Paul - Fundamental Immunology 5 [PDF] | Online Book Share (2024)

2,990 Pages • 973,428 Words • PDF • 14.1 MB

+ Paul+ Fundamental+ immunology

Uploaded at 2021-07-02 12:49

Report DMCA


Fundamental Immunology 5th edition (August 2003): by William E., Md. Paul (Editor) By Lippincott Williams & Wilkins Publishers;

By OkDoKeY

Fundamental Immunology CONTENTS Editors Contributors Dedication Acknowledgements Preface Quotes

Introduction Immunoglobulins and B Lymphocytes T Cells & NK Cells Organization and Evolution of the Immune System Antigen Processing and Presentation Regulation of the Immune Response Effector Mechanisms of Immunity Immunity to Infectious Agents Immunologic Mechanisms in Disease



Chapter 1 The Immune System: An Introduction William E. Paul Chapter 2 History of Immunology Pauline M. H. Mazumdar

Immunoglobulins and B Lymphocytes Chapter 3 Immunoglobulins: Structure and Function Grant R. Kolar and J. Donald Capra


Chapter 4 Antigen–Antibody Interactions and Monoclonal Antibodies Jay A. Berzofsky, Ira J. Berkower, and Suzanne L. Epstein Chapter 5 Immunoglobulins: Molecular Genetics Edward E. Max Chapter 6 B-Lymphocyte Development and Biology Richard R. Hardy Chapter 7 B-Cell Signaling Mechanisms and Activation Michael McHeyzer-Williams

T Cells & NK Cells


Chapter 8 T-Cell Antigen Receptors Mark M. Davis and Yueh-Hsiu Chien Chapter 9 T-Cell Developmental Biology Ellen V. Rothenberg, Mary A. Yui, and Janice C. Telfer Chapter 10 Peripheral T-Lymphocyte Responses and Function Marc K. Jenkins Chapter 11 T-Lymphocyte Activation Arthur Weiss and Lawrence E. Samelson Chapter 12 Natural Killer Cells David H. Raulet Chapter 13 Accessory Molecules and Co-Stimulation Arlene H. Sharpe, Yvette Latchman, and Rebecca J. Greenwald

Organization and Evolution of the Immune System Chapter 14 Lymphoid Tissues and Organs David D. Chaplin Chapter 15 Dendritic Cells Muriel Moser Chapter 16 Macrophages and the Immune Response Siamon Gordon Chapter 17 The Innate Immune System Ruslan Medzhitov


Chapter 18 Evolution of the Immune System Martin F. Flajnik, Kristina Miller, and Louis Du Pasquier

Antigen Processing and Presentation


Chapter 19 The Major Histocompatibility Complex and Its Encoded Proteins David H. Margulies and James McCluskey Chapter 20 The Biochemistry and Cell Biology of Antigen Processing Peter Cresswell

Regulation of the Immune Response Chapter 21 Immunogenicity and Antigen Structure Jay A. Berzofsky and Ira J. Berkower Chapter 22 Fc Receptors Jeffrey V. Ravetch Chapter 23 Type I Cytokines and Interferons and Their Receptors Warren J. Leonard Chapter 24 The Tumor Necrosis Factor Superfamily and Its Receptors Lyle L. Moldawer Chapter 25 Interleukin-1 Family of Ligands and Receptors Charles A. Dinarello Chapter 26 Chemokines Philip M. Murphy Chapter 27 Programmed Cell Death Francis Ka-Ming Chan and Michael J. Lenardo Chapter 28 Immunological Memory David F. Tough and Jonathan Sprent Chapter 29 Immunological Tolerance Ronald H. Schwartz and Daniel L. Mueller Chapter 30 Regulatory/Suppressor T Cells Ethan M. Shevach Chapter 31 The Mucosal Immune System Jiri Mestecky, Richard S. Blumberg, Hiroshi Kiyono, and Jerry R. McGhee


Chapter 32 Neural Immune Interactions in Health and Disease Esther M. Sternberg and Jeanette I. Webster Chapter 33 Immunology of Aging Dan L. Longo

Effector Mechanisms of Immunity


Chapter 34 Complement Wolfgang M. Prodinger, Reinhard Würzner, Heribert Stoiber, and Manfred P. Dierich Chapter 35 Phagocytosis Eric J. Brown and Hattie D. Gresham Chapter 36 Cytotoxic T-Lymphocytes Pierre A. Henkart and Michail V. Sitkovsky Chapter 37 Inflammation Helene F. Rosenberg and John I. Gallin

Immunity to Infectious Agents


Chapter 38 The Immune Response to Parasites Alan Sher, Thomas A. Wynn, and David L. Sacks Chapter 39 Viral Immunology Hildegund C. J. Ertl Chapter 40 Immunity to Intracellular Bacteria Stefan H. E. Kaufmann Chapter 41 Immunity to Extracellular Bacteria Moon H. Nahm, Michael A. Apicella, and David E. Briles Chapter 42 Immunology of HIV Infection Mark Dybul, Mark Connors, and Anthony S. Fauci Chapter 43 Vaccines G. J.V. Nossal

Immunologic Mechanisms in Disease


Chapter 44 Systemic Autoimmunity Philip L. Cohen Chapter 45 Organ-Specific Autoimmunity Matthias G. von Herrath and Dirk Homann Chapter 46 Immunological Mechanisms of Allergic Disorders Marsha Wills-Karp and Gurjit K. Khurana Hershey Chapter 47 Transplantation Immunology Megan Sykes, Hugh Auchincloss, Jr., and David H. Sachs Chapter 48 Tumor Immunology Hans Schreiber Chapter 49 Primary Immunodeficiency Diseases Rebecca H. Buckley Chapter 50 Immunotherapy Ellen S. Vitetta, Elaine Coleman, Maria-Ana Ghetie, Victor Ghetie, Jaroslav Michálek, Laurentiu M. Pop, Joan E. Smallshaw, and Camelia Spiridon

EDITORS Editor WILLIAM E. PAUL, M.D. JAMES MERRITT Acquisitions Editor JULIA SETO Developmental Editor STEVEN P. MARTIN Production Editor COLIN J. WARNOCK Manufacturing Manager CHRISTINE JENNY Cover Designer

Contributors Michael A. Apicella, M.D. Professor and Head, Department of Microbiology, The University of Iowa College of Medicine, Iowa City, IA Hugh Auchincloss, Jr., M.D. Professor of Surgery (Immunology), Harvard Medical School, Department of Surgery, Boston, MA Ira J. Berkower, MD, Ph.D. Chief, Laboratory of Immunogregulation, Office of Vaccines, Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD Jay A. Berzofsky, MD, Ph.D. Chief, Molecular Immunogenetics and Vaccine Research Section, Metabolism Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD Richard S. Blumberg, M.D. Associate Professor of Medicine, Brigham and Women's Hospital, Boston, MA David E. Briles, Ph.D. Professor, Department of Microbiology, University of Alabama at Birmingham, Birmingham, AL Eric J. Brown, M.D. Program in Host-Pathogen Interactions, University of California, San Francisco, San Francisco, CA Rebecca H. Buckley, M.D. Professor, Departments of Pediatrics and Immunology, Duke University Durham, NC J. Donald Capra, M.D. President and Program Head, Department of Molecular Immunogenetics, Oklahoma Medical Research Foundation, Oklahoma City, OK Francis Ka-Ming Chan, Ph.D. Assistant Professor, Department of Pathology, University of Massachusetts Medical School, Worcester, MA David D. Chaplin, M.D., Ph.D. Professor and Chairman, Department of Microbiology University of Alabama at

Birmingham, Birmingham, AL Yueh-Hsiu Chen, Ph.D. Professor, Department of Microbiology and Immunology, Stanford University, Stanford, CA Philip L. Cohen, M.D. Professor, Department of Medicine, University of Pennsylvania, Philadelphia, PA Elaine Coleman, B.S. Graduate Student, Cancer Immunobiology Center University of Texas Southwestern Medical Center, Dallas, TX Mark Connors, M.D. Senior Investigator, Laboratory of Immunoregulation, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Peter Cresswell, Ph.D. Section of Immunobiology, Howard Hughes Medical Institute Yale University School of Medicine, New Haven, CT Mark M. Davis, Ph.D. Professor, Microbiology and Immunology, Stanford University School of Medicine, Stanford CA Manfred P. Dierich, M.D. Professor and Chairman, Institute of Hygiene, Medical University, Innsbruck, Austria Charles A. Dinarello, M.D. Professor of Medicine, Department of Infectious Disease, University of Colorado Health Sciences Center, Denver, CO Mark Dybul, M.D. Assistant Director for Medical Affairs, National Institute of Allergy and Infectious Diseases, The National Institutes of Health, Bethesda, MD Louis E. du Pasquier, Ph.D. Professor, Department of Zoology, University of Basel, Basel, Switzerland Suzanne L. Epstein, Ph.D. Chief, Laboratory of Immunology and Developmental Biology, Division of Cellular and Gene Therapies, Center for Biologics Evaluation and Research, Food and Drug Administration, Rockville, MD Hildegund C. J. Ertl, M.D.

Professor, The Wistar Institute, Philadelphia, PA Anthony S. Fauci, M.D. Director, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Martin F. Flajnik, Ph.D. Professor, Department of Microbiology and Immunology, University of Maryland at Baltimore, Baltimore, MD John I. Gallin, M.D. Laboratory of Host Defenses, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Maria-Ana Ghetie, Ph.D. Assistant Professor, Cancer Immunobiology Center, UT Southwestern Medical Center, Dallas, TX Victor F. Ghetie, Ph.D. Professor, Cancer Immunobiology Center, UT Southwestern Medical Center, Dallas, TX Rebecca J. Greenwald, Ph.D. Instructor, Department of Pathology, Harvard Medical School, Boston, MA Hattie D. Gresham, Ph.D. Associate Professor, Department of Molecular Genetics and Microbiology, University of New Mexico, Albuquerque, NM Siamon Gordon, M.D. Sir William Dunn School of Pathology, University of Oxford, Oxford, England Richard R. Hardy, Ph.D. Institute for Cancer Research, Fox Chase Cancer Center, Philadelphia, PA Pierre A. Henkart, M.D. Experimental Immunology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD Gurjit K. Khurana Hershey, M.D., Ph.D. Assistant Professor, Department of Pediatrics, University of Cincinnati Hospital, Cincinnati, OH Dirk Homann, M.D. Department of Neuropharmacology, The Scripps Research Institute, La Jolla, CA

Marc K. Jenkins, Ph.D. Professor, Department of Microbiology, Center for Immunology, University of Minnesota, Minneapolis, MN Stefan H. E. Kaufmann, Ph.D. Director, Department of Immunology, Max-Planck-Institute for Infection Biology, Berlin, Germany Hiroshi Kiyono, D.D.S., Ph.D. Professor and Director, Division of Mucosal Immunology, The University of Tokyo, Minato-ku, Tokyo Grant R. Kolar, B.S. Graduate Student, Program in Molecular Immunogenetics, Oklahoma Medical Research Foundation, Oklahoma City, OK Yvette Latchman, Ph.D. Department of Pathology, Brigham and Women's Hospital and Harvard Medical School, Boston, MA Michael J. Lenardo, M.D. Chief, Molecular Development Section, Laboratory of Immunology, DIR, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Warren J. Leonard, M.D. Laboratory of Molecular Immunology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD Dan L. Longo, M.D. Scientific Director, National Institute on Aging, National Institutes of Health, Baltimore, MD David H. Margulies, M.D., Ph.D. Chief, Molecular Biology Section, Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Edward Ellis Max, M.D., Ph.D. Associate Director for Research, Office of Therapeutics Research and Review, Center for Biologics Evaluation And Research, Food and Drug Administration, Bethesda, MD Pauline M. H. Mazumdar, M.B., B.S., Ph.D. Department of History of Science and Technology, University of Toronto, Toronto, Ontario James McCluskey, M.D.

Department of Microbiology and Immunology, The University of Melbourne, Victoria, Australia Jerry R. McGhee, Ph.D. Professor and Director, Department of Microbiology/Immunobiology Vaccine Center, University of Alabama at Birmingham, Birmingham, AL Michael G. McHeyzer-Williams, M.D. Associate Professor, Department of Immunology, The Scripps Research Institute, La Jolla, CA Ruslan Medzhitov, Ph.D. Assistant Professor, Department of Immunobiology, Yale University School of Medicine, New Haven, CT Jiri Mestecky, M.D., Ph.D. Professor, Department of Microbiology, University of Alabama at Birmingham, Birmingham, AL Jaroslav Michálek, M.D., Ph.D. Cancer Immunobiology Center, UT Southwestern Medical Center, Dallas, TX Kristina Miller, Ph.D. Research Scientist, Department of Molecular Genetics, Pacific Biological Station, Fisheries and Oceans, Canada, Nanaimo, BC, Canada Lyle Moldawer, M.D. Department of Surgery, College of Medicine, University of Florida, Shands Hospital, Gainesville, FL Muriel Moser, Ph.D. Senior Research Assocaite, Institut de Biologie et Médecine Moléculaires, Université Libre de Bruxelles, Gosselies, Belgium Daniel L. Mueller, M.D. Professor, Department of Medicine, University of Minnesota Medical School, Minneapolis, MN Philip M. Murphy, M.D. Chief, Molecular Signaling Section, Laboratory of Host Defenses, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Moon H. Nahm, M.D. Professor, Departments of Pathology, University of Alabama at Birmingham, Birmingham, AL

G.J.V. Nossal, M.D., Ph.D. Professor Emeritus, Department of Pathology, The University of Melbourne, Australia William E. Paul, M.D. Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, Bethesda, MD Laurentiu M. Pop, M.D. Postdoctoral Fellow, Cancer Immunobiology Center, University of Texas Southwestern Medical Center, Dallas, TX Wolfgang M. Prodinger, M.D. Associate Professor, Institute for Hygiene and Social Medicine, University of Innsbruck, Innsbruck, Austria David H. Raulet, Ph.D. Choh Hao Li Professor and Head, Division of Immunology, Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, CA Jeffrey V. Ravetch, M.D., Ph.D. Theresa and Eugene M. Lang Professor and Head, Laboratory of Molecular Genetics and Immunology, The Rockefeller University, New York, NY Helene F. Rosenberg, M.D., Ph.D. Chief, Eosinophil Pathophysiology Section, Laboratory of Host Defenses, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland Ellen V. Rothenberg, M.D. Division of Biology 156-29, California Institute of Technology, Pasadena, CA David H. Sachs, M.D. Professor of Surgery (Immunology), Harvard Medical School, and Director, Transplantation Biology Research Center, Massachusetts General Hospital, Boston, Massachusetts David L. Sacks, Ph.D. Head, Intracellular Parasite Biology Section, Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Lawrence Samelson, M.D. Chief, Laboratory of Cellular and Molecular Biology, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD Ronald H. Schwartz, M.D., Ph.D.

Chief, Laboratory of Cellular and Molecular Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Hans Schreiber, M.D., Ph.D. Department of Pathology, University of Chicago, IL Arlene H. Sharpe, M.D., Ph.D. Associate Professor, Department of Pathology, Harvard Medical School, Boston, MA Alan Sher, Ph.D. Acting Chief, Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, MD Ethan M. Shevach, M.D. Chief, Cellular Immunology Section, Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Michail V. Sitkovsky, Ph.D. Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD Joan E. Smallshaw, Ph.D. Postdoctoral Researcher, Cancer Immunobiology Center, University of Texas Southwestern Medical Center, Dallas, TX Camelia I. Spiridon, M.D. Postdoctoral Researcher, Cancer Immunobiology Center, University of Texas Southwestern Medical Center, Dallas, TX Jonathan Sprent, M.D., Ph.D. Department of Immunology, IMM4, The Scripps Research Institute, La Jolla, CA Esther M. Sternberg, M.D. Director, Integrative Neural-Immune Program, National Institute of Mental Health, National Institutes of Health, Bethesda, MD Heribert Stoiber, M.D. Associate Professor, Institute of Hygiene, University of Innsbruck, Innsbruck, Austria Megan Sykes, M.D. Professor of Surgery and Medicine, Department of Immunology, Harvard Medical School, Boston, MA Janice C. Telfer, Ph.D. Assistant Professor, Department of Veterinary and Animal Sciences, University of

Massachusetts Amherst, Amherst, MA David F. Tough, Ph.D. Senior Group Leader, The Edward Jenner Institute for Vaccine Research, Newbury, Berkshire, United Kingdom Ellen S. Vitetta, Ph.D. Director, Cancer Immunobiology Center, University of Texas Southwestern Medical Center, Dallas, TX Matthias G. von Herrath, M.D. Associate Professor, Department of Developmental Immunology, La Jolla Institute for Allergy and Immunology, San Diego, CA Jeanette I. Webster, Ph.D. Research Fellow, Section of Neuroendocrine Immunology and Behavior, National Institute of Mental Health, National Institutes of Health, Bethesda, MD Arthur Weiss, M.D., Ph.D. Ephraim P. Engleman Distinguished Professor, Investigator, Howard Hughes Medical Institute, Department of Medicine, University of California, San Francisco, San Francisco, CA Marsha Wills-Karp, Ph.D. Professor and Director, Division of Immunobiology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH Reinhard Würzner, M.D., Ph.D. Professor, Institut für Hygiene und Sozialmedizin, University of Innsbruck, Innsbruck, Austria Thomas A. Wynn, Ph.D. Senior Investigator, Immunopathogenesis Section, National Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, MD Mary A. Yui, Ph.D. Postdoctoral Scholar, Division of Biology, California Institute of Technology, Pasadena, CA

Dedication To Jacob, Sylvie, Julien and Jenna

Acknowledgements The preparation of the Fifth Edition of Fundamental Immunology required the efforts of many individuals. I particularly wish to thank each of the authors. Their contributions, prepared in the midst of extremely busy schedules, are responsible for the value of this book. Julia Seto of Lippincott Williams & Wilkins saw that the process of receiving, editing and assembling the chapters went as smoothly as possible; without her work, the completion of the edition would have been immeasurably more difficult. Steven Martin saw that the complex process of preparing and revising proofs was done with admirable efficiency. I wish to gratefully acknowledge the efforts of each of the members of the editorial and production staffs of Lippincott Williams and Wilkins who participated in the preparation of this edition.

Preface The fifth edition of Fundamental Immunology appears when the importance of the immune response in human health and the prevention of disease was never clearer. Bio-terroism is a world-wide threat, with the possibility that one of the greatest achievements of mankind, the elimination of small pox, may be undone. The HIV pandemic shows no signs of abating and exacts an increasingly frightening toll. Tuberculosis and malaria continue to be major scourges of mankind. The number of infants and children that annually succumb to diarrheal infectious diseases is in the millions. The true impact of autoimmunity is more fully appreciated than ever and we now recognize that inflammation plays a major role in many diseases, not the least of which is atherosclerosis. Childhood asthma and allergies have become a virtual epidemic, particularly in certain parts of the western world. The great promise of transplantation will only be fulfilled when we can induce specific tolerance and avoid the need for long-lasting immunosuppression. The possibility that the immune response can become a major modality for cancer therapy still remains to be determined. These challenges demand a redoubled effort to more fully understand the basis of the immune response and to learn how it can be mobilized or inhibited. Innnovative approaches for the development of new vaccines are needed. A new generation of immunologists will be required to grapple with these issues. Fundamental Immunology and its sister publications play a key role in training those entering our field and in helping current immunologists to be as productive as possible. Fundamental Immunology was first published in 1984; I began to work on it in late 1982. The Fifth Edition thus marks more than 20 years during which I have had the privilege of participating in the preparation of this book. My goal was, and continues to be, to make available to advanced students of immunology and to post-doctoral fellows in immunology and related fields an authoritative treatment of the major areas of immunology. Fundamental Immunology is also designed to provide my colleagues with a simple way to keep current in aspects of immunology outside their immediate area of expertise and to allow scientists in allied fields to rapidly inform themselves of the state of the art to aid them in aspects of their work that impinge on immunology. In agreeing to take responsibility for editing an advanced text in immunology, I was motivated, in part, by my experience as a post-doctoral fellow working on the binding properties of antibodies when I made almost daily use of Kabat and Mayer's Experimental Immunochemistry. I hoped that Fundamental Immunology might serve a similar role for a new generation of immunologists. The degree to which I have succeeded must be judged by the readers. What I failed to anticipate was the unremitting growth of our science. Indeed, immunology has been in a state of continuing revolution throughout my entire career.

Fundamental Immunology, which was 809 pages in its first version, has more than doubled in size and a field that seemed almost too broad to be encompassed in a single volume in 1984 is now far broader. I continue to be impressed with the vibrancy of immunology and with the upwelling of new subjects that gain center stage. Indeed, in the period since the Fourth Edition, virtually every area of immunology has seen major progress. Innate immunity and regulatory T cells, topics that had languished for years, have become the “hottest” of hot subjects. Of course, these are not new areas; the study of innate immunity and the inflammatory response have been central to our discipline since the 19th century. The competing ideas championed by Metchnikoff and by Ehrlich have always been in the minds of immunologists. Nonetheless, the thrust of innate immunity into the forefront of immunological science has been truly remarkable. Similarly, the re-emergence of the study of immunological suppression, with its new name, and the recognition of the central role that regulatory (suppressor) T cells play in control of autoimmunity has been nothing short of spectacular. Fundamental Immunology has changed just as our field has changed. New chapters have been added to represent disciplines that have come to the fore and previous chapters have been dropped, with the material in them reassigned to other chapters. For the Fifth Edition, the previous organizational structure has been generally retained. The opening section, Introduction, provides an overview of contemporary immunology and a portrayal of the history of our field, prepared by a distinguished historian of immunology, so that those with a limited background in the field can productively read the subsequent chapters. The next three sections, Immunoglobulins and B Lymphocytes, T Cells and NK Cells, and Organization and Evolution of the Immune System, introduce the principal cellular components of the immune system and the context in which they act. Emphasizing the centrality of antigen-presentation and of major histocompatibility molecules in the process of T cell recognition of antigen, I have added the section Antigen Processing and Presentation. The book then considers the Regulation of the Immune Response, with 13 individual chapters detailing the critical aspects of this process. Among these are four separate chapters on the central regulatory molecules of the immune system, the cytokines. I then turn to consider how the immune system mediates its functions, deals with infectious agents, and participates in and may prevent or ameliorate a wide range of diseases. The chapters dealing with this are found in the sections Effector Mechanisms of Immunity, Immunity to Infectious Agents (a new section with five chapters) and Immunologic Mechanisms in Disease. In the Preface to each of the previous editions, I reminded readers that Fundamental Immunology grapples with the most current of immunological subjects. In many areas, consensus may not yet have been reached. Each chapter has been written by a leader in the field, but inevitably there will be disagreement among them on certain issues. Rather than striving for an agreement where none yet exists, I ask the reader to take note of the differences and reach their own judgments in these contentious areas. I welcome comments by readers of Fundamental Immunology for ways to improve the book and to increase its value. Such suggestions will be seriously considered in the

preparation of subsequent editions. William E. Paul Bethesda Maryland

Quotes From my teachers I have learned much, from my colleagues still more, but from my students most of all. The Talmud Discovery consists of seeing what everybody has seen and thinking what nobody has thought. Albert Szent-Gyorgyi …the clonal selection hypothesis…[SC]assumes that…[SC]there exist clones of mesenchymal cells, each carrying immunologically reactive sites…complementary…[SC]to one (or possibly a small number of) potential antigenic determinants. Sir Macfarlane Burnet The Clonal Selection Theory of Acquired Immunity In the fields of observation, chance favors only the mind that is prepared. Louis Pasteur Address at the University of Lille In all things of nature there is something of the marvelous. Aristotle Parts of Animals

Chapter 1 The Immune System: An Introduction Fundamental Immunology

Chapter 1 William Paul

Introduction The Immune System: An Introduction

KEY CHARACTERISTICS OF THE IMMUNE SYSTEM Innate Immunity (Chapter 17) Primary Responses (Chapter 6, Chapter 10, and Chapter 14) Secondary Responses and Immunologic Memory (Chapter 6, Chapter 10, Chapter 14, and Chapter 28) The Immune Response Is Highly Specific and the Antigenic Universe Is Vast The Immune System Is Tolerant of Self-Antigens (Chapter 29) Immune Responses Against Self-Antigens Can Result in Autoimmune Diseases (Chapter 44 and Chapter 45) AIDS Is an Example of a Disease Caused by a Virus that the Immune System Generally Fails to Eliminate (Chapter 42) Major Principles of Immunity CELLS OF THE IMMUNE SYSTEM AND THEIR SPECIFIC RECEPTORS AND PRODUCTS B-LYMPHOCYTES AND ANTIBODY B-Lymphocyte Development (Chapter 6) B-Lymphocyte Activation (Chapter 7) B-Lymphocyte Differentiation (Chapter 5, Chapter 7, and Chapter 28) B1 or CD5+ B-Lymphocytes (Chapter 6) B-Lymphocyte Tolerance (Chapter 29) Immunoglobulin Structure (Chapter 3) Immunoglobulin Genetics (Chapter 5) Class Switching (Chapter 5) Affinity Maturation and Somatic Hypermutation (Chapter 5) T-LYMPHOCYTES T-Lymphocyte Antigen Recognition (Chapter 8, Chapter 19, and Chapter 20) T-Lymphocyte Receptors (Chapter 8) T-Lymphocyte Activation (Chapter 11) T-Lymphocyte Development (Chapter 9) T-Lymphocyte Functions (Chapter 10) T Cells That Help Antibody Responses (Chapter 10) Induction of Cellular Immunity (Chapter 10) Regulatory T Cells (Chapter 30) Cytotoxic T Cells (Chapter 36) CYTOKINES (Chapter 23, Chapter 24, Chapter 25, and Chapter 26) Chemokines (Chapter 26) THE MAJOR HISTOCOMPATIBILITY COMPLEX AND ANTIGEN PRESENTATION (Chapter 19 and Chapter 20) Class I MHC Molecules (Chapter 19) Class II MHC Molecules (Chapter 19) Antigen Presentation (Chapter 20) T-Lymphocyte Recognition of Peptide/MHC Complexes Results in MHC-Restricted Recognition (Chapter 8) Antigen-Presenting Cells (Chapter 15) EFFECTOR MECHANISMS OF IMMUNITY

Effector Cells of the Immune Response Monocytes and Macrophages (Chapter 16) Natural Killer Cells (Chapter 12) Mast Cells and Basophils (Chapter 46) Granulocytes (Chapter 37) Eosinophils (Chapter 38 and Chapter 46) The Complement System (Chapter 34) The Classical Pathway of Complement Activation The Alternative Pathway of Complement Activation The Terminal Components of the Complement System CONCLUSION

The immune system is a remarkable defense mechanism. It provides the means to make rapid, specific, and protective responses against the myriad potentially pathogenic microorganisms that inhabit the world in which we live. The tragic example of severe immunodeficiencies, as seen in both genetically determined diseases and in acquired immunodeficiency syndrome (AIDS), graphically illustrates the central role the immune response plays in protection against microbial infection. The immune system also has a role in the rejection of tumors and may exert important effects in regulating other bodily systems, but most immunologists would agree that the evolutionary pressure that has principally shaped the immune system is the challenge to vertebrates of the microbial world. Fundamental Immunology has as its goal the authoritative presentation of the basic elements of the immune system, of the means through which the mechanisms of immunity act in a wide range of clinical conditions, including recovery from infectious diseases, rejection of tumors, transplantation of tissue and organs, autoimmune and other immunopathologic conditions, and allergy; and how the mechanisms of immunity can be martialed by vaccination to provide protection against microbial pathogens. The purpose of this opening chapter is to provide readers with a general introduction to our current understanding of the immune system. It will thus be of particular importance for those with a limited background in immunology, providing them with the preparation needed for subsequent chapters of the book. Indeed, rather than providing extensive references in this chapter, each of the subject headings will indicate the chapters that deal in detail with the topic under discussion. Those chapters will not only provide an extended treatment of the topic but will also furnish the reader with a comprehensive reference list.

KEY CHARACTERISTICS OF THE IMMUNE SYSTEM Innate Immunity ( Chapter 17) Most pathogenic microorganisms attempting to infect an individual encounter powerful nonspecific defenses. The epithelium provides both a physical barrier to the entry of

microbes and produces a variety of antimicrobial factors. Microbes that penetrate the epithelium are met with macrophages and related cells that have receptors for cell-surface molecules found on many microbial agents. These interactions may lead to phagocytosis of the pathogen, activation of the macrophage so that it can destroy the agent and to the induction of an inflammatory response that recruits other cell types, including neutrophils, to the site. Microbial pathogens may also be recognized by components of the complement system leading to the enhanced phagocytosis of the agent and in some instances to its lysis as well as to independent activation of inflammatory responses. The innate immune system also acts to recruit antigen-specific immune responses, not only by attracting cells of the immune system to the site of the infection, but also through the uptake of antigen by dendritic cells that transport antigen to lymphoid tissue where primary immune responses are initiated. Dendritic cells also produce cytokines that can regulate the quality of the immune response so that it is most appropriate to combating the pathogen. Primary Responses ( Chapter 6, Chapter 10, and Chapter 14) Primary immune responses are initiated when a foreign antigenic substance interacts with antigen-specific lymphocytes under appropriate circ*mstances. The response generally consists of the production of antibody molecules specific for the antigenic determinants of the immunogen and of the expansion and differentiation of antigen-specific helper and effector T-lymphocytes. The latter include cells that produce cytokines and killer T cells, capable of lysing infected cells. Generally, the combination of the innate immune response and the primary response are sufficient to eradicate or to control the microbe. Indeed, the most effective function of the immune system is to mount a response that eliminates the infectious agent from the body. Secondary Responses and Immunologic Memory ( Chapter 6, Chapter 10, Chapter 14, and Chapter 28) As a consequence of the initial encounter with antigen, the immunized individual develops a state of immunologic memory. If the same (or a closely related) microorganism is encountered again, a secondary response is made. This generally consists of an antibody response that is more rapid, greater in magnitude, and composed of antibodies that bind to the antigen with greater affinity and are more effective in clearing the microbe from the body. A more rapid and more effective T-cell response also ensues. One effect is that an initial infection with a microorganism initiates a state of immunity in which the individual is protected against a second infection. In the majority of situations, protection is provided by high-affinity antibody molecules that rapidly clear the re-introduced microbe. This is the basis of vaccination; the great power of vaccines is illustrated by the elimination of smallpox from the world and by the complete control of polio in the Western Hemisphere. The Immune Response Is Highly Specific and the Antigenic Universe Is Vast

The immune response is highly specific. Primary immunization with a given microorganism evokes antibodies and T cells that are specific for the antigenic determinants found on that microorganism but that fail to recognize (or recognize only poorly) antigenic determinants expressed by unrelated microbes. Indeed, the range of antigenic specificities that can be discriminated by the immune system is enormous. The Immune System Is Tolerant of Self-Antigens ( Chapter 29) One of the most important features of the immune system is its ability to discriminate between antigenic determinants expressed on foreign substances, such as pathogenic microbes, and potential antigenic determinants expressed by the tissues of the host. The capacity of the system to ignore host antigens is an active process involving the elimination or inactivation of cells that could recognize self-antigens through a process designated immunologic tolerance. Immune Responses Against Self-Antigens Can Result in Autoimmune Diseases ( Chapter 44 and Chapter 45) Failures in establishing immunologic tolerance or unusual presentations of self-antigens can give rise to tissue-damaging immune responses directed against antigenic determinants on host molecules. These can result in autoimmune diseases. It is now recognized that a range of extremely important diseases are caused by autoimmune responses or have major autoimmune components, including systemic lupus erythematosus, rheumatoid arthritis, insulin-dependent diabetes mellitus, multiple sclerosis, myasthenia gravis, and regional enteritis. Efforts to treat these diseases by modulating the autoimmune response are a major theme of contemporary medicine. AIDS Is an Example of a Disease Caused by a Virus that the Immune System Generally Fails to Eliminate ( Chapter 42) Immune responses against infectious agents do not always lead to elimination of the pathogen. In some instances, a chronic infection ensues in which the immune system adopts a variety of strategies to limit damage caused by the organism or by the immune response. One of the most notable infectious diseases in which the immune response generally fails to eliminate the organism is AIDS, caused by the human immunodeficiency virus (HIV). In this instance, the principal infected cells are those of the immune system itself, leading to an eventual state in which the individual can no longer mount protective immune responses against other microbial pathogens. Major Principles of Immunity The major principles of the immune response are: Elimination of many microbial agents through the nonspecific protective mechanisms of the innate immune system Highly specific recognition of foreign antigens coupled with potent mechanisms for

elimination of microbes bearing such antigens A vast universe of distinct antigenic specificities and a comparably vast capacity for the recognition of these antigens The capacity of the system to display immunologic memory Tolerance of self-antigens The remainder of this introductory chapter will describe briefly the molecular and cellular basis of the system and how these central characteristics of the immune response may be explained.

CELLS OF THE IMMUNE SYSTEM AND THEIR SPECIFIC RECEPTORS AND PRODUCTS The immune system consists of a wide range of distinct cell types, each with important roles. The lymphocytes occupy central stage because they are the cells that determine the specificity of immunity. It is their response that orchestrates the effector limbs of the immune system. Cells that interact with lymphocytes play critical parts both in the presentation of antigen and in the mediation of immunologic functions. These cells include dendritic cells, and the closely related Langerhans cells, monocyte/macrophages, natural killer (NK) cells, neutrophils, mast cells, basophils, and eosinophils. In addition, a series of specialized epithelial and stromal cells provide the anatomic environment in which immunity occurs, often by secreting critical factors that regulate migration, growth, and/or gene activation in cells of the immune system. Such cells also play direct roles in the induction and effector phases of the response. The cells of the immune system are found in peripheral organized tissues, such as the spleen, lymph nodes, Peyer’s patches of the intestine, and tonsils, where primary immune responses generally occur (see Chapter 14). A substantial portion of the lymphocytes and macrophages comprise a re-circulating pool of cells found in the blood and lymph, as well as in the lymph nodes and spleen, providing the means to deliver immunocompetent cells to sites where they are needed and to allow immunity that is initiated locally to become generalized. Activated lymphocytes acquire the capacity to enter nonlymphoid tissues where they can express effector functions and eradicate local infections. Some memory lymphocytes are “on patrol” in the tissues, scanning for reintroduction of their specific antigens. Lymphocytes are also found in the central lymphoid organs, thymus, and bone marrow, where they undergo the developmental steps that equip them to mediate the responses of the mature immune system. Individual lymphocytes are specialized in that they are committed to respond to a limited set of structurally related antigens. This commitment exists before the first contact of the immune system with a given antigen. It is expressed by the presence on the lymphocyte’s surface membrane of receptors specific for determinants (epitopes) of the antigen. Each lymphocyte possesses a population of receptors, all of which have identical combining sites. One set, or clone, of lymphocytes differs from another clone in the structure of the combining region of its receptors and thus in the epitopes that it can recognize. The ability of an organism to respond to virtually any non-self antigen is achieved by the existence of a very large number of different lymphocytes, each bearing receptors specific for a distinct epitope. As a consequence, lymphocytes are an

enormously heterogeneous group of cells. Based on reasonable assumptions as to the range of diversity that can be created in the genes encoding antigen-specific receptors, it seems virtually certain that the number of distinct combining sites on lymphocyte receptors of an adult human can be measured in the millions. Lymphocytes differ from each other not only in the specificity of their receptors but also in their functions. There are two broad classes of lymphocytes: the B-lymphocytes, which are precursors of antibody-secreting cells, and the T- (thymus-derived) lymphocytes. T-lymphocytes express important helper functions, such as the ability to aid in the development of specific types of immune responses, including the production of antibody by B cells and the increase in the microbicidal activity of macrophages. Other T-lymphocytes are involved in direct effector functions, such as the lysis of virus-infected cells or certain neoplastic cells. Specialized T-lymphocytes (regulatory T cells) have the capacity to suppress specific immune responses.

B-LYMPHOCYTES AND ANTIBODY B-Lymphocyte Development ( Chapter 6) B-lymphocytes derive from hematopoietic stem cells by a complex set of differentiation events ( Fig. 1). A detailed picture has been obtained of the molecular mechanisms through which committed early members of the B lineage develop into mature B-lymphocytes. These events occur in the fetal liver and, in adult life, principally in the bone marrow. Interaction with specialized stromal cells and their products, including cytokines such as interleukin (IL)-7, are critical to the normal regulation of this process.

FIG. 1. The patterns of gene expression, timing of gene rearrangement events, capacity for self-replenishment and for rapid proliferation of developing B lymphocytes are indicated. Adapted from Hardy RR, Hayakawa K, B cell development pathways, Annu Rev Immunol 2001,19:595–621, with permission.

The key events in B-cell development occur in cells designated pro-B cells and pre-B cells. They center about the assembly of the genetic elements encoding the

antigen-specific receptors of B cells, which are immunoglobulin (Ig) molecules specialized for expression on the cell surface. Igs are heterodimeric molecules consisting of heavy (H) and light (L) chains, both of which have regions (variable [V] regions) that contribute to the binding of antigen and that differ in sequence from one Ig molecule to another (see Chapter 3) ( Fig. 2). In addition, H and L chains contain regions that are nonvariable or constant (C regions).

FIG. 2. A schematic representation of an Ig molecule indicating the means through which the V regions and the CH1 and CL regions of H and L chains pair with one another and how the CH2 and CH3 regions of the H chains pair.

The genetic elements encoding the variable portions of Ig H and L chains are not contiguous in germline DNA or in the DNA of nonlymphoid cells (see Chapter 5) ( Fig. 3). In pro- and pre-B cells, these genetic elements are translocated to construct an expressible V-region gene. This process involves a choice among a large set of potentially usable variable (V), diversity (D), and joining (J) elements in a combinatorial manner. Such combinatorial translocation, together with a related set of events that add diversity in the course of the joining process, results in the generation of a very large number of distinct H and L chains. The pairing of H and L chains in a quasi-random manner further expands the number of distinct Ig molecules that can be formed.

FIG. 3. Organization and translocation of mouse IgH genes. IgH chains are encoded by four distinct genetic elements: Igh-V (V), Igh-D (D), Igh-J (J), and Igh-C. The V, D, and J genetic elements together specify the variable region of the H chain. The Igh-C element specifies the C region. The same V region can be expressed in association with each of the C regions (µ, d, ?3, ?1, ?2ß, ?2a, e, and a). In the germline, the V, D, and J genes are far apart and there are multiple forms of each of these genes. In the course of lymphocyte development, a VDJ gene complex is formed by translocation of individual V and D genes so that they lie next to one of the J genes, with excision of the intervening genes. This VDJ complex is initially expressed with µ and d C genes, but may be subsequently translocated so that it lies near one of the other C genes (e.g. ?1) and in that case leads to the expression of a VDJ ?1 chain.

The H-chain variable region is initially expressed in association with the product of the µ constant (C)-region gene. Together these elements encode the µ IgH chain, which is used in Igs of the IgM class. The successful completion of the process of Ig gene rearrangement and the expression of the resultant IgM on the cell surface marks the transition between the pre-B– and B–cell states ( Fig. 1). The newly differentiated B cell initially expresses surface Ig solely of the IgM class. The cell completes its maturation process by expressing on its surface a second class of Ig composed of the same L chain and the same H chain variable (VDJ) region but of a different H-chain C region; this second Ig H chain is designated d, and the Ig to which it contributes is designated IgD. The differentiation process is controlled at several steps by a system of checks that determines whether prior steps have been successfully completed. These checks depend on the expression on the surface of the cell of appropriately constructed Ig or Ig-like molecules. For, example, in the period after a µ chain has been successfully assembled but before an L chain has been assembled, the µ chain is expressed on the cell surface in association with a surrogate light chain, consisting of VpreB and ?5. Pre-B cells that fail to express this µ/VpreB ?5 complex do not move forward to future differentiation states or do so very inefficiently. B-Lymphocyte Activation ( Chapter 7) A mature B cell can be activated by an encounter with an antigen expressing epitopes that are recognized by its cell-surface Ig ( Fig. 4). The activation process may be a direct one, dependent on cross-linkage of membrane Ig molecules by the antigen ( cross-linkage–dependent B-cell activation), or an indirect one, occurring most efficiently in the context of an intimate interaction with a helper T cell, in a process often referred to as cognate help.

FIG. 4. Two forms of B-cell activation. A: Cognate T-cell/B-cell help. Resting B cells can bind antigens that bear epitopes complementary to their cell-surface Ig. Even if the antigen cannot cross-link the receptor, it will be endocytosed and enter late endosomes and lysosomes where it will be degraded to peptides. Some of these peptides will be loaded into class II MHC molecules and brought to the cell surface, where they can be recognized by CD4+ T cells that bear receptors specific for that peptide/class II complex. This interaction allows an activation ligand on the T cells (CD40 ligand) to bind to its receptor on B cells (CD40) and to signal B-cell activation. In addition, the T cells secrete several cytokines that regulate the growth and differentiation of the stimulated B cell. B: Cross-lineage–dependent B-cell activation. When B cells encounter antigens that bear multiple copies of an epitope that can bind to their surface Ig, the resultant cross-linkage stimulates biochemical signals within the cell leading to B-cell activation, growth, and differentiation. In many instances, B-cell activation events may result from both pathways of stimulation.

Because each B cell bears membrane Ig molecules with identical variable regions, cross-linkage of the cell-surface receptors requires that the antigen express more than one copy of an epitope complementary to the binding site of the receptor. This requirement is fulfilled by antigens with repetitive epitopes. Among these antigens are the capsular polysaccharides of many medically important microorganisms such as pneumococci, streptococci, and meningococci. Similar expression of multiple identical epitopes on a single immunogenic particle is a property of many viruses because they express multiple copies of envelope proteins on their surface. Cross-linkage–dependent B-cell activation is a major protective immune response mounted against these microbes. The binding of complement components (see Chapter 34) to antigen or antigen–antibody complexes can increase the magnitude of the cross-linkage–dependent B-cell activation due to the action of a receptor for complement, which, together with other molecules, increases the magnitude of a B-cell

response to limiting amounts of antigen. Cognate help allows B cells to mount responses against antigens that cannot cross-link receptors and, at the same time, provides co-stimulatory signals that rescue B cells from inactivation when they are stimulated by weak cross-linkage events. Cognate help is dependent on the binding of antigen by the B cell’s membrane Ig, the endocytosis of the antigen, and its fragmentation into peptides within the endosomal/lysosomal compartment of the cell. Some of the resultant peptides are loaded into a groove in a specialized set of cell-surface proteins, the class II major histocompatibility complex (MHC) molecules ( Fig. 5). The resultant class II/peptide complexes are expressed on the cell surface. As will be discussed below, these complexes are the ligands for the antigen-specific receptors of a set of T cells designated CD4+ T cells. CD4+ T cells that have receptors specific for the class II/peptide complex expressed on the B-cell surface recognize and interact with that B cell. That interaction results in the activation of the B cell through the agency of cell-surface molecules expressed by the T cells (e.g., the CD40 ligand [CD154]) and cytokines produced by the T cell ( Fig. 4). The role of the B-cell receptor for antigen is to create the T-cell ligand on the surface of antigen-specific B cells; activation of the B cell derives largely from the action of the T cell. However, in many physiologic situations, receptor cross-linkage stimuli and cognate help synergize to yield more vigorous B-cell responses.

FIG. 5. Illustration of the structure of the peptide-binding domain (a1 and ß1) of a class II MHC molecule (HLA-DR; protein data bank designation 1DLH) bound to an antigenic peptide from influenza hemagglutinin. Adapted by D.H. Margulies from Stern LJ et al., Crystal structure of the human class II MHC protein HLA-DR1 complexed with an influenza virus peptide, Nature 1994;368:215–221, with permission.

B-Lymphocyte Differentiation ( Chapter 5, Chapter 7, and Chapter 28) Activation of B cells prepares them to divide and to differentiate either into antibody-secreting cells or into memory cells, so that there are more cells specific for the antigen used for immunization and these cells have new properties. Those cells that differentiate into antibody secreting cells account for primary antibody responses. Some

of these antibody secreting cells migrate to the bone marrow where they may continue to produce antibody for an extended period of time and may have lifetimes in excess of 1 year. Memory B cells give rise to antibody-secreting cells upon re-challenge of the individual. The hallmark of the antibody response to re-challenge (a secondary response) is that it is of greater magnitude, occurs more promptly, is composed of antibodies with higher affinity for the antigen, and is dominated by Igs expressing ?, a, or e C regions (IgG, IgA, or IgE) rather than by IgM, which is the dominant Ig of the primary response. Division and differentiation of cells into antibody-secreting cells is largely controlled by the interaction of the activated B cells with T cells expressing CD154 and by their stimulation by T-cell–derived cytokines. The differentiation of activated B cells into memory cells occurs in a specialized micro-environmental structure in the spleen and lymph nodes, the germinal center. The process through which increases in antibody affinity occurs also takes place within the germinal center. The latter process, designated affinity maturation, is dependent on somatic hypermutation. The survival of cells within the germinal center depends on the capacity to bind antigen so that as antigen availability diminishes, cells that have higher affinity receptors, either naturally or as a result of the hypermutation process, have a selective survival and growth advantage. Thus, such cells come to dominate the population. The process through which a single H-chain V region can become expressed with genes encoding C regions other than µ and d is referred to as Ig class switching. It is dependent on a gene translocation event through which the C-region genes between the genetic elements encoding the V region and the newly expressed C gene are excised, resulting in the switched C gene being located in the position that the Cµ gene formerly occupied ( Fig. 3). This process also occurs in germinal centers. B1 or CD5+ B-Lymphocytes ( Chapter 6) A second population of B cells (B1 cells) has been described that differs from the dominant B-cell population (sometimes designated B2 cells or conventional B cells) in several important respects. These cells were initially recognized because some express a cell-surface protein, CD5, not generally found on other B cells. In the adult mouse, B1 B cells are found in relatively high frequency in the peritoneal cavity but are present at low frequency in the spleen and lymph nodes. B1 B cells are quite numerous in fetal and perinatal life. Whether B1 B cells derive from a separate set of stem cells found in the fetal liver but absent from (or present only at low frequency in) the adult bone marrow is still a matter of controversy. The alternative view is that B1 B cells are derived from conventional B cells as a result of cross-linkage–dependent B-cell activation. B1 B cells appear to be self-renewing, in contrast to conventional B cells, in which division and memory are

antigen driven. B1 B cells appear to be responsible for the secretion of the serum IgM that exists in nonimmunized mice, often referred to as natural IgM. Among the antibodies found in such “natural” IgM are molecules that can combine with phosphatidyl choline (a component of pneumococcal cell walls) and for lipopolysaccharide and influenza virus. B1 B cells also produce autoantibodies, although they are generally of low affinity and in most cases not pathogenic. It is believed that B1 B cells are important in resistance to several pathogens and may have a significant role in mucosal immunity. B-Lymphocyte Tolerance ( Chapter 29) One of the central problems facing the immune system is that of being able to mount highly effective immune responses to the antigens of foreign, potentially pathogenic, agents while ignoring antigens associated with the host’s own tissues. The mechanisms ensuring this failure to respond to self-antigens are complex and involve a series of strategies. Chief among them is elimination of cells capable of self-reactivity or the inactivation of such cells. The encounter of immature, naive B cells with antigens with repetitive epitopes capable of cross-linking membrane Ig can lead to elimination of the B cells, particularly if no T-cell help is provided at the time of the encounter. This elimination of potentially self-reactive cells is often referred to as clonal elimination. Some self-reactive cells, rather than dying upon encounter with self-antigens, may re-express the proteins needed for immunoglobulin gene rearrangement and undergo a further round of such rearrangement. This process, referred to as receptor editing, allows a self-reactive cell to substitute a new receptor and therefore to avoid elimination. There are many self-antigens that are not encountered by the developing B-cell population or that do not have the capacity to cross-link B-cell receptors to a sufficient degree to elicit the clonal elimination/receptor editing process. Such cells, even when mature, may nonetheless be inactivated through a process that involves cross-linkage of receptors without the receipt of critical co-stimulatory signals. These inactivated cells may be retained in the body but are unresponsive to antigen and are referred to as anergic. When removed from the presence of the anergy-inducing stimulus, such cells may regain responsiveness. Immunoglobulin Structure ( Chapter 3) The antigen-specific membrane receptors and secreted products of B cells are Ig molecules. Igs are members of a large family of proteins designated the immunoglobulin supergene family. Members of the Ig supergene family have sequence hom*ology, a common gene organization, and similarities in three-dimensional structure. The latter is characterized by a structural element referred to as the Ig fold, generally consisting of a set of seven ß-pleated sheets organized into two apposing layers ( Fig. 6). Many of the cell-surface proteins that participate in immunologic recognition processes, including the T-cell receptor (TCR), the CD3 complex, and molecules associated with the B-cell receptor (Iga and Igß), are members of the Ig supergene family.

FIG. 6. Schematic drawing of the V and C domains of an Ig L chain illustrating the “Ig fold.” The ß strands participating in the antiparallel ß-pleated sheets of each domain are represented as arrows. The ß strands of the three-stranded sheets are shaded, whereas those in the four-stranded sheets are white. The intradomain disulfide bonds are represented as black bars. Selected amino acids are numbered with position 1 as the N terminus. From Edmundson AB, Ely KR, Abola EE, et al., Rotational allomerism and divergent evolution of domains in immunoglobulin light chains, Biochemistry 1975;14:3953–3961, with permission.

The Igs themselves are constructed of a unit that consists of two H chains and two L chains ( Fig. 2). The H and L chains are composed of a series of domains, each consisting of approximately 110 amino acids. The L chains, of which there are two types (? and ?), consist of two domains. The carboxy-terminal domain is essentially identical among L chains of a given type and is referred to as the constant (C) region. As already discussed, the amino-terminal domain varies from L chain to L chain and contributes to the binding site of antibody. Because of its variability, it is referred to as the variable (V) region. The variability of this region is largely concentrated in three segments, designated as the hypervariable or complementarity-determining regions (CDRs). The CDRs contain the amino acids that are the L chain’s contribution to the lining of the antibody’s combining site. The three CDRs are interspersed among four regions of much lower degree of variability, designated framework regions (FRs). The H chains of Ig molecules are of several classes (µ, d, ? [of which there are several subclasses], a, and e), as noted above. An assembled Ig molecule, consisting of one or more units of two identical H and L chains, derives its name from the H chain that it possesses. Thus, there are IgM, IgD, IgG, IgA, and IgE antibodies. The H chains each consist of a single amino-terminal V region and three or four C regions. In many H chains, a hinge region separates the first and second C regions and conveys flexibility to the molecule, allowing the two combining sites of a single unit to move in relation to one another so as to promote the binding of a single antibody molecule to an antigen that has more than one copy of the same epitope. Such divalent binding to a single antigenic structure results in a great gain in energy of interaction (see Chapter 4). The H-chain V region, like that of the L chain, contains three CDRs lining the combining site of the antibody and four FRs.

The C region of each H-chain class conveys unique functional attributes to the antibodies that possess it. Among the distinct biologic functions of each class of antibody are the following: IgM antibodies are potent activators of the complement system ( Chapter 34). IgA antibodies are secreted into a variety of bodily fluids and are principally responsible for immunity at mucosal surfaces ( Chapter 31). IgE antibodies are bound by specific receptors (FceRI) on basophils and mast cells. When cross-linked by antigen, these IgE/FceRI complexes cause the cells to release a set of mediators responsible for allergic inflammatory responses ( Chapter 46). IgD antibodies act virtually exclusively as membrane receptors for antigen. IgG antibodies, made up of four subclasses in both humans and mice, mediate a wide range of functions including transplacental passage and opsonization of antigens through binding of antigen–antibody complexes to specialized Fc receptors on macrophages and other cell types ( Chapter 22, Chapter 34, and Chapter 36). IgD, IgG, and IgE antibodies consist of a single unit of two H and L chains. IgM antibodies are constructed of five or six such units, although they consist of a single unit when they act as membrane receptors. IgA antibodies may consist of one or more units. The antibodies that are made up of more than a single unit generally contain an additional polypeptide chain, the J chain, which plays an important role in the ability of these polymeric immunoglobulins to be secreted at mucosal surfaces. Each of the distinct Igs can exist as secreted antibodies and as membrane molecules. Antibodies and cell-surface receptors of the same class made by a specific cell have identical structures except for differences in their carboxy-terminal regions. Membrane Ig possesses a hydrophobic region, spanning the membrane, and a short intracytoplasmic tail, both of which are lacking in the secretory form. Immunoglobulin Genetics ( Chapter 5) The genetic makeup of the Ig H-chain gene has already been alluded to. The IgH-chain gene of a mature lymphocyte is derived from a set of genetic elements that are separated from one another in the germline. The V region is composed of three types of genetic elements: V H, D, and J H. More than 100 V H elements exist; there are more than 10 D elements and a small number of J H elements (4 in the mouse). An H-chain V HDJ H gene is created by the translocation of one of the D elements on a given chromosome to one of the J H elements on that chromosome, generally with the excision of the intervening DNA. This is followed by a second translocation event in which one of the V H elements is brought into apposition with the assembled DJ H element to create the V HDJ H (V region) gene ( Fig. 3). Although it is likely that the choice of the V H, D, and J H elements that are assembled is not entirely random, the combinatorial process allows the creation of a very large number of distinct H-chain

V-region genes. Additional diversity is created by the imprecision of the joining events and by the deletion of nucleotides and addition of new, un-templated nucleotides between D and J H and between V H and D, forming N regions in these areas. This further increases the diversity of distinct IgH chains that can be generated from the relatively modest amount of genetic information present in the germline. The assembly of L-chain genes follows generally similar rules. However, L chains are assembled from V L and J L elements only. Although there is junctional diversity, no N regions exist for L chains. Additional diversity is provided by the existence of two classes of L chains, ? and ?. An Ig molecule is assembled by the pairing of IgH-chain polypeptide with an IgL-chain polypeptide. Although this process is almost certainly not completely random, it allows the formation of an exceedingly large number of distinct Ig molecules, the majority of which will have individual specificities. The rearrangement events that result in the assembly of expressible IgH and IgL chains occur in the course of B-cell development in pro-B cells and pre-B cells, respectively ( Fig. 1). This process is regulated by the Ig products of the rearrangement events. The formation of a µ chain signals the termination of rearrangement of H-chain gene elements and the onset of rearrangement of L-chain gene elements, with ? rearrangements generally preceding ? rearrangements. One important consequence of this is that only a single expressible µ chain will be produced in a given cell, since the first expressible µ chain shuts off the possibility of producing an expressible µ chain on the alternative chromosome. Comparable mechanisms exist to ensure that only one L-chain gene is produced, leading to the phenomenon known as allelic exclusion. Thus, the product of only one of the two alternative allelic regions at both the H- and L-chain loci are expressed. The closely related phenomenon of L-chain isotype exclusion ensures the production of either ? or ? chains in an individual cell, but not both. An obvious but critical consequence of allelic exclusion is that an individual B cell makes antibodies, all of which have identical H- and L-chain V regions, a central prediction of the clonal selection theory of the immune response. Class Switching ( Chapter 5) An individual B cell can continue to express the same IgH-chain V region but, as it matures, can switch the IgH-chain C region that it uses ( Fig. 3). Thus, a cell that expresses receptors of the IgM and IgD classes may differentiate into a cell that expresses IgG, IgA, or IgE receptors and then into a cell-secreting antibody of the same class as it expressed on the cell surface. This process allows the production of antibodies capable of mediating distinct biologic functions but that retain the same antigen-combining specificity. When linked with the process of affinity maturation of antibodies, Ig class switching provides antibodies of extremely high efficacy in preventing re-infection with microbial pathogens or in rapidly eliminating such pathogens. These two associated phenomena account for the high degree of effectiveness of antibodies produced in secondary immune responses.

The process of switching is known to involve a recombination event between specialized switch (S) regions, containing repetitive sequences, that are located upstream of each C region (with the exception of the d C region). Thus, the S region upstream of the µ C H region gene (Sµ) recombines with an S region upstream of a more 3’ isotype, such as S?1, to create a chimeric Sµ/S?1 region resulting in the deletion of the intervening DNA ( Fig. 7). The genes encoding the C regions of the various ? chains (in the human ?1, ?2, ?3, and ?4; in the mouse ?1, ?2a, ?2b, and ?3), of the a chain, and of the e chain are located 3’ of the Cµ and Cd genes.

FIG. 7. Ig class switching. Illustrated here is the process through which a given VDJ gene in a stimulated B cell may switch the C-region gene with which it is associated from µ to another, such as ?1. A recombination event occurs in which DNA between a cleavage point in Sµ and one in S?1 forms a circular episome. This results in C?1 being located immediately downstream of the chimeric Sµ/?1 region, in a position such that transcription initiating upstream of VDJ results in the formation of VDJC?1 mRNA and ?1 H-chain protein.

The induction of the switching process is dependent on the action of a specialized set of B-cell stimulants. Of these, the most widely studied are CD154, expressed on the surface of activated T cells, and bacterial lipopolysaccharide. The targeting of the C region that will be expressed as a result of switching is largely determined by cytokines. Thus, IL-4 determines that switch events in the human and mouse will be to the e C region and to the ?4 (human) or ?1 (mouse) C regions. In the mouse, interferon-gamma (IFN-?) determines switching to ?2a and transforming growth factor-beta (TGF-ß) determines switching to a. A major goal is to understand the physiologic determination of the specificity of the switching process. Because cytokines are often the key controllers of which Ig classes will represent the switched isotype, this logically translates into asking what regulates the relative amounts of particular cytokines that are produced by different modes of immunization. The switching process depends on the RNA-editing, enzyme activation–induced cytidine deaminase (AID). Mice that lack AID fail to undergo immunoglobulin class switching.

AID is also critical in the process of somatic hypermutation. Affinity Maturation and Somatic Hypermutation ( Chapter 5) The process of generation of diversity embodied in the construction of the H- and L-chain V-region genes and of the pairing of H and L chains creates a large number of distinct antibody molecules, each expressed in an individual B cell. This primary repertoire is sufficiently large so that most epitopes on foreign antigens will encounter B cells with complementary receptors. Thus, if adequate T-cell help can be generated, antibody responses can be made to a wide array of foreign substances. Nonetheless, the antibody that is initially produced usually has a relatively low affinity for the antigen. This is partially compensated for by the fact that IgM, the antibody initially made, is a pentamer. Through multivalent binding, high avidities can be achieved even if individual combining sites have only modest affinity (see Chapter 4). In the course of T-cell–dependent B-cell stimulation, particularly within the germinal center, a process of somatic hypermutation is initiated that leads to a large number of mutational events, largely confined to the H-chain and L-chain V-region genes and their immediately surrounding introns. During the process of somatic hypermutation, mutational rates of 1 per 1,000 base pairs per generation may be achieved. This implies that, with each cell division, close to one mutation will occur in either the H- or L-chain V region of an individual cell. This creates an enormous increase in antibody diversity. Although most of these mutations will either not affect the affinity with which the antibody binds its ligand or will lower that affinity, some will increase it. Thus, some B cells emerge that can bind antigen more avidly than the initial population of responding cells. Because there is an active process of apoptosis in the germinal center from which B cells can be rescued by the binding of antigen to their membrane receptors, cells with the most avid receptors should have an advantage over other antigen-specific B cells and should come to dominate the population of responding cells. Thus, upon re-challenge, the affinity of antibody produced will be greater than that in the initial response. As time after immunization elapses, the affinity of antibody produced will increase. This process leads to the presence in immunized individuals of high-affinity antibodies that are much more effective, on a weight basis, in protecting against microbial agents and other antigen-bearing pathogens than was the antibody initially produced. Together with antibody class switching, affinity maturation results in the increased effectiveness of antibody in preventing re-infection with agents with which the individual has had a prior encounter.

T-LYMPHOCYTES T-lymphocytes constitute the second major class of lymphocytes. They derive from precursors in hematopoietic tissue, undergo differentiation in the thymus (hence the name thymus-derived [T] lymphocytes), and are then seeded to the peripheral lymphoid tissue and to the recirculating pool of lymphocytes (see Chapter 14). T cells may be subdivided into two distinct classes based on the cell-surface receptors they express. The majority of T cells express antigen-binding receptors (TCRs) consisting of a and ß chains. A second group of T cells express receptors made up of ? and d chains. Among

the a/ß T cells are two important sublineages: those that express the co-receptor molecule CD4 (CD4+ T cells) and those that express CD8 (CD8+ T cells). These cells differ in how they recognize antigen and mediate different types of regulatory and effector functions. CD4+ T cells are the major helper cells of the immune system. Their helper function depends both on cell-surface molecules such as CD154, induced upon these cells when they are activated, and on the wide array of cytokines they secrete when activated. CD4+ T cells tend to differentiate, as a consequence of priming, into cells that principally secrete the cytokines IL-4, IL-13, IL-5, IL-6, and IL-10 (T H2 cells) or into cells that mainly produce IL-2, IFN-?, and lymphotoxin (T H1 cells). T H2 cells are very effective in helping B cells develop into antibody-producing cells, whereas T H1 cells are effective inducers of cellular immune responses, involving enhancement in the microbicidal activity of macrophages and consequent increased efficiency in lysing microorganisms in intracellular vesicular compartments. T cells also mediate important effector functions. Some of these are determined by the patterns of cytokines they secrete. These powerful molecules can be directly toxic to target cells and can mobilize potent inflammatory mechanisms. In addition, T cells, particularly CD8+ T cells, can develop into cytotoxic T-lymphocytes (CTLs) capable of efficiently lysing target cells that express antigens recognized by the CTLs. T-Lymphocyte Antigen Recognition ( Chapter 8, Chapter 19, and Chapter 20) T cells differ from B cells in their mechanism of antigen recognition. Immunoglobulin, the B-cell’s receptor, binds to individual antigenic epitopes on soluble molecules or on particulate surfaces. B-cell receptors recognize epitopes expressed on the surface of native molecules. Antibody and B-cell receptors evolved to bind to and to protect against microorganisms in extracellular fluids. By contrast, T cells invariably recognize cell-associated molecules and mediate their functions by interacting with and altering the behavior of these antigen-presenting cells (APCs). Indeed, the TCR does not recognize antigenic determinants on intact, undenatured molecules. Rather, it recognizes a complex consisting of a peptide, derived by proteolysis of the antigen, bound into a specialized groove of a class II or class I MHC protein. Indeed, what differentiates a CD4+ T cell from a CD8+ T cell is that the CD4+ T cells only recognize peptide/class II complexes, whereas the CD8+ T cells recognize peptide/class I complexes. The TCR’s ligand (i.e., the peptide/MHC protein complex) is created within the APC. In general, class II MHC molecules bind peptides derived from proteins that have been taken up by the APC through an endocytic process ( Fig. 8). These endocytosed proteins are fragmented by proteolytic enzymes within the endosomal/lysosomal compartment, and the resulting peptides are loaded into class II MHC molecules that traffic through this compartment. These peptide-loaded, class II molecules are then expressed on the surface of the cell where they are available to be bound by CD4+ T cells with TCRs capable of recognizing the expressed cell-surface complex. Thus, CD4+

T cells are specialized to largely react with antigens derived from extracellular sources.

FIG. 8. Pathways of antigen processing. Exogenous antigen (Ea) enters the cell via endocytosis and is transported from early endosomes into late endosome or prelysosomes, where it is fragmented and where resulting peptides (Ea-derived peptides) may be loaded into class II MHC molecules. The latter have been transported from the rough endoplasmic reticulum (RER) through the Golgi apparatus to the peptide-containing vesicles. Class II MHC molecules/Ea-derived peptide complexes are then transported to the cell surface, where they may be recognized by TCR expressed on CD4+ T cells. Cytoplasmic antigens (Ca) are degraded in the cytoplasm and then enter the RER through a peptide transporter. In the RER, Ca-derived peptides are loaded into class I MHC molecules that move through the Golgi apparatus into secretory vesicles and are then expressed on the cell surface where they may be recognized by CD8+ T cells. From Paul WE, Development and function of lymphocytes, in Gallin JI, Goldstein I, Snyderman R, eds. Inflammation, New York: Raven, 1992, 776, with permission.

In contrast, class I MHC molecules are mainly loaded with peptides derived from internally synthesized proteins, such as viral gene products. These peptides are produced from cytosolic proteins by proteolysis within the proteasome and are translocated into the rough endoplasmic reticulum. Such peptides, generally nine amino acids in length, are bound by class I MHC molecules. The complex is brought to the cell surface, where it can be recognized by CD8+ T cells expressing appropriate receptors. This property gives the T-cell system, particularly CD8+ T cells, the ability to detect cells expressing proteins that are different from, or produced in much larger amounts than, those of cells of the remainder of the organism (e.g., viral antigens [whether internal, envelope, or cell surface] or mutant antigens [such as active oncogene products]), even if these proteins, in their intact form, are neither expressed on the cell surface nor secreted. T-Lymphocyte Receptors ( Chapter 8)

The TCR is a disulfide-linked heterodimer ( Fig. 9). The constituent chains (a and ß, or ? and d) are members of the Ig supergene family. The TCR is associated with a set of transmembrane proteins, collectively designated the CD3 complex, that play a critical role in signal transduction. The CD3 complex consists of ?, d (note that the CD3 ? and d chains and the TCR ? and d chains are distinct polypeptides that, unfortunately, have similar designations), and e chains, and is associated with a hom*odimer of two ? chains or a heterodimer of ? and ? chains. CD3 ?, d, and e consist of extracellular domains that are family members of the Ig supergene. The cytosolic domains of CD3 ?, d, and e, and of ? and ?, contain one or more copies of a signaling motif–the immunoreceptor tyrosine-based activation motif (ITAM) (D/ExxYxxLxxxxxxxYxxL/I)–that is found in a variety of chains associated with immune recognition receptors. This motif appears to be very important in the signal transduction process and provides a site through which protein tyrosine kinases can interact with these chains to propagate signaling events.

FIG. 9. The T-cell antigen receptor. Illustrated schematically is the antigen-binding subunit comprised of an aß heterodimer, and the associated invariant CD3 and ? chains. Acidic (-) and basic (+) residues located within the plasma membrane are indicated. The open rectangular boxes indicate motifs within the cytoplasmic domains that interact with protein tyrosine kinases. (This figure also appears as in Chapter 11 as Fig. 2.)

The TCR chains are organized much like Ig chains. Their N-terminal portions are variable and their C-terminal portions are constant. Furthermore, similar recombinational mechanisms are used to assemble the V-region genes of the TCR chains. Thus, the V region of the TCR ß chain is encoded by a gene made of three distinct genetic elements (Vß, D, and Jß) that are separated in the germline. Although the relative numbers of Vß, D, and Jß genes differ from that for the comparable IgH variable-region elements, the strategies for creation of a very large number of distinct genes by combinatorial assembly are the same. Both junctional diversity and N-region addition further diversify the genes, and their encoded products. TCR ß has fewer V genes than IgH but much more diversity centered on the D/J region, which encodes the equivalent of the third

CDR of Igs. The a chain follows similar principles, except that it does not use a D gene. The genes for TCR ? and d chains are assembled in a similar manner except that they have many fewer V genes from which to choose. Indeed, ?/d T cells in certain environments, such as the skin and specific mucosal surfaces, are exceptionally hom*ogeneous. It has been suggested that the TCRs encoded by these essentially invariant ? and d chains may be specific for some antigen that signals microbial invasion and that activation of ?/d T cells through this mechanism constitutes an initial response that aids the development of the more sophisticated response of a/ß T cells. T-Lymphocyte Activation ( Chapter 11) T-cell activation is dependent on the interaction of the TCR/CD3 complex with its cognate ligand, a peptide bound in the groove of a class I or class II MHC molecule, on the surface of a competent antigen-presenting cell. Through the use of chimeric cell-surface molecules that possess cytosolic domains largely limited to the ITAM signaling motif alluded to above, it is clear that cross-linkage of molecules containing such domains can generate some of the signals that result from TCR engagement. Nonetheless, the molecular events set in motion by receptor engagement are complex ones. Among the earliest steps are the activation of tyrosine kinases leading to the tyrosine phosphorylation of a set of substrates that control several signaling pathways. Current evidence indicates that early events in this process involve the Src-family tyrosine kinases p56 lck, and p59 fyn, and ZAP-70, a Syk family tyrosine kinase, that binds to the phosphorylated ITAMs of the ? chain, as well as the action of the protein tyrosine phosphatase CD45, found on the surface of all T cells. A series of important substrates are tyrosine phosphorylated as a result of the action of the kinases associated with the TCR complex. These include (a) a set of adapter proteins that link the TCR to the Ras pathway; (b) phospholipase C?1, the tyrosine phosphorylation of which increases its catalytic activity and engages the inositol phospholipid metabolic pathway, leading to elevation of intracellular free-calcium concentration to the activation of protein, kinase C; and (c) a series of other important enzymes that control cellular growth and differentiation. Particularly important is the phosphorylation of LAT, a molecule that acts as an organizing scaffold to which a series of signaling intermediates bind and upon which they become activated and control downstream signaling. The recognition and early activation events result in the reorganization of cell surface and cytosolic molecules on the T cell, and correspondingly, on the APC to produce a structure, the immunological synapse. The apposition of key interacting molecules involving a small segment of the membranes of the two cells concentrates these molecules in a manner that both strengthens the interaction between the cells and intensifies the signaling events. It also creates a limited space into which cytokines may be secreted to influence the behavior of cells. Indeed, the formation of the immunological synapse is one mechanism through which the recognition of relatively small numbers of ligands by TCRs on a specific T cell can be converted into a vigorous stimulatory process.

