mutated thus avoiding CIITA induced transactivation of HLA-G gene.

(Gobin et al., 1998b, 2001), and consists of four cis-acting elements (Waldburger et al., 2000), referred to as the ‘‘S box’’ (also called W or Z), the X box, comprising the X1 and X2 (also called site) halves, and the Y box, which is an inverted CCAAT box (also called enhancer B). The function of the S box is not fully understood; it could possibly play a role in promoter architecture, rather than as a transcription factor-binding site. The X1 box is the binding site for the ubiquitous regulatory factor X (RFX) complex, consisting of the three subunits RFX5, RFXB/ANK, and RFXAP. The X2 box is bound by the X2BP-CREB/ATF factor, and the Y box by a heterotrimer, the NF-Y factor (also referred to as CBF and CP1). Together, these factors are highly cooperative and are the results of a stable, higher-order enhanceosome

HLA-G MOLECULES

209

complex. HLA-class II expression requires the presence of the multi-protein complex in order to interact with the master co-activator CIITA, which is constitutively expressed only in APCs of the immune system, but which may be inducible in other cell types by IFN- (Waldburger et al., 2000; Masternak and Reith, 2002). Similar to HLA class II genes, the HLA class I gene promoter is transactivated by CIITA (Gobin et al., 1997a, 1998b; Lefebvre et al., 1999a). The S and X1 boxes are the only conserved motifs in the HLA-G gene promoter, suggesting a potential role for them in the regulation of this gene. The X2 box mutation was shown to affect the binding of the cAMP response element binding protein/activation transcription factor (CREB/ATF) family of factors in vitro (Rousseau et al., 2000), and the absence of an intact SXY module in the HLA-G promoter has been demonstrated to prevent CIITA recruitment in vivo (unpublished), and consequently, in CIITA-mediated activation of the gene (Gobin and van den Elsen, 2000). Although the conserved X1 box of the HLA-G promoter is a potential target for RFX5 and Sp1 factors, as revealed by Electrophoretic Mobility Shift Assay (EMSA) (Rousseau et al., 2000) chromatin immunoprecipitation experiments (ChIP) reveal this not to be the case in vivo (unpublished). Therefore, in the absence of regulatory pathways shared between the HLA-G and HLA-class I and class II genes, several groups have investigated other elements that may control HLA-G gene transcription. B. SPECIFIC REGULATORY SEQUENCES

OF

HLA-G GENE

One strategy for analyzing the potential site of HLA-G gene regulation has been the use of transient transfection experiments carried out on the HLA-Gpositive JEG-3 choriocarcinoma cell line with luciferase reporter constructs containing HLA-G promoter fragments of different lengths. This allowed the detection of a sole negative regulatory sequence in the region 450 to 220 bp upstream from the first exon of HLA-G gene. The search for other regulatory sequences outside the HLA-G gene promoter region yielded the detection of weakly induced activity in intron 2 (Gobin et al., 1999a). Another strategy was achieved in vivo, using HLA-G transgenic mice into which HLA-G transgenes of different lengths were introduced. This made it possible to show that a distal upstream regulatory element contained in a 244-bp HindIII/EcoR1 fragment located just over 1.1 kb from the first exon of the HLA-G gene was necessary to direct specific HLA-G transcription in transgenic mouse placenta and resulted in similar levels of extra embryonic HLA-G mRNA as those seen in the human (Schmidt et al., 1993; Yelavarthi et al., 1993; Schmidt et al., 1995). The authors suggested that the 244-bp fragment might contain a control locus region, but such a role remains to be clarified, because of the absence of correlation with results obtained with

210


in vitro transient transfection experiments. However, several data support the contention that this region participates in the control of HLA-G gene transcription (Fig. 2). Indeed, EMSA carried out with the 244-bp HindIII/ EcoRI region incubated with nuclear extracts from cells and tissues exhibiting positive (JEG-3 and PBMC) and negative (YT2C2 and first-trimester fetal liver) HLA-G transcription revealed shared DNA/protein complexes (C2) and specific complexes (C1, C3, C5, C6, and C7), according to the HLA-G expression status (Moreau et al., 1997, 1998). More recently, computer-aided search within the 1438-bp fragment of the HLA-G gene promoter permitted the localization of three functional cAMP response element/TPA-response element (CRE/TRE) elements (CRE1380, CRE930, and CRE770) shown to have a binding affinity for CREB/ATF and Fos/jun proteins. Interestingly, in vivo binding of CREB-1 and c-Jun has been demonstrated around the CRE site located at position 1380 within the 244-bp HindIII/EcoR1 fragment (Gobin et al., 2002). In that case, the binding did not account for the tissue-specific HLA-G transcriptional activity, since it was also observed in the Tera-2 and Raji HLA-G-negative cell lines. This complex might therefore correspond to the C2 complex resulting from the EMSA analysis, since the binding was previously localized between 1438 and 1311, and was not tissue-specific. HLA-G transactivation by CREB-1 via three binding sites appears to be an alternative pathway to the conserved HLA-class I regulatory routes. Interestingly, the in vitro activation of HLA-G promoter by CREB has been shown to be strongly inhibited by inducible cAMP early repressor (ICER), whereas activation by CREB was enhanced by overexpression of the coactivators CBP and p300. In addition, the in situ nuclear localization of CREB/ATF and CREB binding protein (CBP) in extravillous cytotrophoblast strongly supports the contention these transcription factors play a role in the HLA-G transactivation observed in trophoblast cells (Gobin et al., 2002). On the other hand, several groups have investigated the impact of cytokines, stress, and hormones on the modulation of HLA-G gene expression. Despite the absence of a classical ISRE site, an increase in HLA-G protein and HLA-G mRNA after IFN treatment has been described in different studies, suggesting that another HLA-G specific sequence could be involved (Yang et al., 1996; Chu et al., 1999b; Lefebvre et al., 2001). The question is still being debated, since promoter activity of the region encompassing 1.4 kb from ATG appears to be modestly affected by IFN treatment, compared with the classical HLA class-I gene (Gobin et al., 2002). Nevertheless, although a candidate IFN- activated site (GAS) located at position 740 in the HLA-G promoter was shown to be nonfunctional (Chu et al., 1999a), an ISRE sequence located at position 746 is capable of transactivating HLA-G

HLA-G MOLECULES

211

following IFN- treatment (Lefebvre et al., 2001). HLA-G promoter activity was weakly induced in the JEG-3 cell line, but reproducible and clearly significant in the thymic epithelial cell LT-TEC2 line, with an average of 2.8-fold enhancement. Moreover, using EMSA, this work also provided evidence for the binding of IRF-1 to the functional HLA-G ISRE in response to IFN- (Lefebvre et al., 2001). The HLA-G gene may also be activated by stress conditions. Application of heat shock at 42 C or arsenite treatment for 2 h, reverses HLA-G gene repression in the M8 melanoma cell-line without affecting other MHC class I genes. Interestingly, HLA-G6 expression is induced prior to that of the other HLA-G transcripts early during the recovery time after stress treatment, which also indicates tight control of HLA-G alternative splicing. The study also identifies a heat shock element (HSE) located between positions 487 and 466 of the HLA-G gene promoter that binds to heat shock factor 1 (HSF1) in vitro under stress conditions (Ibrahim et al., 2000) (Fig. 2). C. EPIGENETIC MECHANISMS

MAY

REGULATE HLA-G GENE EXPRESSION

DNA methylation and histone modification are interrelated mechanisms known to play a key role in transcriptional control (El-Osta and Wolffe, 2000; Bird, 2002). CpG methylation has been analyzed in the JAR choriocarcinoma cell-line, revealing the activation of HLA-G transcription after treatment with demethylating agent 5-azacytidine. Conversely, no correlation was observed between HLA-G gene transcriptional activity and methylation of CpG islands in the 50 -part of the HLA-G gene in cells that either express HLA-G transcripts (trophoblasts, JEG-3 cells and CD2 þ lymphocytes) or that do not (syncytiotrophoblasts and CD34 þ cells) (Onno et al., 1997). However, these studies are limited to the proximal promoter region of the HLA-G gene, and questions remain concerning potential subsets of CpG sites that might be important in HLA-G gene silencing. A trans-acting demethylating process could also be envisaged that reverses repression of specific transcriptional factors, which in turn would activate HLA-G transcription. Using cell lines negative for HLA-G transcription that exhibits various phenotypes (B-EBV and tumor cell), recent data show that repression of the HLA-G gene by DNA methylation is a more general mechanism than expected. Indeed, although cell exposure to histone deacetylase inhibitors only activated HLA-G transcription in melanoma cells, cell exposure to the demethylating agent 5-aza-20 -deoxycytidine reversed HLA-G gene silencing in all cell-lines studied. In particular, strong HLA-G repression in Raji cells is reversed by 5-aza-20 -deoxycytidine treatment, giving rise to both HLA-G transcription and protein expression (Moreau et al., 2003). Therefore, despite

212


in vivo binding of CREB1 factors to the HLA-G gene promoter in these cells (Gobin et al., 2002), methylation represses HLA-G gene transcription. Whether cis- or trans-acting mechanisms occur and involves particular methyl-binding or cofactors must be elucidated. D. HLA-G CAN BE UPREGULATED BY SOLUBLE MEDIATORS PLACENTAL OR TUMOR MICROENVIRONMENTS

OF

Several groups have investigated the effects of factors secreted locally in placental or tumor microenvironments on HLA-G transcriptional activity and HLA-G cell-surface expression. IL-10 is produced in placenta during all stages of gestation, as well as in tumor cells and may be secreted at high levels in the serum of cancer patients (Dummer et al., 1995; Roth et al., 1996; Bennett et al., 1997; Yue et al., 1997; Urosevic et al., 2002b). Serum IL-10 has also been shown to be elevated in patients with orthotopic liver transplants demonstrating non-acute rejection (Minguela et al., 1999). IL-10 exhibits a broad spectrum of biological activities, including anti-inflammatory and immunosuppressive ones (Goldman and Stordeur, 1997). This interleukin has been shown to selectively upregulate HLA-G transcription in cultured trophoblast explants, and to enhance both HLA-G transcription and cellsurface expression of HLA-G protein in monocytes (Moreau et al., 1999). Accordingly, CMV produces a biologically active IL-10 homolog that can upregulate HLA-G protein expression at the monocyte cell surface (Spencer et al., 2002). Glucocorticoid hormones are also widely expressed in placental tissues (Sun et al., 1998) and have been shown to upregulate HLA-G transcript levels in trophoblast explants treated with dexamethasone and hydrocortisone (Moreau et al., 2001). Transactivation of HLA-G transcription has also been demonstrated by JEG-3 cell exposure to Leukemia Inhibitory Factor (LIF) (Bamberger et al., 2000), a pleiotropic cytokine shown to be expressed at the maternal–fetal interface and which plays an essential role in implantation in mice (Stewart et al., 1992). Moreover, Granulocyte-Macrophage Colony Stimulating Factor (GM-CSF), in combination with IFN- , have been shown to stimulate HLA-G cell-surface expression on the U937 monohistiocytic cell line (Onno et al., 2000a). The mechanisms, by which IL-10, glucocorticoid, LIF, and GMCSF stimulate HLA-G transcription, or both HLA-G transcription and HLA-G protein expression, are still generally unknown, and some of them could involve both transcriptional and posttranscriptional processes. It has notably been observed that these latter factors generally have no effect on HLA-G gene transcription in cells in which the HLA-G gene is strongly repressed (Frumento et al., 2000). This reinforces the hypothesis of

HLA-G MOLECULES

213

the existence of epigenetic mechanisms that may control HLA-G gene activation. In conclusion, evolution has developed very specific and multifactorial pathways for the transcriptional regulation of the HLA-G gene, which, in combination with post-transcriptional mechanisms (mRNA stability and protein processing) contribute to the temporal regulation that occurs in placenta (Solier et al., 2001), and to the constitutive or inducible expression of the HLA-G molecule in other tissues. On the other hand, HLA-G expression may be upregulated in monocytes and T lymphocytes of HIV þ patients, in activated macrophages infected with Human Cytomegalovirus (HCMV), probably due to viral IL-10-like protein, and in HCMV-infected U-373 MG astrocytoma cells, where IE-pp72 and IE-pp86 HCMV proteins are involved (Onno et al., 2000b; Lozano et al., 2002). Therefore, some viruses seem to have developed a strategy for controlling HLA-G expression that consist in taking advantage of its specific immune tolerance function. IV. Processing and Transport of HLA-G Molecules

Folding, assembly, and peptide loading of HLA-class I molecules are critical for the regulation of their cell-surface expression, and involve intracellular protein complexes and chaperone molecules (Neefjes and Momburg, 1993). Schematically, cytosol and nuclear proteins are degraded by the proteasome, and then transported into the lumen of the endoplasmic reticulum (ER) by a heterodimeric peptide transporter, which consists of TAP1 and TAP2. At the same time, the class I chain and 2-microglobulin are assembled in the ER with the help of calnexin and calreticulin. The peptides are then loaded into the groove of the HLA class I heavy-chain, which in turn associates 2-microglobulin. The glycosylation pattern of the class I heavy-chain is modified once the complex has exited the lumen of the ER and the heterotrimer is transported through the golgi apparatus to the cell surface. Due to alternative splicing of the HLA-G primary transcript, HLA-G presents seven isoforms (Fig. 1) (Ishitani and Geraghty, 1992; Fujii et al., 1994; Kirszenbaum et al., 1994; Moreau et al., 1995; Paul et al., 2000a): HLA-G1 is the complete membrane-bound protein; HLA-G2, -G3, and -G4 are membrane-bound isoforms respectively devoid of the 2, 23, and 3 extracellular domains. The intracytoplasmic domains of these isoforms are all reduced, in comparison with those of classical HLA-class I molecules. In addition, the existence of transcripts in which intron 4 or 2 are retained generates three soluble isoforms, known as HLA-G5 (G1-like), -G6 (G2-like), and -G7 (G3-like), respectively (Fig. 1). The use of viruses that affect classical HLA class I protein expression, in combination with biochemical approaches, has revealed common pathways in

214


the assembly and transport of HLA-G and HLA class I molecules (Schust et al., 1999). Nevertheless, the non-classical structural properties of HLA-G isoforms suggest the existence of specific HLA-G trafficking pathways. Most of them have been investigated focusing on HLA-G1 and HLA-G5, for which suitable reagents were available. A. HLA-G BINDS INTRACELLULAR PEPTIDES MECHANISM

WITH A

TAP-DEPENDENT

The question about the ability of the HLA-G molecule to present a peptide was first investigated in comparison with the HLA-A2 molecule. Similarities found in the 1 and 2 domains, which form the peptide pocket, suggested that some peptides presented by HLA-A2 might also be presented by HLA-G1 (Geraghty et al., 1987). Biochemical analysis of the HLA-G1 (membrane-bound) and the HLA-G5 (soluble) forms of the molecule has demonstrated that both proteins, which are expressed in the LCL721.221 lymphoblastoid cell line, consisted of heavy-chain/beta-2-microglobulin/ peptide in a 1 : 1 : 1 ratio. Peptide elution experiments confirmed that the peptides presented by HLA-G1 and HLA-G5 were of the same size as those carried by HLA-A2, and consisted of nine amino acids with the XI/ LPXXXXXL consensus sequence (Lee et al., 1995; Diehl et al., 1996). In accordance with the observed sequence similarities between HLA-G1 and HLA-A2, anchoring positions are located on amino acids 2 (isoleucine or leucine), 3 (proline), and 9 (leucine), although positions 2 and 9 are enough for efficient anchoring (Lee et al., 1995). In transfected cells, the diversity of peptides eluted from HLA-G1 and HLA-G5 was estimated to be around five-fold lower than with classical HLA class I, a property that may be related to the HLA-G molecule’s invariance in the peptide-binding groove region (Lee et al., 1995). On the other hand, in the absence of data obtained with cytotrophoblasts and thymic epithelial cells, the limited peptide diversity could be different from the repertoire of individual peptides bound in vivo. However, it was suggested that HLA-G could bind peptides derived from viruses, a limited number of which infect trophoblasts, and may be presented in order to allow HLA-G anti-viral function (Le Bouteiller and Blaschitz, 1999). Notably, one report described that alloreactive HLA-G-specific CD8 þ CTL clones could be obtained in vitro in a peptide-dependent fashion (Horuzsko et al., 1999). However, whether HLA-G can induce HLA-G restricted CTL responses against viral protein remains to be elucidated. Although there is strong evidence that HLA-G1 and HLA-G5 bind peptide, the question is still open as to whether there are peptides that bind to the other HLA-G isoforms, namely HLA-G2, -G3, -G6, and -G7. The 1 and 2 domains, which form the peptide groove, are present only in HLA-G4, which

HLA-G MOLECULES

215

might therefore be a candidate peptide-presenter. Another hypothesis to explore would be the possible formation of HLA-G2, HLA-G4, and HLA-G3 homodimers, which might permit peptide loading similar to HLA-class II antigens (Ishitani and Geraghty, 1992). Is peptide loading of HLA-G1 and HLA-G5 a TAP-dependent mechanism? Studies seeking HLA-G1 and HLA-G5 expression in the transfected TAPnegative B-LCL 721.131 lymphoblastoid cell line showed that HLA-G1 expression is reduced to around 20%, with respect to expression in the B-LCL721.221 TAP-positive cell line, whereas no HLA-G5 protein could be detected in the supernatant of the TAP-negative cells. HLA-G therefore uses both TAP-dependent (HLA-G1 and HLA-G5) and TAP-independent (HLA-G1) pathways to bind peptide (Lee et al., 1995). Accordingly, a 10fold downregulation was observed with a remainder HLA-G1 cell-surface expression at the surface of HLA-G-LCL721.221 cells transfected with the ICP47 protein of the Herpes simplex virus (HSV), a viral protein that has been shown to block the TAP transporter (Munz et al., 1999). In the JEG-3 choriocarcinoma cell-line infected with (HSV) or transfected with ICP47 protein, HLA-G1 fails to acquire endoglycosidase-H (endo-H) resistance, thus is retained in the endoplasmic reticulum (Schust et al., 1996). In agreement with HSV-related experiments, HLA-G1 has been shown to be sensitive to endo-H treatment in JEG-3 cells transfected with human cytomegalovirus (HCMV)-US6, a gene that encodes a glycoprotein that also prevents peptide-loading of the HLA-class I molecule by inhibiting the TAP complex (Jun et al., 2000). The existence of an interaction between the TAP complex and HLA-G1 has been demonstrated by immunoprecipitation studies (Lee et al., 1995; Gobin et al., 1997b). However, the absence of a detectable association between TAP and HLA-G5 has led to the hypothesis of the possible binding of a cytosolic peptide without interaction. This also suggests that either the residues that interact with TAP are located in the cytoplasmic tail of HLA-G1, or that the unique 21-amino acids located in the C-terminal of HLA-G5 interfere with TAP association. The peptide-loading process thus remains to be investigated, and is apparently associated with either the soluble- or membrane-form state of the molecule. Whatever the loading mechanisms involved may be, the following observations strongly argue that the HLA-G heavy-chain requires peptide association for membrane expression: First, peptide-deficient HLA-G1 molecules are rapidly degraded in the ER of US6-expressing HeLa cells (Park et al., 2001) and, second, HLA-G1 cell-surface expression was entirely abolished in cells with both TAP and proteasome deficiencies (unpublished). The loading of HLA-G1 is not strictly TAP-dependent (as is the case for HLA-G5) and may vary according to cell type. This suggests the existence of

216


an alternative pathway for reaching the cell surface, as described for classical HLA-class I molecules. B. SPECIFIC FEATURES OF THE HLA-G MOLECULE INFLUENCE TRANSPORT TO THE CELL SURFACE

ITS

Compared with classical HLA-class I molecules, HLA-G is likely to have evolved as a more resistant antigen, thus facilitating cell traffic and conferring resistance to some virus products: (a) HLA-G molecules produced by cytotrophoblasts and found in amniotic fluid bear unusual carbohydrate structures, leading to a broad molecular weight range of HLA-G-immunoreactive bands; (b) like classical HLA class I molecules, HLA-G contains a single N-linked glycosylation site. However, digestion of HLA-G proteins from placenta by endo- D-galactosidase suggests that these molecules specifically carry N-acetyllactosamine units, which might stabilize the molecule (McMaster et al., 1998); (c) in both JEG-3 and 2A2 porcine bone marrowderived stromal cells, HLA-G1 has been shown to be resistant to dislocation and degradation mediated by HCMV US2 and US11 gene products, sharing this property with HLA-Cw3, -Cw4. In vitro co-transcription/ translation of class I heavy chains with US2 and US11 gene products followed by immunoprecipitation experiments have demonstrated that these proteins do not associate with either HLA-G or HLA-C molecules (Schust et al., 1998). Another striking feature is the shortened cytoplasmic domain of the HLA-G molecule, due to a premature stop codon in exon 6 (Geraghty et al., 1987). This deletes potential endocytosis signals found in the cytoplasmic tails of all HLA-class I molecules. Consequently, spontaneous endocytosis of HLA-G1 was shown to be reduced, compared to classical class I molecules, resulting in the prolonged retention of molecules at the cell-surface (Davis et al., 1997). In addition, the dilysine residue motif positioned 2–3 amino acids from the C-terminal end of the cytoplasmic domain allows interaction with the coat protein complex (COP), which is known to mediate retrieval from post-ER compartments (Teasdale and Jackson, 1996; Davis et al., 1997). Demonstration that the dilysine motif function acts as a retrieval signal for HLA-G from the post-ER compartment to the endoplasmic reticulum and that it is responsible for the slow transport kinetics of HLA-G was made by Park et al., both in JEG-3 and LCL721.221 transfected cells, which stably express HLA-G1. Interestingly, the loading of HLA-G1 with high-affinity peptides (KIPAQFYIL) instead of low-affinity peptides (KGGAQFYIL), has been shown to avoid retrieval of the HLA-G molecule, resulting in increased cell-surface expression of HLA-G1 (Park et al., 2001). The authors therefore proposed that the loading of these high-affinity peptides into the HLA-G groove could induce conformational changes that might alter recognition of

HLA-G MOLECULES

217

the dilysine motif. This in turn might inactivate the retrieval motif and result in upregulation of HLA-G1 cell-surface expression. The shortened cytoplasmic domain of the HLA-G1 molecule thus plays a critical role in quality control of the HLA-G molecule. Finally, HLA-G may be expressed as three other membrane-bound isoforms, HLA-G2, G3, and G4, all of which are present in cytotrophoblasts (Menier et al., 2000). As will be mentioned in the following chapter, the question of their expression at the cell surface has been investigated by three groups and produced debatable results (Bainbridge et al., 2000b; Mallet et al., 2000; Riteau et al., 2001c). HLA-G may be also be secreted. In particular, the soluble HLA-G5 protein includes the three extracellular domains of HLA-G1, but is shorter than the membrane-bound molecule (Fujii et al., 1994). HLA-G5 thus does not possess a tail endowed with dilysin residues, which are involved in quality control for reaching the cell-surface. Second, HLA-G5 exhibits a unique, 21-amino acid carboxyl terminus that confers solubility to the molecule. It is therefore tempting to suggest that these structural properties could render certain viral protein/HLA-G1 interactions ineffective. In agreement with this hypothesis is the absence of interaction of HLA-G5 with TAP, as mentioned above. Finally, it is noteworthy that the soluble HLA-G6 isoform shares common C-terminal properties with HLA-G5, and may thus also enjoy favored expression. In particular, the HLA-G6 protein isoform has been shown to circulate in maternal blood during pregnancy (Hunt et al., 2000). Therefore, convergent data show that HLA-G1 exhibits the typical structure of an antigen-presenting element. The demonstration that the HLA-Grestricted specific CD8 þ T cell repertoire is selected in HLA-G transgenic mice argues in favor of HLA-G protein being a restricting element. Nevertheless, the primary function of HLA-G might not be peptide presentation to the TCR, since the molecule exhibits low peptide diversity and slow turnover, which would be inefficient in presenting exogenous peptides. As recently proposed, peptide-loading could be a key mechanism in the transport of HLA-G to the cell-surface, acting in a quality-control process (Park et al., 2001). V. Structural and Functional Properties of HLA-G Molecules

The HLA-G primary transcript is alternatively spliced, resulting into seven alternative mRNAs which encode four membrane-bound (HLA-G1, -G2, -G3, -G4), and three soluble (HLA-G5, -G6, -G7) protein isoforms (Fig. 1) (Ishitani, 1992; Kirszenbaum et al., 1994; Paul et al., 2000a). The structural and functional properties of these HLA-G isoforms are presented below.

218


A. THE FULL-LENGTH HLA-G1 PROTEIN HLA-G1 mRNA encodes a 39-kDa full-length protein that contains 1, 2, and 3 extracellular domains linked to the transmembrane domain encoded by exon 5 and to a shortened cytoplasmic tail, due to a premature stop codon in exon 6. This isoform presents a structure similar to that of the other HLA class I molecules, in that it is non-covalently associated with 2-microglobulin and binds a nonapeptide (Lee et al., 1995). The N-linked glycosylation site (Asn 86) and the consensus cysteine pairs in the 2 and 3 domains are also conserved (Geraghty et al., 1987). Of note, the HLA-G1 heavy chain/ 2m/peptide heterotrimer has been recently described in a dimerized form expressed on the cell-surface. This dimerization is mediated through disulfide bonds of Cys-42 of the heavy chain (Boyson et al., 2002). The availability of monoclonal antibodies, such as 87G (Lee et al., 1995), 01G (Paul et al., 2000b), G233 (Loke et al., 1997), and MEM-G/09 (Lozano et al., 2002; Menier et al., 2003), able to specifically react with cell-surface HLA-G1, has widely permitted analysis of both the expression and function of this isoform in vitro and ex vivo. Several lines of evidence indicate that the primary function of HLA-G1 is to serve as an inhibitory ligand for immunocompetent cells, thus contributing to the establishment and maintenance of immune tolerance (Fig. 3): First, HLA-G1 inhibits the CD4 þ T cell proliferative response in allogeneic mixed lymphocyte reactions (Riteau et al., 1999; Bainbridge et al., 2000a). Second, HLA-G1 impairs the ability of HLA-A2-restricted, viral antigenspecific CD8 þ T cells to lyse target cells (Le Gal et al., 1999). Third, HLA-G1 expressed either spontaneously by trophoblasts or following transfection of the encoding cDNA by human cell lines (i.e., K562 erythroleucemic cells, lymphoblastoid LCL721.221 cells, or M8 melanoma cells), porcine endothelial cells, Chinese hamster ovarian cells, and J26 murine fibroblasts, inhibits both decidua and peripheral blood NK cell-mediated cytolysis (Rouas-Freiss et al., 1997a,b; Khalil-Daher et al., 1999; Navarro et al., 1999; Ponte et al., 1999; Sasaki et al., 1999, 1999a; Forte et al., 2000; Riteau et al., 2001a). These inhibitory effects of HLA-G1 are mediated through direct binding to inhibitory receptors, namely the immunoglobulin-like transcript (ILT2/CD85j) expressed by lymphoid and myelomonocytic cells (Colonna et al., 1997; Cosman et al., 1997; Saverino et al., 2000), ILT4 (CD85d) expressed by monocytes, macrophages, and dendritic cells (Colonna et al., 1998), and p49/KIR2DL4 (CD158d) expressed by NK and T cells (Ponte et al., 1999; Rajagopalan and Long, 1999). However, the nature of the HLA-G-KIR2DL4 interaction needs to be explored in more detail since divergent results have been obtained on this point (Boyson et al., 2002). Therefore, through these receptors the HLA-G1 protein can directly interact with T, B, NK, and APC

HLA-G MOLECULES

219

FIG. 3. Functional implications of interactions between HLA-G1 and immune cells such as, T cell, NK cell, and APC.

and exert its immunotolerant functions at different stages of the immune response. The biological relevance of HLA-G1 expression by APC, such as monocytes/macrophages and dendritic cells bearing ILT2 and ILT4 receptors, which have been described as binding soluble HLA-G1 tetrameric complexes (Allan et al., 1999), remains to be assessed in humans. However, we may hypothesize that such functions play an important role in blocking the host’s immune initiation, as recently proposed in a murine model (Horuzsko et al., 2001; Liang et al., 2002). HLA-G1 has been described to bind CD8 with an affinity (150 M) comparable to that of the CD8/HLA-A2 interaction (Gao et al., 2000). Therefore, it is possible that HLA-G1/CD8 interaction may facilitate HLAG-restricted antigen recognition by CTLs. However, although HLA-G1 binds nonamer peptides and CD8, it remains to be ascertained whether HLA-G1 is capable of presenting bacterial or viral peptides to T cells and elicit an HLA-G-restricted cytotoxic T cell-mediated response in a way similar to other HLA class I molecules. Finally, HLA-G1 expressed on transfected cells has been shown to modulate the release of cytokines from peripheral blood and decidual mononuclear cells. The amounts of TNF- and IFN- released from decidual and peripheral blood mononuclear cells were found to be decreased, while the amount of IL-4 was increased (Kanai et al., 2001a,b). Similarly, a Th2-type cytokine response (upregulation of IL-10 and IL-4 with downmodulation of

220


TNF- and IFN- ) has been associated with exposure to high concentrations of HLA-G1 purified from first-trimester trophoblast tissue (Kapasi et al., 2000). However, HLA-G1 has recently been found to affect the general cytokine production of decidual large granular lymphocytes in a manner not consistent with the Th1/Th2 paradigm (i.e., downmodulation of IL-10, IL-13, TNF-, IFN- , and GM-CSF) (Rieger et al., 2002). Interestingly, HLA-G1 protein cell-surface expression may in turn be induced by several cytokines, such as interferons (Yang et al., 1995, 1996; Amiot et al., 1998; Chu et al., 1998, 1999b; Lefebvre et al., 1999b, 2000, 2001; Ugurel et al., 2001), IL-10 (Moreau et al., 1999; Spencer et al., 2002), GM-CSF (Amiot et al., 1998), and LIF (Bamberger et al., 2000). B. THE HLA-G2, -G3,