In general, normal T cells and cloned T-cell lines that are stimulated only by TCR cross-linkage fail to give complete responses. TCR engagement by itself may often lead to a response in which the key T-cell–derived growth factor, IL-2, is not produced and in which the cells enter a state of anergy such that they are unresponsive or poorly responsive to a subsequent competent stimulus (see Chapter 29). Full responsiveness of a T cell requires, in addition to receptor engagement, an accessory-cell–delivered co-stimulatory activity. The engagement of CD28 on the T cell by CD80 and/or CD86 on the APC (or the engagement of comparable ligand receptor pairs on the two cells) provides a potent co-stimulatory activity. Inhibitors of this interaction markedly diminish antigen-specific T-cell activation in vivo and in vitro, indicating that the CD80/86–CD28 interaction is physiologically very important in T-cell activation (see Chapter 13). The interaction of CD80/86 with CD28 increases cytokine production by the responding T cells. For the production of IL-2, this increase appears to be mediated both by enhancing the transcription of the IL-2 gene and by stabilizing IL-2 mRNA. These dual consequences of the CD80/86–CD28 interaction cause a striking increase in the production of IL-2 by antigen-stimulated T cells. CD80/86 has a second receptor on the T cell, CTLA-4, that is expressed later in the course of T-cell activation. The bulk of evidence indicates that the engagement of CTLA-4 by CD80/86 leads to a set of biochemical signals that terminate the T-cell response. Mice that are deficient in CTLA-4 expression develop fulminant autoimmune responses. T-Lymphocyte Development ( Chapter 9) Upon entry into the thymus, T-cell precursors do not express TCR chains, the CD3 complex, or the CD4 or CD8 molecules ( Fig. 10). Because these cells lack both CD4 and CD8, they are often referred to as double-negative (DN) cells. Thymocytes develop from this DN3 pool into cells that are both CD4+ and CD8+ (double-positive cells) and express low levels of TCR and CD3 on their surface. In turn, double-positive cells further differentiate into relatively mature thymocytes that express either CD4 or CD8 (single-positive cells) and high levels of the TCR/CD3 complex.

FIG. 10. Development of a/ß T cells in the thymus. Double-negative T cells (4 -8 -) acquire CD4 and CD8 (4 +8 +) and then express a/ß TCRs, initially at low levels. Thereafter, the degree of expression of TCRs increases and the cells differentiate into CD4 or CD8 cells and are then exported to the periphery. Once the T cells have expressed receptors, their survival depends on the recognition of peptide/MHC class I or class II molecules with an affinity above some given threshold. Cells that fail to do so undergo apoptosis. These cells have failed to be positively selected. Positive selection is associated with the differentiation of 4 +8 + cells into CD4 or CD8 cells. Positive selection involving peptide/class I MHC molecules leads to the development of CD8 cells, whereas positive selection involving peptide/class II MHC molecules leads to the development of CD4 cells. If a T cell recognizes a peptide/MHC complex with high affinity, it is also eliminated via apoptosis (it is negatively selected).

The expression of the TCR depends on complex rearrangement processes that generate TCR a and ß (or ? and d) chains. Once expressed, these cells undergo two important selection processes within the thymus. One, termed negative selection, is the deletion of cells that express receptors that bind with high affinity to complexes of self-peptides with self-MHC molecules. This is a major mechanism through which the T-cell compartment develops immunologic unresponsiveness to self-antigens (see Chapter 9 and Chapter 29). In addition, a second major selection process is positive selection, in which T cells with receptors with “intermediate affinity” for self-peptides bound to self-MHC molecules are selected, thus forming the basis of the T-cell repertoire for foreign peptides associated with self-MHC molecules. It appears that T cells that are not positively selected are eliminated in the thymic cortex by apoptosis. Similarly, T cells that are negatively selected as a result of high-affinity binding to self-peptide/self-MHC complexes are also deleted through apoptotic death. These two selection processes result in the development of a population of T cells that are biased toward the recognition of peptides in association with self-MHC molecules from which those cells that are potentially auto-reactive (capable of high-affinity binding of self-peptide/self-MHC complexes) have been purged. One important event in the development of T cells is their differentiation from

double-positive cells into CD4+ or CD8+ single-positive cells. This process involves the interaction of double-positive thymocytes with peptide bound to class II or class I MHC molecules on accessory cells. Indeed, CD4 binds to monomorphic sites on class II molecules, whereas CD8 binds to comparable sites on class I molecules. The capacity of the TCR and CD4 (or of the TCR and CD8) to bind to a class II MHC (or a class I MHC) molecule on an accessory cell leads either to the differentiation of double-positive thymocytes into CD4+ (or CD8+) single-positive T cells or to the selection of cells that have “stochastically” differentiated down the CD4 (or CD8) pathway. Less is understood about the differentiation of thymocytes that express TCRs composed of ?/d chains. These cells fail to express either CD4 or CD8. However, ?/d cells are relatively numerous early in fetal life; this, together with their limited degree of heterogeneity, suggests that they may comprise a relatively primitive T-cell compartment. T-Lymphocyte Functions ( Chapter 10) T cells mediate a wide range of immunologic functions. These include the capacity to help B cells develop into antibody-producing cells, the capacity to increase the microbicidal action of monocyte/macrophages, the inhibition of certain types of immune responses, direct killing of target cells, and mobilization of the inflammatory response. In general, these effects depend on their expression of specific cell-surface molecules and the secretion of cytokines. T Cells That Help Antibody Responses ( Chapter 10) Helper T cells can stimulate B cells to make antibody responses to proteins and other T-cell–dependent antigens. T-cell–dependent antigens are immunogens in which individual epitopes appear only once or only a limited number of times so that they are unable to cross-link the membrane Ig of B cells or do so inefficiently. B cells bind antigen through their membrane Ig, and the complex undergoes endocytosis. Within the endosomal and lysosomal compartments, antigen is fragmented into peptides by proteolytic enzymes and one or more of the generated peptides are loaded into class II MHC molecules, which traffic through this vesicular compartment. The resulting complex of class II MHC molecule and bound peptide is exported to the B-cell surface membrane. T cells with receptors specific for the peptide/class II molecular complex recognize that complex on the B cell. B-cell activation depends not only on the binding of peptide/class II MHC complexes on the B cell surface by the TCR but also on the interaction of T-cell CD154 with CD40 on the B cell. T cells do not constitutively express CD154; rather, it is induced as a result of an interaction with an activated APC that expresses a cognate antigen recognized by the TCR of the T cell. Furthermore, CD80/86 are generally expressed by activated but not resting B cells so that interactions involving resting B cells and naïve T cells generally do not lead to efficient antibody production. By contrast, a T cell already activated and expressing CD154 can interact with a resting B cell, leading to its up-regulation of CD80/86 and to a more productive T-cell/B-cell interaction with the delivery of cognate help and the development of the B cell into an antibody-producing

cell. Similarly, activated B cells expressing large amounts of class II molecules and CD80/86 can act as effective APC and can participate with T cells in efficient cognate help interactions. Cross-linkage of membrane Ig on the B cell, even if inefficient, may synergize with the CD154/CD40 interaction to yield vigorous B-cell activation. The subsequent events in the B-cell response program, including proliferation, Ig secretion, and class switching either depend on or are enhanced by the actions of T-cell–derived cytokines. Thus, B-cell proliferation and Ig secretion are enhanced by the actions of several type I cytokines including IL-2 and IL-4. Ig class switching is dependent both on the initiation of competence for switching, which can be induced by the CD154/CD40 interaction, and on the targeting of particular C regions for switching, which is determined, in many instances, by cytokines. The best-studied example of this is the role of IL-4 in determining switching to IgG1 and IgE in the mouse and to IgG4 and IgE in the human. Indeed, the central role of IL-4 in the production of IgE is demonstrated by the fact that mice that lack the IL-4 gene or the gene for the IL-4 receptor a chain, as a result of hom*ologous recombination-mediated gene knockouts, have a marked defect in IgE production. Although CD4+ T cells with the phenotype of T H2 cells (i.e., IL-4, IL-13, IL-5, IL-6, and IL-10 producers) are efficient helper cells, T H1 cells also have the capacity to act as helpers. Because T H1 cells produce IFN-?, which acts as a switch factor for IgG2a in the mouse, T H1-mediated help often is dominated by the production of IgG2a antibodies. Induction of Cellular Immunity ( Chapter 10) T cells also may act to enhance the capacity of monocytes and macrophages to destroy intracellular microorganisms. In particular, IFN-? enhances several mechanisms through which mononuclear phagocytes destroy intracellular bacteria and parasites, including the generation of nitric oxide and induction of tumor necrosis factor (TNF) production. T H1-type cells are particularly effective in enhancing microbicidal action because they produce IFN-?. By contrast, two of the major cytokines produced by T H2 cells, IL-4 and IL-10, block these activities. Thus, T H2 cells often oppose the action of T H1 cells in inducing cellular immunity and in certain infections with microorganisms that are intracellular pathogens of macrophages, a T H2-dominated response may be associated with failure to control the infection. Regulatory T Cells ( Chapter 30) There has been a longstanding interest in the capacity of T cells to diminish as well as to help immune responses. Cells that mediate such effects are referred to as regulatory or suppressor T cells. Regulatory T cells may be identified by their constitutive expression of CD25, the IL-2 receptor alpha chain. These cells inhibit the capacity of both CD4 and CD8 T cells to respond to their cognate antigens. The mechanisms through which their suppressor function is mediated are still somewhat controversial. In some instances, it appears that cell–cell contact is essential for suppression, whereas in

other circ*mstances production of cytokines by the regulatory cells has been implicated in their ability to inhibit responses. Evidence has been presented for both IL-10 and TGFß as mediators of inhibition. Regulatory T cells have been particularly studied in the context of various autoimmune conditions. In the absence of regulatory cells, conventional T cells cause several types of autoimmune responses, including autoimmune gastritis and inflammatory bowel disease. Regulatory T cells express cell-surface receptors allowing them to recognize autoantigens and their responses to such recognition results in the suppression of responses by conventional T cells. Whether the T-cell receptor repertoire of the regulatory cells and the conventional T cells are the same has not been fully determined, nor it is completely clear whether regulatory (CD25+) T cells and conventional T cells derive from distinct T-cell lineages or whether regulatory T cells derive from conventional CD4+ T cells that may have been stimulated under certain conditions. Cytotoxic T Cells ( Chapter 36) One of the most striking actions of T cells is the lysis of cells expressing specific antigens. Most cells with such cytotoxic activity are CD8+ T cells that recognize peptides derived from proteins produced within the target cell, bound to class I MHC molecules expressed on the surface of the target cell. However, CD4+ T cells can express CTL activity, although in such cases the antigen recognized is a peptide associated with a class II MHC molecule; often such peptides derive from exogenous antigens. There are two major mechanisms of cytotoxicity. One involves the production by the CTL of perforin, a molecule that can insert into the membrane of target cells and promote the lysis of that cell. Perforin-mediated lysis is enhanced by a series of enzymes produced by activated CTLs, referred to as granzymes. Many active CTLs also express large amounts of Fas ligand on their surface. The interaction of Fas ligand on the surface of the CTL with Fas on the surface of the target cell initiates apoptosis in the target cell. CTL-mediated lysis is a major mechanism for the destruction of virally infected cells. If activated during the period in which the virus is in its eclipse phase, CTLs may be capable of eliminating the virus and curing the host with relatively limited cell destruction. On the other hand, vigorous CTL activity after a virus has been widely disseminated may lead to substantial tissue injury because of the large number of cells that are killed by the action of the CTLs. Thus, in many infections, the disease is caused by the destruction of tissue by CTLs rather than by the virus itself. One example is hepatitis B, in which much of the liver damage represents the attack of HBV-specific CTLs on infected liver cells. It is usually observed that CTLs that have been induced as a result of a viral infection or intentional immunization must be reactivated in vitro through the recognition of antigen on the target cell. This is particularly true if some interval has elapsed between the time of infection or immunization and the time of test. This has led to some question being

raised as to the importance of CTL immunity in protection against re-infection and how important CTL generation is in the long-term immunity induced by protective vaccines. On the other hand, in active infections, such as seen in HIV+ individuals, CTL that can kill their targets cells immediately are often seen. There is much evidence to suggest that these cells play an active role in controlling the number of HIV+ T cells.

CYTOKINES ( Chapter 23, Chapter 24, Chapter 25, and Chapter 26) Many of the functions of cells of the immune system are mediated through the production of a set of small proteins referred to as cytokines. These proteins can now be divided into several families. They include the type I cytokines or hematopoietins that encompass many of the interleukins (i.e., IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-11, IL-12, IL-13, IL-15, IL-21 and IL-23), as well as several hematopoietic growth factors; the type II cytokines, including the interferons and IL-10; the TNF-related molecules, including TNF, lymphotoxin, and Fas ligand; Ig superfamily members, including IL-1 and IL-18; and the chemokines, a growing family of molecules playing critical roles in a wide variety of immune and inflammatory functions. Many of the cytokines are T-cell products; their production represents one of the means through which the wide variety of functions of T cells are mediated. Most cytokines are not constitutive products of the T cell. Rather, they are produced in response to T-cell activation, usually resulting from presentation of antigen to T cells by APCs in concert with the action of a co-stimulatory molecule, such as the interaction of CD80/86 with CD28. Although cytokines are produced in small quantities, they are very potent, binding to their receptors with equilibrium constants of approximately 10 10 M -1. In some instances, cytokines are directionally secreted into the immunological synapse formed between a T cell and an APC. In such cases, the cytokine acts in a paracrine manner. Indeed, many cytokines have limited action at a distance from the cell that produced them. This appears to be particularly true of many of the type I cytokines. However, other cytokines act by diffusion through extracellular fluids and blood to target cells that are distant from the producers. Among these are cytokines that have pro-inflammatory effects, such as IL-1, IL-6, and TNF, and the chemokines, which play important roles in regulating the migration of lymphocytes and other cell types. Chemokines ( Chapter 26) A large family of small proteins that are chemotactic cytokines (chemokines) have been described. While members of this family have a variety of functions, perhaps the most dramatic is their capacity to regulate leukocyte migration and thus to act as critical dynamic organizers of cell distribution in the immune and inflammatory responses. The receptors for chemokines are seven transmembrane-spanning, G-protein coupled receptors. The chemokines are subdivided based on the number and positioning of their highly conserved cysteines. Among chemokines with four conserved cysteines, the cysteines are adjacent in one large group (the CC chemokines) while in a second large group they are separated by one amino acid (CXC chemokines). There are also rare chemokines in

which the cysteins are separated by three amino acids (CX3C) or in which there are only two conserved cysteins (C chemokines). Individual chemokines may signal through more than one chemokine recptor and individual receptors may interact with more than one chemokine, producing a very complex set of chemokine/chemokine receptor pairs and providing opportunities for exceedingly fine regulation of cellular functions.

THE MAJOR HISTOCOMPATIBILITY COMPLEX AND ANTIGEN PRESENTATION ( Chapter 19 and Chapter 20) The MHC has already been introduced in this chapter in the discussion of T-cell recognition of antigen-derived peptides bound to specialized grooves in class I and class II MHC proteins. Indeed, the class I and class II MHC molecules are essential to the process of T-cell recognition and response. Nonetheless, they were first recognized not for this reason but because of the dominant role that MHC class I and class II proteins play in transplantation immunity (see Chapter 47). When the genetic basis of transplantation rejection between mice of distinct inbred strains was sought, it was recognized that although multiple genetic regions contributed to the rejection process, one region played a dominant role. Differences at this region alone would cause prompt graft rejection, whereas any other individual difference usually resulted in a slow rejection of foreign tissue. For this reason, the genetic region responsible for prompt graft rejection was termed the major histocompatibility complex. In all higher vertebrates that have been thoroughly studied, a comparable MHC exists. The defining features of the MHC are the transplantation antigens that it encodes. These are the class I and class II MHC molecules. The genes encoding these molecules show an unprecedented degree of polymorphism. This together with their critical role in antigen presentation explains their central role as the target of the immune responses leading to the rejection of organ and tissue allografts. The MHC also includes other genes, particularly genes for certain complement components. In addition, genes for the cytokines TNF-a and lymphotoxin (also designated TNF-ß) are found in the MHC. Class I MHC Molecules ( Chapter 19) Class I MHC molecules are membrane glycoproteins expressed on most cells. They consist of an a chain of approximately 45,000 daltons noncovalently associated with ß2-microglobulin, a 12,000-dalton molecule ( Fig. 11). The gene for the a chain is encoded in the MHC, whereas that for ß2-microglobulin is not. Both the a chain and ß2-microglobulin are Ig supergene family members. The a chain is highly polymorphic, with the polymorphisms found mainly in the regions that constitute the binding sites for antigen-derived peptides and the contact sites for the TCR.

FIG. 11. Model of the class I HLA-A2 molecule. A schematic representation of the structure of the HLA-A2, class I MHC molecule. The polymorphic a1 and a2 domains are at the top. They form a groove into which antigen-derived peptides fit to form the peptide/MHC class I complex that is recognized by TCRs of CD8+ T cells. From Bjorkman PJ, Saper MA, Sauraomi B, et al., Structure of human class-I histocompatibility HLA-A. Nature 1987;329:506–512, with permission.

The class I a chain consists of three extracellular regions or domains, each of similar length, designated a1, a2, and a3. In addition, a chains have a membrane-spanning domain and a short carboxy-terminal cytoplasmic tail. The crystal structure of class I molecules indicates that the a1 and a2 domains form a site for the binding of peptides derived from antigens. This site is defined by a floor consisting of ß sheets and bounded by a-helical walls. The polymorphisms of the class I molecule are mainly in these areas. In the human, three loci encoding classical class I molecules have been defined; these are designated HLA-A, HLA-B, and HLA-C. All display high degrees of polymorphism. A similar situation exists in the mouse. In addition, there are a series of genes, defined principally in the mouse, that encode class I–like molecules (class Ib molecules). Recently, some of these also have been shown to have antigen-presenting activity for formylated peptides, suggesting that they may be specialized to present certain prokaryotic antigens. In addition, the class Ib molecule CD1 has been shown to have antigen-presenting function for mycobacterial lipids, providing a mechanism through which T cells specific for such molecules can be generated. In the mouse, a-galactosylceramide bound to CD1 is recognized by a novel class of T cells (NK T cells) that produce large amounts of cytokines upon stimulation. Class II MHC Molecules ( Chapter 19)

Class II MHC molecules are heterodimeric membrane glycoproteins. Their constituent chains are designated a and ß; both chains are immunoglobulin supergene family members, and both are encoded within the MHC. Each chain consists of two extracellular domains (a1 and a2; ß1 and ß2, respectively), a hydrophobic domain, and a short cytoplasmic segment. The overall conformation of class II MHC molecules appears to be quite similar to that of class I molecules. The peptide-binding site of the class II molecules is contributed to by the a1 and ß1 domains ( Fig. 5); it is within these domains that the majority of the polymorphic residues of class II molecules are found. A comparison of the three-dimensional structures of class I and class II molecules indicates certain distinctive features that explain differences in the length of peptides that the two types of MHC molecules can bind. Class I molecules generally bind peptides with a mean length of nine amino acids, whereas class II molecules can bind substantially larger peptides. In the mouse, class II MHC molecules are encoded by genes within the I region of the MHC. These molecules are often referred to as I region—associated (Ia) antigens. Two sets of class II molecules exist, designated I-A and I-E, respectively. The a and ß chains of the I-A molecules (Aa and Aß) pair with one another, as do the a and ß chains of I-E (Ea and Eß). In general, cross-pairing between I-A and I-E chains does not occur, although exceptions have been described. In heterozygous mice, a and ß chains encoded on alternative chromosomes (i.e., Aa b and Aß k) may cross-pair so that heterozygous mice can express both parental and hybrid class II molecules. However, the degree of cross-pairing is allele specific; not all hybrid pairs are formed with equal efficiency. In the human, there are three major sets of class II molecules, encoded in the DR, DQ, and DP regions of the HLA complex. Class II molecules have a more restricted tissue distribution than do class I molecules. Class II molecules are found on B cells, dendritic cells, epidermal Langerhans cells, macrophages, thymic epithelial cells, and, in the human, activated T cells. Levels of class II molecule expression are regulated in many cell types by interferons and in B cells by IL-4. Indeed, interferons can cause expression of class II molecules on many cell types that normally lack these cell-surface molecules. Interferons also can cause striking up-regulation in the expression of class I MHC molecules. Thus, immunologically mediated inflammation may result in aberrant expression of class II MHC molecules and heightened expression of class I molecules. Such altered expression of MHC molecules can allow cells that do not normally function as APCs for CD4+ T cells to do so and enhances the sensitivity of such cells to CD8+ T cells. This has important consequences for immunopathologic responses and for autoimmunity. Antigen Presentation ( Chapter 20) As already discussed, the function of class I and class II MHC molecules is to bind and present antigen-derived peptides to T cells whose receptors can recognize the

peptide/MHC complex that is generated. There are two major types of antigen-processing pathways, specialized to deal with distinct classes of pathogens that the T cell system must confront ( Fig. 8). Extracellular bacteria and extracellular proteins enter APCs by endocytosis or phagocytosis. Their antigens and the antigens of bacteria that live within endosomes or lysosomes are fragmented in these organelles and peptides derived from the antigen are loaded into class II MHC molecules as these proteins traverse the vesicular compartments in which the peptides are found. The loading of peptide is important in stabilizing the structure of the class II MHC molecule. The acidic pH of the compartments in which loading occurs facilitates the loading process. However, once the peptide-loaded class II molecules reaches neutral pH, such as at the cell surface, the peptide/MHC complex is stable. Peptide dissociation from such class II molecules is very slow, with a half-time measured in hours. The peptide/class II complex is recognized by T cells of the CD4 class with complementary receptors. As already pointed out, the specialization of CD4+ T cells to recognize peptide/class II complexes is due to the affinity of the CD4 molecule for monomorphic determinants on class II molecules. Obviously, this form of antigen processing can only apply to cells that express class II MHC molecules. Indeed, APCs for CD4+ T cells principally include cells that normally express class II MHC molecules, including dendritic cells, B cells, and macrophages. T cells also can recognize proteins that are produced within the cell that presents the antigen. The major pathogens recognized by this means are viruses and other obligate intracellular (nonendosomal/lysosomal) microbes that have infected cells. In addition, proteins that are unique to tumors, such as mutant oncogenes, or are overexpressed in tumors also can be recognized by T cells. Endogenously produced proteins are fragmented in the cytosol by the proteases in the proteasome. The resultant peptides are transported into the rough endoplasmic reticulum, through the action of a specialized transport system. These peptides are then available for loading into class I molecules. In contrast to the loading of class II molecules, which is facilitated by the acid pH of the loading environment, the loading of class I molecules is controlled by interaction of the class I a chain with ß2-microglobulin. Thus, the bond between peptide and class I molecule is generally weak in the absence of ß2-microglobulin, and the binding of ß2-microglobulin strikingly stabilizes the complex. (Similarly, the binding of ß2-microglobulin to the a chain is markedly enhanced by the presence of peptide in the a chain groove.) The peptide-loaded class I molecule is then brought to the cell surface. In contrast to peptide-loaded class II molecules, that are recognized by CD4+ T cells, peptide-loaded class I molecules are recognized by CD8+ T cells. This form of antigen processing and presentation can be performed by virtually all cells because, with a few exceptions, class I MHC molecules are universally expressed. Although the specialization of class I molecules to bind and present endogenously produced peptides and of class II molecules to bind and present peptides derived from exogenous antigens is generally correct, there are exceptions, many of which have physiologic importance. Particularly important is the re-presentation by class II+ cells of antigens derived from class II- cells.

T-Lymphocyte Recognition of Peptide/MHC Complexes Results in MHC-Restricted Recognition ( Chapter 8) Before the biochemical nature of the interaction between antigen-derived peptides and MHC molecules was recognized, it was observed that T-cell responses displayed MHC-restricted antigen recognition. Thus, if individual animals were primed to a given antigen, their T cells would be able to recognize and respond to that antigen only if the APCs that presented the antigen shared MHC molecules with the animal that had been immunized. The antigen would not be recognized when presented by APCs of an allogeneic MHC type. This can now be explained by the fact that the TCR recognizes the peptide bound to an MHC molecule. MHC molecules display high degrees of polymorphism, and this polymorphism is concentrated in the regions of the class I and class II molecules that interact with the peptide and that can bind to the TCR. Differences in structure of the MHC molecules derived from different individuals (or different inbred strains of mice) profoundly affect the recognition process. Two obvious explanations exist to account for this. First, the structure of the grooves in different class I or class II MHC molecules may determine that a different range of peptides are bound or, even if the same peptide is bound, may change the conformation of the surface of the peptide presented to the TCR. Second, polymorphic sites on the walls of the a-helices that are exposed to the TCR can either enhance or diminish binding of the whole complex, depending on their structure. Thus, priming an individual with a given antigen on APCs that are syngeneic to the individual will elicit a response by T cells whose TCRs are specific for a complex consisting of a peptide derived from the antigen and the exposed polymorphic residues of the MHC molecule. When the same antigen is used with APCs of different MHC types, it is unlikely that the same peptide/MHC surface can be formed, and thus the primed T cells are not likely to bind and respond to such stimulation. Indeed, this process also occurs within the thymus in the generation of the T-cell repertoire, as already discussed. T cells developing within the thymus undergo a positive selection event in which those T cells capable of recognizing MHC molecules displayed within the thymus are selected (and the remainder undergo programmed cell death). This leads to the skewing of the population of T cells that emerges from the thymus so that the cells are specialized to respond to peptides on self-MHC molecules. One of the unsolved enigmas of positive selection within the thymus is how the vast array of T cells with receptors capable of reacting with a very large set of foreign peptides associated with self-MHC molecules are chosen by self-MHC molecules that can only display self-peptides. It is believed that a high degree of cross-reactivity may exist so that T cells selected to bind a given class I (or class II) molecule plus a particular self-peptide can also bind a set of other (foreign) peptides bound to the same MHC molecule. Furthermore, the affinity of an interaction required for positive selection in the thymus appears to be considerably lower that that required for full activation of peripheral T cells. Thus, thymocytes selected by a given self-peptide/self-MHC complex will generally not mount a full response when they encounter the same peptide/MHC complex in the periphery, although they will respond to a set of foreign peptide/MHC

complexes to which they bind with higher affinity. Recognition of the self-peptide/self-MHC complex in the periphery may nonetheless have important consequences, such as sustaining the viability of resting lymphocytes. Our modern understanding of T-cell recognition also aids in explaining the phenomenon of immune response (Ir) gene control of specific responses. In many situations, the capacity to recognize simple antigens can be found in only some members of a species. In most such cases, the genes that determine the capacity to make these responses have been mapped to the MHC. We would now explain Ir gene control of immune responses based on the capacity of different class II MHC molecules (or class I MHC molecules) to bind different sets of peptides. Thus, for simple molecules, it is likely that peptides can be generated that are only capable of binding to some of the polymorphic MHC molecules of the species. Only individuals that possess those allelic forms of the MHC will be able to respond to those antigens. Based on this, some individuals are nonresponders because of the failure to generate a peptide/MHC molecule complex that can be recognized by the T-cell system. This mechanism also may explain the linkage of MHC type with susceptibility to various diseases. Many diseases show a greater incidence in individuals of a given MHC type. These include reactive arthritides, gluten-sensitive enteropathy, insulin-dependent diabetes mellitus, and rheumatoid arthritis (see Chapter 44 and Chapter 45). One explanation is that the MHC type that is associated with increased incidence may convey altered responsiveness to antigens of agents that cause or exacerbate the disease. Indeed, it appears that many of these diseases may be due to enhanced or inappropriate immune responses. Antigen-Presenting Cells ( Chapter 15) T cells recognize peptide/MHC complexes on the surface of other cells. Such cells are often referred to as antigen-presenting cells (APCs). Although effector cells can mediate their functions by recognizing such complexes on virtually any cell type, naïve cells are most efficiently activated by a set of specialized APCs, the dendritic cells (DCs). DCs are a multimember family whose complexity is only now being worked out. Both the common myeloid precursor and the common lymphoid precursor can give rise to immature DCs. In humans, there are two types of immature myeloid DCs emerging from the common myeloid precursor, CD11c+, CD14+ cells and CD11C+, CD14- cells. These cells become interstitial DCs and Langerhans cells. Common myeloid precursors also give rise to monocytes and plasmacytoid cells, which can act as DC precursors in the tissues. DCs can also arise from common lymphoid precursors. In the mouse, this has been demonstrated in vivo; in the human, in vitro. In general, in their immature form, DCs are resident in the tissues where they are efficient at capturing and endocytosing antigen. Their antigen capture activity is dependent upon expression of several surface receptors including Fc receptors, receptors for heat shock proteins, and C-type lectins. If they receive signals, such as various inflammatory stimuli, often mediated by TLRs, they are stimulated to down-regulate the expression of these molecules but to increase their expression of surface MHC molecules and various co-stimulatory molecules such as CD80/86. In

addition, such stimulation induces expression of chemokine receptors such as CCR2 and CCR7. The latter allows cells to follow signals from the chemokines SLC and ELC and to migrate into the T-cell zone of lymph nodes. As part of the maturation process, they may also acquire the capacity to produce cytokines that can aid in determining the polarization of T-cell priming. This includes the production of IL-12 p70 and the production of IFN? itself. Such cells are highly efficient at priming naïve cells to develop into TH1 cells. Other sets of DCs have been reported to favor TH2 development and interaction of developing T cells with immature DCs may induce a state of peripheral tolerance. One important function of DCs is the ability to acquire antigen from virally infected cells and to cross-present it through the class I pathway. This allows DCs to aid in the priming of precursors of cytotoxic T cells specific for viruses that do not infect the DCs themselves.

EFFECTOR MECHANISMS OF IMMUNITY The ultimate purpose of the immune system is to mount responses that protect the individual against infections with pathogenic microorganisms by eliminating these microbes or, where it is not possible to eliminate infection, to control their spread and virulence. In addition, the immune system may play an important role in the control of the development and spread of some malignant tumors. The responses that actually cause the destruction of the agents that initiate these pathogenic states (e.g., bacteria, viruses, parasites, and tumor cells) are collectively the effector mechanisms of the immune system. Several have already been alluded to. Among them are the cytotoxic action of CTLs, which leads to the destruction of cells harboring viruses and, in some circ*mstances, expressing tumor antigens. In some cases, antibody can be directly protective by neutralizing determinants essential to a critical step through which the pathogen establishes or spreads an infectious process. However, in most cases, the immune system mobilizes powerful nonspecific mechanisms to mediate its effector function. Effector Cells of the Immune Response Among the cells that mediate important functions in the immune system are cells of the monocyte/macrophage lineage, NK cells, mast cells, basophils, eosinophils, and neutrophils. It is beyond the scope of this introductory chapter to present an extended discussion of each of these important cell types. However, a brief mention of some of their actions will help in understanding their critical functions in the immune response. Monocytes and Macrophages ( Chapter 16) Cells of the monocyte/macrophage lineage play a central role in immunity. One of the key goals of cellular immunity is to aid the macrophages in eliminating organisms that have established such intracellular infections. In general, nonactivated macrophages are inefficient in destroying intracellular microbes. However, the production of IFN-? and other mediators by T cells can enhance the capacity of macrophages to eliminate such

microorganisms. Several mechanisms exist for this purpose, including the development of reactive forms of oxygen, the development of nitric oxide, and the induction of a series of proteolytic enzymes, as well as the induction of cytokine production. Macrophages can act as APCs and thus can enlist the “help” of activated, cytokine-producing CD4+ T cells in regulating their function. Although macrophages function as APCs for attracting activated T cells, they do not appear to be particularly effective in the activation of naïve CD4 T cells. In instances in which they are the site of infection or have phagocytosed infectious agents or their proteins, antigens from these agents may be transferred to dendritic cells. In such cases the dendritic cells would be the principal antigen-presenting cells that activate naïve or possibly resting-memory CD4 T cells. This process is often described as cross-presentation. Such activated T cells would then be available to help infected macrophages. Natural Killer Cells ( Chapter 12) Natural killer cells play an important role in the immune system. Indeed, in mice that lack mature T and B cells due to the SCID mutation, the NK system appears to be highly active and to provide these animals a substantial measure of protection against infection. NK cells are closely related to T cells. They lack conventional TCR (or Ig) but express two classes of receptors. They have a set of positive receptors that allow them to recognize features associated with virally infected cells or tumor cells. They also express receptors for MHC molecules that shut off their lytic activity. Thus, virally infected cells or tumor cells that escape the surveillance of cytotoxic T cells by down-regulating or shutting off expression of MHC molecules then become targets for efficient killing by NK cells, because the cytotoxic activity of the latter cells is no longer shut off by the recognition of particular alleles of MHC class I molecules. In addition, NK cells express a receptor for the Fc portion of IgG (Fc?RIII). Antibody-coated cells can be recognized by NK cells, and such cells can then be lysed. This process is referred to as antibody-dependent cellular cytotoxicity (ADCC). NK cells are efficient producers of IFN-?. A variety of stimuli, including recognition of virally infected cells and tumor cells, cross-linkage of Fc?RIII and stimulation by the cytokines IL-12 and IL-18, cause striking induction of IFN-? production by NK cells. Mast Cells and Basophils ( Chapter 46) Mast cells and basophils play important roles in the induction of allergic inflammatory responses. They express cell-surface receptors for the Fc portions of IgE (FceRI) and for certain classes of IgG (Fc?R). This enables them to bind antibody to their surfaces, and when antigens capable of reacting with that antibody are introduced, the resultant cross-linkage of FceRI and/or Fc?R results in the prompt release of a series of potent mediators, such as histamine, serotonin, and a variety of enzymes that play critical roles in initiating allergic and anaphylactic-type responses. In addition, such stimulation also causes these cells to produce a set of cytokines, including IL-3, IL-4, IL-13, IL-5, IL-6, granulocyte–macrophage colony-stimulating factor (GM-CSF), and TNFa, which have

important late consequences in allergic inflammatory responses. Granulocytes ( Chapter 37) Granulocytes have critical roles to play in a wide range of inflammatory situations. Rather than attempting an extended discussion of these potent cells, it may be sufficient to say that in their absence it is exceedingly difficult to clear infections with extracellular bacteria and that the immune response plays an important role in orchestrating the growth, differentiation, and mobilization of these crucial cells. Eosinophils ( Chapter 38 and Chapter 46) Eosinophils are bone marrow–derived myeloid cells that complete their late differentiation under the influence of IL-5. They migrate to tissue sites in response to the chemokine eotaxin and as a result of their adhesion receptors. Since TH2 cells can produce IL-5 and stimulate the production of eotaxin, eosinophil accumulation is often associated with TH2-mediated inflammation. Eosinophils store a series of proteins in their secondary granules, including major basic protein, eosinophil cationic protein and eosinophil peroxidase. When released, these proteins are responsible for much of the damage that eosinophils mediate, both to helminthic parasites and to the epithelium. They have been implicated as important in protective responses to helminths and in the tissue damage seen in allergic inflammation in conditions such as asthma. Eosinophils can also produce a set of cytokines. The Complement System ( Chapter 34) The complement system is a complex system of proteolytic enzymes, regulatory and inflammatory proteins and peptides, cell-surface receptors, and proteins capable of causing the lysis of cells. The system can be thought of as consisting of three arrays of proteins. Two of these sets of proteins, when engaged, lead to the activation of the third component of complement (C3) ( Fig. 12). The activation of C3 releases proteins that are critical for opsonization (preparation for phagocytosis) of bacteria and other particles and engages the third set of proteins that insert into biologic membranes and produce cell death through osmotic lysis. In addition, fragments generated from some of the complement components (e.g., C3a and C5a) have potent inflammatory activities.

FIG. 12. The complement system. The classical pathway of complement activation, usually initiated by the aggregation of C1 by binding to antigen–antibody complexes, resulting in the formation of an enzyme, a C3 convertase, that cleaves C3 into two fragments, C3b and C3a. The classical pathway can also be initiated by the aggregation of MBLectin as a result of binding sugars expressed in the capsules of many pathogenic microbes. The components of the MBLectin pathway appear to mimic the function of C1qrs. The alternative pathway of complement activation provides a potent means of activating complement without requiring antibody recognition of antigen. It results in the formation of a distinct C3 convertase. The fragments formed by cleaving C3 have important biologic activities. In addition, C3b, together with elements of the classical pathway (C4b,C2a) or the alternative pathway (Bb, properdin), form enzymes (C5 convertases) that cleave C5, the initial member of the terminal family of proteins. Cleavage of C5 leads to the formation of the membrane attack complex that can result in the osmotic lysis of cells.

The Classical Pathway of Complement Activation The two activation systems for C3 are referred to as the classical pathway and the alternative pathway. The classical pathway is initiated by the formation of complexes of antigen with IgM or IgG antibody. This leads to the binding of the first component of complement, C1, and its activation, creating the C1 esterase that can cleave the next two components of the complement system, C4 and C2. C4 is a trimeric molecule, consisting of a, ß, and ? chains. C1 esterase cleaves the a chain, releasing C4b, which binds to surfaces in the immediate vicinity of the antigen/antibody/C1 esterase complex. A single C1 esterase molecule will cause the deposition of multiple C4b molecules. C2 is a single polypeptide chain that binds to C4b and is then proteolytically cleaved by C1 esterase, releasing C2b. The resulting complex of the residual portion of C2 (C2a) with C4b (C4b2a) is a serine protease whose substrate is C3. Cleavage of C3 by C4b2a (also referred to as the classical pathway C3 convertase) results in the release of C3a and C3b. A single antigen–antibody complex and its associated C1 esterase can lead to the production of a large number of C3 convertases (i.e., C4b2a complexes) and thus to cleavage of a large number of C3 molecules.

The components of the classical pathway can be activated by a distinct, non–antibody-dependent mechanism. The mannose-binding lectin (MBL) is activated by binding to (and being cross-linked by) repetitive sugar residues such as N-acetylglucosamine or mannose. The activation of MBL recruits the MBL-associated serine proteases MASP-1 and MASP-2, which cleave C4 and C2 and lead to the formation of the classical pathway C3 convertase. Because the capsules of several pathogenic microbes can be bound by MBL, this provides an antibody-independent pathway through which the complement system can be activated by foreign microorganisms. The Alternative Pathway of Complement Activation Although discovered more recently, the alternative pathway is the evolutionarily more ancient system of complement activation. Indeed, this system, and the MBL activation of the classical pathway, can be regarded as providing individuals with an innate immune system. The alternative pathway can be activated by a variety of agents such as insoluble, yeast cell–wall preparations and bacterial lipopolysaccharide. Antigen–antibody complexes also can activate the alternative pathway. The C3 convertase of the alternative pathway consists of a complex of C3b (itself a product of cleavage of C3) bound to the b fragment of the molecule factor B. C3bBb is produced by the action of the hydrolytic enzyme, factor D, that cleaves factor B; this cleavage only occurs when factor B has been bound by C3b. Apart from the importance of the alternative pathway in activating the complement system in response to nonspecific stimulants, it also can act to amplify the activity of the classical pathway because the C3 convertase of the classical system (C4b2a) provides a source of C3b that can strikingly enhance formation of the alternative pathway convertase (C3bBb) in the presence of factor D. The Terminal Components of the Complement System C3b, formed from C3 by the action of the C3 convertases, possesses an internal thioester bond that can be cleaved to form a free sulfhydryl group. The latter can form a covalent bond with a variety of surface structures. C3b is recognized by receptors on various types of cells, including macrophages and B cells. The binding of C3b to antibody-coated bacteria is often an essential step for the phagocytosis of these microbes by macrophages. C3b is also essential to the engagement of the terminal components of the complement system (C5 through C9) to form the membrane attack complex that causes cellular lysis. This process is initiated by the cleavage of C5, a 200,000-dalton two-chain molecule. The C5 convertases that catalyze this reaction are C4b2a3b (the classical pathway C5 convertase) or a complex of C3bBb with a protein-designated properdin (the alternative pathway C5 convertase). Cleaved C5, C5b, forms a complex with C6 and then with C7, C8, and C9. This C5b–C9 complex behaves as an integral membrane protein that is responsible for the formation of complement-induced lesions in cell membranes. Such

lesions have a donut like appearance, with C9 molecules forming the ring of the donut. In addition to the role of the complement system in opsonization and cell lysis, several of the fragments of complement components formed during activation are potent mediators of inflammation. C3a, the 9,000-dalton fragment released by the action of the C3 convertases, binds to receptors on mast cells and basophils, resulting in the release of histamine and other mediators of anaphylaxis. C3a is thus termed an anaphylotoxin, as is C5a, the 11,000-dalton fragment released as a result of the action of the C5 convertases. C5a is also a chemoattractant for neutrophils and monocytes. Finally, it is important to note that the process of activation of the complement cascade is highly regulated. Several regulatory proteins (e.g., C1 esterase inhibitor, decay accelerator factor, membrane cofactor protein) exist that function to prevent uncontrolled complement activation. Abnormalities in these regulatory proteins are often associated with clinical disorders such as hereditary angioedema and paroxysmal nocturnal hemoglobinuria.

CONCLUSION This introductory chapter should provide the reader with an appreciation of the overall organization of the immune system and of the properties of its key cellular and molecular components. It should be obvious that the immune system is highly complex, that it is capable of a wide range of effector functions, and that its activities are subject to potent, but only partially understood, regulatory processes. As the most versatile and powerful defense of higher organisms, the immune system may provide the key to the development of effective means to treat and prevent a broad range of diseases. Indeed, the last two sections of this book deal with immunity to infectious agents and immunologic mechanisms in disease. The introductory material provided here should be of considerable help to the uninitiated reader in understanding the immunologic mechanisms brought into play in a wide range of clinical conditions in which immune processes play a major role either in pathogenesis or in recovery.

Chapter 2 History of Immunology Fundamental Immunology

Chapter 2 Pauline Mazumdar

History of Immunology

OVERVIEW VACCINATION THE AGE OF SEROLOGY, 1890–1950 The Side-Chain Theory of Antibody Production Colloid Chemistry and the Template Theory of Antibody Production Allergy and the Clinic Serology at the League of Nations Blood Groups and Transfusion THE CHEMISTRY OF THE ANTIBODY GLOBULINS, 1930–1960 Myeloma Proteins—A Model System CELLULAR IMMUNOLOGY AND THE SELECTION THEORIES 1950s TO 1980s Graft Rejection and Tolerance The Clonal Selection Theory The Biology of the Thymus and the Dictatorship of the Lymphocyte Monoclonal Antibodies MOLECULAR IMMUNOLOGY: DIVERSITY, HISTOCOMPATIBILITY, AND THE T-CELL RECEPTOR, 1980–PRESENT AIDS: THE PUBLIC FACE OF IMMUNOLOGY, 1986 TO THE PRESENT CONCLUSION REFERENCES

OVERVIEW With the important exception of smallpox inoculation, immunology as modern science dates from the 1880s. Its history falls roughly into two periods, before and after World War II. It begins with serology: identification of bacteria, clinical application of vaccines and sera to infectious diseases, and the chemical problems of specificity and antibody diversity. Paul Erlich’s side-chain theory of antibody production was replaced from about 1930 onward by Felix Haurowitz’s template theory. Sources for this period are mainly German or French. After World War II, transplantation rather than infectious disease was paradigmatic. Unlike other biosciences, immunology was not reductionist: The newer work guided by the clonal selection theory concentrated on the activities of clones of cells and on experimental animals, rather than on chemistry. Major growth occurred in the 1960s and 1970s, and there are many memoirs by immunologists from that period. However, with the advent of monoclonal antibodies, interest in specificity was renewed, and serology entered a new period of growth powered by molecular biology and the pharmaceutical industry. Most of the writing by historians dates from the 1990s and deals with social, scientific, and business history. New writing has emphasized the role of experimental systems,

techniques, and instruments, as well as language. As Cambrosio ( 1 ) said, the history of science has as its object a cultural product: It is a history of culture not of nature. Until recently, however, there has been little emphasis on the interaction of the laboratory science with the clinic.

VACCINATION The earliest known smallpox inoculation took place in China, perhaps as early as the 5th century AD. The Chinese method was reported to the Royal Society by an English merchant, John Lister, in 1700. A Jesuit priest, Father d’Entrecolles ( 2 ), provided details of the method, which he said was to collect scabs from the pustules, and blow a powder made from them into an infant’s nose. The scabs or a thread imbibed with the pus could be stored, but the operation was usually done face-to-face with a sick patient. The same method was used in Japan beginning in 1747. In precolonial India, a tika or dot would be made on a child, usually on the sole of the foot, by traditional tikadars who were invited into a home (this professional niche was later blacklisted by colonial-era medical practitioners). The Turkish method was communicated to the Royal Society by Dr. Emmannuel Timoni in 1714. As commonly practiced in Constantinople, a small perforation was made in the skin, and a spot of pus from a benign case introduced with a needle. In 1715, the method famously came to the notice of Lady Mary Montagu, wife of the English ambassador in Constantinople, who used it on her own son, and subsequently talked it up to great effect in aristocratic circles at home in England ( 3 ). Although nationalistic, ethical, and religious objections to this non-European folk practice abounded, the Royal Society with its interest in the empirical recorded many accounts of inoculation presented at its meetings. Dr. James Jurin, its secretary, an early user of the quantitative method, collected large numbers of cases in an effort to compare the risks from inoculation and from the disease. According to his figures, smallpox was both universal and often fatal: He assumed that almost everyone over the age of 2 had had it, and for every person who died, 7 or 8 recovered; inoculation, on the other hand, had a death rate of about 1 in 50. He had not, he said, been able to learn of any person either in England or Turkey, who had been inoculated but still took the disease in the natural way ( 4 ). The mathematician Daniel Bernoulli calculated similarly that if one neglected the point of view of the individual, inoculation would be useful to the state. In 18th-century France, according to Anne-Marie Moulin, the method was discussed, for instance by the Encyclopédistes, but not practiced; it was made illegal by a decree of 1763, and only permitted after the revolution. In England, on the other hand, it seems possible that it was used often enough by the end of the 18th century to affect the incidence and severity of smallpox ( 5 ). The use of Vaccinia (cowpox) as inoculum was suggested several times in the late 1700s; the country doctor and inoculator Edward Jenner tried it out in 1798. He had heard it said that milkmaids who had had cowpox, never caught smallpox, and it struck him that he might be able to propagate the disease as he was accustomed to do with his usual inoculum. It is not clear whether in practice the material actually used was always Vaccinia ( 6 ). Vaccine production was unregulated; the operation was painful and sometimes did not “take.” Nevertheless, public health authorities enforced it, for example, in Prussia and later under the British Compulsory Vaccination Act of 1853. Compulsion led to worldwide antivaccination movements with strong political and

anticolonial overtones ( 7 ). However, the demographer Alex Mercer makes a strong case for its effectiveness: He argues that inoculation and subsequently vaccination were key in the general decline in death rates that took place from the late 18th through the 19th century, as the incidence of smallpox declined. With it went a network of linked respiratory diseases, late sequelae of the damage done by smallpox even when not fatal ( 8 ). It should not be supposed, however, that because vaccination was accepted, an immune theory of disease resistance was an obvious conclusion. The experience of colonial troops in the tropics, where most of them died within a year or two of arrival throughout the 18th and 19th centuries, prompted a racial view of resistance, coupled with the development of acclimatization or seasoning in those few who survived. The constitution of the alien race soon broke down in the unfamiliar conditions of temperature and humidity; the expatriates felt themselves weakened by perspiration, tight clothes, and local miasmas that did not seem to affect the natives. There is a large 19th-century literature advising the displaced European on how to survive a posting to India, the Caribbean, or the Philippines, and on the tragic return home of the soldier or sailor broken in health by the tropics ( 9 ). The importation of Africans to work as slaves in the conditions that were so fatal to Europeans and white Americans was one of the results of the racial view of disease resistance. A theory with such significant historical connotations cannot be ignored ( 10 ). The word vaccine originally applied only to Vaccinia. Anne-Marie Moulin points out that it was Louis Pasteur, who by claiming Jenner as his predecessor, metaphorically included in that word all prophylactic inoculation by attenuated virus-vaccins, organisms attenuated by passage through another species or by treatment with oxygen or antiseptics ( 11 ). Vaccines were prepared in this way against anthrax (1881), which was then a common agricultural problem, and rabies (1885) a frighteningly fatal result of the bite of a rabid animal. These vaccines were dramatically effective, although it was never clear whether the victim of a dog bite had in fact been infected. They led to a flood of donations from a hero-worshipping public, with which the Institut Pasteur was established in 1888. In 1891, Robert Koch too had a dramatic announcement, which also paved the way for the establishment of an institute under his direction. “Koch’s lymph” was a cure for tuberculosis, raising the hopes of sufferers who rushed to Berlin to be treated by the man who had discovered the tubercle bacillus. The reaction was acute and sometimes quite harmful to the patients, and the results were certainly not as good as expected. But it was not the debacle that has sometimes been thought. Koch’s Old Tuberculin continued to be made until the 1940s for use as a treatment for chronic tuberculosis of bones, lymph nodes, and skin. The material was a protein extract of tubercle bacilli, which Koch regarded as an exotoxin similar to that produced by diphtheria bacilli. It was later used under the name of the Mantoux reaction as a skin test for tuberculosis ( 12 ). In 1896, Sir Almroth Wright of St. Mary’s Hospital in London and Richard Pfeiffer and Wilhelm Kolle in Berlin simultaneously prepared a vaccine against typhoid, an important disease in Europe and the colonies. Like the smallpox vaccine, it was very promising, but was attacked passionately by antivaccinationists. Their position was primarily

political and ideological, but typhoid was a water-borne infection, and it was argued that improvements in sanitation and water supplies would eventually make vaccination unnecessary. Hostility focused on Wright’s vaccine especially; it made its recipients feel very ill, and its effectiveness was statistically doubtful. Sir William Leishman of the Royal Army Medical Corps developed a vaccine incorporating typhoid and the newly defined paratyphoids A and B in 1909. Armies in France, Germany, and the United States were beginning to use the newer type, but in Britain compulsion was politically unacceptable, and when World War I came, the Royal Army Medical Corps depended upon pro-vaccination propaganda. As acceptance of the vaccine increased among the troops, the results became more obvious: Compared to dysentery, a disease that was similarly transmitted through infected water supplies, the numbers of enteric cases reported in the field fell steeply ( 13 ). Attempts to develop a dysentery vaccine were unsuccessful. In the 1880s, germ theory had started to sound persuasive ( 14 ). In 1883, the Russian zoologist Élie Metchnikov had suggested that white blood cells attacked invaders from outside the body, an idea based on the Darwinian concept of interspecies struggle for existence, and which he saw as a form of “physiological inflammation” ( 15 ). Pasteur liked Metchnikov’s idea, and invited him to Paris. Alfred Tauber sees Metchnikov’s phagocytosis theory as the foundation of the self–not self concept, later to be central to immunology, and thinks that Metchnikov should be regarded as having founded the discipline ( 16 ). But as Anne-Marie Moulin points out, Metchnikov’s phagocytes had neither specificity nor memory; they simply engulfed particles ( 17 ). In the first half of the 20th century, the practical aspects of immunity, vaccination, and serum therapy defined research in the field. Serology and immunochemistry strove to provide a theoretical basis for these practices. Mechnikov’s phagocytosis theory was briefly at center stage but was soon overtaken by a rush of publications from Koch and colleagues in Berlin—the Franco-Prussian war of 1870 was still being fought by other means ( 18 ). As bacteriologists, the Berlin group favored “humoral immunity” in preference to cellular: They focused on immune sera for their specificity to identify bacteria, and ignored the cells, which seemed to carry a taint of old-fashioned vitalism. Cell-based vaccination systems, however, were to prove popular and very lucrative for their producers, especially in France. At the Institut Pasteur, Metchnikov’s lineage of workers in the cellular style continued to flourish. Alexandre Besredka came to Paris in 1893; he was from Odessa, like Metchnikov, and found work in Metchnikov’s laboratory. In 1918, he succeeded Metchnikov at its head. His interest centered on the then newly described phenomenon of anaphylaxis ( 19 ). He was concerned with sensitization and desensitization of the skin, an interest that was to evolve into his studies of natural resistance and acquired localized immunity. He proposed a system of specific dressings or local injections of a prepared antigen, a parallel to the local injections that desensitized animals to anaphylactic shock. The “terrain,” the skin cells that allowed entry to the infection, was to be made resistant ( 20 ). Besredka’s co-worker, Michel Bardach, was also from Odessa. He began work on an anti-reticuloendothelial serum along the lines suggested by a Russian researcher, Alexander Bogomoletz, who claimed that his serum was effective in a broad range of diseases involving that system. After World War II, the serum was successfully and profitably marketed through the

Institut Pasteur as a nonspecific stimulator of immunity, only to be abandoned in the 1950s as ineffective, perhaps by contrast with the stunning success of penicillin. A rather similar cell-based system had been developed in England. Sir Almroth Wright, originator of an early typhoid vaccine, linked cells with serum in an effort to boost immunity by the preparation of autovaccines from a patient’s own lesion; they were thought to raise a patient’s serum “opsonic index,” and like Bardach’s serum at a later date, to stimulate phagocytosis ( 21 ). Wright’s slogan of 1909, “The physician of the future will be an immunisator,” seems to have been perfectly true for the first decades of the 20th century ( 22 ). Wright’s department at St. Mary’s Hospital London made autovaccines and carried out thousands of index measurements yearly between 1908 and 1945. He built up a practice on a huge, even industrial, scale, out of which the department and the hospital itself were financed. As Wei Chen has commented, his laboratory was a vaccine factory, profitably manufacturing typhoid vaccine as well as the autovaccines that were Wright’s specialty ( 23 ). The effectiveness of autovaccine therapy, like the effectiveness of his typhoid vaccine, was attacked by the statisticians. Even so, laboratory texts until the mid-1940s generally included a chapter on the technique of preparing an autovaccine ( 24 ). Wright’s student George Ross carried both antityphoid and autovaccine manufacture with him to Canada in 1907, to an appointment at the Toronto General Hospital, where his techniques established and funded a new laboratory-based Department of Immunization and Medical Research, a precursor to the Connaught Laboratory, Toronto’s serum institute ( 25 ). The use of Wright’s autovaccines, along with Koch’s Old Tuberculin, persisted more or less up to the appearance of penicillin on the therapeutic scene in 1945, when all such minimally effective treatments were swept away by the brilliance of the first antibiotics. Wei Chen has suggested that Wright’s vaccine program provided a model and a financial goal for his junior colleague Alexander Fleming’s “construction” of penicillin. She shows that penicillin was initially seen as a means of differentially culturing Bacillus influenzae from cases of influenza, and supporting Wright’s claim that a vaccine made from that bacillus would be useful in the disease ( 26 ).

THE AGE OF SEROLOGY, 1890–1950 This period was characterized by the development of serum therapy, most famously diphtheria and tetanus antitoxins, the one affecting children, the other soldiers in the field, both powerfully evocative and important to governments. In its train came the network of serum institutes, problems of standardization, and, on the research front, an outgrowth of studies of the nature of specificity and the chemistry of the antigen–antibody reaction, which dominated the field until after World War II. As Frank Macfarlane Burnet realized in 1959, at a time when this era was giving way to another, very largely under his own influence, The subject matter of immunology has often been unconsciously confined to the high-titre antibodies produced by the immunization of horse or rabbit with diphtheria toxin or some other of the classical antigens. Such antisera react with the antigen by aggregation in the test tube and by neutralization of the biological function of the antigen…. Most of the practical applications of

serology make use of such antisera, and all the classical work in immunology is based on their properties… ( 27 ) Animals were immunized at first with live organisms, then as the concept of immunity was generalized, with killed organisms, and later with tetanus or diphtheria toxin. It was found that antitoxic immunity could be transferred via serum to a second individual. Between 1888 and 1894, Emil von Behring and Shibasaburo Kitasato, working in Koch’s Institute for Infectious Disease, laid the experimental foundations of serum therapy ( 28 ). Antitoxin proved itself clinically in cases of diphtheria in the winter of 1892 ( 29 ). Serum manufacture on a large scale using horses instead of the original guinea pigs and rabbits quickly began at the Institut Pasteur, and a global network soon followed: Instituts Pasteur appeared in the main cities of the French colonial empire. European countries followed by Canada set up their own publicly funded institutions dedicated to the production and distribution of therapeutic antisera: the serum institutes with their laboratories, stables and pastures—a horse-centered world—were to dominate medical research in the decades to come ( 30 ). In Germany, four firms—Schering of Berlin; Meister, Lucius and Brüning, then of Berlin, later moving to Hoechst-am-Main; Merck of Darmstadt; and Ruete-Enoch of Hamburg—were licensed to produce antitoxin ( 31 ). They were soon joined by Burroughs-Wellcome of London. After the introduction of serum therapy, epidemics of diphtheria still continued, but the death rate from the disease dropped steeply. Clinical results of serum therapy, however, were unpredictable. Reliable production required measurement: first the dose of immunizing toxin, and then quantification of the horse serum. A standardized antidiphtheria serum was first produced by Paul Ehrlich in the 1880s, when working in Berlin on the specificity of dyes in histology ( 32 ). The unit he devised, the first bioassay, was defined as the amount of antiserum that just neutralized 100 lethal doses (LDs) of a standard toxin. New batches of either toxin or antitoxin were compared with the old standards. The LD 50 was the dose of toxin that was lethal to 50% of a batch of 250-gram guinea pigs within 4 days. The L 0 dose of a new toxin, L standing for limes or limit, was the number of lethal doses neutralized by one unit of the original antitoxin, and the L + dose of the new toxin was the number of LDs just not neutralized. In theory, L + - L 0 = 1 LD, but the difference in practice was always greater than 1, and as a toxin aged, the gap widened. Ehrlich interpreted the stepped neutralization curve as evidence that toxin was composed of a group of discrete but unstable substances, all of which he named. All of them neutralized antibody irreversibly in proportions of simple chemical equivalence, but affected the toxicity of the mixture to different degrees ( 33 ). [T]he reaction between toxin and anti-toxin takes place in accordance with the proportions of simple equivalence…. A molecule of toxin combines with a definite and unalterable quantity of antitoxin. [Ehrlich’s emphasis] It must be assumed that the ability of toxins to bind antibody must be due to a specific atom group of the toxin complex, which shows a maximum specific relationship to an atom group of the antitoxin complex. They fit together like

lock and key, in the image suggested by Emil Fischer for the specific effect of the ferments ( 34 ). It was typical of Ehrlich’s way of thinking that he was prepared to postulate as many different substances as he needed to accord with the phenomena ( Fig. 1).

FIG. 1. Ehrlich’s standardization of the antidiphtheria serum. Added antitoxin has little effect at first, then toxicity falls rapidly, then does not change any further. The relationships change as the toxin ages. Ehrlich sees the toxin as a mixture of different specific substances: he shows four phases in the breakdown of a single sample, “ Gift No. V.” Each phase contains different breakdown products, which are supposed to react irreversibly with the antitoxin, in the manner of the reactions of organic chemistry. They react with the antitoxin in order of affinity; each substance is named according to its relative affinity and toxicity. Active toxins are proto-, deutero- and trito-toxin, in order of affinity; some, the toxoids and toxones, have lost the toxophore group, and are no longer toxic to guineapigs, but still neutralize antitoxin. The vocabulary is Ehrlich’s own invention. From Paul Ehrlich, “Wertbemessung des Diphtherieheilserums und deren theoretische Grundlagen,” ( 34 ).

The Side-Chain Theory of Antibody Production Ehrlich’s vocabulary and his diagrams of the union of antibody and receptor, and his “side-chain theory” of immunity provided the first general theory for the new science of immunology. His system, modeled on a benzene ring with its attached side chains, linked immunity with nutrition. A cell was nourished by capturing nutrients with an array of different side chains, specific to each nutrient, which could be specifically blocked by

toxin. The blocked side chains were shed by the cell, and then replaced by an excess of new ones as the cell repaired itself. The freed side chains were the antitoxins, released in great numbers into the serum ( 35 ). Later workers have pointed out that this implies a selection theory of immunity: Antigen selects the specific side chains to be released by the cell as antibody. Immunologists themselves have recognized Ehrlich’s side-chain theory as a precursor of the clonal selection theory of antibody production, first introduced in 1957 ( 36 ) ( Fig. 2). Ehrlich himself did not pursue the chemistry of the antigen–antibody reaction, or the nature of the specificity that he postulated, except to claim that the reaction was like those of organic chemistry, firm and irreversible. In the early years of the century, he turned his attention to a new project, the development of chemotherapy, which eventuated in the Salvarsan treatment of syphilis in 1909 ( 37 ). But his cartoon-like diagrams of antigen and antibody oriented thinking around the visual metaphor of the receptor, providing a diagrammatic language for immunology that was to persist long after its supposed chemical basis had been dropped ( 38 ).

FIG. 2. Ehrlich’s side-chain theory of antibody production. Antitoxin production is explained as a special case of cellular nutrition. The cell is equipped with side-chains or receptors to capture specific nutrients. A receptor can be blocked by a matching toxin; the cell then heals itself by shedding the blocked receptors and producing an excess of new ones. Some of the new side-chains are freed into the serum and constitute antibodies specific to the toxin. The vocabulary and the diagrams are Ehrlich’s own invention. His conception of the antigen-antibody reaction is of a firm, specific, irreversible chemical binding. He uses the metaphor of a lock and key: another image suggested by his drawings is a snap-fastener or press-stud. On the influence of Ehrlich’s diagrams and vocabulary, see Cambrosio et al. (n. 38 ). From Paul Ehrlich, “On immunity with special reference to cell life,” ( 35 ).

Ehrlich’s work had immense heuristic power. His method of standardization formed the basis of the activities in the serum institutes over the next half-century. It also set off an era of serological reductionism, in which the chemical nature of the antigen–antibody reaction, rather than the resistance of the body to disease was at the center of interest ( 39 ). Opposition to his views stimulated representatives of other types of chemistry to propose alternative interpretations for the stepped neutralization spectrum. These other workers saw toxin–antitoxin neutralization not as a series of discontinuous steps, representing separate irreversible reactions, but as a smooth curve. The curve might represent either an acid-base type of reaction, in accordance with the dissociation theory of the Swedish chemist Svante Arrhenius, or a colloid reaction, according to Jules Bordet of Brussels and the Viennese immunochemist Karl Landsteiner ( 40 ). Both concepts postulated reversible reactions described by smooth curves, not discrete steps. Both allowed for variable proportions of antigen and antibody in the resulting complex that depended on the concentration of the reacting substances. Bordet said that just because twice as much serum is needed to combine with two as with one dose of bacterial emulsion, some bacteriologists argue that antigen and antibody must combine according to a law of definite proportions. That, he said scornfully, was like claiming that paint must react in definite proportions with a wall ( 41 ). The regular chemical law of definite proportions need not apply. Colloid Chemistry and the Template Theory of Antibody Production All known antigens were proteins, and proteins were colloids. In 1912, the chemist Ernst Peter Pick of Vienna made this a slogan: “Kein Antigen ohne Eiweiss” (no antigen without protein) ( 42 ). The first decades of the century were a time of great excitement about colloid chemistry: This was the chemistry of life itself. It was not unexpected to find that this vital, even mystical, reaction did not obey the rules of ordinary chemistry ( 43 ). Landsteiner’s immunochemistry, and his lifelong opposition to Ehrlich and his theories, began in the early years of the century with an attempt to apply the new colloid chemistry to the problem of the relationship between antigen and antibody. Landsteiner argued that specificity could not be absolute: Ehrlich’s pluralistic approach would require an absurd number of specific substances in the serum, whose significance for the animal body was unclear. In Landsteiner’s words, According to the older view [i.e., Ehrlich’s], for every single effect of a serum, there is a separate substance, or at least a particular chemical group…. A normal serum contained as many different haemagglutinins as it agglutinated different cells. The situation was undoubtedly made much simpler if, to use the Ehrlich terminology… the separate haptophore groups can combine with an extremely large number of receptors, in stepwise differing quantities as a stain does with different animal tissues…. A normal serum would therefore visibly affect such a large number of different blood cells,… not because it contained countless special substances, but because of the colloids in the serum, [that is,]… the agglutinins, by reason of their

chemical constitution and the electrochemical properties resulting from it. That this manner of representation is a considerable simplification is clear; it also opens the way to direct experimental testing by the methods of structural chemistry ( 44 ). Landsteiner’s “simpler” view was that specificity was a matter of “more or less good fit,” which he demonstrated through cross-reacting antibodies against a series of compounds of known structure. His key project began during the 1914–1918 war. Conditions were harsh in Vienna; food and heating were inadequate as the city administration crumbled around the researchers and the Donau monarchy came to its end. Many animals were needed for the project, immunized with many closely related antigens. The animals made low levels of antibody because they were cold and undernourished; the same was true of the researchers. But they were able to conclude that it was highly charged groups such as acid radicals that were most important in determining specificity, a finding that brought them closer to structural rather than colloid chemistry ( 45 ) ( Fig. 3). Landsteiner was able to continue with these immunochemical studies of the antigen–antibody reaction at the Rockefeller Institute in New York, where he worked from 1922 to his death in 1946. Landsteiner and his mother had converted to Catholicism in the 1890s. He had left Vienna before the outpouring of anti-Semitism that led to the Anschluss, the unification of Austria with Nazi Germany, in 1938. He was already in New York, when many of the people he knew were desperately trying to emigrate, or to find jobs in a new and difficult country. Although he did not wish to be seen as Jewish, he was able to help some of them.

FIG. 3. Landsteiner’s conception of specificity. The diagram shows a continuous spectrum of reactions to the benzene-sulphonic acid family of antigens. Starting from the immunizing antigen, stepwise small alterations in the chemistry of the test antigens reduce the strength of the reaction with the antiserum. It is the polar groups that have the most effect on specificity. According to Landsteiner, an antibody has a graded quantitative affinity with a range of different antigenic configurations. His conception of the reaction is one of a reversible, weak binding of broad specificity. Compared to Ehrlich’s tightly bound snap-fastener receptors ( Fig. 2 above), Landsteiner’s view of the reaction suggests a silk scarf draped lightly over the charge outline of the antigen. This conception accords very well with the template theory of antibody production of Breinl and Haurowitz of 1930, in which the polar groups of the antigen control the assembly of antibody globulin (n. 47). Karl Landsteiner and Hans Lampl, “Ueber die

Antigen-eigenschaften von Azoprotein: XI Mitteilung über Antigene,” ( 45 ).

The Rockefeller Institute was a placement that was in many ways ideal for the man and for his program of research. It epitomizes in many ways the typically reductionist immunology carried on outside the ambit of the Institut Pasteur. Writers such as George Corner and René Dubos, who experienced life and work in its laboratories, emphasize the role played by reductionist ideals at the Institute. Writing in 1976, Dubos says that the chemical approach is now more dominant than ever in fields such as cellular biology, genetics, immunology, and experimental pathology ( 46 ). In the 1940s at the Rockefeller Institute, there were six different laboratories working on protein chemistry. Landsteiner’s conception of specificity was that antibody draped itself over the charge outline of its antigen. The antigen–antibody reaction was a charge-based surface adsorption. This suggested to the Prague chemist Felix Haurowitz and his serologist colleague Friedrich Breinl, who had met Landsteiner in New York, that antibody formation might take place by the assembly of the globulin molecule on the antigen. The polarity of the antigenic groups served to orient the amino-acid building blocks of the nascent globulin. These concepts were later known as the template theory of antibody formation ( 47 ). Haurowitz fled from Prague to pass the Nazi period in Turkey, and then like Landsteiner, emigrated to the United States, where he settled in Indiana ( 48 ). He maintained his belief in the template theory to the end of his life. The standardization of sera was key to research and theory building in immunochemistry and to the practical problems of serum production and utilization. It was the source of Ehrlich’s side-chain theory, Landsteiner’s countervailing outline concept, Haurowitz’s related template theory, and the lattice theory of the 1930s ( Fig. 4). This last was proposed by the London serologist J.R. Marrack to account for the relation of antigen–antibody proportions to the appearance and disappearance of precipitation, the so-called zoning effect, which made it difficult to titrate antibody and antigen against each other by precipitation ( 49 ).

FIG. 4. Marrack’s lattice theory of antigen–antibody precipitation. Marrack is explaining the zoning phenomenon whereby the precipitation of antigen by antibody depends on

concentration. Zoning, as Ehrlich had found in the case of toxin neutralization ( Fig. 1 above), made it difficult to standardize antisera for practical use. Increasing concentration of antibody leads to a crowding of antibody molecules around an antigen, forcing the polar groups of the antibody into such close contact that they attract each other instead of molecules of water, and precipitate out of solution. (A in the diagram). Differences in proportion of antigen and antibody in the complexes formed (C, D and E) account for differences between precipitates. If the antibody has more than one absorbing site, the complexes may form a large lattice structure (B). From: J. R. Marrack, The Chemistry of Antigens and Antibodies ( 49 ).

In Germany, the state guaranteed standards for the antisera produced there, based on Ehrlich’s technique and on standards held at Ehrlich’s laboratory in Frankfurt-am-Main ( 50 ). This hegemony was broken up by the outbreak of World War I, when other countries such as Britain found that they could not, indeed must not, rely on Frankfurt any more, and began to develop their own programs. Standardization was one of the first projects to be taken up by the new Medical Research Council of Britain. It was placed under the charge of the young Henry Dale, who had briefly studied under Ehrlich in Frankfurt, and had since been employed by Burroughs-Wellcome in serum manufacture. He was also delegated to supervise the testing of Ehrlich’s Salvarsan and its substitutes, whose German patents had been abrogated at the outbreak of war. Interestingly, these toxic chemicals were treated as if they were bacterial toxins, and assayed by Ehrlich’s LD 50 method. It was a method that did not deal with the common problems that accompanied Salvarsan treatment, Dermatitis exfoliativa, and sometimes sudden death, as well as Icterus lueticus, so-called, later shown to be syringe-transmitted hepatitis. At the time, these side effects were thought to be due to excess toxicity of the drug, but the batches always passed the LD 50 test. In 1930, Henry Dale began to suspect that some of the cases of collapse during Salvarsan treatment were due to anaphylactic shock. Anaphylaxis had been described in 1902 by the eugenist Charles Richet in France. By 1913, when Richet received his Nobel Prize, he had come to see it as a mechanism of natural selection, which maintained the purity of races ( 51 ). Allergy and the Clinic Clinically, anaphylaxis was to be carefully distinguished from allergy and its relations, atopic eczema, asthma, and hay fever, although all of them were agreed to be mediated by substances known as reagins, presumed to be cell-bound antibodies. The earliest suggestion of that came in 1921, with the famous personal experiment of the German medical students Carl Prausnitz and Hans Küstner, who tried to exchange hypersensitivities by exchanging serum with each other in an immunological version of blood brotherhood. Both of them were allergic but only fish sensitivity was transferred. There was also the problem of serum sickness, a reaction to antitetanus and antidiphtheria sera, written up by Clemens von Pirquet and Béla Schick in 1905 ( 52 ). In Britain, the first allergy clinic was set up in 1911, an offshoot of the vaccine department of St. Mary’s Hospital under Sir Almroth Wright, following up on Wright’s

enthusiasm for autovaccines. Like Wright’s immunizations, desensitization was both praised and attacked in the popular press, and in the medical journals. It was also a profitable enterprise, funded by a drug company, Parke Davis, which made the sets of allergens used. By the 1980s, the attacks had intensified: The method had had a longer run than most of the Wright-based procedures. But clinical allergists continued to offer desensitization treatments, in spite of warnings from the Committee on the Safety of Medicines in 1986 ( 53 ). Allergy began to take shape in the United States as a clinical specialty in the course of the 1920s, the allergens here being ragweed and poison ivy. Private clinics were set up, societies were organized, and the clinical Journal of Allergy started in 1929. The more laboratory-oriented papers on the subject still appeared in the older Journal of Immunology, but in most cases, the allergists were not laboratory people, and it was felt that they could not come up to the standard demanded by the Journal of Immunology. In the course of the 1930s, the leaders of the profession began to fear that the specialty might gradually become a kind of medical quackery, focused on a single procedure, the skin test. The professional societies determined that clinics should be certified and controlled. In 1971, board certification was set up through a joint effort by the Boards of Internal Medicine and Pediatrics, and in 1973, the American Academy of Allergy was formed in succession to the two national societies. Founders were Robert Cooke, who had asthma attacks triggered by horses and cows, and Arthur Coca, who suffered from migraine and a large variety of food allergies ( 54 ). Coca was to become medical director at Lederle Laboratories, which was marketing sets of allergens for skin test diagnosis and desensitizing treatments, the clinical allergists’ professional standby. Serology at the League of Nations Serology gained still more prestige during the First World War. In the mud of the trenches and battlefields of 1914–1918, tetanus antitoxin strikingly reduced the incidence of tetanus on both sides, but attempts by the German military to develop an antiserum for gas gangrene were not successful ( 55 ). Inoculation against typhoid had become increasingly accepted, and increasingly effective. After the war, the victorious Allies through the League of Nations and its Health Organization set up their own standardization project at the Statens Seruminstitut in Copenhagen. Postwar arrangements bypassed the German laboratories, which were then suffering under a boycott of all international contacts. However, the League’s laboratories under Thorvald Madsen, a student of Ehrlich, still used the German techniques, and Madsen himself tried to make sure that science remained pure, protected from all national and political interference ( 56 ). The League’s program was a microcosm of practical serology. It began by working over the old sera such as diphtheria and tetanus, including the blood group antisera with their conflicting nomenclatures, and then attempted to add new ones of military importance, such as an antidysentery serum. It also worked at standardizing the serological test for syphilis: The Wassermann reaction was a two-stage complement-fixation test of the type introduced by Jules Bordet and applied by August von Wassermann to the diagnosis of syphilis in 1906 ( 57 ). Prodded by the international organizations and the requirements for a standard procedure for syphilis tracking in seamen under the

Brussels Agreement of 1924, the League’s scientists working through the Statens Seruminstitut began on this most difficult of projects ( 58 ). The Wassermann test had been accepted with great enthusiasm by clinical venereologists, but the laboratory workers saw it as unreliable and difficult to carry out. Newer, simpler versions, often based on colloid chemistry, were tested by the League ( 59 ). The most successful of the new colloid tests was probably the test designed by the American Rudolf Kahn ( 60 ). The Kahn test, however, never completely replaced the Wassermann, except in Kahn’s own laboratory. Clinicians continued to ask for “WR and Kahn” on their patients until the late 1960s, when both tests gradually gave way to a more direct form of immunological screening, and finally to the ELISA test using monoclonal antibody ( 61 ). Blood Groups and Transfusion Karl Landsteiner described the human ABO blood groups in 1901 ( 62 ). For many years, however, he showed no great interest in his discovery. Blood groups probably seemed to be rather a dead end in terms of practice, and, possibly, to imply a sharp specificity rather too close to Ehrlich’s for Landsteiner’s comfort. Further work on blood group serology by the Polish serologist Ludwik Hirszfeld showed that they were inherited as Mendelian unit characters, which he interpreted as two pairs of alleles—A and not-A, B and not-B—along the lines of the then-current Mendelian “presence-and-absence hypothesis.” Working at a front-line hospital in Macedonia during World War I, Hirszfeld and the bacteriologist Hanna Hirszfeld, his wife, were able to show that blood group distribution in the military units was linked to the place of origin of the people studied. These two discoveries rendered blood groups significant as forensic tests of paternity, and as race markers, and Hirszfeld himself tried to use them to elucidate the problem of resistance to disease ( 63 ). Felix Bernstein, a mathematician and director of the Institute for Mathematical Statistics in Göttingen, took up Hirszfeld’s study of the inheritance of the ABO groups, then the only normal human trait for which there was enough family data to perform a satisfactory Mendelian analysis. He argued that the data showed them to be controlled by three alleles, all at the same locus, and not by paired alleles at two separate loci. The test case was that of the AB mother: According to his triple-allele hypothesis, an AB mother could not have an O child, whereas with Hirszfeld’s two-locus hypothesis, such children should have been quite common. The literature was combed for cases in point—before Bernstein made his claim, there were quite a few, but as the triple-allele hypothesis took hold, they disappeared from published results ( 64 ). A Nazi-oriented German Society for Blood Group Research founded in 1928 attempted to use Hirszfeld’s results to define the Aryan race and to map its place in Europe. Official Nazidom, however, paid little attention to its findings. The Society excluded Jews from its membership; that meant that none of the leading researchers, such as Hirszfeld, Bernstein, or Landsteiner were members ( 65 ). Bernstein happened to be in the United States when the Nazi edict stripping him of his directorship arrived; he stayed there until after the war, but was never to reestablish himself and wrote no more after 1933. The Hirszfelds were in Warsaw where Ludwik was director of the State Epidemiological Institute; they managed to survive the Warsaw ghetto. Landsteiner himself went back to the blood groups only when he reached the

Rockefeller Institute in 1922. Over the next 20 years, he and his colleagues Philip Levine and Alexander S. Wiener described several more blood group systems, including M-N, P, and finally the rhesus system, establishing not only an expanded forensic tool, but also a causal mechanism for Erythroblastosis fetalis, hemolytic disease of the newborn ( 66 ). In spite of the efforts of the eugenics movement to show that feeblemindedness was inherited as a single-gene Mendelian recessive, blood groups were for decades the only normal human characteristic that was clearly Mendelian in its pattern of inheritance, and where the data were extensive and reliable enough to use in a mathematical approach to human genetics. For the geneticist, blood groups were the human equivalent of Drosophila. In hindsight, one might have expected that blood grouping would have found immediate application as a condition for blood transfusion. But that was not the case. Transfusion itself was experimental rather than therapeutic, and technical problems abounded. George Crile, professor of surgery at Western Reserve Medical College, investigated the technique and its applications in 1909. Although Ludwig Hektoen had cited Landsteiner’s work and had suggested that isoagglutination of human red corpuscles might be relevant to transfusion of blood, Crile’s personal experience had shown that the occurrence of hemolysis in vitro did not necessarily indicate that it would occur in the vascular system of the recipient after transfusion ( 67 ). Crile’s technique of transfer of blood involved the end-to-end anastomosis of the donor’s and the recipient’s veins, by cutting down and suturing the veins together, or joining them with a cannula. Just before the First World War, a method using a paraffin-coated intermediary bottle was introduced; but it was not until sodium citrate was suggested as anticoagulant that any quantity of blood could actually be transferred. Several individuals suggested it at about the same time, but in practice, it was not used until about 1916–1917, when the American Oswald Robertson and the Canadian Bruce Robertson introduced transfusion in the field ( 68 ). Soldiers who were checked out to act as “professional donors” were generally group O, a so-called universal donor. Grouping tests and cross-matching of donor and recipient were felt to take too long to do; group O donors continued to be the mainstay of transfusion into World War II. Only from about 1944, with the increased demand for blood by the Army, group-to-group transfusion began to take over ( 69 ). Landsteiner received a Nobel Prize in 1930 for his 1901 discovery of the ABO blood groups. It came only when it was clear that they had some practical use; he himself thought that his fundamental work on specificity was more important. Until the outbreak of World War II, transfusions were usually organized on an individual basis. In Britain as in Canada, a service was set up through the Red Cross. A donor would be called to the hospital where the blood was needed, as stored blood even when properly refrigerated was felt to be unsafe ( 70 ). Elsewhere, stored blood was increasingly used. In the Soviet Union, the donors attended centers and blood was stored for use as needed. In the United States, Cook County Hospital of Chicago established a blood banking system in 1937, where a credit balance could be used for a given patient without necessarily using the bottle donated by the patient’s own relatives ( 71 ). Blood transfusion, like standardization, was driven by war and the interests of national governments. The technique was recognized as being of national and military

importance following World War I, as military experience showed that blood could virtually resuscitate the dead. The British government, through the Medical Research Council became involved in developing blood transfusion, and the League of Nations in the interests of collective security put the standardization of grouping sera and of their confused terminology on its schedule. At the outbreak of a new war, the Red Cross Blood Transfusion Service became the state-supported National Blood Transfusion Service, with an expanded mandate to prepare and store sera and blood products, such as freeze-dried plasma, to deal with expected civilian casualties ( 72 ). In 1943, two British workers found that the addition of dextrose to the citrate anticoagulant solution made it possible to store refrigerated whole blood for as long as 21 days ( 73 ). This technique was quickly taken up by the Transfusion Service in wartime Britain, but only gradually adopted in the United States. The U.S. military preferred to use dried bovine albumin as an emergency lifesaver in the field. (See below.) The new availability of stored blood was to make possible an era of large-scale surgery, including dialysis and open-heart surgery with extracorporeal circulation ( 74 ). It also facilitated the large-scale spread of serum hepatitis (hepatitis B) among patients, and technical, medical, nursing and cleaning personnel. In Britain, before the introduction of hepatitis testing in the 1960s, approximately 1 in 10 donated units were infected; in the United States, the numbers were higher ( 75 ). Pooled plasma, with material from several donors in a single bottle made things even worse. In a paragraph that looks forward to the AIDS problems of the 1980s, Vaughan and Panton wrote in 1952: Blood and blood products are highly dangerous materials…. False grouping, the transmission of infectious diseases other than jaundice, the use of the proper kinds and amounts of transfused fluid, the serious danger of infected material, can only be dealt with if the utmost care is taken. The prevention of jaundice is still under investigation and this unsolved problem serves as a reminder that blood transfusion is not in its final phase but is still in urgent need of further research…. The advances stimulated by war in this field have had profound repercussions in many fields of civilian medical practice and are likely to have more ( 76 ). A marker for hepatitis B came in 1966, and a vaccine in 1982. Transfusion, in the course of the 1970s, came to be regarded as almost free of risk, on a par with vitamins. The AIDS crisis was to change that.

THE CHEMISTRY OF THE ANTIBODY GLOBULINS, 1930–1960 The history of protein chemistry is a sequence of developments in technology and instrumentation, each technical innovation opening the door to a new series of interpretations. The techniques centered around the separation of the protein mixtures found in nature, the drive to reduction first defining and naming individual proteins, then protein fragments and chains, and finally amino-acid sequences, focusing down on the nature of the antibody combining site. Like the work on the antigen–antibody reaction, work on the proteins originated in

colloid chemistry. In the first decade of the century, Swedish chemists using the ultramicroscope began differentiating inorganic colloids into separate molecules, arguing that colloids were in fact particulate and not hom*ogeneous aggregates ( 77 ). Theodor (The) Svedberg’s life work began with his project on Brownian motion, which he felt demonstrated ad oculos the reality of molecules. He was attacked from all sides, among others by Albert Einstein, but he stuck to his interpretation and in 1926 won a Nobel Prize. His work on proteins began in the early 1920s with the development of his ultracentrifuge, modified from a dairy cream separator. The addition of an oil turbine rotor gave a speed of 40,000 rpm and a force of 100,000 G, and an optical eyepiece made the boundary of the sedimenting material visible. The results persuaded a skeptical Svedberg that proteins too consisted of molecules, and that they had definite molecular weights. He gave each a sedimentation coefficient S that indicated a relative molecular weight, with serum globulin having a coefficient of 7 S, corresponding to a molecular weight of about 15,000; there was also a small amount of a heavier 18 S globulin ( 78 ). In fact, he went further and suggested that all proteins, like hemoglobin, might be aggregates of identical subunits with molecular weights of about 17,000. Joseph Fruton ( 79 ) ascribes this suggestion to the hypnotic power of numerology. I see it as an example of the need to find simple laws underlying complex phenomena, a principle of scientific research prominent in the work of others of the period, for example, Landsteiner ( 80 ). Svedberg saw ultracentrifugal analysis as classical colloid chemistry. Particle size, aggregation, and dispersal in a medium were central colloid problems, but the vitalistic tone of the earlier colloid enthusiasts was soon lost. Svedberg’s 1926 prize attracted enough government funding and Rockefeller grants to set up a new Institute of Physical Chemistry at Uppsala. Here ultracentrifuges of enormous size could be installed in a “building remarkable for its efficiency: no unnecessary, pointless fittings are to be found,” says Arne Tiselius ( 81 ), in a reflection of the contemporary feeling for unity and simplicity in architecture as in science. Tiselius, Svedberg’s erstwhile research assistant, wrote that the workshop with its highly skilled mechanics was “an increasingly important part of the Institute, for in many investigations, the building of the apparatus is perhaps the most important factor” ( 82 ). Several apparatuses were built for export to the United States and Britain. Tiselius himself followed in his senior’s footsteps by developing another piece of industrial-sized equipment, the electrophoresis apparatus, based on a small-scale apparatus designed by Landsteiner and the colloid chemist Wolfgang Pauli. His apparatus was designed to separate serum proteins by charge, rather than by molecular weight. Tiselius too felt that the study of electrokinetic phenomena was among the most important tasks of colloid chemistry ( 83 ). Lily Kay ( 84 ) has suggested that these large, complicated and very expensive pieces of equipment generated their own research programs as they diffused from Uppsala to other centers, beginning with the Rockefeller Institute in New York. Andrew Ede ( 85 ) goes further and suggests that colloid chemistry itself was a product of the original dialysis apparatus of 1849, the semipermeable membrane that separated colloids from crystalloids. These historians have put their finger on a feature that has been of singular importance in protein chemistry from its mid–19th-century origins to the mid–20th-century work that elucidated the structure of antibodies. At each stage, exploitation of a new separation technique revealed a broad new landscape for the

explorers. Some, like the ultracentrifuge, were products of heavy industry, requiring factory-like laboratories. Others, like the starch-gel electrophoresis setup, were so simple that they could be made at home. All of them contributed to the drive to reduction: Every separation made the fragments smaller, until the smallest possible came into view, and with them, the secret of antibody specificity. The earliest serum fractionation method was “salting out,” by the addition of neutral salts, a technique dating to the mid-19th century, and still in use today for large-scale rough or preliminary separation of a bucket of serum. Here the antibody activity went down with the globulin fraction, leaving the albumin in solution. As late as 1930, however, it was still being argued that antibodies might not actually be globulins; they might simply be precipitated along with the globulins ( 86 ). In 1938, Tiselius collaborated with Elvin Kabat of Columbia University on the fractionation of immune sera, with the significant result that antibody was finally linked to the globulin fraction, which could be seen to separate into three bands, named by the discoverers the a, ß, and ? globulin bands. Antibody activity was located in the ? band ( 87 ). By 1945, the first commercial version of the electrophoresis apparatus had appeared, cost and size were coming down, and the importance of the technique growing as it became more commonplace. The tradition of large-scale fractionation was well established in Uppsala, but these were still analytic rather than preparative techniques. The new methods of the 1940s and 1950s allowed for the preparation of batches of material. Edwin Cohn’s Plasma Fractionation Project, centered during the World War II in his laboratory at Harvard Medical School, reoriented his research on problems of protein structure to the large-scale preparation of plasma fractions for use in battlefield emergencies. Where others had organized a blood transfusion service or used whole plasma, the United States preferred albumin. Cohn’s method was to isolate the albumin from bovine blood by fractionation with alcohol, not by salting out, and to freeze-dry it using a new commercial technique. Purity was guaranteed by inspection of the fractions in the Tiselius apparatus. He was later to change to human serum albumin, as bovine albumin could produce serum sickness. Serum globulin was a useful by-product: It could be used clinically as convalescent serum was used in childhood diseases, that is, as a source of antibodies against common infections, especially hepatitis. Cohn developed a small portable fractionator that could be attached to a donor’s arm for plasmapheresis and the preparation of hyperimmune globulin. Angela Creager ( 88 ) has opened up an interesting pathway here in her studies of Cohn and his practical methods. The name “chromatography” was introduced by the Russian botanist M. Tswett ( 89 ) in 1903 to describe his trick of separating colored plant materials by allowing a drop of the mixture to spread on a piece of blotting paper, producing concentric rings of distinct color. A.J.P. Martin ( 90 ) and his group at St. George’s Hospital in London worked out the first good chromatographic method during the 1940s ( 90 ). They used a filter-paper sheet held vertically as the adsorbent, and allowed the test substance in solution to creep slowly upwards by capillary action, separating into smudges as it went. The separation could be made two-dimensional by turning the paper on its side and dipping it in another solvent, or using an electric current. The method was called “fingerprinting.” It separated mixtures of differently charged peptide fragments, opening the way to

protein genetics. After the war, a new type of chromatography was worked out by Stanford Moore and William Stein at the Rockefeller Institute in New York ( 91 ). The adsorbent matrix this time was an insoluble resin, either acidic or basic, packed into a vertical column. A solvent carrying the mixture trickled down through it, leaving the oppositely charged components attached to the resin, while those with similar charge passed through unhindered. This ion-exchange chromatography was very effective as a preparative procedure, particularly when linked up with the automatic fraction collector that Moore and Stein designed that allowed the experimenters to run their columns overnight and read the results in the morning. But the rough treatment of the protein often resulted in the disappearance of biological activity, “lost on the column.” The lab workers were proud of their sensitivity to the delicate treatment needed for the preservation of antibody. Rough handling that produced foaming often diminished antibody titer, or destroyed it altogether. The “molecular sieve,” another preparative technique from Uppsala, was developed by Jerker Porath in 1960 in collaboration with the Swedish firm Pharmacia. It consisted of a column of Sephadex TM, a cross-linked dextran gel. Separation of the protein mixture was by molecular size, due probably, Porath thought, to steric hindrance as the molecules straggled through the maze of pores in the white fluffy gel ( 92 ). Its mate was the updated, and less destructive, ion-exchange method of column chromatography developed by the American Herbert Sober and his group in 1956 using charged forms of cellulose (diethyl aminoethyl or DEAE cellulose, and carboxymethyl or CM cellulose) and a buffer gradient ( 93 ). Here again separation of proteins and protein fragments was by charge. These two methods were complemented by the practical addition of an elegant Swiss-made automatic fraction collector. Starch gel electrophoresis is a kind of counterexample to the power of the huge machines in creating and controlling their own program of research. A starch gel system could be set up by anyone with a flair for cutting perspex sheets neatly, and boiling up powdered starch and buffer solution in a beaker ( 94 ). It cost virtually nothing, and could be shown to a visiting worker in a few hours. Like paper electrophoresis, it became a favorite in both research and clinical laboratories, partly perhaps because it allowed workers to feel very skilled and sensitive in controlling a simple apparatus that they had made themselves. Technically, starch gel stabilized confusing convection currents and combined charge and sieving properties in one. The thick gels could be stained and desiccated to form thin transparent films that were easy to photograph and store. It was a very effective means of separating complex mixtures such as serum proteins and protein fragments, and detecting genetically determined variants. Oliver Smithies ( 95 ) of the Connaught Laboratory in Toronto in his original paper of 1955 reports just such a finding. One result of the separations was the increasing resolution of globulin types. IgG, IgM, and IgA were distinguished. The question of the nature of reagins was finally solved by the Ishizakas in 1966 as being none of the above, but a new globulin type that they named IgE ( 96 ). The time had come to standardize the nomenclature of the

immunoglobulins. As the League of Nations under Madsen had done in the 1920s and 1930s, the World Health Organization under Howard Goodman applied its immunodiplomacy to come to a general agreement on terminology ( 97 ). It was not easy—I have been told that a scientist would rather use someone else’s toothbrush than their terminology. The accumulation of separation techniques now made it possible to work with fractions of serum and fractions of molecules. The British biochemist Rodney Porter working at the National Institute for Medical Research used DEAE to prepare a sample of immune globulin from whole rabbit serum, then digested it with the proteolytic enzyme papain, and separated the fragments on the ultracentrifuge. There was only one peak. His first stab at globulin structure followed the existing view of it as a long single chain folding on the antigen as a template, as Linus Pauling and Felix Haurowitz had taught ( 98 ). His next attempt involved opening the disulfide bonds of the molecule, adding the Sephadex column to his series of preparations. This produced a suggestion of two pieces, one light and one heavy, which would have normally been joined together by disulfide bonds ( 99 ). The relation of the pieces produced by opening the S–S bonds to the papain pieces was worked out immunologically, using goat precipitating sera raised against what now seemed to be two different fragments of the rabbit globulin. Porter then proposed a second model of the globulin molecule: a pair of heavy chains joined by disulphide bonds, each with a light chain attached, and with the antibody recognition site on papain fragment I, probably on the heavy chain. The model stood up well when new findings accumulated. Different types of heavy chain were found in different classes of immunoglobulin, and the light chains showed genetically determined polymorphisms ( 100 ). Recognition was prompt—in 1968, Porter was awarded the Karl Landsteiner Memorial Award, and in 1972, he shared a Nobel Prize. Porter’s Y-shaped model of the globulin molecule has come to be the symbol of immunology. Myeloma Proteins—A Model System Porter’s model was built up on normal globulin fragments, with their heterogeneous collection of specificities and chain types. In 1965, the protein chemist Frank Putnam, then at the University of Florida could still write: The ?-globulins lack all the prerequisites needed to facilitate study of their primary structure. They are heterogeneous, noncrystallisable and not resolvable into pure components; they are antigenically diverse but share common determinants; many possess… biological activity, but the site of activity has not been defined…. Yet the key question in immunology and protein biosynthesis today still hinges on the determination of whether antibodies of different specificity—or for that matter, antibodies of the same specificity—differ in amino-acid sequence ( 101 ). A new opening was found when it turned out that the abnormal serum protein produced in such quantities by patients with multiple myeloma was a ?-globulin, but unlike the normal serum globulin, each one was absolutely hom*ogeneous and could be resolved into pure components. Some of these globulins had antibody activity. With Smithies’

starch gel separation technique, M.D. Poulik and Gerald Edelmann of the Rockefeller Institute had found in 1961 that if myeloma protein was reduced and alkylated, and the fragments separated, the pattern of components duplicated those of normal globulins, except that the bands that separated on starch gel were very much sharper ( 102 ). The Bence–Jones protein from the same patient’s urine matched the bands for reduced and alkylated fragments of the parent protein, and corresponded to free light chains ( 103 ). Myeloma proteins had antibody activity for a variety of antigens, but they did not have the heterogeneity of the normal protein, raising the possibility that they could be used for detailed studies of antibody chain structure. Human myeloma cells were difficult to grow in tissue culture, but Thelma Dunn, Ragna Rask-Nielsen, and Michael Potter found a way of growing mouse myelomas by transplanting them into inbred mice. Each myeloma derived from a single clone of cells, and produced a single hom*ogeneous globulin. As Michael Potter ( 104 ) remarks, these were cancer workers for whom the idea of a tumor originating from a single cell was not new—the clone maintained its uniqueness through all its transplants. Many myeloma proteins were found later to have antidinitrophenyl activity, perhaps as a result of a cross-reaction with some gut antigen. Separations and amino-acid sequencing showed that all light chains of a given type shared a constant sequence of amino acids at one end, but were variable at the other, suggesting that antibody specificity was a result of a specific amino-acid sequence. Parallel with the information that was building up through the 1950s on the sequence of amino acids in protein chains, there came evidence that the sequences appeared to be genetically controlled. Each amino-acid link in the chain was coded for in the nuclear deoxyribonucleic acid (DNA) and the coding transferred to a messenger ribonucleic acid (RNA), and transcribed as an addition to the chain. The system seemed to be strictly directional—the product could not affect the messenger. No protein could be formed by copying another. Francis Crick of Cambridge called this the “central dogma” of protein synthesis ( 105 ). If it was substantiated, it meant that the template theory of antibody synthesis could not stand. But the theory had deep roots and powerful supporters. American immunochemists such as Linus Pauling, Michael Heidelberger, and Felix Haurowitz, now in Indiana, still held to it. Haurowitz felt that fingerprinting had shown that globulins were all almost alike in sequence, and that even if the sequence was genetically determined, which he doubted, the folding of the chain might still depend on antigen. As he wrote in 1963, It is imaginable that the interference of a template with the folding pattern may affect the sequential pattern and thereby prevent or favour the incorporation of certain amino-acids into particular geometric patterns of the three-dimensional conformation of the globular molecule. The immunochemical observations show quite clearly that not all information required for biosynthesis of proteins is supplied by nucleic acids, and that proteins and other substances may act as templates. Life may then be more than merely the “expression of the chemistry of nucleic acids” ( 106 ). Haurowitz was never to surrender. In fact, he felt that James Watson and Francis Crick’s idea of the replication of DNA strands one from another was something that Watson had picked up while attending his, Haurowitz’s, classes. The idea of a template

for protein synthesis, disconnected from antigen, continued to appear from time to time like a ghost ship. Marshall Nirenberg ( 107 ), in his essay on protein synthesis of 1965, refers to messenger RNA and synthetic polyribonucleotides as “highly active templates,” directing amino acids into nascent proteins.

CELLULAR IMMUNOLOGY AND THE SELECTION THEORIES 1950s TO 1980s Antibiotics, beginning with the strategically important drug penicillin, came in with World War II. At first, it was a secret weapon reserved for the armed forces ( 108 ). New vaccines such as the polio vaccines of the 1950s, famously funded by the American charity March of Dimes, still appeared ( 109 ) (along with a revived antivaccinationist movement [ 110 ]) but the serological treatment of disease had lost its edge. In spite of all its successes, compared with the hopes raised by antibiotics, serology no longer seemed so powerful. Immunochemistry reached a climax with the Porter model of immunoglobulin of the 1960s, but from the 1950s onward, mainstream thinking in immunology became steadily more biological and less reductionist. Few immunologists were interested in both biology and chemistry. For instance, at the first meeting of the International Congress for Immunology, held in Washington, DC in 1971, cellular and chemical sessions ran simultaneously, making it impossible for adepts of either to attend the other’s sessions ( 111 ). It was not the “central dogma” that turned immunologists away from the template theory of antibody production, but the new interest in immunologically competent cells and immunized animals. The leading thinkers of the period, especially the Australian Sir Frank Macfarlane Burnet, saw themselves as biologists and drew on the ideas of contemporary biology, not chemistry. Burnet, a virologist with a childhood love of natural history and of Charles Darwin, took up the directorship of the Walter and Eliza Hall Institute in Melbourne in 1944. His colleague and successor, Sir Gus Nossal, remembered Burnet as having a fundamentally negative attitude to technology: Of course, Burnet was in many ways deeply correct to be mistrustful of technology. Sometimes scientists center their lives around an instrument, they become experts at running an electron microscope, ultracentrifuge or some more sophisticated piece of apparatus until they become prisoners of the instrument and cease asking deep, fact-finding questions. Burnet was wary of that behaviour ( 112 ). In the 1960s, molecular biology was growing exponentially. It had already made its mark on immunology, first through immunochemistry and then through the proliferation of protein separation techniques. Standardization, reduction, and the ideal of molecularization had ruled immunology for decades, often in advance of the effective reduction of other biological sciences, except perhaps the pharmaceutical industry ( 113 ). But Macfarlane Burnet was uncomfortable with biochemistry and its sophisticated equipment, and he discouraged it in his institute. Under his leadership immunology moved in a different direction from most contemporary science, as it rejected

reductionism and returned to the level of the immune animal and the cell. Cellular immunology began in the late 1930s with the attempt to show that skin sensitivity to simple chemicals was due to antibodies or reagins ( 114 ). There was already a tradition of work on skin lesions, usually associated with infections—the tuberculin reaction was first mentioned by Robert Koch in 1891 ( 115 ). According to the Vienna pediatrician Clemens von Pirquet, this was the same reaction that followed smallpox vaccination and other skin infections ( 116 ). Landsteiner’s artificial diazo-protein antigens provided the model: The diazo group would link to body protein, and stimulate antibody production, and hence skin contact sensitivity to the antigen. The elderly Landsteiner and his young colleague Merrill W. Chase at the Rockefeller Institute struggled with serum transfer experiments. Antibody could sometimes be found. Chase assumed that it must be cell bound, since contact sensitivity was not usually transferable by cell-free serum. The same was true of tuberculin hypersensitivity, also known as delayed hypersensitivity. That was usually contrasted with immediate hypersensitivity, in which a small amount of antigen introduced into the skin of a sensitized animal produced local swelling and redness within a few minutes. It could be transferred by serum from the sensitized individual to the normal. Histologically, immediate and delayed hypersensitivity seemed similar: Immediate hypersensitivity or local anaphylaxis was an acute inflammation, with edema, polymorphonuclear leukocytes, and a few lymphocytes, lasting roughly 24 hours. The tuberculin or delayed reaction was slower to develop, with many more lymphocytes and macrophages. It formed a solid red lump on the skin, often breaking down to a black, necrotic center and healing very slowly ( 117 ). It was only when some of the exudates that Landsteiner and Chase were using for transfer were incompletely cleared of cells, that it began to seem as if the transfer of skin hypersensitivity was mediated by the cells, not the serum. A serum factor, called “transfer factor” by New York immunologist Sherwood Lawrence, was mentioned in Chase’s review of 1965, but it sounds from the text as if he did not believe in it. At the time, no one did. As Lawrence said in 1986 ( 118 ), after his factor seemed to have been justified, there was a subtle irony here—the emergence of cellular immunology as a scientific discipline was ushered in by the cataract of soluble factors it released. Graft Rejection and Tolerance The key practical problem of the period was graft rejection, which along with blood transfusion, was important in wartime. Sir Peter Medawar, a professor of zoology at University College, London, made his first attempts at grafting patients with burns in 1943. Comparison of the survival of grafts of the patients’ own skin and skin from donors led him to suggest a genetically determined immune rejection mechanism. Like Landsteiner and Chase, he thought first of antibodies, a system like the blood groups perhaps, with “at least seven antigens” involved ( 119 ). As the surgeon Joseph Murray proved in 1954, between identical twins in the absence of an immunological barrier, a renal autograft could function permanently. He and the urological team at Peter Brent Brigham Hospital in Boston had bypassed the immunological problem, but at the same time, demonstrated its importance. Murray said that organ transplantation revitalized

immunology ( 120 ). Discussion centered on tolerance. In 1949, Frank Fenner and Macfarlane Burnet introduced the concept of self–not self discrimination by suggesting that tolerance for a range of self-markers developed in fetal life ( 121 ). They were able the cite the natural experiment of cattle twins, where exchange of blood precursor cells had taken place in utero, and gone on producing genetically foreign cells throughout life ( 122 ). Burnet’s own attempt to induce tolerance failed, but the demonstration was carried out by Medawar and his colleagues Rupert E. Billingham and Leslie Brent, all three of them zoologists by training and practice. Tolerance could be induced experimentally in embryos, and it persisted after birth ( 123 ). Brent ( 124 ) said that it was only in the mid-1950s that the community of immunologists accepted that their work was relevant to the mainstream, and that they themselves began to regard themselves as immunologists. Burnet ( 125 ) saw the interest in the vital phenomenon of tolerance as one more justification for his view that the new immunology should be biological and not chemical. Tolerance in theory did not solve the clinical problem of graft rejection, which Leslie Brent has called the “search for the holy grail.” A temporary solution was achieved by Byron Waksman and his group with antilymphocyte serum, which suppressed delayed hypersensitivity and acute hom*ograft rejection, though an antiglobulin was soon formed against it ( 126 ). The importance of this work for skin and organ grafting is demonstrated in a peculiar way by the episode of the spotted mouse. William Summerlin, a young researcher at the Sloan–Kettering Institute in New York claimed in 1973 to have achieved a take of grafts between nonsyngeneic mice by culturing the graft cells before setting them. Medawar and colleagues and others tried and failed to replicate the results, and it appeared later that they had been faked. Medawar suggested that perhaps one such graft had taken—perhaps due to a mix-up of mice—and its importance was such that its author could not admit that his result was unrepeatable ( 127 ). The first immunosuppressive drugs turned up initially in the early 1960s as antimitotic agents tested as chemotherapy for cancer, and it was the combination of these with corticosteroid hormones that finally made transplanted organs the commonplace they now are. The next generation of immunosuppressants was based on the cyclosporines, antilymphocytic agents first extracted from the fungus Trichoderma polysporum. They were detected in the laboratories of Sandoz in Basel, in the course of a broad pharmacological screening project that seems to have included all known fungi. Hartmann Stähelin ( 128 ) in discussing this sees it a serendipitous discovery, but it sounds more like the empiricism that once was the preferred program of science. The success of organ grafting depended on the construction of a network of centers carrying waiting lists of patients ready to be correlated with available organs, so that a cadaver organ packed in ice could be rushed to a patient as quickly as possible. Initially, the patients were tissue typed, and organs sought that most nearly matched the antigens on the patients’ leucocytes, the histocompatibility antigens. The mixed

lymphocyte reaction, in which a culture of cells from two genetically different sources responded to each other’s histocompatibility antigens, could be used as a kind of cross-match of donor and patient on the blood transfusion model. A good match improved the survival of the graft, but perhaps not enough, it was argued, to justify the longer period that a patient would have to spend waiting for a matched transplant. From the early 1980s on, the use of cyclosporin improved graft survival so much that typing now seemed less important. The Clonal Selection Theory The insistent question of the generation of antibody diversity went through a biological metamorphosis too. The idea that the sequence of amino acids in antibody globulin, or at least the folding of the chains, was directly molded on antigen had satisfied a generation of chemists and serologists brought up on Landsteiner’s charge outline and more-or-less good fit picture of the antigen–antibody reaction. In 1955, however, Niels Kaj Jerne of the Statens Seruminstitut in Copenhagen had just finished a thesis on the old serological problem of antibody avidity ( 129 ). He now suggested something strikingly different as a theory of antibody production. He proposed that all possible specificities were randomly present in the serum as spontaneously synthesized natural antibodies, and antigen selected—not synthesized—its match from among these natural antibodies. Jerne called this a theory of natural selection. The antigen–antibody complex would be taken up by phagocytic cells, which would be “signaled” to reproduce that same antibody; the antigen, freed from its complex, could go back into circulation and do the same again. For Jerne, it is the antibody, not the antigen, which acts as a template. The crucial point of the natural-selection theory is the postulate that the introduction of antibody molecules into appropriate cells can be the signal for the production of more of their kind. This notion is unfamiliar. However, as nothing is known about the mechanism of antibody synthesis in a cell, it would seem a priori more reasonable to assume that an animal can translate a stimulus, introduced by protein molecules which it has itself at one time produced, into an increased synthesis of this same type of molecules, than to suppose that an animal can utilize all sorts of foreign substances and can build them functionally and semi-permanently into the most intimate parts of its globulin-synthesizing cells ( 130 ). He suggests that the antibody protein can act as a template for the order of nucleotides in the synthesis of RNA, which in turn acts as a template for more of the same protein. Natural selection, he thinks, could account for the increased avidity of antibodies produced later on in the course of immunization. But as Thomas Söderqvist, Jerne’s biographer, has said, for Jerne the expression “natural selection” had only the very faintest of Darwinian overtones ( 131 ). Jerne’s paper initiated a renewed discourse on antibody production. In 1957, two years after Jerne, David Talmage of the University of Colorado in a general review of immunology and its current problems and uncertainties, compared the theories with the available data. He recognized that the template theory was very widely held at the time. But it was beginning to seem strained. Jerne had shown in his thesis that antibody

avidity increased during the secondary response when antibody was being most rapidly produced. Logically, increased avidity should have slowed release of new antibody from an antigen template. And Burnet in 1949 had found a logarithmic rise in antibody production, which suggested that antibody was being produced by something that was replicating, not just being recycled. Burnet, said Talmage, complains that a gap has grown up between immunology and existing knowledge of biology. He has pointed out that nowhere else in nature was there anything analogous to an antigen template. In fact, Burnet and Fenner were already stating programmatically in 1949 that they preferred to approach the problem “on biological rather than chemical or pseudo-chemical lines.” It was a strange and surprising statement. The template theory’s inventor, the chemist Felix Haurowitz, asked in a review of The Production of Antibodies, How can they hope to explain molecular phenomena taking place between molecules of the antigen, the antibody and possibly other substances, without invoking the principles of chemistry?… The words put by the reviewer in quotation marks [those quoted above] demonstrate that Burnet and Fenner use the strong language of men who know they are right ( 132 ). Burnet at the time had a theory of antibody production linked to the enzyme induction well known in bacteria, which could adapt themselves to growing on different substrates. But it was soon abandoned. Talmage welcomed Jerne’s natural selection idea in that it offered an alternative to the template theory, but he proposed a major modification. He saw that it harked back to Ehrlich, and he suggested that, as in Ehrlich’s theory, the recognizing antibody might be on a cell, rather than in the serum. This cellular version of the hypothesis would fit better with the long-continued production of antibody without any further stimulus, and with current views of protein synthesis. It took account of the transfer of active immunity by cells rather than by serum, as serum tended instead to suppress the immune response. Jerne explained by saying that an animal must be able to distinguish between its own globulin and that of another of the same species, or perhaps the globulin was somehow damaged in the course of the transfer. Talmage picked up the Darwinian suggestion in Jerne’s title and gave it a more literally Darwinian content: The process of natural selection requires the selective multiplication of a few species out of a diverse population. As a working hypothesis it is tempting to consider that one of the multiplying units in the antibody response is the cell itself. According to this hypothesis, only those cells are selected for multiplication whose synthesized product has affinity for the antigen injected. This would have the disadvantage of requiring a different species of cell for each species of protein produced, but would not increase the total amount of information required of the hereditary process ( 133 ). In the same year, after reading Talmage’s critique, Burnet also modified Jerne’s theory ( 134 ). It has been said that he purposely sent the paper to a modestly circulated journal, just in case it was embarrassingly dismissed by his colleagues. If that was so, this uncharacteristically tentative approach was soon dropped. Ten years later, he was to

write confidently: It gradually dawned on me that Jerne’s selection theory would make real sense if cells produced a characteristic pattern of globulin for genetic reasons and were stimulated to proliferate by contact with the corresponding antigenic determinant. This would demand a receptor on the cell with the same pattern as antibody and a signal resulting from contact of antigenic determinant and receptor that would initiate mitosis…. Once that central concept was clear, the other implications followed more or less automatically… ( 135 ) Haurowitz had understood him very well: Burnet was a man who knew he was right. Burnet defended his idea in The Clonal Selection Theory of Acquired Immunity of 1959. It is the clonal part that he is at pains to argue. There are many other examples of clones in biology; his models come from bacteriology and cancer research. The spread of the lethal myxomatosis epizootic in Australia, introduced in 1950 to control the rabbit population, provides one model. Here distinct clones of less virulent forms of virus multiplied to become the dominant, keeping the virus circulating as an epizootic. Mutation and selective survival were able to change the character of a population of cells ( 136 ). Another model is multiple myeloma: I hope it is not overstating the case to say that the multiple myeloma findings provide the best possible material for displaying the salient features of the clonal selection approach to the phenomena of antibody production and of malignancy…. The fact that each myeloma patient produces his own characteristic and individual serum protein, with its sharp spike evidence of hom*ogeneity, provides support for what many workers might consider a weak point of the clonal selection theory, that each clone produces a specific antibody globulin whose pattern is genetically determined ( 137 ). Like Burnet himself, contemporary commentators on Burnet’s new approach underlined his Darwinism. According to Gordon Ada and Sir Gustav Nossal, younger colleagues of Burnet at the Walter and Eliza Hall Institute in Melbourne, for Burnet, the immune response was a Darwinian microcosm. Lymphocytes were the individuals in a particular ecological niche, mutating and being selected, like the myxomatosis virus. The fittest, in this case the variant that made rabbits sick but did not kill them outright, survived and kept the epidemic going. In the same way, the cells that made the fittest antibodies, those with the best fit to antigen, multiplied the most. In a Darwinian system, adaptation was not imposed from outside, but was favored by the multiplication of the best adapted ( 138 ). Burnet himself called the template theory “a grossly Lamarckian qualification on what might be described as a strictly Darwinian process at the cellular level” ( 139 ). He did not put his argument in terms of the so-called central dogma of molecular biology. In 1957, soon after the publication of his first statement of the theory, Burnet called his staff together and announced that the whole direction of the Hall Institute would change from virology to immunology. He saw virology as rapidly coming under biochemical influence, which, he said, he “preferred to eschew.” He realized that his theory was the

foundation of a fundamental change in immunology, and the Institute was to work out its implications ( 140 ). Burnet had predicted that each cell would make only one antibody. The first experimental testing was done at the Hall Institute in 1958. Joshua Lederberg, who had arrived in Melbourne hoping to work on virology, joined up with Gus Nossal to develop a micromanipulation system where individual cells could be tested separately. Using rabbits immunized with two different antigens, they found that among 456 cells isolated, 62 made antibody, and each of them made one antibody only ( 141 ). Their system was difficult to reproduce, and only the authors and one or two others were truly able to handle it. But their result was confirmed by Jerne, who invented the ingenious and simple hemolytic plaque technique, something that could be easily learned from his published description ( 142 ). Other findings accumulated that made the template theory less likely, although there was in fact no final disproof. But as Talmage has said, the final acceptance of a theory only comes with utility ( 143 ). The development of hybridoma technology by Georges Köhler and César Milstein (see below), and the commercial production of monoclonal antibodies finally made the clonal selection theory irresistible ( 144 ). The Biology of the Thymus and the Dictatorship of the Lymphocyte As Burnet’s views gained acceptance in the early 1960s, new work focused on populations of cells. The new theory released an avalanche of work. It coincided with the expansion of U.S. funding for science that followed the end of World War II. Clinical applications of cellular immunology included autoimmunity and transplantation surgery. Pharmaceutical companies, until then focused on vaccines and sera, began to develop and patent immunosuppressants, down-regulators of immunity. In the 1970s, with a new field to till the profession expanded, as journals proliferated, congresses national and international were initiated, and symposia and courses organized. The Soviet immunologist Rem Viktorovich Petrov ( 145 ) called this the period of the dictatorship of the lymphocyte. Before the theory, lymphocytes had no known function. “Round cell infiltration” was pathologist’s shorthand for reporting the nonspecific in a tissue section. Now, however, lymphocytes were seen as long-lived cells recirculating through the body’s lymphatic tissue and carrying immune recognition and memory, including the recognition of self. A new and revised anatomy gave a central place to the thymus, which until then was an organ whose histology was described in enormous detail, but whose function was completely blank ( 146 ). At this point, a link-up was made between the activities of cells and two much older fields within immunology, the tuberculin reaction and bacterial or delayed hypersensitivity. Each of these had been the product of a different technique, and had been investigated initially in a different context, but they now came to overlap each other. With the new emphasis on the cell, they appeared in a new light. Old cells-versus-serum controversies resolved as it appeared that T-lymphocytes, developing or maturing in the thymus, mediated cellular immunity, and interacted with

B-lymphocytes from the bone marrow, producers of serum antibody. Prepared mice came to be seen as the experimental system of choice—in studies of thymus function and of tolerance, the system was based on the neonatally thymectomized mouse ( 147 ). Indeed, the elucidation of thymus function depended on the mastery of the difficult technique of neonatal thymectomy: Jacques Miller at the Chester Beatty Research Institute in London found in 1961 that his thymectomized mice had fewer lymphocytes, made no antibody, and tolerated skin allografts. Other laboratories were close behind him. Here the inspiration was at least in part clinical, as the pediatrician Robert A. Good and his students at Minnesota worked through a family of patients with an X-linked absence of antibody globulins, and found that they all lacked plasma cells. They suffered from recurrent infections, but not from tuberculosis—cellular reactions seemed intact. Good could not reproduce the syndrome with thymectomized rabbits, but he and his students Bruce Glick and Timothy Chang found serendipitously that chicks that had had the bursa of Fabricius removed failed to make both antibody and plasma cells. In birds, they decided, the antibody side seemed to be controlled by the bursa, separately from cellular reactions. If thymectomy was carried out early enough in mice, in “hot little newborns,” as Good called them, all immune reactions failed. This reproduced another clinical syndrome, that of the so-called “bubble boy,” who spent his short life enclosed in a germ-free plastic bubble ( 148 ). Other laboratories were all on the same wavelength: Jacques Miller at the Chester Beatty Hospital; Byron Waksman at Harvard; and Delphine Parrott in John Humphrey’s laboratory at the National Institute for Medical Research, Mill Hill. All worked to replicate the clinical syndromes of immune deficiency with thymectomized mice. As immunobiology replaced immunochemistry in the mainstream, the laboratory turned to the inbred mouse as its key instrument ( 149 ). Ilana Löwy and Jean-Paul Gaudillière see genetically hom*ogeneous mice as the equivalent of standardized, chemically pure compounds, produced on an industrial scale ( 150 ). By 1962, syngeneic mice were found to be capable of acting as a cell culture for transplanted clones of mouse myeloma cells, providing a library of monoclonal immunoglobulins for investigation. Monoclonal Antibodies Myelomas were potentially immortal in cell culture, but only about 5% of naturally occurring myeloma proteins had detectable antibody activity. Normal antibody-producing cells, on the other hand, quickly died out in culture. In 1975, this picture changed with Georges Köhler and César Milstein’s fusion of myeloma cells with antibody-producing B cells from a mouse spleen, to produce immortalized cells that secreted monoclonal antibodies of any desired specificity ( 151 ). One lymphocyte clone produced one antibody. It was the epitome of the clonal selection theory: Milstein’s articles include the experimental diagram showing a mouse with a syringe as the source of cells that has typified all cellular immunology since Burnet. But Milstein was an immunochemist, and he saw his invention as answering the questions left behind by the pre–World War II generation of chemists; he cited Ehrlich’s introduction of the problem of diversity and specificity, and Marrack’s lattice theory. One of his examples of a useful application is a superspecific anti-A, able to detect A 2 B, an old blood-group serologist’s

problem. His true interest was not practice, however, but what he saw as the “more fundamental” use of monoclonal antibodies to define and characterize the antigenicity of cell membranes. Like Landsteiner, he was uninterested in the merely useful. In the British tradition, and encouraged by the Medical Research Council, Milstein and Köhler refused to patent their invention. They had received mouse plasmacytoma cells from Michael Potter of the National Institutes of Health in Bethesda, Maryland, and they gave them away freely ( 152 ). But patents were quickly taken out by others—in 1979 for monoclonal antibodies against tumor cells, and 1980, for antibodies to viral antigens, in both cases including Hilary Koprowski of the Wistar Institute in Philadelphia among the patent holders. Legal struggles over the patents and the nature of the innovations patented were fought out through the courts by rival pharmaceutical companies. Milstein and Köhler won their Nobel Prize for this work in 1984, shared with Niels Jerne. The prize was for “a methodological breakthrough that has profound practical significance,” in the case of Milstein and Köhler, and “for theoretical advances that have shaped our concepts of the immune system,” in Jerne’s. Reporting on the prize, the immunologist Jonathan Uhr seemed to feel that the latter was of much more significance ( 153 ). Two of the prizewinners, Jerne and Köhler, were from the Basel Institute for Immunology. It had been funded by the pharmaceutical company Hoffman–La Roche as a vehicle for Jerne, and was the leading center for the fusion of the cellular style with molecular biology. Its reign lasted from 1969 to 2001, when the company decided to close it. Jerne retired in 1980. Writing in 1997, Leslie Brent ( 154 ), an expert on transplantation, said that rarely has a technologic invention affected the course of immunology so dramatically; but that was in the future, and not immediately obvious to the inventors. Milstein and Köhler were to create an industry. By 1984, the date of the prize, the practical and commercial effects of having a purified source of antibody with a single defined specificity had become obvious. Cambrosio and Keating note that by then, according to Index Medicus, there were already 10,000 articles on the subject. The technique was difficult to master, and like much biological manipulation, required a good deal of tacit and local knowledge gained directly from a laboratory or an individual who could make it work. A careful protocol was not always enough ( 155 ). The practical effect of monoclonal antibodies, apart from their many uses in research ( 156 ), was to retool tests for antigenic epitopes, including tumor antigens, viruses, and blood group antigens. The Wassermann test for syphilis, with its theoretical ambiguities, was replaced by a sandwich test, the enzyme-linked immunosorbent assay or ELISA. A similar test was devised for human immunodeficiency virus or HIV, used for both screening blood for transfusion and screening patients from early 1985. Convenient pocket-sized test kits for dozens of clinical problems appeared on the market—it is safe to say that pharmaceutical companies, clinical pathologists and patients took full advantage of them. In blood transfusion, epitope-specific monoclonals elucidated the details of the Rhesus antigen, which turned out to be a mosaic of many epitopes, rather as Alexander Wiener, its discoverer and spokesman, had argued in Landsteiner’s name in the 1940s ( 157 ). Different specificities of monoclonal anti-D identified nine critical residues of the protein molecule, protruding from the cell membrane on four loops. To

produce recognizable binding, two to four residues were needed, either all on one loop, or on two, three, or four loops, so that there must be a very large number of possible combinations. No monoclonal anti-D will react with all of them. For blood donor typing, even the weak variants that could immunize an Rh-negative patient must be classed as Rh D positive. For pregnant women, on the other hand, weak variants and partial Ds should be classed as Rh negative, since they could be immunized by an Rh-positive fetus. The British Blood Transfusion Service, led by the Bristol Institute for Transfusion Sciences, adopted a saline-reactive monoclonal anti-D that detected the common epitopes, along with one that was specific for the subtype VI, whose cells lack most of them ( 158 ). The example from blood-group serology shows monoclonal antibodies in action—older serological tests are refined and the molecular biology of the complex antigen is made visible by the extremely narrow and well-defined specificity of the antibodies. Clonal immunobiology has incorporated the molecular style, and returned to tease out the problems of the past.

MOLECULAR IMMUNOLOGY: DIVERSITY, HISTOCOMPATIBILITY, AND THE T-CELL RECEPTOR, 1980–PRESENT Many years ago, I wrote a short paper, my first on the history of immunology. I used the occasion of the Tenth International Congress of Medicine, held in Berlin on 4 August 1880 as a cross-section of what was important in immunology on that date. It was an important meeting—there were 7,056 people present, a very large number for a meeting at that time. The issue debated was whether immunity was a matter of cells, or of serum ( 159 ). A hundred years later in 1980, the central matter of immunology was still cells and serum. They were now not alternative explanations of immunity, but linked together as part of a single interactive system, represented by T cells, B cells, and antibody. Descriptions of that relationship have surfaced and then disappeared ( 160 ). The current one, still under investigation as I write today, involves the genetics of the immune system, both cellular antigens and globulins. The new methods of the 1980s were those of molecular biology ( 161 ). The elucidation of chain structure and amino-acid sequence by the separation methods of postwar biochemistry had not completely solved the problem of antibody diversity. If the template theory had collapsed, diversity must be genetically determined, but there were two schools of thought on that. If it had appeared far back in evolutionary time and was encoded in the germline, there must be a separate gene for each polypeptide fragment. But if the complete sequence was encoded, with one gene for every possible polypeptide chain, there would have to be an enormous number of genes to cover the enormous number of known and potential specificities, an echo of the problem that had divided Paul Ehrlich and Karl Landsteiner in the first decades of the century, and that was re-emphasized by Macfarlane Burnet in The Clonal Selection Theory of 1959. Another school of thought suggested that diversity might arise during the development of the individual, and might depend on somatic, not germline, inheritance. If diversification could be multiplied up in the somatic cells, for example, in B cells as they matured, fewer germline genes would be needed. But no other examples of

non-germline inheritance were known, so that was a difficult position to support. The first hint of a solution came in 1976 from Susumu Tonegawa and colleagues at the Basel Institute for Immunology, who used restriction enzymes to dissect the DNA, and recombination to identify the fragments. Werner Arber and his group at Basel were to be awarded a Nobel Prize for the invention of genetic engineering—in fact, for the work on restriction enzymes—in 1978 ( 162 ). Recombination, joining the fragmented DNA across species, was introduced in 1972 ( 163 ). Morange sees this paper by Paul Berg as having a foundation value similar to that in 1953 of Watson and Crick on the double helix ( 164 ). However, public alarm was generated by the threat of such genes escaping into the environment, particularly since the Berg experiment was carried out using E. coli, a universal gut inhabitant, and a virus that might be a cause of cancer. Berg himself was well aware of the dangers. At a conference held in 1975, standard operating procedures for the confinement of these potential pathogens were laid out, and were converted into rules by the National Institutes of Health in 1976. Commercial exploitation of the recombination technique was quick to develop, at first by small start-up companies, and later by the well-established pharmaceutical industry. As this got under way, controls were to be loosened ( 165 ). But in the late 1970s, as Tonegawa remembers, Recombinant DNA was just becoming available and was the ideal means for this purpose. Debates on the possible hazards of this type of research were flaring, initially in the USA and shortly afterwards in European countries. In order to make sure that our research would not become a target of controversy, Charlie and I got in touch with Werner Arber at the University of Basel who was coordinating recombinant DNA research activities in Switzerland. A small informal work group was set up by the local researchers interested in this technique. The consensus of the group, which was supported by most of the other Swiss researchers, was that we should follow the practices and guidelines being adopted in the USA ( 166 ). Tonegawa and his colleagues found that the V genes for light chains were split into two segments, separated by joining regions ( 167 ). The Leroy Hood group at the California Institute of Technology found that there were several separate segments with their joining regions coding for the heavy chain. An examination of these regions in inbred BALB/c mice suggested that only one of the regions was always identical in all of them, and so must be the one represented in the germline. The rest differed by single-base changes. There were therefore two separate genetic mechanisms controlling immunoglobulin diversity. In Tonegawa’s words, it turned out that an organism did not inherit even a single complete gene for antibody polypeptide chains. The genetic information was transmitted in the germline as a few hundred gene segments, then reshuffled into tens of thousands of complete genes. Further diversity resulted from hypermutation in these assembled genes. Tonegawa thinks that the initial rather low-affinity antibody response depends on pre-existing germline specificities. The later, higher-affinity antibody is produced by descendants of memory B cells through hypermutation, rearrangement, and splicing of germline genes in the course of B-cell maturation ( 168 ). Each generation of cells fits the antigen better and better—or as Niels Jerne noted in his thesis of 1954, antibody avidity increases with repeated exposure to antigen. In 1987, Susumu Tonegawa received a Nobel Prize for this work.

The field of histocompatibility was of practical importance with the rise of organ transplantation, but it soon grew beyond the practical boundaries of transplantation. A series of Histocompatibility Workshops beginning in 1964 took up the problems of typing and the formation of a nomenclature, hoping perhaps to forestall the bitter struggles over the terminology of the Rhesus blood group system, then still being fought out ( 169 ). New journals served the new field: Transplantation (1962), Tissue Antigens (1971), Immunogenetics (1971), and Journal of Immunogenetics (1974). By 1975, four loci each with a range of specificities had been worked out, and population studies had shown that distribution, as with the blood group antigens, varied globally. In the course of the 1980s, the histocompatibility site was shown to relate to immune response genes—with some simple antigens and some viruses, an all-or-none response can be detected differing between strains of an animal species. Ir gene control affected cellular immunity, and seemed to express itself through cells that collaborated with T cells ( 170 ). This is an area that is the subject of ongoing work, and it is treated elsewhere in this volume. The molecular biology of these antigens was a matter of intense interest. Laboratories at California Institute of Technology under Leroy Hood, at Harvard under J.L. Strominger, and in Uppsala under P.A. Peterson competed to make use of amino-acid analyzers to sequence the molecules. This work led in 1987 to the determination of the sequence and the three-dimensional structure of the molecules. Crystallographic pictures by Pamela Bjorkman from Don Wiley’s laboratory are summarized in her words as follows: The class I histocompatibility antigen… [h]as two structural motifs: the membrane-proximal end of the glycoprotein contains two domains with immunoglobulin folds that are paired in a novel manner, and the region distal from the membrane is a platform of eight anti-parallel ß-strands topped by a-helices. A large groove between the a-helices provides a binding-site for foreign antigens… The groove is located on the top surface of the molecule, and is therefore a likely candidate for the binding site for the foreign antigen recognized by a T-cell receptor ( 171 ). As Leslie Brent said recently, this was no longer the era of sole researchers working alone in their laboratories. These striking results were reached by competing teams of workers equipped with large grants, the heavy machinery of biomedical research. Equally, however, the inspiration for the problem was derived from the clinical importance, however brief, of the transplantation antigens, just as the work on immunochemistry of the 1920s and 1930s was derived ultimately from the requirements for a standardized diphtheria serum. Joseph Murray’s claim that organ transplantation revitalized immunology was no exaggeration. The importance of the thymus-dependent lymphocytes or T cells was first understood in the early 1960s. Several classes of T cell had been defined. They played a part in most immune reactions, turning on effector T and B cells against non-self antigens and suppressing activity directed against self. Since the 1970s, the T cells had been thought to work through the histocompatibility antigens recognized by the T-cell receptor. But

the nature of this receptor was still unknown. From the early 1980s, several groups of molecular immunologists collaborated, or competed, on the problem. They included teams under Ellis Reinherz and Stuart Schlossman at the Laboratory of Molecular Immunology at the Farber Cancer Institute and Medical School at Harvard, James Allison at the University of Texas, John Kappler and Philippa Marrack at the University of Colorado in Denver, Steve Hedrick and Mark Davis at the National Institutes of Health, and Tak Mak and his team at the Ontario Cancer Institute in Toronto. The newest techniques, such as monoclonal antibodies, gene hybridization, and DNA probes, drove the discoveries. First, the Harvard group in 1980 found a monoclonal antibody that blocked human T-cell function—it prevented the generation of cytotoxic T cells in a mixed lymphocyte culture, and stopped them from acting as helpers to B cells. The authors suggested that this might turn out to be useful in autoimmune disorders or in transplantation ( 172 ). In 1982, Allison and his group found another monoclonal antibody that identified a tumor-specific T-cell antigen in mice. The authors speculated that their antigen might be the T-cell equivalent of the B-cell idiotype, and that it might function as an antigen receptor ( 173 ). Also in 1982, Reinherz and his team at Harvard found a direct link between one of their monoclonal anti-T cell antibodies, and antigen recognition by T cells ( 174 ). The following year, John Kappler and Philippa Marrack (a hereditary immunochemist, since her father J.R. Marrack was the proposer of the lattice theory in 1934 [ 175 ]) in Denver, collaborating with Allison and McIntyre from Texas, used a fingerprinting technique to identify the peptides that conferred specificity on the antigen receptor on T cells that recognized the major histocompatibility complex, the key to self–not self recognition. Like the immunoglobulins, the protein sequences showed variable and constant regions linked by joining segments. The T-cell receptor generally resembled a pair of immunoglobulin light chains, a heterodimer of a and ß, joined by a disulfide bond, and with their -COOH ends buried in the cell membrane ( 176 ). By analogy to immunoglobulin, it was predicted that the DNA sequences coding for the T-cell receptor would be in separate regions in the genome, rearranging themselves somatically to form a complete gene. As Tak Mak explains, hybridization to a DNA probe complementary to a sequence encoding a T-cell receptor chain revealed different genomic hybridization patterns in different T-cell clones, all of them different again from the basic germ line pattern in non–T cells. The genetic reconstitution of a T-cell receptor by transfection of the DNA sequences into a recipient cell supported the hypothesis, and it was estimated that a total diversity of about 10 10could be achieved with combinatorial joining and somatic mutation together ( 177 ). A superfamily of Ig-like genetically determined proteins has been proposed. They include the immune globulins themselves, the T-cell receptors and other T markers defining different subsets of T cells, the histocompatibility antigens, some lymphoid–brain-associated antigens (in a piece of discreet advertising, one has been labeled MRC OX-2, associating it with the Medical Research Council’s Immunology Research Unit at Oxford), and other neural-associated antigens. It has been suggested that they have all evolved from a single stable domain, which then produced various sequences ensconced in different cell lines. All are on cell membranes, and seem to be involved with cell recognition and interaction; perhaps they acquired immune functions at about the time of vertebrate evolution ( 178 ). We are still fulfilling Macfarlane Burnet’s ideal of the 1960s, an immunological theory

based on the “simple concepts of biology—reproduction, mutation… and selective survival,” even though we have gone over to the chemical methods he so disliked ( 179 ). But biology itself is different now.

AIDS: THE PUBLIC FACE OF IMMUNOLOGY, 1986 TO THE PRESENT The earliest cases of Acquired Immune Deficiency Syndrome (AIDS) appeared in 1982, as a series of otherwise unusual infections and malignancies in male hom*osexual patients. At first, it seemed to be a problem that concerned only the gay community; it was named Gay-Related Immunodeficiency, or GRID ( 180 ). Social change followed, as gays organized to deal with a sickness that was untreatable, progressive, and ultimately fatal in all cases. Groups such as ACT-UP of New York, an organization of HIV+ people in the arts, demonstrated to demand access to the newest drugs. They brought their anger into the AIDS congresses, to the surprise of the scientists expecting to address a quietly formal scientific meeting. Gay activism altered the image of the hom*osexual from irresponsible hedonist to that of a caring and politically active individual, and the image of the patient from a passive sufferer to an impatient, informed, and critical adversary ( 181 ). Isolated in 1984, human immunodeficiency virus (HIV) was found to affect CD4 lymphocytes, key cells in the orchestration of the immune response ( 182 ). By the early 1990s, through the AIDS activist organizations, their outreach literature, and their brilliant posters, the immune system became part of popular discourse, as Emily Martin found in the streets of her home town. T cells entered the public domain as Mr. T, the killer cell ( 183 ). Historians, as well as their editors and publishers, began to see significance in immunology. A historiography of immunology appeared. The style of the epidemic differed profoundly according to the community affected. The well-organized gay community took safer sex into its own hands; and antiviral drugs when they arrived were carefully studied and diligently taken, in spite of their unpleasant side effects. As a result, death rates dropped, followed by declining new infection rates. But other high-risk groups such as users of illegal intravenous drugs are notoriously difficult to reach. There were even ethical objections raised to so-called harm-reduction initiatives, such as providing the users with clean needles or bleach kits. As with hepatitis B, and later C, the infection was transmitted in transfused blood and blood products, affecting particularly hemophiliacs using concentrates of the blood-clotting Factors VIII and IX. These were made from very large pools of plasma, often from several thousand donors; 60% of the donors were from the United States, where infection rates at the time were the highest in the world. Hemophiliacs, their partners, and their children died in large numbers. In Montreal, 56% of hemophiliacs were infected by 1982; by 1988, 74%. By the time ELISA testing of donor blood began in 1985, 1 in 270 of the blood donors in Toronto were testing positive for HIV ( 184 ). In every country, the seriousness of the epidemic was underestimated. The local blood transfusion organizations, including the trusted Red Cross, hesitated too long for reasons of economy to throw out products they knew were infected, even where heat-treated, infection-free materials were available. Everywhere legal remedies were

demanded. In France, criminal proceedings against several senior members of the organizations resulted in prison sentences. The accusations reached beyond the serological establishment to three former ministers deemed politically responsible ( 185 ). In South Africa, President Thabo Mbeki argued that the form of the epidemic in that country was more dependent on social factors such as poverty than on a virus. He invited a group of scientists who questioned the relevance of the virus, Peter Duesberg among them, to an open debate with a panel of mainstream thinkers. Arguments were presented to suggest that the increased death rates were due to a variety of infections, mainly linked to social stress, deprivation, and poverty. However, the statistician Malegapuru Makgoba of the South African Medical Research Council rejected that. He pointed to rising death rates among young adults of both sexes, beginning with the epidemic in the early 1990s. The significance of this is that a viral etiology makes it sensible to pressure drug companies to provide anti-virals or vaccines at low cost. The numbers of people now infected or dead in Sub-Saharan Africa, as in the Black Death of 14th century Europe, has cut deeply into education, government, and medical services, leaving classes without teachers, government without administrators, hospitals without doctors or nurses, and children without parents ( 187 ). Life expectancy at birth has dropped to below 30 in some areas of Africa, and projections show that the epidemic is not yet leveling off. A vaccine is reported as being in the test phase, but testing is slow with a disease that takes 10 years to develop symptoms. For the present, healthcare activism, including improvements to the status of women and the use of condoms, remain the most effective means of controlling the disease.

CONCLUSION Immunology is a laboratory science; individuals who call themselves immunologists are likely to work in a laboratory. This chapter treats immunology from their point of view, like most of the material on immunology and its history. But it is not a simple history of ideas. Between the lines, a careful reader can perceive that immunology is no abstract science that sets its own goals and wanders wherever science takes it. The force that directs its activities comes from the direction of clinical medicine and in turn, the scientific findings come back to the clinic and its ancillary, the pharmaceutical industry. Before World War II, the serologists responded to the need to understand and to standardize the antisera then in use—immunochemistry tried to answer questions posed by the diphtheria serum. During the 1940s, protein separation methods and plasma fractionation contributed to military medicine and later to civilian needs. After the war, the paradigm was set by transplantation. Work on tolerance and down-regulation succeeded work on immunization. New surgical procedures, the organization of organ supply networks, the development of immunosuppressive drugs, and the teasing out of the linked roles of transplantation antigens and cells were all part of a dialogue with the clinic. Monoclonal antibody research fed on the clinical opportunity, and supplied the pharmaceutical industry’s appetite for neat and accurate test kits, as much as the abstract need to know about the details of epitopes and of the immunoglobulin molecule. Finally, the advent of the AIDS epidemic made immunology a household word, and released the interest of historians in the activities of lymp hocytes. It is to be hoped that a new generation of historians will analyze immunology as an applied clinical

science. There is more to be said. REFERENCES 1. Cambrosio A, Keating P, Tauber AI. Introduction: immunology as a historical object. J Hist Biol 1994;27:531–573. 2. Leung AKC. “Variolation” and vaccination in late imperial China, ca 1570–1911, in Plotkin S, Fantini B, eds, Vaccinia, vaccination and vaccinology: Jenner, Pasteur and their successors , Paris: Elsevier, 1996:65–71. 3. Miller G. The adoption of inoculation for smallpox in France and England in the 18th century, Philadelphia: Pennsylvania University Press, 1957. 4. Rusnock AA. The weight of evidence and the burden of authority: case histories, medical statistics and smallpox inoculation. Clio Medica 1995;29:289–314. 5. Moulin A-M. Le dernier langage de la médecine: histoire de l’immunologie de Pasteur au SIDA, Vendome: Presses Universitaires de France, 1991. 6. Razzell P. Edward Jenner’s cowpox vaccine: the history of a medical myth, 2nd ed. Sussex: Caliban, 1980. 7. Arnold D. Smallpox and colonial medicine in nineteenth-century India, in Arnold D, eds. Imperial medicine and indigenous societies, Manchester: Manchester University Press, 1988:45–65. 8. Mercer AJ. Disease, mortality and population in transition, Leicester: University of Leicester Press, 1990. 9.

Bewell A. Romanticism and colonial disease, Baltimore, MD: Johns Hopkins Press, 1999.