AND

-G4 TRUNCATED PROTEINS

HLA-G2, -G3, and -G4 transcripts, which respectively exclude exon 3, exons 3 and 4, and exon 4, generate truncated isoforms that retain only the 1 domain for HLA-G3, the 1 and 3 domains for HLA-G2, and the 1 and 2 domains for HLA-G4, joined to the transmembrane region. Their ‘‘non conventional’’ conformational structure, exhibiting either one or two extracellular domains (monomers, dimers, heterodimers, or polymers) remains to be elucidated. In contrast to HLA-G1, analysis of the expression pattern of these other HLA-G isoforms remains difficult, due to the absence of reagents able to specifically recognize each of them. However, the availability of the 4H84 monoclonal antibody, raised against peptide 61–83 of the HLA-G 1 domain common to all HLA-G isoforms (McMaster et al., 1998), has allowed characterization of these truncated isoforms. It should be noted that a new mAb exhibiting a similar pattern of recognition to that of 4H84 is now available (MEM-G/01) (Lozano et al., 2002; Menier et al., 2003). By compiling data on the expression of HLA-G2, -G3, and -G4 proteins, several conclusions may be drawn: (i) they are expressed physiologically in cytotrophoblast (Menier et al., 2000), pathologically in tumor cell-lines, such as choriocarcinoma (Adrian Cabestre et al., 1999b), and melanoma (Paul et al., 1998; Adrian Cabestre et al., 1999b), as well as in transfected cells (Mallet et al., 2000; Bainbridge et al., 2000c; Riteau et al., 2001b,c) as glycoproteins of 31-, 22-, and 29-kDa, for HLA-G2, -G3, and -G4, respectively; (ii) their cellsurface expression is likely to be dependent on the cell type in which they are expressed. Indeed, HLA-G2, -G3, and -G4 proteins linked to a tag molecule and expressed in HLA class I-negative cell lines could not be detected on the cell surface (Bainbridge et al., 2000c; Mallet et al., 2000). In contrast, each of these untagged isoforms was found to be expressed on the cell-surface of a transfected HLA-A, -B, -C, and -E-positive melanoma cellline as endoglycosidase-H (Endo-H)-sensitive (i.e., exhibiting an immature

HLA-G MOLECULES

221

glycosylation pattern) cell-surface glycoproteins after a 2 h chase period. It has been suggested that in the latter cell model, an optimal environment is provided, allowing the possible association of these isoforms with chaperone molecule(s), and their consecutive cell-surface expression. Although cellsurface proteins containing immature oligosaccharides is unusual (HLA-G1 and classical HLA class I are Endo-H-resistant cell-surface molecules), it has been reported for other proteins, such as for the HLA class I-like CD1d, and for the mouse cytomegalovirus gp34 protein which escapes retention from endoplasmic reticulum (ER) by associating with folded class I MHC molecules. Similarly, truncated HLA-G isoforms may associate with a chaperone protein, such as other HLA class I molecules, allowing their escape from ER retention, and their cell-surface expression. By using the HLA class I-positive melanoma cell line in which HLA-G2, -G3, and -G4 isoforms were expressed on the cell surface, functional properties of these isoforms could be investigated. These truncated isoforms were found to inhibit both NK and antigen-specific CTL cytolysis through an HLA-E-independent pathway (Riteau et al., 2001c). Indeed, the lytic activity of the YT2C2-PR NK clone, which does not express CD94/NKG2A, the HLA-E-specific inhibitory receptor, was inhibited by the HLA-G isoforms, and the HLA-G isoformmediated inhibition of polyclonal NK lysis was not reversed by blocking HLA-E/CD94NKG2A interactions. The extracellular 1 domain, which is shared by all HLA-G isoforms, is thought to mediate these inhibitory properties. These truncated isoforms might become relevant in situations in which HLA-G1 expression is hampered, such as in 2-m-deficient cells, since HLA-G2, -G3, and -G4 may not require association with 2-m for their expression (Riteau et al., 2001c), and in individuals homozygous for the HLAG* 0105N ‘‘null allele’’ (Ober et al., 1998; Castro et al., 2000c; Arnaiz-Villena et al., 2001; Moreau et al., 2002). The potential functionality of this allele, which might indeed produce functional truncated HLA-G molecules at the fetal–maternal interface, has recently been discussed (Moreau et al., 2002). C. THE SOLUBLE HLA-G PROTEINS The soluble HLA-G isoforms are devoid of transmembrane and cytoplasmic parts, due to the presence of a stop codon in intron 4 (-G5, and -G6) or intron 2 (-G7) leading to a C-terminal tail specific for these soluble forms (Fig. 1). The soluble full-length HLA-G5 isoform is a 37-kDa glycoprotein that retains identical leader, 1, 2, and 3 domains, but includes an intron 4 sequence, yielding a specific open reading frame that encodes 21-amino acids linked to the 3 domain and excludes the transmembrane domain (Lee et al., 1995). Similarly, the HLA-G6 isoform retains an intron 4 sequence, yielding a 27-kDa soluble protein that lacks the 2 domain (Paul et al., 2000a). HLA-G7,

222


the most recently described isoform, is a 17-kDa soluble protein produced by a spliced variant in which the open reading frame continues in intron 2, which contains a stop codon. Thus, HLA-G7 is formed by the 1 domain linked to two specific amino acids encoded by intron 2 (Paul et al., 2000a). Due to the availability of antibodies, such as the monoclonal antibody 16G1 (Lee et al., 1995) and the polyclonal PAG5-6 antibody (Paul et al., 2000b) that are specific for the C-terminal part (i.e., intron-4-encoded residues) of both HLA-G5 and HLA-G6, the expression of these isoforms can now be investigated. Furthermore, the antibodies described above to recognize HLAG1 cell-surface molecules also react with the soluble HLA-G5 protein. In this regard, various assays based on the use of these antibodies have been described for measuring the level of soluble HLA-G in biological fluids. However, most of these assays do not distinguish between soluble HLA-G1 protein (i.e., HLA-G1s) released by the shedding of membrane-bound HLA-G1, and the soluble HLA-G5 protein, produced from a specific spliced transcript. HLA-G5 protein has been described in various body fluids, such as amniotic fluid and serum, from pregnant women (Puppo et al., 1999; Rebmann et al., 1999; Hamai et al., 1999a), cancer patients (Adrian Cabestre et al., 1999a; Ugurel et al., 2001), and transplanted patients (Lila et al., 2000, 2002). HLA-G5 has been also described in trophoblast (Chu et al., 1998; Fournel et al., 2000b), thymus (Mallet et al., 1999a,b), oocytes (Menicucci et al., 1999), pre-implantation embryo (Jurisicova et al., 1996b; Fuzzi et al., 2002), and during human cytomegalovirus reactivation (Onno et al., 2000b). It has been suggested that soluble HLA-G6 is present in maternal blood during pregnancy (Hunt et al., 2000), and until now has been clearly detected in vivo in serum from heart-transplanted patients (Lila et al., 2000, 2002). The presence of HLA-G7 as a secreted protein in both transfected cell supernatants and body fluids remains to be detected. While the functional properties of the HLA-G6 and -G7 soluble proteins remain to be characterized, those of the HLA-G5 soluble protein have been studied, using either HLA-G5 recombinant protein produced by prokaryotic (Marchal-Bras-Goncalves et al., 2001) or eukaryotic (Fournel et al., 1999, 2000a; Kanai et al., 2001a; Contini et al., 2003) cells, or HLA-G5 naturally produced by trophoblast cells (Kapasi et al., 2000) and T cells (Lila et al., 2001). Recombinant HLA-G5 exhibits inhibitory properties and binds to CD8 and inhibitory receptors, as does HLA-G1. Indeed, the HLA-G5 soluble form inhibits NK cell- and CD8 þ T cell-mediated lysis (Marchal-Bras-Goncalves et al., 2001; Contini et al., 2003) and the allogeneic CTL response (Kapasi et al., 2000). Moreover, HLA-G5 naturally produced by alloreactive CD4 þ T cells inhibits their proliferative response, thus also exerting its inhibitory effects through a feedback mechanism (Lila et al., 2001). On the other hand,

223

HLA-G MOLECULES

conflicting information has been reported regarding the ability of HLA-G5 to induce apoptosis of activated CD8 þ T cells. While recombinant HLA-G5 purified from transfected cells has been found to induce apoptosis of phytohemagglutinin-activated CD8 þ cells through binding to CD8 and in a Fas/FasL-dependent-manner (Fournel et al., 2000a; Contini et al., 2003), this soluble form naturally produced by alloreactive CD4 þ T cells during the mixed lymphocyte reaction did not enhance their apoptosis (Lila et al., 2001). Furthermore, an HLA-G-negative glioma cell-line transfected with HLA-G5 cDNA and co-cultured with freshly isolated CD4 þ or CD8 þ lymphocytes did not induce apoptosis in these effector cell populations (Wiendl et al., 2002). Finally, recombinant HLA-G5 has been found to influence the release of cytokines from peripheral blood mononuclear cells by stimulating the release of IL-10, TNF-, and IFN- (Kanai et al., 2001a). Such a soluble protein may have particular relevance in vivo, as it exerts its functions in the environment near its site of origin, and/or at distant sites, because of distribution via the circulatory system. D. RELATIONS BETWEEN HLA-G

AND

HLA-E EXPRESSION

AND

FUNCTION

In addition to its direct inhibitory role in interacting with the abovementioned immune receptors, HLA-G may also exert its immunosuppressive effects via an indirect pathway by favoring the expression of another non-classical HLA class I molecule, HLA-E. Indeed, peptides derived from the leader sequences of several HLA class I molecules, including HLA-G, bind to HLA-E and stabilize its conformation, allowing its stable cell-surface expression. Thus, when HLA-G is expressed in cells or tissues, co-expression of cell-surface HLA-E molecules may also occur. HLA-E inhibits both NK and T cells by interacting with the CD94/NKG2A inhibitory receptor, present on both these immune cells (Braud et al., 1998; Lee et al., 1998a,b). The contributions of the direct and indirect pathways to HLA-G-mediated inhibition have been evaluated in the context of the co-expression of HLA-G and naturally expressed HLA-E and the classical HLA class I molecules. Experiments carried out with polyclonal NK cells and CTL have demonstrated that when HLA-G1 is co-expressed with HLA-E, the blockade of HLA-E/CD94/NKG2A interaction does not reverse the inhibition of NK and CTL cytolysis (Le Gal et al., 1999; Riteau et al., 2001a,c). These experiments provide evidence that, in this context, HLA-G1 becomes the major NK and CTL inhibitory ligand. Furthermore, the leader peptide of HLA-G2, -G3, and -G4 might also permit stable HLA-E cell-surface expression, thus constituting an indirect inhibitory pathway by blocking lysis by CD94/NKG2A þ effector cells. In this regard, the triple transfection of human 2-microglobulin, HLA-E, and HLA-G3 genes into swine endothelial cells leads to HLA-E cell-surface expression, as well as inhibition

224


of human polyclonal NK cell-mediated lysis of such transfected cells (Matsunami et al., 2002). However, such a hypothesis for HLA-G2 and HLAG4 remains to be verified. Interestingly, HLA-E is also able to interact with specific activating receptors, such as CD94/NKG2C, present on NK cells. Of particular interest is that recognition of HLA-E by both CD94/NKG2A and CD94/NKG2C receptors is influenced by the HLA class I leader peptide sequence bound to HLA-E (Vale´s-Gomez et al., 1999). HLA-E loaded with the HLA-G-derived nonamer can interact with both receptors; it very efficiently inhibits the cytolytic activity of CD94/NKG2A þ NK clones and promotes a strong cytotoxic response by CD94/NKG2C þ NK clones by interacting with this activating receptor with a high affinity (Llano et al., 1998). Although the in vivo functional role of such an HLA-E/HLA-G peptide complex remains to be elucidated, it should be noted that target cells exhibiting both HLA-E and HLA-G cell-surface molecules are strongly protected from the lytic activity of polyclonal peripheral blood NK cells, showing a dominant inhibitory activity. In this case, as mentioned above, the major NK lysis inhibitory effect is mediated by the interaction of HLA-G with ILT-2 present on peripheral blood NK cells. E. HLA-G, CYTOKINES,

AND INFLAMMATION

Cytokine-interleukins are regulators of the host response during the inflammatory processes occurring during immune reactions, infections, and trauma. Pro-inflammatory cytokines and anti-inflammatory cytokines intervene during this process; the first acts mainly on TNF, IL1 , IFN , IL5, IL8, IL12, IL16, IL17, IL18, MIP1, MIP1 , and G-CSF; the second on IL4, IL6, IL10, IL11, IL13, TGF , and EGF. The majority of these molecules play a major role in the immune response and intervene in the lymphocyte balance Th1-Th2 (Dinarello, 2000; Opal and DePalo, 2000). Th1 cells are implicated in cellular immunity and macrophage-dependent inflammation. A dominant reply of this T cell subpopulation is thought to be involved in the pathogenicity of auto-immune specific organ diseases, of acute graft reject, of inflammatory dermatoses (psoriasis and atopic dermatitis), and of spontaneous recurrent miscarriage. On the other hand, Th2 cells are implicated in the antibody response (including IgE) and in macrophageindependent inflammations (accumulation of eosinophiles and inhibition of phagocytic action). Th2 cells are also responsible for allergies and for a faster evolution of HIV infection (Shurin et al., 1999). Studies carried out on inflammatory diseases, such as psoriasis (Aractingi et al., 2001), atopic dermatitis (Khosrotehrani et al., 2001), myopathic inflammations (Wiendl et al., 2000), and cutaneous lymphomas (Urosevic et al., 2002b), as well as on the inflammatory centers of other tumors (Urosevic et al.,

HLA-G MOLECULES

225

2001), and even on HIV-positive patients (Lozano et al., 2002), show the accrued presence of HLA-G protein (soluble or membranous). The expression of this molecule is principally induced by IL10, contrary to that of class I and II antigens, whose expression is inhibited by IL10 and increased by proinflammatory interleukins such as TNF, IFN , and IL1 (Moreau et al., 1999). These observations, both ex vivo and in vitro, lead to the proposal that the HLA-G molecule probably acts as a tissue defense mechanism in cytolysis against target cells (Carosella et al., 2001). The HLA-G molecule would therefore block the reaction of T effector and APC cells, thus avoiding the destruction of healthy tissues, and would play a regulatory role in limiting inflammatory processes. VI. Role of HLA-G in Normal and Pathological Pregnancies

Human reproduction begins by the union of paternal and maternal gametes, yielding a fetus that is semiallogenic with respect to the maternal body in which it will develop. The immunogenetically different fetus should elicit immune responses by the potentially hostile maternal organism. In fact, pregnancy is a prime example of immunotolerance. While at first glance, this maternal tolerance to the semiallogenic fetus appears to violate the classical rules of immunology, when fetal cells are confronted with maternal immunity, classical immune responses are subdued or suppressed. Thorough knowledge of the structure of the fetal–maternal interface and of HLA antigen distribution on maternal and fetal cells will provide insight into understanding the establishment of this privileged-immune status, as well as of its disturbances. A. THE FETAL–MATERNAL INTERFACE A fertilized human egg, called an embryo during the first eight weeks, and then a fetus, divides in two parts. One becomes the embryo proper and the other forms the trophoblast, which develops into all the cell types found in the human placenta. The placenta is now acknowledged to be a selective filter between the mother and the developing embryo/fetus. The trophoblast differentiates from a proliferative, undifferentiated, mononuclear cytotrophoblast—the stem cells of the placenta—into three distinct, differentiated forms of trophoblast cells: villous syncytiotrophoblasts, extravillous anchoring trophoblasts, and invasive trophoblasts. The trophoblast layer of human placental villi infiltrates the uterine lining (endometrium and myometrium) and penetrates the maternal arteries; it is the extension of fetal tissue into the mother. The essential roles of trophoblast cells in the fetal–maternal exchanges implicated in successful pregnancy occur during implantation, placentation,

226


hormone production, invasion of maternal blood vessels into the placenta, and protection of the fetus against attack by infiltrating maternal immune cells. Disturbance of any one of these functions can lead to pregnancy failure. One would expect contact by fetal cells with the maternal immune system to lead to rejection and destruction of the foreign fetus. To understand the absence of rejection, which is an atypical immune reaction, it is important to define the cell population present in the gravid uterus and the antigenic status of the placental cells (Johnson et al., 1999). Lymphomyeloid populations in the endometrium essentially consist of CD56 þ CD16 NK cells (distinct from peripheral NK cells), as well as a small number of CD8 þ T cells, CD16 þ NK cells, macrophages, and mast cells. The ubiquity of classical class I HLA-A, -B, and -C molecule expression should be restricted to non-trophoblastic somatic cells. Indeed, trophoblast tissues displays a particular HLA protein distribution, characterized by the absence of HLA class I molecules, and the presence of HLA-G (Kovats et al., 1990; McMaster et al., 1995; Hutter et al., 1996; Tarrade et al., 2001), of HLA-E (King et al., 2000; Menier et al., 2003), and to a lesser extent, of HLA-C (King et al., 1996). B. PROTECTIVE EFFECT

OF

HLA-G TOWARD MATERNAL CYTOTOXIC CELLS

During pregnancy, HLA-G expression is spatially restricted to trophoblast cells and temporally restricted to the first trimester (as well as to the third trimester at a decreased level) (Hiby et al., 1999b). A molecular construction that includes the HLA-G promoter in transgenic mouse models confirms this temporal expression, attributing it to control at its promoter level (Solier et al., 2001). Thus, the extinction of classical HLA class I antigens, the predominant expression of HLA-G in fetal trophoblast cells, and the presence of HLA-G receptors on decidual immune cells makes this gene an excellent candidate for explaining immune tolerance observed at the fetal– maternal interface. To examine this hypothesis, we conducted (for the first time) a study on the immunological role of HLA-G, carrying out in vitro cytotoxicity assays that confronted peripheral blood NK cells (obtained from 50 donors) and the K562 cell line, which does not express either classical or non-classical HLA class I molecules, transfected by HLA-G1. Lysis of the K562 cell line by NK cells was inhibited when the target K562 cell line expressed HLA-G1. The correlation between the presence of HLA-G1 on the target and inhibition of NK lysis was strengthened by the restoration of NK lysis after blocking of HLA-G1 with anti-HLA-G1 antibodies. In order to confirm these results in the physiological context of pregnancy, we carried out ex vivo experiments, using fetal and maternal tissues obtained from voluntary first-trimester terminations of normal pregnancies. Cytotoxic assays were carried out

HLA-G MOLECULES

227

under semiallogenic conditions (uterine NK cells and trophoblast cells obtained from the same mother), as well as under allogenic conditions (uterine NK cells and trophoblast cells from a different mother). The inhibition of cytolysis under both conditions demonstrated the ability of trophoblast cells to protect themselves against the lytic activity of decidual NK cells of the mother or of another pregnant woman. Taken together, these experiments demonstrate the protective role of the HLA-G molecule present on the surface of cytotrophoblast cells towards lysis carried out by uterine NK cells. HLA-G was thus shown to be responsible for the absence of maternal rejection during the time the fertilized egg implants into the uterus (RouasFreiss et al., 1997a). The natural physical barrier between fetal tissues and maternal organs is not impermeable; trophoblast cells encounter maternal immune blood cells. It is possible for maternal cells to enter the fetus, and fetal cells are able to cross the trophoblast and enter the maternal circulation (Hahn and Holzgreve, 2002). The presence of HLA-G on these fetal cells could favor their escape from maternal immunosurveillance, and to persist in maternal tissues. In this context, a study was done concerning the polymorphic eruptions of pregnancy (PEP), a cutaneous gestational pathology that occurs during the third trimester of pregnancy, when peripheral blood chimerism is very high. Interestingly, by studying samples of skin from women with PEP who were carrying male fetuses, Y-chromosome DNA was detected in dermis that also expressed HLA-G protein. In contrast, male DNA was not detected in any non-pregnant women. Moreover, expression of HLA-G by the cytotrophoblasts allows them to migrate to the maternal circulation and their infiltration into epidermal tissue (Aractingi et al., 1998). Similarly, a soluble HLA-G isoform not associated with light-chain 2-microglobulin was detected in a serum sample from pregnant women at higher level than in serum from non-pregnant women (Hunt et al., 2000). Moreover, the recent attribution of the increased HLA-G level encountered in peripheral maternal blood to fetal trophoblast cells detectable at the ninth week of gestation has permitted noninvasive prenatal diagnosis of numeric chromosomal aberrations (van Wijk et al., 2001). C. FAILURES

OF

PREGNANCY

During pregnancy, fetus and mother live in symbiosis, and the outset of this relationship is most important. Its disruption often ends in the failure of pregnancy. During this period, immunological interactions between fetal trophoblast cells and maternal decidual leukocytes at the fetal–maternal interface are primordial. Since HLA-G is the most abundant HLA protein expressed by fetal trophoblast, invading the decidua and the walls of spiral arteries, and because HLA-G is a powerful immunosuppressor (Menier et al.,

228


2000; Riteau et al., 2001a), research aimed at explaining fetal–maternal tolerance—and even successful pregnancy—focuses principally on the HLA-G gene. An interruption or failure in reproduction can take place at many—but not random—temporal and spatial levels during pregnancy. Indeed, abnormal development may occur either within the embryo or in the extra-embryonic placenta at the beginning of pregnancy. When a fertilized egg implants outside the uterine cavity, placental formation is wrongly located and the pregnancy is said to be ectopic. A welltargeted implantation can be disrupted during placentation, which is the invasion of trophoblast tissue into a well-structured, functional placenta. These dysfunctions often end in miscarriage. The three main complications of pregnancy related to the trophoblast and placenta are hyper- and hypoinvasion of trophoblast, immunological rejection, and infection by microorganisms. 1. Normal and Ectopic Implantation The detection of HLA-G mRNA from blastocysts suggested an important role for HLA-G in embryo viability and implantation (Jurisicova et al., 1996a,b). This suggestion was confirmed by the detection of soluble HLA-G proteins deriving from blastocysts (Menicucci et al., 1999), then by the identification of this soluble form as deriving from HLA-G5 (Fuzzi et al., 2002). Furthermore, the authors of this study correlated successful implantation and detection of HLA-G5 from pre-implantation embryos after in vitro fertilization or intracytoplasmic sperm injection. In fact, pregnancy was achieved when in vitro transfers were obtained with growing embryos that secreted HLA-G5 in their supernatant, and pregnancy did not occur with preimplantation embryos that did not secrete HLA-G5. In contrast, chromosome abnormalities, such as trisomy, do not induce deficient HLA-G expression in trophoblast (Rabreau et al., 2000). Study of HLA-G expression during pregnancy is carried out with cells that are in contact with the maternal decidua, but these two tissues could not be dissociated in any intrauterine implantation, which would have made it possible to determine whether these tissues were dependent on HLA-G expression. In ectopic tubal pregnancies, which represent a decidual layer-free tissue, although disturbed trophoblast differentiation is visible, all extra-villous trophoblast cells express HLA-G (Proll et al., 2000; Emmer et al., 2002). As in the case for implantation in the uterus, HLA-G mRNA and protein were found to be expressed in ectopic pregnancies and in extra-villous trophoblast from placental accreta, which develops without decidua (Goldman-Wohl et al., 2000; Rabreau et al., 2000). These results suggest that HLA-G expression in trophoblast cells does not depend on specific

HLA-G MOLECULES

229

environmental factors present in the decidua that permit implantation in other tissues. 2. Recurrent Miscarriage Recurrent spontaneous abortion (RSA) remains unexplained in a large proportion of cases. Numerous biological causes could lead to this result, but studies of HLA-G in RSA showed it to be involved at different levels in cellular interactions, in the cytokine pattern, and in genetic predisposition (Hviid et al., 2002). HLA-G immunostaining at the fetal–maternal interface in the columns of proliferating cytotrophoblast cells of recurrent miscarriage tissues revealed— as in the normal situation—increasing staining from the embryo to the decidua, but this staining appeared to be weaker in recurrent miscarriage tissues. It has been suggested that this decreased HLA-G expression could be related to an abnormal maternal NK cell phenotype in RSA cases, notably in the appearance of a CD16 population in pathological tissues (Emmer et al., 2002). Lower HLA-G expression in RSA could influence cytokine release. Cytokine assay in co-cultures reveals that cytokine release by peripheral mononuclear cells after contact with HLA-G-expressing cells is perturbed, whether mononuclear cells come from women with recurrent abortion, compared with mononuclear cells from fertile women (Hamai et al., 1998). Allele frequency studies of HLA-G and RSA in Finnish and Japanese populations did not detect significant differences (Karhukorpi et al., 1997; Yamashita et al., 1999). Other authors obtained the same results when they analyzed data from all RSA couples without distinction, but it was possible to establish a correlation between HLA-G genotype and RSA in women who had undergone five or more RSAs, with a significant increase in the HLA-G* 01013 and 0105N allele distribution (Pfeiffer et al., 2001). The presence of either HLA-G* 0104 or HLA-G* 0105N has also been associated with unexplained recurrent miscarriage (Aldrich et al., 2001). 3. Gestational Trophoblastic Disease In contrast with pre-eclampsia, gestational trophoblastic disease represents increased and uncontrolled trophoblast invasion. This exaggerated proliferation may not have pathological consequences, but can result in placentalsite trophoblastic tumors, molar pregnancy, or even choriocarcinoma. The deeper trophoblast cells invade the decidua, the more HLA-G they express. Shallow trophoblast invasion, encountered in pre-eclampsia, is associated with reduced HLA-G expression. Molar pregnancies develop morphologically and genetically in two different forms: complete and partial. A complete hydatidform mole (CHM) appears when paternal chromosomes dominate the genome in the absence of the maternal component. In CHM, there is no

230


fetus and the trophoblast is hyperplastic. Partial hydatidform moles (PHM) contain both abnormal trophoblastic cells, and often a fetus with severe defects and a triploid karyotype consisting of one paternal and two maternal chromosomes. In PHM, the trophoblast is generally not hyperplastic (Paradinas et al., 1996). Hyperplastic clusters of extravillous trophoblast cells from both CHM and PHM express a greater number of HLA-G molecules than a normal egg does. In PHM, a less invasive trophoblast population is associated with HLA-G expression lower than that of CHM (Rabreau et al., 2000). This reduction of HLA-G expression in PHM compared with CHM can go as far as the complete absence of HLA-G when PHM is associated with pre-eclampsia (Goldman-Wohl et al., 2001). Thus, HLA-G expression is greatly enhanced in molar pregnancies, presenting deeper infiltration of extravillous trophoblast cells in the decidua. Furthermore, the exclusively paternal genetic status of CHM confirms that the expression of HLA-G molecules—whether of paternal or maternal origin—confer trophoblast cells with the capacity to protect themselves from NK cells. Choriocarcinoma is a rare form of primary placental cancer. The JEG-3, BeWo, and JAR human choriocarcinoma cell lines are widely used as models to study trophoblast cells and HLA-G (Kovats et al., 1990; Copeman et al., 2000; Sivori et al., 2000; Easterfield et al., 2001). Only one study has reported the results of biopsies of choriocarcinoma and trophoblastic tumors. Positive HLA-G immunoreactivity was specific for intermediate trophoblast (a subpopulation of extravillous trophoblast cells) in choriocarcinoma and trophoblastic tumors, and the authors suggest using HLA-G as a marker in the differential diagnosis of gestational trophoblastic disease (Singer et al., 2002). 4. Pre-eclampsia Pre-eclampsia (PE) is a materno-placental disease in which utero-placental blood pressure increases dramatically. In pre-eclampsia, invasion by trophoblast from maternal spiral arteries is decreased or absent, resulting in inadequate perfusion of the placental bed and a reduced supply of oxygen and nutrients to the fetus (Khong et al., 1986; Genbacev et al., 1997). These clinical and anatomic observations are consequences whose causes remain to be identified and explained (Roberts and Redman, 1993). PE appears related to an enhancement of maternal immunological defenses, preventing invasion into the spiral arteries. Indeed, at the pre-eclamptic fetal–maternal interface, trophoblasts do not progress beyond the endometrium and myometrium (Taylor, 1997). Familial predisposition to PE suggests genetic control and inheritance in this pathology (Arngrimsson et al., 1990). Expression of several genes outside the HLA complex has been linked to disorders in PE, but the cause of inadequate trophoblast invasion was not elucidated. HLA-G seems to be an

HLA-G MOLECULES

231

ideal candidate gene to explain the ability of trophoblast cells to differentiate and proliferate in maternal tissue. A preliminary study observed a lower mRNA HLA-G level in the pre-eclamptic population, but the authors attribute this lower expression to a reduced number of trophoblast cells in this pathological tissue (Colbern et al., 1994). The development of specific antibodies against HLA-G protein have allowed improved experiments and to confirm impaired HLA-G expression in pre-eclamptic trophoblast (Hara et al., 1996; Lim et al., 1997). Immunohistochemical characterization of trophoblast and lymphocyte populations in the pre-eclamptic fetal–maternal interface has led to the observation of CD56 þ NK/CD8 þ T cell imbalance with enhancement of CD8 þ T cells, and led to the description of two different trophoblast types; one in the decidua, observed in normal placenta, and the other unable to migrate among NK cells (Stallmach et al., 1999). These two different trophoblast behaviors were distinguished by their HLA-G expression. Indeed, molecular analysis by RNA in situ hybridization in normal placentae correlates increased HLA-G expression with increased trophoblast invasiveness, and allow distinguishing two trophoblast populations in pre-eclamptic placentae, one consisting of HLA-G-negative trophoblasts that failed to invade maternal tissue, and another consisting of a small quantity of HLA-G-positive trophoblasts that had reached the decidua. An evaluation of mRNA HLA-G expression in PE and normal placental tissue revealed the significant overall deficit of HLA-G transcription and protein expression to be associated with a disturbed HLA-G isoform expression profile and the absence or very low expression of the spliced HLA-G3 mRNA (O’Brien et al., 2001). The cytokine environment is also changed at the pre-eclamptic fetal– maternal interface, with a pro-inflammatory Th1/anti-inflammatory Th2 cytokine imbalance observed in patients with predominant Th1-type immunity, an increase in IFN- secreting cells (Th1), and a decrease in IL10- and IL4-secreting cells (Th2) (Hennessy et al., 1999; Saito et al., 1999). A reciprocal effect is now known to exist between the release of cytokines and HLA-G expression. In fact, cytokines can raise HLA-G expression, and conversely. A co-culture of decidual lymphocytes with HLA-G-expressing cells reduced the release of cytokines, including IFN- and IL10, indicating an immune-suppressive effect of HLA-G on decidual lymphocyte activity (Rieger et al., 2002). Genotyping of HLA-G in various populations has led to a change in its initial description as a non-polymorphic gene to that of a low polymorphic one. The relationship between HLA-G genotype and susceptibility to the development of diseases of pregnancy has since been investigated in the search for possible predominance among the 15 known HLA-G alleles. Deletion or insertion of a 14-nucleotide (D/I-14) sequence in the untranslated region of exon 8 seems to play a role in its transcriptional and