10. Anderson W, Jackson M, Rosenkrantz B. Towards an unnatural history of immunology. J Hist Biol 1994;27:575–594; Anderson W, Immunities of empire: race, disease and the new tropical medicine, 1900–1920, Bull Hist Med 1996;70:94–118. 11. Moulin A-M. La métaphore vaccine. De l’inoculation à la vaccinologie. Hist Phil Life Sci 1992;14:271–297. For a highly unflattering account of Pasteur’s work on attenuation of organisms, see Geison GL, The private science of Louis Pasteur, Princeton, NJ: Princeton University Press, 1995, 145–176, 206–233. 12. Long ER. Tuberculin and the tuberculin reaction, in Jordan EO, Falk IS, eds. The newer knowledge of bacteriology and immunology, Chicago: University of Chicago Press, 1928:1016–1034. 13. Hardy A. Straight back to barbarism: anti-typhoid inoculation and the Great War. Bull Hist Med 2000;74:265–290. 14.

Salomon-Bayet C, ed. Pasteur et la révolution pastorienne, Paris: Payot, 1986.


Metchnikov E. Phagocytosis and immunity, London: British Medical Association, 1891.

16. Tauber AI, Chernyak L. History of immunology. The birth of immunology: I. Metchnikoff and his critics. Cell Immunol 1989;121:447–473; Tauber AI. The immune self: theory or metaphor? Cambridge: Cambridge University Press, 1994. It is true that Macfarlane Burnet, who formally introduced the self–not

self concept in 1949, referred both to Metchnikov and to Darwin as his predecessors. 17.

Moulin. Le dernier langage, 49–66.

18. Mazumdar PMH. Immunity in 1890. J Hist Med Allied Sci 1972;27:312–324; Besredka A, Histoire d’une idée: l’oeuvre de Metchnikov, Paris: Masson, 1921. 19. Portier C, Richet C. De l’action anaphylactique de certains venins, Comptes Rendus de la Société de Biologie 1902;54:170–172; Kröker K, Immunity and its other: the anaphylactic selves of Charles Richet, Stud Hist Phil Biol Biomed Sci 1999;30:273–296. 20. Besredka A. Local immunization, specific dressings, Baltimore: Williams, 1927, Plotz H, translator; Löwy I, “The terrain is all”: Mechnikov’s heritage at the Pasteur Institute, from Besredka’s antivirus to Bardach’s orthobiotic serum, in Lawrence C, Weisz G, eds. Greater than the parts: holism in biomedicine 1920–1950, Oxford: Oxford University Press, 1998:257–282. 21. Dunhill M. The Plato of Praed Street: the life and times of Almroth Wright, London: Royal Society of Medicine, 2000:120–127; Keating P, Vaccine therapy and the problem of opsonins. J Hist Med Allied Sci 1988;43:275–296; Worboys M, Vaccine therapy and laboratory medicine in Edwardian England, in Pickstone J, ed., Medical innovations in historical perspective, Basingstoke: Macmillan, 1992:84–103. 22. Wright AE. Studies on immunisation and their application to the diagnosis and treatment of bacterial infections, London: Constable, 1909. 23. Chen W. The laboratory as business: Sir Almroth Wright and the construction of penicillin, in Cunningham A, Williams P, eds. The laboratory revolution in medicine, Cambridge: Cambridge University Press, 1992:245–292. 24. Panton PN, Marrack JR. Clinical pathology London: Churchill, 1st ed. 1913, 6th ed. 1951; Gillett HT. Vaccine therapy in acute and chronic infections . London: Lewis, 1933. 25. Connor JTH. Doing good: the life of Toronto’s General Hospital, Toronto: University of Toronto Press, 2000. 26.

Chen. The laboratory as business, 287–290.

27. Burnet FM. The clonal selection theory of acquired immunity, London: Cambridge University Press, 1959:36. 28. Von Behring E, Kitasato S. Ueber das Zustandkekommen der Diphtherieimmunität und der Tetanusimmunität beir Thieren. Deutsche medizinische Wochenschrift 1890;16:1113. 29. Weindling P. From medical research to clinical practice: serum therapy for diphtheria in the 1890s, in Pickstone JV, eds. Medical innovations in historical perspective, Basingstoke: Macmillan, 1992:72–83; Throm C, Das Diphtherieserum: ein neues Therapieprinzip, seine Entwicklung und Markteinführung, Stuttgart: Wissenschaftliche Verlagsgesellschaft, 1995, 50–52. 30. Madsen T. Institutets udvikling 1902–1927. Madsen. Statens Seruminstitut: Institutets Udvikling 1902–1940, Copenhagen: Bianco Lunos, 1940:7–37; DeFries RD, The Connaught Medical Research Laboratories 1914–1948, Can J Public Health 1948;39:348–360. 31. Otto R. Die staatliche Prüfung der Heilsera. Arbeiten aus dem königlichen Institut für experimentelle Therapie zu Frankfurt-a-M 1906;2:7–86.


Bäumler E. Paul Ehrlich: scientist for life, New York: Holmes, 1984, Edwards G, translator.

33. Ehrlich P. The assay of the diphtheria-curative serum and its theoretical basis, 1897, translation in Himmelweit F, Marquardt M, Dale H, eds. Collected papers of Paul Ehrlich, vol. 2, London: Pergamon, 1957:107–125. 34. Ehrlich P. Die Wertbemessung des Diphtherieheilserums, und deren theoretische Grundlagen, 1897, translation in Himmelweit, Marquardt, Dale, eds. Collected papers, vol. 2, London: Pergamon, 1957:86–106 (German), 107–125 (English). 35. Ehrlich P. On immunity with special reference to cell life, The Croonian Lecture to the Royal Society of London, 1900, in Himmelweit, Marquardt, Dale, eds. Collected papers, vol. 2, London: Pergamon, 1957:178–195. 36.

Ada GL, Nossal GJV. The clonal selection theory. Sci Am 1987;257:62–69.

37. Parascandola J. Jasensky R. Origins of the receptor theory of drug action. Bull Hist Med 1974;48:199–220. 38. Cambrosio A, Jacobi D, Keating P. Ehrlich’s “beautiful pictures” and the controversial beginnings of immunologic imagery. Isis 1993;84:662–699. 39.

Silverstein A. A history of theories of antibody formation. Cell Immunol 1985;91:263–283.

40. Mazumdar PMH. Species and specificity: an interpretation of the history of immunology, Cambridge: Cambridge University Press, 1995:202–213, 214–236. 41. Bordet J. Sur le mode d’action des antitoxines sur les toxines, 1903, in Gay FP, ed. and transl., Studies in immunity by Professor Jules Bordet and his collaborators, New York: Wiley, 1909:161–186. 42. Pick EP. Biochemie der Antigene, mit besonderer Berücksichtigung der chemischen Grundlagen der Antigenspezifizität, in Kolle W, von Wassermann A, eds. Handbuch der pathogenen Organismen, 2nd ed. Jena, Austria: Fischer, 1912:685–868. 43. Mazumdar PMH. The antigen—antibody reaction and the physics and chemistry of life. Bull Hist Med 1974;48:1–21. 44. Landsteiner K. Die Theorien der Antikörperbildung. Wiener klinische Wochenscrift 1909;22:1623–1631. 45. Landsteiner K, Lampl H. Ueber die Antigen-eigenschaften von Azoprotein: XI. Mitteilung über Antigene. Zeitschrift f. Immuniyätsforschung 1917;26:293–304. 46. Dubos RJ. Chemistry in medical research, in Dubos, The professor, the Institute and DNA, New York: Rockefeller University Press, 1976:35–46; Corner G, The laboratories are organised, in Corner, A history of the Rockefeller Institute 1901–1953: origins and growth, New York: Rockefeller Institute Press, 1964:56–87. 47. Mazumdar PMH. The template theory of antibody formation and the chemical synthesis of the twenties, in Mazumdar, eds. Immunology 1930–1980: essays on the history of immunology, Toronto: Wall & Thompson, 1989:13–32. 48.

Putnam EW, Felix Haurowitz. Biographies of Members of the National Academy of Science,

Washington, DC: National Academy of Science, 1994;64:135–163. 49.

Marrack JR, The chemistry of antigens and antibodies, London: HMSO, 1934:113–123.

50. Throm C. Das Diphtherie-serum: ein neues Therapieprinzip, seine Entwicklung und Markteinführung, Stuttgart: Wissenschaftliche Verlagsgesellschaft, 1995. Ehrlich was never eligible for a university position, since he was a Jew. The offer by an institute came about because of the funding provided by an influential group of Frankfurt citizens. 51. Kröker K. Immunity and its other: the anaphylactic selves of Charles Richet. Stud Hist Phil Biol Biomed Sci 1999;30:273–296. 52. Prausnitz C. Küstner H, Studien über die Überempfindlichkeit. Zentralblatt f. Bakt. Originalien 1921;86:160–169; von Pirquet CE, Schick B, Die Serumkrankheit, Leipzig: Deuticke, 1905. 53. Jackson M. Between scepticism and wild enthusiasm: the chequered history of allergen immunotherapy in Britain, in Moulin A-M, Cambrosio A, eds. Singular selves: historical Issues and contemporary debates in immunology, Amsterdam: Elsevier, 2001:155–164. 54. Chase MW. Irreverent recollections from Cooke and Coca, 1928–1978. J Allergy Clin Immunol 1979;64:306–320. 55. Linton DS. The obscure object of knowledge: German military medicine confronts gas gangrene during World War I. Bull Hist Med 2000;74:291–316. 56. Mazumdar PMH. “In the silence of the laboratory”: the League of Nations standardises syphilis tests, Social History of Medicine, 2003, ( in press). On the ban, see Schröder-Gudehus B, Deutsche Wissenschaft und internationale Zusammenarbeit 1914–1928: ein Beitrag zum Studium kultureller Beziehungen in politischen Krisenzeiten [thesis], University of Geneva, 1966. 57. Bordet J, Gengou O. Sur l’existence des substances sensibilatrices dans la plupart des sérums antimicrobiens, Ann l’Inst Pasteur, 1901, translated as On the existence of sensitising substances in the majority of antimicrobial sera, in Gay FP, eds. Studies in immunity by Professor Jules Bordet and his collaborators, New York: Wiley, 1909:217–227; Wassermann A, Neisser M, Brücke K, Eine serodiagnostisches Reaktion bei Syphilis, Deutsche medizinische Wochenschrift 1906;32:745–746; Löwy I, Testing for a sexually transmissible disease, in Berridge V, Strong P, eds. AIDS and contemporary history, Cambridge: Cambridge University Press, 1993:74–92. 58. Weindling P. The politics of international co-ordination to combat sexually transmitted diseases, 1900–1890s, in Berridge V, Strong P, eds. AIDS and contemporary history, Cambridge: Cambridge University Press, 1993:93–107. 59. Fleck L. Genesis and development of a scientific fact, Chicago: University of Chicago Press, 1979, Trenn TJ, translator (of 1935 edition). 60. Kahn RL. Serum diagnosis of syphilis by precipitation: governing principles, procedure and clinical application of the Kahn Precipitation Test, Baltimore: Williams, 1926. 61. Slatkin M. Trends in the diagnosis and treatment of syphilis. Med Clin North Am 1965;49:823–847; Nicholas L, Beerman H, Present day serodiagnosis of syphilis: a review of some of the recent literature, Am J Med Sci 1965;249:466–483. 62. Landsteiner K. Ueber Agglutinationerscheinungen normalen menschlichen Blutes. Wiener klinische Wochenschrift 1901;14:1132–1134. 63.

Hirszfeld L. Konstitutionsserologie und Blutgruppenforschung , Berlin: Springer, 1928, Keating P,

Holistic bacteriology: Ludwik Hirszfeld’s doctrine of serogenesis between the two world wars, in Lawrence C, Weisz G, eds., Greater than the parts: holism in biomedicine 1920–1950, Oxford: Oxford University Press, 1998:283–302. 64. Mazumdar PMH. Two models for human genetics: blood grouping and psychiatry between the wars, Bull Hist Med 1996;70:609–657. 65. Mazumdar PMH. Blood and soil: the serology of the Aryan racial state. Bull Hist Med 1990;64:187–219. 66. Zimmermann DR. Rh: the intimate history of a disease and its conquest, New York: Macmillan, 1973. 67. Crile GW. Hemorrhage and transfusion: an experimental and clinical research, New York: Appleton, 1909:318–319, 323–324. 68. Robertson LB. The transfusion of whole blood. BMJ 1916;p:38; Robertson OH, A method of citrated blood transfusion, BMJ 1918;t:477. For a conspectus of transfusion at the end of World War I, see Keynes G, Blood transfusion, London: Oxford University Press, 1922. 69. Vaughan JM, Panton PN. The civilian blood transfusion service, in Dunn CL, eds. The emergency medical services, vol. 1, London: HMSO, 1952:334–355; Medical Research Council, War memorandum no. 9: determination of blood groups, London: HMSO, 1943. 70.

Riddell V. Blood transfusion, London: Oxford University Press, 1939:300–302.

71. Diamond LK. A short history of blood banking in the United States. JAMA 1965;193:128–132; Diamond LK, A history of blood transfusion, in Wintrobe MM, ed. Blood pure and eloquent: a story of discovery, of people and of ideas, New York: McGraw-Hill, 1980:659–688. 72. Vaughan JM, Panton PN. The civilian blood transfusion service, in Dunn CL, eds. The emergency medical services, London: HMSO, 1952;1:334–355, 2:68–71. 73. Loutit JF, Mollison PL. Advantages of a disodium citrate–glucose mixture as blood preservative. BMJ 1943;p:744–745. 74.

Norris RF. Hospital programs of blood banking. JAMA 1965;193:133–135.

75. On the ethical problems and dangers of the U.S. system of paid donors from a British perspective, see Titmuss RM, The gift relationship: from human blood to social policy, London: Allen, 1970. On the political problems of vaccine development, see Muraskin W, The war against hepatitis B, Philadelphia: University of Pennsylvania Press, 1995. 76.

Vaughan and Panton, Civilian blood transfusion service, 355.


Kerker M. The Svedberg and molecular reality. Isis 1976;67:190–216.

78. Rånby B, eds. Physical chemistry of colloids and macromolecules, Proceedings of the International Symposium on Physical Chemistry and Macromolecules to Celebrate the 100th Anniversary of the Birth of Professor Theodor Svedberg on 30 August 1884… Uppsala University, Sweden, 22–24 August 1984, Oxford: Blackwell, 1985, The proceedings contain several reminiscences and anecdotes by scientists who worked with Svedberg. 79.

Fruton J. Molecules and life: historical essays on the interplay of chemistry and biology, New York:

Wiley, 1972:138–139. 80.

Mazumdar. Species and specificity, 152–178.

81. Brohult S. Gralén N. The Institute of Physical Chemistry at the University of Uppsala: some notes on its history and activities, in Tiselius A, ed. [Festschrift for] The Svedberg, Uppsala: Almqvist, 1944:623–638. On unity and simplicity, see Mazumdar. Species and specificity, 152–178. 82.

Brohult and Gralén. The Institute of Physical Chemistry, 629.


Tiselius A. Electrophoretische Messungen am Eiweiss. Kolloid Zeitschrift 1938;85:119–128.

84. Kay LE. Laboratory technology and biological knowledge: the Tiselius electrophoresis apparatus, 1930–1945. Hist Phil Life Sci 1988;10:51–72; See also Ede A, Colloids and quantification: the ultracentrifuge and its transformation of colloid chemistry. Ambix 1996;43:32–45. 85. Ede A. When is a tool not a tool? Understanding the role of equipment in the early colloidal laboratory. Ambix 1993;40:11–24. 86.

Spiegel-Adolf M. Die Globuline, Leipzig: Steinkopf, 1930:422–429.

87. Kabat EA, Tiselius E. An electrophoretic study of immune sera and purified antibody preparations. J Exper Med 1939;69:119–131. 88. Creager ANH. Producing molecular therapeutics from human blood: Edwin Cohn’s wartime enterprise, in de Chadarevian S, Kamminga H, eds. Molecularizing biology and medicine, Amsterdam: Harwood, 1998:107–138; Creager ANH, From blood fractions to antibody structure: gamma globulin research growing out of World War II, in Moulin A-M, Cambrosio A, eds. Singular selves: historical issues and contemporary debates in immunology, Amsterdam: Elsevier, 2000:140–154; Cohn EJ, The history of plasma fractionation, Adv Mil Med 1948;1:364–443. 89. Tswett M. Physikalisch-chemische Studien über Chlorophyll. Berichte d. botanische Gesellschaft 1906;24:316–323. 90. Consden R, Gordon AH, Martin AJP. Quantitative analysis of proteins: a partition chromatographic method using paper. Biochem J 1944;38:476–484. 91. Moore S, Stein WH. Chromatography of amino-acids on sulfonated polystyrene resins. J Biol Chem 1951;192:663–681. 92. Porath J. Gel filtration of proteins, peptides and amino-acids. Biochim Biophys Acta 1960;39:193–207. 93. Sober HA, Gutter FJ, Wyckoff MM, et al. Chromatography of proteins. II. Fractionation of serum protein on anion-exchange cellulose. J Am Chem Soc 1956;78:756–763; Sober R, Hartley W Jr, Carroll WR, et al., Fractionation of proteins, The Proteins 1965;3:1–97. For a contemporary overview of protein separation methods, see Morris CJOR, Morris P, Separation methods in biochemistry, London: Pitman, 1964, especially on the choice of separation methods, 838–853. 94.

Morris and Morris, Separation methods in biochemistry, 701–705.

95. Smithies O. Grouped variations in the occurrence of new protein components in normal human serum. Nature 1955;175:307–308; Although this is the first publication to use a starch gel, Smithies does not describe the method, only the results.

96. Ishizaka K, Ishizaka T. Human reaginic antibodies and immunoglobulin E. J Allergy 1968;42:330–363; See Chase, Irreverent recollections, 315. 97. Goodman H. Immunodiplomacy: the story of the World Health Organisation’s immunological research programme, in Mazumdar PMH, eds. Immunology 1930–1980: essays on the history of immunology, Toronto: Wall & Thompson, 1989:253–272. 98. Porter RR. The hydrolysis of rabbit ?-globulin and antibodies with crystalline papain. Biochem J 1959;73:119–127. See also Porter RR, Immunochemistry, Annu Rev Biochem 1962;31:625–652. 99.

Porter RR. The structure of antibodies. Scientific American, 1967;217:81–90.


Porter RR. Structural studies of immunoglobulins. Science 1973;180:713–716.


Putnam FW. Structure and function of the plasma proteins. The Proteins 1965;3:153–267.

102. Poulik MD, Edelman GM. Comparison of reduced alkylated derivatives of some myeloma globulins and Bence–Jones proteins. Nature 1961;191:1274–1276. 103.

Kunkel HG. Myeloma proteins and antibodies. Harvey Lectures 1965;59:219–242.

104. Potter M. Some reminiscences on the contributions of cancer research to immunology in the 1950s, in Mazumdar PMH, eds. Immunology 1930–1980: essays on the history of immunology, Toronto: Wall & Thompson, 1989:95–105. 105.

Crick F. On protein synthesis. Symp Soc Exp Biol 1957;12:138–167.

106. Haurowitz F. The chemistry and functions of proteins, New York: Academic Press, 1963,439. The quotation is from B. Commoner, in Science 1961;133:1747. 107.

Nirenberg M. Protein synthesis and the RNA code. Harvey Lectures 1965;59:155–185.

108. Neushul P. Science, government and the mass production of penicillin. J Hist Med Allied Sci 1993;48:371–395. 109. Smith JS. Patenting the sun: polio and the Salk vaccine, New York: Morrow, 1990:279–312. On polio and other viral vaccines made in cell culture, see Beale AJ, The development of IPV, in Plotkin SA, Fantini B, Vaccinia, vaccination, vaccinology: Jenner, Pasteur and their successors. International Meeting on the History of Vaccinology, December 1995, Amsterdam: Elsevier, 1996:211–292. 110.

Coulter HL, Fisher BL. DPT: a shot in the dark, New York: Harcourt, 1985.

111. Amos B, eds. Progress in immunology: proceedings of the First International Congress of Immunology, Washington, DC, 1971, New York: Academic Press, 1971, This arrangement of the sessions scuttled my attempt to use the congress to revise for my exam. 112. Nossal’s statement dates from 1985 and is cited in Sexton C, Burnet: a life, 2nd ed. Melbourne: Oxford University Press, 1999:131–132. 113. Sturdy S. Reflections: molecularization, standardization and the history of science, in de Chadarevian S, Kamminga H, eds. Molecularizing biology and medicine: new practices and alliances, 1910s–1970s, Amsterdam: Harwood, 1998:273–292.


Chase MW. Hypersensitivity to simple chemicals. The Harvey Lectures, 1965–1966;61:169–203.


Koch R. Weitere Mittheilungen über das Tuberkulin. Deutsche med Wschr 1891;17:1189–1192.

116. von Pirquet CF. Quantitative experiments with the cutaneous tuberculin reaction, J Pharm 1909–1910;1:151–174; von Pirquet CF, Allergy. Arch Int Med 1911;7:259–288. 117. Humphrey JH, White RG. Immunology for Students of Medicine, 2nd ed. Oxford: Blackwell, 1963:337–377. 118. Lawrence HS. The early history of soluble factors, in Mazumdar PMH, eds. Immunology 1930–1980: essays on the history of immunology, Toronto: Wall and Thompson, 1989:221–230. 119.

Brent L. A history of transplantation immunology, San Diego: Academic Press, 1997, 70–115.

120. Murray JE. Organ transplantation and the revitalization of immunology, in Gallagher RB, Gilder J, Nossal GJV, eds. Immunology: the making of a modern science, London: Academic Press, 1995:179–189. 121. Burnet FM, Fenner F. The production of antibodies, 2nd ed. New York: Macmillan, 1949:102–105; Tauber AI, Podolsky SH, Frank Macfarlane Burnet and the immune self. J Hist Biol 1994;27:531–573. 122. Own RD. Immunogenetic consequences of vascular anastomoses between bovine twins. Science 1945;102:400–401; Cinader BH, Down-regulation and tolerance: the trail from the past, in Mazumdar PMH, ed. Immunology 1930–1980: essays on the history of immunology, Toronto: Wall & Thompson, 1989:51–66. 123. Billingham RE, Brent L, Medawar PB. Actively acquired tolerance of foreign cells. Nature 1953;172:603–606. 124.

Brent L. History of transplantation immunology, 73.

125. Burnet FM. A discussion of immunological tolerance under the leadership of Sir Frank Macfarlane Burnet. Proc R Soc Lond B Biol Sci 1957;146:1–92. 126.

Brent. History of transplantation, 230–305.

127. Medawar PB. The strange case of the spotted mice, in Medawar, in D. Pyke, eds. The threat and the glory: reflections on science and scientists, Oxford: Oxford University Press, 1991:71–82. 128. Stähelin H. The development of immunosuppressive agents from x-rays to cyclosporin, in Mazumdar PMH, eds. Immunology 1930–1980: essays on the history of immunology, Toronto: Wall & Thompson, 1989:185–201. 129.

Jerne NK. A study of avidity. Acta Pathol Microbiol Scand Suppl 1951;87:1–183.

130. Jerne NK. The natural-selection theory of antibody formation. Proc Nat Acad Sci 1955;41:849–857. 131. Söderqvist T. Darwinian overtones: Niels K. Jerne and the origin of the clonal selection theory of antibody formation. J Hist Biol 27:481–529; Söderqvist T, Hvilken Kamp for at Undslippe: en Biografi om Immunologen og Nobelpristrageren Niels Kaj Jerne, Copenhagen: Borgens Forlag, 1998. Parabasis: the selection theory as a personal confession of its author, and Chapter 15, I just sit at my desk and thin k

about life in general. 132.

Haurowitz F. Review of The production of antibodies, 2nd ed. 1949. J Immunol 1951;66:485–486.


Talmage DW. Allergy and immunology. Annu Rev Med 1957;8:239–256.


Burnet FM. A modification of Jerne’s theory of antibody formation. Aust J Sci 1957;20:67–68.


Burnet FM. The impact of ideas on immunology. Cold Spring Harb Symp Quant Biol 1967;32:1–8.

136. Burnet FM. The clonal selection theory of acquired immunity, Melbourne: Cambridge University Press, 1959:23–27. 137.

Burnet. Clonal selection theory, 182–183.

138. Ada GL, Nossal GJV. How cells make antibody: the clonal selection theory of immunity. Sci Am 1987;257:62–69. 139. Burnet. Clonal selection theory, 64; Burnet FM, A Darwinian approach to immunity. Nature 1964;203:451–454. 140.

Sexton, Burnet: a life, 134–141.


Nossal GJV, Lederberg J. Antibody production by single cells. Nature 1958;182:1383–1384.

142. Jerne NK, Noordin AA. Plaque formation in agar by single antibody producing cells. Science 1963;141:405. 143. Talmage DW. Is this theory necessary?, in Mazumdar PMH, eds. Immunology 1930–1980: essays on the history of immunology, Toronto: Wall & Thompson, 1989:67–72. 144. Cambrosio A, Keating P. Exquisite specificity: the monoclonal antibody revolution, (New York, NY: Oxford University Press 1995) 1995. 145. Petrov RV. Me or not-me: immunological mobiles, (Moscow: Mir Publishers, 1987) 1987, 99–110, Degtyaryova, GY, translator. 146. Gowans JL. The mysterious lymphocyte, in Gallagher RB, Gilder J, Nossal GJV, et al., eds. Immunology: the making of a modern science, London: Academic Press, 1992:65–74; Miller JFAP, The discovery of thymus function, in Gallagher RB, Gilder J, Nossal GJV, et al., eds. Immunology: the making of a modern science, London: Academic Press, 1995;75–84. 147.

Stillwell CR. Thymectomy as an experimental system in immunology. J Hist Biol 1994;27:379–401.

148. Good RA. The Minnesota scene: a crucial portal of entry to modern cellular immunology, in Szentivanyi A, Friedman H, eds. The immunologic revolution: facts and witnesses, Boca Raton, FL: CRC Press, 1993:105–168. 149. Morse HC. Origins of inbred mice, New York: Academic Press, 1978; Foster J, Henry L, Small D, et al., The mouse in biomedical research, New York: Academic Press, 1981. 150. Löwy I, Gaudillière J-P. Disciplining cancer: mice and the practice of genetic purity, in Gaudillière J-P, Löwy I, eds. The invisible industrialist: manufactures and the production of scientific knowledge , Houndmills, Basingstoke, Hampshire: Macmillan Press; New York: St. Martin’s Press, 1998:209–249.

151. Milstein C. Monoclonal antibodies. Sci Am 1980;243:66–74; Milstein, Monoclonal antibodies from hybrid myelomas, Wellcome Foundation Lecture, 1980. Proc R Soc Lond B Biol Sci, 211:393–412; Cambrosio A, Keating P, Between fact and technique: the beginnings of hybridoma technology. J Hist Biol 1992;25:175–230. 152.

Potter M. Myeloma proteins. Experientia 1986;42:967–1068.


Uhr JW. The 1984 Nobel Prize for Medicine. Science 1984;226:1025.


Brent. History of transplantation immunology, 16–46.

155. Cambrosio A, Keating P. Going monoclonal: art, science and magic in the day-to-day use of hybridoma technology. Soc Problems, 1988;35:244–260. 156.

Springer TA. Hybridoma technology in the biosciences and medicine, New York: Plenum, 1985.


Mazumdar. Species and specificity, 337–378.

158. Scott ML, Voak D, Liu W, et al. Using monoclonal antibodies and site-directed mutagenesis to map the epitopes of the blood group D antigen, in Eibl M, Mayr WR, Thorbecke GJ, eds. Epitope recognition since Landsteiner’s discovery, Berlin: Springer, 2002:83–102. 159.

Mazumdar PMH. Immunity in 1890. J Hist Med Allied Sci 1972;27:312–324.

160. Jerne NK. Towards a network theory of the immune system. Ann d’Immunologie 1974;125C:373–389; Moulin A-M, Le dernier langage, 333–339. 161. Morange M. Histoire de la biologie moléculaire, Paris: La Découverte, 1994, translated as A history of molecular biology, Cambridge, MA: Harvard University Press, 1998. 162.

Arber W. Promotion and limitation of genetic exchange. Science 1979;205:361–365.

163. Jackson DA, Symons RH, Berg P. Biochemical method for inserting new genetic information into DNA of Simian Virus 40: circular SV40 molecules containing Lambda Phage genes and the galactose operon of E. coli. Proc Nat Acad Sci USA 1972;69:2904–2909. 164.

Morange, Histoire de la biologie moléculaire, 241–245.

165. Morange, Histoire de la biologie moléculaire, 245–251; Krimsky S, Genetic alchemy: the social history of the recombinant DNA controversy, Cambridge, MA: MIT Press, 1982; Wright S, Recombinant DNA technology and its social transformation, 1972–1982, Osiris 2nd series, 1986;2:303–360. 166. Tonegawa S. Somatic generation of antibody diversity, in Gallagher RB, Gilder J, Nossal GJV, et al., eds. Immunology: the making of a modern science, London: Academic Press, 1995:145–161. 167. Tonegawa S, Maxam AM, Tizard R, et al. Sequence of a mouse germ line gene for a variable region of an immunoglobulin light chain. Proc Nat Acad Sci USA 1978;74:3171. Kurosawa Y, Tonegawa S, Organization, structure and assembly of the immunoglobulin heavy chain diversity DNA segments. J Exp Med 1982;155:201–218. 168. Tonegawa. Somatic generation, 159–160; Leder P, The genetics of antibody diversity. Sci Am 1982;246:102. For a contemporary textbook account of genetic control of diversity, see Hood LE, Weissman IL, Wood WB, et al., Immunology, 2nd ed. Menlo Park, CA: Benjamin, 1984:81–111.


Mazumdar. Species and specificity, 337–378.


Paul WE. Immunogenetics New York: Raven, 1984:151–167.

171. Bjorkman PJ, Saper MA, Samraoui B, et al. Structure of the human class I histocompatibility antigen, HLA-2, Nature 1987;329:506–512; Bjorkman PJ, Saper MA, Samraoui B, et al., The foreign antigen binding site and T cell recognition regions of class I histocompatibility antigens, Nature 1987;329:512–518. Also cited by Brent, History of transplantation immunology, 150–151. 172. Reinherz ER, Hussey RE, Schlossman SF. A monoclonal antibody blocking human T cell function. Eur J Immunol 1980;10:758–762. 173. Allison JP, McIntyre BW, Bloch D. Tumor-specific antigen of murine T-lymphoma defined with monoclonal antibody. J Immunol 1982;129:2293–2299. 174. Reinherz EL, Meuer S, Fitzgerald KA, et al. Antigen recognition by human T lymphocytes is linked to surface expression of the T3 molecular complex. Cell 1982;30:735–743. 175.

Marrack. The chemistry of antigens and antibodies, 113–123.

176. Kappler J, Kubo R, Haskins K, et al. The major histocompatibility complex–restricted antigen receptor on T cells in mouse and man: identification of constant and variable peptides. Cell 1983;35:295–302. 177. Caccia N, Toyonaga B, Kimura N, et al. The a and ß chains of the T-cell receptor, in Mak TW, eds. The T-cell receptors, New York: Plenum, 1988:9–51. 178. Barclay AN, Johnson P, McCaughan GW, et al. Immunoglobulin-related structures associated with vertebrate cell surfaces, in Mak TW, eds. The T-cell receptors, New York: Plenum, 1988:53–87. 179.

Burnet FM. The impact of ideas on immunology. Cold Spring Harb Symp Quant Biol 1967;32:1–8.


Kahn AD. AIDS: the winter war, Philadelphia, PA: Temple University Press, 1993:116–153.

181. Altman D. Legitimation through disaster: AIDS and the gay movement, in Fee E, Fox DM, eds. AIDS: the burdens of history, Berkeley: University of California Press, 1988:301–315. 182. Nicholson JKA, Mawle AC, McDougal JS. The effects of HIV-1 on the immune system, in Madhok R, Forbes CD, Evatt BL, Blood, blood products and HIV, 2nd ed. London: Chapman, 1994:75–100. 183. Martin E. Flexible bodies: tracking immunity in American culture from the days of polio to the age of AIDS, Boston, MA: Beacon, 1993:23–112. 184. Gilmore N, Somerville MA. From trust to tragedy: HIV/AIDS and the Canadian blood system, in Feldman EA, Bayer R, Blood feuds: AIDS, blood and the politics of medical disaster, New York: Oxford University Press, 1999:127–159; Canadian Ministry of Public Works and Government Services, Commission of inquiry on the blood system in Canada, Commissioner Hon. Justice Horace Krever (Ottawa: Canadian Goverment Publishing 1997). 185. Steffen M. The nation’s blood: medicine justice and the state in France, in Feldman EA, Bayer R, eds. Blood feuds: AIDS, blood and the politics of medical disaster, New York: Oxford University Press, 1999:95–126. 186.

South Africa. Presidential AIDS Advisory Panel Report, a synthesis report of the deliberations by

the panel of experts invited by the President of the Republic of South Africa, the Hon. Mr Thabo Mbeki, March 2001. 187. Nolen S. Continent’s educated can no longer run things: they’re dead or can’t cope, report datelined Nkhotakota, Malawi, Globe and Mail June 19, 2002, A13.

Chapter 3 Immunoglobulins: Structure and Function Fundamental Immunology

Chapter 3 Grant Kolar

Immunoglobulins and B Lymphocytes Immunoglobulins: Structure and Function


INTRODUCTION All who approach the study of the structure and function of immunoglobulins eventually marvel at the duality of the problem: There is variability and there is constancy, and an appreciation of both is critical to understand this class of proteins that are the prototypic members of the “immunoglobulin superfamily.” Having co-authored this chapter in all previous editions of this volume, providing a fresh perspective may seem difficult, but it is not. The study of the structure and function of antibodies remains as fresh as it was three decades ago, as new and existing experiments continue to be published that provide critical new insights into these molecules. The relationship between the structure and the function of the immunoglobulin molecule is a tribute to the power of molecular evolution. Via the duplication and diversification of the immunoglobulin hom*ology domain, a family of molecules with diverse biological functions has been derived. The antibody molecule plays the central role in humoral immunity by attaching to pathogens and then recruiting effector systems to stem the invader. In doing so, as noted above, it embodies two antagonistic tendencies—diversify and commonality—since it must possess both a variable surface to recognize different foreign antigens, and a constant surface that its own effector systems can recognize.

The fundamental tertiary structure of antibodies is called the immunoglobulin fold—the basic three-dimensional structure that thematically describes the structure of both the variable and constant domains of an immunoglobulin, as well as other members of the immunoglobulin superfamily. It was the recognition of this repeating structure that many years ago led to one of the first descriptions of “families” of molecules, some of which bore only minimal amino-acid sequence similarity but at the same time had profound three dimensional structural hom*ology. The major difference between the variable and constant domains from a structural perspective is the “loops” between the sandwich-like layers. In the variable domains these loops represent the amino acids, which by and large are in contact with antigen. As such they tend to be longer than the loops in the constant region where they serve, in general, as the structures that interact with certain effector molecules such as cell-surface receptors, or serum proteins such as certain complement components. The antibody molecule is also called “the B-cell receptor,” as at early stages of the immune response with a membrane exon at its carboxyl terminus, immunoglobulins serve as cell-surface receptors for antigen. Further stimulation of B cells with antigen leads to differentiation of the B cell such that antibody is secreted into the serum to make up the circulating antibody pool. Antibodies are one of the major plasma proteins and are often referred to as the “first line of defense” against infection. Additional stimulation of cell-surface antibodies leads to a “class switch” such that antibodies of different classes are produced, first as a cell-surface form and then as a secreted form. Thus, the antibody molecule has yet two other functions: to serve as a receptor molecule to transmit the signal from antigen capture to downstream signaling events that instruct the cell to perform such functions as division, secretion, and differentiation. Finally, a slightly modified immunoglobulin eventually enters the serum antibody pool. This chapter will attempt to provide a framework for understanding the various structural elements of the immunoglobulin classes and subclasses and connect those structural elements with the discrete biologic functions commonly attributed to antibody molecules. Other chapters will deal with the molecular events leading to the formation of this remarkable set of molecules and yet others will deal with specific effector functions.

A HISTORICAL PERSPECTIVE The structure and function of immunoglobulins is inexorably connected with the knowledge of the times in which discoveries were made in the field. Thus, in the 1940s and 1950s, when antibodies were known as “antitoxins” and “antisera” and the immune response was primarily studied as a serum response to antigenic challenge (largely from deliberate immunization although to some extent from response to disease) it was sufficient to label them simply as “antibodies.” It is important to remember that until the 1950s, there were few ways to partition serum proteins, and most relied on techniques that separated albumins from globulins (in medicine this became known as the A/G ratio). In the 1960s, once electrophoresis became commonplace, the globulins were divided into a 1 , a 2 , ß, and ? globulins. The connection between antibodies and ? globulins followed. “Sizing” columns were required to distinguish immunoglobulins into those that were “heavy” (IgM), “regular” (IgA, IgE, IgD, IgG), and “light” (light chain

dimers). Only after immunoelectrophoresis was it clear that there were other “classes” of immunoglobulins. Finally, with the discovery of myeloma proteins as “gamma globulins” or “immunoglobulins,” the clear class and subclass (isotype) distinctions that we know today became commonplace. When hybridomas and the immortalization of B cells became commonplace, further distinctions became evident. However, from a historical perspective it should be appreciated that significant “structure–function” issues were solved by biochemists. With the advent of molecular biology, we gained great insight into the genomic structure of antibodies, learned a great deal of how the information was stored for the variable regions and so on, but no new classes, subclasses, allotypes and the like were discovered. In addition, no new functions of antibodies were uncovered. Thus, we owe a great debt to the immunochemists of the 1940s through the 1960s for laying out the basic structure/function relationship of arguably the most significant molecule of our field.

INTRODUCTION TO STRUCTURE AND NOMENCLATURE The immunoglobulin molecule is a complex structure of four polypeptide chains. The central structural component of the molecule is the Ig domain. This key structure is discussed in a subsequent section. The four-polypeptide chains are organized as a hom*odimeric structure of a heterodimer between a heavy and light chain. Both chains contain variable and constant domains, with the heavy chain having two or three more constant domains than the light chain. Dimerization between the heavy and light chain variable domains and the first constant domain occurs as a result of hydrophobic interactions as well as a set of disulfide bonds at the carboxy-terminal end. A hom*odimer of this heavy–light chain configuration is then produced and held together by disulfide bonds in the hinge and tight hydrophobic interactions of the other constant domains. Therefore, an immunoglobulin contains two heavy chains (typically 55 kD each) and two light chains (25 kD each) ( 1 ). This forms an overall “Y” or “T” conformation that is the most widely recognized feature of immunoglobulin structure. By enzymatic and/or chemical cleavage, the immunoglobulin molecule can be broken into a number of “sections” or “fragments.” The experiments that created these fragments and later those that resulted in an understanding of these fragments are still among the most elegant experiments in our field. They continue to influence our descriptions of the molecule today and will be critical for understanding the structure/function issues discussed in this chapter. Porter ( 2 ) found that papain would cleave the antibody into two species of protein. The Fab (fragment antigen binding) portion acquired in this cleavage would monovalently bind to antigen. Two of these regions are produced per immunoglobulin as cleavage occurs N terminal to the disulfide bonds of the hinge ( 3 , 4 ). The remaining portion, the Fc (fragment crystallizable), was found to crystallize under low ionic conditions. Nisonoff et al. ( 5 ) and Palmer and Nisonoff ( 6 ) found that pepsin cleavage produced the bivalent F(ab) 2 that upon exposure to reducing conditions could be separated into Fab monomeric units. Further study revealed other subdivisions that are described in Table 1, Fig. 1, and Colorplate 1.

TABLE 1. Definitions of key immunoglobulin structure nomenclature

FIG. 1. Schematic representing the major features of a serum immunoglobulin (i.e., IgG). IgM and IgE have an extra CH domain in place of the hinge. Adapted from Carayannopoulos and Capra ( 153 ), with permission.

Immunoglobulins are glycosylated as they are secreted from the endoplasmic reticulum (ER). These glycosylation sites are illustrated (along with the most common disulfide bonds) in Fig. 2. Most immunoglobulins contain at least one asparagine (or N-linked glycosylation group), while others contain a significant amount of O-linked glycosylation particularly in the hinge. The N-linked glycosylation at Asn 267 is thought to have different orientations between various immunoglobulins (e.g., IgA and IgG) ( 7 ). As we discuss below, these features have important physiological consequences. Transmembrane domains and cytoplasmic tails are present in membranous forms of antibodies. The extent of these structures varies with the immunoglobulin isotype. Finally, both IgM and IgA have a tailpiece important in polymerization that is present immediately carboxy terminal to the last constant region domain of the Fc (Cµ4 and

Ca3, respectively).

FIG. 2. Illustration of the potential glyosylation sites and disulfide bonds in immunoglobulin isotypes. From Putnam ( 158 ), with permission.

The study of myeloma proteins led to a great leap in our understanding of immunoglobulin function. These “single” or “monoclonal” antibodies obtained from the sera of patients with the disease multiple myeloma were used in many of the serologic and biochemical studies of the 1950s and 1960s. They remained the major source of hom*ogeneous immunoglobulins until the development of the hybridoma in 1974. The serologists injected them into animals and produced antisera that were used to detail some of the basic divisions of antibodies. For example, the immune sera were absorbed with other myeloma proteins and were used to identify isotypic, allotypic, and idiotypic specificities ( 8 ). The isotype of an antibody refers to the particular light or heavy chain-constant region that is used. Isotypes are present in all members of a species. The allotype refers to allelic differences in both the variable (particularly rabbit) and constant region. Allotypes are present in some but not all members of a species and are inherited in a simple Mendelian fashion. Idiotype refers to a specificity that is associated with the variable region and generally is a marker for the antigen-combining site ( 9 ). Anti-idiotypic antibodies generally prevent antigen–antibody interaction. Myeloma proteins were also used for the first amino acid sequencing of immunoglobulins and provided our first introduction to the idea of sequence variability (and indeed the definition of) the variable and the constant regions. Finally, myelomas were the first immunoglobulins that were subjected to crystallographic studies and provided the first glimpses of the domain structure of the prototypic immunoglobulin ( 10 , 11 and 12 ).

At first, immunologists thought of the antibody molecule in static terms, but increasingly there has been an appreciation of the motion of the immunoglobulin. The major motions of the Fab can be illustrated in Fig. 3. We now know that rotation about the hinge as well as segmental flexibility varies more than previously thought. The elbow peptide is important in the orientation of the Fv for antigen binding. Finally, flexibility in the Fc (fragment crystallizable) region, particularly between the constant domains as well as perpendicular to the plane of the constant region is increasingly appreciated in certain immunoglobulin–receptor interactions (particularly in IgE) ( 13 ).

FIG. 3. Illustration of the motions and flexibility of the immunoglobulin. Axial and segmental flexibility are determined by the hinge. The switch peptide (elbow) also contributes flexibility to the Fab. The measure of its angle is defined as between the symmetry of the Fv and Fb axes. From Carayannopoulos and Capra ( 153 ), with permission.

The following sections focus on particular structural concepts important to understanding the functions of immunoglobulin beginning with Ig, the core domain of the immunoglobulin superfamily.

THE Ig DOMAIN The immunoglobulin domain (Ig domain) is the central structural unit that defines members of the immunoglobulin superfamily (IgSF) (reviewed in Williams and Barclay [ 14 ] and Harpaz and Chothia [ 15 ]). This domain is composed of two sandwiched ß pleated sheets. Each sheet is composed of an arrangement of ß strands whose

particular composition is based on the type of domain used in the molecule. There are two general types of domains in immunoglobulins, V and C. The ß strand conformation in V-type domains consists of nine antiparallel strands with five strands in the first sheet and four strands in the second. C-type domains have seven antiparallel strands distributed as three strands in the first and four strands in the second sheet. The core of the domain is formed through the b, c, e, f and part of the c’ or c-d loops. These portions comprise 31 amino acid residues. The edges formed of a, g, c’, and c” maintain conformational flexibility (see Fig. 4). Despite differences in conformation, these domains—in the case of most immunoglobulins—share a common set of cysteine residues that form a disulfide bridge linking the two sheets (see Colorplate 2). This disulfide bridge forms the nuclear portion of what is called “the pin region” and provides structural stability to the unit. Disulfide bridges are common to most of the IgSF members but vary in number and placement. Interestingly, in molecular biological experiments in which the disulfide bonds are removed (cysteines replaced by serines), there is remarkably little overall alteration in antibody function. Thus, while almost universal among the domains of members of the IgSF, the disulfide bridges seem like evolutionary “add-ons.” The tryptophan residue that packs against the disulfide bridge is also common to members of the IgSF. Beyond the common cysteine and tryptophan amino acids, the Ig domain can vary widely in the primary amino acid sequence. Despite this variability, however, a common secondary and tertiary structure characteristic to the Ig domain is preserved. The region between the two sheets maintains a hydrophobic character. Nonpolar amino acids occupy most of the positions where side chains are pointing into the domain. Other residues in this area participate in the formation of hydrogen bonds. Residues in the edges of the domain are solvent exposed. The variation among the size of residues that occupy the central portion of the domain are considerable, but instead of being compensated strictly by local conformational changes and complementary mutations, the movement of the sheets relative to each other as well as the insertion of side chains from the periphery provide the majority of the changes required ( 16 , 17 ). These mechanisms allow considerable variation to occur, as in the process of somatic hypermutation while maintaining the structural conformation of the molecule. The Ig domain bears unusual functional properties by maintaining structural stability while providing extreme variability in binding specificity through its loops rather than secondary structural elements, as in the case of the binding site formed by two V domains (heavy and light chains) for antigen. Usually binding domains of protein are considered the driving force of evolutionary conservation, unlike the Ig domain case.

FIG. 4. Schematic of the secondary structural topology of the two major types of Ig domains present in the immunoglobulin (V and C). Horizontal lines are beta strands, and vertical lines are loops connecting them. Lines with large dots represent CDRs. C domains contain 7 strands and V domains contain 9 strands. Residue numbers are according to Kabat et al. ( 154 ). Numbering of beta strands (in parentheses) is according to Edmundson or Hood. From Carayannopoulos and Capra ( 153 ), with permission.

FAB STRUCTURE AND FUNCTION In order to combat a seemingly infinite range of potential pathogens, the humoral immune system is equipped with a highly structured yet extremely versatile weapon—the Fab domain of the immunoglobulin. This domain shows an amazing array of binding capabilities while maintaining a highly hom*ologous scaffold. This section first describes the characteristics of this domain, the relationship of the variable segments, structurally important features, and finally, some of the important characteristics of the antigen interface. The antigen-binding fragment (Fab) is comprised of heavy and light chains that are both divided into a constant region (Fb) and a variable region (Fv). Other than minor allotypic differences, the constant region does not vary for a given isotype in the heavy chain or for each class of light chain (? or ?). However, the variable region exhibits significant plasticity. Gene segments are assembled in an ordered fashion by recombination to encode the Fv but these mechanisms are beyond the scope of this chapter (see Chapter 5). However, it is important, in the context of antibody structure and function to appreciate that two or more genes in virtually every species encode variable regions. Each variable region is approximately 120 to 130 amino acids long, and is generated by two light [L] chain and three heavy [H] chain gene segments. The “V gene segment” encodes the majority of the variable region while the D (H chain) and J (H or L chain) gene segments encode the rest. Multiple V, D, and J gene segments provide ample genetic information, which can be used in virtually every combination to provide the diversity required to respond to a limitless array of antigens (see Chapter 5). When only a few amino acid sequences of Ig variable domains were available, a comparison of a number of sequences led to the observation that some regions of the immunoglobulin sequence are more variable than others. A method of quantifying these differences was derived. Variability for a given residue is defined as the ratio of the number of different amino acids that are found at a given position to the frequency of the most common residue at that position. Thus, a residue that is always present will have a variability of 1, but variability at a position in which all amino acid residues are present at an equal frequency is 400 ( 18 ). By comparing regions of the immunoglobulin in this way, it was found that certain segments of the variable region varied more than others (see Fig. 5). From these early comparisons the concept of “framework” and “hypervariable” regions entered the lexicon of immunology (see Table 2). It was soon hypothesized that the hypervariable regions would play a prominent role in antigen recognition. In subsequent studies, most of the hypervariable regions were indeed the major antigen contact points and they were termed complementarity determining

regions or “CDRs” ( 9 ).

TABLE 2. Residues defining framework and CDR regions of immunoglobulin chains

FIG. 5. Representation of the variability of amino acids in the primary sequence of the human heavy chain variable region. Framework and CDR regions are labeled. The hypervariable regions can be identified as the regions with large grouped peaks. For comparison, cytochrome C variability is also shown. Adapted from Kabat et al. ( 154 ), with permission.

Thus, molecular biology (showing that two or three gene segments generate variable regions), primary amino acid sequence analysis (revealing highly variable and reasonably constant areas within the variable region) and x-ray crystallography came together in the early 1970s to provide us with a view of the antibody variable regions that neither alone could provide: The framework regions determined by protein

sequencing were seen to be the ß strands forming the Ig domain of the Fv and are far less variable than the CDRs, which comprise the loops that make up the majority of the antigen-binding region of the Fab. Thus, three very different disciplines converged to provide an insight that has stood the test of 3 more decades of study largely intact; that is, the hypervariable regions represent those portions of the antibody molecule that directly interact with antigen and the framework regions provide the scaffold for the interaction to take place. The variable regions of both the heavy and light chains are held together through the interaction of frameworks 2 and 4 of the heavy and light chains. Framework 2 in the heavy chain contains a specific sequence (Gly-Leu-Glu-Trp-hydrophobic) that interacts with a light-chain–specific stretch (Pro-hydrophobic-Leu-hydrophobic) in framework 2 as well, to help the two immunoglobulin folds of each chain to properly dimerize ( 19 ). In addition, the sequence Trp/Phe-Gly-X-Gly in framework 4 creates a beta bulge that is necessary for dimerization of the variable heavy- and light-chain domains ( 20 ). Regions in the CH1 domain are important for dimerization and effectively bind the Fab at the opposite end from the antigen-binding domain. The five-stranded, beta sheet face is used as the dimerization surface for the Fb region although only four of the strands participate ( c” is excluded). The three-stranded face generally participates in V-region dimerization ( 21 ). This orientation requires nearly a 180° rotation in comparison to the Fv region that is facilitated by the “elbow peptide” or “switch” region located as a spacer in between the Fv and Fb ( 22 ). Interactions between the Ig domains in both Fv and Fb regions of the Fab in addition to the CH3 or hom*ologous structure have an interrupted alternating hydrophobic and hydrophilic residue patterns that are usually seen for other protein–protein interactions with Ig domains and replace it with hydrophobic residues that form a core for binding ( 10 , 22 , 23 and 24 ). Bulges in the g strand as well as the c’ strands of the variable regions protrude into the interior of the dimer and prevent tight adherence between the two variable regions. A hydrophilic groove is produced, which is lined by residues of the hypervariable region and other CDR loops forming the antigen-binding site. The core hydrophobic regions exist between contacts of frameworks 2 and 4 as mentioned previously, as well as between CDR3s of the heavy and light chain or between framework 2 and the CDR3 of the other variable region ( 25 ). These arrangements provide stable associations between the components of the Fab, maintaining structural integrity of the molecule, while at the same time allowing it the freedom to perform its antigen-binding function and conformational changes that might need to occur to facilitate this capacity. The tight interaction between these roles is illustrated in the difficulty of successfully creating engineered antibodies with framework and CDR regions from separate species ( 26 , 27 ). The sequence similarity among V regions can be used to place them into related groups. Framework-1 amino acid residues 6 to 24 divide them into three clans (see Fig. 6). In a further subdivision, framework 3 residues can be used to distinguish among family members within a clan ( 28 , 29 and 30 ). Families are then divided into individual members based on the rest of the differences in the germline repertoire that divide them (see Chapter 5). At the amino acid level, in general, different families display a similarity of 75% at the amino acid level, but between families display less than a 70% hom*ology ( 31 ). There are seven families of VH genes, four families of V? genes, and ten families of V? genes in the human. In the mouse, these numbers are expanded to currently 14 VH

and 20 V? families. The ability to group this large number of variable region structures into families and clans based on relatively strict similarity requirements reflects the concept that the origin of family members is likely to be through the duplication of an originally smaller set of genes.

FIG. 6. Immunoglobulins can be organized into clans according to their amino acid hom*ology. Clans can be further subdivided into families in different species as illustrated. Adapted from Kirkham and Schroeder ( 30 ), with permission.

While the framework regions exhibit high degrees of similarity, the CDR regions are characterized by their divergence. While some characteristics among families can be seen in parameters such as length, variability is the hallmark of the CDR. Many factors influence the construction of the CDRs. These include the length of the V gene used and the presence of somatic mutations and insertions or deletions that produce a sequence that differs from the germline. The latter two processes occur in the peripheral lymphoid compartments. The conformation of the CDR loops in a three-dimensional context is influenced by interactions that occur with neighboring framework residues ( 32 , 33 ), other CDRs from both the VH and VL chains, and even glycosylation that has been reported at CDR asparagines ( 34 , 35 ). As we will see, not only the immediate context of the antigen-binding region constructed by the CDR loops is important, but changes in the extended structure can also have profound influences upon the affinity of the antibody. The third complementarity-determining region of the heavy chain, HCDR3, lies at the center of the classic antigen-binding site. HCDR3 is the direct product of nonhom*ologous gene rearrangement; D gene segments that have the potential to be read in any one of three reading frames; by deletion; and N addition, which has the potential to introduce a totally random sequence into the antigen-binding site (see Chapter 5). Together, these factors make HCDR3 the focus of somatic diversification of the antibody repertoire ( 36 , 37 ). In practice, however, as Schroeder and his group have shown (see Fig. 7), the sequence composition of HCDR3 is constrained, with a preference for tyrosine, glycine, and serine and underrepresentation of positively charged (Arg, Lys) and hydrophobic (e.g., Val, Ileu, Leu) amino acids ( Fig. 7). Thus, the HCDR3 is enriched for neutral, hydrophilic sequences ( 38 , 39 ). In large part, these preferences reflect nonrandom representation of amino acids in D and J gene segment sequences. Whether these preferences reflect structural concerns or evolutionary selection for a specific range of antigen-binding sites remains a focus of active

investigation ( 40 ).

FIG. 7. A: Amino acid representation in the three deletional reading frames in mouse D segments and HCDR3. The first column for each amino acid corresponds to that amino acid’s frequency of occurrence (in percentages) in all available Genbank protein sequences ( 36 ). The second column shows the amino acid content of the germline sequence of the D segments (in all three deletional RFs) ( 40 ). The third column shows the amino acid representation in adult mouse spleen HCDR3 ( 37 ). Z = Stop codons. B: Representation of three groups of amino acids (in percentages). Shading scheme as in A. C: Hydropathicity of HCDR3. Average hydropathicity of mouse HCDR3 intervals ( 40 ) and their frequency of occurrence (in percentages).

Not only does the variable domain have the capacity to bind to antigens using its antigen-binding site using complementarity-determining regions, but the Fv can also bind bacterial virulence factors without the classical antigen–antibody interactions. Antigens that bind to immunoglobulins (and T-cell receptors) outside the classical binding sites, and therefore react with a large number of different antibodies, are referred to as “superantigens.” Staphylococcal protein A (SpA) is an example of a B-cell superantigen that binds to human VH3-encoded immunoglobulins (Igs) independently of the D- and JH-encoded regions or light-chain sequences ( 41 ). The exact SpA-binding structure formed by VH3-encoded Igs was first elucidated by work done in our laboratory in which we expressed a VH3-encoded Ab in baculovirus that bound SpA and then produced mutant Abs in which regions of the human VH3 Ab were exchanged with those from a mouse Ab of the J558 family—a family not associated with SpA binding. The pattern of SpA binding indicated not only that residues in FR1, CDR2, and FR3 were involved, but also that the three regions were required to interact simultaneously with SpA for binding to occur. When any one of the three regions was replaced with the corresponding region from the nonbinding Ab, SpA binding was severely disrupted. The data indicated that SpA required simultaneous interaction with three distinct regions of a VH3 structure, which together in three-dimensional space presumably formed an extended solvent-exposed surface ( 42 ). The crucial finding of these experiments was that framework residues played a central role in binding. Recent crystallographic studies have confirmed and extended these studies ( 43 ) (see Colorplate 3). Moreover, the VH surface-bound SpA seems to have been conserved in the B-cell repertoires of

amphibian, avian, and mammalian species ( 44 ). In the mouse, Fab-mediated SpA-binding interactions are commonly displayed by 5% to 10% of mature B cells, which express genes from the clan III set of related VH families ( 45 46 , ). Of the murine analogs of human VH3 genes, certain J606-, 7183-, and DNA4-encoded VH regions commonly convey binding activity, while VH encoded by products of clan III/S107 VH genes commonly convey among the highest affinity for SpA, and this binding activity is independent of specific VL region usage (reviewed in Silverman and Goodyear [ 47 ]). Thus, once again we see a duality in structure/function. A B-cell superantigen binds to certain VH genes primarily through interaction with framework residues, and at the same time, the same B-cell superantigen binds to the Fc region of certain immunoglobulins.

ANTIGEN–ANTIBODY INTERACTIONS Recent studies have begun to refine our understanding of the types of interactions that occur at the antigen–antibody interface. These studies have come out of a body of work devoted to visualizing antigen–antibody complexes at resolutions down to 1.7 Å, at which levels the role of solvent is being elucidated. Some of these studies have been performed on monoclonal antibody interactions with their haptens in both complexed and uncomplexed states and with both germline and high-affinity configurations. In addition, stepwise manipulation of the antigen-binding sites by single and grouped sets of mutations has also been performed. The lessons learned through these studies are numerous and represent an exciting development in the study of antibody structure. For example, the differences in antigen–antibody interactions in germline versus high-affinity, somatic, mutated counterparts provide a new glimpse into the nature of these interactions. The difference in the affinity between these two forms has approached a 30,000-fold higher affinity in the mutated antibody than in the germline antibody. This enormous change for some antibodies in their affinity for antigen arises from the small additive changes that the mutations have contributed. Three common themes for the nature of the role of mutational differences in the forming of an affinity-matured antibody have arisen. The first is the contribution of direct interaction of the mutated base with the hapten. In some circ*mstances, a base-pair substitution contributes, for instance, a new hydrogen bond or creates a local hydrophobic region that is additive to the affinity. Furthermore, somatic mutation can result in an amino acid substitution that alters the flexibility of the antibody-binding site. This is illustrated in studies of both antigen-bound and antigen-unbound germline and high-affinity antibodies ( 48 ). Germline antibodies often undergo a localized antigen-combining, site-conformational change as antigen binds. The movement can approach the range of 4 to 5 Å for some antibodies. This is in contrast to what is seen with high-affinity antibodies to the same antigen in which the combining site has very little movement on binding. Finally, somatically mutated antibodies versus germline antibodies often show differences in the geometry of the hapten in relation to the antibody-combining site. These changes in orientation are the result of the addition, deletion, or replacement of

interactions not only between the hapten and antibody but also between the peripheral loops and the most proximal loops to the hapten in the combining region ( 48 , 49 and 50 ). It should be noted, however, that some high-affinity antibodies have been reported that undergo exceptionally dramatic conformational changes upon binding to their ligand ( 51 ). Taken together, these studies reveal new depth to our ideas about antigen–antibody interactions. While we normally think of antibody–hapten complexes as a single hand-in-glove fit, this is not necessarily the case. Recent studies have also indicated that several high-affinity conformations are possible and occur for a given antigen–antibody interaction. In other words, redundancy can exist for high-affinity antibodies ( 52 , 53 ). In addition, there is some evidence for the ability of an antibody to bind to a region on antigen that is not necessarily solvent exposed. Small, localized conformational changes may therefore occur in antigen that allow these regions to be exposed perhaps for only brief periods of time ( 54 ). Such an activity is reinforced by the nature of catalytic antibodies. At this point, there are over 20 different antigen–antibody crystal structures. While this represents one of the largest numbers of crystal structures within a family of proteins, few of the structures are at a resolution high enough to solve questions related to the role of water in the interaction with antigen as well as the thermodynamic questions that arise from its influence on the system. The recent publication of a few much-higher-resolution crystal structures is already beginning to demonstrate the role of water in these protein, carbohydrate, and DNA interactions. The residues involved in the combining site include amphipathic amino acids such as Tyr and Trp at a high frequency. Water molecules in a number of studies have been shown to be present and involved with antigen interactions with the crystal structures. In general, water in areas where close protein–protein contacts are not occurring has been shown to participate in additional hydrogen bonds other than those that exist directly between antigen and antibody ( 55 , 56 and 57 ). The presence of water molecules in these areas then has the capacity to increase the interactions particularly for antigens that fit less well into the combining site. These interactions have been observed for both carbohydrate ( 58 , 59 ) and protein haptens ( 55 ). In other cases as well, water molecules bound to the surface of either antigens or antibodies are excluded. The release of these molecules participates to increase randomization in the solvent environment and provide for another means to enhance the affinity of antigen binding ( 57 ). The primary view then of the role of water seems to be to provide a better “fit” for antigen by filling space in the binding region and to participate in extended hydrogen bonding. This contribution increases the enthalpy of the system and, in combination with hydrogen bonds between the interacting proteins as well as van der Waal forces, contributes to an enthalpy-driven antigen–antibody interaction. While this view of the presence of water contributing to an enthalpy-driven reaction is common to many studies, it should not be discounted that the exclusion of certain water molecules may play a role in driving the strength of affinity through entropic means. Studies of HIV protease inhibitors as well as some antibody–antigen interactions as indicated above suggest that the increase of randomization of the solution will contribute to the affinity of an antibody and may be worth pursuing as an additional strategy in antibody

engineering studies (



Antigen recognition, as we have seen and as will be elaborated further in subsequent chapters, depends on diversification from a number of processes. V(D)J recombination, somatic hypermutation, gene conversion, and other such mechanisms generate nearly an infinite variety of molecules designed to recognize antigen. Class-switch recombination from “upstream” to “downstream” isotypes results in the generation of an antibody with the same antigen recognition capacity but different effector capacities to facilitate antigen elimination. The various immunoglobulin classes have certain unique properties that when taken together allow for a wider range of host defenses than would be possible if only a single class of heavy chain constant regions existed. This is illustrated in the breakdown of immune defense as seen in hyper-IgM syndrome (the only isotype present in the patient is IgM) or IgA deficiency (the most common immunodeficiency in humans, complete absence of IgA). While many of the differences in Ig function can be localized to the CH2, CH3, or (if present) the CH4 domain, surprisingly, many of the structural properties of these classes can be attributed to the hinge region.

THE IMMUNOGLOBULIN HINGE While not all immunoglobulin classes have a hinge that is separately encoded (see Chapter 5), all Ig classes have a structure that recognizably fulfills its function, albeit to various degrees. Classes that do not have genetic hinges use an extra C domain in its place. The genetic hinges that are encoded in the other classes have a great variety of lengths and structural properties. The most dramatic of these, IgG3, serves as an illustration of the construction of the hinge. The IgG3 hinge is divided into upper, middle, and lower regions that can be separated based on both structural (amino acid sequence) and genetic components. Structurally, the upper hinge (UH) stretches from the C terminal end of CH1 to the first hinge disulfide bond. The middle hinge (MH) stretches from the first cysteine to the last cysteine in the hinge. The lower hinge (LH) extends from the last cysteine to the glycine of CH2 ( 60 ). The cysteines present in the hinge form interchain disulfide bonds that link the two immunoglobulin monomers. Table 3 compares some of the structural features of hinges from various isotypes.

TABLE 3. Properties of hinges in IgG, IgA, and IgD

The structural differences among the hinges are reflected in the various properties of

the heavy chains. In a simplistic way, one can think of the hinge as the structural unit that links the functions of the Fab and Fc fragments. With greater flexibility, antibodies can bind antigens on the surface of targets with varying degrees of distance between them. In addition, the steric position of these two components may directly affect Fc binding to cellular receptors. Using a similar argument, the hinge may also be involved in the modulation of complement binding (see below). These hinge properties will be explored in more depth by focusing on the individual isotypes. The motion about the IgG hinge has been extensively studied and serves as a basis of comparison for other isotypes (reviewed in Schumaker et al. [ 61 ]). With the exception of IgG3, IgG hinges are encoded by single exons (see Chapter 5). The upper and lower portions of the hinge are the most flexible and allow for motions such as bending between the Fc and Fab in both parallel and perpendicular planes. In contrast, the middle hinge is a rigid structure that is thought to provide spacing between the Fab and Fc domains. It can be seen from Table 3 that this structure is greatly extended in IgG3 compared to the other isotypes. It has been shown that for some functions, this large hinge region can even replace a missing CH domain. In addition, rotation about the long axis of the Fab leads to additional levels of flexibility. A number of electron microscopic studies using immune complexes of IgG show that the Fab–Fab angles range from a very narrow “Y” with an apparent separation of 10° to a “T” with angles of 180°. An addition rotational flexibility of up to 180° is also required to account for some of the observed complexes ( 62 ). Thus, the remarkable conformational plasticity of the immunoglobulin molecule allows it to bind epitopes spaced at various distances on the surface of a target. In addition to altering the Fab angles, the flexibility of the hinge plays a role in Fc function. While recent evidence suggests that other parts of the Ig constant region influence complement binding more than thought in the past, the role of the hinge still appears to be critical. In general, the more flexible hinges allow less steric hindrance and better binding of complement. The greater flexibility of some hinges allows them to either expose or sterically hinder certain complement-binding regions in CH2. While once thought to be a sequence-specific interaction between complement and portions of the CH2 domain, it has been recently shown that the accessibility of the site is more important for determining the activation ability of the various Ig subclasses. This can be illustrated by the trend of complement activation to follow the order of IgG3>IgG1>IgG4>IgG2 (reviewed in Brekke et al. [ 63 ] and Feinstein et al. [ 64 ]) in the human (unless otherwise specified all subclass designations will refer to human immunoglobulins). Following a similar trend, hinge flexibility and relief of steric hindrance have also been shown to modulate to some degree the binding of the CH2 domain to certain Fc receptors ( 65 ). While tip-to-tip separation of the Fabs in IgG structures varies from 13 to 16 nm, IgA1 has a spread of up to 23 nm. This additional distance that IgA1 is able to span may confer advantages for more efficient recognition of epitopes that are widely separated. IgA2, of decreased abundance in the serum (see below), has a hinge with a length equivalent to the shorter IgG subclasses. A study of the differences between IgA1 and the IgG subclasses is instructive about several aspects of the hinge. The extended

structure of the IgA1 hinge is a combination of two properties: the abundance of O-linked glycosylation coupled with the location of the disulfide linkages between the Ig monomers near the top of the CH2 (Ca2) domain ( 66 ). Most other isotypes contain some N-linked glycosylation. In immunoglobulins, O-linked glycosylation is unique to IgA1 and IgD. This extensive glycosylation has two potential advantages for these molecules. The extended hinge of IgA1 is probably protected from proteolysis by many bacterial enzymes because of its glycosylation. There are several pathogenic bacteria that exploit two contiguous, repeating amino-acid octamers consisting of proline, serine, and threonine as a binding site for IgA1 proteases—an important virulence factor. It is thought that the resistance to common bacterial proteases allows IgA1 as a major secretory isotype to survive among the flora that also colonize the mucosa (reviewed in Kilian et al. [ 67 ]). Coupled with the separation of Fab arms, this extensive glycosylation most likely also aids in separating the Fab domains. Eliciting a similar effect, the disulfide bonds at the top of the CH2 domain rather than in the hinge itself allows the extended structure to give the molecule a greater spread than could be accomplished if disulfide bonds were contained within the hinge itself. The other IgA isotype, IgA2, lacks the extended hinge that IgA1 possesses and is also not heavily glycosylated ( 66 ). However, the smaller hinge in IgA2 does not contain the proteolytic motifs that are recognized by the enzymes produced by certain bacteria in IgA1 and may therefore be maintained in the isotype repertoire for this specialized niche. This may explain why, in serum, the ratio of IgA1:IgA2 is about 6:1, but in most secretions, it is close to 1:1. The hinge of IgD, like IgA1, has extensive O-linked glycosylation on an extended hinge structure. The hinge is divided into two major subregions (and encoded by two exons) that are either rich in alanine and threonine or glutamate and lysine. This latter subregion is highly sensitive to proteolytic enzymes and is even sensitive to a yet unidentified enzyme and has thus been dubbed “spontaneously” proteolytic ( 68 , 69 , 70 and 71 ). Both IgA1 and IgD possess another characteristic that is linked to their O-linked glycosylation. Jacalin (jackfruit lectin) binds to these O-linked carbohydrates with high affinity and can be used to precipitate these specific isotypes. Interestingly, the cell-surface receptor for human IgD binds to these O-linked oligosaccharides. This receptor binds both IgD and IgA1 ( 72 , 73 and 74 ). Thus, this is an example of the hinge region and in particular the oligosaccharides of the hinge region playing a critical role in cellular binding.

Fc STRUCTURE AND FUNCTION While the hinge is essential to modulate many properties of the immunoglobulin, the Fc portion is the primary effector domain of the molecule. While the antigen-binding function of the Fab domains allows the immunoglobulin to specifically recognize diverse antigens, the Fc domain allows an antibody at the same time to elicit host responses. This requires that the Fc region provide binding sites for both cellular receptors and complement—the two primary effector response types to antibody–antigen complexes. This property requires that the Fc domains maintain considerable conservation especially in structural support regions. While there is substantial conservation within species (some allotypes vary in function), many “Fc functions” seem to be conserved with only modest structural similarity—especially within the primary amino acid sequence. In humans, the subclasses are even closer in structure, being over 90%

identical in amino acid sequences. (The degree of difference between the IgG isotypes varies widely in mammals—in the human, the IgG subclasses are among the most closely related and presumably are of very recent evolutionary origin.) However, the regions between the conserved Ig domain structures (generally small loops) also serve as a target region for the binding of certain bacterial and viral virulence factors and can be involved in binding to cell-surface receptors (see below). Each constant region consists of 3 CH domains for IgM, IgA, and IgD or 4 CH domains in the case of IgM, and IgE. In the latter isotypes, the CH2 domain replaces the hinge structurally and to some degree functionally. Each CH domain contains a core of an Ig domain with 7 antiparallel beta sheets oriented by 3 in one direction and 4 in the opposite direction. This is in contrast to the structure of the Ig domain in the V region. Looking at all mammals, in general there is approximately 30% amino-acid sequence identity between the constant regions (IgM, IgD, IgG, IgA, IgE) and 60% to 90% hom*ology among the subclasses. The majority of the hom*ology is present in the ß strands forming the Ig domains, disulfide-bonding cysteines, and tryptophans. CH domains are numbered from the first domain located in the Fab and positioned above the hinge, to the CH2 and CH3 domains that are increasingly distal to the hinge. The CH domains contain several general features that contribute to the structure of the Fc region. N-linked oligosaccharides are positioned in the middle of the CH2 domain that protect a hydrophobic patch in this region, and therefore increase the solubility of the molecule. For IgA, this N-linked glycosylation is thought to be located near the base of the CH2 domain ( 75 ). Longitudinal contact between CH2 and CH3 prevents binding between monomer chains at this junction. The CH3 domain uses 4 strands of the ß sheet to dimerize between chains. In IgA and IgM, a tailpiece is added to this domain to create higher-order structures. Glycosylation is an important component of immunoglobulin molecules and the isotypes vary in the extent and type of glycosylation present (see Fig. 2). Between 3% and 17% of the mass of an immunoglobulin is due to glycosylation. While the pattern of glycosylation varies among isotypes, certain conserved sites are preserved. The N-linked glycosylation on Asn297 is conserved for all mammalian IgGs and hom*ologous portions of IgM, IgD, and IgE. This oligosaccharide is thought to project from the inner face of the CH2 domain ( 7 ). Major characteristics of the immunoglobulin isotypes are listed in Table 4.

TABLE 4. Properties of immunoglobulin isotypes

IgM IgM is the most versatile of the antibody classes. It is first expressed as a surface immunoglobulin on immature lymphocytes and as such is the first “B-cell receptor.” B-cell maturation is critically dependent on the presence of immunoglobulin on the surface ( 76 ). The µ chain is the first to be produced upon heavy-chain rearrangement. Initially, the µ chain is expressed with a surrogate light chain, which allows a B cell to continue maturation in the bone marrow. Finally the µ chain is paired with a functional light chain and the naïve B cell leaves the bone marrow ( 77 , 78 ). In the periphery, IgM can be expressed by immature, mature, memory, and plasma cells. Of these, expression on immature and maturing cells is the most common where it remains as a surface receptor. Its presence on the surface of these cells provides a receptor for B-cell activation along with the Iga and Igß accessory molecules ( 79 ). Following activation, the B cell undergoes the critical process of affinity maturation. As well, when these cells enter peripheral lymphoid tissues they also acquire IgD by differential RNA splicing. Thus, the two surface receptors (IgM and IgD) have the same antigen-binding capabilities. The Cµ4 domain contains the transmembrane and cytoplasmic regions of IgM that undergo RNA processing to be removed for the production of secreted IgM. While the membrane-bound form of IgM is most common, IgM plasma cells secrete polymeric IgM that serves important functions as well. Polymeric IgM is an important complement activator, and thus participates in phagocytosis. IgM forms hexamers or pentamers, the latter upon the incorporation of a J chain, arranged in a star pattern with the Cµ4 domain at the center. Cµ4 and part of Cµ3 have been implicated in the formation of this structure, including certain aspartic acid, lysine, and histidine residues. (The next section on complex immunoglobulin structures contains further discussion of this structure.) While monomeric IgM itself has low affinity for antigen, in its polymeric form it has considerable avidity for antigen. It is this increased avidity that makes IgM an important complement activator and mediator of opsonization. The Cµ3 domain binds C1q with the essential participation of its carbohydrate residues ( 80 ). Aspartic acid, lysine, and proline residues in two clusters have been implicated for this activity in the mouse. However, the only hom*ologous region to be shown with this activity between IgG and IgM is a single proline. Cµ1 as well interacts with the C3b ( 81 ) complement component and helps to mediate phagocytosis of opsonized antigens by macrophages. IgM is second only to IgA in its contribution to mucosal immunity. It can be secreted through similar means by polymer association with a J chain, and like IgA, is transported

by the polymeric Ig receptor (pIgR). In many patients with IgA deficiency, IgM adequately substitutes for IgA in mucosal protection. IgD IgD is perhaps the most enigmatic of the immunoglobulin isotypes. It is present on all naïve B cells and serves as a better receptor in terms of activation than IgM. It also requires the co-expression of Iga and Igß to elicit a cellular signal due to its short lysine-valine-lysine cytoplasmic domain identical to that of IgM. However, other participating co-receptors differ between the two membrane-bound isotypes. Coexpression on naïve B cells of these two receptors occurs by differential mRNA splicing ( 82 ) (see Chapter 5). Two laboratories have made IgD knockouts and the phenotype is somewhat ambiguous. Although there are fewer lymphoid follicles and overall a slower process of affinity maturation, one does not see defects that would otherwise indicate an irreplaceable role ( 83 , 84 and 85 ). Yet at the same time, ligation of IgD can activate, delete, or anergize B cells independent of IgM ( 86 , 87 and 88 ). In mice, the overexpression of IgD creates a greater induction of APCs, up-regulation of B7-1 and B7-2, and increased class switching. Despite its short serum half-life and less abundant mRNA, membrane IgD density exceeds that of membrane IgM on naïve B cells. This phenomenon is thought to be due to a greater stability of IgD mRNA than IgM ( 85 , 89 , 90 and 91 ). Through the process of maturation, most B cells lose IgD expression ( 92 , 93 ). However, there are notable exceptions. IgD-only plasma cells are present in various compartments and secrete IgD into the serum. These plasma cells are in high concentration in the nasal mucosa ( 94 , 95 , 96 and 97 ). The serum half-life of IgD is quite limited, however, being only about 2.8 days. As mentioned before, the extended hinge of IgD is prone to proteolytic enzyme activity and this sensitivity extends into the C terminus of the Cd3 domain as well. While the Cd1 and Cd2 domains are similar in structure to that of other isotypes, Cd3 lacks several key proline residues that play structural roles in the loops between beta strands. In addition this domain contains two N-linked carbohydrates at asparagines 316 and 347 that are not present in other immunoglobulins in this location. These structural differences appear to play a role in binding of IgD to the IgD receptor, at least in the mouse (recognizing the N-linked sugars of Cd1 and Cd3 in a Ca 2+-dependent manner), but may have other functional properties in the human where the receptor binds to the Jacalin-binding domains (O-linked sugars) of the hinge without the requirement of calcium (reviewed in Preud’homme et al. [ 91 ]). IgD plasma cells and a particular set of IgD+IgM- B cells from germinal centers exhibit an unusually high level of somatic hypermutation ( 98 , 99 and 100 ). In addition, it has recently been found that this population of cells is quite prone to a VH-region event called receptor revision ( 101 ).

IgA IgA is the major isotype of mucosal secretions. In addition, it is also the most prominent isotype in colostrum and breast milk ( 102 ). A number of features make this molecule suitable to the mucosal environment. First, the secreted forms are dimerized by their tailpieces and stabilized by J chains. The polymeric Ig receptor (pIgR) transports dimeric IgA across the epithelial barrier where a portion of the pIgR is cleaved to result in the formation of the secretory component. Secretory component remains attached to dimeric IgA, and J chain dimerized IgA with a secretory component are called “secretory IgA.” There are two subclasses or isotypes of IgA in humans (IgA1 and IgA2) ( 102 ). IgA2 is the main component of secreted IgA ( 103 ) and as noted earlier has a truncated hinge that is resistant to most bacterial proteases ( 104 ). IgA is an important component of the first line of defense from organisms entering by mucosal routes. While most secretory immunoglobulin is IgA, it also accounts for 10% to 15% of serum immunoglobulin, making it the third most plentiful. Serum IgA tends to be mostly IgA1 ( 103 ). As mentioned earlier, IgA1 has an extended hinge and can bind antigens at a variety of spacings. In addition, extensive O-linked glycosylation prevents cleavage by most bacterial proteases. Serum and secretory IgA are derived from separate pools of B cells, but antigenic exposure at any given site primes the development of both secretory and serum IgA ( 102 ). Inflammatory responses are not efficiently generated upon antigen binding with IgA. Such a response would most likely be damaging to the mucosa. Instead, IgA elicits protection primarily through exclusion, binding, and cross-linking of pathogens. As well, IgA has been shown to be able to fix the complement by the alternative pathway and most recently by the lectin-binding cascade. Targets that are opsonized by IgA are removed by FcaR-mediated phagocytosis. FcaR, although more distantly related, is most similar in structure to the Fc?RII and FceRI receptors. However, its binding site on the IgA Fc does not follow the pattern in either of these other two receptors. It has been shown that the FcaR (CD89) binds to regions on the loops present between the Ca2 and Ca3 domains ( 105 , 106 ) (see Fig. 8). Mutation of certain residues in these regions eliminates FcaR binding. In addition, N-linked glycosylation at Asn263 appears necessary by some accounts for the interaction between IgA and its receptor ( 105 ). Others have found that this interaction may not be necessary ( 106 ). While we do not as yet have an x-ray structure of IgA, it has been postulated that the oligosaccharide at this location, unlike IgG, is oriented with the glycosylation pointing away from the cavity formed by the heavy chains and therefore released from protein interactions that otherwise would prevent its further modification. Indeed, these oligosaccharides exhibit more sialation than similar oligosaccharides in IgG ( 107 ).

FIG. 8. Illustration of the residues essential for the binding of IgA to FcaR (CD89). Residues represent mutations made in IgA constant, heavy-chain regions as mapped on an Fc? fragment. Of the residues, mutated L465 and L266 were found to be important for binding to CD89. From Carayannopoulos et al. ( 105 ), with permission.

An NMR solution structure of IgA1 has been obtained with molecular modeling to IgG and has provided some insights into the differences between the IgA1 hinge and the IgG1 hinge (murine) ( 66 ) (see Colorplate 4). The placement of interchain disulfide linkages between the immunoglobulin monomers is also a feature that sets IgA apart from other isotypes. Most isotopes contain interchain disulfide bonds in the hinge or, in the absence of the hinge, in the CH2 domain that replaces it. In IgA, however, these bonds are made at the top of Ca2 below (or carboxy-terminal to) the hinge. While it has been confirmed that the Cys241–Cys241 bridge is common among the IgA molecules studied, the other three or four cysteine–cysteine disulfide bonds that form out of a possible six candidates is unclear at the present time ( 66 ). Two cysteine residues remain exposed (one being Cys311), which are the likely to be covalently bonded with J chains. Cys471 forms another interchain disulfide bond between monomers outside of the cluster mentioned above ( 107 ). IgE IgE is the least abundant of all of the immunoglobulins. It is present 25-fold to 3,000-fold less than the other isotypes and has the shortest free serum half-life. IgE is primarily produced in plasma cells in the lung and skin. It is quickly taken up by the high-affinity FceRI. Thus, while the serum half-life of IgE is short, it remains for several weeks or months attached by this receptor to the surface of basophils and mast cells. Once a multivalent antigen is bound to the IgE–IgE receptor complex, the release of inflammatory mediators such as histamine and chemoattractants results in violent reactions (see Chapter 45). These reactions are involved in the clearance of parasites ( 108 ) and are intimately involved in allergy and anaphylaxis. People who suffer from atopy have an inappropriately high synthesis of IgE and almost always a high serum level of

IgE. Mast cells and basophiles express the high-affinity FceRI receptor, as do Langerhans cells and eosinophils, although the reason for its presence on the latter two cell types is unknown ( 109 , 110 ). The interaction between IgE and the FceR is among the strongest known with a k d of 10 -9 to 10 -10 M. There are two distinct binding sites on IgE for its high-affinity receptor. They use identical residues from each Ce3 ( 111 ) with some involvement of Ce2 ( 112 ). The glycosylation status of IgE does not seem to play a role in this interaction. These bound regions of IgE are found on the surface loops of the Ce3 domain. Although there are two sites on the receptor, a 1:1 stoichiometry is maintained and both crystal structures of the interaction as well as biphasic dissociation rates in kinetic studies show that both sites are involved in IgE binding to its high-affinity receptor. Several aromatic amino acids as well as a buried interface surface contribute to the stability of this interaction ( 111 ). A large conformational change has been postulated to take place in the Ce3 domain as IgE is bound to its receptor ( 13 , 113 ). Site 1 appears to provide specificity for IgE in binding to its receptor, while site 2 appears to contain certain conserved residues between IgG receptors and high-affinity IgE receptors. Pro426 from the IgE Fc is sandwiched between FceRI residues Trp87 and Trp110. These three residues are absolutely conserved between IgG, IgE, and their receptors in the binding sites. Leu425 is also absolutely conserved. This illustrates the relation of the IgE and IgG receptors as well as the utility of the “proline” sandwich motif between the immunoglobulin and its receptor, which themselves are members of the Ig superfamily ( 111 ) (see Fig. 9).

FIG. 9. Certain residues are conserved between Fc?Rs and FceRI as well as IgG and IgE that facilitate binding. Two sites participate: site 1 in (a) and site 2 in (b). Heavy lines indicate the highest number of contacts and dashed lines indicate the least. Of considerable note are residues W87 and W110 in site 2 of the receptors and P426 in the immunoglobulin that form a core “proline sandwich” in the interaction between immunoglobulin and receptor. From Garman ( 111 ), with permission.

IgE also binds to a low-affinity receptor, FceRII (CD23). This receptor is a type-II integral-membrane glycoprotein that is involved in a number of activities. The low-affinity IgE receptor binds IgE with 1,000 times less affinity than the high-affinity receptor. Unlike the high-affinity receptor, the low-affinity receptor is expressed on monocytes ( 114 ). Ce3 is essential for binding of the low-affinity receptor, and a major determinant of binding appears to be Lys352 in the AB loop. IgG IgG is the most abundant isotype in the blood as well as in the lymph and peritoneal fluids. Seventy-five percent of the serum immunoglobulin is comprised of IgG. IgG has a long half-life of 3 weeks in the serum. This makes it the most stable antibody in the serum ( 115 ). High-affinity IgG signifies the humoral immune response. Among all isotypes, IgG may at first appear to be a bit bland in its function due to the absence of special properties, such as multimerization and secretion like IgM and IgA, or enigmatic roles like IgD, or extremely high-affinity interactions with receptors and unique modes of effect like IgE. But such a conclusion about this molecule would be an oversight. Of all isotypes, IgG has been the most studied structurally. This is the result of early crystal structures of two hinge-deletion mutants ( 116 , 117 and 118 ). More recently, two complete antibody structures have been reported of murine IgG1 and IgG2a ( 119 , 120 ). The most obvious lesson these structures have taught us is that the mobility of an otherwise perceived static molecule is quite striking. IgG must simultaneously bind with very high affinity to three independent sites in order to effect its immune response function. These recent crystal structures will help us continue to refine our view of the immunoglobulin molecule and will serve as a basis of comparison for structures of immunoglobulins of other isotypes that are most likely not far from being solved. IgG subclasses bind and activate the complement with different efficiencies, as discussed previously ( 121 , 122 ). However, all subclasses carry a core C1q binding motif at Glu337, Lys339, and Lys341 on the fairly mobile C?2 domain ( 123 ). This indicates that the variability is due rather to the steric properties that the various hinge conformations impart. The presence of carbohydrate on IgG has been shown to be absolutely necessary for complement activation (reviewed in Furukawa and Kobata [ 124 ]) and galactosylation has been shown to be especially important ( 125 ). Bacterial proteins A and G have been classically known to bind IgG. This occurs at the C?2–C?3 junction involving residues 264 to 267 ( 24 ). These residues are consistently oriented between structures to C?3 in the same manner. In addition residues 330 and 465 are important as well. These regions overlap the binding site of the neonatal Fc?Rn and, in addition, may produce inhibition of other Fc?Rs ( 120 ). Fc Receptors IgG can bind to four types of receptors. These receptors vary in their affinity for IgG as well as their expressed location. We will discuss the receptors as two

major groups. The first group consists of the high-affinity and lower-affinity receptors that are IgSF members. These receptors are the Fc?RI (high affinity), Fc?RII, and Fc?RIII. The second contains the neonatal IgG receptor (Fc?Rn) that is related to the MHC class I. The IgSF family of receptors, although differing in portions of their binding sites, shares an important motif. A “proline sandwich” is produced between Pro329 and two tryptophans in the receptor. This motif is even shared with the hom*ologous high-affinity IgE receptor described previously in this chapter (see Fig. 9). In addition, the IgG receptors show a dependence on residues Leu234–Pro238 of the lower hinge ( 126 ). Although the individual amino acids involved vary between the receptors, this appears to be another common interaction. These lower-hinge residues account for four of the six hydrogen-bond interactions in crystal structures depicting Fc?RIII interactions with IgG1 ( 127 ). Van der Waal interactions are plentiful between the receptor and the Fc. Finally, all receptors are dependent as well on the presence of a carbohydrate at Asn297 although the interaction is not direct. This carbohydrate is thought to stabilize the lower hinge by producing a hydrophobic core in the (C?2) domain by filling its cavity ( 126 , 127 , 128 and 129 ). Elimination of the branching mannose residues from the glycosylated IgG Fc produce a linear trisaccharide core that severely decreases affinity for the Fc?RII, indicating the importance of this structure for proper conformation of the Fc ( 126 ). The other regions involved in binding with these receptors are varied but important to the individual receptors. Fc?RI binds to IgG1 Fc with 100 times greater affinity than the other IgG receptors. Residues within the stretch Gly316–Ala339 have been mapped with differential importance to binding interaction. In addition, a separate chain on the Fc?RI receptor is involved in augmenting the binding chains of the receptor without making direct contact with the IgG1 Fc. While the Fc?RI receptor uses essentially the same region for binding as the other receptors, the difference in affinity may be attributable to either conformational changes that are made, or differences in particular amino acids used for the actual binding of the receptor. Fc?RII has been shown to require the presence of two identical IgG heavy chains in order to elicit binding. Residues in the loops of the C?2 domain are important for this interaction in addition to the lower hinge. For Fc?RIII there are two important binding regions. The first are the class-1 residues of the hinge proximal region of C?2. The second is the C?2–C?3 interface. For both Fc?RII and Fc?RIII, several residues at the “bottom” of the C?3 domain influence the binding to IgG ( 126 ). The Fc?Rn (neonatal receptor) has the interesting property of binding maternal IgG and transporting it across the epithelia of the placenta ( 115 ). Binding of the IgG occurs in the cells of this barrier at a pH of 6.5. Then the Fc is released by the receptor into the blood at a pH of 7.4. This sharp pH dependence is a function of the titration of ligand residues on the Fc of IgG. Several histidine residues in the C?2 and C?3 interface are involved that bind negatively charged residues at acidic but not basic pH. A further interesting element is the structural relationship of this receptor to the MHC class-I molecule ( 130 ).

HIGHER-ORDER STRUCTURE While IgM and IgA have activities as monomeric immunoglobulins, both have the ability to form multimeric structures that fill yet other biological niches. IgA usually forms dimers through its tailpiece, an extra 18 amino acids at the end of the Ca2 domain, and, as noted above, is complexed with another B-cell protein, the J chain ( 131 ). This complex can then be bound by the polymeric Ig receptor (pIgR) and transported across

the mucosal epithelial layers to provide important primary immune defense roles ( 132 ). This dimeric IgA can then bind with greater avidity to polymeric epitopes to increase its effectiveness in eliminating these targets from the mucosal surface. Similarly, IgM also forms a polymer, but is most commonly in the form of a pentamer. This configuration allows IgM to bind polymeric low-affinity epitopes and efficiently activate the complement to opsonize and eliminate its target. While the J chain is present in most IgM pentamers and binding to the pIgR is possible, dimeric IgA is the primary antibody of most mucosal surfaces. Polymeric IgM has significant activity in the serum. The J chain is a 137 amino acid/15kDa protein that serves to link two immunoglobulin monomers covalently ( 133 ). It contains eight cysteine residues that participate in a disulfide bond with each tailpiece in addition to stabilizing its own structure ( 134 ). Whether the J chain forms other disulfide bonds with the immunoglobulins is still not clear. The J chain is a highly conserved molecule among a range of species and even predates the presence of the antibody ( 135 , 136 , 137 and 138 ). While thought to be arranged as a single domain in a beta barrel formation, it has not yet been crystallized and does not show sequence hom*ology to an immunoglobulin domain. The J chain is proteolytically labile and contains a high amount of negatively charged residues ( 134 ). Both tailpieces of each IgM and IgA carboxy-terminal domain contain a cysteine (the penultimate cysteine residues 575 and 495, respectively) that are involved in multimerization. One of these residues from each immunoglobulin is paired “at the tail” to form a direct disulfide bond between the monomers. The other two residues (one from each monomer) bind to separate cysteines in the J chain ( 139 , 140 , 141 and 142 ). The tailpieces are thought to form two extra beta strands on one face of the terminal domain that facilitate this interaction ( 143 ). In addition, there is evidence that Cµ3 and Cµ4 of IgM, as well as the hom*ologous regions in IgA, are also required for interaction with the J chain ( 144 ). The structure of the carbohydrate at Asn563 (in IgM and at the hom*ologous region in IgA) is also important. Usually this carbohydrate contains a large amount of high mannose glycans, which indicates that it is protected by polymerization occurring before exposure to the Golgi-complex enzyme, mannosidase II ( 145 ). In the case of IgA, the J chain is required for polymerization, although some reports of multimers in its absence have been reported. It exists in all forms of IgA polymers (not just dimers), including, as well, some reports of the secreted monomers. Domain-swapping experiments have shown that the propensity for IgA dimer formation and its binding by the J chain is due to the presence of the IgA tailpiece in the context of its own heavy chain. Tailpieces spliced to IgA from the µ chain result in higher-order multimers than the simple dimer ( 144 ). IgM, unlike IgA, has no requirement for the J chain in its polymeric forms, although J chain is often present and essential for secretion. Two other disulfides besides the penultimate disulfide mentioned above are involved in the formation of multimers. Like Cys575, Cys414 forms intermonomeric disulfide bonds. Cys377 is most likely to form intramonomeric bonds. While pentameric IgM is the most common form of IgM, there are many reports of IgM hexamers ( 146 ). The latter almost never have incorporated the J chain, but are highly dependent on Cys414–Cys414 bonds between monomers.

Hexameric IgM is rarely found in vivo except in the case of cold agglutinin disease and Waldenstrom’s macroglobulinemia ( 147 , 148 and 149 ). Hexameric IgM has been reported to be far more efficient at complement activation than pentameric IgM ( 146 ). Pentameric IgM is regularly associated with one or more J chains. It has been hypothesized that pentameric IgM with J chains is more thermodynamically favorable than hexameric IgM as a possible explanation of why it is more common. Thus, for the formation of multimeric immunoglobulin, we see that the presence and sequence of the tailpiece is important in the context of the proper heavy chain. IgM tailpieces incorporated into IgA will cause higher numbers of IgA multimers, but the reverse substitution does not induce dimers in IgM. It has therefore been proposed that IgM polymerization is more efficient than IgA ( 143 ). As mentioned above, a J chain is essential for the secretion of IgA and IgM. The pIgR receptor binds to the J chain, and through clathrin-coated vesicle transport, moves dimeric IgA across the epithelial cell barrier of the mucosa ( 150 ). This receptor contains seven domains with five extracellular regions similar to the V regions of the immunoglobulin, a sixth transmembrane domain, and a seventh cytoplasmic domain ( 151 ). pIgR is synthesized on epithelial cells of respiratory, gastrointestinal, and genitourinary tracts, and is expressed on the basolateral aspect. Tight interactions with the J chain and the IgA Fc occur. Cys309 of IgA (hom*ologous to Cys414 in IgM) forms a disulfide bond with the receptor ( 152 ). After transcytosis, the pIgR is cleaved between its fifth and sixth domains to release dimeric IgA, J chain, and the rest of the receptor referred to as the secretory component (SC) as a complex ( 150 ). The remaining SC helps to provide protection for the secreted immunoglobulin from proteolysis on the mucosal surfaces.

AN EVOLUTIONARY PERSPECTIVE From an evolutionary perspective, antibodies are easily traceable to the beginnings of the vertebrate radiation well over 400 million years ago. While the molecular biological events that bring VDJ and VJ together, and bring V domains in the context of C domains, have varied greatly over evolutionary time, the basic structure of the Ig fold and the concept of a variable and a constant region remain intact. Indeed, with the exception of the myriad ways in which diversity is generated within the V domain (somatic hypermutation, multiple germline genes, a variety of gene segments, and the like), the most profound events are rather remarkably similar: Proteins are required to splice various sections of the molecule together, and the hinge region seems required to transmit signals from one part of the molecule to another. Indeed, the functions we attribute to the Fc region—complement binding and binding to phagocytic cells—are very old in evolutionary time. In essence, once evolution solved the problem of linking a common biologic function (recruiting proteins—like complement; and cells—like neutrophils) to an inflammatory site by a specific molecule (and Fv domain), the system seems to have been duplicated over and over, with remarkable constancy by a variety of vertebrates and perhaps some


CONCLUSION Immunoglobulins are extremely versatile molecules that carry out many biological activities simultaneously. The duality of the structure between preparation to recognize unique antigen structures a priori and maintenance of host cell receptor or complement recognition properties presents a truly unique task for the system. As has been described, many varieties of antibodies have different biological niches, but the overall design for these molecules is the same. As the science of our field progresses, attention will be given ever more closely to the engineering of antibodies for multiple tasks. Many therapeutic applications are already in various stages of development and various parts of immunoglobulins are being used for biotechnology applications. Thus, there has been a resurgence of interest in the structure–function aspects of antibodies as we approach “designer antibodies.” It is reasonable to assume that at some point in time, therapeutics will be designed with, for example, the same variable region but with different constant regions depending on the desired effector function (complement binds vs. phagocytosis). Indeed, some effector function could be engineered out of antibody molecules as the need develops. Thus, the study of the structure and function of antibodies is ever more urgent as we take fundamental principles of protein chemistry to the bedside. Color Plates

COLORPLATE 1. Ribbon diagram of a complete IgG1 crystal (1hzh in PDB from data of Harris et al. [ 119 ]). The major regions of the immunoglobulin are illustrated. The heavy-chain constant regions (green) also include the hinge (yellow) between the first two domains. Cg2 is glycosylated (also seen in yellow). The heavy- and light-chain variable regions (red and dark blue, respectively) are N terminal to the heavy- (green) and light-chain (light blue) constant regions. CDR loops in the heavy- and light-chain variable regions (yellow and white) are illustrated as well.

COLORPLATE 2. Ribbon diagrams of side and face on views of Ig domains from VH and Cl regions. Strands are labeled according to Hood nomenclature. The “pin” composed of a disulfide bond between two cysteines is illustrated (yellow) along with the conserved tryptophan residue (red).

COLORPLATE 3. Schematic representation of the complex between SpA domain D and Fab 2A2 from a human IgM. A side view shows the peptide backbone of SpA domain D (red) bound to the framework region of the Fab heavy chain (cyan). The VL domain, which is not involved in this interaction, is shown in dark blue. The binding site for SpA is remote from the CDR loops, which are highlighted in magenta. This model is based on the superposition of helix I and II of SpA domains in the Fab-domain D complex reported here and in the previously determined Fcg-domain B complex. From Graille et al. ( 43 ), with permission.

COLORPLATE 4. A comparison of an x-ray and neutron-solution-scattering theoretical model (human IgA1) and x-ray crystal (murine IgG1 and IgG2a) structures. Light chains (yellow), heavy chains (red and dark blue), and glycosylation (light blue) are illustrated. The extended length of IgA1 over that of IgG can be seen along with extensive glycosylation that characterizes this isotype. From Boehm et al. ( 66 ), with permission.

REFERENCES 1. Hill RL, Delaney R, Fellows RE, et al. The evolutionary origins of the immunoglobulins. Proc Natl Acad Sci USA 1966;56:1762–1769. 2. Porter RR. Separation and isolation of fractions of the rabbit gamma-globulin containing the antibody and antigenic combining sites. Nature 1958;182:670–671. 3. Plaut AG, Wistar R Jr, Capra JD. Differential susceptibility of human IgA immunoglobulins to streptococcal IgA protease. J Clin Invest 1974;54:1295–1300. 4. Plaut AG, Gilbert JV, Artenstein MS, et al. Neisseria gonorrhoeae and neisseria meningitidis: extracellular enzyme cleaves human immunoglobulin A. Science 1975;190:1103–1105. 5. Nisonoff A, Wissler FC, Lipman LN. Properties of the major component of a peptic digest of rabbit antibody. Science 1960;132:1770–1771. 6. Palmer JL, Nisonoff A. Dissociation of rabbit gamma-globulin into half-molecules after reduction of one labile disulfide bond. Biochemistry 1964;3:863–869. 7. Beale D, Feinstein A. Structure and function of the constant regions of immunoglobulins. Q Rev Biophys 1976;9:135–180. 8.

Kindt TJ, Capra JD. The Antibody Enigma, New York: Plenum Press, 1984.

9. Capra JD, Kehoe JM. Hypervariable regions, idiotypy, and the antibody-combining site. Adv Immunol 1975;20:1–40. 10. Davies DR, Padlan EA, Segal DM. Three-dimensional structure of immunoglobulins. Annu Rev Biochem 1975;44:639–667. 11. Edmundson AB, Wood MK, Schiffer M, et al. A crystallographic investigation of the Mcg myeloma protein. ANL Rep 1969:283–185. 12. Edmundson AB, Schiffer M, Wood MK, et al. Crystallographic studies of an IgG immunoglobulin and the Bence–Jones protein from one patient. Cold Spring Harb Symp Quant Biol 1972;36:427–432. 13.

Wurzburg BA, Garman SC, Jardetzky TS. Structure of the human IgE-Fc Cepsilon 3-Cepsilon4

reveals conformational flexibility in the antibody effector domains. Immunity 2000;13:375–385. 14. Williams AF, Barclay AN. The immunoglobulin superfamily—domains for cell surface recognition. Annu Rev Immunol 1988;6:381–405. 15. Harpaz Y, Chothia C. Many of the immunoglobulin superfamily domains in cell adhesion molecules and surface receptors belong to a new structural set which is close to that containing variable domains. J Mol Biol 1994;238:528–539. 16. Bork P, Holm L, Sander C. The immunoglobulin fold: structural classification, sequence patterns and common core. J Mol Biol 1994;242:309–320. 17. Lesk AM, Chothia C. Evolution of proteins formed by beta-sheets. II. The core of the immunoglobulin domains. J Mol Biol 1982;160:325–342. 18. Wu TT, Kabat EA. An analysis of the sequences of the variable regions of Bence Jones proteins and myeloma light chains and their implications for antibody complementarity. J Exp Med 1970;132:211–250. 19. Chothia C, Novotny J, Bruccoleri R, et al. Domain association in immunoglobulin molecules: the packing of variable domains. J Mol Biol 1985;186:651–663. 20. Colman PM. Structure of antibody–antigen complexes: implications for immune recognition. Adv Immunol 1988;43:99–132. 21. Saul FA, Amzel LM, Poljak RJ. Preliminary refinement and structural analysis of the Fab fragment from human immunoglobulin new at 20 A resolution. J Biol Chem 1978;253:585–597. 22. Edmundson AB, Ely KR, Girling RL, et al. Binding of 2,4-dinitro-phenyl compounds and other small molecules to a crystalline lambda-type Bence–Jones dimer. Biochemistry 1974;13:3816–3827. 23. Amzel LM, Poljak RJ. Three-dimensional structure of immunoglobulins. Annu Rev Biochem 1979;48:961–997. 24. Deisenhofer J. Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8-A resolution. Biochemistry 1981;20:2361–2370. 25.

Padlan EA. Anatomy of the antibody molecule. Mol Immunol 1994;31:169–217.

26. Jones PT, Dear PH, Foote J, et al. Replacing the complementarity-determining regions in a human antibody with those from a mouse. Nature 1986;321:522–525. 27. Foote J, Winter G. Antibody framework residues affecting the conformation of the hypervariable loops. J Mol Biol 1992;224:487–499. 28. Schroeder HW Jr, Hillson JL, Perlmutter RM. Structure and evolution of mammalian VH families. Int Immunol 1990;2:41–50. 29. Kirkham PM, Mortari F, Newton JA, et al. Immunoglobulin VH clan and family identity predicts variable domain structure and may influence antigen binding. Embo J 1992;11:603–609. 30. Kirkham PM, Schroeder HW Jr. Antibody structure and the evolution of immunoglobulin V gene segments. Semin Immunol 1994;6:347–360.

31. Pascual V, Capra JD. Human immunoglobulin heavy-chain variable region genes: organization, polymorphism, and expression. Adv Immunol 1991;49:1–74. 32. Chothia C, Lesk AM, Tramontano A, et al. Conformations of immunoglobulin hypervariable regions. Nature 1989;342:877–883. 33. Strong RK, Campbell R, Rose DR, et al. Three-dimensional structure of murine anti-p-azophenylarsonate Fab 36-71.1. X-ray crystallography, site-directed mutagenesis, and modeling of the complex with hapten. Biochemistry 1991;30:3739–3748. 34. Wright A, Tao MH, Kabat EA, et al. Antibody variable region glycosylation: position effects on antigen binding and carbohydrate structure. Embo J 1991;10:2717–2723. 35. Middaugh CR, Litman GW. Atypical glycosylation of an IgG monoclonal cryoimmunoglobulin. J Biol Chem 1987;262:3671–3673. 36. Nakamura Y, Gojobori T, Ikemura T. Codon usage tabulated from international DNA sequence databases: status for the year 2000. Nucleic Acids Res 2000;28:292. 37. Feeney AJ. Lack of N regions in fetal and neonatal mouse immunoglobulin V-D-J junctional sequences. J Exp Med 1990;172:1377–1390. 38. Kyte J, Doolittle RF. A simple method for displaying the hydropathic character of a protein. J Mol Biol 1982;157:105–132. 39. Eisenberg D. Three-dimensional structure of membrane and surface proteins. Annu Rev Biochem 1984;53:595–623. 40. Inanov I, et al. Constraints on the hydropathicity and sequence composition of HCDR3 are conserved evolution, In: The antibodies, New York: Taylor and Francis, 2002;43–68. 41. Sasso EH, Silverman GJ, Mannik M. Human IgM molecules that bind staphylococcal protein A contain VHIII H chains. J Immunol 1989;142:2778–2783. 42. Potter KN, Li Y, Capra JD. Staphylococcal protein A simultaneously interacts with framework region 1, complementarity-determining region 2, and framework region 3 on human VH3-encoded Igs. J Immunol 1996;157:2982–2988. 43. Graille M, Stura EA, Corper AL, et al. Crystal structure of a Staphylococcus aureus protein A domain complexed with the Fab fragment of a human IgM antibody: structural basis for recognition of B-cell receptors and superantigen activity. Proc Natl Acad Sci USA, 2000;97:5399–5404. 44. Cary SP, Lee J, Wagenknecht R, et al. Characterization of superantigen-induced clonal deletion with a novel clan III-restricted avian monoclonal antibody: exploiting evolutionary distance to create antibodies specific for a conserved VH region surface. J Immunol 2000;164:4730–4741. 45. Cary S, Krishnan M, Marion TN, et al. The murine clan V(H) III related 7183, J606 and S107 and DNA4 families commonly encode for binding to a bacterial B cell superantigen. Mol Immunol 1999;36:769–776. 46. Silverman GJ, Nayak JV, Warnatz K, et al. The dual phases of the response to neonatal exposure to a VH family–restricted staphylococcal B cell superantigen. J Immunol 1998;161:5720–5732. 47.

Silverman GJ, Goodyear CS. A model B-cell superantigen and the immunobiology of B

lymphocytes. Clin Immunol 2002;102:117–134. 48. Yin J, Mundorff EC, Yang PL, et al. A comparative analysis of the immunological evolution of antibody 28B4. Biochemistry 2001;40:10764–10773. 49. Mundorff EC, Hanson MA, Varvak A, et al. Conformational effects in biological catalysis: an antibody-catalyzed oxy-cope rearrangement. Biochemistry 2000;39:627–632. 50. Romesberg FE, Spiller B, Schultz PG, et al. Immunological origins of binding and catalysis in a Diels–Alderase antibody. Science 1998;279:1929–1933. 51. Guddat LW, Shan L, Anchin JM, et al. Local and transmitted conformational changes on complexation of an anti-sweetener Fab. J Mol Biol 1994;236:247–274. 52. Decanniere K, Transue TR, Desmyter A, et al. Degenerate interfaces in antigen–antibody complexes. J Mol Biol 2001;313:473–478. 53. Tribbick G, Edmundson AB, Mason TJ, et al. Similar binding properties of peptide ligands for a human immunoglobulin and its light chain dimer. Mol Immunol 1989;26:625–635. 54. Afonin PV, Fokin AV, Tsygannik IN, et al. Crystal structure of an anti-interleukin-2 monoclonal antibody Fab complexed with an antigenic nonapeptide. Protein Sci 2001;10:1514–1521. 55. Bhat TN, Bentley GA, Boulot G, et al. Bound water molecules and conformational stabilization help mediate an antigen–antibody association. Proc Natl Acad Sci USA 1994;91:1089–1093. 56. Li Y, Li H, Smith-Gill SJ, et al. Three-dimensional structures of the free and antigen-bound Fab from monoclonal antilysozyme antibody HyHEL-63. Biochemistry 2000;39:6296–6309. 57. Faelber K, Kirchhofer D, Presta L, et al. The 1.85 A resolution crystal structures of tissue factor in complex with humanized Fab D3h44 and of free humanized Fab D3h44: revisiting the solvation of antigen combining sites. J Mol Biol 2001;313:83–97. 58. Sigurskjold BW, Altman E, Bundle DR. Sensitive titration microcalorimetric study of the binding of Salmonella O-antigenic oligosaccharides by a monoclonal antibody. Eur J Biochem 1991;197:239–246. 59. Kelley RF, O’Connell MP, Carter P, et al. Antigen binding thermodynamics and antiproliferative effects of chimeric and humanized anti-p185HER2 antibody Fab fragments. Biochemistry 1992;31:5434–5441. 60. Roux KH, Strelets L, Brekke OH, et al. Comparisons of the ability of human IgG3 hinge mutants, IgM, IgE, and IgA2, to form small immune complexes: a role for flexibility and geometry. J Immunol 1998;161:4083–4090. 61. Schumaker VN, Phillips ML, Hanson DC. Dynamic aspects of antibody structure. Mol Immunol 1991;28:1347–1360. 62. Wade RH, Taveau JC, Lamy JN. Concerning the axial rotational flexibility of the Fab regions of immunoglobulin G. J Mol Biol 1989;206:349–356. 63. Brekke OH, Michaelsen TE, Sandlie I. The structural requirements for complement activation by IgG: does it hinge on the hinge? Immunol Today 1995;16:85–90. 64.

Feinstein A, Richardson N, Taussig MJ. Immunoglobulin flexibility in complement activation.

Immunol Today 1986;7:169–174. 65. Canfield SM, Morrison SL. The binding affinity of human IgG for its high affinity Fc receptor is determined by multiple amino acids in the CH2 domain and is modulated by the hinge region. J Exp Med 1991;173:1483–1491. 66. Boehm MK, Woof JM, Kerr MA, et al. The Fab and Fc fragments of IgA1 exhibit a different arrangement from that in IgG: a study by x-ray and neutron solution scattering and hom*ology modelling. J Mol Biol 1999;286:1421–1447. 67. Kilian M, Reinholdt J, Lomholt H, et al. Biological significance of IgA1 proteases in bacterial colonization and pathogenesis: critical evaluation of experimental evidence. Apmis 1996;104:321–338. 68. Griffiths RW, Gleich GJ. Proteolytic degradation of IgD and its relation to molecular conformation. J Biol Chem 1972;247:4543–4548. 69. Lin LC, Putnam FW. Structural studies of human IgD: isolation by a two-step purification procedure and characterization by chemical and enzymatic fragmentation. Proc Natl Acad Sci USA 1979;76:6572–6576. 70. Goyert SM, Hugli TE, Spiegelberg HL. Sites of “spontaneous” degradation of IgD. J Immunol 1977;118:2138–2144. 71. Ishioka N, Takahashi N, Putnam FW. Analysis of the mechanism, rate, and sites of proteolytic cleavage of human immunoglobulin D by high-pressure liquid chromatography. Proc Natl Acad Sci USA 1987;84:61–65. 72. Aucouturier P, Mihaesco E, Mihaesco C, et al. Characterization of jacalin, the human IgA and IgD binding lectin from jackfruit. Mol Immunol 1987;24:503–511. 73. Aucouturier P, Pineau N, Brugier JC, et al. Jacalin: a new laboratory tool in immunochemistry and cellular immunology. J Clin Lab Anal 1989;3:244–251. 74. Zehr BD, Litwin SD. Human IgD and IgA1 compete for D-galactose–related binding sites on the lectin jacalin. Scand J Immunol 1987;26:229–236. 75. Deisenhofer J, Colman PM, Epp O, et al. Crystallographic structural studies of a human Fc fragment. II. A complete model based on a Fourier map at 3.5 A resolution. Hoppe Seylers Z Physiol Chem 1976;357:1421–1434. 76. Williams GT, Venkitaraman AR, Gilmore DJ, et al. The sequence of the mu transmembrane segment determines the tissue specificity of the transport of immunoglobulin M to the cell surface. J Exp Med 1990;171:947–952. 77. Karasuyama H, Kudo A, Melchers F. The proteins encoded by the VpreB and lambda 5 pre–B cell-specific genes can associate with each other and with mu heavy chain. J Exp Med 1990;172:969–972. 78. Tsubata T, Reth M. The products of pre–B cell-specific genes (lambda 5 and VpreB) and the immunoglobulin mu chain form a complex that is transported onto the cell surface. J Exp Med 1990;172:973–976. 79. Ishihara K, Wood WJ Jr, Damore M, et al. B29 gene products complex with immunoglobulins on B lymphocytes. Proc Natl Acad Sci USA 1992;89:633–637. 80.

Arya S, Chen F, Spycher S, et al. Mapping of amino acid residues in the C mu 3 domain of mouse

IgM important in macromolecular assembly and complement-dependent cytolysis. J Immunol 1994;152:1206–1212. 81. Du Pasquier L, Wabl MR. Antibody diversity in amphibians: inheritance of isoelectric focusing antibody patterns in isogenic frogs. Eur J Immunol 1978;8:428–433. 82. Forster I, Vieira P, Rajewsky K. Flow cytometric analysis of cell proliferation dynamics in the B cell compartment of the mouse. Int Immunol 1989;1:321–331. 83. Roes J, Rajewsky K. Cell autonomous expression of IgD is not essential for the maturation of conventional B cells. Int Immunol 1991;3:1367–1371. 84. Roes J, Rajewsky K. Immunoglobulin D (IgD)-deficient mice reveal an auxiliary receptor function for IgD in antigen-mediated recruitment of B cells. J Exp Med 1993;177:45–55. 85. Nitschke L, Kosco MH, Kohler G, et al. Immunoglobulin D-deficient mice can mount normal immune responses to thymus-independent and -dependent antigens. Proc Natl Acad Sci USA 1993;90:1887–1891. 86. Brink R, Goodnow CC, Crosbie J, et al. Immunoglobulin M and D antigen receptors are both capable of mediating B lymphocyte activation, deletion, or anergy after interaction with specific antigen. J Exp Med 1992;176:991–1005. 87. Morris SC, Lees A, Finkelman FD. In vivo activation of naive T cells by antigen–presenting B cells. J Immunol 1994;152:3777–3785. 88. Brink R, Goodnow CC, Basten A. IgD expression on B cells is more efficient than IgM but both receptors are functionally equivalent in up-regulation CD80/CD86 co-stimulatory molecules. Eur J Immunol 1995;25:1980–1984. 89. Havran WL, DiGiusto DL, Cambier JC. mIgM:mIgD ratios on B cells: mean mIgD expression exceeds mIgM by 10- fold on most splenic B cells. J Immunol 1984;132:1712–1716. 90. Finkelman FD, Holmes JM, Dukhanina OI, et al. Cross-linking of membrane immunoglobulin D, in the absence of T cell help, kills mature B cells in vivo. J Exp Med 1995;181:515–525. 91. Preud’homme JL, Petit I, Barra A, et al. Structural and functional properties of membrane and secreted IgD. Mol Immunol 2000;37:871–887. 92. Black SJ, van der Loo W, Loken MR, et al. Expression of IgD by murine lymphocytes. Loss of surface IgD indicates maturation of memory B cells. J Exp Med 1978;147:984–996. 93. McHeyzer-Williams MG, Nossal GJ, Lalor PA. Molecular characterization of single memory B cells. Nature 1991;350:502–505. 94. Surjan L Jr, Brandtzaeg P, Berdal P. Immunoglobulin systems of human tonsils. II. Patients with chronic tonsillitis or tonsillar hyperplasia: quantification of Ig-producing cells, tonsillar morphometry and serum Ig concentrations. Clin Exp Immunol 1978;31:382–390. 95. Korsrud FR, Brandtzaeg P. Quantitative immunohistochemistry of immunoglobulin- and J-chain–producing cells in human parotid and submandibular salivary glands. Immunology 1980;39:129–140. 96.

Korsrud FR, Brandtzaeg P. Immunohistochemical evaluation of J-chain expression by intra- and

extra-follicular immunoglobulin-producing human tonsillar cells. Scand J Immunol 1981;13:271–280. 97. Brandtzaeg P. Overview of the mucosal immune system. Curr Top Microbiol Immunol 1989;146:13–25. 98. Liu YJ, de Bouteiller O, Arpin C, et al. Normal human IgD+IgM- germinal center B cells can express up to 80 mutations in the variable region of their IgD transcripts. Immunity 1996;4:603–613. 99. Billian G, Bella C, Mondiere P, et al. Identification of a tonsil IgD+ B cell subset with phenotypical and functional characteristics of germinal center B cells. Eur J Immunol 1996;26:1712–1719. 100. Arpin C, de Bouteiller O, Razanajaona D, et al. The normal counterpart of IgD myeloma cells in germinal center displays extensively mutated IgVH gene, Cmu-Cdelta switch, and lambda light chain expression. J Exp Med 1998;187:1169–1178. 101. Wilson PC, Wilson K, Liu YJ, et al. Receptor revision of immunoglobulin heavy chain variable region genes in normal human B lymphocytes. J Exp Med 2000;191:1881–1894. 102. Lehner T, Bergmeier LA, Tao L, et al. Targeted lymph node immunization with simian immunodeficiency virus p27 antigen to elicit genital, rectal, and urinary immune responses in nonhuman primates. J Immunol 1994;153:1858–1868. 103. Mestecky J, McGhee JR. Immunoglobulin A (IgA): molecular and cellular interactions involved in IgA biosynthesis and immune response. Adv Immunol 1987;40:153–245. 104. Qiu J, Brackee GP, Plaut AG. Analysis of the specificity of bacterial immunoglobulin A (IgA) proteases by a comparative study of ape serum IgAs as substrates. Infect Immun 1996;64:933–937. 105. Carayannopoulos L, Hexham JM, Capra JD. Localization of the binding site for the monocyte immunoglobulin (Ig) A-Fc receptor (CD89) to the domain boundary between Calpha2 and Calpha3 in human IgA1. J Exp Med 1996;183:1579–1586. 106. Pleass RJ, Dunlop JI, Anderson CM, et al. Identification of residues in the CH2/CH3 domain interface of IgA essential for interaction with the human Fc alpha receptor (FcalphaR) CD89. J Biol Chem 1999;274:23508–23514. 107. Mattu TS, Pleass RJ, Willis AC, et al. The glycosylation and structure of human serum IgA1, Fab, and Fc regions and the role of N-glycosylation on Fc alpha receptor interactions. J Biol Chem 1998;273:2260–2272. 108. Gounni AS, Lamkhioued B, Ochiai K, et al. High-affinity IgE receptor on eosinophils is involved in defence against parasites. Nature 1994;367:183–186. 109. Wang B, Rieger A, Kilgus O, et al. Epidermal Langerhans cells from normal human skin bind monomeric IgE via Fc epsilon RI. J Exp Med 1992;175:1353–1365. 110. Bieber T, de la Salle H, Wollenberg A, et al. Human epidermal Langerhans cells express the high affinity receptor for immunoglobulin E (Fc epsilon RI). J Exp Med 1992;175:1285–1290. 111. Garman SC, Wurzburg BA, Tarchevskaya SS, et al. Structure of the Fc fragment of human IgE bound to its high-affinity receptor Fc epsilonRI alpha. Nature 2000;406:259–266. 112. McDonnell JM, Calvert R, Beavil RL, et al. The structure of the IgE Cepsilon2 domain and its role in stabilizing the complex with its high-affinity receptor Fc epsilonRI alpha. Nat Struct Biol

2001;8:437–441. 113. Sayers I, Helm BA. The structural basis of human IgE-Fc receptor interactions. Clin Exp Allergy 1999;29:585–594. 114. Hashimoto S, Koh K, Tomita Y, et al. TNF-alpha regulates IL-4–induced Fc epsilon RII/CD23 gene expression and soluble Fc epsilon RII release by human monocytes. Int Immunol 1995;7:705–713. 115. Saji F, Koyama M, Matsuzaki N. Current topic: human placental Fc receptors. Placenta 1994;15:453–466. 116. Guddat LW, Herron JN, Edmundson AB. Three-dimensional structure of a human immunoglobulin with a hinge deletion. Proc Natl Acad Sci USA 1993;90:4271–4275. 117. Silverton EW, Navia MA, Davies DR. Three-dimensional structure of an intact human immunoglobulin. Proc Natl Acad Sci USA 1977;74:5140–5144. 118. Sarma VR, Davies DR, Labaw LW, et al. Crystal structure of an immunoglobulin molecule by x-ray diffraction and electron microscopy. Cold Spring Harb Symp Quant Biol 1972;36:413–419. 119. Harris LJ, Skaletsky E, McPherson A. Crystallographic structure of an intact IgG1 monoclonal antibody. J Mol Biol 1998;275:861–872. 120. Harris LJ, Larson SB, Hasel KW, et al. Refined structure of an intact IgG2a monoclonal antibody. Biochemistry 1997;36:1581–1597. 121. Tan LK, Shopes RJ, Oi VT, et al. Influence of the hinge region on complement activation, C1q binding, and segmental flexibility in chimeric human immunoglobulins. Proc Natl Acad Sci USA 1990;87:162–166. 122. Sandlie I, Michaelsen TE. Engineering monoclonal antibodies to determine the structural requirements for complement activation and complement mediated lysis. Mol Immunol 1991;28:1361–1368. 123.

Duncan AR, Winter G. The binding site for C1q on IgG. Nature 1988;332:738–740.

124. Furukawa K, Kobata A. IgG galactosylation—its biological significance and pathology. Mol Immunol 1991;28:1333–1340. 125. Tsuchiya N, Endo T, Matsuta K, et al. Effects of galactose depletion from oligosaccharide chains on immunological activities of human IgG. J Rheumatol 1989;16:285–290. 126. Shields RL, Namenuk AK, Hong K, et al. High resolution mapping of the binding site on human IgG1 for Fc gamma RI, Fc gamma RII Fc gamma RIII, and FcRn and design of IgG1 variants with improved binding to the Fc gamma R. J Biol Chem 2001;276:6591–6604. 127. Sondermann P, Huber R, Oosthuizen V, et al. The 3.2-A crystal structure of the human IgG1 Fc fragment–Fc gammaRIII complex. Nature 2000;406:267–273. 128. Radaev S, Motyka S, Fridman WH, et al. The structure of a human type III Fcgamma receptor in complex with Fc. J Biol Chem 2001;276:16469–16477. 129. Radaev S, Sun PD. Recognition of IgG by Fcgamma receptor: the role of Fc glycosylation and the binding of peptide inhibitors. J Biol Chem 2001;276:16478–16483.

130. Martin WL, West AP Jr, Gan L, et al. Crystal structure at 2.8 A of an FcRn/heterodimeric Fc complex: mechanism of pH-dependent binding. Mol Cell 2001;7:867–877. 131. Davis AC, Roux KH, Shulman MJ. On the structure of polymeric IgM. Eur J Immunol 1988;18:1001–1008. 132. Brandtzaeg P, Prydz H. Direct evidence for an integrated function of J chain and secretory component in epithelial transport of immunoglobulins. Nature 1984;311:71–73. 133. Mole JE, Bhown AS, Bennett JC. Primary structure of human J chain: alignment of peptides from chemical and enzymatic hydrolyses. Biochemistry 1977;16:3507–3513. 134. Koshland ME. The coming of age of the immunoglobulin J chain. Annu Rev Immunol 1985;3:425–453. 135. Max EE, Korsmeyer SJ. Human J chain gene. Structure and expression in B lymphoid cells. J Exp Med 1985;161:832–849. 136. Cann GM, Zaritsky A, Koshland ME. Primary structure of the immunoglobulin J chain from the mouse. Proc Natl Acad Sci USA 1982;79:6656–6660. 137. Hughes GJ, Frutiger S, Paquet N, et al. The amino acid sequence of rabbit J chain in secretory immunoglobulin A. Biochem J 1990;271:641–647. 138. Mikoryak CA, Margolies MN, Steiner LA. J chain in Rana catesbeiana high molecular weight Ig. J Immunol 1988;140:4279–4285. 139. Mestecky J, Schrohenloher RE, Kulhavy R, et al. Site of J chain attachment to human polymeric IgA. Proc Natl Acad Sci USA 1974;71:544–548. 140. Mestecky J, Schrohenloher RE. Site of attachment of J chain to human immunoglobulin M. Nature 1974;249:650–652. 141. Bastian A, Kratzin H, Fallgren-Gebauer H, et al. Intra- and inter-chain disulfide bridges of J chain in human S-IgA. Adv Exp Med Biol 1995:581–583. 142. Frutiger S, Hughes GJ, Paquet N, et al. Disulfide bond assignment in human J chain and its covalent pairing with immunoglobulin M. Biochemistry 1992;31:12643–12647. 143. Sorensen V, Sundvold V, Michaelsen TE, et al. Polymerization of IgA and IgM: roles of Cys309/Cys414 and the secretory tailpiece. J Immunol 1999;162:3448–3455. 144. Sorensen V, Rasmussen IB, Sundvold V, et al. Structural requirements for incorporation of J chain into human IgM and IgA. Int Immunol 2000;12:19–27. 145. Cals MM, Guenzi S, Carelli S, et al. IgM polymerization inhibits the Golgi-mediated processing of the mu- chain carboxy-terminal glycans. Mol Immunol 1996;33:15–24. 146. Randall TD, King LB, Corley RB. The biological effects of IgM hexamer formation. Eur J Immunol 1990;20:1971–1979. 147. Hughey CT, Brewer JW, Colosia AD, et al. Production of IgM hexamers by normal and autoimmune B cells: implications for the physiologic role of hexameric IgM. J Immunol

1998;161:4091–4097. 148.

Metzger H. Structure and function of gamma M macroglobulins. Adv Immunol 1970;12:57–116.

149. Eskeland T, Christensen TB. IgM molecules with and without J chain in serum and after purification, studied by ultracentrifugation, electrophoresis, and electron microscopy. Scand J Immunol 1975;4:217–228. 150. Mostov KE, Blobel G. A transmembrane precursor of secretory component. The receptor for transcellular transport of polymeric immunoglobulins. J Biol Chem 1982;257:11816–11821. 151. Mostov KE, Friedlander M, Blobel G. The receptor for transepithelial transport of IgA and IgM contains multiple immunoglobulin-like domains. Nature 1984;308:37–43. 152. Fallgren-Gebauer E, Gebauer W, Bastian A, et al. The covalent linkage of the secretory component to IgA. Adv Exp Med Biol 1995:625–628. 153. Carayannopoulos L, Capra JD. Immunoglobulins: structure and function. In: Fundamental Immunology. Paul WE, ed. New York: Raven Press, 1993:283–314. 154. Kabat EA, Wu TT, Perry HM, et al. Sequences of proteins of immunological interest, Washington, DC: U.S. Department of Health and Human Services, 1991. 155. Lefranc MP, Lefranc G. The Immunoglobulin Factsbook. FactsBook Series. San Diego: Academic Press, 2001. 156.

Kuby J. Immunology, 3rd ed. New York: W.H. Freeman and Company, 1997.

157. Janeway CA Jr, Travers P, Walport M, et al. Immunobiology, 4th ed. New York: Garland Publishing, 1999. 158. Putnam FW. The Plasma Proteins: Structure, Function, and Genetic Control, 2nd ed. vol. 3, New York: Academic Press, 1977.

Chapter 4 Antigen–Antibody Interactions and Monoclonal Antibodies Fundamental Immunology

Chapter 4 Jay Berzofsky

Antigen–Antibody Interactions and Monoclonal Antibodies

THERMODYNAMICS AND KINETICS The Thermodynamics of Affinity Kinetics of Antigen–Antibody Reactions AFFINITY Interaction in Solution with Monovalent Ligand Two-Phase Systems RADIOIMMUNOASSAY AND RELATED METHODS Separation of Bound and Free Antigen Optimization of Antibody and Tracer Concentrations for Sensitivity Analysis of Data: Graphic and Numerical Representation Nonequilibrium Radioimmunoassay Enzyme-Linked Immunosorbent Assay Enzyme-Linked Immunospot Assay SPECIFICITY AND CROSS-REACTIVITY Multispecificity OTHER METHODS Quantitative Precipitin Immunodiffusion and the Ouchterlony Method Immunoelectrophoresis Hemagglutination and Hemagglutination Inhibition Immunoblot (Western Blot) Surface Plasmon Resonance MONOCLONAL ANTIBODIES Derivation of Hybridomas Applications of Monoclonal Antibodies Specificity and Cross-Reactivity CONCLUSION ACKNOWLEDGMENTS REFERENCES

The basic principles of antigen–antibody interaction are those of any bimolecular reaction. Moreover, the binding of antigen by antibody can, in general, be described by the same theories and studied by the same experimental approaches as the binding of a hormone by its receptor, of a substrate by enzyme, or of oxygen by hemoglobin. There are several major differences, however, between antigen–antibody interactions and these other situations. First, unlike most enzymes and many hormone-binding

systems, antibodies do not irreversibly alter the antigen they bind. Thus, the reactions are, at least in principle, always reversible. Second, antibodies can be raised, by design of the investigator, with specificity for almost any substance known. In each case, it is possible to find antibodies with affinities as high as and specificities as great as those of enzymes for their substrates and receptors for their hormones. The interaction of antibody with antigen can thus be taken as a prototype for interactions of macromolecules with ligands in general. In addition, the same features of reversibility and availability of a wide variety of specificities have made antibodies invaluable reagents for identifying, quantitating, and even purifying a growing number of substances of biological and medical importance. One other feature of antibodies that in the past created difficulty in studying and using them—in comparison with, say, enzymes—is their enormous heterogeneity. Even “purified” antibodies from an immune antiserum, all specific for the same substance and sharing the same overall immunoglobulin structure (see Chapter 9), are a heterogeneous mixture of molecules of different subclass, different affinity, and different fine specificity and ability to discriminate among cross-reacting antigens. The advent of hybridoma monoclonal antibodies ( 1 , 2 and 3 ) has made available a source of hom*ogeneous antibodies to almost anything to which antisera can be raised. Nevertheless, heterogeneous antisera are still in widespread use and even have advantages for certain purposes, such as precipitation reactions. Therefore, it is critical to keep in mind throughout this chapter, and indeed much of this book, that the principles derived for the interaction of one antibody with one antigen must be modified and extended to cover the case of heterogeneous components in the reaction. In this chapter, we examine the theoretical principles necessary for analyzing, in a quantitative manner, the interaction of antibody with antigen and the experimental techniques that have been developed both to study these interactions and to make use of antibodies as quantitative reagents. Furthermore, we discuss the derivation, use, and properties of monoclonal antibodies.

THERMODYNAMICS AND KINETICS The Thermodynamics of Affinity The basic thermodynamic principles of antigen–antibody interactions, as indicated previously, are the same as those for any reversible bimolecular binding reaction. We review these as they apply to this particular immunological reaction. Chemical Equilibrium in Solution For this purpose, S is the antibody binding sites, L is the ligand (antigen) sites, and SL is the complex of the two. Then for the reaction

according to the mass action law,

where K A is the association constant (or affinity) and square brackets in the equation

indicate molar concentration of the reactants enclosed. The importance of this equation is that, for any given set of conditions such as temperature, pH, and salt concentration, the ratio of the concentration of the complex to the product of the concentrations of the reactants at equilibrium is always constant. Thus, changing the concentration of either the antibody or the ligand invariably causes change in the concentration of the complex, provided that neither reactant is limiting—that is, neither has already been saturated—and provided that sufficient time is allowed to reach a new state of equilibrium. Moreover, because the concentrations of antibody and ligand appear in this equation in a completely symmetrical manner, doubling either the antibody concentration or the antigen concentration results in a doubling of the concentration of the antigen–antibody complex, provided that the other reactant is in sufficient excess. This proviso, an echo of the first one mentioned previously, is inherent in the fact that [ S] and [ L] refer to the concentrations of free S and free L, respectively, in solution, not the total concentration, which would include that of the complex. Thus, if L is not in great excess, doubling [ S] results in a decrease in [ L], because some of it is consumed in the complex; therefore, the net result is less than a doubling of [ SL]. Similarly, halving the volume results in a doubling of the total concentration of both antibody and ligand. If the fraction of both reactants tied up in the complex is negligibly small (as might be the case for low-affinity binding), the concentration of the complex quadruples. However, in most practical cases, the concentration of complex is a significant fraction of the total concentration of antigen or antibody or both; therefore, the net result is an increase in the concentration of complex, but by a factor of less than 4. The other important, perhaps obvious, but often forgotten principle to be gleaned from this example is that, because it is concentration, not amount, of each reactant that enters into the mass action law ( Equation 2), putting the same amount of antigen and antibody in a smaller volume increases the amount of complex formed, and diluting them in a larger volume greatly decreases the amount of complex formed. Moreover, these changes occur approximately as the square of the volume; therefore, volumes are critical in the design of an experiment. The effect of increasing free ligand concentration [ L], at constant total antibody concentration, on the concentration of complex, [ SL], is illustrated in Fig. 1. The mass action law ( Equation 2) can be rewritten or

where [ S] t = total antibody site concentration: that is, [ S] + [ SL]. Initially, when the complex [ SL] is a negligible fraction of the total antibody [ S] t , the concentration of complex increases nearly linearly with increasing ligand. However, as a larger fraction of antibody is consumed, the slope tapers off, and the concentration of complex, [ SL], asymptotically approaches a plateau value of [ S] t as all the antibody becomes saturated. Thus, the concentration of antibody-binding sites can be determined from such a saturation binding curve ( Fig. 1), in which the concentration of (radioactively or otherwise labeled) ligand bound at saturation is a measure of the concentration of antibody sites. 1 This measurement is sometimes referred to as antigen-binding capacity.

FIG. 1. Schematic plot of bound ligand concentration as a function of free ligand concentration at a constant total concentration of antibody combining sites, [ S] t . The curve asymptotically approaches a plateau at which [bound ligand] = [ S] t . The total concentration of ligand at which the antibody begins to saturate is a function not only of the antibody concentration but also of the association constant, K A , also called the affinity. This constant has units of M -1, or L/mol, if all the concentrations in Equation 2 are molar. Thus, the product K A [ L] is unitless. The value of this product relative to 1 determines how saturated the antibody is, as can be seen from Equation 3'. For example, an antibody with an affinity of 10 7 M -1 is not saturated if the ligand concentration is 10 -8 M (product K A [ L] = 0.1) even if the total amount of ligand is in great excess over the total amount of antibody. According to Equation 3', the fraction of antibody occupied would be only 0.1/1.1, or about 9%, in this example. These aspects of affinity and the methods for measuring affinity are analyzed in greater detail in the next section. Free Energy With regard to thermodynamics, the affinity, K A , is also the central quantity, because it is directly related to the free energy, d F, of the reaction by the equations and

where R is the so-called gas constant (1.98717 cal/°K·mol), T is the absolute temperature (in degrees Kelvin), ln is the natural logarithm, and e is the base of the natural logarithms. The minus sign is introduced because of the convention that a negative change in free energy corresponds to positive binding. d F° is the standard free-energy change defined as the d F for 1 mol antigen + 1 mol antibody sites combining to form 1 mol of complex, at unit concentration. It is also instructive to note an apparent discrepancy in Equation 4 and Equation 4'. As defined in Equation 2, K A has dimensions of M -1 (i.e., L/mol), whereas in Equation 4' it is dimensionless. The reason is that for Equation 4' to hold strictly, K A must be expressed in terms of mole fractions rather than concentrations. The mole fraction of a solute is the ratio of moles of that solute to the total number of moles of all components in the solution. Because water (55 M) is by far the predominant component of most aqueous solutions, for practical

purposes, K A can be converted into a unitless ratio of mole fractions by dividing all concentrations in Equation 2 by 55 M. This transformation makes Equation 4' strictly correct, but it introduces an additional term, - RT ln 55 (corresponding to the entropy of dilution), into Equation 4. This constant term cancels out when d F values are subtracted out, but not in ratios of d F values. An important rule of thumb can be extracted from these equations. Because ln 10 = 2.303, a 10-fold increase in affinity of binding corresponds to a free-energy change d F of only 1.42 kcal/mol at 37°C (310.15°K). (The corresponding values for 25°C and 4°C are 1.36 and 1.27 kcal/mol, respectively.) This is less than one-third the energy of a single hydrogen bond (about 4.5 kcal/mol). Looked at another way, a very high affinity of 10 10M -1 corresponds to a d F of only 14.2 kcal/mol, approximately the bonding energy of three hydrogen bonds. (Of course, because hydrogen bonds with water are broken during the formation of hydrogen bonds between antigen and antibody, the net energy per hydrogen bond is closer to 1 kcal/mol.) It is apparent from this example that of the many interactions (hydrophobic and ionic as well as hydrogen bonding) that occur between the contact residues in an antibody-combining site and the contacting residues of an antigen (such as a protein), almost as many are repulsive as attractive. It is this small difference of a few kilocalories between much larger numbers corresponding to the total of attractive interactions and the total of repulsive interactions that leads to net “high-affinity” binding. If d F were any larger, binding reactions would be of such high affinity as to be essentially irreversible. Viewed in this way, it is not surprising that a small modification of the antigen can result in an enormous change in affinity. A single hydrogen bond can change the affinity many-fold, and similar arguments apply to hydrophobic interactions and other forms of bonding. This concept is important in later discussions of specificity and antigen structure. Effects of Temperature, pH, and Salt Concentration It was mentioned earlier that K A is constant for any given set of conditions such as temperature, pH, and salt concentration. However, it varies with each of these conditions. We have already seen that the conversion of free energy to affinity depends on temperature. However, the free energy itself is also a function of temperature: where d H is change in enthalpy (the heat of the reaction) 2 and d S is the entropy (the change in disorder produced by the reaction), 2 and T is the absolute temperature (in degrees Kelvin). It can be shown that the association constant K A will thus vary with temperature as follows:

or, equivalently,

The derivation of these equations is beyond the scope of this book [see Moore ( 4 )]. However, the practical implications are as follows. First, the standard enthalpy change d H° of the reaction can be determined from the slope of a plot of ln K A versus 1/ T. Second, for an interaction that is primarily exothermic (i.e., driven by a large negative d

H, such as the formation of hydrogen bonds and polar bonds), the affinity decreases with increasing temperature. Thus, many antigen–antibody interactions have a higher affinity at 4°C than at 25°C or 37°C, and so maximum binding for a given set of concentrations can be achieved in the cold. In contrast, apolar or hydrophobic interactions are driven largely by the entropy term, Td S, and d H° is near zero. In this case, there is little effect of temperature on the affinity. As for the effects of pH and salt concentration (or ionic strength) on the affinity, these vary depending on the nature of the interacting groups. Most antigen–antibody reactions are studied at a pH near neutral and at physiologic salt concentrations (0.15 M NaCl). If the interaction is dominated by ionic interactions, high salt concentration lowers the affinity. Kinetics of Antigen–Antibody Reactions A fundamental connection between the thermodynamics and kinetics of antigen–antibody binding is expressed by the relationship

where k 1 and k -1 are the rate constants for the forward (association) and backward (dissociation) reactions. The forward reaction is determined largely by diffusion rates (theoretical upper limit, 10 9 L/mol·sec) and by the probability that a collision will result in binding: that is, largely the probability that both the antigen and the antibody will be oriented in the right way to produce a good fit, as well as the activation energy for binding. The diffusive rate constant can be shown ( 5 ) to be approximated by the Smoluchowski equation:

where a is the sum of the radii in centimeters of the two reactants, D is the sum of the diffusion constants in cm 2 /sec for the individual reactants, and the constant 6 × 10 20 is necessary to convert the units to M -1·sec -1. For example, if a = 10 -6 cm and D = 10 -7 cm 2 /sec, then k dl ˜ 7.5 × 10 8 M -1·sec -1. Association rates are generally slower for large protein antigens than for small haptens. This observation may be due to the smaller value of D, to the orientational effects in the collision, and to other nondiffusional aspects of protein–protein interactions. Therefore, association rates for protein antigens are more frequently on the order of 10 5 to 10 6 M -1·sec -1 (see later discussion). However, this observation can also be partly understood from diffusion-limited rates alone. If the radii of hypothetically spherical reactants are r 1 and r 2 , then in Equation 7a, a = r 1 + r 2 , whereas D is proportional to 1/ r 1 + 1/ r 2 . The diffusive rate constant is

therefore proportional to

From this result, it can be seen that if r 1 = r 2 = r, then r cancels out and the whole term in Equation 7b is simply equal to 4. Thus, for the interaction between two molecules of equal size, the diffusive rate constant is the same regardless of whether those molecules are large or small ( 6 ). However, if one molecule is large and the other small, the rate is greater than if both molecules are large. This difference occurs because reducing the radius r 1 while keeping r 2 constant (and larger than r 1 ) has a greater effect on increasing the diffusion constant term D, proportional to 1/ r 1 + 1/ r 2 , in which the smaller radius produces the larger term, than it has on the term a, which is still dominated by the larger radius r 2 . For example, if r 2 = r, as previously, but r 1 = 0.1 r, then the numerator in Equation 7b is reduced only from 4 r 2 to 1.21 r 2 , whereas the denominator is reduced from 1 r 2 to 0.1 r 2 . Thus, the ratio is increased from 4 to 12.1. Viewed another way, the greater diffusive mobility of the small hapten outweighs its diminished target area relative to a large protein antigen, inasmuch as the larger target area of the antibody is available to both. The dissociation rate (or “off rate”) k -1 is determined by the strength of the bonds (as it affects the activation energy barriers for dissociation) and the thermal energy kT (where k is Boltzmann’s constant), which provides the energy to surmount this barrier. The activation energy for dissociation is the difference in energy between the starting state and the transition state of highest energy to which the system must be raised before dissociation can occur. As pointed out by Eisen ( 7 ), if one of a series of related antigens, of similar size and other physical properties, is compared for binding to an antibody, all the association rates are very similar. The differences in affinity largely correspond to the differences in dissociation rates. A good example is that of antibodies to the protein antigen staphylococcal nuclease ( 8 ). Antibodies to native nuclease were fractionated on affinity columns of peptide fragments to isolate a fraction specific for residues 99 through 126. The antibodies had an affinity of 8.3 × 10 8 M -1 for the native antigen and an association rate constant, k on, of 4.1 × 10 5 M -1·sec -1. This k on was several orders of magnitude lower than had been observed for small haptens ( 9 ), as discussed previously. A value of k off of 4.9 × 10 -4 sec -1 was calculated by using these results in Equation 7. This is a first-order rate constant from which it is possible to calculate a half-time for dissociation (based on t 1/2 = ln 2/ k off) of 23 minutes. These rates are probably typical for high-affinity ( K A ˜ 10 9 M -1) antibodies to small protein antigens such as nuclease (molecular weight ˜ 17,000).