232


post-transcriptional functions. A study of persons affected by pre-eclampsia (PE patients, their husbands, and individuals born of a PE pregnancy) revealed no significant difference in the polymorphism of these 14 nucleotides in the 30 UT region (Humphrey et al., 1995), although placenta from mild preeclampsia, compared with normal placenta revealed significant differences in the silent CAC-CAT polymorphism at codon 93 (C ! T-93), in exon 3 with an excess of T-93, and in exon 8, with an excess of I-14 alleles in PE samples (O’Brien et al., 2001). Conversely, distribution of the G* 0106 allele (T-93 and I-14) screened for the presence of polymorphism in codon 258 showed no obvious association between PE and recurrent spontaneous abortion (Hviid et al., 2001). Furthermore, a recent study suggests an association between PE and the HLA-G* 0104 allele, which has the C-93 and D-14 genotype (Carreiras et al., 2002). Therefore, it is now clear that there is a relationship between PE and HLA-G polymorphism, but the identity of the implicated alleles remains to be clarified. Discrepancies observed among the various studies may be due to the fact that their authors did not take the degrees of seriousness of the PE diagnosis into account. In fact, it is possible to distinguish degrees of pre-eclampsia, from mild to severe, the latter often being associated with retarded fetal development. It seems important to not dissociate the importance of polymorphism from the level of progression of the disease. By labeling invading trophoblast cells, the HLA-G distribution pattern reflects the heterogeneity of the trophoblast at the fetal–maternal interface. Through its ability to prevent allo-recognition by inhibiting maternal NK and CTL activity, HLA-G plays a key role in successful implantation and placentation. Alteration of HLA-G expression and HLA-G allele inheritance has often been correlated with major disturbances and diseases of pregnancy. The impact of HLA-G molecule expression on the success and failure of pregnancy is now studied in relation to full-length HLA-G1 and its soluble HLA-G5 isoform. It would be interesting to introduce analysis of the other spliced isoforms, but specific antibodies against these HLA-G molecules do not exist. Finally, these results considered together can reveal the involvement of HLA-G in disease states only if pathologies are meticulously diagnosed and when patients are divided into groups, according to the disease state. VII. HLA-G in Organ Transplantation

The role of HLA-G in, and its potential use for the prevention of graft rejection are being increasingly investigated. The reasons for this interest are that while HLA-G differs from classical HLA Class I molecules in being of low polymorphism and so far described as non-allo-stimulatory, HLA-G is a ligand

HLA-G MOLECULES

233

of inhibitory receptors. What makes it even more attractive is that all cell subsets involved in graft rejection bear at least one receptor for HLA-G. HLAG therefore has the potential to be tolerogenic, by exerting its inhibitory functions on all cells responsible for graft rejection without triggering an allogeneic response. The potential involvement of HLA-G in the establishment and maintenance of graft tolerance has been investigated in vitro and in vivo, in the context of allo- and xeno-transplantation, in the human and in the animal system, and when HLA-G is expressed by various cell subsets. We summarize here the latest data on this topic. A. IN VIVO RELEVANCE

OF

HLA-G

HLA-G is a molecule that under normal circumstances is not expressed, except at the feto-maternal interface and by thymic epithelial cells. However, ectopic HLA-G expression has been demonstrated in heart-transplanted patients. Two studies by Lila et al. (2000, 2002) showed that in the absence of inflammation, HLA-G was expressed by myocardiac cells of 18% of hearttransplanted patients. It was systematically possible to detect the soluble HLA-G5 and HLA-G6 isoforms in the serum of these HLA-G-positive patients. Finally, HLA-G expression by myocardiac cells of heart-transplanted patients proved to be stable over time. Most interestingly, expression of HLAG by heart-transplanted patients might be linked to better graft acceptance. HLA-G expression significantly correlated with a reduced number of acute rejection episodes (1.2 1.1 in HLA-G positive patients vs 4.5 2.8 in HLAG-negative patients) with no chronic rejection. HLA-G-negative patients occurred and had a higher number of acute rejection episodes and chronic rejection in 27% of the cases. The mechanisms by which HLA-G neo-expression occurs in some transplanted patients are still unknown, but there are data suggesting that an allo-reaction itself might trigger it. Indeed, Lila et al. (2001) showed that in vitro, HLA-G5 was produced and secreted by alloreactive CD4 þ T cells from some mixed lymphocyte reactions. Furthermore, this alloreactive T cellderived soluble HLA-G was functionally active and capable of inhibiting CD4 þ T cells during in vitro allo-proliferation. These data were the first and remain the only linking natural HLA-G expression with long-term allograft acceptance in vivo in humans. They are strengthened by recent work in the murine system (Horuzsko et al., 2001; Liang et al., 2002), which shows that human HLA-G can bind to the murine ILT4 homolog PIR-B receptor, and that this engagement of PIR-B by human HLA-G significantly improved skin allograft survival. In the first report, Horuzsko et al. showed that HLA-G-transgenic mice exhibited reduced cellular immune responses. Indeed, allograft survival in HLA-G mice was

234


prolonged around two-fold, compared with non-transgenic controls, even though the T cell function in HLA-G-transgenic mice was normal. The authors demonstrated that the prolonged survival of allografts was due to an HLA-G-induced maturation defect of the antigen-presenting cells of HLA-G transgenic mice, leading to impaired stimulation of allo-reactive T cells. Following up on these findings and working with non-transgenic animals, Liang et al. (i) confirmed the HLA-G-induced APC maturation defect, and (ii) demonstrated that immunization of recipient animals with HLA-G-coated latex microbeads 1 day prior to, and again 4 days after skin allo-transplantation increased graft survival by about two-fold as well. These latter data are of crucial importance because they show for the first time that (i) human HLA-G can bind to the murine homolog of a human HLA-G receptor on APCs, even though no murine HLA-G homolog that has been found to date, and (ii) in vivo immunization with human HLA-G is effective and improves allograft survival. The mechanisms by which HLA-G can alter allo-transplantation outcome have not been investigated in vivo. However, in vitro studies produced results clear enough to shed light on what might happen, indicating that HLA-G might improve graft acceptance by acting on multiple parameters and cellular actors in graft rejection. B. POTENTIAL MECHANISMS

OF

GRAFT PROTECTION

BY

HLA-G

Rejection of an allo- or a xeno-graft is a complex, multi-stepped event. If the potentially reactive cells are NK cells, they will need to reach the graft, adhere to the tissue, and finally lyse it. However, if the potentially reactive cells are T cells, an allogeneic response must be initiated through (i) graft allo-antigen uptake and presentation by infiltrating APCs; (ii) maturation of these APC; (iii) mutual stimulation of allo-specific CD4 þ , CD8 þ T cells and APCs; (iv) migration of the cytolytic cells to the graft; and finally, (v) lysis. In vitro data seem to indicate that HLA-G is capable of interfering with each of these steps, providing multi-level protection to the graft. 1. Inhibition of Adhesion and Transendothelial Migration Rolling, and firm adhesion of NK cells to their targets is a prerequisite to their cytolytic function. The influence of HLA-G on rolling adhesion and transendothelial migration of NK cells has been investigated in the context of xeno-recognition, since NK seem to play a crucial role in the rejection of xenografts. Forte et al. (2000, 2001) have shown that expression of HLA-G at the surface of porcine endothelial cells (PEC) monolayers inhibits the rolling adhesion of activated human NK cells. In these experiments, rolling adhesion of activated NK cell lines to HLA-G-transfected PEC monolayers was

HLA-G MOLECULES

235

inhibited 50%, compared with their rolling adhesion to non-transfected PEC. The effect was not completely mediated by the HLA-G receptor ILT2, since masking ILT2 on NK cells produced no, or only partial reversal of the HLA-G-mediated inhibition of rolling adhesion. This might suggest that NK cell-expressed KIR2DL4 might have been involved principally in the inhibition of NK rolling adhesion, or that a yet-unknown HLA-G receptor was. Using a similar system, Dorling et al. demonstrated HLA-G-mediated inhibition of the transendothelial migration of human NK cells (Dorling et al., 2000). Kinetics experiments showed that HLA-G delayed human NK transendothelial migration across PEC monolayers for up to half an hour. In this study, the inhibitory effect of HLA-G could be completely reversed by blocking ILT2 on NK cells. These results indicate that HLA-G might improve pig-to-human xenograft survival, by limiting human NK–pig tissue interactions. The same experiments have not been carried out in the context of allo-recognition, or for cytotoxic T lymphocytes. However, since the basic requirements for cytolysis by NK cells or CTLs with respect to migration and adhesion are similar, these results may very well hold true outside the xeno-recognition system. 2. Inhibition of CD4 þ , CD8 þ , APC Mutual Activation/Differentiation and Effector Functions The generation of an allo-response involves at least three different cell subsets: alloreactive CD4 þ T cells, alloreactive CD8 þ T cells, and antigenpresenting cells. Proper activation and functional maturation of APCs and allo-specific CD4 þ T cells is necessary for generation of allo-specific CTLs. HLA-G can act on CD4 þ T cells and APCs, leading to the inhibition of in vitro allo-responses. The inhibition of CD4 þ T cell allo-proliferation by membrane-bound HLA-G1 was investigated in two studies (Riteau et al., 1999; Bainbridge et al., 2000a) that demonstrate that HLA-G1-transfected MHC class II-positive cells inhibited the allo-proliferation of CD4 þ T cells by 80%, compared to nontransfected controls. Similar results were obtained when HLA-G1 was presented by HLA-negative, HLA-G1-transfected K562 cells used as third party cells in mixed lymphocyte reactions (Riteau et al., 1999), or when soluble HLA-G5 was present (Lila et al., 2001). It should be noted that in these three types of experiments, HLA-G could potentially bind to CD4 þ T cells (through ILT2) as well as to antigen-presenting cells (through ILT2 and/or ILT4). It is therefore difficult to determine whether HLA-G acted on CD4 þ T cells and prevented them from being activated, on APCs and prevented them from fully mature and properly stimulate CD4 þ T cells, or on both cell subsets.

236


The question of the direct effect of HLA-G on CD4 þ T cells is still unanswered, but as stated above, recent work has shed light on the effects of HLA-G on APCs. Binding to murine PIR-B receptors, HLA-G induces a defect in murine dendritic cell functional maturation in vivo in HLA-Gtransgenic mice, as well as in vitro in non-transgenic mice (Horuzsko et al., 2001; Liang et al., 2002). This functional effect of HLA-G on APCs is further strengthened by a recent study by Chang et al. (2002), which shows that ILT4 over-expressing APCs do not induce and support allo-specific CD4 þ T cell proliferation. If data from these two groups are put together in an all-human system, it seems that by binding to ILT4 receptors on dendritic cells, HLA-G should be capable of preventing APC functional maturation and the generation of an allogeneic response. This effect might even be increased by a direct effect of HLA-G on T lymphocytes. On the other hand, the inhibition of NK function by HLA-G-transfected xenogeneic cells is now well documented. Human NK cells can lyse allogeneic, as well as xenogeneic targets. HLA-G, being a human HLA-Class I molecule of low polymorphism, can act as a ligand for the NK ILT2 and/or KIR2DL4 inhibitory receptors, and protect allo- and xenogeneic cells from NK cell lysis, as will be reviewed next. Transfection of HLA-G1 cDNA or genomic HLA-G DNA into porcine endothelial cells (PEC) or Chinese hamster ovarian cells (CHO) lead to a ‘‘partial’’ to complete inhibition (ranging from 30 to 100%) of the cytolytic function of human polyclonal NK cells and of that of most human NK cell lines (Sasaki et al., 1999a,b; Forte et al., 2000, 2001; Matsunami et al., 2001). Furthermore, two studies have shown a correlation between the level of HLA-G1 cell-surface expression by transfected PEC and the extent of NK functional inhibition (Forte et al., 2001; Matsunami et al., 2001). In conclusion, all these experiments demonstrate that membrane-bound and soluble HLA-G are expressed in the context of allo-transplantation and are capable of inhibiting NK cytolytic function, as well as CD4 þ and CD8 þ T cell allo-responses (Fig. 4). It seems that HLA-G may participate in the down-modulation of a rejection episode by acting at all levels of the allo-response. It is not known which effect of HLA-G is functionally the most important, but the fact that in vivo immunization with HLA-G induces prolonged allograft survival in mice is certainly highly promising.

VIII. HLA-G in Malignancies

Over the past few years, HLA-G has been proposed as playing a potential role in the molecular mechanisms underlying immune escape strategies

HLA-G MOLECULES

237

FIG. 4. Potential inhibitory effect of HLA-G on allogeneic response. Center: Schematic representation of the sequence of events involved in transplant rejection. In the course of a graft rejection reaction, antigens from the transplant (Ag) are captured and processed for presentation by immature APCs in the context of HLA class I and class II molecules. APCs mature and interaction with antigen-specific CD4 þ and CD8 þ T cells leads to the activation of all three cell subsets and to the differentiation of CD4 þ T cells and CD8 þ T cells into CD4 þ Th and CD8 þ CTLs effectors respectively. CTLs then migrate to the graft and may lyse graft cells. NK cells migrate to the graft and may lyse graft cells, based on the compatibility between graft HLA Class I antigens and NK cells inhibitory receptors. Sides: Summary of HLA-G functions in the context of transplantation. G represents HLA-G. A: Human HLA-G interacts with murine APCs and inhibit their maturation in vitro and in vivo. B: Allo-specific CD4 þ T cells express and secrete soluble HLA-G5 in vitro. C: HLA-G inhibits CD4 þ T cells proliferation in vitro and may inhibit concomitant maturation. D: HLA-G inhibits the generation of allo-specific CTLs in vitro. E: HLAG inhibits rolling adhesion and transendothelial migration of NK cells in the xenogeneic context. F: HLA-G inhibits the cytolytic function of NK cells in the allo- and in the xenogeneic context, and inhibits antigen-specific CTL cytolytic function.

utilized by tumor cells. Although divergent information has been provided about HLA-G expression by some types of tumors, many investigations have shown HLA-G mRNA and protein expression to be associated with the malignant transformation of cells. We discuss below the possible reasons for such conflicting results and present studies emphasizing that HLA-G, the engagement of which generates inhibitory signals in various immune cells, has functional significance in the host’s immunosurveillance of tumors.

238


A. EXPRESSION OF HLA-G CANCER PATIENTS

IN

TUMORS OBTAINED FROM

By carrying out immunohistochemical analysis of tumor biopsies obtained from cancer patients, upregulation of HLA-G protein expression has been found in various types of malignant lesions, such as melanoma (Paul et al., 1998, 1999; Carosella et al., 2000; Wagner et al., 2000), cutaneous lymphoma (Urosevic et al., 2002b), non-Hodgkin lymphomas (Dreńou et al., 2002), glioblastoma (Wiendl et al., 2002), breast cancer (Lefebvre et al., 2002), lung carcinoma (Pangault et al., 1999; Urosevic et al., 2001; Pangault et al., 2002), colorectal carcinoma (Fukushima et al., 1998), renal cell carcinoma (Ibrahim et al., 2001a), bladder carcinoma (Carosella et al., 2000), and hydatidform moles (Rabreau et al., 2000; Goldman-Wohl et al., 2001). According to type of tumor, the expression of HLA-G was found in tumor cells or/and in infiltrating cells (predominantly monocytes/macrophages, but also lymphocytes), but not in the corresponding healthy tissue. Studies of interest describing ex vivo expression of HLA-G at the tumor site are reviewed below. Among them, one deserves particular attention: Sequential biopsies of healthy skin, primary cutaneous tumor, lymph node metastasis, and a tumor regression site within the skin primary tumor were all obtained from one patient. HLA-G protein was detected in both primary and metastatic tumor sites, but not in either healthy skin that had been resected from the vicinity of the tumor or at a tumor regression site in which an efficient immune response may develop, in the absence of HLA-G (Paul et al., 1999). Another study described a correlation between poor prognosis of melanoma treated with interferon alpha (IFN-)-2b and the presence of HLA-G in melanoma cells before the treatment (Wagner et al., 2000). This may be important for selection of cancer patients likely to benefit from IFN- therapy. Recently, various types of malignant epithelial tumors and benign cutaneous lesions arising in kidney transplant recipients have been investigated. The results revealed HLA-G expression in malignant and pre-malignant cutaneous diseases (i.e., 35% of squamous cell carcinomas, 47% of Bowen disease, 27% of actinic keratoses, 14% of basal-cell carcinomas) but not in the 24 benign lesions studied (Aractingi et al., 2003). Interestingly, when both benign and malignant cutaneous lesions developed at distinct sites in the same transplanted patient, upregulation of HLA-G expression was only found at the malignant site. This study confirms the link between ectopic HLA-G expression and tumor development in patients with cancer. The HLA-G gene was found to be transcribed in 100% of 45 cases of primary cutaneous lymphoma and the HLA-G protein was detected in 70% and 45% of cutaneous B-cell and T-cell lymphomas, respectively (Urosevic et al., 2002b). In T-cell cutaneous lymphomas, HLA-G protein expression

HLA-G MOLECULES

239

was associated with high-grade histology and an advanced stage of the disease. Two populations of cells expressing HLA-G were identified: lymphoid cells (i.e., neoplastic and infiltrating T and B lymphocytes) and cells of myeloid origin (i.e., intra-lesional macrophages and dendritic cells). Interestingly, the presence of the immunosuppressive IL-10 cytokine, which is known to be secreted by cutaneous lymphomas and to induce HLA-G expression, could be correlated with HLA-G protein expression in these cutaneous tumors. Analysis of tumor and adjacent normal renal tissue samples showed tumorspecific HLA-G expression in 61% of renal cell carcinomas (Ibrahim et al., 2001a). HLA-G has been shown to be expressed in 26–33% of lung carcinoma lesions, either in tumor cells and/or intratumoral-infiltrating cells, such as activated macrophages and dendritic cells (Pangault et al., 1999; Urosevic et al., 2001; Pangault et al., 2002). HLA-G protein was recently detected in brain tumors, such as glioblastomas and anaplastic oligoastrocytoma (Wiendl et al., 2002), as well as in breast cancer (38% of cases) (Lefebvre et al., 2002). In the latter study, both HLA-G and its inhibitory receptor, ILT2, were found to be expressed in tumor tissue samples. Interestingly, in vivo expression of the ILT2 inhibitory receptor has already been reported on tumor-specific CTLs in melanoma (Ikeda et al., 1997; Bakker et al., 1998; Speiser et al., 1999), and renal cell carcinoma (Guerra et al., 2000; Gati et al., 2001) leading to the downmodulation of the lytic activity of such CTLs. In conclusion, the number of lesions analyzed in the various types of malignancies ranges from about 10 lesions in basal cell carcinoma to more than 100 lesions in melanoma. The percentage of positive lesions varies markedly among the various types of malignancies. It is noteworthy that, although HLA-G is activated at the transcriptional level in most of the tumor specimens analyzed, no systematic correlation with protein expression is found. This may be explained by several mechanisms, including mutations in the HLA-G transcript, blocking protein translation and strong posttranscriptional regulatory mechanisms. Moreover, not all HLA-G isoforms are usually detected at the transcriptional level in tumors, as is the case in trophoblasts. An illustrative example is the differential transcription of HLA-G1 and HLA-G5 in melanoma biopsies. This may reflect an adaptive tumor mechanism that enables selective expression of membrane-bound or soluble HLA-G isoforms (Paul et al., 1999). The generation of soluble HLA-G molecules has also been associated to certain HLA-G alleles, implying that the levels of soluble HLA-G secretion are under genetic control (Rebmann et al., 2001). However, divergent information has been provided concerning the expression of HLA-G protein in malignant tumors. Indeed several authors

240


could not detect HLA-G expression in a variety of solid tumors (Pangault et al., 1999; Real et al., 1999; Frumento et al., 2000; Davies et al., 2001; Hurks et al., 2001; Palmisano et al., 2002). These contrasting results may be explained by the fact that most of these studies were conducted on frozen tissue sections, allowing less discrimination of cell morphology, obtained from a small number of biopsies from a tumor-type analyzed, and using mAbs that react only with the HLA-G1, and -G5 isoforms. Sharing of reagents as well as standardization of techniques will avoid the generation of conflicting data in future literature. Studies along this line will benefit from comparative analysis of the reactivity of anti-HLA-G antibodies in frozen versus paraffin-embedded tissues and from the use of anti-HLA-G (Fab0 )2 fragments in immunohistochemical experiments. In this regard, the second international workshop on HLA-G, which will be held in Paris in 2003, will allow pursuing our efforts in the standardization of tools and protocols. B. DETECTION PATIENTS

OF

SOLUBLE HLA-G

IN

SERUM OBTAINED

FROM

CANCER

An increase in the serum HLA-G level has been described in patients with melanoma that was enhanced upon treatment with IFN-. The serum HLA-G level was found to be correlated with advanced disease stage and tumor load. However, it was not associated with recurrence-free or overall survival. Interestingly, cell-surface expression of HLA-G1 was detected on peripheral blood monocytes of two melanoma patients with an elevated soluble HLA-G serum level. Such expression was upregulated after systemic IFN- immunotherapy. Monocytes may thus constitute a possible source of elevated serum soluble HLA-G levels in these patients (Ugurel et al., 2001). Such serum HLA-G antigens, which may be derived from the release of membrane-bound HLA-G1 and/or from the secretion of soluble HLA-G5 may affect anti-tumor immune both locally at the tumor site, as well as systemically by distribution via the circulation (see above-described functional properties of HLA-G). C. HLA-G EXPRESSION

BY

TUMOR CELL-LINES

With the exception of choriocarcinoma cell-lines (i.e., JEG-3 and BeWo), only a few tumor cell-lines have been found to express HLA-G on their cell membranes (Pangault et al., 1999; Real et al., 1999; Polakova and Russ, 2000; Frumento et al., 2000; Davies et al., 2001; Hurks et al., 2001). The discrepancy between the data obtained with tumor cell-lines and surgically removed lesions may reflect the lack of factors in tissue culture media that

HLA-G MOLECULES

241

induce HLA-G expression in the tumor microenvironment. In fact, tumor cells may definitively lose their initial HLA-G expression after long-term in vitro cell culture. Factors that participate in ectopic activation of HLA-G gene transcription and protein expression in tumor cells remain to be defined and may implicate several cytokines, including IL-10, IFN-, , and . IL-10 is often secreted by malignant cells, and according to its immunosuppressive properties, also constitutes a way for tumor cells to evade tumor antigen-specific immune responses (Botti et al., 1998). More relevant is the demonstration that IFN- treatment induces or enhances HLA-G expression in several tumor cell lines, such as the U937 leukemia monohistiocytic cell-line (Yang et al., 1996; Amiot et al., 1998), the T98G glioblastoma cellline (Maier et al., 1999; Viendl et al., 2002), the JEG-3 (Yang et al., 1995), and BeWo (Hamai et al., 1999b) choriocarcinoma cell lines. It should be noted that the U937 monohistiocytic cell-line expresses HLA-G1 when incubated with IFN- , IFN- plus IL-2, or IFN- plus GM-CSF (Amiot et al., 1998). However, constitutive expression of HLA-G has been described in few tumor cell-lines. First, melanoma cell-lines have been described to express HLA-G isoforms (Paul et al., 1998; Adrian Cabestre et al., 1999b). Also, HLA-G1 cell-surface expression and HLA-G5 secretion were maintained in a tumor cell line established from an HLA-G-positive renal cell carcinoma lesion (Ibrahim et al., 2001a). Of note, IFN-(, , and ) treatment of this renal tumor cell line enhances HLA-G1 cell-surface expression (Ibrahim et al., 2001a). Recently, several glioma cell-lines were found to highly transcribe the HLA-G gene and to express the HLA-G1 protein on their cell-surfaces (Wiendl et al., 2002). Interestingly, HLA-G1 cell-surface expression could be induced on initially HLA-G protein-negative glioma cell lines upon IFN- treatment (Wiendl et al., 2002). Short-term tumor cell-lines from patients with advanced ovarian carcinoma were described as expressing an HLA-G protein whose level is enhanced after IFN- treatment (Malmberg et al., 2002). Analysis of MHC abnormalities in non-Hodgkin lymphoma showed that three out of 50 lymphomas tested expressed surface HLA-G1. These HLA-G1positive cases correspond to HLA class I-defective lymphomas (Dreńou et al., 2002). In the same way, HLA-G1 cell-surface expression was analyzed in 40 leukemia samples of various subtypes (i.e., acute lymphoblastic, acute myeloblastic, chronic myeloblastic, and chronic lymphocytic leukemia). Although none of the leukemia cells spontaneously expressed surface HLA-G1, 21% of them became HLA-G1-positive upon IFN- stimulation (Mizuno et al., 2000). Taken together, these results show that, although it is a rare event, HLA-G may be found in tumor cell-lines, allowing investigation of its functional role in human cancer.

242


D. FUNCTIONAL SIGNIFICANCE

OF

HLA-G EXPRESSION

IN

TUMORS

The immunosuppressive properties of HLA-G, as described above, has led to the hypothesis that malignant cells could be protected by HLA-G expression from anti-tumor immune responses. In vitro studies with melanoma cell-lines that express various HLA-G isoforms were first found to be resistant to lysis from the NK cell-line YT2C2-PR (Paul et al., 1998; Adrian Cabestre et al., 1999b). This NK cell-line does not express known inhibitory receptors that interact with HLA-E and/or classical HLA class I molecules, but bears the KIR2DL4 receptor, which specifically binds HLA-G (Rouas-Freiss et al., 1999; Riteau et al., 2001a). In addition, several glioma cell-lines expressing the full-length HLA-G1 isoform on their cell-surfaces, either constitutively or after induction by IFN- , were found to be protected from alloreactive cytolysis (Wiendl et al., 2002). This protective effect could be reversed by blocking HLA-G1 with an anti-HLA-G1 mAb. Finally, these results were confirmed by utilizing HLA-G-negative tumor cell-lines, such as melanoma (Riteau et al., 2001c) or glioma (Wiendl et al., 2002) cell-lines in which HLA-G1, -G2, -G3, -G4, or -G5 cDNA has been transfected. In both cases, such gene transfer rendered the tumor cells highly resistant to lysis by NK- and Ag-specific CTL cells (Riteau et al., 2001c), inhibited the alloproliferative response, and prevented efficient priming of cytotoxic T cells (Wiendl et al., 2002). Of particular interest, analysis of the balance between the activating signal delivered by the stress-inducible molecule MICA and the inhibitory signal generated by HLA-G1 on NK cell-mediated lysis of a melanoma cell-line expressing both molecules, showed that HLA-G1 counteracts the triggering signal of MICA (Menier et al., 2002). This finding suggests that in vivo, overexpression of inhibitory ligands, such as HLA-G, by tumor cells may bypass activating signal(s), such as that mediated by MICA, thereby favoring tumor progression. Interestingly, the tumor tissue-distribution of MICA is common with that of HLA-G (Groh et al., 1999; Pende et al., 2001). Another tumor situation in which HLA-G expression would have a biological relevance is 2-microglobulin-deficient tumors. Indeed, while HLA-A, -B, -C, -E, -G1, and -G5 molecules will not be properly folded and expressed on the cell-surface of 2-microglobulin-deficient tumors, expression of the truncated HLA-G isoforms might still occur. In such cases, these isoforms would facilitate escape from NK cell immunosurveillance of these otherwise NK-susceptible tumor targets. Finally, cytokines, such as IFN and IL-10, which have been described to upregulate HLA-G1 cell-surface expression, are also detected in tumor sites (Knoefel et al., 1997; Asadullah et al., 1998; Botti et al., 1998) and found to correlate with HLA-G protein expression by malignant cells (Urosevic et al.,

HLA-G MOLECULES

243

2001, 2002b). It should be stressed that some of these cytokines are used for immunotherapy of malignant diseases (Agarwala and Kirkwood, 1996) and may contribute to the induction of HLA-G expression at the tumor site, thus conferring immune-protection to tumor cells and facilitating their spreading in the host (Wagner et al., 2000; Ugurel et al., 2001). REFERENCES Adrian Cabestre, F., Lefebvre, S., Moreau, P., Rouas-Freiss, N., Dausset, J., Carosella, E. D., and Paul, P. (1999a). Sem. Cancer Biol. 9, 27. Adrian Cabestre, F., Moreau, P., Riteau, B., Ibrahim, E. C., Le Danff, C., Dausset, J., RouasFreiss, N., Carosella, E. D., and Paul, P. (1999b). J. Reprod. Immunol. 43, 183. Agarwala, S. S., and Kirkwood, J. M. (1996). Curr. Opin. Oncol. 8, 167. Aikhionbare, F. O., Hodge, T., Kuhn, L., Bulterys, M., Abrams, E. J., and Bond, V. C. (2001). Aids 15, 2196. Aldrich, C., Wambebe, C., Odama, L., Di Rienzo, A., and Ober, C. (2002). Hum. Immunol. 63, 405. Aldrich, C. L., Stephenson, M. D., Karrison, T., Odem, R. R., Branch, D. W., Scott, J. R., Schreiber, J. R., and Ober, C. (2001). Mol. Hum. Reprod. 7, 1167. Allan, D. S. J., Colonna, M., Lanier, L., Churakova, T., Abrams, J. S., Ellis, S. A., McMichael, A. J., and Braud, V. M. (1999). J. Exp. Med. 189, 1149. Alvarez, M., Santos, P., Martinho, A., Simoes, O., Abade, A., and Breda-Coimbra, H. (1999). Transplant Proc. 31, 1829. Amiot, L., Onno, M., Drenou, B., Monvoisin, C., and Fauchet, R. (1998). Hum. Immunol. 59, 524. Amiot, L., Onno, M., Renard, I., Drenou, B., Guillaudeux, T., Le Bouteiller, P., and Fauchet, R. (1996). Tissue Antigens 48, 609. Aractingi, S., Berkane, N., Bertheau, P., Le Goue´, C., Dausset, J., Uzan, S., and Carosella, E. D. (1998). Lancet 352, 1898. Aractingi, S., Briand, N., Le Danff, C., Viguier, M., Bachelez, H., Michel, L., Dubertret, L., and Carosella, E. D. (2001). Am. J. Pathol. 159, 71. Aractingi, S., Kanitakis, J., Euvrard, S., Le Danff, C., and Carosella, E. D. (2003). Int. J. Cancer 106, 232. Arnaiz-Villena, A., Martinez-Laso, J., Alvarez, M., Castro, M. J., Varela, E., Gomez-Casado, E., Suarez, B., Recio, M. J., Vargas-Alarcon, G., and Morales, P. (1997). Immunogenetics 46, 251. Arnaiz-Villena, A., Martinez-Laso, J., Castro, M. J., Morales, P., Moscoso, J., Varela, P., GomezCasado, E., and Allende, L. M. (2001). Immunol. Rev. 183, 65. Arnaiz-Villena, A., Morales, P., Gomez-Casado, E., Castro, M. J., Varela, P., Rojo-Amigo, R., and Martinez-Laso, J. (1999). J. Reprod. Immunol. 43, 111. Arngrimsson, R., Bjornsson, S., Geirsson, R. T., Bjornsson, H., Walker, J. J., and Snaedal, G. (1990). Br. J. Obstet. Gynaecol. 97, 762. Asadullah, K., Docke, W. D., Volk, H. D., and Sterry, W. (1998). Exp. Dermatol. 7, 314. Bainbridge, D. R., Ellis, S. A., and Sargent, I. L. (2000a). J. Reprod. Immunol. 48, 17. Bainbridge, D. R., Ellis, S. A., and Sargent, I. L. (2000b). J. Reprod. Immunol. 47, 1. Bainbridge, D. R. J., Ellis, S. A., and Sargent, I. L. (2000c). J. Reprod. Immunol. 47, 1. Bakker, A., Phillips, J., Figdor, C., and Lanier, L. (1998). J. Immunol. 160, 5239. Bamberger, A. M., Jenatschke, S., Schulte, H. M., Loning, T., and Bamberger, M. C. (2000). J. Clin. Endocrinol. Metab. 85, 3932. Bennett, W. A., Lagoo-Deenadayalan, S., Whitworth, N. S., Brackin, M. N., Hale, E., and Cowan, B. D. (1997). Early Pregnancy 3, 190.