The dissociation rate is important to know in designing experiments to measure binding, because if the act of measurement perturbs the equilibrium, the time for making the measurement (e.g., to separate bound and free) is determined by this half-time for dissociation. For instance, a 2-minute procedure that involves dilution of the antigen–antibody mixture can be completed before significant dissociation has occurred if the dissociation half-time is 23 min. However, if the “on” rate is the same but the affinity is 10-fold lower, still a respectable 8 × 10 7 M -1, then the complex could be 50% dissociated in the time required to complete the procedure. This caution is very relevant in the later discussion of methods of measuring binding and affinity. Because knowledge of the dissociation rate can be very important in the design of experiments, techniques to measure it should be understood. Perhaps the most widely applicable one is the use of radiolabeled antigen. After equilibrium is reached and the equilibrium concentration of bound radioactivity determined, a large excess of unlabeled antigen is added. Because any radioactive antigen molecule that dissociates is quickly replaced by an unlabeled one, the probability that a radioactive molecule will associate again is very low. Therefore, the decrease in radioactivity bound to antibody with time can be measured to determine the dissociation rate. 3

AFFINITY It is apparent from the preceding discussion that a lot of information about an antigen–antibody reaction is packed into a single value: its affinity. In this section, we examine affinity more closely, including methods for measuring affinity and the heterogeneity thereof, the effects of multivalency of antibody and of antigen, and the special effects seen when the antigen–antibody interaction occurs on a solid surface (two-phase systems). Interaction in Solution with Monovalent Ligand The simplest case is that of the interaction of antibody with monovalent ligand. This category may include both antihapten antibodies reacting with truly monovalent haptens and antimacromolecule antibodies, which have been fractionated to obtain a population that reacts only with a single, nonrepeating site on the antigen. 4 In the latter case, the antigen behaves as if monovalent in its interaction with the particular antibody population under study. The proviso that the site recognized (antigenic determinant) be nonrepeating—that is, it occurs only once per antigen molecule—is, of course, critical. If the combining sites on the antibody are independent (i.e., display no positive or negative cooperativity for antigen binding), then for many purposes these combining sites, reacting with monovalent ligands, can be treated as if they were separate molecules. Thus, many, but not all, of the properties we discuss can be analyzed in terms of the concentration of antibody-combining sites, independent of the number of such sites per antibody molecule [two for immunoglobulins G and A (IgG and IgA), 10 for immunoglobulin M (IgM)]. In general, to determine the affinity of an antibody, the equilibrium concentrations of

bound and free ligand are determined, at increasing total ligand concentrations but at constant antibody concentration. Alternatively, the antibody concentration can be varied, but then the analysis is slightly more complicated. Perhaps the theoretically most elegant experimental method to determine these quantities is equilibrium dialysis ( 10 , 11 ), depicted and explained in Fig. 2, in which ligand (antigen) is allowed to equilibrate between two chambers, only one of which contains antibody, separated by a semipermeable membrane impermeable to antibody. The important feature of this method, as opposed to most others, is that the concentrations of ligand in each chamber can be determined without perturbing the equilibrium. The disadvantage of this method is that it is applicable only to antigens small enough to freely permeate a membrane that will exclude antibody. Another technical disadvantage is that bound antigen, determined as the difference between bound plus free antigen in one chamber and free antigen in the other, is not measured independently of free antigen.

FIG. 2. Equilibrium dialysis. Two chambers are separated by a semipermeable membrane that is freely permeable to ligand but not at all to antibody. Antibody is placed in one chamber (chamber B), and ligand in one or both chambers. Regardless of how the ligand is distributed initially, after sufficient time to reach equilibrium, it is distributed as follows. The concentration of free ligand is identical in both chambers, but chamber B has additional ligand bound to antibody. The concentration of bound ligand is thus the difference between the ligand concentrations in the two chambers, whereas the free concentration is the concentration in chamber A. Because these concentrations must obey the mass action law, Equation 2, they can be used to determine the affinity K A , from Equation 3 or Equation 3', by any of several graphical procedures, such as Scatchard analysis (described in the text).

Another category of method involves using radiolabeled ligand in equilibrium with antibody and then physically separating free antigen bound to antibody and quantitating each separately. The methods used to separate bound and free antigen are discussed later in the section on radioimmunoassay. These methods generally allow independent measurement of bound and free antigen but may perturb the equilibrium. Scatchard Analysis Once data are obtained, there are a number of methods of computing the affinity, of which we shall discuss two. Perhaps the most widely used is

that described by Scatchard ( 12 ) [ Fig. 3; see Berzofsky et al. ( 13 )]. The mass action equilibrium law is plotted in the form of Equation 3, and B is substituted for [ SL] and F for [ L], referring to bound and free ligand, respectively. Then the Scatchard equation is

FIG. 3. Scatchard analysis of the binding of [ 3 H]–sperm whale myoglobin by a monoclonal antibody to myoglobin (A) and by the serum antibodies from the same mouse whose spleen cells were fused to prepare the hybridoma (B). The monoclonal antibody (clone HAL 43-201E11, clone 5) produces a linear Scatchard plot, whose slope, -1.6 × 10 9 M -1, equals - K A and whose intercept on the abscissa indicates the concentration of antibody-binding sites. In contrast, the serum antibodies produce a curved (concave-up) Scatchard plot, indicative of heterogeneity of affinity. From ( 13 ), with permission. Note that a critical implicit assumption was made in this seemingly very simple conversion. The [ SL] within the parentheses in Equation 3 was intended to be the concentration of bound antibody sites, so that ([ S] t - [ SL]) = free [ S]. However, in Equation 8, we have substituted B, the concentration of bound ligand. If the ligand behaves as if it were monovalent, then this substitution is legitimate, because every bound ligand molecule corresponds to an occupied antibody site. However, if the ligand is multivalent and can bind more than one antibody site, then Equation 8 is valid only in ligand excess, in which the frequency of ligands with more than one antibody bound is very low. In this section, we are discussing only monovalent ligands, but this proviso must be kept in mind when the Scatchard analysis is applied in other circ*mstances. From Equation 8, we see that a plot of B/ F versus B should yield a straight line (for a single affinity), with a slope of - K A and an intercept on the abscissa corresponding to antibody-binding site concentration ( Fig. 3). This is the so-called Scatchard plot. An alternative version that is normalized for antibody concentration is especially useful if the data were obtained at different values of total antibody concentration, [ A] t, instead

of constant [ A] t. However, for this version, an independent measure of total antibody concentration, other than the intercept of the plot, is required. Then Equation 8 is divided by the total concentration of antibody molecules (with no assumptions about the number of sites per molecule) to obtain

where r is defined as the number of occupied sites per antibody molecule, n is defined as the total number of sites per antibody molecule, and c is free ligand concentration; that is, c = F. Thus,


where [ A] t = total molar antibody concentration. In this form of the Scatchard plot, r/ c versus r, the slope is still - K A and the intercept on the r axis is n. Thus, the number of sites per molecule can be determined. Of course, if [ S] t is determined from the intercept of Equation 8, the number of sites per molecule by dividing [ S] t can also be calculated by any independent measure of antibody concentration. Thus, the only advantage of normalizing all the data points first to plot the r/ c form arises when the data were obtained at varying antibody concentrations. If the antibody concentration is unknown but held constant, then the B/ F form is more convenient and actually provides one measure of antibody (site) concentration. Because today the value of n for each class of antibody is known (two for IgG and serum IgA, 10 for IgM), the concentration of sites and that of antibody are easily converted in many cases. Heterogeneity of Affinity The next level of complexity involves a mixture of antibodies of varying affinity for the ligand. This is the rule, rather than the exception, with antibodies from immune serum, even if they are fractionated to be monospecific: that is, all specific for the same site on the antigen. Contrast, for example, the linear Scatchard plot for a hom*ogeneous monoclonal antibody to myoglobin ( Fig. 3A) with the curved Scatchard plot for the serum antibodies from the same mouse used to prepare the hybridoma monoclonal antibody ( Fig. 3B). This concave-up Scatchard plot is typical for heterogeneous antibodies. In a system such as hormone receptor–hormone interaction, in which negative cooperativity can occur between receptor sites (i.e., occupation of one site lowers the affinity of its neighbor), a concave-up Scatchard plot can be produced by negative cooperativity in the absence of any intrinsic heterogeneity in affinity. However, in the case of antibodies, for which no such allosteric effect has been demonstrated, a concave-up Scatchard plot indicates heterogeneity of affinity. Ideally, the tangents all along the curve correspond (in slope) the affinities of the many subpopulations of antibodies. Mathematically, this is not strictly correct, but it is true that the steeper part of the curve corresponds to the higher affinity antibodies and the shallower part of the

curve to the lower affinity antibodies. Graphical methods have been developed to analyze more quantitatively the components of such curves ( 14 , 15 ), and a very general and versatile computer program (LIGAND), developed by Munson and Rodbard ( 16 ), can fit such curves when any number of subpopulations of different affinity is used. For purposes of this chapter, we discuss only the case of two affinities and then examine the types of average affinities that have been proposed for much greater heterogeneity. We also examine mathematical estimates of the degree of heterogeneity (analogous to a variance). When an antibody population consists of only two subpopulations of different affinities, K 1 and K 2 , the component Equation 3' can be added to obtain

so that

where the subscripts correspond to the two populations. Then the graph of r/ c versus r can be shown to be a hyperbola whose asymptotes are, in fact, the linear Scatchard plots of the two components ( Fig. 4). This situation was analyzed graphically by Bright ( 17 ). If the limits are c ? 0 and c ? 8, it can easily be shown that the intercept on the abscissa is just n 1 + n 2 (or, in the form B/ F vs. B, the intercept is the total concentration of binding sites [ S] t ), and the intercept on the ordinate is n 1 K 1 + n 2 K 2 . Thus, it is still possible to obtain the total value of n or [ S] t from the intercept on the abscissa. The problem is in obtaining the two affinities, K 1 and K 2 , and the concentrations of the individual antibody subpopulations (corresponding to n 1 and n 2 ). If K 1 is greater than K 2 , the affinities can be approximated from the slopes of the tangents at the two intercepts ( Fig. 4); however, these are not, in general, exactly parallel to the two asymptotes, which give the true affinities, and so some error is always introduced, depending on the relative values of n 1 , n 2 , K 1 , and K 2 . A graphical method for solving for these exactly was worked out by Bright ( 17 ), and computer methods were worked out by Munson and Rodbard ( 16 ).

FIG. 4. Analysis of a curved Scatchard plot produced by a mixture of two antibodies with different affinities. The antibodies have affinities K 1 and K 2 and have n 1 and n 2 binding sites per molecule, respectively. r is the concentration of bound antigen divided by the total antibody concentration (i.e., bound sites per molecule), and c is the free antigen concentration. The curve is a hyperbola that can be decomposed into its two asymptotes, which correspond to the linear Scatchard plots of the two components in the antibody mixture. The tangents to the curve at its intercepts only approximate these asymptotes, so that the slopes of the tangents provide an estimate of but do not accurately correspond to the affinities of the two antibodies. However, the intercept on the r axis corresponds to n 1 + n 2 . Note that in this case n 1 and n 2 must be defined in terms of the total antibody concentration, not that of each component. Average Affinities In practice, of course, it is rarely known that exactly two subpopulations are involved, and most antisera are significantly more heterogeneous than that. Therefore, the case just discussed is more illustrative of principles than of practical value. When faced with a curved Scatchard plot, the investigator usually asks what the average affinity is, and perhaps some measure of the variance of the affinities, without being able to define exactly how many different affinity populations exist. Suppose there are m populations each with site concentration [ S i ] and affinity S i, so that at free ligand concentration [ L], the fraction of each antibody that has ligand bound is given by an equation of the form of Equation 3':

Then the bound concentrations sum as follows:

Substituting F for [ L] and dividing through by this quantity yields

or, equivalently,

These can be seen to be generalizations of Equations 10 and 10'. If the limits are F ? 0 and F ? 8,


Therefore, it is still possible to obtain the total antibody site concentration from the intercept on the abscissa ( Fig. 5) ( 18 ).

FIG. 5. Types of average affinities for a heterogeneous population of antibodies, as defined on a Scatchard plot. K 0 is the slope of the tangent to the curve at a point where B = [ S] t /2: that is, where half the antibody sites are bound. Thus, K 0 corresponds to a median affinity. K av is the slope of the chord between the intercepts and corresponds to a weighted average of the affinities, weighted by the concentrations of the antibodies with each affinity. Adapted from ( 18 ), with permission. Two types of average affinity can be obtained graphically from the Scatchard plot ( 18 ). A term perhaps the more widely used, K 0 , is actually more accurately a median affinity rather than a mean affinity. It is defined as the slope of the tangent at the point on the curve where half the sites are bound: that is, where B = [ S] t /2 ( Fig. 5). A second type

of average affinity, which we call K av , is a weighted mean of the affinities, each affinity weighted by its proportional representation in the antibody population. Thus, the ratio is

From Equations 13 and 14, it is apparent that K av is simply the ratio of the two intercepts on the B/ F and B axes: that is, the slope of the chord ( Fig. 5). This type of weighted mean affinity, K av , is therefore actually easier to obtain graphically in some cases than is K 0 , and it is useful in other types of plots as well. Indices of Heterogeneity: the Sips Plot For a heterogeneous antiserum, it is desirable to have some idea of the extent of heterogeneity of affinity. For instance, if the affinities are distributed according to a normal (gaussian) distribution, it is helpful to know the variance ( 19 , 20 ). More complex analyses have been developed that do not require as many assumptions about the shape of the distribution ( 21 , 22 and 23 ), but the first and most widely used index of heterogeneity arbitrarily assumes that the affinities fit a distribution, first described by Sips ( 24 ), which is similar in shape to a normal distribution. This was applied to the case of antibody heterogeneity by Nisonoff and Pressman ( 25 ) and was summarized by Karush and Karush ( 26 ). The data are fit to the assumed binding function

which is analogous to Equations 3' and 11 (the Langmuir adsorption isotherm) except for the exponent a, which is the index of heterogeneity. This index, a, is allowed to range from 0 to 1. For a = 1, Equation 16 is equivalent to Equation 3 and there is no heterogeneity. As a decreases toward 0, the heterogeneity increases. To obtain a value for a graphically, the algebraic rearrangement of Equation 16 is plotted as follows:

so that the slope of log [ r/( n - r)] versus log c is the heterogeneity index a. C. DeLisi (personal communication) derived the variance (second moment) of the Sips distribution in terms of the free energy RT ln K 0 , about the mean of free energy. The result (normalized to RT) gives the dispersion or width of the distribution as a function of a:

This is useful for determining a quantity, s Sips, which can be thought of as analogous to a standard deviation, if one keeps in mind that this is not a true gaussian distribution. In addition, as noted previously, the use of the Sips distribution requires the assumption that the affinities (really the free energies) are continuously distributed symmetrically about a mean, approximating a gaussian distribution. This assumption frequently is not valid.

The Plot of B/F Versus F or T Another graphical method that is useful for estimating affinities is the plot of bound/free versus free or total ligand concentration, denoted F and T, respectively ( 18 ) ( Fig. 6). To simplify the discussion, we define the bound/free ratio, B/ F, as R and define R 0 as the intercept, or limit, as free ligand F ? 0. First, for the case of a hom*ogeneous antibody, from Equation 3', and

FIG. 6. Schematic plot of R, the bound/free ratio, as a function of free ( F) or total ( T) antigen concentration. The curves have a similar sigmoidal shape, but the midpoint (where R = R 0 /2) of the plot of R versus T has a term dependent on antibody site concentration ([ S] t ), whereas the midpoint of the plot of R versus F is exactly 1/ K, independent of antibody concentration. Adapted from ( 18 ), with permission. We define the midpoint of the plot ( Fig. 6) as the point at which R decreases to half its initial value, R 0 : that is, at which R = K[ S] t /2. For the case of hom*ogeneous antibody (i.e., a single affinity), simple algebraic manipulation ( 18 ), substituting K[ S] t /2 (i.e., R 5 0 /2) for B/ F in Equation 8, will show that at this midpoint

and so that the total concentration, T, is

Thus, if B/ F versus F is plotted, the midpoint directly yields 1/ K. However, it is frequently more convenient experimentally to plot B/ F versus T. In this case, the midpoint is no longer simply the reciprocal of the affinity. As seen from Equation 23, the

assumption that the midpoint is 1/ K will result in an error equal to half the antibody-binding site concentration. Thus, in plots of B/ F versus T, the midpoint is a good estimate of the affinity only if [ S] t /2 2h was required for the progression of antigen-specific Th-cell differentiation. Immune synapse formation can be considered one of the central checkpoints underlying the cognate regulation of developing immune responses in vivo. When activated DC first contact naïve antigen-specific Th cells in the T zones, the resultant immune synapse (synapse I) ( Fig. 5) communicates the nature of the original antigenic insult to the adaptive immune system. Initial antigen-specific Th-cell development is heavily influenced by the DC expression pattern of cytokines (such as IL-12 and IL-6) and co-stimulatory molecules (such as CD80 and CD86) that reflect the initial inflammatory context of its activation. The strength and kinetics of the TCR–pMHC interaction can significantly impact Th-cell fate. Therefore, the available pre–immune TCR repertoire influences the outcome of this developing immune response. The antigen-primed Th cells can also deliver signals to the DC by way of cell contact (such as CD40L) and immediate early cytokine production such as TNF-a before dissociating the initial contact. Thus, immune synapsis encourages efficient local exchange of complex molecular information in an antigen-specific manner. Clonal Selection, Expansion, and Effector Th-Cell Differentiation Synapse I interactions result in extensive clonal expansion of the antigen-specific Th-cell compartment ( Fig. 5). There are 100- to 500-fold increases in the numbers of antigen-responsive Th cells during the first week after initial antigen encounter ( 148 ). Selection for Th-cells with preferred TCR features can occur very rapidly between days

3 to 5 after priming. Early selection events are consolidated through preferential clonal expansion and appear to be driven by the kinetics of TCR–pMHC binding ( 149 ). Effector cell differentiation accompanies clonal expansion in the T zones ( Fig. 5). Developmental programs initiated at synapse I are consolidated over this period of Th-cell expansion and differentiation. Autocrine and paracrine influences of cytokines may also play a major role in shaping the mix of Th-cell effector functions within the responsive population. Studies on cytokine production in vivo demonstrate a wide spectrum of effector Th-cell functions associated with the regulation of antigen-specific B-cell responses ( 150 ). Differential changes in cell-surface phenotype and alterations in T-cell physiology during this stage of development will impact the quality of T-cell help delivered to B cells. Over this first week after priming, clonally expanded effector Th cells migrate towards the T/B borders of the secondary lymphoid tissues. Ansel et al. ( 151 ) have demonstrated this to be due to the up-regulation of CXCR5 and response to CCL13/BLC in the B zones and a concomitant decrease in the response to the T-cell zone chemokines CCL19/ELC and CCL21/SLC. This combined pull-and-release mechanism allows the activated Th cells to enter the microenvironments in which they are most likely to encounter antigen-primed B cells ( Fig. 5). Recruitment and Activation of Naïve Antigen-Specific B Cells To receive cognate T-cell help, antigen-specific B cells must have contacted their specific antigen, internalized, processed, and presented antigenic peptides in the context of MHC class II. The efficiency of antigen capture is a major determinant in the density of pMHC molecules presented on the cell surface. Early studies established that BCR-mediated processing is ˜10 4 -fold more sensitive that nonreceptor-mediated uptake through fluid pinocytosis. This is likely due to enhanced endocytosis as well as more efficient targeting to the intracellular class II peptide-loading compartment ( 152 ). Both antigen recognition and BCR signaling events appear to be involved as inhibitors of kinase activity block antigen processing. B cells do not efficiently present antigen that is in the form of immune complexes. Most likely, this leads to an aborted BCR signal in naïve B cells due to co-clustering of Fc?RIIb and recruitment of SHP-1. Mutant sIgM that do not associate with the Iga/Igß also lose much of their endocytic capacity and process antigen inefficiently. Addition of the cytoplasmic tail of Igß is sufficient to restore normal processing in these cells. Further, BCR mutants containing a deletion in their cytoplasmic tail can consititutively associate with the lipid rafts but do not internalize antigen upon cross-linking ( 153 ). Hence, antigen recognition, BCR signaling, and endocytosis are required to accelerate appropriate antigen targeting, processing, and presentation of pMHC II ( Fig. 5). Antigen-activated B cells rapidly relocate to the T/B-cell interface of secondary lymphoid organs ( Fig. 5). These earliest events in B-cell activation are difficult to access experimentally in normal nontransgenic animals due to the extremely low pre-immune precursor frequency for any known antigens. Adoptive transfer of BCR-transgenic B cells and TCR-transgenic Th cells has helped to overcome this limitation. Garside et al. ( 154 ) directly demonstrated these early B and Th-cell migration patterns and were able to visualize Th/B-cell conjugate formation that was antigen-specific. More recently, Reif et al. ( 155 ) demonstrated a chemokine-driven basis for the B-cell migration to these areas. B cells were shown to up-regulate the chemokine receptor CCR7 (specific for T-zone chemokines CCL19/ELC and CCL21/SLC) and this was sufficient to relocate activated B cells to the T/B borders. Curiously, the continued expression of CXCR5 (specific for B-zone chemokine CCL13/BLC) on the B cells seemed to create the right counterbalance for this migration event. Thus, at around

days 5 to 7 after initial antigen priming, both the antigen-activated–effector Th cells and antigen-primed B cells are translocated to the same microenvironment to continue the regulation and development of the humoral immune response. Delivery of T-Cell Help to Antigen-Primed B Cells—Phase II The delivery of cognate T-cell help to the antigen-primed B cells requires the formation of Immune synapse II ( Fig. 6). These interactions involve receptor-counter-receptor pairs of the TNFR and CD28/B7 families of molecules as well as the focal secretion of soluble factors that impact subsequent B-cell and Th-cell development. The cellular outcome for the B-cell bifurcates at this point with a GC pathway that proceeds in the B zone (discussed below) and a T-zone pathway that involves isotype switch recombination and the development of short-lived plasma cells.

FIG. 6. Delivery of T-cell help to antigen-specific B cells. Immune synapse II forms between clonally expanded and differentiated antigen-specific Th cells and antigen-experienced B cells at the T-B cell borders of secondary lymphoid organs. The sets of co-receptor and co-stimulatory molecules potentially involved in this complex interaction are covered in detail in this chapter. The immediate impact of synapse II is rapid and substantial B-cell clonal expansion either within the B-cell zones or T-cell zones. The B-zone pathway leads to the formation of secondary follicles, the precursors of the germinal center reaction while the T-zone pathway involves the development of short-lived plasma cells and isotype switch recombination in the absence of somatic hypermutation. Some antigen-specific Th cells also remain in the T zones while the majority migrate to the secondary follicles to participate in the germinal center reaction. Immune Synapse II Immune synapse II forms between antigen-experienced and clonally expanded effector Th cells and antigen-primed B cells ( Fig. 6). This intercellular contact is qualitatively and quantitatively distinct from the interaction between activated DC and naïve Th cells discussed in Phase I. While cytokines are sufficient to promote B-cell differentiation in bulk cell culture, it soon became apparent that inter-cellular contact is also a crucial factor that regulates B-cell fate in vivo. A signal through the pMHC expressed on these activated B cells could promote early biochemical changes, as well as cell cycle entry and plasma cell differentiation. However, the tumor necrosis family receptor (TNFR) family member, CD40, was one of the earliest co-stimulatory molecules identified as indispensable to effective Th-cell–regulated B-cell response ( 156 ). In the absence of this molecule, T-dependent immune responses generated IgM

plasma cell formation without isotype switch recombination, germinal center formation, or affinity maturation ( 157 ). CD40 is constitutively expressed on B cells and its ligand, CD154 (CD40L), is expressed during the synapse I interactions of DC-induced Th-cell activation. At synapse II, there is an exchange of information through CD40–CD40L interactions with impact on both activated Th cells and antigen-primed B cells. In CD40-deficient animals, delivering a signal to the CD40L on the activated Th cells overcomes the defect in B-cell responsiveness ( 158 , 159 ). Hence, these specific molecular interactions significantly impact continued progression of lymphocyte development ( Fig. 6). Other TNF/TNFR family members play significant roles in shaping the fate of the B-cell response ( 160 ). OX40 (CD134) is expressed on activated Th cells, and its counter-receptor OX40-L (CD134L) is expressed on activated B cells. Mice deficient in CD134L have substantially reduced isotype switch recombination ( 161 ). Unlike the CD40-deficient animals, the CD134L-deficient mice can promote GC formation. Signals through CD134L on activated B cells can enhance the rate of IgG production in vitro by anti-CD40, IL-4, and IL-10–stimulated B cells suggesting a role in the regulation of antibody production distinct from CD40. Another TNFR family member, CD27, has been a useful marker of memory B cells in humans ( 162 , 163 ). The CD27 counter-receptor, CD70—a TNF family member expressed on T cells relatively late in activation—appears important in the regulation of plasma cell differentiation. CD27 is also expressed by many T cells and may serve to regulate the effect of CD70 on B cells as a decoy function. Two other sets of receptors in this family act as negative regulators of B-cell immunity. Signals through CD30 (TNFR family) or its ligand CD153 (TNF family) on activated B cells are reported to inhibit isotype switch and limit the extent of the B-cell response in vivo ( 164 ). However, the response to vesicular stomatitis virus in CD30-deficient animals appears to be normal and may suggest that there are redundant controlling mechanisms for the action of these molecules in vivo ( 165 ). Finally, CD95 (Fas) (TNF family) and its ligand CD95L(TNFR family) have well-characterized effects on the B-cell response. Deficiencies in either member of this pair lead to marked lymphoproliferative defects together with autoimmune susceptibility. CD40 signals induce this molecule on activated B cells, increasing their susceptibility to apoptosis through CD95L expression on activated T cells ( 166 ). Thus, all these sets of molecules clearly have an impact on the emerging B-cell response with substantial ability to regulate the quality of its outcome in vivo. However, the precise temporally and spatially defined events and how their molecular activities are distributed across antigen-specific–effector Th-cell and B-cell subsets remains to be thoroughly analyzed. CD28 is another well-characterized T-cell co-stimulatory molecule primarily involved in sustained T-cell activation and appears critical for the initial DC-naïve Th interaction through the counter-receptors B7-1 (CD80) and B7-2 (CD86) ( 167 ). These counter-receptors are also up-regulated on antigen-primed B cells offering a means to increase stability of synapse II interactions or more directly enhance the TCR-pMHC interactions at this later stage of development in vivo. CTLA4 is another well-characterized member of this CD28/B7 family of molecules that has a dramatic impact on the decline of immune responses ( 168 ). In the absence of CTLA4, the animal develops a fatal lymphoproliferative disorder that is characterized by the presence of a huge number of infiltrating activated CD4 Th cells due to uncontrolled B7-1/B7-2 stimulation ( 169 ). Hence, a negative signal through CTLA4 appears important to reestablish homeostasis following antigen-driven clonal expansion. Another more recently identified member of this CD28/B7 family is the inducible T-cell co-stimulator

(ICOS) ( 167 , 170 ). ICOS is hom*ologous to CD28 with a distinct ligand-binding motif and cytoplasmic tail and no detectable B7-1 or B7-2 binding. ICOS is not expressed constitutively on Th cells but is rapidly up-regulated on TCR engagement. ICOS-L is expressed at low levels on resting B cells and is not strongly up-regulated upon activation with BCR or anti-CD40. Unlike CD28, interfering with ICOS/ICOS-L interactions with soluble ICOS-Ig has modest effects at the initiation of Th-cell responses in vivo, but appear to be more important in promoting sustained T-cell expansion. ICOS-deficient mice have clear defects in class switch recombination and cannot form GC to T-dependent antigen ( 171 , 172 and 173 ). CD40 stimulation can overcome these defects and suggest that ICOS interactions are upstream from CD40–CD40L interactions in vivo. These newly described sets of interactions clearly contribute to the ongoing development of effective B-cell immunity and display some level of temporal organization in a cascade of cellular activities and outcomes. Isotype Switch Recombination Antigen-specific B-cell development divides spatially and functionally at the end of immune synapse II ( Fig. 6). One group of antigen-primed B cells clonally expands in the T zones and differentiates into short-lived plasma cells, while the second group returns to the B zones to initiate the GC reaction. Both pathways involve immunoglobulin class-switch recombination (CSR), while only the GC pathway undergoes somatic hypermutation and affinity-based maturation. The switch of Ig isotype from IgM to IgG, IgE, or IgA is accompanied by CSR. CSR is regulated by Th-cell signals and is critical for the generation of functional diversity in the humoral immune response. CD40–CD154 (CD40L) are required for CSR, as the absence of these signals leads to elevated serum-IgM levels (hyperIgM syndrome) in the absence of IgG, IgE, and IgA ( 174 ). Soluble T-cell–derived factors are also implicated in this differentiation event. IL-4 drives the high-efficiency switch to IgG1 and IgE in vitro ( 175 ). However, IL-4–deficient animals display residual IgG1 production but absent IgE responses in vivo ( 176 ). TGF-ß has been implicated in the regulation of IgA, while IFN-? is thought to induce IgG2a switch and to counter-regulate the influence of IL-4. CSR is an intrachromosonal deletional process between the switch (S) regions that reside 5' of each constant-region gene in B cells (except Cd) ( 177 ). Signaling through CD40 and cytokine receptors induces germline transcription through the targeted S regions. The Sµ region and the targeted S region is then cleaved by a putative DNA-cleaving enzyme. The activation-induced deaminase (AID), a putative RNA-editing cytidine deaminase, is required and sufficient for the initiation of the CSR reaction in the activated locus ( 178 ); however, it is not yet clear if it does the cleaving or regulates the cleaving activity of a separate complex. AID-deficient animals ( 179 ) and humans ( 180 ) display no CSR or somatic hypermutation of the Ig genes. Repair and ligation through nonhom*ologous end joining completes the process and results in the looping-out and replacement of the Cµ heavy-chain constant-region gene (C H) with other downstream C H genes. While isotype switch proceeds without somatic hypermutation in the T-zone pathway, the process within the GC reaction is thought to use the same molecular machinery. Development of Short-lived Plasma Cells The T-zone pathway to plasma cell differentiation induces the rapid production of germline-encoded antigen-specific antibody ( Fig. 6). Within the first 3 to 5 days of a T-dependent response, small foci of B-cell blasts can be seen in the T zones ( 181 ). They expand and differentiate into plasma cells of multiple Ig isotypes that migrate via the bridging channel in the splenic

marginal zones and lodge in the red pulp, and are found in the lymphatic sinus of the medullary cords in LN responses ( 182 ). In contrast to their GC counterparts, these T-zone B cells do not diversify their Ig receptors, and once differentiated into plasma cells, have short half-lives of 3 to 5 days. Plasma cells are terminally differentiated, post-mitotic, antibody-producing factories. They display a marked increase in IgH and IgL mRNA and prominent amounts of rough ER to accommodate translation and secretion of abundant Ig. They have reduced or lost numerous cell-surface molecules, including MHC II, B220, CD19, CD21, and CD22, with an increase in the proteoglycan syndecan-1 (CD138), often used as a distinguishing marker for plasma cells. Plasma cells decrease the expression of CXCR5 and CCR7, and up-regulate responsiveness to CXCL12/SDF-1, the CXCR4 ligand that is localized more to the red pulp and medullary cords ( 183 ). Several transcription factors are also decreased or absent in plasma cells including B-cell lineage-specific activator (BSAP), the Pax-5 gene product and the class II transactivator (CIITA). Early B-cell factor (EBF), A-Myb, and BCL-6 that are also associated with B-cell development are down-regulated in plasma cells. In contrast, some transcription factors increase upon terminal differentiation. B-lymphocyte–induced maturation protein 1 (BLIMP-1) is induced upon cytokine-induced, plasma-cell differentiation of a murine B-cell lymphoma BCL-1 in vitro ( 184 ). This zinc-finger–containing transcriptional repressor has three identified targets. The repression of c-Myc may provide a mechanism for the cessation of cell cycle. The repression of CIITA may explain the decrease of MHC II expression and inability to receive further cognate T-cell signals, and Pax-5 repression may be required to release its control of XBP-1 transcription ( 185 ). While XBP-1–deficient mice are embryonic lethal, using the RAG complementation system, XBP-1–deficient B cells could develop normally in vivo, but were severely blocked in their ability to differentiate into plasma cells ( 186 ). XBP-1 is induced upon activation in splenic B cells and remains at high concentrations in plasma cells. IRF4 is another transcription factor up-regulated in plasma cells and IRF4-deficient animals cannot mount antibody responses ( 187 ). Calame ( 185 ) proposes a complex regulatory cascade in antigen-activated B cells that integrates the action of XBP-1, BLIMP-1, and IRF4 to induce terminal differentiation and plasma cell commitment. The Germinal Center Reaction—Phase III The second broad cellular outcome of synapse II interactions is the GC reaction ( Fig. 7). Antigen-primed B cells migrate to the B-zone follicular area after the delivery of cognate T-cell help and rapidly expand as sIgD - B220 + B cells. This massive and rapid clonal expansion displaces mature resting B cells, creating regions within the B-cell follicular area now referred to as secondary follicles. At some point, the secondary follicle polarizes into the “dark zone” region of rapidly proliferating sIgM/D - cells proximal to the T-cell areas, and a “light zone” region of nondividing cells that express downstream sIg at the opposite pole ( Fig. 7). The polarization of the secondary follicle signifies the beginning of the germinal center reaction. The dividing cells are referred to as centroblasts and the nondividing cells are centrocytes. Multiple specialized cell types participate in the GC reaction, giving rise to a cycle of activity that is focused on the development of high-affinity B-cell memory. Recruitment into the GC cycle involves massive clonal expansion and the random somatic diversification of the BCR. GC B

cells expressing high-affinity variants are then selected for either re-entry into the GC cycle or export into the long-lived memory B-cell compartment.

FIG. 7. Phase III: the germinal center reaction. Cells involved in the GC reaction and their location in the germinal center (left), and cycle of activity that accompanies these cellular events in vivo (right). The GC reaction officially begins when the secondary follicle polarizes into the dark zone of proliferating centroblasts and the light zone of centrocytes that have dropped out of the cell cycle and express variant Ig receptors. Antigen-specific B cells are initially recruitment into the GC pathway followed by massive clonal expansion and BCR diversification via somatic hypermutation. Cells expressing variant BCR then exit the cell cycle and migrate to light zones that are filled with follicular dendritic cells (FDC) displaying copious immune-complexed antigen and occasional antigen-specific Th cells. The majority of centrocytes die locally by apoptosis and are cleared rapidly by tingible body macrophage (Mf). Some centrocytes with high-affinity variant BCR are selected to either re-enter the GC cycle of expansion, diversification, and selection or exit the GC reaction to enter the long-lived memory B-cell pool. This cell fate decision appears dependent on the quality of the BCR–antigen interaction and the nature of T-cell help delivered in a cognate manner by the resident antigen-specific GC Th cells (immune synapse III). The memory B-cell compartment exists in at least two major cellular fractions, the long-lived affinity-matured plasma cells and the affinity-matured memory-response precursors that are primed to respond rapidly to secondary antigen challenge. Formation of Germinal Center Microenvironment The formation of the GC reaction is generally thought to require T-cell help, as it is absent in athymic nude mice, CD40- and CD154 (CD40L)-deficient animals, and is diminished using reagents that block or deplete Th-cell function, such as anti-CD4, anti-CD40, and anti-CD28 ( 188 , 189 and 190 ). However, some T-independent antigens do induce a GC response and they can arise in T-deficient animals. In the latter case, the GC reaction appears with truncated kinetics and does not support somatic hypermutation ( 191 ). Using T-dependent antigens such as hapten–protein conjugates, there is evidence for the initiation of GC by 5 to 7 days after priming. Each GC develops as a discrete entity imposing the cycles of diversification and then selection upon an oligoclonal set of antigen-specific B cells. There is some evidence for secondary selection of the subset of antigen-specific B cells that enter the GC reaction compared with those in the T-zone pathway; however, these issues have not yet been carefully addressed. Upon committing to the GC pathway, clonal expansion that forms the secondary follicle proceeds with a B-cell doubling time

of about 6 to 8 hours. This rapid doubling time continues in the GC dark zone over the course of the primary response GC reaction (˜21 days) ( 182 ). Somatic Hypermutation and BCR Diversification Clonal expansion in the GC is associated with somatic diversification of the Ig receptor by a hypermutator mechanism ( Fig. 7). Somatic hypermutation underpins affinity maturation in the B-cell compartment; however, the molecular mechanism that drives this process remains largely unresolved. The B cells recruited into the GC reaction have already been selected based on the germline specificity of their rearranged V-region genes ( Fig. 7). Upon expansion in the GC reaction, the B cells down-regulate this germline sIg and somatically diversify their variable-region genes. Single base substitutions, rare insertions, and deletions are introduced into a region spanning 1.5 to 2.0 kb downstream of the transcription initiation site ( 192 ). Activity peaks within the V(D)J region and decreases within the J-C intronic region of IgH and IgL V genes. The mutation rate approaches 10 -3 per base pair per generation at six orders of magnitude higher than spontaneous mutation frequencies. Approximately one mutation is introduced with each cell division. Analysis of mutation in “passenger” Ig transgenes that are not under selection pressure indicates intrinsic sequence hot spots for the mutator mechanism ( 193 ). These analyses identify a motif referred to as RGYW (where R = A or G; Y = C or T; W = A or T), with AGC/T triplets for serine identified as preferred targets. Transitions are more frequent than transversions and A nucleotides in the coding strand are replaced more frequently than T nucleotides (referred to as strand bias). However, it is clear that mutation does not rely on the sequence of the target genes, but does require transcription of the target locus. Double stranded DNA breaks (DSB) have been identified in the V(D)J regions of mutating B cells ( 194 ). Most of these DSB also occur preferentially at RGYW motifs and most often 5' of the G and R residues. These DSB may represent the reaction intermediates of the hypermutation mechanism or entry points for an endonuclease to cleave DNA strands and initiate hypermutation. Mutations could be introduced into these lesions as mismatched nucleotides by an error-prone DNA polymerase during the repair process. Zan et al. ( 195 ) demonstrate a role for polymerase ? (POLZ) in a human B-cell line that can hypermutate in vitro. POLZ effectively extends DNA past mismatch base insertions and is up-regulated in this B-cell line upon BCR engagement in the presence of T-cell help. Blocking its action impaired hypermutation frequency. Polymerase ? (POLH) is also highly error prone, and while humans with defective enzyme appear to have normal frequency of hypermutation in their B cells, the pattern of mutation is significantly altered to less A/T and more G/C mutations ( 196 ). This suggests that multiple error-prone POLs may be involved in B cells. Polymerase ? (POLI) and µ (POLM) are two other highly error-prone polymerases that may play a role in hypermutation. However, after the introduction of mutations, there must be some means to subvert the mismatch repair mechanisms present in all cells. Curiously, mice deficient in nucleotide excision repair, mismatch repair, and base excision repair displayed normal levels of hypermutation, albeit with some differences in the overall patterns ( 197 ). A most remarkable finding in this field by Muramatsu et al. ( 198 ) is a role for the activation-induced cytidine deaminase (AID) as a central component for both somatic hypermutation and CSR. Originally discovered through cDNA subtraction focused on novel genes in GC B cells ( 198 ), it was then found to be the defect associated with an autosomal recessive form of hyper-IgM syndrome ( 180 ). Mice deficient in AID were able to form the GC reaction but were unable to undergo CSR or somatic hypermutation ( 179 ). AID may function as a

catalytic subunit of an RNA-editing enzyme complex as other members of this deaminase family and as such it may edit the RNA of a putative hypermutator and exert an indirect effect. Alternatively, it may function more directly as the enzyme that introduces nicks or single nucleotide gaps, as it can also deaminate deoxycytosines to uracil. If these lesions are not effectively repaired they have an increased spontaneous mutation rate causing G/C to A/T transitions. Although the precise role AID plays in hypermuation has yet to be resolved; it is clearly an early intermediate in two critical aspects of antigen-driven B-cell development—somatic hypermutation and CSR. Antigen-Driven Selection and GC Th Cells—Immune Synapse III Centroblast expansion and receptor diversification introduce different point mutations into the V-region genes of clonal progeny. GC microdissection, PCR amplification, and V-region sequence analysis directly identifies these clonally related genealogies in vivo ( 199 , 200 ). Noncycling centrocytes express the variant receptors and move into the light zones of the GC, a region rich in follicular dendritic cells (FDC) and scattered antigen-specific GC Th cells. FDC are nonphagocytic stromal cells that are involved in the organization of primary follicles. These cells have been implicated in antigen-based GC selection events due to expression of Fc?R and complement receptor-1 (CR-1) influencing their ability to trap native antigen as immune complexes. In contrast, animals lacking complement receptors, C3, or treatment with anti-CR1/2 displayed diminished or absent GCs and slower maturation of the humoral response ( 201 ). Using a BCR transgenic mouse model with B cells expressing only membrane-bound antibody that were unable to secrete (and thereby unable to form immune complexes; IC), Hannum et al. ( 202 ) demonstrated normal GC formation and affinity maturation. These data argue against the requirement for IC in GC selection events. Antigen-specific Th cells are also enriched in the GC environment of an ongoing immune response ( Fig. 7). These GC Th cells express low to negative levels of CD90 (Thy 1), unlike their non-GC counterparts, and are very sensitive to apoptosis induction via CD3 signaling ( 203 ). Anti-CD154 (CD40L) or anti-B7-2 antibody can disrupt ongoing GC reactions, presumably by interfering with cognate GC Th-cell–GC B-cell interactions. As the surface phenotype of the GC Th cells and the GC B cells are substantially different from their T-zone counterparts, it is reasonable to consider this distinct cognate cellular interchange as immune synapse III. The specialized function of the GC Th cells is not yet well understood, but these GC Th-B interactions serve to propagate B-cell memory and may help to interpret secondary antigen-selection events. Most mutational events in the BCR are deleterious to antigen binding and result in loss of the variants by apoptosis in the GC reaction. In support of this idea, overexpression of anti-apoptotic molecules, Bcl-2 or Bcl-xL, results in the accumulation of low-affinity and/or autoreactive B cells as either long-lived plasma cells or memory B cells ( 204 , 205 ). CD95 (Fas) is also expressed in GC B cells and these cells are highly susceptible to CD95L-induced apoptosis. Mice with defective CD95 function (lpr mutation) display a lymphoproliferative disorder with accumulation of autoreactive B cells. There is evidence for an impact on clonal selection in the GC and subsequent entry into the memory B-cell compartment of these animals. GC B cells appear poised to undergo apoptosis expressing high levels of CD95, c-Myc, P 53, and Bax and low levels of bcl-2. Studies using human tonsillar GC B cells indicate preformed death-inducing signaling complex (DISC) that are held inactivated by c-FLIP ( 206 ). The c-FLIP is rapidly down-regulated in vitro in the absence of stimuli and can be prevented by CD40 ligation. Most studies support the susceptibility of GC B cells to apoptosis unless rescued by antigen and the appropriate T-cell help.

The B-Cell Memory Response—Phase IV Exit from the GC reaction is one consequence of positive selection based on the increased affinity for antigen ( Fig. 7). These post–GC B cells exist long term as multiple cellular subsets that form the memory B-cell compartment. Stable maintenance of B-cell memory requires cell longevity that does not appear to need the continued expression of the affinity-matured BCR. These data argue that a continued antigen depot is not needed for memory B-cell survival. Finally, accelerated cellular expansion and rapid differentiation to high-affinity plasma cells is the hallmark of the memory B-cell response to antigen recall. This rapid cellular response is regulated by antigen-specific memory Th cells and constitutes Phase IV of the B-cell response that is controlled by the formation of immune synapse IV between memory Th cells and antigen-activated memory B cells ( Fig. 8).

FIG. 8. Phase IV: memory response to antigen re-challenge. Memory B-cell responses to T-dependent antigens also require T-cell help. It is thought that the memory B cell acts as the main APC in these memory responses. Hence, immune synapse IV can be thought to occur between antigen-activated memory B cells and memory Th cells and most likely occurs in the T-cell zones of secondary lymphoid tissue. Massive and rapid clonal expansion ensues in both the memory B-cell and memory Th-cell compartments with substantial plasma cell production evident in the T zones of these organs. These plasma cells appear short-lived and secrete the affinity-matured range of BCR expressed by their memory cell precursors. There is also evidence of secondary-response germinal centers that may be seeded by memory response precursors or naïve B cells with a different spectrum of BCR. Memory B-Cell Subsets Post–GC B cells can be considered to persist in two broad categories: long-lived plasma cells and memory response precursors. These long-lived plasma cells display evidence for somatic hypermutation, produce isotype-switched, high-affinity antibody, and preferentially home to the bone marrow with greatly extended half-lives compared to their T-zone/red-pulp plasma cell counterparts ( 207 ). Long-lived plasma cells do not appear to self-replenish but can survive in the absence of transferred antigen with half-lives ˜140 days as estimated by Slifka et al. ( 208 ). Animals deficient in CXCR4 expression have increased antigen-specific plasma cells in the peripheral blood but a reduced cell number in the bone marrow. Plasma cells express high levels of the integrin a4ß1, and animals deficient in its ligand VCAM1 are depleted

of mature recirculating B cells in the bone marrow and display decreased T-dependent B-cell responses ( 209 ). Interestingly, CD22-deficient mice also have a decreased number of recirculating B cells in the bone marrow, and injecting normal mice with a soluble CD22-Fc to block interactions of CD22 with its ligands (a2,6-linked sialic acids on glycans) decreases the numbers of plasma cells in the bone marrow ( 210 ). Hence, there may be multiple mechanisms for regulating plasma-cell homing to the bone marrow that may represent signals for both homing and long-term survival in this specialized microenvironment. Memory response precursors can be broadly defined as a residual population of antigen-experienced B cells that are not actively secreting antibody. Antigen experience preprograms these memory B cells to respond rapidly to secondary encounters with antigen under the regulation of memory Th cells. Accelerated clonal expansion and exaggerated plasma cell differentiation are the cardinal cellular characteristics of a memory response. Invariably, memory B cells will express isotype-switched and mutated BCR with evidence for affinity-based selection ( 211 ). Martin and Goodnow ( 212 ) recently demonstrated that the cytoplasmic tail of IgG was sufficient to increase the clonal burst potential of naïve B cells upon primary encounter with antigen. These data indicate one mechanism for quantitatively altering the memory B-cell response to antigen. Our group has recently identified multiple subsets of antigen-specific memory-response precursors based on the expression of cell-surface phenotype (mainly the expression of particular CD45 glycoform seen by mAb 6B2). The two main subsets (6B2 + and 6B2 -) differ in localization (peripheral lymphoid tissue versus bone marrow, respectively), and proliferative and differentiative capacity upon adoptive transfer and antigen re-exposure ( 213 , 214 ). The cell-surface phenotype of these two main memory B-cell subsets also suggests they have overtly differing responses to BCR triggering with differences in co-receptor and complement-receptor expression (6B2 +CD19 +CD21 +CD22 +; 6B2 -CD19 -CD21 + CD11b +CD22 -). Linton et al. ( 215 ) have identified multiple potentials for antigen responsiveness in the pre-immune compartment. They demonstrate subsets of naïve B cells that have a greater propensity to produce antibody-secreting cells (CD24 high) and others more likely to form memory-response precursors (CD24low). While the understanding of B-cell memory is still rudimentary, it is clear that its heterogeneous cellular organization indicates complex molecular regulation for both survival and response to recall. Memory B-Cell Maintenance The longevity of memory B cells appears independent of antigen. Initial transfer analyses by Gray and Skarvall ( 216 ) suggested a constant source of antigen was required for long-term memory B-cell survival. Maruyama et al. ( 217 ) recently addressed this question in an elegant Cre-LoxP–mediated genetic manipulation of an animal model. Animals were engineered to express two Ig heavy-chain alleles with expected specificity to two separate antigens (the hapten, NP, and the protein phycoerythrin, PE with the heavy chain in the opposite orientation) and then immunized to NP. Once NP-specific memory B cells were produced after immunization, the BCR heavy chain was switched by Cre-mediated recombination to lose the NP specificity and express the PE heavy chain. These BCR-switched memory B cells survived for 15 weeks in the absence of any exposure to PE, thus indicating no requirement for specific antigen. The PE-specific response had been used previously by the same group to demonstrate that memory B cells survived for extended periods with very low cell turnover. The Response to Antigen Recall—Immune Synapse IV The role of memory Th cells

in the regulation of memory B-cell responses has not been clearly addressed to date. Abrogation of CD4 Th cells using ablative anti-CD4 treatment clearly decreases subsequent memory B-cell responses in vivo. It is also apparent that all memory Th cells do not express similar functions, at least with respect to cytokine secretion ( 150 ). The APC for the memory B-cell response are most likely antigen-specific memory B cells. These cells have high-affinity receptors for antigen, and memory responses can be triggered with very low doses of antigen in the absence of any adjuvant. In a similar manner to the APC impact on naïve Th cells in the primary response, the quality of the initial memory B-cell response to antigen recall could significantly impact the differentiation of memory T-cell help. Hence, the nature of cognate help in the induction of memory B-cell responses may again be qualitatively and quantitatively distinct from any previous synaptic interaction and can be considered immune synapse IV ( Fig. 8). Synapse IV interactions will still be antigen specific and require pMHC expression by the memory B cells. However, the constellation of co-stimulatory/co-receptor molecules involved in this specialized molecular exchange has yet to be carefully investigated. In summary, B-cell activation in vivo takes on multiple forms that depend on the nature of the antigen and the developmental program of the B cell recruited to respond. Some subsets of B cells appear to be specialized to respond to T-independent forms of antigen in a manner that suggests the presence of a natural, evolutionarily conserved B-cell memory. In contrast, the response to most protein antigens requires the activation of the innate immune system and the recruitment of antigen-specific Th cells. These T-dependent B-cell responses develop as elaborate cellular cascades of interdependent activity, critically regulated over time and spatially organized within specialized microenvironments. Cognate interactions of intercellular information exchange act as the key checkpoints to this complex process. The primary response can be divided into three broad phases of cellular development culminating in the development of high-affinity B-cell memory as the cellular product of the germinal center reaction. Long-term protection is served through the ability of these memory B cells to respond very rapidly to subclinical re-challenge with antigen under the regulation of memory Th cells and the heightened reactivity of this adaptive arm of the immune system. REFERENCES 1. Reth M, Wienands J. Initiation and processing of signals from the B cell antigen receptor. Annu Rev Immunol 1997;15:453–479. 2. DeFranco AL. The complexity of signaling pathways activated by the BCR. Curr Opin Immunol 1997;9:296–308. 3. Kurosaki T. Genetic analysis of B cell antigen receptor signaling. Annu Rev Immunol 1999;17:555–592. 4. Shaw AC, Mitchell RN, Weaver YK, et al. Mutations of immunoglobulin transmembrane and cytoplasmic domains: effects on intracellular signaling and antigen presentation. Cell 1990;63:381–392. 5. Schamel WW, Reth M. Monomeric and oligomeric complexes of the B cell antigen receptor. Immunity 2000;13:5–14. 6. Campbell KS, Hager EJ, Cambier JC. Alpha-chains of IgM and IgD antigen receptor complexes are differentially N-glycosylated MB-1–related molecules. J Immunol 1991;147:1575–1580.

7. Venkitaraman AR, Williams GT, Dariavach P, et al. The B-cell antigen receptor of the five immunoglobulin classes. Nature 1991;352:777–781. 8.

Reth M. Antigen receptor tail clue. Nature 1989;338:383–384.

9. Sanchez M, Misulovin Z, Burkhardt AL, et al. Signal transduction by immunoglobulin is mediated through Ig alpha and Ig beta. J Exp Med 1993;178:1049–1055. 10. Matsuuchi L, Gold MR. New views of BCR structure and organization. Curr Opin Immunol 2001;13:270–277. 11. Lam KP, Kuhn R, Rajewsky K. In vivo ablation of surface immunoglobulin on mature B cells by inducible gene targeting results in rapid cell death. Cell 1997;90:1073–1083. 12.

Simons K, Toomre D. Lipid rafts and signal transduction. Nat Rev Mol Cell Biol 2000;1:31–39.


Pierce SK. Lipid rafts and B-cell activation. Nature Rev Immunol 2002;2:96–105.


Simons K, Ikonen E. Functional rafts in cell membranes. Nature 1997;387:569–572.

15. Zhang W, Trible RP, Samelson LE. LAT palmitoylation: its essential role in membrane microdomain targeting and tyrosine phosphorylation during T cell activation. Immunity 1998;9:239–246. 16. Guo B, Kato RM, Garcia-Lloret M, et al. Engagement of the human pre–B cell receptor generates a lipid raft- dependent calcium signaling complex. Immunity 2000;13:243–253. 17. Weintraub BC, Jun JE, Bishop AC, et al. Entry of B cell receptor into signaling domains is inhibited in tolerant B cells. J Exp Med 2000;191:1443–1448. 18. Bray D, Levin MD, Morton-Firth CJ. Receptor clustering as a cellular mechanism to control sensitivity. Nature 1998;393:85–88. 19. Gold MR, Matsuuchi L, Kelly RB, et al. Tyrosine phosphorylation of components of the B-cell antigen receptors following receptor crosslinking. Proc Natl Acad Sci U S A 1991;88:3436–3440. 20. Cherukuri A, Dykstra M, Pierce SK. Floating the raft hypothesis: lipid rafts play a role in immune cell activation. Immunity 2001;14:657–660. 21. Field KA, Holowka D, Baird B. Structural aspects of the association of FcepsilonRI with detergent-resistant membranes. J Biol Chem 1999;274:1753–1758. 22. Friedrichson T, Kurzchalia TV. Microdomains of GPI-anchored proteins in living cells revealed by crosslinking. Nature 1998;394:802–805. 23. Xavier R, Brennan T, Li Q, et al. Membrane compartmentation is required for efficient T cell activation. Immunity 1998;8:723–732. 24. Kabouridis PS, Janzen J, Magee AL, et al. Cholesterol depletion disrupts lipid rafts and modulates the activity of multiple signaling pathways in T lymphocytes. Eur J Immunol 2000;30:954–963. 25. Campbell MA, Sefton BM. Protein tyrosine phosphorylation is induced in murine B lymphocytes in response to stimulation with anti-immunoglobulin. EMBO J 1990;9:2125–2131. 26.

Carter RH, Park DJ, Rhee SG, et al. Tyrosine phosphorylation of phospholipase C induced by

membrane immunoglobulin in B lymphocytes. Proc Natl Acad Sci U S A 1991;88:2745–2749. 27. Flaswinkel H, Reth M. Dual role of the tyrosine activation motif of the Ig-alpha protein during signal transduction via the B cell antigen receptor. EMBO J 1994;13:83–89. 28. Wang J, Koizumi T, Watanabe T. Altered antigen receptor signaling and impaired Fas-mediated apoptosis of B cells in Lyn-deficient mice. J Exp Med 1996;184:831–838. 29. Chen T, Repetto B, Chizzonite R, et al. Interaction of phosphorylated FcepsilonRIgamma immunoglobulin receptor tyrosine activation motif-based peptides with dual and single SH2 domains of p72syk. Assessment of binding parameters and real time binding kinetics. J Biol Chem 1996;271:25308–25315. 30. Kimura T, Sakamoto H, Appella E, et al. Conformational changes induced in the protein tyrosine kinase p72syk by tyrosine phosphorylation or by binding of phosphorylated immunoreceptor tyrosine-based activation motif peptides. Mol Cell Biol 1996;16:1471–1478. 31. Kurosaki T, Johnson SA, Pao L, et al. Role of the Syk autophosphorylation site and SH2 domains in B cell antigen receptor signaling. J Exp Med 1995;182:1815–1823. 32.

Weiss A, Littman DR. Signal transduction by lymphocyte antigen receptors. Cell 1994;76:263–274.

33. Turner M, Mee PJ, Costello PS, et al. Perinatal lethality and blocked B-cell development in mice lacking the tyrosine kinase Syk. Nature 1995;378:298–302. 34.

Kurosaki T, Tsukada S. BLNK: connecting Syk and Btk to calcium signals. Immunity 2000;12:1–5.


Leo A, Schraven B. Adapters in lymphocyte signalling. Curr Opin Immunol 2001;13:307–316.

36. Ishiai M, Kurosaki M, Pappu R, et al. BLNK required for coupling Syk to PLC gamma 2 and Rac1-JNK in B cells. Immunity 1999;10:117–125. 37. Xu S, Tan JE, Wong EP, et al. B cell development and activation defects resulting in xid-like immunodeficiency in BLNK/SLP-65–deficient mice. Int Immunol 2000;12:397–404. 38. Takata M, Sabe H, Hata A, et al. Tyrosine kinases Lyn and Syk regulate B cell receptor-coupled Ca 2+ mobilization through distinct pathways. EMBO J 1994;13:1341–1349. 39. Fluckiger AC, Li Z, Kato RM, et al. Btk/Tec kinases regulate sustained increases in intracellular Ca 2+ following B-cell receptor activation. EMBO J 1998;17:1973–1985. 40. Okada T, Maeda A, Iwamatsu A, et al. BCAP: the tyrosine kinase substrate that connects B cell receptor to phosphoinositide 3-kinase activation. Immunity 2000;13:817–827. 41. Beitz LO, Fruman DA, Kurosaki T, et al. SYK is upstream of phosphoinositide 3-kinase in B cell receptor signaling. J Biol Chem 1999;274:32662–32666. 42. Tamir I, Dal Porto JM, Cambier JC. Cytoplasmic protein tyrosine phosphatases SHP-1 and SHP-2: regulators of B cell signal transduction. Curr Opin Immunol 2000;12:307–315. 43. Yamazaki T, Takeda K, Gotoh K, et al. Essential immunoregulatory role for BCAP in B cell development and function. J Exp Med 2002;195:535–545. 44.

Takata M, Kurosaki T. A role for Bruton’s tyrosine kinase in B cell antigen receptor-mediated

activation of phospholipase C-gamma 2. J Exp Med 1996;184:31–40. 45. Ward SG, Cantrell DA. Phosphoinositide 3-kinases in T lymphocyte activation. Curr Opin Immunol 2001;13:332–338. 46. Tsukada S, Rawlings DJ, Witte ON. Role of Bruton’s tyrosine kinase in immunodeficiency. Curr Opin Immunol 1994;6:623–630. 47. Thomas JD, Sideras P, Smith CI, et al. Colocalization of X-linked agammaglobulinemia and X-linked immunodeficiency genes. Science 1993;261:355–358. 48. Rawlings DJ, Saffran DC, Tsukada S, et al. Mutation of unique region of Bruton’s tyrosine kinase in immunodeficient XID mice. Science 1993;261:358–361. 49. Liu W, Quinto I, Chen X, et al. Direct inhibition of Bruton’s tyrosine kinase by IBtk, a Btk-binding protein. Nat Immunol 2001;2:939–946. 50. Law CL, Chandran KA, Sidorenko SP, et al. Phospholipase C-gamma1 interacts with conserved phosphotyrosyl residues in the linker region of Syk and is a substrate for Syk. Mol Cell Biol 1996;16:1305–1315. 51. Wang D, Feng J, Wen R, et al. Phospholipase Cgamma2 is essential in the functions of B cell and several Fc receptors. Immunity 2000;13:25–35. 52. Miyakawa T, Maeda A, Yamazawa T, et al. Encoding of Ca 2+ signals by differential expression of IP3 receptor subtypes. EMBO J 1999;18:1303–1308. 53.

Lewis RS. Calcium signaling mechanisms in T lymphocytes. Annu Rev Immunol 2001;19:497–521.


Clapham DE. Calcium signaling. Cell 1995;80:259–268.

55. Hagar RE, Burgstahler AD, Nathanson MH, et al. Type III InsP3 receptor channel stays open in the presence of increased calcium. Nature 1998;396:81–84. 56. Sugawara H, Kurosaki M, Takata M, et al. Genetic evidence for involvement of type 1, type 2 and type 3 inositol 1,4,5-trisphosphate receptors in signal transduction through the B-cell antigen receptor. EMBO J 1997;16:3078–3088. 57. Bubien JK, Zhou LJ, Bell PD, et al. Transfection of the CD20 cell surface molecule into ectopic cell types generates a Ca 2+ conductance found constitutively in B lymphocytes. J Cell Biol 1993;121:1121–1132. 58. Peng JB, Chen XZ, Berger UV, et al. Molecular cloning and characterization of a channel-like transporter mediating intestinal calcium absorption. J Biol Chem 1999;274:22739–22746. 59. Yue L, Peng JB, Hediger MA, et al. CaT1 manifests the pore properties of the calcium-release-activated calcium channel. Nature 2001;410:705–709. 60. Putney JW. “Kissin’ cousins”: intimate plasma membrane–ER interactions underlie capacitative calcium entry. Cell 1999;99:5–8. 61. Ma HT, Patterson RL, van Rossum DB, et al. Requirement of the inositol trisphosphate receptor for activation of store-operated Ca 2+ channels. Science 2000;287:1647–1651. 62.

Bikah G, Pogue-Caley RR, McHeyzer-Williams LJ, et al. Regulating T helper cell immunity through

antigen responsiveness and calcium entry. Nat Immunol 2000;1:402–412. 63. Serafini AT, Lewis RS, Clipstone NA, et al. Isolation of mutant T lymphocytes with defects in capacitative calcium entry. Immunity 1995;3:239–250. 64. Rao A, Luo C, Hogan PG. Transcription factors of the NFAT family: regulation and function. Annu Rev Immunol 1997;15:707–747. 65. Neilson J, Stankunas K, Crabtree GR. Monitoring the duration of antigen-receptor occupancy by calcineurin/glycogen-synthase-kinase-3 control of NF-AT nuclear shuttling. Curr Opin Immunol 2001;13:346–350. 66. Beals CR, Clipstone NA, Ho SN, et al. Nuclear localization of NF-ATc by a calcineurin-dependent, cyclosporin- sensitive intramolecular interaction. Genes Dev 1997;11:824–834. 67. Beals CR, Sheridan CM, Turck CW, et al. Nuclear export of NF-ATc enhanced by glycogen synthase kinase-3. Science 1997;275:1930–1934. 68. Ranger AM, Oukka M, Rengarajan J, et al. Inhibitory function of two NFAT family members in lymphoid homeostasis and Th2 development. Immunity 1998;9:627–635. 69. Hodge MR, Ranger AM, Charles de la Brousse F, et al. Hyperproliferation and dysregulation of IL-4 expression in NF-ATp–deficient mice. Immunity 1996;4:397–405. 70. Peng SL, Gerth AJ, Ranger AM, et al. NFATc1 and NFATc2 together control both T and B cell activation and differentiation. Immunity 2001;14:13–20. 71. Karin M, Ben-Neriah Y. Phosphorylation meets ubiquitination: the control of NF-[kappa]B activity. Annu Rev Immunol 2000;18:621–663. 72.

Karin M, Lin A. NF-kappaB at the crossroads of life and death. Nat Immunol 2002;3:221–227.

73. Dolmetsch RE, Lewis RS, Goodnow CC, et al. Differential activation of transcription factors induced by Ca 2+ response amplitude and duration. Nature 1997;386:855–858. 74. Dolmetsch RE, Xu K, Lewis RS. Calcium oscillations increase the efficiency and specificity of gene expression. Nature 1998;392:933–936. 75. Dorn GW 2nd, Mochly-Rosen D. Intracellular transport mechanisms of signal transducers. Annu Rev Physiol 2002;64:407–429. 76. Leitges M, Schmedt C, Guinamard R, et al. Immunodeficiency in protein kinase cbeta-deficient mice. Science 1996;273:788–791. 77. Kang SW, Wahl MI, Chu J, et al. PKCbeta modulates antigen receptor signaling via regulation of Btk membrane localization. EMBO J 2001;20:5692–5702. 78. Sidorenko SP, Law CL, Klaus SJ, et al. Protein kinase C mu (PKC mu) associates with the B cell antigen receptor complex and regulates lymphocyte signaling. Immunity 1996;5:353–363. 79. Fruman DA, Ferl GZ, An SS, et al. Phosphoinositide 3-kinase and Bruton’s tyrosine kinase regulate overlapping sets of genes in B lymphocytes. Proc Natl Acad Sci U S A 2002;99:359–364. 80. Gold MR, Scheid MP, Santos L, et al. The B cell antigen receptor activates the Akt (protein kinase B)/glycogen synthase kinase-3 signaling pathway via phosphatidylinositol 3-kinase. J Immunol

1999;163:1894–1905. 81. Stokoe D, Stephens LR, Copeland T, et al. Dual role of phosphatidylinositol-3,4,5-trisphosphate in the activation of protein kinase B. Science 1997;277:567–570. 82. Tsuruta F, Masuyama N, Gotoh Y. The PI3K-Akt pathway suppresses Bax translocation to mitochondria. J Biol Chem 2002;12:12. 83. Graziadei L, Riabowol K, Bar-Sagi D. Co-capping of ras proteins with surface immunoglobulins in B lymphocytes. Nature 1990;347:396–400. 84. Harmer SL, DeFranco AL. The src hom*ology domain 2-containing inositol phosphatase SHIP forms a ternary complex with Shc and Grb2 in antigen receptor-stimulated B lymphocytes. J Biol Chem 1999;274:12183–12191. 85. Tridandapani S, Kelley T, Cooney D, et al. Negative signaling in B cells: SHIP Grbs Shc. Immunol Today 1997;18:424–427. 86. Carpino N, Wisniewski D, Strife A, et al. p62(dok): a constitutively tyrosine-phosphorylated, GAP-associated protein in chronic myelogenous leukemia progenitor cells. Cell 1997;88:197–204. 87. Tamir I, Stolpa JC, Helgason CD, et al. The RasGAP-binding protein p62dok is a mediator of inhibitory FcgammaRIIB signals in B cells. Immunity 2000;12:347–358. 88. Hashimoto A, Okada H, Jiang A, et al. Involvement of guanosine triphosphatases and phospholipase C-gamma2 in extracellular signal-regulated kinase, c-Jun NH2-terminal kinase, and p38 mitogen-activated protein kinase activation by the B cell antigen receptor. J Exp Med 1998;188:1287–1295. 89. Jacinto E, Werlen G, Karin M. Cooperation between Syk and Rac1 leads to synergistic JNK activation in T lymphocytes. Immunity 1998;8:31–41. 90. Werlen G, Jacinto E, Xia Y, et al. Calcineurin preferentially synergizes with PKC-theta to activate JNK and IL-2 promoter in T lymphocytes. EMBO J 1998;17:3101–3111. 91. Latour S, Veillette A. Proximal protein tyrosine kinases in immunoreceptor signaling. Curr Opin Immunol 2001;13:299–306. 92. Cloutier JF, Veillette A. Cooperative inhibition of T-cell antigen receptor signaling by a complex between a kinase and a phosphatase. J Exp Med 1999;189:111–121. 93. Kawabuchi M, Satomi Y, Takao T, et al. Transmembrane phosphoprotein Cbp regulates the activities of Src-family tyrosine kinases. Nature 2000;404:999–1003. 94. Brdicka T, Pavlistova D, Leo A, et al. Phosphoprotein associated with glycosphingolipid-enriched microdomains (PAG), a novel ubiquitously expressed transmembrane adaptor protein, binds the protein tyrosine kinase csk and is involved in regulation of T cell activation. J Exp Med 2000;191:1591–1604. 95. Hermiston ML, Xu Z, Majeti R, et al. Reciprocal regulation of lymphocyte activation by tyrosine kinases and phosphatases. J Clin Invest 2002;109:9–14. 96. Batista FD, Iber D, Neuberger MS. B cells acquire antigen from target cells after synapse formation. Nature 2001;411:489–494. 97.