244


Bird, A. (2002). Genes Dev. 16, 6. Blaschitz, A., Lenfant, F., Mallet, V., Hartmann, M., Bensussan, A., Geraghty, D. E., Le Bouteiller, P., and Dohr, G. (1997). Eur. J. Immunol. 27, 3380. Botti, C., Seregni, E., Ferrari, L., Martinetti, A., and Bombardieri, E. (1998). Int. J. Biol. Markers, 13, 51. Boyson, J. E., Erskine, R., Whitman, M. C., Chiu, M., Lau, J. M., Koopman, L. A., Valter, M. M., Angelisova, P., Horejsi, V., and Strominger, J. L. (2002). Proc. Natl. Acad. Sci. USA 99, 16180–16185. Braud, V. M., Allan, D. S., O’Callaghan, C. A., Soderstrom, K., D’Andrea, A., Ogg, G. S., Lazetic, S., Young, N. T., Bell, J. I., Phillips, J. H., Lanier, L. L., and McMichael, A. J. (1998). Nature 391(6669), 795. Cabestre, F. A., Lefebvre, S., Moreau, P., Rouas-Friess, N., Dausset, J., Carosella, E. D., and Paul, P. (1999). Semin. Cancer Biol. 9, 27. Carosella, E. D., Moreau, P., Aractingi, S., and Rouas-Freiss, N. (2001). Trends Immunol. 22, 553. Carosella, E. D., Paul, P., Moreau, P., and Rouas-Freiss, N. (2000). Immunol. Today 21, 532. Carreiras, M., Montagnani, S., and Layrisse, Z. (2002). Am. J. Reprod. Immunol. 48, 176. Castro, M. J., Morales, P., Martinez-Laso, J., Allende, L., Rojo-Amigo, R., Gonzalez-Hevilla, M., Varela, P., Moreno, A., Garcia-Berciano, M., and Arnaiz-Villena, A. (2000a). Hum. Immunol. 61, 1157. Castro, M. J., Morales, P., Martinez-Laso, J., Allende, L., Rojo-Amigo, R., Gonzalez-Hevilla, M., Varela, P., Moscoso, J., Garcia-Berciano, M., and Arnaiz-Villena, A. (2000b). Hum. Immunol. 61, 1164. Castro, M. J., Morales, P., Rojo-Amigo, R., Martnez-Laso, J., Allende, L., Varela, P., GarciaBerciano, M., Guillen-Perales, J., and Arnaiz-Villena, A. (2000c). Tissue Antigens 56, 232. Chang, C. C., Ciubotariu, R., Manavalan, J. S., Yuan, J., Colovai, A. I., Piazza, F., Lederman, S., Colonna, M., Cortesini, R., Dalla-Favera, R., and Suciu-Foca, N. (2002). Nat. Immunol. 3, 237. Chu, W., Fant, M. E., Geraghty, D. E., and Hunt, J. S. (1998). Hum. Immunol. 59, 435. Chu, W., Gao, J., Murphy, W. J., and Hunt, J. S. (1999a). Hum. Immunol. 60, 1113. Chu, W., Yang, Y., Geraghty, D. E., and Hunt, J. S. (1999b). J. Reprod. Immunol. 42, 1. Colbern, G. T., Chiang, M. H., and Main, E. K. (1994). Am. J. Obstet. Gynecol. 170, 1244. Colonna, M., Navarro, F., Bellon, T., Llano, M., Samaridis, J., Angman, L., Cella, M., and LopezBotet, M. (1997). J. Exp. Med. 186, 1809. Colonna, M., Samaridis, J., Cella, M., Angman, L., Allen, R. L., O’Callaghan, C., Dunbar, R., Ogg, G. S., Cerundolo, V., and Rolink, A. (1998). J. Immunol. 160, 3096. Contini, P., Ghio, M., Poggi, A., Filaci, G., Indiveri, F., Ferrone, S., and Puppo, F. (2003). Eur. J. Immunol. 33, 125. Copeman, J., Han, R. N., Caniggia, I., McMaster, M., Fisher, S. J., and Cross, J. C. (2000). Biol. Reprod. 62, 1543. Corell, A., Morales, P., Martinez-Laso, J., Martin-Villa, J., Varela, P., Paz-Artal, E., Allende, L. M., Rodriguez, C., and Arnaiz-Villena, A. (1994). Hum. Immunol. 41, 52. Cosman, D., Fanger, N., Borges, L., Kubin, M., Chin, W., Peterson, L., and Hsu, M.-L. (1997). Immunity 7, 273. Crisa, L., McMaster, M. T., Ishii, J. K., Fisher, S. J., and Salomon, D. R. (1997). J. Exp. Med. 186(2), 289. Davies, B., Hiby, S., Gardner, L., Loke, Y. W., and King, A. (2001). Am. J. Reprod. Immunol. 45, 103. Davis, D. M., Reyburn, H. T., Pazmany, L., Chiu, I., Mandelboim, O., and Strominger, J. L. (1997). Eur. J. Immunol. 27, 2714. Diehl, M., Munz, C., Keilholz, W., Stevanovic, S., Holmes, N., Loke, Y. W., and Rammensee, H. G. (1996). Curr. Biol. 6, 305.

HLA-G MOLECULES

245

Dinarello, C. A. (2000). Chest 118, 503. Dorling, A., Monk, N., and Lechler, R. (2000). Transplant Proc. 32, 938. Dreńou, B., Le Friec, G., Bernard, M., Pangault, C., Grosset, J. M., Lamy, T., Fauchet, R., and Amiot, L. (2002). Br. J. Haematol. 119, 417. Dummer, W., Becker, J. C., Schwaaf, A., Leverkus, M., Moll, T., and Brocker, E. B. (1995). Melanoma. Res. 5, 67. Easterfield, A. J., Austen, B. M., and Westwood, O. M. (2001). J. Reprod. Immunol. 50, 19. Ellis, S. A., Sargent, I. L., Redman, C. W. G., and McMichael, A. J. (1986). Immunology 59, 595. El-Osta, A., and Wolffe, A. P. (2000). Gene Expr. 9, 63. Emmer, P. M., Steegers, E. A., Kerstens, H. M., Bulten, J., Nelen, W. L., Boer, K., and Joosten, I. (2002). Hum. Reprod. 17, 1072. Forte, P., Matter-Reissmann, U. B., Strasser, M., Schneider, M. K., and Seebach, J. D. (2000). Hum. Immunol. 61, 1066. Forte, P., Pazmany, L., Matter-Reissmann, U. B., Stussi, G., Schneider, M. K., and Seebach, J. D. (2001). J. Immunol. 167, 6002. Fournel, S., Aguerre-Girr, M., Campan, A., Salauze, L., Berrebi, A., Lone, Y. C., Lenfant, F., and Le Bouteiller, P. (1999). Am. J. Reprod. Immunol. 42, 22. Fournel, S., Aguerre-Girr, M., Huc, X., Lenfant, F., Alam, A., Toubert, A., Bensussan, A., and Le Bouteiller, P. (2000a). J. Immunol. 164, 6100. Fournel, S., Huc, X., Aguerre-Girr, M., Solier, C., Legros, M., Praud-Brethenou, C., Moussa, M., Chaouat, G., Berrebi, A., Bensussan, A., Lenfant, F., and Le Bouteiller, P. (2000b). Tissue Antigens 55, 510. Frumento, G., Franchello, S., Geraghty, E., and Ferrara, G. B. (1999). Transplant Proc. 31, 1847. Frumento, G., Franchello, S., Palmisano, G. L., Nicotra, M. R., Giacomini, P., Loke, Y. W., Geraghty, D. E., Maio, M., Manzo, C., Natali, P. G., and Ferrara, G. B. (2000). Tissue Antigens 56, 30. Fujii, T., Ishitani, A., and Geraghty, D. E. (1994). J. Immunol. 153, 5516. Fukushima, Y., Oshika, Y., Nakamura, M., Tokunaga, T., Hatanaka, H., Abe, Y., Yamazaki, H., Kijima, H., Ueyama, Y., and Tamaoki, N. (1998). Int. J. Mol. Med. 2, 349. Fuzzi, B., Rizzo, R., Criscuoli, L., Noci, I., Melchiorri, L., Scarselli, B., Bencini, E., Menicucci, A., and Baricordi, O. R. (2002). Eur. J. Immunol. 32, 311. Gao, G. F., Willcox, B. E., Wyer, J. R., Boulter, J. M., O’Callaghan, C. A., Maenaka, K., Stuart, D. I., Jones, E. Y., Van Der Merwe, P. A., Bell, J. I., and Jakobsen, B. K. (2000). J. Biol. Chem. 275, 15232. Gati, A., Guerra, N., Giron-Michel, J., Azzarone, B., Angevin, E., Moretta, A., Chouaib, S., and Caignard, A. (2001). Cancer Res. 61, 3240. Genbacev, O., Zhou, Y., Ludlow, J. W., and Fisher, S. J. (1997). Science 277, 1669. Geraghty, D. E., Koller, B. H., and Orr, H. T. (1987). Proc. Natl. Acad. Sci. USA 84, 9145. Gobin, S. J., Biesta, P., De Steenwinkel, J. E., Datema, G., and Van Den Elsen, P. J. (2002). J. Biol. Chem. 277, 39525. Gobin, S. J., Keijsers, V., Cheong, C., van Zutphen, M., and Van den Elsen, P. J. (1999a). Transplant Proc. 31, 1857. Gobin, S. J., Keijsers, V., van Zutphen, M., and van den Elsen, P. J. (1998a). J. Immunol. 161, 2276. Gobin, S. J., Peijnenburg, A., Keijsers, V., and van den Elsen, P. J. (1997a). Immunity 6, 601. Gobin, S. J., Peijnenburg, A., van Eggermond, M., van Zutphen, M., van den Berg, R., and van den Elsen, P. J. (1998b). Immunity 9, 531. Gobin, S. J., and van den Elsen, P. J. (2000). Hum. Immunol. 61, 1102. Gobin, S. J., van Zutphen, M., Westerheide, S. D., Boss, J. M., and van den Elsen, P. J. (2001). J. Immunol. 167, 5175.

246


Gobin, S. J., van Zutphen, M., Woltman, A. M., and van den Elsen, P. J. (1999b). J. Immunol. 163, 1428. Gobin, S. J., Wilson, L., Keijsers, V., and Van den Elsen, P. J. (1997b). J. Immunol. 158, 3587. Goldman, M., and Stordeur, P. (1997). Eur. Cytokine Netw. 8, 301. Goldman-Wohl, D., Ariel, I., Greenfield, C., Hochner-Celnikier, D., Lavy, Y., and Yagel, S. (2001). Am. J. Obstet. Gynecol. 185, 476. Goldman-Wohl, D. S., Ariel, I., Greenfield, C., Hanoch, J., and Yagel, S. (2000). Mol. Hum. Reprod. 6, 535. Groh, V., Rhinehart, R., Secrist, H., Bauer, S., Grabstein, K. H., and Spies, T. (1999). Proc. Natl. Acad. Sci. USA 96, 6879. Guerra, N., Guillard, M., Angevin, E., Echchakir, H., Escudier, B., Moretta, A., Chouaib, S., and Caignard, A. (2000). Blood 95, 2883. Hahn, S., and Holzgreve, W. (2002). Clin. Obstet. Gynecol. 45, 649. Hamai, Y., Fujii, T., Miki, A., Geraghty, D. E., Harada, I., Takai, Y., Kozuma, S., Tsutsumi, O., and Taketani, Y. (1999a). Am. J. Reprod. Immunol. 41, 293. Hamai, Y., Fujii, T., Yamashita, T., Miki, A., Hyodo, H., Kozuma, S., Geraghty, D. E., and Taketani, Y. (1999b). Am. J. Reprod. Immunol. 41, 153. Hamai, Y., Fujii, T., Yamashita, T., Miki, A., Kozuma, S., Geraghty, D. E., and Taketani, Y. (1998). Am. J. Reprod. Immunol. 40, 408. Hammer, A., Hutter, H., Blaschitz, A., Mahnert, W., Hartmann, M., Uchanska-Ziegler, B., Ziegler, A., and Dohr, G. (1997a). Am. J. Reprod. Immunol. 37, 161. Hammer, A., Hutter, H., and Dohr, G. (1997b). Am. J. Reprod. Immunol. 38, 150. Hara, N., Fujii, T., Yamashita, T., Kozuma, S., Okai, T., and Taketani, Y. (1996). AJRI 36, 349. Harrison, G. A., Humphrey, K. E., Jakobsen, I. B., and Cooper, D. W. (1993). Hum. Mol. Genet. 2, 2200. Hennessy, A., Pilmore, H. L., Simmons, L. A., and Painter, D. M. (1999). J. Immunol. 163, 3491. Hiby, S., King, A., Sharkey, A., and Loke, Y. W. (1999a). Tissue Antigens 53, 1. Hiby, S. E., King, A., Sharkey, A., and Loke, Y. W. (1999b). Tissue Antigens 53, 1. Horuzsko, A., Lenfant, F., Munn, D., and Mellor, A. (2001). Int. Immunol. 13, 385. Horuzsko, A., Portik-Dobos, V., Hansen, K. A., Markowitz, R. B., Helman, S. W., and Mellor, A. L. (1999). Transplant Proc. 31, 1860. Houlihan, J. M., Biro, P. A., Harper, H. M., Jenkinson, H. J., and Holmes, C. H. (1995). J. Immunol. 154, 5665. Humphrey, K. E., Harrison, G. A., Cooper, D. W., Wilton, A. N., Brennecke, S. P., and Trudinger, B. J. (1995). Br. J. Obstet. Gynaecol. 102, 707. Hunt, J. S., Jadhav, L., Chu, W., Geraghty, D. E., and Ober, C. (2000). Am. J. Obstet. Gynecol. 183, 682. Hurks, H. M., Valter, M. M., Wilson, L., Hilgert, I., van den Elsen, P. J., and Jager, M. J. (2001). Invest. Ophthalmol. Vis. Sci. 42, 3081. Hutter, H., Hammer, A., Blaschitz, A., Hartmann, M., Ebbesen, P., Dohr, G., Ziegler, A., and Uchanska-Ziegler, B. (1996). Cell Tissue Res. 286, 439. Hviid, T. V., Christiansen, O. B., Johansen, J. K., Hviid, U. R., Lundegaard, C., Moller, C., and Morling, N. (2001). Immunogenetics 53, 48. Hviid, T. V., Hylenius, S., Hoegh, A. M., Kruse, C., and Christiansen, O. B. (2002). Tissue Antigens 60, 122. Hviid, T. V., Meldgaard, M., Sorensen, S., and Morling, N. (1997). J. Reprod. Immunol. 35, 31. Hviid, T. V., Sorensen, S., and Morling, N. (1999). Hum. Immunol. 60, 1237. Ibrahim, E., Guerra, N., Lacombe, M., Angevin, E., Chouaib, S., Carosella, E., Caignard, A., and Paul, P. (2001a). Cancer Res. 61, 6838.

HLA-G MOLECULES

247

Ibrahim, E. C., Guerra, N., Lacombe, M. J., Angevin, E., Chouaib, S., Carosella, E. D., Caignard, A., and Paul, P. (2001b). Cancer Res. 61, 6838. Ibrahim, E. C., Morange, M., Dausset, J., Carosella, E. D., and Paul, P. (2000). Cell Stress Chaperones 5, 207. Ikeda, H., Lethe´, B., Lehmann, F., Van Baren, N., Baurain, J.-F., DeSmet, C., Chambost, H., Vitale, M., Moretta, A., Boon, T., and Coulie, P. (1997). Immunity 6, 199. Ishitani, A., Kishida, M., Sageshima, N., Yashiki, S., Sonoda, S., Hayami, M., Smith, A. G., and Hatake, K. (1999). Immunogenetics 49, 808. Johnson, P. M., Christmas, S. E., and Vince, G. S. (1999). Hum. Reprod. 2(14 Suppl), 26. Jun, Y., Kim, E., Jin, M., Sung, H. C., Han, H., Geraghty, D. E., and Ahn, K. (2000). J. Immunol. 164, 805. Jurisicova, A., Casper, R. F., MacLusky, N. J., and Librach, C. L. (1996a). Fertil. Steril. 65, 997. Jurisicova, A., Casper, R. F., MacLusky, N. J., Mills, G. B., and Librach, C. L. (1996b). Proc. Natl. Acad. Sci. USA 93, 161. Kanai, T., Fujii, T., Kozuma, S., Yamashita, T., Miki, A., Kikuchi, A., and Taketani, Y. (2001a). Mol. Hum. Reprod. 7, 195. Kanai, T., Fujii, T., Unno, N., Yamashita, T., Hyodo, H., Miki, A., Hamai, Y., Kozuma, S., and Taketani, Y. (2001b). Am. J. Reprod. Immunol. 45, 94. Kapasi, K., Albert, S. E., Yie, S.-M., Zavazava, N., and Librach, C. L. (2000). Immunol. 101, 191. Karhukorpi, J., Laitinen, T., and Tiilikainen, A. S. (1997). Br. J. Obstet. Gynaecol. 104, 1212. Khalil-Daher, I., Riteau, B., Menier, C., Sedlik, C., Paul, P., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (1999). J. Reprod. Immunol. 43, 175. Khong, T. Y., De Wolf, F., Robertson, W. B., and Brosens, I. (1986). Br. J. Obstet. Gynaecol. 93, 1049. Khosrotehrani, K., Le Danff, C., Reynaud, B., Dubertret, L., Carosella, E. D., and Aractingi, S. (2001). J. Invest. Dermatol. 117, 750. King, A., Allan, D. S., Bowen, M., Powis, S. J., Joseph, S., Verma, S., Hiby, S. E., McMichael, A. J., Loke, Y. W., and Braud, V. M. (2000). Eur. J. Immunol. 30, 1623. King, A., Boocock, C., Sharkey, A. M., Gardner, L., Beretta, A., Siccardi, A. G., and Loke, Y. W. (1996). J. Immunol. 156, 2068. Kirszenbaum, M., Djoulah, S., Hors, J., Le Gall, I., de Oliveira, E. B., Prost, S., Dausset, J., and Carosella, E. D. (1997). Hum. Immunol. 53, 140. Kirszenbaum, M., Moreau, P., Gluckman, E., Dausset, J., and Carosella, E. (1994). Proc. Natl. Acad. Sci. USA 91, 4209. Kirszenbaum, M., Moreau, P., Teyssier, M., Lafon, C., Gluckman, E., Dausset, J., and Carosella, E. (1995). Hum. Immunol. 43, 237. Knoefel, B., K., N., Steiner, T., Junker, K., Kosmehl, H., Rebstock, K., Reinhold, D., and Junker, U. (1997). J. Interferon Cytokine Res. 17, 95. Kovats, S., Main, E. K., Librach, C., Stubblebine, M., Fisher, S. J., and DeMars, R. (1990). Science 248, 220. Langat, D. K., Morales, P. J., Fazleabas, A. T., Mwenda, J. M., and Hunt, J. S. (2002). Immunogenetics 54, 164. Le Bouteiller, P., and Blaschitz, A. (1999). Immunol. Rev. 167, 233. Le Gal, F. A., Riteau, B., Sedlik, C., Khalil-Daher, I., Menier, C., Dausset, J., Guillet, J. G., Carosella, E. D., and Rouas-Freiss, N. (1999). Int. Immunol. 11, 1351. Lee, N., Goodlett, D. R., Ishitani, A., Marquardt, H., and Geraghty, D. E. (1998a). J. Immunol. 160, 4951. Lee, N., Llano, M., Carretero, M., Ishitani, A., Navarro, F., Lopez-Botet, M., and Geraghty, D. E. (1998b). Proc. Natl. Acad. Sci. USA 95(9), 5199.

248


Lee, N., Malacko, A. R., Ishitani, A., Chen, M. C., Bajorath, J., Marquardt, H., and Geraghty, D. E. (1995). Immunity 3, 591. Lefebvre, S., Adrian, F., Moreau, P., Gourand, L., Dausset, J., Berrih-Aknin, S., Carosella, E. D., and Paul, P. (2000). Hum. Immunol. 61, 1095. Lefebvre, S., Antoine, M., Uzan, S., McMaster, M., Dausset, J., Carosella, E. D., and Paul, P. (2002). J. Pathol. 196, 266. Lefebvre, S., Berrih-Aknin, S., Adrian, F., Moreau, P., Poea, S., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (2001). J. Biol. Chem. 276, 6133. Lefebvre, S., Moreau, P., Dausset, J., Carosella, E. D., and Paul, P. (1999a). Placenta 20, 293. Lefebvre, S., Moreau, P., Guiard, V., Ibrahim, E. C., Adrian-Cabestre, F., Menier, C., Dausset, J., Carosella, E. D., and Paul, P. (1999b). J. Reprod. Immunol. 43, 213. Liang, S., Baibakov, B., and Horuzsko, A. (2002). Eur. J. Immunol. 32, 2418. Lila, N., Amrein, C., Guillemain, R., Chevalier, P., Latremouille, C., Fabiani, J. N., Dausset, J., Carosella, E. D., and Carpentier, A. (2002). Circulation 105, 1949. Lila, N., Carpentier, A., Amrein, C., Khalil-Daher, I., Dausset, J., and Carosella, E. D. (2000). Lancet 355, 2138. Lila, N., Rouas-Freiss, N., Dausset, J., Carpentier, A., and Carosella, E. D. (2001). Proc. Natl. Acad. Sci. USA 98, 12150. Lim, K. H., Zhou, Y., Janatpour, M., McMaster, M., Bass, K., Chun, S. H., and Fisher, S. J. (1997). Am. J. Pathol. 151, 1809. Llano, M., Lee, N., Navarro, F., Garcia, P., Albar, J. P., Geraghty, D. E., and Lopez-Botet, M. (1998). Eur. J. Immunol. 28, 2854. Loke, Y. W., King, A., Burrows, T., Gardner, L., Bowen, M., Hiby, S., Howlett, S., Holmes, N., and Jacobs, D. (1997). Tissue Antigens 50, 135. Lozano, J. M., Gonzalez, R., Kindelan, J. M., Rouas-Freiss, N., Caballos, R., Dausset, J., Carosella, E. D., and Pena, J. (2002). Aids 16, 347. Maier, S., Geraghty, D., and Weiss, E. (1999). Transplant Proc. 31, 1849. Mallet, V., Blaschitz, A., Crisa, L., Schmitt, C., Fournel, S., King, A., Loke, Y. W., Dohr, G., and Le Bouteiller, P. (1999a). Int. Immunol. 11, 889. Mallet, V., Fournel, S., Schmitt, C., Campan, A., Lenfant, F., and Le Bouteiller, P. (1999b). J. Reprod. Immunol. 43, 225. Mallet, V., Proll, J., Solier, C., Aguerre-Girr, M., DeRossi, M., Loke, Y. W., Lenfant, F., and Le Bouteiller, P. (2000). Hum. Immunol. 61, 212. Malmberg, K. J., Levitsky, V., Norell, H., Matos, C. T., Carlsten, M., Schedvins, K., Rabbani, H., Moretta, A., Soderstrom, K., Levitskaya, J., and Kiessling, R. (2002). J. Clin. Invest. 110, 1515. Marchal-Bras-Goncalves, R., Rouas-Freiss, N., Connan, F., Choppin, J., Dausset, J., Carosella, E., Kirszenbaum, M., and Guillet, J. G. (2001). Transplant Proc. 33, 2355. Masternak, K., Muhlethaler-Mottet, A., Villard, J., Peretti, M., and Reith, W. (2000). Rev. Immunogenet. 2, 267. Masternak, K., and Reith, W. (2002). EMBO J. 21, 1379. Matsunami, K., Miyagawa, S., Nakai, R., Murase, A., and Shirakura, R. (2001). Clin. Exp. Immunol. 126, 165. Matsunami, K., Miyagawa, S., Nakai, R., Yamada, M., and Shirakura, R. (2002). Transplantation 73, 1582. Matte, C., Lacaille, J., Zijenah, L., Ward, B., and Roger, M. (2000). Hum. Immunol. 61, 1150. Matte, C., Lacaille, J., Zijenah, L., Ward, B., and Roger, M. (2002). Hum. Immunol. 63, 495. McMaster, M., Zhou, Y., Shorter, S., Kapasi, K., Geraghty, D., Lim, K. H., and Fisher, S. (1998). J. Immunol. 160, 5922. McMaster, M. T., Librach, C. L., Zhou, Y., Lim, K. H., Janatpour, M. J., DeMars, R., Kovats, S., Damsky, C., and Fisher, S. J. (1995). J. Immunol. 154, 3771.

HLA-G MOLECULES

249

Menicucci, A., Noci, I., Fuzzi, B., Criscuoli, L., Scarselli, G., Baricordi, O., and Mattiuz, P. (1999). Hum. Immunol. 60, 1054. Menier, C., Riteau, B., Carosella, E. D., and Rouas-Freiss, N. (2002). Int. J. Cancer 100, 63. Menier, C., Riteau, B., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (2000). Hum. Immunol. 61, 1118. Menier, C., Saez, B., Horejsi, V., Martinozzi, S., Krawice-Radanne, I., Bruel, S., Le Dannff, C., Reboul, M., Hilgert, I., Rabreau, M., Larrad, M. L., Pla, M., Carosella, E. D., and Rouas-Freiss, N. (2003). Hum. Immunol. 64, 315. Messer, G., Zemmour, J., Orr, H. T., Parham, P., Weiss, E. H., and Girdlestone, J. (1992). J. Immunol. 148, 4043. Minguela, A., Torio, A., Marin, L., Muro, M., Villar, L. M., Diaz, J., Ramirez, P., Parrilla, P., Garcia-Alonso, A. M., and Alvarez-Lopez, M. R. (1999). Hum. Immunol. 60, 500. Mizuno, S., Emi, N., Kasai, M., Ishitani, A., and Saito, H. (2000). Br. J. Haematol. 111, 280. Moreau, P., Adrian-Cabestre, F., Menier, C., Guiard, V., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (1999). Int. Immunol. 11, 803. Moreau, P., Carosella, E., Teyssier, M., Prost, S., Gluckman, E., Dausset, J., and Kirszenbaum, M. (1995). Hum. Immunol. 43, 231. Moreau, P., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (2002). Biol. Reprod. 67, 1375. Moreau, P., Faure, O., Lefebvre, S., Ibrahim, E. C., O’Brien, M., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (2001). Transplant Proc. 33, 2277. Moreau, P., Lefebvre, S., Gourand, L., Dausset, J., Carosella, E. D., and Paul, P. (1998). Hum. Immunol. 59, 751. Moreau, P., Paul, P., Gourand, L., Prost, S., Dausset, J., Carosella, E., and Kirszenbaum, M. (1997). Hum. Immunol. 52, 41. Moreau, P., Mouillot, G., Rousseau, P., Marcou, C., Dausset, J., and Carosella, E. D. (2003). Proc. Natl. Acad. Sci. USA 100, 1191. Munz, C., Stevanovic, S., and Rammensee, H. G. (1999). J. Reprod. Immunol. 43, 139. Navarro, F., Llano, M., Bellon, T., Colonna, M., Geraghty, D. E., and Lopez-Botet, M. (1999). Eur. J. Immunol. 29, 277. Neefjes, J. J., and Momburg, F. (1993). Curr. Opin. Immunol. 5, 27. Ober, C., Aldrich, C., Rosinsky, B., Robertson, A., Walker, M. A., Willadsen, S., Verp, M. S., Geraghty, D. E., and Hunt, J. S. (1998). Placenta 19, 127. O’Brien, M., McCarthy, T., Jenkins, D., Paul, P., Dausset, J., Carosella, E. D., and Moreau, P. (2001). Cell. Mol. Life Sci. 58, 1943. Onno, M., Amiot, L., Bertho, N., Drenou, B., and Fauchet, R. (1997). Tissue Antigens 49, 356. Onno, M., Guillaudeux, T., Amiot, L., Renard, I., Drenou, B., Hirel, B., Girr, M., Semana, G., Le Bouteiller, P., and Fauchet, R. (1994). Hum. Immunol. 41, 79. Onno, M., Le Friec, G., Pangault, C., Amiot, L., Guilloux, V., Drenou, B., Caulet-Maugendre, S., Andre, P., and Fauchet, R. (2000a). Hum. Immunol. 61, 1086. Onno, M., Pangault, C., Le Friec, G., Guilloux, V., Andre, P., and Fauchet, R. (2000b). J. Immunol. 164, 6426. Opal, S. M., and DePalo, V. A. (2000). Chest 117, 1162. Palmisano, G. L., Pistillo, M. P., Fardin, P., Capanni, P., Nicolo, G., Salvi, S., Spina, B., Pasciucco, G., and Ferrara, G. B. (2002). Hum. Immunol. 63, 969. Pangault, C., Amiot, L., Caulet-Maugendre, S., Brasseur, F., Burtin, F., Guilloux, V., Drenou, B., Fauchet, R., and Onno, M. (1999). Tissue Antigens 53, 335. Pangault, C., Le Friec, G., Caulet-Maugendre, S., Lena, H., Amiot, L., Guilloux, V., Onno, M., and Fauchet, R. (2002). Hum. Immunol. 63, 83. Paradinas, F. J., Browne, P., Fisher, R. A., Foskett, M., Bagshawe, K. D., and Newlands, E. (1996). Histopathology 28, 101.

250


Park, B., Lee, S., Kim, E., Chang, S., Jin, M., and Ahn, K. (2001). Immunity 15, 213. Paul, P., Adrian Cabestre, F., Le Gal, F.-A., Khalil-Daher, I., Le Danff, C., Schmid, M., Mercier, S., Avril, M.-F., Dausset, J., Guillet, J.-G., and Carosella, E. D. (1999). Cancer Res. 59, 1954. Paul, P., Cabestre, F. A., Ibrahim, E. C., Lefebvre, S., Khalil-Daher, I., Vazeux, G., Quiles, R. M., Bermond, F., Dausset, J., and Carosella, E. D. (2000a). Hum. Immunol. 61, 1138. Paul, P., Rouas-Freiss, N., Khalil-Daher, I., Moreau, P., Riteau, B., Le Gal, F. A., Avril, M. F., Dausset, J., Guillet, J. G., and Carosella, E. D. (1998). Proc. Natl. Acad. Sci. USA 95, 4510. Paul, P., Rouas-Freiss, N., Moreau, P., Adrian Cabestre, F., Menier, C., Khalil-Daher, I., Pangault, C., Onno, M., Fauchet, R., Martinez-Laso, J., Morales, P., Arnaiz-Villena, A., Giacomini, P., Natali, P. G., Frumento, G., Ferrara, G. B., McMaster, M., Fisher, S., Schust, D., Ferrone, S., Dausset, J., Geraghty, D., and Carosella, E. D. (2000b). Hum. Immunol. 61, 1177. Pende, D., Cantoni, C., Rivera, P., Vitale, M., Castriconi, R., Marcenaro, S., Nanni, M., Biassoni, R., Bottino, C., Moretta, A., and Moretta, L. (2001). Eur. J. Immunol. 31(4), 1076. Pfeiffer, K. A., Fimmers, R., Engels, G., van der Ven, H., and van der Ven, K. (2001). Mol. Hum. Reprod. 7, 373. Polakova, K., and Russ, G. (2000). Neoplasma 47, 342. Ponte, M., Cantoni, C., Biassoni, R., Tradori-Cappai, A., Bentivoglio, G., Vitale, C., Bertone, S., Moretta, A., Moretta, L., and Mingari, M. C. (1999). Proc. Natl. Acad. Sci. USA 96, 5674. Proll, J., Bensussan, A., Goffin, F., Foidart, J. M., Berrebi, A., and Le Bouteiller, P. (2000). Tissue Antigens 56, 479. Proll, J., Blaschitz, A., Hutter, H., and Dohr, G. (1999). Am. J. Reprod. Immunol. 42, 30. Puppo, F., Costa, M., Contini, P., Brenci, S., Cevasco, E., Ghio, M., Norelli, R., Bensussan, A., Capitanio, G. L., and Indiveri, F. (1999). Transplant Proc. 31, 1841. Rabreau, M., Rouas-Freiss, N., Landi, M., Le Danff, C., and Carosella, E. D. (2000). Hum. Immunol. 61, 1108. Rajagopalan, S., and Long, E. O. (1999). J. Exp. Med. 189, 1093. Real, L. M., Cabrera, T., Collado, A., Jimenez, P., Garcia, A., Ruiz-Cabello, F., and Garrido, F. (1999). Int. J. Cancer 81, 512. Rebmann, V., Pfeiffer, K., Passler, M., Ferrone, S., Maier, S., Weiss, E., and Grosse-Wilde, H. (1999). Tissue Antigens 53, 14. Rebmann, V., van der Ven, K., Passler, M., Pfeiffer, K., Krebs, D., and Grosse-Wilde, H. (2001). Tissue Antigens 57, 15. Rieger, L., Hofmeister, V., Probe, C., Dietl, J., Weiss, E. H., Steck, T., and Kammerer, U. (2002). Mol. Hum. Reprod. 8, 255. Riteau, B., Menier, C., Khalil-Daher, I., Martinozzi, S., Pla, M., Dausset, J., Carosella, E. D., and Rouas-Freiss, N. (2001a). Int. Immunol. 13, 193. Riteau, B., Menier, C., Khalil-Daher, I., Sedlik, C., Dausset, J., Rouas-Freiss, N., and Carosella, E. D. (1999). J. Reprod. Immunol. 43, 203. Riteau, B., Moreau, P., Menier, C., Khalil-Daher, I., Khosrotehrani, K., Bras-Goncalves, R., Paul, P., Dausset, J., Rouas-Freiss, N., and Carosella, E. D. (2001b). Transplant Proc. 33, 2360. Riteau, B., Rouas-Freiss, N., Menier, C., Paul, P., Dausset, J., and Carosella, E. D. (2001c). J. Immunol. 166, 5018. Roberts, J., and Redman, C. (1993). Lancet 341, 1447. Roth, I., Corry, D. B., Locksley, R. M., Abrams, J. S., Litton, M. J., and Fisher, S. J. (1996). J. Exp. Med. 184, 539. Rouas-Freiss, N., Goncalves, R. M., Menier, C., Dausset, J., and Carosella, E. D. (1997a). Proc. Natl. Acad. Sci. USA 94, 11520. Rouas-Freiss, N., Khalil-Daher, I., Riteau, B., Menier, C., Paul, P., Dausset, J., and Carosella, E. D. (1999). Sem. Cancer Biol. 9, 3.

HLA-G MOLECULES

251

Rouas-Freiss, N., Marchal, R. E., Kirszenbaum, M., Dausset, J., and Carosella, E. D. (1997b). Proc. Natl. Acad. Sci. USA 94, 5249. Rousseau, P., Paul, P., O’Brien, M., Dausset, J., Carosella, E. D., and Moreau, P. (2000). Hum. Immunol. 61, 1132. Ryan, A. F., Grendell, R. L., Geraghty, D. E., and Golos, T. G. (2002). J. Immunol. 169, 673. Saito, S., Sakai, M., Sasaki, Y., Tanebe, K., Tsuda, H., and Michimata, T. (1999). Clin. Exp. Immunol. 117, 550. Sasaki, H., Xu, X. C., and Mohanakumar, T. (1999). J. Immunol. 163, 6301. Sasaki, H., Xu, X. C., Smith, D., Shenoy, S., Lowell, J., Howard, T., and Mohanakumar, T. (1999a). Transplant Proc. 31, 953. Sasaki, H., Xu, X. C., Smith, D. M., Howard, T., and Mohanakumar, T. (1999b). Transplantation 67, 31. Saverino, D., Fabbi, M., Ghiotto, F., Merlo, A., Bruno, S., Zarcone, D., Tenca, C., Tiso, M., Santoro, G., Anastasi, G., Cosman, D., Grossi, C. E., and Ciccone, E. (2000). J. Immunol. 165, 3742. Schmidt, C. M., Chen, H. L., Chiu, I., Ehlenfeldt, R. G., Hunt, J. S., and Orr, H. T. (1995). J. Immunol. 155, 619. Schmidt, C. M., Ehlenfeldt, R. G., Athanasiou, M. C., Duvick, L. A., Heinrichs, H., David, C. S., and Orr, H. T. (1993). J. Immunol. 151, 2633. Schust, D. J., Hill, A. B., and Ploegh, H. L. (1996). J. Immunol. 157, 3375. Schust, D. J., Tortorella, D., and Ploegh, H. L. (1999). Semin. Cancer Biol. 9, 37. Schust, D. J., Tortorella, D., Seebach, J., Phan, C., and Ploegh, H. L. (1998). J. Exp. Med. 188, 497. Shurin, M. R., Lu, L., Kalinski, P., Stewarts-Akers, A. M., and Lotze, M. T. (1999). Springer Semin. Immunopathol. 21, 339. Singer, G., Kurman, R. J., McMaster, M. T., and Shih Ie, M. (2002). Am. J. Surg. Pathol. 26, 914. Sivori, S., Parolini, S., Marcenaro, E., Millo, R., Bottino, C., and Moretta, A. (2000). Hum. Immunol. 61, 1055. Slukvin, II, Lunn, D. P., Watkins, D. I., and Golos, T. G. (2000). Proc. Natl. Acad. Sci. USA 97, 9104. Solier, C., Mallet, V., Lenfant, F., Bertrand, A., Huchenq, A., and Le Bouteiller, P. (2001). Immunogenetics 53, 617. Speiser, D. E., Pittet, M. J., Valmori, D., Dunbar, R., Rimoldi, D., Lienard, D., MacDonald, H. R., Cerottini, J.-C., Cerundolo, V., and Romero, P. (1999). J. Exp. Med. 190, 775. Spencer, J. V., Lockridge, K. M., Barry, P. A., Lin, G., Tsang, M., Penfold, M. E., and Schall, T. J. (2002). J. Virol. 76, 1285. Stallmach, T., Hebisch, G., Orban, P., and Lu, X. (1999). Virchows Arch. 434, 207. Stewart, C. L., Kaspar, P., Brunet, L. J., Bhatt, H., Gadi, I., Kontgen, F., and Abbondanzo, S. J. (1992). Nature 359, 76. Suarez, M. B., Morales, P., Castro, M. J., Fernandez, V., Varela, P., Alvarez, M., Martinez-Laso, J., and Arnaiz-Villena, A. (1997). Immunogenetics 45, 464. Sun, K., Yang, K., and Challis, J. R. (1998). Placenta 19, 353. Tarrade, A., Lai Kuen, R., Malassine, A., Tricottet, V., Blain, P., Vidaud, M., and Evain-Brion, D. (2001). Lab. Invest. 81, 1199. Taylor, R. N. (1997). Am. J. Reprod. Immunol. 37, 79. Teasdale, R. D., and Jackson, M. R. (1996). Annu. Rev. Cell. Dev. Biol. 12, 27. Ugurel, S., Rebmann, V., Ferrone, S., Tilgen, W., Grosse-Wilde, H., and Reinhold, U. (2001). Cancer 92, 369. Ugurel, S., Reinhold, U., and Tilgen, W. (2002). Onkologie 25, 129. Urosevic, M., Kurrer, M. O., Kamarashev, J., Mueller, B., Weder, W., Burg, G., Stahel, R. A., Dummer, R., and Trojan, A. (2001). Am. J. Pathol. 159, 817.

252


Urosevic, M., Trojan, A., and Dummer, R. (2002). J. Pathol. 196, 252. Urosevic, M., Willers, J., Mueller, B., Kempf, W., Burg, G., and Dummer, R. (2002). Blood 99, 609. Vale´s-Gomez, M., Reyburn, H. T., Erskine, R. A., Lopez-Botet, M., and Strominger, J. L. (1999). EMBO J. 18, 4250. van den Elsen, P. J., Gobin, S. J., van Eggermond, M. C., and Peijnenburg, A. (1998a). Immunogenetics 48, 208. van den Elsen, P. J., Peijnenburg, A., van Eggermond, M. C., and Gobin, S. J. (1998b). Immunol. Today 19, 308. van der Ven, K., Skrablin, S., Engels, G., and Krebs, D. (1998). Hum. Immunol. 59, 302. van Wijk, I. J., Griffioen, S., Tjoa, M. L., Mulders, M. A., van Vugt, J. M., Loke, Y. W., and Oudejans, C. B. (2001). Am. J. Obstet. Gynecol. 184, 991. Wagner, S. N., Rebmann, V., Willers, C. P., Grosse-Wilde, H., and Goos, M. (2000). Lancet 356, 220. Waldburger, J. M., Masternak, K., Muhlethaler-Mottet, A., Villard, J., Peretti, M., Landmann, S., and Reith, W. (2000). Immunol. Rev. 178, 148. Watkins, D. I., Chen, Z. W., Hughes, A. L., Evans, M. G., Tedder, T. F., and Letvin, N. L. (1990). Nature 346, 60. Wiendl, H., Behrens, L., Maier, S., Johnson, M. A., Weiss, E. H., and Hohlfeld, R. (2000). Ann. Neurol. 48, 679. Wiendl, H., Mitsdoerffer, M., Hofmeister, V., Wischhusen, J., Bornemann, A., Meyermann, R., Weiss, E. H., Melms, A., and Weller, M. (2002). J. Immunol. 168, 4772. Yamashita, T., Fujii, T., Tokunaga, K., Tadokoro, K., Hamai, Y., Miki, A., Kozuma, S., Juji, T., and Taketani, Y. (1999). Am. J. Reprod. Immunol. 41, 159. Yang, Y., Chu, W., Geraghty, D. E., and Hunt, J. S. (1996). J. Immunol. 156, 4224. Yang, Y., Geraghty, D. E., and Hunt, J. S. (1995). J. Reprod. Immunol. 29, 179. Yelavarthi, K. K., Schmidt, C. M., Ehlenfeldt, R. G., Orr, H. T., and Hunt, J. S. (1993). J. Immunol. 151, 3638. Yue, F. Y., Dummer, R., Geertsen, R., Hofbauer, G., Laine, E., Manolio, S., and Burg, G. (1997). Int. J. Cancer 71, 630.


The Zebrafish as a Model Organism to Study Development of the Immune System DAVID TRAVER,* PHILIPPE HERBOMEL,y E. ELIZABETH PATTON,* RYAN D. MURPHEY,* ´ CATIC,* CHRIS T. AMEMIYA,yy JEFFREY A. YODER,z, x,** GARY W. LITMAN,x,** ANDRE ,zz LEONARD I. ZON,* AND NIKOLAUS S. TREDE* ,1 * Division of Hematology, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115, USA; yUnite´ Macrophages et De´veloppement de l’Immunite´, Institut Pasteur, 25 rue du Dr Roux, 75724 Paris cedex 15, France; zDepartment of Biology, University of South Florida, 4202 East Fowler Avenue, SCA110, Tampa, FL 33620, USA; xImmunology Program, H. Lee Moffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Tampa, FL 33612, USA; ** Department of Pediatrics, Children’s Research Institute, University of South Florida, 140 Seventh Avenue South, St. Petersburg, FL 33701, USA; yyBenaroya Research Institute, Molecular Genetics Dept., 1201 Ninth Avenue, Seattle, WA 98101, USA; and zzHoward Hughes Medical Institute, Children’s Hospital, 320, Longwood Avenue, Boston, MA 02115, USA

Abbreviations WISH hpf dpf ENU NITR *

whole mount in situ hybridization hour post fertilization day post fertilization ethylnitrosourea Novel Immune-Type Receptor indicates movies, to be found at: http://www-alt.pasteur.fr/ herbomel/

I. Introduction

Over the past 25 years the field of immunology has greatly benefited from the advances in molecular biology. For example two seminal techniques, Southern blotting and DNA sequencing were employed shortly after their discovery by S. Tonegawa (Tonegawa, 1983) and others to unravel one of the great secrets of biology that had mystified legions of immunologists: the generation of somatic diversity in antigen receptors. Transgenic mice (Le Meur et al., 1985; Yamamura et al., 1985) and targeted gene inactivation by homologous recombination (Zijlstra et al., 1989) or conditional inactivation (Gu et al., 1994; reviewed in Rajewsky et al., 1996) were successfully employed for analysis of genes with pivotal roles in the mammalian immune system. Furthermore, the study of mice and patients with immune deficiencies has greatly enhanced our understanding of the molecular processes involved in developmental aspects of the immune system (reviewed 1

E-mail: [email protected] 253 Copyright ß 2003 by Elsevier (USA) All rights of reproduction in any form reserved. 0065-2776/03 $35.00

254

D. TRAVER ET AL.

in Fischer and Malissen, 1998; Fischer, 2001). However, mammalian immunology is handicapped by the fact that the earliest steps in the ontogeny of the immune system occur in-utero, and are therefore difficult to study in an in vivo system. Prominent among these early steps are thymic organogenesis and T-cell lymphopoiesis, crucial interdependent processes that establish a functional vertebrate immune system. Current understanding of vertebrate thymic development during embryogenesis remains incomplete and would benefit from novel approaches. Despite much innovative work, immunologists have been hesitant to use novel model systems to study these earliest steps in immune ontogeny. In this chapter, we propose to use the zebrafish (Danio rerio) as a model organism for immunology as an alternative to study humans or mice. Studies in several species of teleosts have demonstrated presence of all major blood lineages. As teleosts do not have bone marrow, blood cells are produced in the kidney (Willett et al., 1999). The immune system of teleosts has been previously explored in the channel catfish (Ictalurus punctatus), in particular by the groups of L. W. Clem and N. W. Miller (Miller et al., 1985, 1987; Ellsaesser et al., 1988; Shen et al., 2002) and in rainbow trout (Oncorhynchus mykiss) by J. D. Hansen (Hansen, 1997; Hansen and Strassburger, 2000; Hansen and La Patra, 2002) and others. Given their large size and abundance of specific cell types and tissues, catfish and trout have become extensively used for cell biologic studies involved in infection and immunity. These studies established that teleosts have T cells, B cells, antigen presenting cells, and natural killer cells and can mount antibody-mediated and cellular immune responses to infections. Recently, efforts were initiated to establish the rainbow trout as a genomic model ‘‘to address areas such as the evolution of the immune system and duplicated genes’’ (Thorgaard et al., 2002). Despite its small size and comparative paucity of molecular reagents for immunological research, the zebrafish has many advantages over the other vertebrates in studying the immune system. Several lines of evidence suggest that zebrafish have a complete set of genes required for the establishment of a fully functional adaptive immune system. We and others have cloned genes involved in early T-cell development, including ikaros, GATA-3, Rag-1, Rag-2, and lck (Willett et al., 1997, 2001; Trede et al., 2001). Recently, the genomic area encompassing the zebrafish TCR alpha locus has been fully sequenced (C.T.A. et al., unpublished). Sequence analysis revealed a similar organization of the variable (V) segments to those in humans and mice. Furthermore, the high degree of conservation of the heptamer and nonamer sequences in all species of jawed vertebrates as well as the length of the spacer in the recombination signal sequences strongly suggests conservation of the mechanism of the V(D)J recombination machinery. Zebrafish have a thymic organ, which remains bilateral in the third pharyngeal pouch (Willett et al.,

THE ZEBRAFISH AS A MODEL ORGANISM

255

1999; Lam et al., 2002). Expression of early T cell genes can be observed in the bilateral thymi starting on day 3 post fertilization (pf ) (Willett et al., 1997, 1999; Trede and Zon, 1998; Schorpp et al., 2000; Trede et al., 2001; Lam et al., 2002). Zebrafish also have B cells as evidenced by expression of VH genes in the kidney, the bone marrow equivalent of the zebrafish (Danilova and Steiner, 2002). Recently, zebrafish B cells have been reported to first arise in the pancreas (Danilova and Steiner, 2002). Genes encoding class I and II major histocompatibility (MHC) molecules have been isolated from the zebrafish genome (Ono et al., 1992; Sultmann et al., 1993; Takeuchi et al., 1995) and the presence of macrophages has been documented by whole mount in situ hybridization (WISH) and by differential interference contrast (DIC) video microscopy (Herbomel et al., 1999, 2001). Taken together, these data strongly suggest that zebrafish can mount innate and adaptive immune responses to pathogens much like higher vertebrates. The zebrafish provides a unique vertebrate model system for the analysis of developmental processes due to the transparency of the larvae, the short generation time and the ease with which mutations can be created and detected by various types of screens. Standard ethylnitrosourea (ENU)-based large-scale screens have identified a large number of mutants affecting early development and organogenesis (Haffter et al., 1996). Gynogenetic diploid embryos can be obtained thus permitting the expeditious detection of heterozygous individuals (Beattie et al., 1999). Zebrafish produce a large number of embryos and can be mated every week, facilitating the swift accumulation of meiotic recombinants for the purpose of positional cloning of mutant genes. Several tools are available for mapping and positional cloning in the zebrafish. These include a dense map of sequence-specific length polymorphisms (SSLP), an increasing number of expressed sequence tags (ESTs) mapped on genetic and radiation hybrid panels, and a variety of genomic libraries. Furthermore, the zebrafish genome has been sequenced in its entirety and a high degree of synteny to the human and pufferfish (Fugu rubripes) genomes has become apparent. These reagents make the zebrafish an extremely versatile genetic system for the generation of mutant phenotypes and cloning of the corresponding genes. Given these advantages the zebrafish is an excellent genetic system to study vertebrate immunology, and to define novel factors that participate in the development of the immune system. Zebrafish researchers receive frequent inquiries from biologists in various fields, including immunology, because during homology, searches of the gene of interest, zebrafish ESTs are found to be highly represented in the Genbank database. A number of recent publications have promoted the zebrafish as a powerful vertebrate model organism where, for example, transient gene

256

D. TRAVER ET AL.

inactivation using morpholinos can be done speedily (see special issue of Genesis, volume 30, July 2001). This raises the expectation of many researchers outside the zebrafish field that quick answers can be found for their biologic questions using this model organism. In this chapter we intend to introduce the immunologically inclined reader to the structural background and developmental aspects of the zebrafish immune system. We propose that the zebrafish system offers many opportunities to explore vertebrate immunology. For example, forward genetic and expression screens can uncover novel genes required for immune cell development and function. We also point to the limitations of the zebrafish system, where relatively few traditional reagents (such as specific antibodies directed against surface molecules) are available, and where functional properties of the various immune cells are at the early stages of investigation. Furthermore, if one considers that the entire kidney (the bone marrow equivalent of the zebrafish) contains only about 106 cells, 40% of which are mature red cells, the number of cells of a particular hematopoietic lineage that can be obtained per fish is rather limited. However, the ready availability of a large number of fish can circumvent this problem. This chapter provides an overview of the current knowledge accumulated— much of it still unpublished—in the following aspects of zebrafish immunology. Section II introduces fish innate immunity with an emphasis on the early origin of effector cells. Here we point to the power of DIC video microscopy in determining cell trafficking and behaviors to gain insight into potential functions of early macrophages. A separate subsection (Section II.C) is devoted to a novel family of immune type receptors (NITRs), which have been characterized extensively in zebrafish. This exciting novel finding has multiple implications, which are discussed in detail. The origins of adaptive immunity with an emphasis on primary and secondary lymphoid organs in fishes, as well as their antigen receptors are outlined in Section III. This section, which also treats the specific zebrafish immune system, points to several immunologic peculiarities, such as absence of lymph nodes and conventional class switching encountered in fishes. There is a paucity of reagents available for the study of zebrafish immune cells. For example, antibodies directed to immune surface receptors, which have been generated against phylogenetically related catfish molecules, do not cross-react in zebrafish and multiple attempts at raising antibodies against zebrafish blood cells have failed in our laboratories. Section IV discusses these shortcomings and offers possible alternative imaging and cell isolation techniques. A detailed account of the various types of mutagenesis screens that can be accomplished in zebrafish to obtain mutant phenotypes is the subject of Section V, followed by a description of lymphoid screens carried out to date (Section VI). The emerging field of reverse genetic


257

approaches in zebrafish, which allow for transient gene knockdowns as well as permanent gene inactivation, is exposed in Section VII. An alternative screening method combining the power of genomics and early development of the zebrafish to identify genes involved in a particular biologic process is detailed in Section VIII. Here, cDNA libraries from various stages of development, or from specific blood cell lineages, can be used to identify genes expressed in the tissue of interest throughout early development. Genomic analysis exploring syntenic relationships and functional assays involving gene knockdown experiments can make this approach a powerful tool for detecting immune genes and probing their function. Finally we provide a draft of a hands-on manual of how to proceed as a newcomer in the field of zebrafish genomics with a focused interest in mind (Section IX). As zebrafish immunology is still in its infancy and published literature on this subject to date remains rather scant, we include in this chapter a number of unpublished results to demonstrate the power and potential of the zebrafish system. We hope that this contribution will entice stimulating discussions and will be instrumental in forging fruitful collaborations between established immunologists and scientists in the emerging field of zebrafish immunology. II. Innate Immunity of Teleosts

A. THE COMPONENTS

OF INNATE IMMUNITY OF

BONY FISH

1. Leukocytes Teleosts have monocytes, granulocytes and tissue macrophages. Monocytes and macrophages appear similar to those of mammals in terms of ultrastructure and cytochemistry, and phagocytic and secretory abilities (Secombes and Fletcher, 1992; Beattie et al., 1999; Neumann et al., 2001). In addition, fish have pigmented macrophages, called melanomacrophages, that are found as aggregates in lymphoid tissues (Herraez and Zapata, 1991). On the other hand, there has been much debate about granulocyte typology. The classes of granulocytes that could be discerned seem to vary among different fish species, and establishing their correspondence to mammalian types is not straightforward, partly because of insufficient functional studies (Rowley et al., 1988; Press et al., 1998). Among cyprinids, the taxon to which zebrafish belong, two species have been mainly studied, carp and goldfish (Press et al., 1998; Barreda et al., 2000). In both, and in zebrafish (Jagadeeswaran et al., 1999; Bennett et al., 2001; Lieschke et al., 2001), the head kidney, blood and spleen were found to contain neutrophilic/heterophilic and eosinophilic granulocytes. Neutrophils (>90% of blood granulocytes) have an eccentric,

258

D. TRAVER ET AL.

pluri-lobed nucleus, and two kinds of granules: many cigar-shaped granules, ca. 0.4 m in length, the mature form of which have a crystalline, striated axial core structure, and less numerous, round and homogeneous granules. Eosinophils have an eccentric nucleus attached to the plasma membrane, and are packed with larger, roughly spherical granules (0.7 m mean diameter). Basophils were not detected in zebrafish, but were found in carp, while in goldfish their presence has been disputed (Rowley et al., 1988). Tissue mast cells have been detected in various fish species including cyprinids (Reite, 1998), but not yet in zebrafish (Lieschke et al., 2001). In the striped bass (perciforms) they were shown to produce potent broad-spectrum peptide antibiotics called piscidins (Silphaduang and Noga, 2001). 2. Complement Fish have a highly developed complement system, with the three known— lectin-dependent, alternative, and classical (antibody-dependent)—pathways of complement C3 activation, followed by the common lytic pathway leading to self-assembly of the membrane attack complex. Most of the 25 proteins of the system found in mammals arose by successive duplications from four primordial genes, and this led to the emergence of distinct activation pathways (Zarkadis et al., 2001). Some of these duplications happened after the fish/ tetrapod divergence. Thus, fish have no factor B/C2 dichotomy, but a factor B that functions in both the alternative and the classical pathways. On the other hand, fish have developed multiple isoforms of C3, the central factor of the system, and of factor B. These isoforms are the products of different genes, and differ in structure and functional characteristics. Furthermore the genes are quite polymorphic, and can differ vastly among fish species. Therefore, several authors have suggested that the low-affinity antibody response in fish (see Section III.D) may have been compensated by the diversification of innate immune mechanisms such as the complement system. In support of this notion, whereas classical pathway titers in fish serum are similar to those found in mammals, alternative pathway titers are 5–10 times higher in fish (Zarkadis et al., 2001). The number of complement genes differs among the various fish species. Carp have five C3 and three factor B genes, the gilthead sea bream, a diploid fish, has five C3, while zebrafish have three C3 and two factor B genes (Gongora et al., 1998; Nakao et al., 2000; Zarkadis et al., 2001). The C3 isoforms display often widely different affinities for various microbial cell walls. Strikingly, in the carp, a histidine residue at the catalytic site of C3, so far thought to be required for the covalent binding of C3 to biological surfaces (the key-point of the whole system), is substituted in three out of the five C3 isoforms by a serine or a glutamine. The serine-substituted isoform actually