Kelm S, Pelz A, Schauer R, et al. Sialoadhesin, myelin-associated glycoprotein and CD22 define a

new family of sialic acid-dependent adhesion molecules of the immunoglobulin superfamily. Curr Biol 1994;4:965–972. 98. Engel P, Nojima Y, Rothstein D, et al. The same epitope on CD22 of B lymphocytes mediates the adhesion of erythrocytes, T and B lymphocytes, neutrophils, and monocytes. J Immunol 1993;150:4719–4732. 99. Tedder TF, Tuscano J, Sato S, et al. CD22, a B lymphocyte-specific adhesion molecule that regulates antigen receptor signaling. Annu Rev Immunol 1997;15:481–504. 100. Doody GM, Justement LB, Delibrias CC, et al. A role in B cell activation for CD22 and the protein tyrosine phosphatase SHP. Science 1995;269:242–244. 101. Sato S, Miller AS, Inaoki M, et al. CD22 is both a positive and negative regulator of B lymphocyte antigen receptor signal transduction: altered signaling in CD22-deficient mice. Immunity 1996;5:551–562. 102. Cyster JG, Goodnow CC. Protein tyrosine phosphatase 1C negatively regulates antigen receptor signaling in B lymphocytes and determines thresholds for negative selection. Immunity 1995;2:13–24. 103. Fujimoto M, Bradney AP, Poe JC, et al. Modulation of B lymphocyte antigen receptor signal transduction by a CD19/CD22 regulatory loop. Immunity 1999;11:191–200. 104. Tedder TF, Zhou LJ, Engel P. The CD19/CD21 signal transduction complex of B lymphocytes. Immunol Today 1994;15:437–442. 105. Dempsey PW, Allison ME, Akkaraju S, et al. C3d of complement as a molecular adjuvant: bridging innate and acquired immunity. Science 1996;271:348–350. 106. Cherukuri A, Cheng PC, Sohn HW, et al. The CD19/CD21 complex functions to prolong B cell antigen receptor signaling from lipid rafts. Immunity 2001;14:169–179. 107. Fujimoto M, Fujimoto Y, Poe JC, et al. CD19 regulates Src family protein tyrosine kinase activation in B lymphocytes through processive amplification. Immunity 2000;13:47–57. 108.

Ravetch JV, Bolland S. IgG Fc receptors. Annu Rev Immunol 2001;19:275–290.

109. D’Ambrosio D, Fong DC, Cambier JC. The SHIP phosphatase becomes associated with Fc gammaRIIB1 and is tyrosine phosphorylated during “negative” signaling. Immunol Lett 1996;54:77–82. 110.

Hardy RR, Hayakawa K. B cell development pathways. Annu Rev Immunol 2001;19:595–621.

111. Berland R, Wortis HH. Origins and functions of B-1 cells with notes on the role of cd5. Annu Rev Immunol 2002;20:253–300. 112. Arnold LW, Pennell CA, McCray SK, et al. Development of B-1 cells: segregation of phosphatidyl choline-specific B cells to the B-1 population occurs after immunoglobulin gene expression. J Exp Med 1994;179:1585–1595. 113. Clarke SH, Arnold LW. B-1 cell development: evidence for an uncommitted immunoglobulin (Ig)M+ B cell precursor in B-1 cell differentiation. J Exp Med 1998;187:1325–1334. 114. Martin F, Kearney JF. B1 cells: similarities and differences with other B cell subsets. Curr Opin Immunol 2001;13:195–201. 115.

Bendelac A, Bonneville M, Kearney JF. Autoreactivity by design: innate B and T lymphocytes.

Nature Rev Immunol 2001;1:177–186. 116. Hayakawa K, Asano M, Shinton SA, et al. Positive selection of natural autoreactive B cells. Science 1999;285:113–116. 117. Ansel KM, Harris RB, Cyster JG. CXCL13 is required for B1 cell homing, natural antibody production, and body cavity immunity. Immunity 2002;16:67–76. 118. Martin F, Kearney JF. Positive selection from newly formed to marginal zone B cells depends on the rate of clonal production, CD19, and btk. Immunity 2000;12:39–49. 119. Guinamard R, Okigaki M, Schlessinger J, et al. Absence of marginal zone B cells in Pyk-2–deficient mice defines their role in the humoral response. Nat Immunol 2000;1:31–36. 120. Wang JH, Avitahl N, Cariappa A, et al. Aiolos regulates B cell activation and maturation to effector state. Immunity 1998;9:543–553. 121. Medzhitov R, Janeway CA Jr. Innate immunity: the virtues of a nonclonal system of recognition. Cell 1997;91:295–298. 122. Medzhitov R, Preston-Hurlburt P, Janeway CA Jr. A human hom*ologue of the Drosophila Toll protein signals activation of adaptive immunity. Nature 1997;388:394–397. 123. Poltorak A, He X, Smirnova I, et al. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 1998;282:2085–2088. 124. Takeuchi O, Hoshino K, Kawai T, et al. Differential roles of TLR2 and TLR4 in recognition of gram-negative and gram-positive bacterial cell wall components. Immunity 1999;11:443–451. 125. Fearon DT, Locksley RM. The instructive role of innate immunity in the acquired immune response. Science 1996;272:50–53. 126. Dintzis HM, Dintzis RZ, Vogelstein B. Molecular determinants of immunogenicity: the immunon model of immune response. Proc Natl Acad Sci U S A 1976;73:3671–3675. 127. Martin F, Oliver AM, Kearney JF. Marginal zone and B1 B cells unite in the early response against T-independent blood-borne particulate antigens. Immunity 2001;14:617–629. 128. Ochsenbein AF, Fehr T, Lutz C, et al. Control of early viral and bacterial distribution and disease by natural antibodies. Science 1999;286:2156–2159. 129. von Bulow GU, Bram RJ. NF-AT activation induced by a CAML-interacting member of the tumor necrosis factor receptor superfamily. Science 1997;278:138–141. 130. Laabi Y, Egle A, Strasser A. TNF cytokine family: more BAFF-ling complexities. Curr Biol 2001;11:R1013–R1016. 131. von Bulow GU, van Deursen JM, Bram RJ. Regulation of the T-independent humoral response by TACI. Immunity 2001;14:573–582. 132. Gross JA, Johnston J, Mudri S, et al. TACI and BCMA are receptors for a TNF hom*ologue implicated in B-cell autoimmune disease. Nature 2000;404:995–999. 133.

Marsters SA, Yan M, Pitti RM, et al. Interaction of the TNF hom*ologues BLyS and APRIL with the

TNF receptor hom*ologues BCMA and TACI. Curr Biol 2000;10:785–788. 134. Wu Y, Bressette D, Carrell JA, et al. Tumor necrosis factor (TNF) receptor superfamily member TACI is a high affinity receptor for TNF family members APRIL and BLyS. J Biol Chem 2000;275:35478–35485. 135. Gross JA, Dillon SR, Mudri S, et al. TACI-Ig neutralizes molecules critical for B cell development and autoimmune disease. impaired B cell maturation in mice lacking BLyS. Immunity 2001;15:289–302. 136. fa*garasan S, Honjo T. T-Independent immune response: new aspects of B cell biology. Science 2000;290:89–92. 137.

Ravetch JV, Lanier LL. Immune inhibitory receptors. Science 2000;290:84–89.

138. Nishimura H, Nose M, Hiai H, et al. Development of lupus-like autoimmune diseases by disruption of the PD-1 gene encoding an ITIM motif-carrying immunoreceptor. Immunity 1999;11:141–151. 139. Lanzavecchia A, Sallusto F. The instructive role of dendritic cells on T cell responses: lineages, plasticity and kinetics. Curr Opin Immunol 2001;13:291–198. 140. Sallusto F, Mackay CR, Lanzavecchia A. The role of chemokine receptors in primary, effector, and memory immune responses. Annu Rev Immunol 2000;18:593–620. 141. Cyster JG. Chemokines and cell migration in secondary lymphoid organs. Science 1999;286:2098–2102. 142. Nakano H, Yanagita M, Gunn MD. CD11c(+)B220(+)Gr-1(+) cells in mouse lymph nodes and spleen display characteristics of plasmacytoid dendritic cells. J Exp Med 2001;194:1171–1178. 143. Monks CR, Freiberg BA, Kupfer H, et al. Three-dimensional segregation of supramolecular activation clusters in T cells. Nature 1998;395:82–86. 144. Grakoui A, Bromley SK, Sumen C, et al. The immunological synapse: a molecular machine controlling T cell activation. Science 1999;285:221–227. 145. Bromley SK, Iaboni A, Davis SJ, et al. The immunological synapse and CD28-CD80 interactions. Nat Immunol 2001;2:1159–1166. 146. Wulfing C, Davis MM. A receptor/cytoskeletal movement triggered by costimulation during T cell activation. Science 1998;282:2266–2269. 147. Lee KH, Holdorf AD, Dustin ML, et al. T cell receptor signaling precedes immunological synapse formation. Science 2002;295:1539–1542. 148. McHeyzer-Williams MG, Davis MM. Antigen-specific development of primary and memory T cells in vivo. Science 1995;268:106–111. 149. McHeyzer-Williams LJ, Panus JF, Mikszta JA, et al. Evolution of antigen-specific T cell receptors in vivo: preimmune and antigen-driven selection of preferred complementarity-determining region 3 (CDR3) motifs. J Exp Med 1999;189:1823–1838. 150. Panus JF, McHeyzer-Williams LJ, McHeyzer-Williams MG. Antigen-specific T helper cell function: differential cytokine expression in primary and memory responses. J Exp Med 2000;192:1301–1316. 151.

Ansel KM, McHeyzer-Williams LJ, Ngo VN, et al. In vivo–activated CD4 T cells upregulate CXC

chemokine receptor 5 and reprogram their response to lymphoid chemokines. J Exp Med 1999;190:1123–1134. 152. Watts C. Capture and processing of exogenous antigens for presentation on MHC molecules. Annu Rev Immunol 1997;15:821–850. 153. Cheng PC, Dykstra ML, Mitchell RN, et al. A role for lipid rafts in B cell antigen receptor signaling and antigen targeting. J Exp Med 1999;190:1549–1560. 154. Garside P, Ingulli E, Merica RR, et al. Visualization of specific B and T lymphocyte interactions in the lymph node. Science 1998;281:96–99. 155. Reif K, Ekland EH, Ohl L, et al. Balanced responsiveness to chemoattractants from adjacent zones determines B-cell position. Nature 2002;416:94–99. 156. Calderhead DM, Kosaka Y, Manning EM, et al. CD40-CD154 interactions in B-cell signaling. Curr Top Microbiol Immunol 2000;245:73–99. 157. Banchereau J, Bazan F, Blanchard D, et al. The CD40 antigen and its ligand. Annu Rev Immunol 1994;12:881–922. 158. Grewal IS, Xu J, Flavell RA. Impairment of antigen-specific T-cell priming in mice lacking CD40 ligand. Nature 1995;378:617–620. 159. van Essen D, Kikutani H, Gray D. CD40 ligand-transduced co-stimulation of T cells in the development of helper function. Nature 1995;378:620–623. 160. Bishop GA, Hostager BS. B lymphocyte activation by contact-mediated interactions with T lymphocytes. Curr Opin Immunol 2001;13:278–285. 161. Murata K, Ishii N, Takano H, et al. Impairment of antigen-presenting cell function in mice lacking expression of OX40 ligand. J Exp Med 2000;191:365–374. 162. Tangye SG, Liu YJ, Aversa G, et al. Identification of functional human splenic memory B cells by expression of CD148 and CD27. J Exp Med 1998;188:1691–1703. 163. Agematsu K, Hokibara S, Nagumo H, et al. CD27: a memory B-cell marker. Immunol Today 2000;21:204–206. 164. Cerutti A, Schaffer A, Shah S, et al. CD30 is a CD40-inducible molecule that negatively regulates CD40-mediated immunoglobulin class switching in non–antigen-selected human B cells. Immunity 1998;9:247–256. 165. Amakawa R, Hakem A, Kundig TM, et al. Impaired negative selection of T cells in Hodgkin’s disease antigen CD30-deficient mice. Cell 1996;84:551–562. 166.

Krammer PH. CD95’s deadly mission in the immune system. Nature 2000;407:789–795.


Sharpe AH, Freeman GJ. The B7-CD28 superfamily. Nature Rev Immunol 2002;2:116–126.

168. Walunas TL, Lenschow DJ, Bakker CY, et al. CTLA-4 can function as a negative regulator of T cell activation. Immunity 1994;1:405–413. 169.

Waterhouse P, Penninger JM, Timms E, et al. Lymphoproliferative disorders with early lethality in

mice deficient in Ctla-4. Science 1995;270:985–988. 170. Coyle AJ, Lehar S, Lloyd C, et al. The CD28-related molecule ICOS is required for effective T cell-dependent immune responses. Immunity 2000;13:95–105. 171. McAdam AJ, Greenwald RJ, Levin MA, et al. ICOS is critical for CD40-mediated antibody class switching. Nature 2001;409:102–105. 172. Dong C, Juedes AE, Temann UA, et al. ICOS co-stimulatory receptor is essential for T-cell activation and function. Nature 2001;409:97–101. 173. Tafuri A, Shahinian A, Bladt F, et al. ICOS is essential for effective T-helper-cell responses. Nature 2001;409:105–109. 174. Armitage RJ, Fanslow WC, Strockbine L, et al. Molecular and biological characterization of a murine ligand for CD40. Nature 1992;357:80–82. 175. Coffman RL, Seymour BW, Lebman DA, et al. The role of helper T cell products in mouse B cell differentiation and isotype regulation. Immunol Rev 1988;102:5–28. 176. Kuhn R, Rajewsky K, Muller W. Generation and analysis of interleukin-4 deficient mice. Science 1991;254:707–710. 177. Honjo T, Kinosh*ta K, Muramatsu M. Molecular mechanism of class switch recombination: linkage with somatic hypermutation. Annu Rev Immunol 2002;20:165–196. 178. Okazaki I, Kinosh*ta K, Muramatsu M, et al. The AID enzyme induces class switch recombination in fibroblasts. Nature 2002;727:1–5. 179. Muramatsu M, Kinosh*ta K, fa*garasan S, et al. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 2000;102:553–563. 180. Revy P, Muto T, Levy Y, et al. Activation-induced cytidine deaminase (AID) deficiency causes the autosomal recessive form of the Hyper-IgM syndrome (HIGM2). Cell 2000;102:565–575. 181. Jacob J, Kassir R, Kelsoe G. In situ studies of the primary immune response to (4-hydroxy-3-nitrophenyl)acetyl. I. The architecture and dynamics of responding cell populations. J Exp Med 1991;173:1165–1175. 182. Liu YJ, Zhang J, Lane PJ, et al. Sites of specific B cell activation in primary and secondary responses to T cell-dependent and T cell-independent antigens. Eur J Immunol 1991;21:2951–2962. 183. Hargreaves DC, Hyman PL, Lu TT, et al. A coordinated change in chemokine responsiveness guides plasma cell movements. J Exp Med 2001;194:45–56. 184. Turner CA Jr, Mack DH, Davis MM. Blimp-1, a novel zinc finger-containing protein that can drive the maturation of B lymphocytes into immunoglobulin-secreting cells. Cell 1994;77:297–306. 185. Calame KL. Plasma cells: finding new light at the end of B cell development. Nat Immunol 2001;2:1103–1108. 186. Reimold AM, Iwakoshi NN, Manis J, et al. Plasma cell differentiation requires the transcription factor XBP-1. Nature 2001;412:300–307.

187. Mittrucker HW, Matsuyama T, Grossman A, et al. Requirement for the transcription factor LSIRF/IRF4 for mature B and T lymphocyte function. Science 1997;275:540–543. 188. Jacobson EB, Caporale LH, Thorbecke GJ. Effect of thymus cell injections on germinal center formation in lymphoid tissues of nude (thymusless) mice. Cell Immunol 1974;13:416–430. 189. Kawabe T, Naka T, Yoshida K, et al. The immune responses in CD40-deficient mice: impaired immunoglobulin class switching and germinal center formation. Immunity 1994;1:167–178. 190. Renshaw BR, Fanslow WC 3rd, Armitage RJ, et al. Humoral immune responses in CD40 ligand–deficient mice. J Exp Med 1994;180:1889–1900. 191. de Vinuesa CG, Cook MC, Ball J, et al. Germinal centers without T cells. J Exp Med 2000;191:485–494. 192.

Neuberger MS, Milstein C. Somatic hypermutation. Curr Opin Immunol 1995;7:248–254.

193. Jolly CJ, Wagner SD, Rada C, et al. The targeting of somatic hypermutation. Semin Immunol 1996;8:159–168. 194. Bross L, f*ckita Y, McBlane F, et al. DNA double-strand breaks in immunoglobulin genes undergoing somatic hypermutation. Immunity 2000;13:589–597. 195. Zan H, Komori A, Li Z, et al. The translesion DNA polymerase zeta plays a major role in Ig and bcl-6 somatic hypermutation. Immunity 2001;14:643–653. 196. Gearhart PJ, Wood RD. Emerging links between hypermutation of antibody genes and DNA polymerases. Nature Rev Immunol 2001;1:187–192. 197. Jacobs H, Bross L. Towards an understanding of somatic hypermutation. Curr Opin Immunol 2001;13:208–218. 198. Muramatsu M, Sankaranand VS, Anant S, et al. Specific expression of activation-induced cytidine deaminase (AID), a novel member of the RNA-editing deaminase family in germinal center B cells. J Biol Chem 1999;274:18470–18476. 199. Jacob J, Kelsoe G, Rajewsky K, et al. Intraclonal generation of antibody mutants in germinal centres. Nature 1991;354:389–392. 200. Berek C, Berger A, Apel M. Maturation of the immune response in germinal centers. Cell 1991;67:1121–1129. 201. Fearon DT, Carroll MC. Regulation of B lymphocyte responses to foreign and self-antigens by the CD19/CD21 complex. Annu Rev Immunol 2000;18:393–422. 202. Hannum LG, Haberman AM, Anderson SM, et al. Germinal center initiation, variable gene region hypermutation, and mutant B cell selection without detectable immune complexes on follicular dendritic cells. J Exp Med 2000;192:931–942. 203. Zheng B, Han S, Kelsoe G. T helper cells in murine germinal centers are antigen-specific emigrants that downregulate Thy-1. J Exp Med 1996;184:1083–1091. 204. Smith KG, Light A, O’Reilly LA, et al. bcl–2 transgene expression inhibits apoptosis in the germinal center and reveals differences in the selection of memory B cells and bone marrow antibody-forming

cells. J Exp Med 2000;191:475–484. 205. Takahashi Y, Cerasoli DM, Dal Porto JM, et al. Relaxed negative selection in germinal centers and impaired affinity maturation in bcl-xL transgenic mice. J Exp Med 1999;190:399–410. 206. Hennino A, Berard M, Krammer PH, et al. FLICE-inhibitory protein is a key regulator of germinal center B cell apoptosis. J Exp Med 2001;193:447–458. 207. McHeyzer-Williams MG, Ahmed R. B cell memory and the long-lived plasma cell. Curr Opin Immunol 1999;11:172–179. 208. Slifka MK, Antia R, Whitmire JK, et al. Humoral immunity due to long-lived plasma cells. Immunity 1998;8:363–372. 209. Koni PA, Joshi SK, Temann UA, et al. Conditional vascular cell adhesion molecule 1 deletion in mice: impaired lymphocyte migration to bone marrow. J Exp Med 2001;193:741–754. 210. Nitschke L, Floyd H, Ferguson DJ, et al. Identification of CD22 ligands on bone marrow sinusoidal endothelium implicated in CD22-dependent homing of recirculating B cells. J Exp Med 1999;189:1513–1518. 211. McHeyzer-Williams MG, Nossal GJ, Lalor PA. Molecular characterization of single memory B cells. Nature 1991;350:502–505. 212. Martin SW, Goodnow CC. Burst-enhancing role of the IgG membrane tail as a molecular determinant of memory. Nat Immunol 2002;3:182–188. 213. McHeyzer-Williams LJ, Cool M, McHeyzer-Williams MG. Antigen-specific B cell memory: expression and replenishment of a novel b220(-) memory b cell compartment. J Exp Med 2000;191:1149–1166. 214. Driver DJ, McHeyzer-Williams LJ, Cool M, et al. Development and maintenance of a B220 memory B cell compartment. J Immunol 2001;167:1393–1405. 215. Linton PL, Decker DJ, Klinman NR. Primary antibody-forming cells and secondary B cells are generated from separate precursor cell subpopulations. Cell 1989;59:1049–1059. 216. Gray D, Skarvall H. B-cell memory is short-lived in the absence of antigen. Nature 1988;336:70–73. 217. Maruyama M, Lam KP, Rajewsky K. Memory B-cell persistence is independent of persisting immunizing antigen. Nature 2000;407:636–642.

Chapter 8 T-Cell Antigen Receptors Fundamental Immunology

Chapter 8 Mark Davis

T Cells & NK Cells T-Cell Antigen Receptors

T-CELL RECEPTOR POLYPEPTIDES T-CELL RECEPTOR STRUCTURE aß T-Cell Receptor Structure ?d T-Cell Receptor Structure THE CD3 POLYPEPTIDES Sequence and Structure of the CD3 Polypeptides Intracellular Assembly and Degradation of the T-Cell Receptor–CD3 Complex CD3 Structure T-CELL RECEPTOR GENES Organization of the T-Cell Receptor a/d Locus Organization of the T-Cell Receptor ß Locus Organization of the T-Cell Receptor ? Locus Transcriptional Control of the T-Cell Receptor Genes Chromosomal Locations of T-Cell Receptor Genes and Translocations Associated with Disease Allelic Exclusion Commitment to the aß Lineage versus the ?d Lineage Other Genetic Mechanisms BIOCHEMISTRY OF aß T-CELL RECEPTOR–LIGAND INTERACTIONS Role of CD4 and CD8 TOPOLOGY OF T-CELL RECEPTOR–PEPTIDE/MAJOR HISTOCOMPATIBILITY COMPLEX INTERACTIONS T-Cell Receptor Plasticity aß T-CELL RECEPTOR AND SUPERANTIGENS A SECOND TYPE OF RECEPTOR: ?d-CD3 Identification of ?d T Cells ?d T Cells Contribute to Host Immune Defense Differently than aß T Cells Antigen Recognition by ?d T Cells Does Not Require Processing Other Antigen Specificities of ?d T Cells Complementarity-Determining Region 3 Length Distribution Analysis Shows ?d T-Cell Receptors Are More Immunoglobulin-like Multivalence of the Ligands Is Required for Activation through the ?d T-Cell Receptor COMPLEMENTARITY-DETERMINING REGION 3 DIVERSIFICATION: A GENERAL STRATEGY FOR T-CELL RECEPTORS AND IMMUNOGLOBULIN COMPLEMENTARITY TO ANTIGENS? CONCLUSIONS ACKNOWLEDGMENTS Color Plates REFERENCES

The characteristics of T-lymphocyte recognition and the nature of T-cell antigen receptors (TCRs) has been a difficult and controversial area for immunologists. However, since 1983 there has been tremendous progress in identifying the molecules

and genes that govern T cell recognition, and in more recent years, researchers have obtained the first concrete information on their biochemistry and structure. Although TCRs share many similarities, both structural and genetic, with B cell antigen receptors (immunoglobulins), they also possess a number of unique features related to their specific biological functions. For classically defined helper and cytotoxic T cells, the most important of these differences was first suggested by the experiments of Zinkernagel and Doherty, who showed that viral antigen recognition by cytotoxic T cells was possible only with a certain major histocompatibility complex (MHC) haplotype on the infected cell ( 1 , 2 ). Evidence for this phenomenon of MHC restricted “recognition” was also demonstrated for helper T cells ( 3 , 4 ). It is now known that this type of T-cell recognition involves fragments of antigens (e.g., peptides) bound to specific MHC molecules (see Chapter 19 and Chapter 20). Because all antigens must eventually be degraded, this form of T-cell recognition is very complementary to that of B cells, in which pathogens can escape recognition by obscuring an antibody binding site or employing “decoy” molecules. T-cell receptors occur as either of two distinct heterodimers, aß or ?d, both of which are expressed with the nonpolymorphic CD3 polypeptides ?, d, e, and ? and, in some cases, the ribonucleic acid (RNA) splicing variant of ?, ?, or Fce chains. The CD3 polypeptides, especially ? and its variants, are critical for intracellular signaling ( 5 ). The aß TCR heterodimer–expressing cells predominate in most lymphoid compartments (90% to 95%) of humans and mice, and they are responsible for the classical helper or cytotoxic T cell responses. In most cases, the aß TCR ligand is a peptide antigen bound to a class I or class II MHC molecule. T cells bearing ?d TCR are less numerous than the aß type in most cellular compartments of humans and mice. However, they make up a substantial fraction of T lymphocytes in cows, sheep, and chickens ( 6 ). Studies of the structural characteristics and specificity of ?d TCRs indicate that they are much more like immunoglobulins than like aß TCRs in their antigen recognition properties. In particular, they do not seem to require MHCs or other molecules to present antigens but instead appear to recognize antigens directly ( 7 ). Although it is not yet clear what role they play in the immune response, this is a very active area of current research, and many interesting leads are being pursued.

T-CELL RECEPTOR POLYPEPTIDES The search for the molecules responsible for T-cell recognition first focused on deriving antisera or monoclonal antibodies specific for molecules on T-cell surfaces. Ultimately, a number of groups identified “clonotypic” sera ( 8 ) or monoclonal antibodies ( 9 , 10 , 11 , 12 and 13 ). A number of these antibodies were able to block antigen specific responses by the T cells they were raised against or, when coated on a surface, could activate the T cells for which they are specific. They were also able to immunoprecipitate 85,000- to 90,000—molecular weight (MW) disulfide-bonded heterodimers from different T cell clones or hybridomas consisting of two 40,000- to 50,000-MW glycosylated subunits referred to as a and ß. Peptide mapping studies showed that there was a striking degree of polymorphism between heterodimers isolated from T cells of differing

specificity, which thus suggests that these antigen recognition molecules may be akin to immunoglobulins ( 14 , 15 ). Work in parallel to these serological studies exploited the small differences (˜2%) observed between B- and T-cell gene expression ( 16 ) and isolated both a murine ( 17 , 18 ) and a human ( 19 ) T-cell specific gene that had antibody-like V, J, and C region sequences and could rearrange in T-lymphocytes ( 18 ). This molecule was identified as TCRß by partial sequence analysis of immunoprecipitated materials ( 20 ). Subsequent subtractive cloning work rapidly identified two other candidate TCRs’ complementary deoxyribonucleic acids (cDNAs) identified as TCRa ( 21 , 22 ) and TCR? ( 23 ). It was quickly established that all antigen-specific helper or cytotoxic T-cell expressed TCRaß heterodimers. Where TCR? fit in remained a puzzle until work by Brenner et al. ( 24 ) showed that it was expressed on a small (5% to 10%) subset of peripheral T cells together with another polypeptide, TCRd. The nature of TCRd remained unknown until it was discovered within the TCRa locus, between the Va and Ja regions ( 25 ). Formal proof that the TCRa and TCRß subunits were sufficient to transfer antigen/MHC recognition from one T cell to another came from gene transfection experiments ( 26 , 27 ), and equivalent experiments have also been performed with ?d TCRs ( 28 ). As shown in Fig. , all TCR polypeptides have a similar primary structure, with distinct variable (V) and diversity (D) regions in the case of TCRß and TCRd, and with joining (J), and constant (C) regions exactly analogous to their immunoglobulin counterparts. They also share many of the amino acid residues thought to be important for the characteristic variable and constant domains of immunoglobulins ( 29 ). The Cß region is particularly hom*ologous, sharing 40% of its amino acid sequence with CK and C?. The TCR polypeptides all contain a single C region domain (versus up to four for immunoglobulins) followed by a connecting peptide or hinge region, usually containing the cysteine for the disulfide linkage, which joins the two chains of the heterodimer [some human TCR?d isoforms lack this cysteine and consequently are not disulfide linked ( 30 )]. N-linked glycosylation sites vary from two to four for each polypeptide, with no indications of O-linked sugar addition. C-terminals to the connecting peptide sequences are the hydrophobic transmembrane regions, which have no similarity to those of heavy immunoglobulin genes but instead have one or two positively charged residues that appear to be important for interaction with the CD3 molecules and T cell signaling, through interaction with the acidic residues found in all CD3 transmembrane regions. The newest member of the TCR polypeptide family is the pre–Ta chain, which serves as a chaperone for TCRß in early thymocytes, which is similar to the role of ?5 in pre–B cells. It was first identified and cloned by Groettrup et al. ( 31 ) and Saint-Ruf et al. ( 32 ). It has an interesting structure that consists of a single immunoglobulin constant region–like domain followed by a cysteine-containing connecting peptide and a transmembrane region containing two charged residues: an arginine and a lysine spaced identically to those on the TCRa transmembrane region. The cysteine in the connecting peptide is presumably what allows heterodimer formation with TCRß, and the similarity to TCRa in the transmembrane region is most likely to accommodate the CD3 polypeptides. In both mice and humans, the cytoplasmic tail is much longer than any of the TCR chains (37 and 120 amino acids, respectively), and the murine sequence contains two likely phosphorylation sites and sequences hom*ologous to an

SH3 domain–binding region. These are not present in the human sequence, however, and their functional significance is therefore questionable ( 32 ). Thus, the pre–Ta molecule could function as signaling intermediate independent of any of the CD3 polypeptides.

FIG. 1. Structural features of T-cell receptors and pre–T a polypeptides. Leader (L), variable (V), diversity (D), joining (J), and constant region (C) gene segments are indicated. TM and bold horizontal lines delineate the putative transmembrane regions; CHO indicates potential carbohydrate addition sites; C and S refer to cysteine residues that form interchain and intrachain disulfide bonds; R and K indicate the positively charged amino acids (arginine and lysine, respectively) that are found in the transmembrane regions.

T-CELL RECEPTOR STRUCTURE As just discussed, the sequences of TCR polypeptides show many similarities to immunoglobulins, and thus it has long been suggested that both heterodimers would be antibody-like in structure ( 18 , 19 , 33 ). These similarities include the number and spacing of specific cysteine residues within domains, which in antibodies form intrachain disulfide bonds. Also conserved are many of the interdomain and intradomain contact residues; in addition, secondary structure predictions are largely consistent with an immunoglobulin-like “ß barrel” structure. This consists of three to four antiparallel ß strands on one side of the “barrel” facing a similar number on the other side, with a disulfide bridge (usually) connecting the two ß “sheets” (sets of ß strands in the same plane) ( Fig 2A). A diagrammatic representation of a typical V-region structure is shown in Fig. 2B. All immunoglobulin V- and C-region domains have this structure, with slight variations in the number of ß strands in V-region domains (by convention, including V, D, and J sequences) in comparison with C-region domains.

FIG. 2. T-cell receptors ß and Va. A: Ribbon diagram of the first T-cell receptor crystal structure ( 34 ), showing the antiparallel ß sheets of a Vß-Cß polypeptide. The Vß and Cß domains show the classical eight and seven ß-strand “barrels” characteristic of immunoglobulin V and C domains, respectively. Also shown are the positions of the complementarity-determining region loops 1, 2, and 3 at the end of Vß and the fourth loop, which has been implicated in superantigen interactions. B: Schematic of the ß strands in typical V region domain, which contrast with the alterations found in a Va domain.

aß T-Cell Receptor Structure Efforts to derive x-ray crystal structures of TCR heterodimers and fragments of heterodimers have encountered many technical hurdles. One reason is that it requires engineering the molecules into a soluble form. A second is that many of the TCRs are heavily glycosylated, and it was necessary to eliminate most or all of the carbohydrates on each chain to obtain high-quality crystals. An alternative is to express soluble TCRs in insect cells, which have compact N-linked sugars, or in Escherichia coli, which have none. The first successes in TCR crystallization come from the work of Bentley et al. ( 34 ), who solved the structure of a Vß Cß polypeptide, and Fields et al. ( 35 ), who then solved the structure of a Va fragment. In general, these domains all are very immunoglobulin-like, with the classical ß-barrel structure in evidence in all three domains. At each end of the barrel in each V-region domain, there are four loops between the ß sheets, three of which form the complementarity-determining regions (CDRs) of immunoglobulins. The fourth loop, between the D and E strands, has been implicated in superantigen binding. The six CDR loops from the two variable domains form the antigen-binding surface of immunoglobulins and, as discussed later, TCRs as well. Whereas the Vß domain depicted in Fig. 2A follows the canonical V domain ß sheet structure, Va differs significantly in that one of the sheets has been translocated to the other half of the barrel (as schematized in Fig. 2B). This acts to remove a bulge in the side of the Va domain, and it has been suggested that this would allow dimers of TCRs or perhaps higher order structures to assemble ( 35 ). Ultimately, Garcia et al. ( 36 ) were able to solve the structure of the Ca in the context of a complete heterodimer, and it has a remarkable variation of the classical immunoglobulin-like domain ( Colorplate 1). Here there is only one half of the classical ß-barrel—that is, one set (or “sheet”) of ß strands—whereas the rest of the somewhat truncated domain exhibits random coils. This type of structure is unprecedented in the immunoglobulin. The functional significance of such a variant structure in unknown, but it has been suggested that this incompletely formed immunoglobulin-like domain may be responsible for the observed lability of TCRa, and this may allow greater flexibility in the regulation of its expression. Another possible explanation is that this alteration may be designed to accommodate

one or more of the CD3 molecules. With regard to complete heterodimer structures, there are now data from four aß ( 36 , 37 , 38 , 39 and 40 ) and one ?d heterodimer ( 41 ), and they largely resemble the crystallized fragment of an antibody (Fab). Although many features of these structures are shared with their antibody counterparts, several unusual features in the aß molecules may be significant. These include the following: (a) In one structure ( 36 ), four of seven N-linked sugars diffracted to high resolution, which indicates that they are not free to move very much and thus are likely to play a structural role, particularly in Ca:Cß interactions. This correlates with mutagenesis data indicating that certain Ca sugars could not be eliminated without abolishing protein expression in mammalian cells ( 42 ) and the disordered state of a Ca domain in the structure of a TCR lacking glycosylation ( 37 ). (b) There is much more contact between Vß and somewhat more between Va and Ca than in the equivalent regions of antibodies. (c) The geometry of the interaction of Va and Vß more closely resembles that of the CH3 domains of antibodies than VH and VL domains. (d) Between the CDR3 loops of Va and Vß, there can be a pocket that, in at least one case ( 37 )] accommodates a large side chain from the peptide bound to an MHC. Another key question is whether any conformational change occurs in the TCR upon ligand binding. Conformational changes in the TCR or in the CD3 polypeptides in particular may hold important clues as to the mechanism of signal transduction across the membrane after TCR engagement. ?d T-Cell Receptor Structure The crystal structure of a ?d TCR from a human T cell clone ( 41 ) that can be stimulated by small phosphate-containing compounds has been solved. Although both of its CDR3 loops are similar in length to those of aß TCRs whose structures have been determined, they protrude significantly from the rest of the putative binding surface and create a cleft between them. Portions of the CDR1?, CDR1d, and CDR2? combine with the clefts between the CDR3 loops to form a pocket that may be the phosphoantigen binding site. This is because its structure is similar to those of pockets that are found in antibodies that bind phosphate-containing antigens, and it is surrounded by positively charged amino acid residues contributed by CDR2?, CDR2d, and CDR3?, which is consistent with binding the negatively charged phosphate compounds ( 41 ). A unique feature of this structure is the unusually small angle between the variable and constant regions of the ?d TCR, in comparison with aß TCRs and antibodies. In addition, structural differences in C? and Cd and the locations of the disulfide bond between them may indicate distinct recognition and signaling properties in comparison with aß TCRs.

THE CD3 POLYPEPTIDES Immunoprecipitation of the TCR with anti-idiotypic antibodies after solubilization with the nonionic detergent noniodet P 40 (NP 40) shows only the a- and ß-chain heterodimer. However, the use of gentler detergents, such as digitonin or Triton-X100, reveals five

other proteins [as reviewed by Terhorst et al. ( 5 ) and Klausner et al. ( 43 )]. This is shown most clearly in a form of two-dimensional gel electrophoresis in which the first dimension is run without a reducing agent, whereas the second gel is run with one such agent ( Fig. 3). The result is that most proteins can be graphed along a diagonal, whereas the subunits from disulfide-bonded multimeric proteins fall below the diagonal. Analysis of murine T cells by this technique shows the two TCR subunits (a, ß) running at 40,000 MW together with CD3? (20,000), CD3d (25,000), CD3e (20,000), and a fourth running below the diagonal at 16,000 MW ( 44 ) (?). The fact that the e chain runs above the diagonal indicates that it migrates faster when disulfide bonds are intact than when they are broken. This in turn implies that there are intrachain disulfide bonds that hold the molecule in a compact configuration. The migration of the ? chain is indicative of a disulfide-bonded hom*odimer; however, further studies have shown that the ? chain can be part of a heterodimer in at least two forms. In murine T cells, the ? chain can disulfide-bond with a minor variant called the ? chain ( 45 , 46 ). This latter chain is an alternate splicing variant of the ?-chain gene ( 47 ). This alternatively spliced species of the ? chain is not found in significant quantities in human T cells ( 48 ). The second type of ? chain containing heterodimer contains the ? chain associated with the FceRI (FceRI?) and Fc?RIII (CD16) receptors ( 49 ).

FIG. 3. T-cell receptor (TCR) aß CD3 complexes. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) analysis of TCR-CD3 complexes studied with immunoprecipitation and the two-dimensional “diagonal” gel method of Goding et al. ( 291 ). T-cell hybridoma cells were surface labeled with iodine-125 and immunoprecipitated with an anti-TCR antibody. The first dimension was run on SDS-PAGE without reducing agents, and the second dimension included the reducing agents. Molecules that are not disulfide-linked multimers cluster along a diagonal, whereas those that are disulfide-linked multimers “fall off” the diagonal as their molecular weight decreases, as they dissociate into their component chains. Shown here are TCRa; TCRß; and CD3 ?, d, e, and ? chains, from Samelson et al. ( 44 ).

With regard to overall stoichiometry, current evidence suggests that there are two TCR heterodimers per CD3 cluster. This is based on a number of findings, particularly the work of Terhorst et al. ( 5 ) who showed that in a T-T hybridoma, a monoclonal antibody against one TCRaß pair could comodulate a second aß heterodimer. In addition, sucrose gradient centrifugation of TCR/CD3 showed a predicted molecular weight of 300 kD, more than 100 kD larger than expected from a minimal d subunit complex (a, ß, ?, d, e 2 , ? 2 ) ( 50 ). Another study suggesting that there are least two TCRs in a given CD3 complex is the Scatchard analysis, which indicated that the number of CD3e molecules on a T-cell surface equals the number of aß TCRs ( 51 , 52 and 53 ). Finally, Fernandez-Miguel et al. ( 53a) showed that in T cells that have two transgenic TCRß chains, antibodies to one Vß can immunoprecipitate the other. It was also found that they are often close enough to allow fluorescence energy transfer, which means that the two TCRß chains in a cluster are within 50 Å of each other ( 53a). Interestingly, it appears that the TCR complexes with CD3 have either CD3 ? or CD3d but not both, and these two receptor types are expressed in different ratios in different cells. Furthermore, in cell types that express the FceRI? chain, these two forms of the receptor can be further divided into those that contain the ?? hom*odimer and those that have the ?FceRI? heterodimer ( 49 ). Thus, as shown in Fig. 4, much of the evidence to-date suggests a stoichiometry of the core cluster being [aß] 2 [?/de] 2 [??] 4 , with a number of the variations involving FceR, as discussed previously. However, this has recently been disputed by Call and colleagues ( 53b), whose data support a single TCR heterodimer/CD3 complex.

FIG. 4. Structural features of the CD3 molecules. As in Figure 1, transmembrane regions (TM) carbohydrate addition sites (CHO) and cysteine residues (C) are indicated. In addition, negatively charged transmembrane residues (D for aspartic acid and E for

glutamic acid) as well putative phosphorylation sites are shown.

Sequence and Structure of the CD3 Polypeptides Figure 4 illustrates the principal structural features of the ?-, d-, e-, and ?-chain polypeptides as derived from gene cloning and sequencing [as reviewed by Terhorst et al. ( 5 ) and Clevers et al. ( 54 )]. The extracellular domains of the ?-, d-, and e-chains show a significant degree of similarity to one another. These domains retain the cysteines that have been shown to form intrachain disulfide bonds and each consists of a single immunoglobulin superfamily domain. The spacing of the cysteines in these domains indicates a compact immunoglobulin fold, similar to a constant region domain. All of the extracellular domains contain a pair of closely spaced cysteines just before the predicted membrane-spanning regions, and these are likely candidates for the formation of intermolecular disulfide bonds as described previously. The extracellular domain of the ? chain consists of only nine amino acids and contains the only cysteine, which is responsible for the disulfide linkage of the ?? hom*odimer or the ?FceRI? heterodimer. In the transmembrane regions, it is particularly striking that all of the CD3 polypeptides have a conserved negatively charged amino acid, complementary to the positive charges seen in the TCR transmembrane regions and also necessary for proper assembly ( 55 , 56 and 57 ). The intracellular domains of the ?, d, e, and ? chains are the intracellular signaling “domains” of the TCR heterodimer. Each of these molecules contains an amino acid sequence motif that can mediate cellular activation ( 58 ). In T cells that are defective in ?-chain expression, a small but significant amount of interleukin-2 production can be elicited through the use of either the superantigen SEA or an antibody specific for thy-1. However, the ? chain is required for optimal stimulation by antigen, and the intracellular sequences responsible for this activation are contained within as few as 18 amino acids with the sequence X2YX2L/IX7YX2L/I. Both of the tyrosines in this sequence motif are absolutely required to mediate signal transduction, because mutation of either completely prevents the mobilization of free Ca 2+ or cytolytic activity. This sequence occurs three times in the ? chain and once in each of the CD3 ?, d, e, and F ceRI? chains. There are also pairs of tyrosines present in the cytoplasmic domains of the ?, d, e, and ? chains ( Fig. 5). This sequence motif is also present in the mß-1 and B29 chains associated with the (immunoglobulin) ß-cell receptor and in the F ceRIß chain. The tyrosines in these cytoplasmic sequences are substrates for tyrosine phosphorylation, which is one of earliest steps in T cell signaling ( 58 ) and is thought to occur aberrantly in nonproductive T cell responses (e.g., antagonism, described later). Serine phosphorylation of the CD3? also occurs upon antigen or mitogenic stimulation of T cells ( 59 ) and thus may play a role as well.

FIG. 5. A model of the T-cell receptor aß–CD3 complex. Although the precise arrangement of T-cell receptor polypeptides and CD3 molecules in a given cluster is not known, the fragmentary data that exist have been schematized by Terhorst et al. ( 5 ) into this working model of the complex. This model is also consistent with the data of Fernandez-Miguel et al. ( 53a), but disputed by Call et al. ( 53b). From ( 5 ), with permission.

Intracellular Assembly and Degradation of the T-Cell Receptor–CD3 Complex The assembly of newly formed TCRa- and TCRß chains with the CD3 ?, d, e, and ? chains and their intracellular fate have been studied in detail ( 5 , 43 , 60 ). Studies have focused on mutant hybridoma lines that fail to express TCR on their cell surface, and in transfection studies, cDNA has been used for the different chains in the receptor. Experiments in a nonlymphoid cell system ( 61 ) have shown that TCRa can assemble with CD3d and CD3e but not with CD3? and CD3?. In contrast, the TCRß chain can assemble with any of the CD3 chains except the ? chain. When the CD3? chain was transfected with a or ß chain genes or with any of the three CD3 chains, no pairwise interaction occurred. Only when all six cDNAs were cotransfected was it shown that the ? chain could be coprecipitated with the other chains ( 61 ). On the basis of these data, a model has been proposed to suggest that the TCRa chain pairs with CD3d and CD3e chains and that the TCRß chain pairs with the CD3? and CD3e chains in the completed molecule. The CD3? chain is thought to join the TCR and other CD3 polypeptides in that last stage of assembly. Pulse-chase experiments have shown that all six chains are assembled in the endoplasmic reticulum, transported to the Golgi apparatus, and then transferred to the plasma membrane. It also appears that the amount of ? chain is rate limiting, as it is synthesized at only 10% the level of the other chains. This results in the vast majority of newly synthesized a, ß or CD3 components being degraded within 4 hours of their synthesis. The remaining nondegraded chains are long-lived and form complete TCR/CD3 complexes with the limiting ? chain ( 74 ). TCR/CD3 complexes lacking CD3? chains migrate through the endoplasmic reticulum and Golgi apparatus intact but then are transported to the lysosomes and degraded. Analysis using transfectants of individual chains or pairs of chains has shown that CD3? and CD3d chains contain endoplasmic reticulum retention signals. If these signals are removed, the chains are transported through the Golgi apparatus and rapidly degraded in the lysosomes. The

immunological significance of this pre-Golgi degradation pathway is most evident in CD4 +CD8 + thymocytes, in which, despite high levels of synthesis of both messenger RNA and protein for all the TCR, CD3, and ? chains, surface expression is relatively low. The TCR chains in immature thymocytes seem to be selectively degraded ( 62 ). Thus, posttranslation regulation appears to be an important means of controlling the cell surface expression of TCR heterodimers. CD3 Structure It is very important, ultimately, to know the structure and dynamics of an entire TCR/CD3 complex. Although a number of TCR structures are now known, only recently have the first CD3 structures been solved ( 63 ). As expected, this CD3 e? heterodimer has immunoglobulin domains and yields some possible clues as to how it “fits” with a TCR heterodimer. It is hoped to be the harbinger of bigger pieces to this puzzle.

T-CELL RECEPTOR GENES As shown in Fig. 6, TCR gene segments are organized similarly to segments of immunoglobulins, and the same recombination machinery is responsible for joining separate V and D segments to particular J and C segments. This was initially indicated by the fact that the characteristic seven- and nine-nucleotide conserved sequences adjacent to the V, D, and J regions with the 12- or 23-nucleotide spacing between them, first described for immunoglobulin genes, are also present in TCRs ( 64 ). The most conclusive evidence of this common rearrangement mechanism has been shown by the fact that both a naturally occurring recombination-deficient mouse strain [severe combined immune deficiency ( 65 )] and mice engineered to lack recombinase activating genes (RAG) 1 ( 66 ) or 2 ( 67 ) are unable to rearrange either TCR or immunoglobulin gene segments properly. As with immunoglobulins, if the V region and J region gene segments are in the same transcriptional orientation, the intervening DNA is deleted during recombination. DNA circles of such material can be observed in the thymus ( 68 , 69 ), the principal site of TCR recombination (see later discussion). In the case of TCRß and TCRd, there is a single V region 3' to the C region in the opposite transcriptional orientation to J and C regions. Thus, rearrangement of these gene segments occurs through an inversion. Variable points of joining are seen along the V, D, and J gene segments, as are random nucleotide addition (N regions) in postnatal TCRs. The addition of several nucleotides in an inverted repeat pattern, referred to a P element insertion, at the V-J junction of the TCR? chains has also been observed ( 70 ).

FIG. 6. T-cell receptor gene organization in mice and humans. Schematic of V, D, J, and C elements of the T-cell receptor genes. Transcriptional orientation is from left to right except where noted. The overall size of each locus is indicated on the right side. E, enhancers; S, silencer elements.

Organization of the T-Cell Receptor a/d Locus In humans and in mice, there is a single a-chain C-region gene that is composed of four exons encoding (a) the constant region domain; (b) 16 amino acids, including the cysteine that forms the interchain disulfide bond; (c) the transmembrane and intracytoplasmic domains; and (d) the 3' untranslated region ( Fig. 6). The entire a/d locus in humans spans about 1.1 Mb ( 96 ). The murine a/d locus appears to be similar in size. There are 50 different J-region gene segments upstream of the C region in the murine locus. At least eight of the J-region gene elements are nonfunctional because of in-frame stop codons or rearrangement and splicing signals that are likely to be defective. A similar number of a-chain J regions are present in the human locus. This very large number of a-chain J regions, in comparison with the immunoglobulin loci, may indicate that the functional diversity contributed by the J segment of the TCR (which constitutes a major portion of the CDR3 loop) makes a special contribution to antigen recognition (see later discussion). Both the murine and human Cd, Jd, and two Dd gene segments are located between the Va and Ja gene segments. In the murine system, there are two Jd and two Dd gene segments on the 5' side of Cd, and the Cd gene is approximately 75 kb upstream of the Ca gene but only approximately 8 kb upstream of the most 5' known Ja gene segments. The human organization is similar, with three Da and two Jd gene segments. Surprisingly, in both species, all of the D elements can be used in one rearranged gene rather than alternating, as is the case with TCRß or heavy immunoglobulin; that is, Vd, D 1 , D 2 , and Jd rearrangements are frequently found in mice ( 71 ) and Vd, D 1 , D 2 , D 72 3 , and Jd are frequently found in humans ( ). This greatly increases the junctional or CDR3 diversity that is available, especially because of the potential for N-region addition in between each gene segment. This property makes TCRd the most diverse of any of the antigens receptors known, with approximately 10 12 to 10 13 different amino acid sequences in a relatively small (10- to 15—amino acid) region ( 71 ). The location of Dd, Jd, and Cd genes between Va and Ja gene segments raises the possibility that TCRd and TCRa could share the same pool of V gene segments. There is some overlap in V gene usage; however, in the murine system, four of the commonly used Vd genes (Vd1, Vd2, Vd4, and Vd5) are very different than known Va sequences and they have not been found to associate with Ca ( 73 ). The other four Vd gene families overlap with or are identical to Va subfamilies (Vd3, Vd6, Vd7, and Vd8 with

Va6, Va7, Va4, and Va11, respectively). The mechanisms that account for the preferential usage of certain gene segments to produce d versus a chain are not known. Although some Vd genes are located closer to the Dd and Jd fragments than Va genes (such as Vd1), other Vd genes (such as Vd6) are rarely deleted by Va-Ja rearrangements and thus seem likely to be located 5' of many Va gene segments. One of the Vd gene segments, Vd5, is located approximately 2.5 kb to the 3' of Cd in the opposite transcriptional orientation and rearranges by inversion. Despite its close proximity to Dd-Jd gene segments, Vd5 is not frequently found in fetal ?d T cells. Instead, the Vd5?DJd rearrangement predominates in adult ?d T cells. An implicit characteristic of the a/d gene locus is that a rearrangement of Va to Ja deletes the entire D-J-C core of the d-chain locus. In many aß T cells, the a-chain locus is rearranged on both chromosomes, and thus no TCRd could be made. In most cases, this results from Va?Ja rearrangement, but evidence suggesting an intermediate step in the deletion of TCRd has been reported ( 74 ). This involves rearrangements of an element termed TEA to a pseudo-Ja 3' of Cd. The rearrangement of TEA to this pseudo-Ja would eliminate the d-chain locus in aß T cells. Gene targeting of the TEA element has resulted in normal levels of aß and ?d T cells, but usage of the most Ja genes was severely restricted ( 75 ), which suggests that its function has more to do with governing the accessibility of the most 5' Ja genes for recombination. Organization of the T-Cell Receptor ß Locus The entire human 685-kb ß-chain gene locus was originally sequenced by Rowen et al. ( 76 ), and the organization is shown in Fig. 6. One interesting feature is the tandem nature of Jß-Cß in the TCRß locus. This arrangement is preserved in all higher vertebrate species that have been characterized thus far (mouse, human, chicken, and frog). The two Cß coding sequences are identical in the mouse and nearly so in humans and other species. Thus, it is unlikely that they represent two functionally distinct forms of Cß. However, the Jß clusters have relatively unique sequences, and this may thus be a mechanism for increasing the number of Jß gene segments. Together with the large number of Ja gene segments, there is far more combinatorial diversity (Ja × Jß = 50 × 12 = 600) provided by J regions in aß TCRs than in immunoglobulins. Most of the V regions are located upstream of the J and C regions and in the same transcriptional orientation as the D and J gene element, and they rearrange to Dß-Jß genes through deletion. As in the case of Vd5, a single Vß gene, Vß14, is located 3' to C regions and in the opposite transcriptional orientation; thus, rearrangements involving Vß14 occur through inversion. In the NZW strain of mouse, there is a deletion in the ß chain locus that spans from Cß1 up to and including the Jß2 cluster ( 77 ). In SJL, C57BR, and C57L mice, there is a large deletion ( 78 ) in the V-region locus from Vd5 to Vß9. These mice also express a V gene, Vß17, which is not expressed in other strains of mice. Deletion of about half of the V

genes (in SJL, C57BR, and C57L mice) does not seem to have any particular effect on the ability of these mice to mount immune responses whereas mice which have deleted the Jß2 cluster show impaired responses ( 79 ). Organization of the T-Cell Receptor ? Locus The organization of the murine and human ?-chain loci are shown in Fig. 6. The human ? genes span about 150 kb ( 29 ) and are organized in a manner similar to that of the ß chain locus with two J?-C? regions. An array of V? genes in which at least six of the V regions are pseudogenes (filled in) are located 5' to these J?-C? clusters, and each of the V genes are potentially capable of rearranging to any of the five J regions. The sequences of the two human C? regions are very similar overall and differ significantly only in the second exon. In C?2, this exon is duplicated two or three times, and the cysteine that forms in the interchain disulfide bond is absent. Thus, C?2-bearing human T cells have an extra large ? chain (55,000 MW) that is not disulfide-bonded to its d-chain partner. The organization of the murine ? chain genes is very different from that of the human genes in that there are three separate rearranging loci that span about 205 kb ( 117 ). Of four murine C? genes, C?3 is apparently a pseudogene in BALB/c mice, and the J?3-C?3 region is deleted in several mouse strains, including C57 Bl/10. C?1 and C?2 are very similar in coding sequences. The major differences between these two genes is in the five–amino acid deletion in the C?2 gene, which is located in the C II exon at the amino acid terminal of the cysteine residue used for the disulfide formation with the d chain. The C?4 gene differs significantly in sequences from the other C? genes (in 66% overall amino acid identity). In addition, the C?4 sequences contains a 17–amino acid insertion (in comparison with C?1) in the C II exon located at similar position as that of the five–amino acid deletion of the C?2 gene (G. Kershard and S. M. Hedrick, unpublished results). Each of the C? genes is associated with a single J? gene segment. The sequences of J?1 and J?2 are identical at the amino acid level, whereas J?4 differs from J?1 and J?2 at 9 of 19 amino acid residues. The murine V? genes usually rearrange to the J?-C? gene that is most proximal and in the same transcription orientation. Thus, V?1.1 rearranges to J?4; V?1.2 rearranges to J?2; and V?2, V?3, V?4, and V?5 rearrange to J?1. Interestingly, it appears that some V? genes are rearranged and expressed preferentially during ?d T cell ontogeny and in different adult tissues as well ( 80 ). Transcriptional Control of the T-Cell Receptor Genes Transcriptional regulation of the TCR genes has been studied extensively; enhancer sequences were first identified in the TCRß locus, 3' of Cß2 ( 81 , 82 ) and subsequently for the other TCR loci as well [reviewed by Lefranc and Lefranc ( 29 )], as indicated in Fig. 6. These TCR enhancers all share sequence similarities. Some of the transcriptional factors that bind to the TCR genes are also found to regulate

immunoglobulin gene expressions. Work by Sleckman et al. ( 83 ) has shown that the TCRa enhancer (Ea) is not only important for normal rearrangement and expression for the a chain locus but is also required for a normal expression level of mature TCRd transcripts. Also interesting is the work of Lauzurica and Krangel ( 84 , 85 ), who showed that a human TCRd enhancer containing mini-locus in transgenic mice is able to rearrange equally well in aß T cells as in ad T cells but that an Ea-containing construct was active only in aß-lineage T cells. Like immunoglobulin genes, promoter sequences are located 5' to the V gene segments. Although D?Jß rearrangement and transcription occur fairly often in B cells and in B-cell tumors ( 86 ), Vß rearrangement and transcription appear highly specific to T cells. In addition to enhancers, there are also “silencer” sequences 3' of Ca ( 87 , 88 ) and in the C?1 locus ( 89 ). It has been suggested that these “repressor sites” could turn off the expression of either of these genes, influencing T cell differentiation toward either the aß or the ?d T-cell lineage. Chromosomal Locations of T-Cell Receptor Genes and Translocations Associated with Disease The chromosomal locations of the different TCR loci have been delineated in both mice and humans, and the results are summarized in Table 1. One significant factor in cancers of hematopoietic cells are chromosomal translocations, which result in the activation of genes that are normally turned off or in the inactivation of genes that are normally turned on. Thus, B or T lymphocyte neoplasia is frequently associated with interchromosomal or intrachromosomal rearrangements of immunoglobulin or TCR loci and, in some cases, both ( 90 , 91 ).

TABLE 1. Chromosomal locations of T-cell receptor, immunoglobulin, and related loci in mouse and human

These translocations seemed to mediated by the V(D)J recombinase machinery, indicating the inherent danger and need for tight regulation of this pathway. Such rearrangements are particularly common in the a/d locus, perhaps because this locus

spans the longest developmental window in terms of gene expression, with TCRd being the first and TCRa the last gene to rearrange during T cell ontogeny (as discussed in more detail later). In addition, the a/d locus is in excess of 1 Mb in size, and this provides a larger target for rearrangement than does either TCRß or TCR?. Interestingly, in humans, TCRad is on the same chromosome as the heavy immunoglobulin locus, and VH?Ja rearrangements (by inversion) have been observed in some human tumor material ( 92 , 93 ). The functional significance of this is not known. Particularly frequent is the chromosome 8–14 translocation [t(8;14) (q24;q11)], which joins the a/d locus to the c-myc gene, analogous to the C-myc?heavy immunoglobulin translocation in many murine myeloma tumors and in Burkitt’s lymphomas in humans. In one cell line, a rearrangement occurs between the Ja-region coding sequences and in a region 3' of c-myc ( 94 ). In both B- and T-cell malignancies, the translocation of c-myc into heavy immunoglobulin or TCRa/ß appears to increase the expression of c-myc and may be a major factor in the unregulated cell growth that characterizes cancerous cells. Other putative proto-oncogenes that have been found translocated into the TCRa/ß locus are the LIM domain–containing transcription factors Ttg-1 ( 95 ) and Ttg-2 ( 96 , 97 ), which are involved in neural development; the helix-loop-helix proteins Lyl-1 ( 98 ) and Scl ( 99 ), which are involved in early hematopoietic development; and the homeobox gene Hox 11 ( 100 ), which is normally active in the liver. How these particular translocations contribute to malignancy is unknown, but they presumably causes aberrations in gene expression that contribute to cell growth or escape from normal regulation. In patients infected with the human T-cell leukemia virus type I, there are large numbers of similar translocations, and it is thought that this virus itself is not directly leukemogenic but acts by causing aberrant rearrangements in the T cells that it infects, some of which become malignant. Another disorder that exhibits frequent TCR and immunoglobulin locus translocations is ataxia telangiectasia, a autosomal recessive disorder characterized by ataxia, vascular telangiectasis, immunodeficiency, an increased incidence of neoplasia, and an increased sensitivity to ionizing radiation. Peripheral blood lymphocytes from patients with ataxia telangiectasia have an especially high frequency of translocations involving chromosomes 7 and 14 ( 101 ). These sites correspond to the TCR?, TCRß, and TCRa loci and to the immunoglobulin heavy-chain locus. Thus, it appears as though one of the characteristics of patients with ataxia telangiectasia is a relatively error-prone rearrangement process that indiscriminately recombines genes that have the TCR and immunoglobulin rearrangement signals ( 102 ). Allelic Exclusion In immunoglobulins, only one allele of the heavy chain locus and one of the light chain alleles are normally productively rearranged and expressed; this phenomenon is termed allelic exclusion (see Chapter 5). With regard to aß TCR expression, current data indicate that, although TCRß exhibits allelic exclusion ( 103 ), TCRa does not ( 104 , 105 ) and that some mature T cells express two functional TCRa chains. As the chances of forming an in-frame joint with any antigen receptor is only one in three, the probability that a T cell would have two productively rearranged TCRa genes is only 1/3 ×x 1/3

(1/9), or 11%. However, even when this happens, the two TCRa chains may not form heterodimers equally well with the single TCRß that is expressed, and thus only one heterodimer may be expressed. Data strongly suggest an important role for the pre-TCR heterodimer (e.g., pre-Ta:TCRß) in blocking further TCRß rearrangement and thus ensuring allelic exclusion at that locus ( 106 , 107 ). In particular, pre-Ta–deficient mice had a significant increase in the number of cells with two productive TCRß rearrangements, in comparison with wild-type mice ( 106 ). Commitment to the aß Lineage versus the ?d Lineage One important issue in T-cell development concerns the lineage relationship between aß and ?d T cells: What governs the differentiation of the thymic stem cells to become either aß or ?d T cells? Two models have been proposed. In one, which could be termed the sequential rearrangement model ( 24 ), the precursor cells first rearrange the ?- and d-chain genes. The cells that fail to made a functional TCR? or TCRd would progress to the aß lineage and attempt to rearrange the TCRß- and TCRa-chain loci. According to the second model, referred to as the separate lineage model, T cells differentiate into two lineages before rearrangement. One line of evidence that supports the sequential rearrangement model is a study in which d-chains were often found to be rearranged on chromosomes that undergo an a-chain rearrangement ( 108 ), but a subsequent more extensive investigation revealed most unrearranged sequences ( 109 ). Further evidence in favor of the separate lineage model comes from transgenic mice bearing rearranged TCR?- and TCRd-chain genes. In these mice, although all of the precursor cells express functional ?d genes, there are normal numbers of aß T cells in the thymus ( 110 ). This is the opposite of what would be expected if successful ?d TCR expression blocked the rearrangement of the a and ß loci. In another study of early aß precursor thymocytes, it was found that, in half the cells, TCRd had not rearranged at all but the TEA transcript was being expressed ( 111 ), presumably just before Va?Ja rearrangement. In mice that are defective in either aß TCR or ?d TCR, there is no obvious effect on the development of the remaining lineage ( 112 , 113 and 114 ). Taken together, almost all of the data in the literature supports a separate lineage model and not sequential rearrangement. Other Genetic Mechanisms One important mechanism of antibody diversification that has not been reproducibly found in TCR genes is somatic hypermutation. In antibodies, this form of mutation typically raises the affinities of antigen-specific immunoglobulins several orders of magnitude, typically from the micromolar range (10 -6 M) to the nanomolar range (10 -9 M) for protein antigens. It is now known that most cell surface receptors that bind ligands on other cell surfaces, including TCRs, typically have affinities in the micromolar range but that they compensate for this relatively low affinity by engaging multiple receptors simultaneously (e.g., increasing the valency) and by functioning in a confined, largely two-dimensional volume (e.g., between two cells). Cells employing such receptors most probably require weak (but highly specific) interactions so that they can

disengage quickly ( 115 , 116 ). The rapid “off” rate seen with TCRs has even been postulated to amplify the effects of small numbers of ligands (i.e. “the serial engagement model” described later). There has also been no enduring evidence for a naturally secreted form of either an aß or ?d TCR. Again, it can be argued that such a molecule would have no obvious use because it is too low in affinity to bind ligands efficiently. In the case of most TCRs, the concentration of protein would have to be very high to achieve an effect similar to soluble antibodies (in the milligram/milliliter range). A third mechanism seen in antibodies but not TCRs is CH switching, which allows different immunoglobulin isotypes to maintain a given V region specificity and associate it with different C regions that have different properties in solution (such as complement fixation and basophil binding). Because there is no secreted form of the TCR, it is not obvious how this would be useful.

BIOCHEMISTRY OF aß T-CELL RECEPTOR–LIGAND INTERACTIONS Although it has long been established that T cells recognize a peptide in association with an MHC molecule, a formal biochemical demonstration that this was caused by TCR binding to a peptide/MHC complex took many years to establish. Part of the difficulty in obtaining measurements of this type has been the intrinsically membrane-bound nature of MHC and TCR molecules. Another major problem is that the affinities are relatively low, in the micromolar range, which is too unstable to measure by conventional means. The problem of normally membrane-bound molecules can be circumvented by expressing soluble forms of TCR and MHC, which is also essential for structural studies (see previous discussion). For TCRs, there have been many successful strategies, including replacement of the transmembrane regions with signal sequences for glycolipid linkage ( 117 ), expression of chains without transmembrane regions in either insect or mammalian cells ( 36 , 118 ), or a combination of cysteine mutagenesis and E. coli expression ( 37 ). Unfortunately, no one method seems to work for all TCR heterodimers, although the combination of insect cell expression and leucine zippers at the C-terminal to stabilize heterodimer expression has been successful in many cases ( 119 ). The production of soluble forms of MHC molecule has a much longer history, starting with the enzymatic cleavage of detergent-solubilized native molecules ( 120 ), as well as some of the same methods employed for TCR such as glycophosphotidyl inositol linkage ( 121 ), E. coli expression and refolding ( 122 , 123 ), and insect cell expression of truncated (or leucine zippered) molecules ( 124 ). One interesting variant that seems necessary for the stable expression of some class II MHC molecules in insect cells has been the addition of a covalent peptide to the N-terminal of the ß chain ( 125 ). The first measurements of TCR affinities for peptide/MHC complexes were made by Matsui et al. ( 126 ) and Weber et al. ( 127 ). Matsui et al. used a high concentration of

soluble peptide/MHC to block the binding of a labeled anti-TCR Fab to T cells specific for those complexes, obtaining a binding constant (K d ) of approximately 50 µM for several different T cells and two different cytochrome peptide/IE K complexes (as shown in Table 2). Weber et al. used a soluble TCR to inhibit the recognition of a flu peptide/IE d complex by a T cell and obtained a K value of approximately 10 µM. Although these d measurements were an important start in TCR biochemistry, they gave no direct information about the kinetics of TCR-ligand interactions. Fortunately, the development of surface plasmon resonance instruments, particularly the BIAcore TM (Pharmacia Biosensor) with its remarkable sensitivity to weak macromolecular interactions ( 128 ), has allowed rapid progress in this area. In the BIAcore technique, one component is covalently cross-linked to a surface, and then buffer containing the ligand is passed in solution over it. The binding of even approximately 5% of the surface-bound material is sufficient to cause a detectable change in the resonance state of gold electrons on the surface. This method allows the direct measurement of association and dissociation rates—that is, kinetic parameters—and also has the advantage of being completely cell free. Figure 7 shows the type of resonance profile obtained contrasting the weak but specific binding of a particular peptide/MHC complex in solution to a bound TCR with the binding pattern of an antibody to the same TCR. The affinity of cytochrome c/IE K /2B4 TCR measured with this instrument ( 129 ) matches well ( Table 1) with previous results obtained from cell-based measurements. These and other data [reviewed by Davis et al. ( 116 )] showed definitively that TCR and peptide-loaded MHC molecules alone are able to interact and also that expression in a soluble form has not altered their ability to bind to each other. Because of its sensitivity and ease of use, the surface plasmon resonance technique has become the method of choice for measuring the kinetics of TCR binding to its ligands. As shown in Table 2, these measurements show that although the “on” rates of TCRs binding to peptide/MHC molecules vary from very slow (1,000 M per second) to moderately fast (200,000 M per second), their “off” rates fall in a relatively narrow range (0.5 to 0.01 second -1) or a t 1/2 of 12 to 30 seconds at 25°C. This is in the general range of other membrane-bound receptors that recognize membrane molecules on other cells ( 114 ), but it has also been noted that most TCRs have very slow “on” rates ( 130 ), which seems to reflect a flexibility in the binding site that might help to foster cross-reactivity (see later discussion). In the case of the class I MHC-restricted TCR, 2C, this relatively fast “off” rate may be stabilized (10-fold) if soluble CD8 is introduced ( 131 ), but this result is controversial ( 132 ). CD8 stabilization of TCR binding has been seen by Renard et al. ( 133 ) in their unique cell-based TCR labeling assay; however, no enhancement of TCR binding has been seen with soluble CD4 ( 134 ). Although most of the BIAcore measurements cited earlier were performed at 25°C because of instrument limitations, the “off” rates are likely to be much faster (10 to 20 times) at 37°C ( 135 ).