259

showed a three-fold increased hemolytic activity (Nakao et al., 2000). Complement C3-dependent phagocytosis has been described in several fish species (Secombes and Fletcher, 1992), but a phagocyte CR3 receptor has not yet been characterized at the molecular level. 3. Inflammatory Processes As in mammals, upon wounding or local infection, the first leukocytes that operate in fish are tissue macrophages, and possibly mast cells. Then if signals are released to launch a systemic inflammatory process, the first inflammatory cells to arrive at the site of injury are neutrophils, detected from 6 h to 4 days, followed about a day later by monocytes that then differentiate into macrophages (Ellis, 2001; Neumann et al., 2001). The major proteins and corresponding genes involved in inflammation and its resolution in mammals have recently been found in fish: TNF-, NF-B, COX2, IL-1, IL-8 and other C-C and C-X-C chemokines (Engelsma et al., 2001; Secombes et al., 2001). Several signaling pathways appear conserved. For instance, bacterial LPS induce IL-1, TNF-, and iNOS transcription in macrophages by an NF-B dependent mechanism, and sequences that would be typical NF-B response elements in mammals are found in the carp iNOS gene promoter. Besides peptide chemokines, eicosanoids, particularly lipoxins, produced by fish macrophages, appear to be potent chemoattractants for fish leukocytes (Sharp et al., 1992). Functional studies strongly suggest that both monocytes/ macrophages and neutrophils possess receptors for lectins, complement and immunoglobulins. Recently an Fc receptor gene was cloned in the carp (Fujiki et al., 2000). In terms of microbicidal weapons, the two main phagocytes, macrophages and neutrophilic granulocytes, produce reactive oxygen intermediates and H2O2 through a well-documented inducible respiratory burst, that seems to involve a membrane-bound phagocyte (NADPH) oxidase similar to that of mammals (Secombes and Fletcher, 1992; Neumann et al., 2001). Upon infection, fish macrophages, as their mammalian homologs, express inducible nitric oxide synthase (iNOS, see above), leading to increased production of NO and reactive nitrogen intermediates. One originality found in goldfish and carp, relative to mammals, is that contact of the macrophage with the microbes appears sufficient to trigger induction of iNOS, with no need for exogenous cytokine signals such as IFN- . Peptides derived from the proteolysis of transferrin appear to be involved in this induction of iNOS (Neumann et al., 2001). In mammalian monocytes and neutrophils, myeloperoxidase uses the H2O2 produced by the respiratory burst to catalyze the formation in the phagosome of hypochlorous acid. In fish, monocytes, neutrophils, eosinophils, and macrophages in some species (e.g., goldfish) were found to be peroxidase-positive

260

D. TRAVER ET AL.

(Rowley et al., 1988; Belcourt et al., 1995). The first fish myeloid peroxidase gene cloned, in zebrafish, suggests that the diversification of myeloid peroxidases found in mammals (three members: myelo-, lacto- and eosinophil peroxidase) occurred after the fish/tetrapod divergence (Bennett et al., 2001; Lieschke et al., 2001). Comparison of the zebrafish and fugu genome sequences should tell us whether teleosts have independently diversified their myeloid peroxidases. 4. Connection of the Innate and Adaptive Immune Systems Several studies have shown that fish macrophages can activate T and B lymphocytes. Consistently, teleost fish have polymorphic class I and class II MHC genes, and overall, the latter were found to be expressed primarily in lymphoid tissues (Sultmann et al., 1993; Rodrigues et al., 1995; Press et al., 1998). However, in mammals, it has recently become clear that a distinct set of myeloid cells, the dendritic leukocytes, and not macrophages, are able to prime naive T lymphocytes. In fish, the potential distinction between macrophages and dendritic leukocytes has not yet been approached. It will certainly be a major question in fish immunology for the near future. 5. Non-specific, non-phagocytic cytotoxic cells (NCC) Teleosts have non-specific or natural-cytotoxic cells (NCC) that appear to function as mammalian natural killer (NK) cells (Jaso-Friedmann et al., 2001). They can kill allogeneic and xenogeneic tumor cells, virus-infected cells, and protozoan parasites. Unlike mammalian NK cells, they are small, agranular lymphocytes, which display constitutively a spectrum of cytotoxicities that lies between that of naive and cytokine-activated NK cells. In addition, leukocytes other than NCC appear to have cytotoxic capacity. In the channel catfish, normal peripheral blood leukocytes negative for a well-defined NCC cell surface marker showed potent cytotoxic activity against allogeneic and virusinfected autologous cells (Yoshida et al., 1995; Hogan et al., 1996). No marker is presently available to characterize these non-NCC cytotoxic cells. Section IV revisits zebrafish NCC, and discusses novel surface receptors instrumental in their characterization. B. THE ZEBRAFISH INNATE IMMUNE SYSTEM One of the prime strengths of the zebrafish is its optical clarity, which allows direct inspection during early development. For this reason the zebrafish has been a particularly useful model for the developmental biologist. The main contribution of zebrafish to the understanding of vertebrate innate immunity has therefore come from the opportunity to study early developmental stages of the cells that compose the innate immune system, which were hardly accessible in more classical vertebrate model species. These studies revealed


261

that the embryo already had a fully operational, macrophage-based immune system. By the time this book is published, the availability of the zebrafish genome sequence will have provided a much more comprehensive account of which of the numerous molecular components relevant to the innate immune system known in mammals are also found in fish. Thus, the first contribution of zebrafish to innate immunity was to reveal the origin and functional characteristics of a lineage of early macrophages that arises well before any other leukocytes, and even precedes the emergence of dedicated hematopoietic organs. 1. The Early Macrophage Lineage Macrophages are an ancient cell type, older than blood itself, as initially identified by Metchnikoff over a century ago. Metchnikoff, then a comparative embryologist, was wondering about the function of mesoderm, and found that the phylogenetic origin of mesoderm coincides with the origin of phagocytes, which he named macrophages (Metchnikoff, 1892). From this, Metchnikoff then developed his ‘‘phagocytic theory of inflammation’’ that founded immunology. It is common knowledge that in mammalian or avian ontogeny the first leukocytes to arise are macrophages. Their appearance on the yolk sac coincides with development of the first erythroid blood islands. 10–15 years ago, a few pioneering teams went further, providing evidence that these early macrophages differentiated along a pathway that bypassed monocytic intermediates and granulocyte precursors (Takahashi et al., 1989, 1996; Sorokin et al., 1992a,b; Cuadros et al., 1992, 1993). The work of these groups collectively showed that from the yolk sac, these primitive macrophages quickly invaded the embryo’s mesenchyme. Invasion was especially prominent in the head, and from there they entered the still unvascularized, developing organs, starting with the brain (see also Kurz and Christ, 1998; Knabe and Kuhn, 1999; Lichanska et al., 1999). Mitotic figures were seen in these early tissue macrophages, again at odds with ‘‘classical,’’ monocytederived macrophages, which are thought to be mostly post-mitotic cells. These findings contradicted the dogma that in ontogeny, the first tissue macrophages were derived from circulating monocytes, extravasating from blood vessels into tissues to become macrophages, as occurs during inflammation in adults (reviewed in Sorokin et al., 1992a and Takahashi et al., 1996). In the zebrafish embryo, analogous early macrophages were detected, and could be studied in detail by combining fluorescent cell tracing and videoenhanced DIC microscopy in live embryos, with WISH in fixed embryos. They were found to differentiate in the yolk sac before the onset of blood circulation, and from there to invade cephalic mesenchyme, followed by brain,

262

D. TRAVER ET AL.

FIG. 1. Hematopoietic regions in the zebrafish embryo. Two hematopoietic regions can be distinguished in the lateral plate mesoderm of the zebrafish embryo at the 22-somite stage: the rostral region gives rise to the ‘‘early macrophages’’; the caudal hematopoietic region gives rise to embryonic erythroblasts and probably also to the progenitors of all definitive blood cell lineages, which will colonize the definitive hemopoietic organs, thymus and head kidney, 2–4 days later.

retina, and epidermis, in a pattern remarkably similar to that described in mammals and birds (Herbomel et al., 1999, 2001). The yolk sac, where the early macrophages differentiate, has a very simple histological structure. It is essentially one giant yolk cell (actually a thin ‘‘yolk syncytial layer’’ enclosing a large yolk mass), covered by a thin ‘‘skin,’’ made of two closely juxtaposed thin cell monolayers, the epidermis and a protective ‘‘periderm.’’ Early macrophages are initially found in the free space between the yolk cell and skin, most often spread out on or tethered to the basal side of the overlying epidermis (see Fig. 1). Thus, using video-enhanced DIC microscopy, these cells can be imaged and followed in vivo at high magnification (see movie 4* ). Furthermore, the zebrafish displays a highly convenient peculiarity among teleosts. At the start of blood circulation, the venous blood, as it arrives on the yolk sac, is no longer contained in a blood vessel, but flows freely over the surface of the yolk cell, to be collected in the single atrium of the zebrafish heart. Thus many of the early macrophages present on the yolk sac are actually standing in the blood flow, and their interaction with other blood cells or with intravenously injected microbes can be imaged conveniently. In mammals and birds, the origin of early macrophage precursors before they differentiate in the yolk sac is not known. In the zebrafish, their embryologic origin could be traced back to gastrulation. The surprising discovery was that unlike all other hemopoietic lineages in amniotes—and probably also in zebrafish—their precursors did not originate from the caudal lateral plate mesoderm, but from the rostral-most lateral mesoderm, just anterior to the cardiac field in the developing head region (Herbomel et al., 1999). This rostral site of origin of early macrophages is not specific to the zebrafish. In Xenopus, early macrophages were recently found


263

to arise from the same region (Smith et al., 2002). It is worth noting here that in Drosophila, embryonic macrophages also arise from the cephalic mesoderm (Tepass et al., 1994). Hence the production of early macrophages directly from cephalic lateral mesoderm in vertebrate ontogeny might be considered as a case of ontogeny recapitulating phylogeny—since the developmentally later, elaborate system of pluripotent stem cells giving rise to all leukocytes within a dedicated hematopoietic stroma (often centered on macrophages) was probably discovered in vertebrates. It is tempting to suggest that a similar site of origin will also be found for mammalian and avian early macrophages. In zebrafish, this small region of rostral-most lateral mesoderm (about 50 cells) from which the early macrophages originate also gives rise to endothelial cells, that will notably make up the carotids (Herbomel et al., 1999). It thus conforms to the rule derived from mammalian and avian studies that hematopoietic lateral plate mesoderm is always also vasculogenic (Pardanaud et al., 1996). In the rostral territory, vascular and macrophage precursors segregate from each other during the final convergence of this rostral-most lateral mesoderm to the midline, beneath the paraxial mesoderm and neural tube at approximately 14–16 hpf. Several genes coding for transcription factors (TFs) were found to be turned on specifically in mesodermal macrophage precursors just before or during this short period. Two of these code for the zebrafish homologs of PU.1 and C/EBP- , both key transcription factors for the commitment of hematopoietic cells to a myeloid fate in mammals (Lieschke et al., 2002; P.H., B. Thisse and C. Thisse, unpublished results). DIC videomicroscopy reveals that macrophage precursors, still with mesenchymal morphology, subsequently almost reverse the direction of migration and emigrate, now as single cells, to the neighboring, anterior aspect of the yolk sac (Herbomel et al., 1999). During this emigration they express other TFs of the C/EBP family (Lyons et al., 2001a,b), and in addition now start transcribing genes coding for markers of leukocyte or macrophage differentiation. Among these are L-plastin, a leukocyte-specific actin-bundling protein likely involved in the amoeboid motility that characterizes leukocytes, lysozyme C, and the M-CSF (CSF-1) receptor—an important marker of the macrophage lineage in mammals (Herbomel et al., 1999, 2001; Liu and Wen, 2002). In Xenopus embryo, two other genes were recently found to also mark the early macrophage population: POX2, a new member of the myeloid peroxidase gene family, and LURP-1, a novel and likely secreted member of the Ly-6 family (Smith et al., 2002). Once on the anterior aspect of the yolk sac, the macrophage precursors evolve into ‘‘pre-macrophages’’ that are characterized as round cells of homogeneous size (12 m) with little cytoplasm, and a consistent mini-blebbing

264

D. TRAVER ET AL.

behavior observed by DIC video-microscopy (Herbomel et al., 1999; Movie 6* ). These pre-macrophages differentiate into ‘‘wandering’’ immature macrophages, which will become competent for phagocytosis in the next few hours (see below). These ‘‘pre-’’ and ‘‘young’’ early macrophages are likely homologous to the ‘‘primitive’’ and ‘‘fetal’’ macrophages described by Takahashi et al. (Takahashi et al., 1989) in the mouse yolk sac. Thus, early macrophages arise rapidly from the rostral-most lateral mesoderm, after only one morphologically distinguishable intermediate, the ‘‘pre-macrophage.’’ Such a direct process makes physiologic sense, for a population of wandering macrophages is thus produced in only a few hours and is now ready to remodel tissues and to fight pathogens, prior to the onset of adaptive immunity (see Section III.B). The finding that cells of a special early macrophage lineage are the first resident macrophages in the tissues, and that they constitute autonomous immune sentries before any other leukocytes are produced (see below), raises a number of questions. How similar are they from definitive monocyte-derived macrophages? Are they involved in developmental processes? Do they persist in the adult, as suggested by data obtained in rodents (reviewed in Takahashi et al., 1996; see also Ginsel, 1993; Leenen and Campbell, 1993), to constitute a significant, self-renewing fraction of tissue macrophages and/or dendritic leukocytes, possibly endowed with a distinct reactivity and functional connection to the other cells of the adult immune system? A recent study of goldfish macrophages (Barreda et al., 2000) suggests that a rapid macrophage differentiation pathway bypassing monocytes may also be operative in adults, arising from blast-like progenitors in the head kidney, and coexisting with the better known pathway involving monocytes. a. Functional Characteristics of Early Macrophages. In the yolk sac, early macrophages become phagocytic, ingesting apoptotic erythroblasts as well as injected microbes (see below), at 23–26 hpf, coinciding with the onset of blood circulation (P.H., unpublished data). Subsequently, their deployment in embryonic tissues is accompanied by tissue- and stage-specific variations in gene expression patterns and levels of endocytic activity (Herbomel et al., 2001; B. Thisse and C. Thisse, unpublished data), thus starting to delineate local differentiation patterns, a typical feature of tissue macrophages in mammals (Leenen and Campbell, 1993; Gordon, 1995). Some are transient, and are associated with a defined tissue-remodeling task, such as the elimination of the remnants of the hatching gland after hatching has occurred, others are permanent. The most striking example for the latter is a sudden phenotypic transition that exclusively affects all brain and retinal macrophages


265

at 55 hpf, regardless of how much time they have already spent in this tissue. The resulting cells were termed ‘‘early microglial cells’’ (Herbomel et al., 2001). They are highly endocytic, and notably express high levels of apolipoprotein E (apoE). While apoE is a typical secretion product of many tissue macrophages in adult mammals, in zebrafish embryos, the early microglial cells are the only macrophages that express transcripts of the gene by WISH. DIC video-microscopy revealed that once in the epithelial tissues (brain, epidermis), most early macrophages wander restlessly among epithelial cells, showing a remarkable ability to wind through tissues as dense as the neuroepithelium of the developing brain (Herbomel et al., 2001; Movies 1 and 2* ). So far, in mammals, the movement of macrophages had largely been considered in the context of inflammation, and towards inflammatory foci. The live observations in the zebrafish embryo suggest that resident macrophages may well be constantly ‘‘patrolling’’ in vertebrate tissues at steady state— conceivably for immune surveillance, and more broadly as guardians of homeostasis (Gordon, 1995). b. Functional Relevance of the M-CSF Receptor. As in mammals, the zebrafish has an M-CSF receptor gene (c-fms) that is expressed in all early macrophages (Herbomel et al., 2001). In the panther mutant, a loss-offunction defect in fms (Parichy et al., 2000) early macrophages differentiate and behave normally in the yolk sac. This suggests that the fms gene is dispensable in terms of proliferation, survival, motility, and endocytic or phagocytic abilities in early macrophages. However, they subsequently fail to invade embryonic tissues, and remain in the yolk sac and in the blood (Herbomel et al., 2001). This suggests that the tissues that normally become colonized attract macrophages actively, by secreting M-CSF as a chemoattractant. A similar finding was recently reported in Drosophila embryos, where VEGF-R deficient macrophages are specified but fail to migrate properly (Cho et al., 2002). Another great asset of the zebrafish is the wide temperature range under which zebrafish embryos and adults can live (20–33 C), making it possible to screen for temperature-sensitive mutants. Thus a thermosensitive M-CSF receptor (fms) mutant was recently obtained (Parichy and Turner, 2003). It will likely be useful for studying monocyte/macrophage differentiation, physiology and contribution to immunity in embryos, fry and adults. c. Autonomous Immune Function of Early Macrophages. When massive amounts of live gram þ (B. subtilis) or gram (E. coli) bacteria are injected in the bloodstream of zebrafish embryos, they are quickly phagocytosed

266

D. TRAVER ET AL.

and killed by the early macrophages. It was shown by DIC video-microscopy that three hours after injection, there were no more free bacteria, and still a few hours later, no more traces of bacteria were seen in the macrophages. In addition, the macrophages now show increased ruffling and also start phagocytosing live, apparently healthy erythroblasts—often catching them in the middle of their mitosis (Herbomel et al., 1999; Movie 8* ). Thus the infection triggered a type of macrophage activation that is not rapidly downregulated and remains to be defined at the molecular level. All microorganisms tested so far were readily phagocytosed: live gram and gram þ bacteria, and yeast (Saccharomyces cerevisiae) (P.H., unpublished data). This suggests that early macrophages have pattern-recognition receptors for antibody-independent phagocytosis of most microbes. Their capacity to phagocytose live yeast suggests they possess at least a mannose/fucose receptor akin to that of mammals. Serum proteins, such as complement may well play an adjunct role for killing or opsonizing the microbes. However, in adults these proteins are produced by the liver, and the macrophage-based immune system of the embryo is operational before the development of a functional liver. One putative source of complement in the embryo would be the yolk syncytial layer, which assumes many aspects of the role of an embryonic liver, releasing nutrients into the blood, and thus quite possibly also defense-oriented proteins. Another potential source of opsonizing complement factors could be the macrophages themselves (Ezekowitz et al., 1985). To further study the functionality of early macrophages, bacteria were injected into the hindbrain ventricle, a closed cavity remote from the blood. This experiment was carried out at 21 hpf, a developmental stage that precedes vascularization and seeding of the head with early macrophages. Five hours after the injection, many macrophages had entered the hindbrain ventricle and cleared the bacteria (Herbomel et al., 1999). Because of the initial lack of resident leukocytes of any kind at the infection site, the 24 hpf zebrafish embryo offers the unique opportunity to directly assess the chemoattractive nature of microbes in an entire live organism—a question that is not easily accessible in mammalian systems. So what attracts the macrophages to the infected brain ventricle? One possibility is that the bacteria in the ventricle release some molecules small enough to permeate the ventricle walls, reach the macrophages 100 microns away and chemoattract them (N-formylated peptides would be possible candidates). Alternatively, microbial molecules are first detected by the neuroepithelial and/or meningeal walls of the ventricle, which then secrete chemokines that are sensed by the distant macrophages. The latter would imply that vertebrate epithelial tissues are equipped with microbe sensing, possibly Toll-like receptors and chemokine signaling abilities early in their


267

ontogeny, even before they are fully differentiated. Interestingly, all macrophages in the embryo acquire an activated morphology (abundant ruffling) and behavior upon infection of the hindbrain ventricle. This is true not only for macrophages that home to the infected brain ventricle but also for those in the blood flow that now start phagocytosing apparently healthy erythroblasts (Herbomel et al., 1999). Such concerted behavioral change in activation suggests cytokine-mediated communication within the macrophage population throughout the organism. d. Attraction of Early Macrophages to Wounds. Infliction of an aseptic wound anywhere in the embryo as early as 40 hpf (and possibly earlier) attracts fms þ early macrophages to the wound site within the next four hours. Macrophages are still found gathered at the wound site 35 h later (P.H., B. Thisse and C. Thisse, unpublished data). The initial absence of leukocytes at several of the tested wound sites introduces a novel situation compared to mammalian systems. In this setting the macrophages must have been attracted to the wound by signals emanating from the dead or dying cells themselves, possibly relayed or amplified by healthy neighboring cells in the epidermis. Despite the intense scientific interest in apoptosis, it is still not known if dying cells can directly chemoattract macrophages in vivo. The zebrafish embryo could be instrumental in elucidating this point. 2. Granulocytes The next type of immune cells that arise during zebrafish development are neutrophilic granulocytes. Immature neutrophils are first reliably detected by E.M. at 48 hpf, 24 h after the onset of blood circulation and early macrophage phagocytic functions. The initial site of appearance is the caudal/axial vein and the mesenchyme surrounding it—in the tail proper and more rostral along the ventral trunk (see Fig. 1) (Willett et al., 1999; Lieschke et al., 2001). Their detection in the vein suggests that some neutrophils must be circulating at that time in the blood, though they likely represent a small minority relative to the early macrophages, and even to the thrombocytes, that appear around 36 hpf (Gregory and Jagadeeswaran, 2002). Since lymphocytes will appear even later (see Section III.B.2), the four cell types appear developmentally in the same order as in mammalian embryos. The initial location of the neutrophilic granulocytes in the mesenchyme around the caudal vein suggests that they, as well as the thrombocytes, arise from the hematopoietic stroma that is steadily expanding at this stage in between the caudal artery and caudal vein, and has been presumed to be equivalent to the mammalian AGM region (Thompson et al., 1998). When a wound is introduced in the caudal area of the embryo/ larva, i.e., close to the site where neutrophils arise, these neutrophilic

268

D. TRAVER ET AL.

granulocytes home to the wound within minutes (M. Redd, personal communication) to hours (Lieschke et al., 2001). Whether they can respond to microbial infections and are phagocytic remains to be tested. The second type of granulocytes found in adult zebrafish, eosinophils, were not detected up to five dpf (Lieschke et al., 2001). It is therefore quite possible that their generation is delayed and initiates in the definitive hematopoietic organ, the head kidney. The head kidney is also the site where neutrophilic granulocytes will become the most numerous myeloid leukocytes, and where monocytes likely arise for the first time. 3. Impact of Zebrafish on Understanding Innate Immunity Notwithstanding much accumulated knowledge about the in-and-out of our immune system, relatively, little is known about the repertoire of behaviors by which leukocytes residing in the tissues at steady state control microorganisms on a daily basis. In particular it is unclear how resident macrophages contain infections, most of the time successfully and silently, i.e., without launching a systemic immune response. The constant activity of these resident leukocytes scattered in the tissues, and the basis of their variable efficiency at controlling a given microorganism—among different individuals, or species—are difficult to approach in mammals. To explore further the potential of the zebrafish model along these lines, the establishment of transgenic zebrafish expressing different fluorochromes under the control of promoters specific for the various types of leukocytes will be instrumental. By working at progressively later stages of development, the various cellular components that will make up the mature immune system can be implemented one by one, a prospect also inaccessible in mammals. To take full advantage of the system, zebrafish lines will have to be generated that remain transparent into adulthood—as was successfully done in the medaka fish (Wakamatsu et al., 2001). C. NATURAL KILLER CELLS, NATURAL CYTOTOXICITY AND ROLE OF NOVEL IMMUNE-TYPE RECEPTOR GENES

THE

POSSIBLE

Whereas zebrafish Ig and TCR genes are easily identified as orthologs of their mammalian counterparts (see Sections III.B and III.C) and can be used to identify B and T cells respectively, the identity of NK-like receptors in zebrafish remains in question. This observation, compounded with the apparent absence of a morphological equivalent to NK cells in bony fish makes the study of natural killing in zebrafish challenging, but highly intriguing. Nevertheless, several studies have demonstrated that fish possess at least three types of cytotoxic cells. Non-specific cytotoxic cells (NCC) which primarily populate the pronephros (the hematopoietic head


269

kidney), have been described in catfish, trout (Evans et al., 1984; Greenlee et al., 1991) and other fish species (Shen et al., 2002) and have been referred to as ‘‘NK-like’’ although their morphology contrasts with that of mammalian NK cells (NCC are small and agranular). NCC have been reported to recognize and kill a wide variety of target cells including allogeneic and xenogeneic tumor cells (Jaso-Friedmann et al., 1997). An undefined population of cells within the peripheral blood leukocytes (PBL) of catfish have been shown to kill allogeneic target cells, but fail to recognize xenogeneic targets (Yoshida et al., 1995). In addition, a population of cells isolated from the pronephros in carp (C. carpio) kill a profile of xenogeneic target cells that differs from the profile recognized by NCCs (Kurata et al., 1995); these cells have been morphologically characterized as neutrophilic granulocytes, and thus are distinct from NCC. Zebrafish spleens and kidneys possess similar populations of cytotoxic cells that also can recognize and lyse specific mammalian tumor target cells (J.A.Y., S. Wei and J. Djeu, unpublished observations). NK activity in mammals also is associated with a number of different cell lineages (McQueen and Parham, 2002) and recent advances in understanding the molecular biology of NK cell function in man and mouse have identified diverse receptor gene families that are associated with the recognition of different ligands. Mammalian NK receptors can be classified into two broad categories based on the structure of their extracellular domains: (1) immunoglobulin (Ig) domain containing receptors [primarily encoded at the leukocyte receptor complex (LRC)] and (2) C-type lectin (CTL) receptors [primarily encoded at the natural killer complex (NKC)] (Barten et al., 2001), both of which include activating and inhibitory forms. Despite overall differences in extracellular structure, mammalian NK receptors share common modes of intracellular signaling; activating signaling leads to target cell lysis and inhibitory signaling represents a means to control misdirected lysis. Activating signaling is based on the association of NK receptors with adaptor proteins (e.g., DAP12, CD3), that typically possess cytoplasmic immunoreceptor tyrosine-based activation motifs (ITAMs), although other adaptor proteins function through alternative pathways (e.g., DAP10). This protein–protein interaction involves the association of a positively charged residue in the transmembrane region of an NK receptor with a negatively charged residue in the adaptor molecule. Inhibitory signaling is based on the presence of cytoplasmic immunoreceptor tyrosine-based inhibition motifs (ITIMs) in certain (non-activating) NK receptor molecules. It is becoming clear that NK cells utilize both activating and inhibitory receptors to maintain a signaling balance; a shift towards the activating pathway can override the inhibitory pathway and lead to target destruction (Ravetch and Lanier, 2000; Cerwenka and Lanier, 2001).

270

D. TRAVER ET AL.

Although gene families resembling either the LRC or the NKC encoded receptors have not been reported in species at or below the phylogenetic level of bony fish, Novel Immune-Type Receptors (NITRs), which have been characterized in pufferfish, zebrafish, catfish and trout (Strong et al., 1999; Yoder et al., 2001, 2002a; Hawke et al., 2001), encode receptors that share overall structural characteristics with certain innate Ig-type NK receptors (KIR-type) but possess diversified variable (V) domains reminiscent of the adaptive immune genes. NITRs may be the bony fish equivalent to NK receptors (Litman et al., 2001). Interestingly, a single copy CTL-type receptor maps next to an NITR locus in trout, which may provide a link between the Ig-type and CTL-type of NK receptors (Yoder et al., 2002a). Most NITRs identified to date, including those identified in zebrafish, possess two extracellular Ig domains, a transmembrane region and a cytoplasmic tail containing ITIMs (Strong et al., 1999; Litman et al., 2001; Yoder et al., 2001, 2002b). The N terminal Ig ectodomain of the prototypic NITR is of the V-type and contains a contiguous Ig/TCR-like joining (J) region motif; the C terminal ectodomain shares characteristics of both V- and C2-type domains. The presence of authentic V domains with contiguous J-like sequences is a hallmark of NITR genes in multiple fish species (Strong et al., 1999; Yoder et al., 2001, 2002a; Hawke et al., 2001; Litman et al., 2001) and distinguishes them from other Ig-like innate receptors (e.g., KIRs which possess C2 domains). As discussed elsewhere in this review, highly diversified T cell receptor and immunoglobulin gene families have been identified in zebrafish (see Sections III.B and III.C), which strongly argues that the NITRs do not supplant the function of genes of the adaptive immune system. Furthermore, radiation hybrid mapping data indicate that the identified zebrafish TCR, MHC and Ig loci are on different chromosomes from the NITR gene cluster (Yoder et al., 2001; reviewed in Yoder et al., 2002b). Basic questions exist about the complexity, extent of diversity and transcriptional regulation of NITR genes in zebrafish. The Zebrafish Genome Project (ZGP; being conducted at the Wellcome Trust Sanger Institute) has prioritized the characterization of the NITR gene cluster and, while not complete, already has resulted in the identification of new families of NITR genes. Taken together with earlier studies, several conclusions can be drawn from the emerging genomic DNA dataset as regards the structure of NITRs and the organization of the NITR gene cluster in zebrafish: (1) the NITRs are highly diverse; seven families of V regions have been identified (Yoder et al., 2001; J. Y., unpublished observations); (2) sequence data acquired to date are consistent with the possibility of a single gene cluster as at least four NITR gene families map to a single genomic region based on STS mapping using


271

radiation hybrid panels (Yoder et al., 2001); (3) >100 NITR genes, including polymorphic variants, currently have been identified; (4) the vast majority of these possess ITIM-containing cytoplasmic tails; however, the type 5 (nitr5) genes possess an ITAM in the cytoplasmic tail (see below); (5) efforts to determine if the map position of zebrafish NITRs (LG7) share conserved synteny with the human 19q13 LRC are inconclusive (owing to disrupted synteny) (Yoder and Litman, 2000; Yoder et al., 2001); (6) emerging data from the ZGP, which is based on PAC sequences of pooled ‘‘inbred’’ AB line embryos, have defined extensive allelic polymorphism in the polygenic NITRs, resembling that seen with the KIRs (Uhrberg et al., 1997; Vilches and Parham, 2002), and are consistent with RFLP analyses of zebrafish genomic DNA employing NITR family-specific probes (J.A.Y. and G.W.L., unpublished observations); and (7) NITRs encoding only one (vs two) V-type ectodomains have been identified (Hawke et al., 2001; Yoder et al., 2001). Based on earlier work with pufferfish and catfish, the most unexpected findings relate to the receptors encoded by the nitr5 genes, which may possess a novel activating mechanism in an IgSF-related putative NK-type gene product. The nitr5 genes encode two extracellular Ig domains (the N-terminal ectodomain is of the V-type and the C-terminal ectodomain is of the V/C2 or I (intermediate) type), a transmembrane domain that has a neutral charge and a cytoplasmic tail that contains an ITAM. A current analysis of the Zebrafish Genome Project database has identified six different nitr5 genes (‘‘alleles’’), and the RNA splice junctions have been confirmed through analysis of corresponding cDNAs (J. Y. and G. W. L., unpublished observations). It has not escaped our notice that the presence of an ITAM within the cytoplasmic region of an NITR may allow for direct activating function, independent of an adaptor protein. Such a mechanism is in contrast with the two-protein complexes utilized in the mammalian NK system. In this regard Nitr5 receptors are distinguished from any known activating component of the NK system in mammals. Nitr5 receptors might fulfill an activating function that is coupled with inhibitory functions of the ITIM-containing Nitrs 1–4, 6, and 7. Although the ligand binding of NITRs is presently uncharacterized, it has been possible to make a number of observations regarding the cytoplasmic signaling and developmental expression of NITRs, which may facilitate understanding their function. Specifically: (1) studies using vacciniatransfected, FLAG-tagged Nitr3.1 indicate that epitope-crosslinking results in ITIM-mediated inhibition of both the phosphorylation of MAP kinase and NK cytotoxicity; (2) the earliest NITR expression in somatic cells can be detected at seven dpf, with the exception of nitr3, which can be detected at 72 hpf; in addition, maternal stores of nitr3 mRNA have been observed (J.A.Y. and G.W.L., unpublished observations); and (3) embryonic expression

272

D. TRAVER ET AL.

of nitr3 appears to be downregulated (disrupted) in a zebrafish mutant line (H75) that displays delayed hematopoietic development (J. A. Y. and N. S. T., unpublished observations). The hypothesis that the NITR genes may represent another variant form of NK-related receptors (Litman et al., 2001) is based on the general organizational features of NITRs, as well as the patterned sequence diversity in the V ectodomains, which is consistent with ligand binding. Furthermore, the polygenic and polymorphic character of NITRs, the preponderance of inhibitory to activating forms and the close similarity in the ectodomains of some activating and inhibitory forms of NITRs are characteristic of certain NK receptors (Litman et al., 2001). Specifically, the variable nature of the N-terminal ectodomains, character of the transmembrane and cytoplasmic regions, as well as the polygenic and polymorphic nature of NITRs, collectively have led us to draw analogies between these genes and KIRs (Litman et al., 2001). Despite the presence of diversified V region, NITRs resemble KIRs more closely than do the C-type lectins (e.g., Ly49) that mediate NK function in mouse. Although NKp30 and NKp44 are V domain-containing NK receptors (Pessino et al., 1998; Cantoni et al., 1999; Pende et al., 1999), they are single copy as opposed to the highly diversified multigenic families encoding the NITRs. The studies conducted to date with the various families of NITR genes in zebrafish have paved the way for future research aimed at (1) identifying additional NITRs as well as related genes, (2) defining the ligands recognized by NITRs, and (3) characterizing the functional role NITRs may play in the natural killing of tumor cells. The ability to produce transgenic zebrafish and perform mutagenesis screens makes zebrafish amenable to studies addressing the role of NITRs in the development of the immune system. The continued investigations into the forms and functions of the NITRS may contribute to a better understanding of how NK receptors function in mammals and likely will shed light on the evolutionary pathway(s) that have led to the unusual diversity found in mammalian NK receptors. III. Ontogeny of Adaptive Immunity from Fishes to Mammals

As mentioned above, the zebrafish as a model system for immunology is a relative latecomer. Thus, it is prudent to review the abundant literature on comparative vertebrate immunity when considering the zebrafish adaptive immune system. In particular, it should be realized that the components of the adaptive immune system, namely the immunoglobulin (Ig), T cell receptor (TCR) and major histocompatibility complex (MHC), are not necessarily identical to those of mouse and human (where they have been described most

273


extensively) and may exhibit notable differences between various vertebrate lineages (Du Pasquier and Flajnik, 1998). These differences primarily involve variations in genomic organization of the respective gene families, and manifest in presumptive alterations in the manner in which these genes are regulated. In addition, there are poignant differences in the types of hematolymphoid tissues within and among the various vertebrate classes, including the primary sites of production of T and B cells. This section is not meant to be a comprehensive treatise of the evolution and phylogeny of the adaptive immune system; rather, we review what is known about adaptive immunity amongst the poikilothermic (cold-blooded) vertebrates and highlight significant departure points from other systems. Where known, data on primary and secondary lymphoid organs, as well as the Ig and TCR systems of zebrafish are presented and discussed. A. THYMIC DEVELOPMENT IS CONSERVED THROUGHOUT EVOLUTION OF JAWED VERTEBRATES

THE

1. Phylogeny of Thymic Organogenesis The phylogeny of extant chordate taxa, all of which have been examined, to some degree, with respect to the adaptive immune system, is outlined in Fig. 2. It is important to note that although lymphocyte-like cells are seen in protovertebrate taxa such as urochordates and cephalochordates, it was not until the emergence of the gnathostomes (jawed vertebrates), roughly 500 million years ago, that the presence of Ig, TCR and MHC are first observed (Du Pasquier and Flajnik, 1998; Litman et al., 1999). During evolution the development of the jaw in vertebrates coincides with the emergence of adaptive immunity and the thymic organ. The thymus is the centerpiece of the adaptive immune system because it provides the environment for generation and selection of large numbers of antigen-specific T cells. The appearance of the thymic organ in evolution therefore marks an important milestone in the ability to fight infections with a specific cell- and/or antibody-mediated immune response. The phylogenetically oldest vertebrates, lamprey and hagfish, are jawless and therefore lack a thymus and evidence of secondary lymphoid organs. It was long held that these agnathan species of fish rely solely on granulocytes (Rowley et al., 1988) and C3-like complement (Nonaka, 2000; Zarkadis et al., 2001) for protection against infections. Support for this notion comes from the fact that the pivotal components of adaptive immunity, the T- and B-cell receptors for antigen, as well as MHC molecules have not been identified in this class of fishes (Klein et al., 1999). However, lymphocyte-like cells were detected in these jawless fishes, particularly in the intestinal tract (Kampmeier, 1969; Cohen, 1977; Marchalonis, 1977). Recently, the demonstration of the presence of members of the

274

D. TRAVER ET AL.

FIG. 2. Inter-relationships of the phylum Chordata. This figure was modified from Nelson (1984) and Zapata and Amemiya (2000). Taxa are shown above the tips of the cladogram and welldefined monophyletic assemblages are listed below the cladogram. The zebrafish, Danio rerio, is a member of the teleost (bony) fish lineage. This cladogram merely provides a framework for the major extant chordate clades. With regard to the immune system, salient features of immunological import are listed briefly in Table I. The adaptive immune system appears to be a gnathostome (jawed vertebrate) invention, corresponding possibly to the co-option and utilization of RAG-based somatic diversification in rearranging genes (Du Pasquier and Flajnik, 1998; Plasterk, 1998; Litman et al., 1999; Harsen and McBlane, 2000). It is important to note that the Agnatha, a group that includes the extant cyclostomes (M) Myxini (hagfishes) and (P) Petromyzontiformes (lampreys), is thought to be a ‘‘paraphyletic’’ group. Several advanced anatomical characters exist in lampreys that are absent in hagfishes, although the presence or absence of neural crest cells (which, in part, define the Vertebrata) has not been adequately assessed in hagfishes due to the paucity of embryos. Lastly, lungfish are placed with the coelacanth as they both are ‘‘lobe-finned’’ fishes; however, lungfish are not technically ‘‘crossopterygians.’’

Ikaros (Mayer et al., 2002b) and Spi family (Shintani et al., 2000) of DNA-binding molecules in the lamprey genome and their expression in lymphocyte-like cells in the gastrointestinal tract bolsters the notion that lymphocytes may predate the development of the thymus. The functionality of these lymphocyte-like cells remains to be elucidated as no obvious homologs of Ig superfamily genes involved in immune recognition or lymphocyte differentiation were identified in a screen of over 8000 ESTs (Uinuk-Ool et al., 2002; Mayer et al., 2002a).


275

It has been speculated that with the appearance of the jaw dietary changes as well as increased injuries and infections occurred, creating selective pressure for the development of an organ of high output of immune cells with a diverse and anticipatory repertoire (Matsunaga and Rahman, 2001). Thus the cartilaginous (chondrichthyes) and bony fishes (osteichthyes) are the first vertebrates in which a thymus can be identified. In both classes of jawed fishes the thymus is a bilateral organ located close to the gill cavity. Amongst teleosts, the bony fish division possesses thymi that remain in continuity with the pharyngeal epithelium. Furthermore, there is debate as to whether the teleost thymus is compartmentalized into cortex and medulla (Romano et al., 1999a,b; Lam et al., 2002), or if it lacks a clear demarcation (Castillo et al., 1990; Zapata et al., 1996, p. 359; Willett et al., 1997). In non-teleostean fishes the thymus appears to be composed of a clearly distinguishable cortex and medulla (Zapata and Cooper, 1990). Gene expression patterns (see below) and the capacity to mount T cell dependent immune responses, such as specific antibody production (Palm et al., 1998; Espelid et al., 2001), graft rejection and GvHD (Nakanishi and Ototake, 1999) suggest that there is no functional difference between the teleost thymus and that of other fishes or mammals. The compartmentalized thymus is a constant feature of higher vertebrates from frogs to birds and mammals, and also includes sharks and skates. 2. Ontogeny Thymic development from lower vertebrates to mammals requires the contribution of tissues from all three embryonic germlayers: ectoderm, mesoderm and endoderm (see Fig. 3). Two steps in this process can be distinguished-formation of the thymic rudiment and the interaction between T cells and thymic epithelial cells (TECs) (excellent reviews in Manley, 2000; Rodewald, 2003). Common to all vertebrates is the first step in the formation of the thymic anlage. It initiates with the migration of neurectodermally derived neural crest cells from the sixth hindbrain rhombomere to the pharyngeal area. It has been proposed that interaction of neural crest cellderived mesenchyme with endodermally derived epithelial cells of the pharyngeal pouches is crucial for thymic development (Anderson et al., 1993; Shinohara and Honjo, 1997; Suniara et al., 2000). This concept was first proposed when the seminal experiments of Bockman and Kirby (Bockman and Kirby, 1984) demonstrated that ablation of the sixth hindbrain rhombomere in chicks leads to the absence of the thymus. Positional information indispensable for proper neural crest cell migration is provided by the homeobox gene Hoxa-3 and the paired box gene Pax-3. Inactivation of these genes leads to defects in thymic and cardiac development similar to those observed in the human DiGeorge syndrome.

276

D. TRAVER ET AL.

FIG. 3. Expression of genes involved in zebrafish T cell lymphopoiesis and thymic organogenesis. A. The hindbrain rhombomeres (left panel) from which neural crest cells arise are stained with Krox-20 (third and fifth rhombomere, r3 and r5). Neural crest cell (NCC) expression of forkhead 6 ( f kd6) in 15-somite (a developmental stage, approximately 20 hpf, at which 15 body segments can be counted) wild-type embryo is shown in the middle panel. Arrow indicates the third, postotic (behind the ear vesicle) stream of neural crest that invades the third and fourth pharyngeal pouches. In d2 wild-type embryos (right panel) Hoxa-3 is expressed in rhombomeres 5 and 6 (arrow) and pharyngeal endoderm (arrowhead). B. Expression of Nkx2.3 in the pharyngeal endoderm in d2 wild-type larva. Expression is seen in pharyngeal pouches (arrow) and in anterior gut (asterisk). C. Mesodermally derived hematopoietic progenitor cells, stained with c-myb (arrow), presumably arise from the dorsal aorta in zebrafish embryos at 36 hpf. Pro-T cells probably originate from this area and invade the thymic rudiment. D. Alcion blue staining of skeletal elements of the head region in a d7 wild-type zebrafish shows arches p1 to p7. Asterisk indicates the third pharyngeal pouch, where thymus originates. E. Thymus (grey arrow) of d7 wild-type zebrafish with heterogeneous cell populations, consisting of lymphoblasts and thymic epithelial cell. F. Expression of Rag-1 in d8 wild-type embryo is seen in the olfactory pit (grey horizontal arrows) and bilateral thymi (arrows). (Reproduced with modifications from Trede et al. (2001) with permission.)

A number of genes are expressed at the critical developmental time point (E9.5–10.5) for neural crest cell–endoderm interaction in the pharyngeal pouches between arches 3, 4, and 6. For example, inactivation of the T box gene Tbx-1, as well as Pax-1 and Pax-9 similarly lead to defects in the formation of proper pharyngeal arch and artery architecture and result in a


277

DiGeorge-like phenotype of variable severity. These results emphasize the importance of pharyngeal endoderm in the formation of the thymic rudiment. Hoxa-3 is expressed in both migrating neural crest cells and endoderm (Manley and Capecchi, 1995) (see Fig. 3) and the absence of this gene in both tissues might be responsible for the severe thymic defects seen in Hoxa-3/ mice (Su and Manley, 2000; reviewed in Manley, 2000). Furthermore, attempts have been made at trying to place the above genes in a hierarchical order. For example, an interaction of Hoxa-3 and Pax-1 is suggested by the severity of thymic defects in double-mutant Hoxa-3 þ / Pax-1/ mice (Su and Manley, 2000). A picture emerges where Hoxa-3 is placed ‘‘upstream’’ of Pax-1 (Su and Manley, 2000; reviewed in Manley, 2000) and Pax-9 (Peters et al., 1998; Hetzer-Egger et al., 2002) in a pathway to form the endodermal basis of the thymic rudiment (Su and Manley, 2000; reviewed in Manley, 2000). A functional whn gene is indispensable for TEC differentiation. This is borne out in studies with the whn-deficient nude mouse, which exhibits differentiation arrest and lack of migration of T cell precursors into the thymic rudiment (Nehls et al., 1996). Whether the whn gene product acts in a separate pathway from the Hoxa-3/Pax-1 pathway (Su and Manley, 2000) or is placed between Hoxa-3 and Pax-9 as recently suggested by Tom Boehm’s group (Hetzer-Egger et al., 2002) remains an open question. 3. Thymic Organogenesis in Zebrafish a. Formation of the Thymic Rudiment. Based on the above data obtained in chicks and mammals we obtained molecular markers to probe evolution of the thymic rudiment during zebrafish development (Fig. 3). Neural crest cells contributing to thymic organogenesis are expected to arise from hindbrain rhombomeres 5 and 6 and migrate in the post-otic stream to the pharyngeal area. Krox-20 (red dye) is expressed in hindbrain rhombomeres 3 and 5 (Fig. 3A, left panel) and is useful as a landmark to identify the region of the hindbrain of interest. Prior to and during migration, neural crest cells express the transcription factor forkhead 6 (Fkd6, Fig. 3A, middle panel). In agreement with data from mice, Hoxa-3 is expressed in the fifth and sixth rhombomere and also stains pharyngeal endoderm (Fig. 3A right panel) (Su and Manley, 2000). Nkx2.3, a marker of pharyngeal endoderm, is expressed throughout pouches two to six (Fig. 3B). Similarly to mammals, formation of pharyngeal arches in zebrafish (Fig. 3D) is critically dependent on interaction of neural crest cells and pharyngeal epithelium and endoderm. For example, the zebrafish mutant sucker (suc) has defects in the four most anterior pharyngeal arches (Piotrowski and Nusslein-Volhard, 2000). Cloning of the defective suc gene identified a mutation in the endothelin-1 (et-1) gene,

278

D. TRAVER ET AL.

which is expressed in a central core of arch paraxial mesoderm and in arch epithelia, both surface ectoderm and pharyngeal endoderm, but not in skeletogenic neural crest (Miller et al., 2000). Miller et al. demonstrated that suc/et-1 functions nonautonomously in neural crest cells, and is thus required in the environment of post migratory neural crest cells to specify ventral arch fates. These results confirm and extend previous findings observed in the mouse knockout of et-1 which results in a disruption of pharyngeal arch architecture and cardiovascular defects reminiscent of the human CATCH 22 syndrome (Kurihara et al., 1994). An identical phenotype is caused by targeted inactivation of the Endothelin Converting Enzyme-1 (ECE-1) necessary for proteolytic activation of Endothelin-1 in mouse embryos (Yanagisawa et al., 1998). Finally, disruption of the Endothelin-A Receptor (ETAR), which is expressed by the neural crest-derived ectomesenchymal cells of pharyngeal arches and cardiac outflow tissues, causes a phenocopy of the et-1 and ECE-1 knockouts (Clouthier et al., 1998). A pathway similar to the et-1–ETAR interaction has to date not been described for the more caudal pharyngeal arches. In contrast to mammals, the zebrafish thymi remain in continuity with the pharyngeal endoderm. The thymic rudiment arises between pharyngeal arches 3–5, in the third and fourth pharyngeal pouch (our own observations and Lam et al., 2002). It can first be detected between 54 and 60 hpf as a small outgrowth of pharyngeal epithelium (Willett et al., 1997, 1999). At this stage, the thymic rudiment consists of two layers of epithelial cells surrounded by a basement membrane (Willett et al., 1999) and is devoid of lymphocytes. In mouse, the prospective TECs express Whn/Foxn1, the gene disrupted in the nude mouse (Nehls et al., 1996), which is absolutely required for further epithelial differentiation into thymic tissue. The initially reported zebrafish whn/Foxn1 homolog (Schlake et al., 1997) is expressed in the developing eye and in brain structures (our own observations and Schorpp et al., 2002), as well as in a few scattered epithelial cells in the adult thymus. However, this gene was never expressed during early thymus development, initiating a search for the true zebrafish ortholog of whn/Foxn-1. Recently, a second zebrafish whn gene (designated whnb) was cloned by the group of Tom Boehm (Schorpp et al., 2002). The whnb gene appears to be the true orthologue of Foxn1 as it is expressed in TECs on day 3 of development and is furthermore more closely related to the mammalian Foxn1 gene by phylogenetic analysis and synteny (Schorpp et al., 2002). The original zebrafish whn gene (now designated whna) is now thought to be the orthologue of Foxn4 (Schorpp et al., 2002). T cell precursors in mammals are derived from hematopoietic stem cells of mesodermal origin in the yolk sac (Yoder et al., 1997) and aortagonad-mesonephros (AGM) region (Cumano et al., 1996, 2000). There is


279

evidence to suggest that the AGM equivalent in zebrafish is the dorsal aorta (Thompson et al., 1998). For example, markers of early hematopoietic differentiation, such as c-myb, are expressed in the dorsal aorta between 36 and 48 hpf (Thompson et al., 1998) (Fig. 3C). It is therefore possible that T cell precursors in the zebrafish are derived from progenitor cells in the ventral wall of the dorsal aorta. There are currently no markers available to corroborate the presence of T cell precursors in the dorsal aorta. Furthermore, transplantation or in vitro differentiation experiments are only beginning to be explored in zebrafish. Ultrastructural analysis (Willett et al., 1999) identifies the first T cell precursors in the thymic rudiment at 65 hpf and WISH at 72 hpf (Willett et al., 1999; and our own observation). This leads to a bilateral staining pattern with lymphoid probes such as Rag-1 (see Fig. 3F). A section of a day (d) 7 wild-type thymus displaying a heterogeneous population of cells of epithelial and hematopoietic origin is shown in Fig. 3E. In summary, genes required for the formation of the thymic rudiment in chicks and mammals are expressed in zebrafish in the correct temporo-spatial frame. This suggests that early thymic development is conserved throughout vertebrate evolution. b. Thymus Development Beyond the Larval Stage. Subsequent to its initial formation during the first week of development, the zebrafish thymus undergoes a morphologic transformation, which has been examined by detailed morphometric analysis (Lam et al., 2002). From its original pouch-like shape the thymus rapidly grows into a cone-like structure between weeks two and six of life. At week three a demarcation into thymic cortex and medulla is first discernible (Lam et al., 2002; Schorpp et al., 2002) evidenced by absence of Rag-staining in the medulla, paucity of TCR-alpha in the cortex, which remains in close proximity to the pharyngeal cavity (Schorpp et al., 2002) and strong TCR-alpha expression in the medulla. Furthermore, the cortex is far more cellular than the medulla, in keeping with findings in mammals, where positive selection occurs predominantly in the cortex, while negative selection is effected at the cortico-medullary zone. In addition, septa-like structures were observed in the 6-week-old thymus at the cortico-medullary junction. Whnb is expressed in epithelial cells of both, cortex and medulla (Schorpp et al., 2002). Lam et al. also described involution of the zebrafish thymus as measured by a decrease in thymic volume as a consequence of atrophy and replacement with connective tissue. The initiation of this process coincides with sexual maturity, a phenomenon also known to occur in other fishes (Chilmonczyk, 1992) and mammals (Turner, 1994).

280

D. TRAVER ET AL.

B. T CELL DEVELOPMENT 1. T Cell Receptor Genes in Teleosts Unlike the immunoglobulins (see below), much less is known regarding the TCRs of primitive vertebrates. The higher molecular evolutionary rate of TCRs precludes routine cross-hybridization screening approaches and the vast majority of studies have required highly degenerate PCR strategies as a means to isolate the respective TCR genes (Rast et al., 1995; Hawke et al., 1996, 1999; De Guerra and Charlemagne, 1997; Wilson et al., 1998; Haire et al., 2000). While scant data are available regarding the TCRs in primitive vertebrates, it is presumed that all four classes of TCRs (, , , ) will be identified since they have been found in the skates, a taxon which represents the most basal vertebrate that possesses an adaptive immune system (Rast et al., 1997; Litman et al., 1999). Thus far, TCR genes have been isolated from the zebrafish (Haire et al., 2000) and their expression patterns are consistent with early development of the T cell compartment (Trede et al., 2001; Lam et al., 2002). As an alternative to the PCR based strategy, we have taken a genomic sequencing approach in order to identify and characterize zebrafish TCR genes. Using probes to both V and C we constructed a PAC contig encompassing the entire TCR locus (T. Ota et al., unpublished). Partial annotations are listed for GenBank nos. AL591481, AL591511 and AL592550. Thus far the sequencing data suggest that the locus contains over 150 non-allelic V genes and that these can be classified into at least 86 different V families. Characterization of the expression patterns of many of these families is in progress (L. Steiner, unpublished). In terms of genomic organization, it appears that the zebrafish locus is very similar to that of the freshwater pufferfish, Tetraodon nigroviridis, whose TCR locus has been sequenced in its entirety (Fischer et al., 2002). The linkage of TCR genes to the TCR locus in zebrafish has not yet been established, however, it is predicted to be tightly linked as in the pufferfish. The gene complexity of the zebrafish TCR locus, however, is considerably greater than that found in either Tetraodon or Fugu (Fischer et al., 2002; Fugu genome supplementary online materials: www.sciencemag.org/cgi/content/full/1072104/DC1). 2. T Cell Ontogeny in Zebrafish In mouse, T cell precursors can be found on embryonic day 9.5 (E9.5) in the yolk sac (Yoder et al., 1997) and on E10.5 in the AGM region (Cumano et al., 1996; 2000). As mentioned above (Section III.A.3.b) the anatomic region corresponding to the AGM region in zebrafish may be the dorsal aorta circa 36 h of development. Data to corroborate this line of reasoning come from gene expression studies, demonstrating expression of markers of


281

hematopoietic progenitors, such as c-myb, in the dorsal aorta (Thompson et al., 1998). In the mouse, T lineage-committed (Thy-1 þ c-kitlow) cells have been identified in fetal blood circulation (Rodewald et al., 1994). This type of analysis in the zebrafish awaits identification of the Thy-1 and c-kit orthologs and/or availability of antibodies directed against the corresponding surface proteins. Transplantation of hematopoietic precursors from various stages of development derived from transgenic fish expressing green fluorescent protein (GFP) under the control of the GATA-1 or LMO2 promoters into Rag1 deficient recipients (see Sections IV.B and VII.A) could be instrumental for the identification of the origin of pro-T cells in zebrafish. Identifying GFP positive cells in the thymus of the Rag-1 deficient recipients, followed by FACS sorting and morphologic as well as gene-expression analysis would be a straightforward assay in this system. By ultramicroscopy and gene expression studies T cell-specific genes can first be detected at 72 h in the thymus, making it the first zebrafish organ to become lymphoid. Analysis of T cell development by WISH during the first week of wild-type zebrafish development is shown in Fig. 4. By analogy with mammals, T cells are expected to express genes such as Ikaros, GATA-3, Rag-1, lck and TCR during early development. Ikaros belongs to the family of zinc-finger transcription factors that act early in lymphoid development (Georgopoulos et al., 1994) and is expressed in the brain and thymus (Willett et al., 2001). GATA-3, which acts at the level of the T cell progenitor is expressed in brain, thymus and pharyngeal endoderm. Rag-1 (Fig. 4) and Rag-2 (not shown) are expressed in the thymus and the olfactory placode throughout the lifespan of zebrafish. The presence of Rag-1 and Rag-2 in the olfactory placode of zebrafish (Jessen et al., 1999, 2001) prior to the onset of thymic population with T cells is intriguing, but so far gene expression studies (our own observations) and histologic analysis (Jessen et al., 2001) have failed to demonstrate the presence of T cell precursors in this organ. For example, the src kinase p56lck is exclusively expressed in the thymus (Fig. 4). The thymus is permissive for the development of non-T cells belonging to the myeloid and B cell lineages (Barcena et al., 1994). In the absence of standard T cell markers of differentiation in zebrafish, such as CD4 and CD8, or B cell markers such as CD19, it is not possible to attribute the lymphoid cells in the thymus to the T or B cell lineage definitively. However, the expression of GATA-3 and TCR, combined with the absence of Pax-5 expression (Fig. 4, bottom panel), in the zebrafish thymus strongly argues for a T cell identity of these cells. Furthermore, B cells of higher vertebrates develop primarily in the bone marrow. The pronephros/kidney of the zebrafish, which is the larval and adult site of hematopoiesis in this organism, is devoid of Rag-1 expression until 21 days of development (Willett et al., 1999). The paucity of available reagents in the zebrafish, such as antibodies directed against T cell surface markers, or

282

D. TRAVER ET AL.

FIG. 4. Expression of lymphoid genes in zebrafish larvae. The top row shows whole mount in situ hybridization (WISH) analysis of day (d) 3.5 wild-type zebrafish larvae with the indicated probes. A grey arrowhead indicates expression of Ikaros and GATA-3 in brain, black arrows indicate thymic expression with all probes, and a grey vertical arrow indicates Rag-1 expression in the olfactory pit. Expression of the indicated probes is shown in the middle row in d7 wild-type larvae. The bottom row shows expression of TCR and Pax-5 in d8 and d4 wild-type larvae, respectively. The grey arrowhead indicates expression of Pax-5 at the midbrain/hindbrain boundary. Given the transparency and positioning of the larvae, contralateral thymic staining at d3.5 with the Rag-1 and at d8 with the TCR probes is seen. (Reproduced with modifications from Trede et al. (2001) with permission.)

genetic probes for markers of T cell activation and differentiation, currently renders a detailed analysis of T cell development difficult. We are in the process of creating a zebrafish lymphoid library (see Section III.A.1), which will provide some of the reagents necessary to gain new insights into zebrafish lymphopoiesis. C. B CELL DEVELOPMENT

IN

TELEOSTS

1. Immunoglobulin Genes in Teleosts Among the receptor molecules of the adaptive immune system, immunoglobulins have been the most extensively studied among vertebrates. This is primarily due to the fact that Ig gene probes will cross-hybridize across


283

wide phylogenetic distances, a situation not possible for MHC or TCR genes. Indeed, the first shark Ig VH genes were isolated via cross-hybridization with a mouse VH segment (VH-S107) known to be specific for phosphorylcholine (Litman et al., 1985). Thus, using heterologous gene probes it has been possible to isolate and characterize IgH and IgL genes from myriad vertebrate species both at the cDNA and genomic levels. This has resulted in a fairly comprehensive survey, albeit with some taxon sampling biases, of Ig genes across the vertebrate radiation. Table I lists features of the Ig system from the major lineages of vertebrates and highlights salient similarities and differences among and within the group. While the overall structure and function of immunoglobulin molecules per se are highly similar across all gnathostomes, the genomic organization of immunoglobulin genes can be quite disparate. That is, the V, D (heavy chain only), J and C components that comprise Ig genes are highly similar between vertebrate lineages. The genomic structure of the Ig families, however, can be radically different. This is evinced by the different genomic organizations observed within and among the vertebrates. The two general forms observed are the ‘‘clustered’’ arrangement first seen in the IgH genes of cartilaginous fishes and the ‘‘translocon’’ arrangement described for mouse and human (see Table I). The translocon arrangement is the organization found in mouse and human and consists of a long, extended locus comprised of tandem segments that are recombined during B cell development (Vn-Dn-Jn-Cn for the heavy chain, where n can be anywhere from 1 to >100). In contrast, the clustered arrangement is fundamentally different in that each cluster contains essentially one or a few respective segments, e.g., (V-D-J-C)n, where the cluster is reiterated throughout the genome (Du Pasquier and Flajnik, 1998; Litman et al., 1999; Bengten et al., 2000). In both situations, translocon and clustered, RAG-driven recombination between the V, (D) and J segments are required. Another salient feature of humoral immunity in lower vertebrates is that the tissues associated with development and differentiation of B lymphocytes vary between major lineages (Table I). For example, in fishes (including cartilaginous fishes) there is no bone marrow; B lymphocyte development thus occurs at other sites such as kidney, spleen and elsewhere. The bone marrow can thus be considered a ‘‘derived’’ tissue within vertebrates that has been co-opted for B lymphocyte development (B cell receptor production) in the more recent phylogenetic lineages (Table I). With zebrafish, we need to be cognizant of these sorts of differences if we are to use this system as an immunological model. Most recently, it was suggested that the pancreas is the primary site of early B cell development and Ig expression in the zebrafish (see Section III.C.2), a surprising finding given that the pancreas has never been recruited for this function in other vertebrates.

284

TABLE 1. IMMUNOGLOBULINS AND ASSOCIATED FEATURES AMONG THE VERTEBRATES.1 Ig isotypes

Gene organization3

Germinal centers

Class switching

kidney, protovertebral arch (fat column), intestine epigonal organ, Leydig organ, spleen, kidney, gonads

none

n/a

n/a

n/a

clustered for both heavy and light chain loci

no

no

head kidney, spleen, pancreas, gonads, heart

IgM, IgNAR, IgW; three or more different light chain isotypes IgM, IgD; two or three light chain isotypes

no

no

yes

liver, kidney, spleen, gonads, spiral valve in lungfish; not known in coelacanth

IgM, two other isotypes known in lungfish; light chain isotypes unknown

no

?

amphibians

yes

bone marrow, liver, kidney

no

yes

reptiles birds

yes yes

bone marrow, other? Bursa of Fabricius

translocon translocon

no yes

yes? yes

mammals

yes

bone marrow

IgM, IgY, IgX; two or three different light chain isotypes IgM, others? IgM, IgY, IgG; two light chain isotypes IgM, IgD, IgG, IgA, IgE; two light chain isotypes (, )

translocon for heavy chain; clustered for light chain coelacanth has an IgM organization intermediate between clustered and translocon translocon

translocon

yes

yes

Adaptive immunity

cyclostomes

no

cartilaginous fishes

yes

teleost fishes

yes

lobe-finned fishes

1 Data taken and/or summarized from numerous sources including: (Amemiya et al., 1993, Aparicio et al., 2002, Bengten et al., 2000, Turchin and Hsu, 1996, Marchalonis et al., 1998, Zapata and Cooper, 1990, Danilova and Steiner, 2002, Zapata and Amemiya, 2000, Litman et al., 1993, Du Pasquier and Flajnik, 1998, Flajnik, 2002, Warr, 1997). This table is a broad generalization only and is not meant to be all-encompassing. 2 ‘‘Bone marrow equivalents’’ refers to the histological observation of hematolymphoid tissue resembling bone marrow, the site of mammalian B-cells. Note, some of the tissues listed in this column may actually represent secondary sites of development of B-cells. 3 The mammalian type, extended locus with multiple tandem segments is referred to as the ‘‘translocon’’ organization; the shark organization whereby individual segmental elements are organized in discrete units (e.g., V-D-J-C for the heavy chain) is referred to as the ‘‘clustered’’ organization (Litman et al., 1999, Bengten et al., 2000).

D. TRAVER ET AL.

Bone marrow equivalents2

Vertebrate taxon


285

Immunoglobulin Ig VH and C genes have been isolated from the zebrafish and are similar in inferred amino acid sequences to those of other vertebrate species (Danilova et al., 2000). Preliminary data from genomic mapping and sequencing studies of zebrafish IgH-containing PAC clones suggests that the zebrafish, as predicted, has a translocon organization of its heavy chain locus (T. Ota, unpublished). Isolation of its light chain genes as well as genomic characterization of its other heavy chain isotype (IgD) (Wilson et al., 1998; Bengten et al., 2000) is in progress. It is predicted, however, that the results will be unremarkable based on the genome sequence of Fugu (Aparicio et al., 2002), a comparatively closely related teleost fish. An area of fundamental importance in humoral immunity is the regulation of gene expression of the Ig genes. Considerable progress has been made in understanding the IgH enhancer in the channel catfish (Magor et al., 1994, 1997; Cioffi et al., 2001). It is important to note that the IgH locus in the catfish exhibits a translocon organization similar to that of mammals, yet its major IgH enhancer is in a vastly different locale. In the mouse and human, this enhancer is found in the intron between the JH region and the first exon of C1. In the catfish, however, the strongest IgH enhancer has been functionally localized to a region near the second transmembrane exon (Magor et al., 1994; Warr, 1995). Using the same mammalian cell culture-based system for assaying enhancer activity, B. Magor and colleagues (Univ. Alberta) have shown that the major zebrafish enhancer is not localized to the same region as in either the mouse or catfish (B. Magor, personal communication). This is somewhat surprising given the close phylogenetic relationship of the zebrafish and catfish within the Ostariophysans (Nelson, 1984). It will be interesting to determine whether the zebrafish system has enhancers and other cis-regulatory elements that specifically drive IgM expression in the pancreas. 2. B Cell Development Initiates in the Zebrafish Pancreas Detection of T cells by WISH using probes such as Rag-1 (see Section III.B.2) starting at 72 hpf is facilitated by the very superficial position of the thymus. B cell development is more difficult to examine by WISH given the central location of the pronephros, the site where lymphocytes are produced starting at three weeks post fertilization (Willett et al., 1999). At that stage of development WISH is unable to detect transcripts except for the most superficial regions of the juvenile zebrafish. Transcripts of Rag-1 as well as c were detectable on sections in adult zebrafish in the kidney and intestine (Danilova and Steiner, 2002), consistent with other teleost species (Zapata and Cooper, 1990; Zapata et al., 1996; Fournier-Betz et al., 2000; Zapata and Amemiya, 2000) and higher vertebrates, where B cells are produced in the bone marrow and are also found in the intestine. Intestinal B lymphocytes in

286

D. TRAVER ET AL.

teleosts are found in the loosely organized gut associated lymphoid tissue (Picchietti et al., 1997; Fournier-Betz et al., 2000; Zapata and Amemiya, 2000). Surprisingly, Danilova and Steiner also found co-expression of Rag-1 and c in the pancreas of adult zebrafish. Tracing this gene expression pattern back in ontogeny, the pancreas was shown to be the initial site of B cell development in zebrafish larvae, initiating as early as 10 dpf by sectioning and 4 dpf by WISH (Danilova and Steiner, 2002). By RT-PCR, membrane IgM was first expressed on day 7 pf. Although double-stains with pancreas-specific (e.g., insulin) and lymphocyte-specific probes (e.g., Rag-1, Ig) were not done, these findings provide the first suggestion of initial B cell development in pancreatic tissue and open the way for future studies. For example, it will be interesting to study other teleosts and higher vertebrates for a role of the pancreas in B cell development. As the developing B cells are presumably not produced in situ in the pancreas, but are rather derived from hematopoietic stem cells, it will be interesting to determine the soluble factors secreted by the developing pancreas, which attract B cell progenitors to the organ. Furthermore, zebrafish mutants affecting pancreas development can be studied to ascertain the necessity of this organ for physiologic B cell development. Analogous experiments can be carried out by transient inhibition of pancreatic development using morpholinos (Huang et al., 2001; Yee et al., 2001) followed by analysis of this delay on B cell development. D. SECONDARY LYMPHOID ORGANS 1. The Spleen is the Major Secondary Lymphoid Organ in Teleosts Primary lymphoid organs are the sites where T and B cell production occurs. Clusters of lymphoid cells can first be observed in lampreys (Zapata et al., 1981; Hagen et al., 1983) and in the Atlantic hagfish in the ‘‘central mass’’ of the pronephros (Zapata et al., 1984; Zapata and Amemiya, 2000), which is the site of bone marrow hematopoiesis in many lower vertebrates. However, there are no distinct secondary lymphoid organs in these primitive vertebrates (Zapata et al., 1996; Zapata and Amemiya, 2000). It is only with the advent of modern fishes that primary and secondary lymphoid organs are truly discernible. In chondrichthyes and osteichthyes B cells are produced in different organs that represent sites of hematopoiesis. These can range from the meninges to the gonads, but in most species involve the kidney (reviewed in Zapata and Amemiya, 2000), while T cells are always produced in the thymus (see above). Secondary lymphoid organs in vertebrates serve as trapping and processing devices for antigen. It is here that antigen-presenting cells communicate with T cells, and where T cell–B cell interactions take place. The major secondary lymphoid organ consistently found in fishes is the spleen. The mammalian


287

spleen is compartmentalized into areas of erythroid predominance (red pulp) and lymphoid follicles (white pulp). At the border between red and white pulp is the marginal zone, where antigen trapping occurs and cells transit in and out of the white pulp. Upon exposure to antigen, B–T cell cooperation leads to activation of both cell types. Activated B cells along with the antigen-specific T helper cells migrate to the primary follicles of the spleen, where the B cells form distinct structures, the germinal centers. It is here that B cell repertoire selection occurs through isotype class switching and somatic hypermutationmediated affinity maturation of the B cell receptor for antigen (BCR). In birds and rabbits gene conversion in secondary lymphoid organs is the major event that confers diversification to the Ig repertoire. There appears to be no direct correlation between germinal center formation and efficient affinity maturation. For example in birds, which form germinal centers during secondary immune response, affinity maturation is rather poor (Du Pasquier et al., 1998). In contrast to higher vertebrates, the marginal zone separating the splenic white and red pulp is not fully developed in fishes. White pulp, consisting of lymphocytes, APCs and plasma cells are intermingled with red pulp (Zapata and Amemiya, 2000). Following antigen stimulation white pulp in teleosts increases in size. However, germinal centers are conspicuously absent in ectotherms, including fishes. Class switching is a process that has its earliest evolutionary roots in amphibians and does not occur in fishes. On the other hand, somatic hypermutation has been clearly demonstrated in lower vertebrates, including fishes. However, affinity maturation of immunoglobulins is inferior in fishes compared to higher vertebrates and the anatomic site where this process occurs in ectotherms is unknown. The three processes affecting diversification of the immunoglobulin repertoire, somatic hypermutation, class switching and gene conversion (found in birds and rabbits) depends on the activity of a recently discovered enzyme, activation-induced cytidine deaminase (AID). The presence of this enzyme has been documented and its activity studied in mammals. Given the evidence for hypermutation of the shark heavy and in particular light chains (Lee et al., 2002b), AID has been presumed to be present in lower jawed vertebrates (see above). It is conceivable that the mechanism of hypermutation evolutionarily precedes adaptive immunity (Flajnik, 2002). Verification of this hypothesis awaits cloning of the AID gene and study of its function in agnathan vertebrates. The gut is the other anatomic site where lymphocytes are consistently found in fishes. Lymphoid aggregates are found in the lamina propria of the teleostian gut, but they are not encapsulated and hence do not represent true, isolated lymphoid organs. These aggregates represent mostly Ig positive

288

D. TRAVER ET AL.

plasma cells (Rombout et al., 1993). DLT15 þ T cells were also found in the intestine of the sea bass (Dicentrarchus labrax) (Picchietti et al., 1997; Romano et al., 1997) and probably represent a majority of the Ig negative intraepithelial lymphocytes (Rombout et al., 1993). 2. Secondary Lymphoid Tissues in Zebrafish As mentioned above, teleosts do not have lymph nodes. This makes the spleen the major secondary lymphoid organ that can be traced throughout vertebrate phylogeny (reviewed in Zapata and Amemiya, 2000). In mammals, spleen development has been shown to be critically dependent on the homeobox gene Hox-11 (Roberts et al., 1994). During zebrafish development, Tom Look’s group first identified the putative splenic primordium by WISH using the Hox-11/Tlx-1 probe (Langenau et al., 2002). Figure 5 shows developmental expression of Hox-11/Tlx-1 in wild-type zebrafish larvae. We first detected evidence of Hox-11 as an asymmetric focus of expression on the left anterior gut, in the region of the putative spleen primordium on day 4 pf. Over the next several days, signal intensity increased in this area. Crosssections confirm the identity of this focus of expression with an organ on the left side of the larva in proximity to the gut (Fig. 5, T. Palomero, D. Langenau, A. Ferrando, J. Kanki and Tom Look, personal communication). In adult zebrafish the spleen is a highly cellular organ. Danilova and Steiner found predominantly erythrocytes in the adult spleen and no significant expression of either Rag-1 or Ig (Danilova and Steiner, 2002). Figure 7 shows a FACS profile of cells populating the spleen and confirms the erythroid predominance of the teleostian spleen. However, there is a subpopulation of lymphocytes, which can be distinguished by light scatter characteristics (see Section IV.A). These cells presumably represent mature lymphocytes, which do not rearrange their receptors for antigen (TCR and BCR, respectively), explaining the absence of Rag-expression, which we also confirmed in Rag-2GFP transgenics (Fig. 7C). Germinal centers are not found in teleost spleen and class switching does not occur due to the absence of isotypes other than IgM and an IgD equivalent (see Section III.C.1). However, evidence for somatic hypermutation is found in teleost B cells (reviewed in Flajnik, 2002). We have therefore conducted a search for a zebrafish homolog of AID. A stretch of sequence homology to human AID was found in the zebrafish database Sanger Center. Further examination of the genetic region surrounding the sequence homology revealed the entire ZF AID gene encompassing 5700 bp. Several lines of evidence support the identity of zebrafish AID as the true ortholog of human AID. There is a high degree of sequence homology between zebrafish AID with human (60% identity at the protein level) and mouse AID (64% identity at the protein level). Compared with human AID,


289

FIG. 5. Expression of Hox-11 in Zebrafish Larvae. A–D. Hox-11 expression in wild-type zebrafish. WISH analysis reveals Hox-11 in pharyngeal endoderm (asterisks) in side view on d4 (A) and 6 (C), brain in dorsal view on d6 (D), and in the left anterior gut (arrows). E. Cross-section of d4 wild-type zebrafish larva after WISH. Black arrowhead indicates Hox-11 expression on left side in proximity of gut (G). YC ¼ yolk cell, NC ¼ notochord, NT ¼ neural tube. (Cross-section courtesy of T. Palomero, D. Langenau, A. Ferrando, J. Kanki, T. Look.)

zebrafish AID has conserved intron-exon boundaries for exons 1–5. Expression studies are under way. We were able to amplify a correctly spliced AID transcript from day 2 embryos. This implies that zebrafish AID has a role beyond B cell antigen receptor maturation, as B cells are first detected in the zebrafish pancreas at 4 dpf (Danilova and Steiner, 2002) and

290

D. TRAVER ET AL.

predicts that AID may also be found to be expressed in more primitive vertebrates, where it could fulfill a role in DNA metabolism. The gut is a site of lymphoid activity in zebrafish, although no distinct secondary lymphoid structures, such as Peyer’s patches, can be identified histologically. Thus, Ig and Rag-1 positive cells were detected in the lamina propria of the straight part of the intestine (Danilova and Steiner, 2002) reminiscent of Rag-1 expression detected in the activated B cells that populate murine Peyer’s patch germinal centers (Han et al., 1996). Interestingly, Danilova and Steiner report that patches co-expressing Ig and Rag-1 in the zebrafish gut resembled Peyer’s patches, previously not reported in lower vertebrates (Danilova and Steiner, 2002).

IV. Phenotypic Characterization of Zebrafish Hematolymphoid Cells

Mutagenesis screens in zebrafish have led to the discovery of a wide array of mutants that fail to correctly develop embryonic blood cells. Due to the early time points analyzed, the vast majority of these mutants show defects in the maturation of primitive erythrocytes. As discussed above, definitive, multilineage hematopoiesis does not occur until several days post fertilization in larval development, making simple visualization of nonerythroid mutants difficult. Current screens aimed to uncover deficiencies in definitive hematopoiesis have thus relied upon in situ-based approaches for both the lymphoid (Trede and Zon, 1998; Schorpp et al., 2000; Trede et al., 2001) and myeloid lineages (Bennett et al., 2001; Lyons et al., 2001a,b; Lieschke et al., 2001, 2002). These screens have yielded mutants that fail to specify early larval lymphoid and myeloid populations. It remains to be determined whether these mutants also show defects in adult blood cell production in the kidney, the teleost equivalent of mammalian bone marrow. Additionally, since primitive erythrocytes are thought to derive from different populations of hematopoietic stem cells (HSCs) than their adult, definitive counterparts, it remains to be determined whether the embryonic blood mutants also show defects in producing adult red blood cells. To this end, we have undertaken a thorough characterization of the adult zebrafish hematopoietic system. A. LINEAGE SEPARATION

BY

LIGHT SCATTER CHARACTERISTICS

Blood production in adult zebrafish, like other teleosts, occurs in the kidney, which supports both renal functions and multilineage hematopoiesis (zebrafish blood cell types are depicted in Fig. 6). It appears that all blood cell lineages develop here from HSCs, with the exception of mature T lymphocytes, which


291

FIG. 6. Model of definitive hematopoiesis in adult zebrafish. Cell types shown are the actual cells found in the zebrafish kidney. May-Grunwald/ Giemsa stain.

292

D. TRAVER ET AL.

are educated in the thymus. To assess the parameters of steady state definitive hematopoiesis, we have recently performed differential cells counts in adult hematolymphoid tissues, including the kidney, spleen, and blood (D. T. and L. I. Z., unpublished). Additionally, we have analyzed these tissues by flow cytometry. Examination of kidney marrow, spleen, and blood by light scatter characteristics reveals distinct profiles for each (Fig. 7A). Surprisingly, when compared to mammalian bone marrow scatter characteristics, there exist several discrete ‘‘scatter’’ populations in the zebrafish kidney. From these profiles, the major blood lineages can be isolated to purity from each tissue by cell sorting (Fig. 7B). Mature erythroid cells are found exclusively within a forward scatter (FSC)low fraction, myelomonocytic cells within only a FSChigh, side scatter (SSC)high population, lymphoid cells within a FSCint SSClow subset, and immature precursors within a FSChigh SSCint subset. Percentages of cells within each scatter population match those obtained by morphological cell counts, demonstrating that this flow cytometric assay is accurate in measuring the relative percentages of each of the major blood lineages. This new technique can thus be used to precisely identify and quantitate defects in definitive blood mutants. In this way, we have analyzed the zebrafish ‘‘embryonic’’ blood mutants. Since many of the primitive mutants are embryonic lethal when homozygous, we have analyzed heterozygous carriers of one mutant allele in adult fish. We find that several mutants show haploinsufficiency as evidenced by aberrant kidney erythropoiesis (D. T. and L. I. Z., unpublished). In general, mutant phenotypes of adult heterozygotes appear intermediate to those of their wild-type siblings and embryonic homozygous mutants. This suggests that many of the gene functions required to make embryonic erythrocytes are similarly required in their adult counterparts. It will be interesting to similarly assess additional lymphoid- and myeloid-specific blood mutants arising from ongoing mutagenesis screens. B. USE

OF

GFP TRANSGENES

TO

DEFINE ZEBRAFISH BLOOD LINEAGES

To complement the mutant analysis described above and to test autonomy of mutant gene function, we have developed zebrafish hematopoietic cell transplantation (HCT). As a donor cell marker for transplants, we have made use of adult zebrafish expressing GFP under control of the erythroid-specific GATA-1 promoter. Since erythroid cells are short-lived, they are produced continuously from kidney precursor populations. Continued production of GFP þ erythrocytes from transplanted kidney cells thus serves as a surrogate marker of donor-derived hematopoietic stem and progenitor cell activity. One potential complication of transplantation experiments in zebrafish is immune rejection by the host, since it has not been possible to create inbred strains of genetically identical individuals. To lessen this risk, we have chosen 2-day-old


293

FIG. 7. Separation of definitive blood lineages by flow cytometry. A. Single-cell suspensions of adult kidney cells form distinct populations when analyzed by size (forward scatter; FSC) and granularity (side scatter; SSC). B. Sorting of each population reveals that cells within the red gate are comprised of only mature erythrocytes (upper left panel), that the blue gate contains only lymphocytes (lower left panel), that the purple gate contains immature precursors of all mature blood lineages (lower right panel), and that the green gate contains only myelomonocytic cells (upper right panel). C. Scatter profiles can also be used to identify and quantitate each lineage in the adult spleen (panel 1), and in the peripheral blood (panel 2). Panel 3 shows the kidney profile of a transgenic zebrafish expressing GFP by the lymphoid-specific RAG-2 promoter. Approximately 30% of the cells within the lymphoid scatter population are GFP-positive (panel 4), whereas all other populations are uniformly negative (not shown). All numbers shown are relative percentages of each respective tissue.

embryos as transplant recipients. Lymphocyte development, as judged by in situ expression patterns of Rag1/2 genes as well as T and B antigen receptor genes, does not begin until 4 dpf (Willett et al., 1997; Trede and Zon, 1998; Willett et al., 1999; Trede et al., 2001). Transplantation before this stage may thus tolerize the host to the graft. Forty-eight hour embryos

294

D. TRAVER ET AL.

are also transparent, allowing transplant success to be easily visualized by circulating GFP þ cells. Transplantation of whole kidney cells from GATA-1eGFP adults into mutant vlad tepes embryos, for example, rescues multilineage hematopoiesis for at least six months in these animals that normally die by two weeks of age from erythropoietic failure (D. T. and L. I. Z., unpublished). This demonstrates that vlad tepes is required in a cell autonomous manner and strongly suggests that long-term survival of reconstituted individuals is mediated by stable hematopoietic stem cell engraftment. The ability to isolate the cells of each of the major hematopoietic lineages has facilitated the development of novel reagents to better understand blood development in zebrafish. We have generated cDNA libraries from highly purified myeloid, lymphoid, and precursor cell populations. We are currently sequencing at least 10,000 clones from each library to identify both novel genes and zebrafish orthologues of known mammalian genes. In an approach pioneered by the Thisse lab in zebrafish (Donovan et al., 2002), interesting clones will be used to generate in situ probes in order to perform high-throughput expression screens over early embryonic and larval development to identify genes expressed in lymphoid, myeloid, or erythroid cell subsets. Genes with interesting expression patterns can then be tested functionally in gain- or loss-of-function experiments by RNA or morpholino injections, respectively. The relative ease of testing functionality in zebrafish is an important advantage over similar approaches in mammals. Transgenic technology is now commonly used in zebrafish, and many investigators have made a variety of fluorescent reporter lines to mark cell lineages of interest. Lin and colleagues (Long et al., 1997) reported the first blood-specific transgenic line in which GFP was driven by the erythroid-specific GATA-1 promoter. Another GFP line marking early lymphocytes has been created using the zebrafish Rag-2 promoter (Jessen et al., 2001; and D. Langenau and T. Look, unpublished). Analysis of adult kidney marrow shows that the Rag-2 promoter is active only within the lymphoid scatter population (Fig. 7C, panels 3, 4), and is expressed in approximately 29% of the cells within this fraction. This expression is presumably in early B lymphocytes in the process of rearranging their antigen receptor genes since the adult kidney has been shown to be the site of B cell production in other teleosts (Daggfeldt et al., 1993; Hansen and Zapata, 1998; Zapata and Amemiya, 2000). The Rag-2 promoter is also active in the zebrafish thymus in developing T lymphocytes, but expression is absent in the adult spleen. We have also developed a transgenic line in which GFP is targeted to thrombocytes, zebrafish platelet equivalents, by the CD41 promoter (H.-F. Lin, D. T., C. Abraham, L. I. Z., and R. I. Handin, unpublished). These


295

transgenic animals are useful not only in tracking in vivo a population of interest but can also be used to create lineage-specific cDNA libraries in order to identify lineage-specific genes, expression patterns, and rapid functional testing as discussed above. These lines also afford the possibility of identifying populations enriched for hematopoietic stem and progenitor cells by flow cytometry and transplantation assays. C. IMAGING TECHNIQUES The use of labeled cells also facilitates following specific cell populations in optically transparent zebrafish in vivo. Thus, recent developments in the field of intravital microscopy (IVM) have led to important insights into the function and kinetics of the immune system. Using fluorescent dyes or antibodies, cells can be readily tracked in vivo. While the applicability of this approach is limited in mammals, it can be a powerful tool to follow circulation and homing of cells in translucent animals, such as the zebrafish. Pigmentdeficient strains as well as PTU-treated embryos are convenient subjects for studies. Even in wild-type fish, labeled cells can be easily analyzed in superficial vessels or in the fins (data not shown). Given the paucity of reagents such as specific antibodies, we utilized the dye CFSE to label harvested cells from spleen and kidney in combination with Fluorescence Microscopy. We decided to use CFSE because of its low toxicity, its uncomplicated use and the excellent staining properties for leukocytes. Particularly in murine lymphocytes, this marker has proven to be an excellent tool for subsequent FACS analysis or IVM and cell tracking for up to six months, as well as for cell division studies (Lyons, 1999; Parish, 1999). Toxicity or specific alterations of cell functions have not been described at recommended concentrations. Prior to injection into recipient fish, labeled cells can be FACS-sorted and divided into different fractions (Fig. 8A). Thus, labeled populations of lymphocytes or myeloid cells can be injected separately into recipient fish. Migration patterns as well as homing to superficial wounds can then be observed in the living animal by UV fluoroscopy (Fig. 8B, C). In necropsies, the final destination of homing cells can be analyzed. The jelly-like consistency of most internal organs and their translucence also facilitate optical studies (Fig. 8D, E). In our preliminary studies, we used CFSE to label harvested spleen and kidney cells of zebrafish. On average, a combined 1–2 million cells of hematopoietic origin can be obtained per sacrificed animal. Due to lack of appropriate culture conditions for zebrafish cells, we immediately inject the labeled cells into recipient fish either intraperitoneally or intracardially. These approaches now enable us to address questions of homing of leukocytes to

296

D. TRAVER ET AL.

wounds or sites of infection (see Fig. 8B, C and Section II.B.1.d). Allogeneic recipients can be used to study graft rejection as well as GvHD after irradiation of the host. We tested if a GFP signal emanating from internal organs could be detected in live zebrafish. For this purpose we chose a transgenic line, which expresses GFP under the control of the insulin promoter. Consistently, a signal could be observed close to the midline anterior to the anal fins even at low magnification (Fig. 8F). To ascertain the origin of the signal, the fish in Fig. 8F was sacrificed and internal organs were visualized after opening the abdominal cavity. The signal originated from discrete foci in an organ that by macroscopic inspection and position relative to adjacent structures (gut) was identified as the pancreas (Fig. 8G). The ability to visualize GFP signals from internal organs in living adult zebrafish will be a great asset for future immunology screens in adult zebrafish.

V. The Zebrafish as a Vertebrate Model System for Forward Genetic Screens

A. GENETIC SCREENS ARE A POWERFUL METHOD FOR IDENTIFYING MULTIPLE MUTATIONS WITHIN A GENETIC PATHWAY Forward genetic screening methods are used to identify mutations in novel and known genes within a specific genetic pathway. Large-scale genetic screens in non-vertebrate organisms, such as yeast, worms, plants, and flies have identified hundreds of genes and loci involved in a plethora of biological pathways including the cell division cycle, axon guidance, and embryology, to name a mere few. Selected mouse mutants, dominant screens in mice, and mutations associated with human genetic traits and disease have provided a wealth of knowledge about specific genes. More recently, genetic screens in the zebrafish have revealed hundreds of loci essential for early vertebrate development (Driever et al., 1996; Haffter et al., 1996; Amsterdam et al., 1999; Golling et al., 2002). Many mutations can be grouped in specific phenotypic classes, suggesting that mutant gene products may function within the same genetic pathway. Importantly, a number of mutations reside in genes that are highly conserved between fish and humans, and often zebrafish mutant phenotypes resemble human disease conditions (Driever and Fishman, 1996; Dooley and Zon, 2000). With both the zebrafish and human genome sequencing projects completed, determining the function of the predicted 40–60,000 genes in the human genome can be approached by analyzing the effects of mutations in model organisms, such as zebrafish. Zebrafish characteristics and life cycle facilitate genetic screening (Streisinger et al., 1981; Patton and Zon, 2001; Grunwald and Eisen, 2002).


297

298 D. TRAVER ET AL.

FIG. 8. Intravital microscopy. A. Light scatter characteristics of kidney and spleen cells. After labeling of whole kidney and spleen cells with CFSE, fractions were separately analyzed according to their light scatter characteristics. While red blood cells—although nucleated—show only little staining, myelomonocytic and lymphocytic cells are highly fluorescent. The background fluorescence of unstained cells is

Advances in Immunology Volume 81

Advances in Immunology Volume 81

Advances in Agronomy, Volume 81

Advances in Immunology Volume 45

Advances in Immunology Volume 34

Advances in Immunology Volume 1

Advances in Immunology Volume 10

Advances in Immunology Volume 30

Advances in Immunology Volume 82

Advances in Immunology Volume 31

Advances in Immunology Volume 52

Advances in Immunology Volume 48

Advances in Immunology Volume 56

Advances in Immunology Volume 12

Advances in Immunology Volume 21

Advances in Immunology Volume 41

Advances in Immunology Volume 71

Advances in Immunology Volume 84

Advances in Immunology Volume 43

Advances in Immunology Volume 73

Advances in Immunology, Volume 61

Advances in Immunology Volume 14

Advances in Immunology Volume 46

Advances in Immunology, Volume 106

Advances in Immunology Volume 19

Advances in Immunology Volume 99

Advances in Immunology Volume 4

Advances in Immunology Volume 80

Advances in Immunology Volume 64

Advances in Immunology Volume 62

Advances in Immunology Volume 81