FIG. 7. T-cell receptor (TCR) binding to peptide/major histocompatibility complex (MHC). Top: A typical surface plasmon resonance analysis of the binding characteristics of a TCR specific for a cytochrome c peptide bound to the mouse class II MHC molecule, IE K. Here the soluble TCR is fixed to a solid support and different peptide/MHCs are passed over it in solution. The most robust profile represents the original peptide MCC (residues 88 to 103) complexes to IE K, a strong agonist, whereas T102s represents a weak agonist, K99A, a null peptide (see also Table 1 and Table 2). These profiles are compared to the bottom trace in the figure, which shows an antibody specific for Ca binding to the same TCR. Note the sharper initial phase, which is a measure of the association rate, and the very stable decay phase, which is a measure of the dissociation rate. The x-axis is the time in seconds, and the y-axis is in arbitrary resonance units. Figure courtesy of D. S. Lyons.

TABLE 2. T-cell receptor–ligand binding

To what extent can a T-cell response be predicted on the basis of the binding characteristic of its TCR to a ligand? One of the most intriguing discoveries concerning T-cell reactivity has been the phenomenon of altered peptide ligands. These are single–amino acid variants of antigenic peptides that change either the nature or the degree of the T-cell response (partial agonists) or prevent a response to a normally stimulating ligand (antagonists) ( 136 , 137 ). Discussions concerning the mechanism of

these “altered peptide” responses have focused on whether they are caused by some conformational phenomenon involving TCRs or CD3 molecules or both or caused by affinity or kinetic characteristics. The data now available indicate that most, but not all, T-cell responses correlate very well with the binding characteristics of their T-cell receptors. In particular, Sykulev et al. ( 138 ) first noted that higher affinity peptide variants elicited more robust T-cell responses. Subsequently, Matsui et al. ( 129 ) found that in a series of three agonist peptides, increasing dissociation rates correlated with decreasing agonist activity. Lyons et al. ( 139 ) found that this correlation extended to antagonist peptides in the same antigen system (moth cytochrome c/E k). They also showed that although an antagonist peptide might differ only slightly in affinity in comparison with the weakest agonist, its dissociation rate differed by 10-fold or more (see Table 2). This data in a class II MHC-restricted system is largely supported by the studies of Alam et al. ( 140 ) in a class I MHC system, who also saw a drop-off in affinities and an increase in “off” rates (with one exception, as noted in Table 2) with antagonist versus agonist ligands. In the cell-based TCR labeling system of Luescher, a survey of related peptide ligands of varying potency also revealed a general, but not absolute, correlation between receptor occupancy and stimulatory ability ( 141 ). Thus, although there is a general trend toward weaker T-cell responses and faster “off” rates and lower affinities, this does not seem to be an absolute rule, and thus other factors may be important in some cases. Alternatively, Holler et al. ( 142 ) suggested that some or all of the discrepancies may derive from differences in peptide stability (in the MHC) between the relatively short (minutes) time scale of BIAcore analysis at 25°C, in comparison with the much longer (days) cellular assays at 37°C. How might the relatively small differences in the binding characteristics of the ligands summarized in Table 2 and Table 3 cause such different T-cell signaling outcomes as agonism or antagonism? As McKeithan ( 143 ) and Rabinowitz et al. ( 144 ) noted, any multistep system such as T-cell recognition has an inherent ability to amplify small differences in signals that are received on the cell surface to much larger differences at the end of the pathway—in this case, gene transcription in the nucleus. Thus, antagonism may occur at one threshold and an agonist response at another. Alternatively, an antagonist ligand may traverse the activation pathway just far enough to use up some critical substrate, as proposed by Lyons et al. ( 139 ). Yet another possibility that has also been suggested is that some antagonists may act even earlier, by blocking TCR clustering at the cell surface ( 145 ).

TABLE 3. Weak agonist–antagonist binding

One controversy that bears on this data is the serial engagement model of Vallitutti et al. ( 146 ) and Viola and Lanzavecchia ( 147 ), which proposes that one way in which a small number of peptide/MHC complexes can initiate T-cell activation is by transiently binding many TCRs in a sequential manner. Estimates based on TCR down-regulation have suggested that one peptide/MHC complex could bind to as many as 200 TCR molecules in succession ( 147 ). Although the dissociation rates reviewed here show that TCR binding is likely to be very transient, they do not in fact, support the statement that more interactions are better. This is because, in most cases, improvements in TCR-peptide/MHC stability within any one system result in a more robust T cell response. This has been shown most spectacularly in the work of Holler et al. ( 142 ), who selected a nanomolar-affinity TCR from a mutagenized library expressed in yeast. With an approximately 100-fold slower “off” rate than the original, this TCR should have been only poorly stimulatory, according to the serial engagement model. Instead, T cells bearing it were considerably more sensitive to antigen. Role of CD4 and CD8 What is the role of CD4 and CD8 with regard to the T-cell response to agonist and antagonist peptides? In the case of a T helper cell response, the presence of CD4 greatly augments the amount of cytokine produced and, in some cases, determines whether there is a response at all [as reviewed by Janeway ( 148 )]. Much of the effect of CD4 seems to come from the recruitment of Lck to the TCR/CD3 complexes. In addition, there is a significant positive effect even with CD4 molecules that are unable to bind Lck, and thus there appears to be an affect on TCR-ligand interaction as well. Nonetheless, although a weak binding of CD4 to class II MHC has been observed ( 134 ), there is no apparent cooperativity with regard to TCR binding to peptide/MHC, in contrast to the case of CD8 and class I–specific TCRs (see later discussion). Together with the low-resolution structure of CD4–class II MHC ( 149 ), the classical model of CD4 binding to the same MHC as a TCR that it is associated with ( 148 ) seems untenable. However, there is abundant evidence that CD4 molecules do associate with TCRs, especially on previously activated T cells ( 150 ). Thus, models in which CD4 cross-linking to class II MHC indirectly supports TCR binding to peptide/MHCs and potentiates signaling through the delivery of Lck seem more likely (see later discussion). In addition, Irvine et al. ( 151 ), using a single-peptide labeling technique, showed an appreciable T-cell response to even one agonist peptide, resulting in a “stop” signal for the T cell and a small but detectable rise in intracellular calcium. Both of these effects are attenuated by antibody blockade of CD4, in such a way that many more (25 to 30) peptides are required in order to elicit a stop signal and a calcium flux. How could CD4 be facilitating the recognition of small numbers of peptides? Irvine et al. ( 151 ) proposed a “pseudodimer” model that suggests that a CD4 molecule associated with a TCR binding to an agonist peptide/MHC could bind laterally to an endogenous peptide/MHC that is also being bound by an adjacent TCR. This takes advantage of the apparent abundance of endogenous peptide/MHCs that can be bound by a given TCR ( 152 ) and uses two weak interactions (CD4? class II MHC and TCR?endogenous peptide/MHC) to

help create a dimeric “trigger” for activation. CD8 also greatly augments the response of class I MHC-specific T cells ( 148 ) and binds to class I MHC in much the same manner as CD4 ( 153 ). Overall, it seems likely that each of these co-receptor molecules has two roles: to stabilize TCR–ligand interactions physically and to aid in signaling by recruiting Lck. Consistent with this are data showing that CD4 can convert an antagonist peptide into a weak agonist ( 154 , 155 ), although CD4 has no apparent effect on antagonism ( 156 , 157 ). These results indicate that CD4 acts to augment T cell responses, even of very weak ligands, but that antagonism per se exerts its effects before CD4 engagement.

TOPOLOGY OF T-CELL RECEPTOR–PEPTIDE/MAJOR HISTOCOMPATIBILITY COMPLEX INTERACTIONS An analysis of TCR sequence diversity has shown that most amino acid variation resides in the region between the V- and J-region gene segments, which corresponds to the CDR3 regions of antibodies ( 158 ). This has led to models in which the CDR3 loops of Va and Vß make the principal contacts with the antigenic peptide bound to the MHC ( 158 , 159 and 160 ). Support for such a model has come from many studies that have shown that the CDR3 sequences of TCRs are important predictors of specificity [as reviewed by Davis and Bjorkman ( 158 )] as well as the elegant mutagenesis studies of Engel and Hedrick ( 161 ), who showed that a single CDR3 point mutation could alter the specificity of a TCR, and Katayama et al. ( 162 ), who showed also that a CDR3 “transplant” could confer the specificity of the donor TCR onto the recipient. In addition, a novel approach to TCR-ligand interactions was developed by Jorgensen et al. ( 163 , 164 ), who made single–amino acid changes in an antigenic peptide at positions that affect T-cell recognition but not MHC binding. These variant peptides are then used to immunize mice that express either the a or ß chain of a TCR that recognizes the original peptide, and the responding T cells are analyzed. Using these hemitransgenic mice allows the resulting T cells to keep half of the receptor constant while allowing considerable variation in the chain that pairs with it. The results from this study and from work in another system by Sant’Angelo et al. ( 165 ) are very similar in that every mutation at a TCR-sensitive residue triggered a change in the CD3 sequence of Va, Vß, or both and, in some cases, changed the Va or Vß gene segment as well (as summarized in Fig. 8). One of the more striking examples of a CDR3-peptide interaction occurred in the cytochrome c system, in which a Lys?Glu change in the central TCR determinant on the peptide triggered a Glu?Lys charge reversal in the Va CDR3 loop, which argues for a direct Lys?Glu contact between the two molecules ( 199 ).

FIG. 8. Sensitivity of T-cell receptor (TCR) complementarity-determining region 3 (CDR3) sequences and Va/Vß usage to changes in the antigen peptide. This figure summarizes the data of Jorgensen et al. ( 163 , 164 ) and Sant’Angelo et al. ( 165 ), who immunized single-chain transgenic mice (TCRa or TCRß) with antigenic peptides (MCC or CVA) altered at residues that influence T-cell recognition but not major histocompatibility complex binding. These data show that such changes invariably affect the CDR3 sequences of Va or Vß or both and that there appears to be a definite topology in which Va governs the N-terminal region and Vß seems more responsible for the c-terminal portion of the peptide.

Another interesting finding was the order of Va?Vß preference going from the N-terminal to the C-terminal residues of the peptides. This led Jorgensen et al. ( 163 , 164 ) to propose a “linear” topology of TCR-peptide/MHC interaction in which the CDR3 loops of Va and Vß line up directly over the peptide. Sant’Angelo et al. ( 165 ) proposed an orientation of the TCR in which the CDR3 loops are perpendicular to the peptide. This was based partially on intriguing data they found that suggested an interaction between the CDR1 of Va and an N-terminal residue of the peptide. A third orientation was proposed by Sun et al. ( 166 ) on the basis of the analysis of a large number of class I MHC mutants and their effect on TCR reactivity. This produced a roughly diagonal footprint of TCRs over the MHC, in comparison with the two previous models. On the other hand, an extensive class II MHC mutagenesis study failed to reveal a consistent “footprint” of TCR interaction and furthermore revealed that the pattern of TCR sensitivity was remarkably labile and highly dependent on sequences in the TCR CDR3 region or the peptide ( 167 ). This controversy has been largely resolved by the work of Garcia et al. ( 36 ) and Garboczi et al. ( 37 ), who, nearly simultaneously, solved the crystal structures of two different TCR-peptide/class I MHCs. These studies show a TCR binding surface much like an antibody fitting down between the two opposite “high points” of the class I MHC a helices, in a roughly diagonal configuration. In these structures, one of which is shown in Colorplate 2, the CDR3 loops are centrally located over the peptide, but the Va CDR1 and the Vß CDR1 are also in a position to contact the N-terminal and C-terminal peptide residues, respectively. Such a contact between Va CDR1 and an N-terminal residue was seen in the structure of Garboczi et al., whereas that of Garcia et al. has insufficient resolution at this point. There are now many additional structures including two involving class II MHCs, all of which exhibit a similar orientation, albeit with an approximately 20° variation in orientation ( 168 , 169 ). This oriented recognition constitutes a major departure from antibody–antigen interactions and may reflect a need to accommodate other molecules into a particular configuration that is optimal for signaling.

T-Cell Receptor Plasticity As aß T-cell receptor heterodimers are first selected in the thymus for reactivity to self-peptides bound to MHC molecules (see Chapter 9), all foreign peptide–reactive TCRs could be considered to be inherently cross-reactive. Indeed, a number of T cells have reactivity to very different peptide sequences, as shown by Nanda et al. ( 170 ). It has also been argued by Mason ( 171 ) that the universe of peptides is so large that each T cell must, on average, be cross-reactive to approximately 10 6 different peptides (although many of the differences in peptide sequence in this calculation would not be accessible to the TCR, being buried in the MHC binding groove). A large-scale screen of a random nonamer-peptide library with different T cells does turn up a great many stimulatory peptides, most with nonaccessible residues, but some with significantly different sequences, so that, in some cases, peptides with completely different sequences can activate the same T cell ( 172 ). Analyses of a T-cell hybrid that could recognize either a lysine or a glutamic acid residue in the center of a cytochrome c peptide on a panel of MHC mutants revealed that a different MHC “footprint” was evident, depending on which peptide was recognized ( 167 , 173 ) (as shown in Fig. 9). This suggests a plasticity of TCR binding to particular peptide/MHC complexes. More direct evidence of TCR plasticity was obtained by Garcia et al. ( 174 ), who, in comparing the x-ray crystal structures of the same TCR bound to two different peptide/MHC ligands, found a large conformational change in the CDR3 loop and a smaller one in the CDR1a loop. An even larger onformational change (13 Å) has been found in the CDR3ß residue of another TCR as it binds to a peptide/MHC complex ( 175 ). That each TCR may have many different conformations of its CDR3 loops is suggested by the two-dimensional nuclear magnetic resonance studies of Hare et al. ( 176 ) ( Fig. 10), who found that the CDR3 regions of a TCR in solution were significantly more mobile than the rest of the structure. That this may be a general feature of most TCRs is supported by thermodynamic analyses of various TCRs binding to their peptide/MHC ligands, both class I and class II. This binding is invariably accompanied by a substantial loss of entropy ( 130 , 177 ) and, at least in some cases, an “induced fit” mechanism ( 178 ). This seems to be a situation in which an inherently flexible binding site achieves greater order upon binding. This mechanism is also employed by DNA recognition proteins; Boniface et al. ( 178 ) suggested that it might represent a common mechanism of “scanning” an array of very similar molecular structures (MHCs or DNA) rapidly for the few that “fit” properly. As mentioned previously, the association rates are remarkably slow, in the range of 1,000 to 10,000 M per second ( Table 2). This indicates either that a multistep process is occurring before stable binding can be achieved or that only a fraction of the TCRs in solution have the correct conformation. Just how such a scanning mechanism might work for TCRs has been shown by Wu and colleagues ( 179 ), who found that a cytochrome c/class II MHC-specific TCR derived most of its stability of binding, but very little of its initial activation energy, from antigenic peptide residues. In contrast, MHC residues contributed by far the most to the initial binding but had relatively modest effects on stability. This indicates that “scanning” may be a process as shown in Fig. 11, first involving contact with (and orientation by) the a helices of the MHC and then a “fitting” process with and stabilization by peptide residues that involves a substantial loss of entropy. This model of TCR binding may help explain the striking

efficiency and sensitivity of T-cell recognition with the MHC helices guiding the TCR into the correct orientation. It may also be the structural basis for cross-reactivity with structurally very different peptides binding to the same MHC, inasmuch as the CDR3 regions of TCR could “fold” into the peptide in many possible configurations.

FIG. 9. A ?d T-cell receptor does not recognize the same epitope as aß T-cell receptors. Shown here are the effects of a panel of mutation located on the a helices of the IE K molecule on T-cell recognition. Inhibition of recognition is denoted by a filled circle. The one ?d T cell is this survey, LBK5, does not recognize a part of the central peptide-binding groove. This is also consistent with its indifference to what peptides occupy this site (see text). From ( 244 ), with permission.

FIG. 10. T-cell receptor complementarity-determining region 3 (CDR3) loops are more mobile than other CDRs in the binding site. Two-dimensional nuclear magnetic resonance studies of a murine T-cell receptor by Reinherz and Wüthrich and colleagues ( 176 ) show greater mobility in the central CDR loop (CDR3a and CDR3ß) than in the outer loops (CDR1 and CDR2 of TCRa and TCRß).

FIG. 11. As shown by Wu et al. ( 179 ), mutational analysis of T-cell receptor (TCR)–peptide/major histocompatibility complex (MHC) binding indicates that the TCR first contacts MHC residues (in the transition state), and the peptide has very little influence. Subsequently, however, the peptide residues contribute greatly to the stability of the complex. Thus, we have proposed that the transition state largely involves TCR-MHC contact followed by stabilization of mobile complementarity-determining region 3 residues into a stable state, usually involving significant conformational change and loss of entropy.

aß T-CELL RECEPTOR AND SUPERANTIGENS One of the most interesting and unexpected discoveries to emerge from the study of aß T-cell reactivities is the that of superantigens. Whereas a particular antigenic peptide might be recognized by only 1 or fewer in 100,000 T cells in a naive organism, a given superantigen might stimulate 1% to 20% of the T cells ( 180 , 181 , 182 and 183 ). As discussed in more detailed later, the physical basis for this is that the superantigen binds to a Vß domain of the TCR on T cells while simultaneously binding to a class II MHC molecule on an antigen-presenting cell (although not in the peptide-binding groove). This allows a single superantigen, such as SEA in Table 4, to stimulate virtually every murine T-cell–bearing Vß 1, 3, 10, 11, 12, or 17 (˜15% of all aß T cells), in most cases regardless of what Va it is paired with or what CDR3 sequence is expressed. This is clearly a unique class of T-cell stimulatory molecule.

TABLE 4. Vß specificity of exogenous and endogenous superantigens

The first indication of a superantigen effect was the discovery of minor lymphocyte stimulating determinants by Festenstein ( 184 ) in the early 1970s. Many years later, Kappler et al. ( 185 ) characterized a mouse strain–specific deletion of T cells expressing a specific TCR Vßs that were attributable to these loci. It emerged that these effects were caused by endogenous retroviruses of the murine mammary tumor virus (MMTV) family ( 186 , 187 , 188 , 189 and 190 ). Different family members bind different TCR Vß domains (as shown in Table 4) and stimulate T cells expressing them. Meanwhile, Janeway et al. ( 191 ) showed that Staphylococcus enterotoxins could polyclonally activate naive T cells in a Vß-specific manner without a requirement for antigen processing. Many of these enterotoxins have been characterized extensively ( 180 , 182 , 183 ). Unlike the MMTV proteins, which are a type II membrane protein, the enterotoxins are secreted. Subsequently, proteins having similar properties have been isolated from other bacteria, such as Yersinia pseudotuberculosis and Y. enterocolitica ( 192 , 193 ) and Streptococcus ( 194 ), and from Mycoplasma ( 195 , 196 ). There is also evidence of superantigen-like activities in other mammalian viruses such as rabies ( 197 ), cytomegalovirus ( 198 ), herpes virus ( 199 ), and Epstein-Barr virus ( 200 ) and also in Toxoplasma gondii ( 201 ). Because so many pathogenic or parasitic organisms possess these molecules, apparently by convergent evolution, there must be some selective advantage, but in most cases, there is no conclusive evidence as to what this might be. The one exception is the case of the MMTV superantigens, in which it has been shown that polyclonal T cell stimulation allows the virus to much more efficiently infect the B lymphocytes that are activated by the T cells ( 202 , 203 ). This may be a special case, however, and most authors writing on this subject have suggested that superantigens primarily serve to confuse and occupy the immune system while the pathogen escapes specific targeting and elimination. Large doses of superantigens have also been implicated in various “shock” syndromes, such as food poisoning or toxic shock ( 180 ), but this is probably not their everyday purpose, because it would violate the general rule that the host and parasite should coexist. It has also been suggested that superantigens may be involved in triggering autoimmune diseases. The hypothesis is that a large number of some Vß-bearing T cells are activated by a pathogenic superantigen and that subsequently self-reactive T cells within those activated cells are more easily stimulated by a particular tissue antigen. That this may occur in some cases is supported by the work of Stauffer et al. ( 204 ) on a human endogenous retrovirus which specifically stimulates Vß7 T cells and is

implicated in the initiation of type I diabetes. Another report implicates a superantigen in Crohn’s disease, another autoimmune disorder ( 205 ). Although the biochemistry of superantigen binding to TCR and MHC is similar to that of TCR peptide/MHC interactions ( 206 ), mutagenesis data and, in particular, x-ray structural data have shown that the topologies are both quite different and variable ( 207 ). In particular, it has been found that Mls-la presentation to T cells is most affected by mutations on the “outside” surface of the Vß domain that do not affect peptide/MHC recognition ( 207 ). In contrast, CDR1 and CDR2 of regions of Vß chains are involved in bacterial superantigen reactivity. An example of these data is shown in Colorplate 3, which shows how a model TCR-superantigen-MHC complex (derived from separate structures) would displace the TCR somewhat (but not entirely) away from the MHC binding groove ( 208 ), thus making the interaction largely insensitive to the TCR-peptide specificity. Other TCR-superantigen-MHC complexes have very different geometries ( 209 , 210 , 211 and 212 ). Why do all the many independently derived superantigens interact with only the TCRß-chain? One possibility is that the ß-chain offers the only access to the TCR, perhaps because the CD4 molecules hinders access to the Va side, as suggested by the antibody blocking studies of Rojo and Janeway ( 213 ).

A SECOND TYPE OF RECEPTOR: ?d-CD3 Identification of ?d T Cells Although aß T cells were originally defined on the basis of functional characteristics, such as providing T cell “help” or initiating cytotoxicity, ?d TCR–bearing cells were not discovered through any cellular assay or by serological analysis but instead were identified through gene cloning. Thus, most work on these cells has been devoted to the understanding of what they recognize and how they function within the immune system. Although there has been substantial progress, these questions are still largely unresolved. We review here some of the salient characteristics of these enigmatic cells. In the mouse, ?d T cells first appear in the fetal thymus fully 2 days before aß T cells, but in later weeks, aß T cells quickly predominate. In both mouse and human adults, ?d T cells represent only a small fraction (1% to 5%) of thymocytes ( 214 , 215 ) and lymphocytes in all of the secondary lymphoid organs. However, they are found in larger numbers in the mucus membranes of a variety of tissues such as the skin ( 216 ), small intestine ( 217 ), female reproductive tract ( 218 ), and lung ( 219 ). One population of ?d T cells that has been studied intensively are the CD4 -CD8 - ?dT lymphocytes, which have a dendritic structure and are embedded in the epidermis ( 216 , 217 ). These cells have been termed dendritic epidermal cells (DECs). Curiously, 90% of

these cells express a TCR with identical V? and Vd sequences ( 217 ). It has been shown that most DECs arise during days 15 to 17 of fetal life ( 119 ). At this stage in development, there is a preference for V?3 rearrangement, and little or no terminal deoxynucleotidyl transferase is expressed, and N-region diversity is consequently absent. In addition, the mechanism of gene rearrangement has been shown to be biased by nucleotide hom*ologies between the end of the V region and the beginning of (in this case) the J region ( 221 ). Thus, there may be a limited repertoire of ?d sequences at this stage, but the presence of so many identical ones so reproducibly indicates there is either some additional recombinational mechanism other than those cited or a strong selection for this particular outcome. As to what these DEC cells “see,” experiments have shown that they can respond to mouse keratinocytes or to an extract of keratinocytes added directly to the DECs ( 222 ). The nature of the determinant recognized is currently unknown. Other intraepithelial lymphocytes (IELs) show distinct receptor expression as well. The ?d T cells found in the female reproductive epithelia and tongues of mice preferentially express V?4 and Vd1 ( 218 ). In the BALB/c strain of mice, most of the TCRd sequences are the same ( 223 ), but others are diverse, and this phenomenon has not been seen in other strains. Another population of ?d T cells that has been studied extensively is resident in the epithelium of the small intestine ( 217 ). The gut IELs consist of a population of aß T cells and a population of ?d T cells. They are phenotypically CD4 -CD8 - or CD4 -CD8 +. Unlike CD8 + aß T cells, the CD8 molecules on ?d IELs contain a chains and no ß chains ( 224 , 225 ). IEL?d TCRs use different V? and Vd chains, and the CDR3 regions of both the ? and d chains show significant diversity both in length and sequence, which suggests that they can “see” a wide variety of ligands. How does this correlation between ?d TCR expression and anatomically different epithelia reflect an immune function? Is it the result of a unique homing process, or does it reveal some aspect of ontogeny? No concrete answers to these questions are yet at hand. ?d T Cells Contribute to Host Immune Defense Differently than aß T Cells Earlier studies showed that ?d T cells can secrete a variety of lymphokines and mount cytolytic responses and therefore have the potential to function like aß T cells. Their preferential localization in the epithelium also suggested that they may be responsible for a first line of defense [reviewed by Allison and Havran ( 226 )]. This hypothesis is supported by the increase of ?d T lymphocytes occurring early in infections by some bacteria and a virulent Sendai virus strain, before aß T cell responses are observed ( 227 , 228 ). However, in other infection models, ?d T cells accumulate within the inflammatory lesions late in the infection after the virus have been cleared [reviewed by Kaufmann ( 229 )], which suggests that they may be responding to cells that are damaged or stressed by the infection. Consistent with this is the demonstration that some ?d T cells can kill virus infected cells in vitro but that the recognition is not virus specific ( 230 ).

In addition, mice with deficiencies of aß or ?d T cells have been used to dissect the role of these cells in the immune defense against intracellular pathogens (bacteria, protozoa, and viruses) ( 231 , 232 and 233 ). These T-cell deficiencies were induced by either the administration of a monoclonal antibody against aß or ?d T lymphocytes or by disruption of a TCR gene through hom*ologous recombination. It was found that the effect of a ?d T cell deficiency differs, depending on the type of infection. In case of bacille Calmette-Guérin or Salmonella administration, aß but not ?d T cells are essential in controlling the infection. In other cases, such as Mycobacterium tuberculosis and Listeria monocytogenes, ?d T cells are able to compensate for the absence of aß T cells. Interestingly, in L. monocytogenes and Hartmannella vermiformis infections, a lack of ?d T cells does not change the pathogen load but instead results in a different pathological process in the infected tissue ( 231 , 232 , 233 and 234 ). This has led to the suggestion that ?d T cells may somehow regulate immune and nonimmune cells to maintain host tissue integrity ( 235 ). This possibility is supported by data showing that certain ?d T cells can produce keratinocyte growth factor and chemokines ( 236 ), as well as regulate the development of epithelial cells ( 237 ) and influence aß T cell responses ( 238 239 240 , , and 241 ). It is also compatible with the analysis of ?d T cell recognition requirements in that these cells can mediate cellular immune functions without a requirement for antigen processing and specialized antigen-presenting cells [reviewed by Hein and Mackay ( 6 )]. Therefore, they have the capacity to initiate immune responses by recognizing other lymphoid cells or damaged tissue cells directly. To gain insight into the scope of ?d IEL responses, the gene expression profiles of ?d IELs were surveyed with DNA microarrays (Affymetrix) ( 242 ) and the serial analysis of gene expression ( 243 ). These data suggest that ?d IELs may modulate local immune responses and participate in the intestinal metabolism and physiology by using mechanisms not previously appreciated. More strikingly, the transcription profiles show that whereas lymph node CD8 + aß T cells must be activated to become cytotoxic effectors, ?d IELs are constitutively transcribing genes associated with activation and effector functions. In particular, even in uninfected animals, ?d IELs constitutively express very high levels of granzyme A and B transcripts as well as natural killer cell–activating and inhibitory receptors. Thus, a cytolytic program could be readily turned on with little or no de novo transcription. An important implication is that the lytic activity of ?d IELs may be induced without a requirement for TCR ligand recognition. This would allow ?d IELs to deal with a broad range of pathological situations quickly, despite the diversity of the ?d TCRs expressed by these cells. The expression of the T-cell receptor may give IELs an alternative route to induce cytotoxicity, such as by recognizing pathogens directly or by utilizing additional or different sets of effector programs, depending on the method of target recognition. Although all these experiments point to an unique role for ?d T cell in the immune system, ?d T cell specificity and their exact effector functions in any pathological situation remains undefined. It is interesting to note that the function of ?d T cells has been studied mainly in mouse and human, but they are significantly more abundant in birds and artiodactyls ( 214 , 226 ). Thus, ?d T cells in these species may encompass other functions as well. Antigen Recognition by ?d T Cells Does Not Require Processing

Since 1994, a number of studies have shown that ?d T cells have profound differences in their antigen recognition requirements in comparison with aß T cells. Some ?d T cells also seem to recognize an entirely different types of antigens. More specifically, these experiments suggest that the antigens recognized by many ?d T cells do not have to be processed and presented and that they also do not have to be proteins [as reviewed by Hein and Mackay ( 6 )]. Because most aß T cells recognize protein antigens processed inside the cell and presented by MHC molecules, it was originally assumed that ?d T cells follow the same general pattern. Despite early work showing that classical MHC molecules are not involved in antigen recognition by ?d T cells, it was assumed that nonclassical MHC molecules, heat shock proteins, or yet-unidentified surface proteins may play a similar role. To date, the recognition requirements for ?d T cells have been evaluated in three model systems that allow a precise interpretation of the results. They are the recognition of the mouse class II MHC molecules IE K by the T cell clone LBK5 ( 244 ); the recognition of the murine nonclassical class I MHC molecules T10 and the closely related T22 molecule (94% identity) by the T cell clone G8 ( 244 , 245 ); and the recognition of a herpes simplex virus glycoprotein, gI, by the T cell clone TgI4.4 ( 246 ). The IE K encoded protein has been shown to bind peptides, whereas both biochemical ( 247 ) and structural studies ( 248 ) have shown that T10 and T22 do not. Furthermore, all three proteins have the potential to be degraded into peptides and “presented” for recognition. Strikingly, in all three cases, neither peptides bound to these proteins nor peptides derived from them are recognized by the ?d T cell clones. Instead, protein antigens are recognized directly without any requirement for antigen processing. An example of these data is shown in Table 5, which shows the effect of temperature-sensitive endocytic compartment mutants on aß T cell recognition of a protein antigen versus the recognition of IE K by LBK5 ( 244 ). Note that the endosomal mutants disrupt processing of cytochrome c but have no effect on ?d T cell recognition. In addition, epitope mapping with mutant IE molecules shows that amino acid residues in the a helices of the IEa and IEß chains that affect aß T cell recognition do not affect LBK5 stimulation ( 244 ).

TABLE 5. Effect of temperature-sensitive endocytic compartment mutants on aß T-cell recognition of antigen versus recognition of IE k by LBK-5

Research on LBK5 recognition ( 250 ) has also shown a remarkable sensitivity to changes in N-linked glycosylation of the IE K molecule. This is despite the fact that E. coli expressed (e.g., unglycosylated molecules) can be recognized. Because cells that are stressed, infected, or transformed often change the posttranslational modifications of their surface proteins, these findings suggest a way to regulate a ?d T cell response by qualitative changes of self antigens. T22 Tetramers Stain a Relatively Large Fraction of ?d T Cells Figure 12 shows the results of Crowley et al. ( 251 ), who used a T22 tetrameric straining reagent to show that a surprisingly large fraction (0.4% to 2.0%) of splenic ?d T cells could be stained. More than 90% of these cells are CD4 -CD8 -, whereas the rest are either CD4 or CD8 single positive (about 3% to 4% each). A similar frequency of tetramer positive ?d T cells was also found in the intestinal IEL population. This represents a much higher frequency of this particular ?d T cell specificity than is true of unimmunized aß T cells, which are in the range of 0.001 to 0.0001%. Also interesting is the finding that the T10 molecule is expressed at very low levels in the periphery (and T22 is not expressed at all) but is induced on activated cells (B and T lymphocytes, macrophages, and dendritic cells). This has led to the suggestion that T10/T22-specific ?d T cells could regulate these cells during an immune response ( 251 ). Whether other ?d T cell specificities occur in such large numbers is not known but seems very likely.

FIG. 12. T10/T22-specific ?d T cells can be detected in normal mice through use of a tetrameric T22 staining reagent. As shown by Crowley et al. ( 247 ), a T22 tetrameric flow cytometry staining reagent, which was generated by similar methods as tetrameric peptide/MHC reagents, stained approximately 0.6% of splenic ?d T cells in normal animals. More than 90% of these cells are CD4 -CD8 -; the rest are either CD4 or CD8 single positive (about 3% to 4% each). A similar frequency of tetramer-positive ?d T cells was also found in the intestinal intraepithelial lymphocyte (IEL) population (data not

shown). A human hom*ologue of T10/T22, MICA/MICB, was found to stimulate human ?d T cell lines derived from intestinal intraepithelial lymphocytes. Subsequent experiments demonstrated that MICA/MICB is a ligand for the natural killer cell activating receptor NKG2D ( 252 ). The reactivity of ?d T cell line to MICA/MICB-expressing cells is inhibited by antibodies to NKG2D. It has been proposed that MICA/MICB may also act as a ligand for the ?d TCR, because antibodies to the receptor also inhibit the reactivity. ?d T Cells Can Be Stimulated by Nonpeptide Antigens ?dT cells from healthy human peripheral blood and from patients with tuberculoid leprosy or rheumatoid arthritis respond to heat-killed mycobacteria. The major T-cell stimulatory components in the former are not the mycobacterial heat-shock proteins but instead have been identified to be phosphate-containing, nonpeptide molecules ( 253 , 254 , 255 , 256 , 257 and 258 ). Although the consensus is that phosphate is a necessary component, compounds identified from various laboratories with different mycobacteria-responsive clones appear to have distinctive structures ( Table 6). These nonphosphate moieties include unusual carbohydrate and phosphate groups; a 5'-triphosphorylated thymidine or uridine substituted at its ?-phosphate group by a yet-uncharacterized low-molecular-weight structure; isopentenyl pyrophosphate and related prenyl pyrophosphate derivatives; synthetic alkenyl and prenyl derivatives of phosphate; and pyrophosphate and ?-monoethyl derivatives of nucleoside and deoxynucleoside triphosphates ( 253 ). Although the relative biological importance of these compounds remains to be determined, it is clear that a major class of stimulants are phosphate-containing nonpeptides. It is also clear that multiple phosphate-containing compounds are able to stimulate different clones with different efficacy.

TABLE 6. Nonpeptide mycobacterial antigens that stimulate human V?9Vd2(V?2Vd2) T cells a An important finding is that all of these compounds can be found in both microbial and mammalian cells. Constant et al. ( 254 ) proposed that the mammalian TTP-X and UTP-X conjugate may be involved in a “salvage pathway” in DNA and RNA synthesis and thus could be involved in a metabolic pathway related to DNA or RNA synthesis such as cell proliferation. Such a molecule would fit with the “stress antigen” or “conserved primitive stimulus” expected for ?d T cell ligands ( 226 ). Tanaka et al. ( 255 , 256 ) proposed that a link in the recognition of both microbial pathogens and hematopoietic tumor cells by these ?d T cells is provided by the common set of prenyl pyrophosphate intermediates, isopentenyl, and related prenyl pyrophosphate derivatives. These compounds are present in normal mammalian cells as precursors in lipid metabolism for the synthesis of farnesyl pyrophosphate. In mammalian cells, farnesyl addition has been proposed to be a critical modification for the membrane association of the ras protein and is required for transforming activity. The observation that this ?d T-cell population accumulates in lesions caused by mycobacterial infections in humans ( 257 , 258 ) and is able to respond

to virally and bacterially infected cells suggests that these ?d cells respond to a class of antigens shared by a number of pathogens and transformed, damaged, or stressed cells. Other Antigen Specificities of ?d T Cells Even the very earliest studies of ?d T cell reactivities showed that classical MHC molecules are not the major ligands for these cells ( 214 ). Although some that can recognize either classical MHC or related molecules such as TL, Qa d , or CD1 have been found, the frequency of such clones derived from a mixed lymphocyte reaction is low (about 1 in 100,000), which is much lower than the frequency of aß alloreactive generated in such reactions (1 in 10 to 100). In many cases, these ?d T cells also show a broad cross-reactivity that is not seen for aß alloreactive T cells, which is consistent with the suggestion that there is a fundamental difference in their recognition properties. There are also two reports indicating that the ?d T-cell recognition may involve a “complexed antigen” on the cell surface: a ?d T cell hybridoma that responded to synthetic copolymer Glu-Tyr (GT) in the presence of stimulator cells expressing the Qa-1b (but not the Qa-1a) molecule ( 259 ). Also, a human ?d T cell clone from synovial fluid of a patient with early rheumatoid arthritis responding to a fragment C of tetanus toxin ( 260 ). The tetanus toxin response requires the presence of cells expressing a class II MHC molecule, DRw53, and can be inhibited by an anti-DRw53 antibody. In these two cases, it is not clear whether the Glu-Tyr copolymer and the tetanus toxin are “processed” and, if so, what kind of antigen processing is required. In addition to these specificities, ?d T cells that are responsive to mycobacterial 60-kD heat-shock protein and peptide derived from it ( 261 ), staphylococcal enterotoxin A ( 262 ), and an immunoglobulin light chain–derived peptide in the context of the heat-shock protein have also been reported ( 263 ). Complementarity-Determining Region 3 Length Distribution Analysis Shows ?d T-Cell Receptors Are More Immunoglobulin-like In an effort to find a molecular basis for these surprising differences in ?d versus aß T cell recognition, Rock et al. ( 264 ) characterized the length distribution of CDR3 regions in three immune receptor chains: immunoglobulin, aß TCR, and ?d TCR. Rock et al. found that the CDR3 lengths of both a and ß TCR polypeptides are nearly identical and have very constrained length distributions. In contrast, CDR3 lengths of immunoglobulin heavy chains are long and variable, whereas those of light chains are much shorter and more constrained. As discussed previously, the CDR3 loops of aß TCRs are critical for recognizing antigenic peptides bound to MHC molecules. The constraints on a and ß CDR3 length may reflect this functional requirement. Surprisingly, d-chain CDR3 lengths are long and variable, but those of ? TCR chains are much shorter and constrained. In this regard, ?d TCR and CDR3 length distributions are similar to those of immunoglobulins and distinct from those of aß TCR, as also indicated by the x-ray crystal structure of a ?d TCR ( 41 ).

It has been observed that the frequency of ?d T cell clones recognizing allogeneic MHC molecules in a mixed lymphocyte reaction is very low (in comparison with aß alloreactive clones) and that the majority of these clones show a high degree of cross-reactivity (only rarely seen with aß alloreactive clones) [reviewed by Hein and Mackay ( 6 )]. These observations are consistent with the proposal that ?d TCR recognition is more immunoglobulin-like, focusing on the common features shared by MHC molecules. It is noteworthy that the specificity of LBK5 (IE B and IE K but not IE D) is the same as two previously described anti-IE antibodies ( 265 , 266 ). Along these lines, human ?d T cell clones from healthy donors that respond to mycobacteria extract have been found to express V?9 and Vd2 with diverse junctional (CDR3) sequences ( 267 ). This is reminiscent of the immunoglobulin receptor usage in naturally occurring murine B cells that recognize phosphorylcholine. There, it was found that only very restricted immunoglobulin heavy-chain (VH11, VH12, or Q52) and light-chain V gene segments are used, coupled with variable CDR3 junctional sequences ( 267a). In the latter case, the restricted usage of the V genes may be more significant, inasmuch as several hundred to a thousand VH gene segments are available to mount an immunoglobulin response. The suggestion that ?d TCR recognition is more immunoglobulin-like does not preclude the possibility that some ?d T cells may recognize similar or identical ligands as aß T cells. It is clear, for example, that it is possible to make antibodies that are specific for different subtypes of MHC molecules or even particular peptide/MHC complexes ( 268 , 269 and 270 ). As discussed earlier, by considering all elements that contribute to the variability of the junctional (CDR3) region, such as the numbers of D and J elements used, D-element reading frame, junctional diversity, and N-region nucleotide addition, it was calculated that the number of possible CDR3 sequences is the greatest for ?d TCR, the least for immunoglobulin (irrespective of somatic mutation), and intermediate for aß TCR ( 158 ). This suggests that ?d T cells have the potential to recognize a wide variety of different antigens. Multivalence of the Ligands Is Required for Activation through the ?d T-Cell Receptor ?d TCR, as with to aß TCR, needs to associate with CD3 molecules for cell surface expression. Therefore, signaling through the antigen receptor may utilize a multivalent form of the antigen so that the engaged receptors can be cross-linked. Cell surface molecules can be recognized as such, but soluble antigen must be rendered polyvalent. A demonstration of this requirement is that in the three cases of ?d T cells recognizing cell surface molecules—IE K, T10/T22, and HSV gI protein—a soluble form of the protein can be recognized only when bound to plastic plates: for example, presented in a multivalent form ( 244 , 245 and 246 ). Interestingly, the stimulation of mycobacterial extract reactive ?dT-cell clones by small phosphate-containing compounds requires

cell–cell contact, and all cell types are able to induce the recognition ( 271 ,



This apparent requirement for multivalent antigens would then suggest that soluble antigens—such as the phosphate-containing compounds—must be associated with certain cell surface molecules for their recognition. It is important to know whether the binding and display of soluble antigens is achieved by a variety of different molecules on the surface or by a limited set of molecules and whether they normally form part of the epitope recognized by the antigen receptors. Although the recognition requirements just discussed are derived largely from observations with model systems, the identification of the Mycobacterium antigen clearly stems from a “physiologically relevant” event. It will be interesting to determine the generality of these rules in other systems, especially pathological ones. This should lead to a much better definition of the role or roles of ?d T cells. This includes the identification of what ?d T cells recognize and the consequence of such recognition in pathological situations. The issue of ?d T cell specificity is also important in understanding the development of these cells. Whereas some experiments with ?d TCR transgenic mice have suggested that they are both positively and negatively selected much the same way as aß T cells, others have shown an entirely different mode of selection ( 273 ). Interestingly, the phosphate-containing compounds isolated from mycobacterial extracts can be found both in pathogens and in mammalian cells. Thus, they are “self” as well as “nonself.” However, ?d T cells with this specificity seem not to have be eliminated from the normal repertoire.

COMPLEMENTARITY-DETERMINING REGION 3 DIVERSIFICATION: A GENERAL STRATEGY FOR T-CELL RECEPTORS AND IMMUNOGLOBULIN COMPLEMENTARITY TO ANTIGENS? One interesting observation that emerges from a detailed analysis of the gene rearrangements that create both TCRs and immunoglobulins is how the diversity of the CDR3 loop region in one or both of the chains in a given TCR is so much greater than that available to the other CDRs. A schematic of this skewing of diversity is shown in Fig. 13 for human immunoglobulins and for aß and ?d TCR heterodimers. In the case of aß TCRs, this concentration of diversity occurs in both Va and Vß CDR3 loops, and structural data ( 43 , 45 ) has confirmed that these loops sit largely over the center of the antigenic peptide (see previous section). Although this concentration of diversity in aß TCRs in the regions of principal contact with the many possible antigenic peptides seems reasonable, it is much harder to explain for immunoglobulin or ?d TCRs. Clearly, there must be some chemical or structural “logic” behind this phenomenon. A clue as to what this might be comes from the elegant studies of Cunningham and Wells ( 274 ) and Clackson and Wells ( 275 ), who systematically mutated all of the amino acids (to alanine) at the interface of human growth hormone and its receptor, as determined by x-ray crystallography. Interestingly, only a fourth of the approximately 30 mutations on either side had any effect on the binding affinity, even in cases in which the x-ray structural analysis showed that the amino acid side chains of most of the residues were “buried” in

the other. These studies illustrate an important caveat to the interpretation of protein crystal structures: Although they are invaluable for identifying which amino acids could be important in a given interaction, they do not indicate which ones are the most important. This is presumably because the “fit” at that many positions is not “exact” enough to add significant binding energy to the interaction. In this context, Davis et al. ( 276 ) proposed a new model in which the principal antigen specificity of an immunoglobulin or TCR is derived from its most diverse CDR3 loops. In the case of antibodies, we imagine that most of the specific contacts (and hence the free energy) with antigen are made by the VH CDR3 and that the other CDRs provide “opportunistic” contacts that make, in general, only minor contributions to the energy of binding and specificity. Once antigen has been encountered and clonal selection activates a particular cell B, somatic mutation would then “improve” the binding of the CDR1 and CDR2 regions to convert the typically low-affinity antibodies to the higher affinity models, as observed by Berek and Milstein ( 277 ) and also by Patten et al. ( 278 ). As a test of this model, Xu and Davis ( 279 ) analyzed mice that had a severely limited immunoglobulin V-region repertoire, consisting of one VH and effectively two VL chains (V?1 and V?2). These mice are able to respond to a wide variety of protein and haptenic antigens, even with this very limited complement of V regions. In several cases, hybridomas specific for very different antigens (e.g., ovalbumin vs. dinitrophenol) differ only in the V H CDR3. A limited V-region repertoire also seemed no barrier to deriving high-affinity antibodies with somatic mutation, inasmuch as repeated immunizations produced immunoglobulin G monoclonal antibodies with very high affinities (10 -9 to 10 -10 M). The major immune deficit in these mice was their ability to produce antibodies to carbohydrates, which may require a special type of binding site or specific V region. Thus, although these experiments involved only one VH, the results are highly suggestive about the inherent malleability of VHVl in general, at least with regard to protein and haptenic epitopes. With respect to aß TCRs, we expect that most of the energy of the interaction with a typical ligand resides in the CDR3-peptide contacts, and here again the CDR1 and CDR2 regions make less energetically important contacts. The case of ?d TCRs would be more like an antibody only without the affinity improvements that are gained from somatic hypermutation. We have only an ad hoc explanation for the extremes of diversity seen in the TCRd CDR3: It has to recognize both protein surfaces and small nonpeptidic molecules with a high degree of specificity. Perhaps the lack of somatic mutation forces it to provide more diversity in the initial repertoire.

FIG. 13. Diversity “map” of immunoglobulins and T-cell receptors, showing the

theoretical potential for sequence diversity in human antigen receptor molecules. N region addition is assumed to contribute 0 to 6 nucleotides to the junction of each gene segment, except for immunoglobulin K chains, in which this form of diversity is seldom utilized.

CONCLUSIONS Since TCR genes were first identified in the early 1980s, information about their genetics, biochemistry, structure, and function has accumulated to become almost a field unto itself. Despite this very real progress, many issues still remain unsolved: What do ?d T cells normally “see,” and what function do they serve? What do superantigens actually do during the course of a normal response, and how is this of benefit to the pathogen/parasite? What is the structural/chemical basis of TCR specificity? What sort of rearrangements or conformational charges occur in the TCR CD3 molecular ensemble upon ligand engagement? These and other questions should serve as a source of entertainment for many years to come.

ACKNOWLEDGMENTS We are very grateful to Dr. Stephen Hedrick for allowing us to build so freely on his excellent previous chapters for this volume. This work was supported by grants from the National Institutes of Health (to Mark M. Davis and Yueh-Hsiu Chien) and from the Howard Hughes Medical Institute (to Mark M. Davis). Color Plates

COLORPLATE 1. Complete T-cell receptor ab structure. A ribbon diagram of the first T-cell receptor ab heterodimer structure from Garcia et al. ( 36 ). In all domains, b strands are indicated by letters and the complementarity-determining regions 1, 2, 3, and 4 loops by numbers.

COLORPLATE 2. T-cell receptor (TCR)dpeptide/major histocompatibility complex (MHC) crystal structure of a TCR-peptide/MHC complex. Peptide and complementary-determining regions are portrayed in different colors. From ( 36 ), with permission.

COLORPLATE 3. Crystal structure of a T-cell receptor (TCR) b/superantigen (SAg) complex. Fields et al. ( 208 ) crystallized TCR-SAg complexes and from the structure of the same superantigens with a class II major histocompatibility complex (MHC) molecule and were able to deduce the relative spatial arrangement of the three molecules. This model suggests that TCR does not contact the MHC very strongly, which is consistent with the relative peptide insensitivity of SAg activation

REFERENCES 1. Zinkernagel RM, Doherty PC. Immunological surveillance against altered self components by sensitized T lymphocytes in lymphocytic choriomeningitis. Nature 1974;251:547. 2. Zinkernagel R., Doherty P. H-2 compatibility requirement for T cell–mediated lysis of target cells infected with lymphocytic choriomeningitis virus: Different cytotoxic T cell specificities are associated with structures from H-3K or H-2D. J Exp Med 1975;141:1427. 3. Shevach EM, Rosenthal AS. Function of macrophages in antigen recognition by guinea pig T lymphocytes. II. Role of the macrophage in the regulation of genetic control of the immune response. J Exp Med 1973;138:1213. 4. Katz DH, Hamaoka T, Benacerraf B. Cell interactions between histoincompatible T and B lymphocytes. II. Failure of physiological cooperative interactions between T and B lymphocytes from allogeneic donor strains in humoral response to hapten-protein conjugates. J Exp Med 1973;137:1405. 5. Terhorst C, Spits H, Stall F, et al. T lymphocyte signal transduction. In: Hames BD, Glover DM, eds. Molecular immunology, 2nd ed. Oxford, UK: IRL Press, 1996:132–188. 6. Hein WR, Mackay CR. Prominence of gamma delta T cells in the ruminant immune system. Immunol Today 1991;12:30. 7.

Chien Y-H, Jores R, Crowley MP. Recognition by ?/d T cells. Annu Rev Immunol 1996;14:511.

8. Infante AJ, Infante PD, Gillis S, et al. Definition of T cell idiotypes using anti-idiotype antisera produced by immunization with T cell clones. J Exp Med 1982;155:1100.

9. Allison JP, McIntyre BW, Bloch D. Tumor-specific antigen of murine T lymphoma defined with monoclonal antibody. J Immunol 1982;129:2293. 10. Meuer SC, Fitzgerald KA, Hussey RE, et al. Clonotypic structures involved in antigen specific human T cell function: relationship to the T3 molecular complex. J Exp Med 1983;157:705. 11. Haskins K, Kubo R, White J, et al. The major histocompatibility complex-restricted antigen receptor on T cells. I. Isolation with a monoclonal antibody. J Exp Med 1983;157:1149. 12. Kaye J, Procelli S, Tite J, et al. Both a monoclonal antibody and antisera specific for determinants unique to individual cloned helper T cell lines can substitute for antigen and antigen-presenting cells in the activation of T cells. J Exp Med 1983;158:836. 13. Samelson LE, Germain RN, Schwartz RH. Monoclonal antibodies against the antigen receptor on a cloned T cell hybrid. Proc Natl Acad Sci U S A 1983;80:6972. 14. MacIntyre BW, Allison JP. Biosynthesis and processing of murine T-cell antigen receptor. Cell 1984;38:654. 15. Kappler J, Kubo R, Haskins K, et al. The major histocompatibility complex–restricted antigen receptor on T cells in mouse and man. V. Identification of constant and variable peptides. Cell 1983;35:295. 16. Davis MM, Cohen DI, Nielsen EA, et al. The isolation of B and T cell–specific genes. In: Vitteta E, ed. UCLA symposia on molecular and cellular biology: B and T cell tumors, vol. 24, New York: Academic Press, 1982:215. 17. Hedrick SM, Cohen DI, Nielsen EA, et al. Isolation of cDNA clones encoding T cell–specific membrane-associated proteins. Nature 1984;308:149. 18. Hedrick SM, Nielsen EA, Kavaler J, et al. Sequence relationships between putative T-cell receptor polypeptides and immunoglobulins. Nature 1984;308:153. 19. Yanagi Y, Yoshikai Y, Leggett K, et al. A human T cell–specific cDNA clone encodes a protein having extensive hom*ology to immunoglobulin chains. Nature 1984;308:145. 20. Acuto O, Fabbi M, Smart J, et al. Purification and NH2-terminal amino acid sequencing of the beta subunit of a human T-cell antigen receptor. Proc Natl Acad Sci U S A 1984;81:3851. 21. Chien Y, Becker DM, Lindsten T, et al. A third type of murine T-cell receptor gene. Nature 1984;312:31. 22. Saito H, Kranz D, Takagaki Y, et al. A third rearranged and expressed gene in a clone of cytotoxic T lymphocytes. Nature 1984;312:36. 23. Saito H, Kranz DM, Takagaki Y, et al. Complete primary structure of a heterodimeric T-cell receptor deduced from cDNA sequences. Nature 1984;309:757. 24. Brenner MB, McLean J, Dialynas DP, et al. Identification of a putative second T-cell receptor. Nature 1986;322:145. 25. Chien YH, Iwashima M, Kaplan KB, et al. A new T-cell receptor gene located within the alpha locus and expressed early in T-cell differentiation. Nature 1987;327:677.

26. Dembic Z, Haas W, Weiss S, et al. Transfer of specificity by murine a and ß T-cell receptor genes. Nature 1986;320:232. 27. Saito T, Weiss A, Miller J, et al. Specific antigen-Ia activation of transfected human T cells expressing murine Ti ab-human T3 receptor complexes. Nature 1987;325:125. 28. Havran WL, Chien YH, Allison JP Recognition of self antigens by skin-derived T cells with invariant gamma delta antigen receptors. Science 1991;252:1430. 29.

Lefranc M-P, Lefranc G. The T cell receptor: FactsBook, London: Academic Press, 2001.

30. Brenner MB, McLean J, Scheft H, et al. Two forms of the T cell receptor gamma protein found on peripheral blood cytotoxic T lymphocytes. Nature 1987;325:689. 31. Groettrup M, Ungewiss K, Azogui O, et al. A novel disulfide-linked heterodimer on pre–T cells consists of the T cell receptor ß chain and a 33 kDa glycoprotein. Cell 1993;75:283. 32. Saint-Ruf C, Ungewiss K, Groettrup M, et al. Analysis and expression of a cloned pre–T cell receptor gene. Science 1994;206:1208. 33. Novotny J, Tonegawa S, Saito H, et al. Secondary, tertiary, and quaternary structure of T-cell–specific immunoglobulin-like polypeptide chains. Proc Natl Acad Sci U S A 1986;83:742. 34. Bentley GA, Boulot G, Karjalainen K, et al. Crystal structure of the beta chain of a T cell antigen receptor. Science 1995;267:1984. 35. Fields BA, Ober B, Malchiodi EL, et al. Crystal structure of the V alpha domain of a T cell antigen receptor. Science 1995;270:1821–1824. 36. Garcia KC, Decagon M, Stanfield RL, et al. The structure of an aß T-cell receptor at 2.5Å. Science 1996;274:209. 37. Garboczi DN, Ghosh P, Utz U, et al. Structure of the complex between human T-cell receptor, viral peptide and HLA-A2Å. Nature 1996;384:134. 38. Housset D, Mazza G, Gregoire C, et al. The three-dimensional structure of a T-cell antigen receptor V alpha V beta heterodimer reveals a novel arrangement of the V beta domain. EMBO J 1997;14:4205. 39. Reiser JB, Darnault C, Guimezanes A, et al. Crystal structure of a T cell receptor bound to an allogeneic MHC molecule. Nat Immunol 2002;1:291. 40. Hennecke J, Carfi A, Wiley DC. Structure of a covalently stabilized complex of a human aß T-cell receptor, influenza HA peptide and MHC class II molecule, HLA-DR1. EMBO J 2000;19:5611. 41. Allison TJ, Winter CC, Fournie JJ, et al. Structure of a human gammadelta T-cell antigen receptor. Nature 2001;411:820. 42. Strong RK, Penny DM, Feldman RM, et al. Engineering and expression of a secreted murine T cell receptor with reduced N-linked glycosylation. J Immunol 1994;153:4111. 43. Klausner RD, Lippincott-Schwartz J, Bonifacino JS. The T cell antigen receptor: insights into organelle biology. Annu Rev Cell Biol 1990;6:403. 44.

Samelson LE, Harford HB, Klausner RD. Identification of the components of the murine T cell

antigen receptor complex. Cell 1985;43:223. 45. Orloff DG, Frank SJ, Robey FA, et al. Biochemical characterization of the eta chain of the T-cell receptor. J Biol Chem 1989;264:14812. 46. Jin YJ, Clayton LK, Howard FD, et al. Molecular cloning of the CD3 eta subunit identifies a CD3 zeta–related product in thymus-derived cells. Proc Natl Acad Sci U S A 1990;87:3319. 47. Clayton LK, D’Adamio L, Howard FD, et al. CD3 eta and CD3 zeta are alternatively spliced products of a common genetic locus and are transcriptionally and/or post-transcriptionally regulated during T-cell development. Proc Natl Acad Sci U S A 1991;88:5202. 48. Rodewald HR, Arulanandam AR, Koyasu S, et al. The high affinity Fc epsilon receptor gamma subunit (Fc epsilon RI gamma) facilitates T cell receptor expression and antigen/major histocompatibility complex-driven signaling in the absence of CD3 zeta and CD3 eta. J Biol Chem 1991;266:15974. 49. Orloff DG, Ra C S, Frank SJ, et al. Family of disulphide-linked dimers containing the zeta and eta chains of the T-cell receptor and the gamma chain of Fc receptors. Nature 1990;347:189. 50. Koyasu S, D’Adamio L, Arulanandam AR, et al. T cell receptor complexes containing Fc epsilon RI gamma hom*odimers in lieu of CD3 zeta and CD3 eta components: a novel isoform expressed on large granular lymphocytes. J Exp Med 1992;175:203. 51. Exley M, Wileman T, Mueller B, et al. Evidence for multivalent structure of T cell antigen receptor complex. Mol Immunol 1995;32:829. 52. Blumberg RS, Ley S, Sancho J, et al. Structure of the T-cell receptor: Evidence for two CD3 epsilon subunits in the T-cell receptor–CD3 complex. Proc Natl Acad Sci U S A 1990;87:7220. 53. Jin YJ, Koyasu S, Moingeon P, et al. A fraction of CD3 epsilon subunits exists as disulfide-linked dimers in both human and murine T lymphocytes. J Biol Chem 1990;265:15850. 53a. Fernandez-Miguel G, Alarcón B, Iglesias A, et al. Multivalent structure of an aß T cell receptor. Science 1999;96:1547. 53b. Call ME, Pyrdol J, Wiedmann M, et al. The organizing principal in the formation of the T cell receptor-CD3 complex. Cell 2002;111:967. 54. Clevers H, Alarcón B, Wileman T, et al. The T cell receptor/CD3 complex: a dynamic protein ensemble. Annu Rev Immunol 1988;6:629. 55. Cosson P, Lankford SP, Bonifacino JS, et al. Membrane protein association by potential intramembrane charge pairs. Nature 1991;351:414. 56. Hall C, Berkhout B, Alarcón B, et al. Requirements for cell surface expression of the human TCR/CD3 complex in non–T cells. Int Immunol 1991;3:359. 57. Alarcón B, Ley SC, Sanchez-Madrid F, et al. The CD3-gamma and CD3-delta subunits of the T cell antigen receptor can be expressed within distinct functional TCR/CD3 complexes. EMBO J 1991;10:903. 58. Klausner RD, Samelson LE. T cell antigen receptor activation pathways: the tyrosine kinase connection. Cell 1991;64:875. 59. Cantrell DA, Davies AA, Crumpton MJ. Activators of protein kinase C down-regulate and phosphorylate the T3/T–cell antigen receptor complex of human T lymphocytes. Proc Natl Acad Sci U S A

1985;82:8158. 60. Ashwell JD, Klausner RD. Genetic and mutational analysis of the T-cell antigen receptor. Annu Rev Immunol 1990;8:139. 61. Minami Y, Weissman AM, Samelson LE, et al. Building a multichain receptor: synthesis, degradation and assembly of the T-cell antigen receptor. Proc Natl Acad Sci U S A 1987;84:2688. 62. Bonifacino JS, Suzuki CK, Klausner RD. A peptide sequence confers retention and degradation in the endoplasmic reticulum. Science, 1990;247:79. 63. Sun ZJ, Kim KS, Wagner G, et al. Mechanisms contributing to T cell receptor signaling and assembly revealed by the solution structure of an ectodomain fragment of the CD3e? heterodimer. Cell 2001;105:913. 64. Chien Y, Gascoigne NRJ, Kavaler J, et al. Somatic recombination in a murine T-cell receptor gene. Nature 1984;309:322. 65.

Bosma MJ. B and T cell leakiness in the SCID mouse mutant. Immunodefic Rev 1992;3:261.

66. Mombaerts P, Iacomini J, Johnson RS, et al. RAG-1–deficient mice have no mature B and T lymphocytes. Cell 1992;68:869. 67. Shinkai Y, Rathbun G, Lam KP, et al. RAG-2–deficient mice lack mature lymphocytes owing to inability to initiate V(D)J rearrangement. Cell 1992;68:855. 68. Fujimoto S, Yamagishi H. Isolation of an excision product of T-cell receptor alpha-chain gene rearrangements. Nature 1987;327:242. 69. Okazaki K, Davis DD, Sakano HT. Cell receptor beta gene sequences in the circular DNA of thymocyte nuclei: direct evidence for intramolecular DNA deletion in V-D-J joining. Cell 1987;49:477. 70. Heilig JS, Tonegawa S. Diversity of murine gamma genes and expression in fetal and adult T lymphocytes. Nature 1986;322:836. 71. Elliott JF, Rock EP, Patten PA, et al. The adult T-cell receptor d-chain is diverse and distinct from that of fetal thymocytes. Nature 1988;331:627. 72. Hata S, Satyanarayana K, Devlin P, et al. Extensive junctional diversity of rearranged human T cell receptor d genes. Science 1988;250:1541. 73. Raulet DH. The structure, function, and molecular genetics of the ?/d T cell receptor. Annu Rev Immunol 1989;7:175. 74. de Villartay JP, Lewis D, Hockett R, et al. Deletional rearrangement in the human T-cell receptor a-chain locus. Proc Natl Acad Sci U S A 1987;84:8608. 75. Villey I, Caillol D, Seiz F, et al. Defect in rearrangement of the most 5'TCR-Ja (TEA): implications for TCR a locus accessibility. Immunity 1996;5:331. 76. Rowen L, Koop BF, Hood L. The complete 685-kilobase DNA sequence of the human beta T cell receptor locus. Science 1996;272:1755. 77.

Noonan DJ, Kofler R, Singer PA, et al. Delineation of a defect in T cell receptor beta genes of NZW

mice predisposed to autoimmunity. J Exp Med 1986;163:644. 78. Chou HS, Nelson CA, Godambe SA, et al. Germline organization of the murine T cell receptor beta-chain genes. Science 1987;238:545. 79. Woodland DL, Kotzin B, Palmer E. Functional consequences of a T cell receptor Dß2 and Jß2 gene segment deletion. J Immunol 1990;14:379. 80. Havran WL, Allison JP. Developmentally ordered appearance of thymocytes expressing different T cell antigen receptors. Nature 1988;335:443. 81. McDougall S, Peterson CL, Calame K. A transcriptional enhancer 3' of C ß 2 in the T cell receptor ß locus. Science 1988;241:205. 82. Krimpenfort P, de Jong R, Uematsu Y, et al. Transcription of T cell receptor ß-chain genes is controlled by a down-stream regulatory element. EMBO J 1988;7:745. 83. Sleckman BP, Bardon CG, Ferrini R, et al. Function of the TCRa enhancer in aß and ?d T cells. Immunity 1997;7:505. 84. Lauzurica P, Krangel MS. Temporal and lineage-specific control of T cell receptor alpha/delta gene rearrangement by T cell receptor alpha and delta enhancers. J Exp Med 1994;179:1913. 85. Lauzurica P, Krangel MS. Enhancer-dependent and -independent steps in the rearrangement of a human T cell receptor delta transgene. J Exp Med 1994;179:43. 86. Waldmann TA, Davis MM, Bongiovanni KF, et al. Rearrangements of genes for the antigen receptor on T cells as markers of lineage and clonality in human lymphoid neoplasms. N Engl J Med 1985;313:776. 87. Winoto A, Baltimore D. aß lineage-specific expression of the T cell receptor a gene by nearby silencers. Cell 1989;59:649. 88. Diaz P, Cado D, Winoto A. A locus control region in the T cell receptor a/d locus. Immunity 1994;1:207. 89. Ishida I, Verbeek S, Bonneville M, et al. T-cell receptor ?d and ? transgenic mice suggest a role of a? transgenic mice suggest a role of a ? gene silencer in the generation of aß T cells. Proc Natl Acad Sci U S A 1990;87:3067. 90. Korsmeyer SJ. Chromosomal translocations in lymphoid malignancies reveal novel proto-oncogenes. Annu Rev Immunol 1992;10:785. 91. Finger LR, Harvey RC, Moore RC, et al. A common mechanism of chromosomal translocation in Tand B-cell neoplasia. Science 1986;234:982. 92. Baer R, Chen K-C, Smith SD, et al. Fusion of an immunoglobulin variable gene and a T cell receptor constant gene in the chromosome 14 inversion associated with T cell tumors. Cell 1986;43:705. 93. Denny CT, Yoshikai YU, Mak TW, et al. A chromosome 14 inversion in a T cell lymphoma is caused by site-specific recombination between immunoglobulin and T cell receptor loci. Nature 1986;320:549. 94. Boehm T, Greenberg JM, Buluwela L, et al. An unusual structure of a putative T cell oncogene which allows production of similar proteins from distinct mRNAs. EMBO J 1990;9:857.

95. McGuire EA, Hockett RD, Pollock KM, et al. The t(11;14)(p15;q11) in a T cell acute lymphoblastic leukemia cell line activates multiple transcripts, including Ttg-1, a gene encoding a potential zinc finger protein. Mol Cell Biol 1989;9:2124. 96. Royer-Pokora B, Loos U, Ludwig W-D. Ttg-2, a new gene encoding a cysteine-rich protein with the LIM motif, is overexpressed in acute T-cell leukaemia with the t(11;14)(p13;q11). Oncogene 1991;6:1887. 97. Boehm T, Foroni L, Kaneko Y, et al. The rhombotin family of cysteine-rich LIM-domain oncogenes: distinct members are involved in T-cell translocations to human chromosomes 11p15 and 11p13. Proc Natl Acad Sci U S A 1991;88:4367. 98. Visvader J, Begley CG, Adams JM. Differential expression of the LYL, SCL and E2A helix-loop-helix genes within the hemopoietic system. Oncogene 1991;6:187. 99. Finger LR, Kagan J, Christopher G, et al. Involvement of the TCL5 gene on human chromosome 1 in T-cell leukemia and melanoma. Proc Natl Acad Sci U S A 1989;86:5039. 100. Hatano M, Roberts CWM, Minden M, et al. Deregulation of a homeobox gene HOXII, by the t(10;14) in T cell leukemia. Science 1991;253:79. 101. O’Conner RD, Brown MG, Francke U. Immunologic and karyotypic studies in ataxia-telangiectasia: specificity of break points on chromosomes 7 and 14 in lymphocytes from patients and relatives. In: Bridges BA, Harnden DG, eds. Ataxia-telangiectasia—a cellular and molecular link between cancer, neuropathology and immune deficiency. New York: John Wiley and Sons, 1982;259. 102. Lipkowitz S, Stern MH, Kirsch IR. Hybrid T cell receptor genes formed by interlocus recombination in normal and ataxia-telangiectasis lymphocytes. J Exp Med 1990;172:409. 103. Uematus Y, Ryser S, Dembic Z, et al. In transgenic mice the introduced functional T cell receptor beta gene prevents expression of endogenous beta genes. Cell 1988;52:831. 104. Padovan E, Casorati G, Dellabona P, et al. Expression of two T cell receptor alpha chains: dual receptor T cells. Science 1993;262:422. 105. Malissen M, Trucy J, Jouvin-Marche E, et al. Regulation of TCR alpha and beta gene allelic exclusion during T-cell development. Immunol Today 1992;13:315. 106. Aifantis I, Buer J, von Boehmer H, et al. Essential role of the pre–T cell receptor in allelic exclusion of the T cell receptor ß locus. Immunity 1997;7:601. 107. O’Shea CC, Thornell AP, Rosewell IR, et al. Exit of the pre-TCR from the endoplasmic reticulum/cis-Golgi is necessary for signaling differentiation, proliferation, and allelic exclusion in immature thymocytes. Immunity 1997;7:591. 108. Takesh*ta S, Toda M, Yamagishi H. Excision products of T cell receptor gene support a progressive rearrangement model of the alpha/delta locus. EMBO J 1989;8:3261. 109. Winoto A, Baltimore D. Separate lineages of T cells expressing the alpha beta and gamma delta receptors. Nature 1989;338:430. 110. Dent AL., Matis LA, Hooshmand F, et al. Self-reactive gamma delta T cells are eliminated in the thymus. Nature 1990;343:714. 111.

Wilson A, de Villartay J-P, MacDonald HR. T cell receptor d gene rearrangement and T early a

(TEA) expression in immature aß lineage thymocytes: implications for aß/?d lineage commitment. Immunity 1996;4:37. 112. Mombaerts P, Clarke AR, Rudnicki MA, et al. Mutations in T-cell receptor genes a and ß block thymocyte development at different stages. Nature 1992;360:225. 113. Philpott KL, Viney JL, Kay G, et al. Lymphoid development in mice congenitally lacking T cell receptor aß-expressing cells. Science 1992;256:1448. 114. Itohara S, Mombaerts P, Lafaille J, et al. T cell receptor delta gene mutant mice: independent generation of alpha beta T cells and programmed rearrangements of gamma delta TCR genes. Cell 1993;72:337. 115. van der Merwe PA, Barclay AN. Transient intercellular adhesion: the importance of weak protein–protein interactions. Trends Biochem Sci 1994;19:354. 116. Davis MM, Boniface JJ, Reich Z, et al. Ligand recognition by aß T cell receptors. Annu Rev Immunol 1998;16:523. 117. Lin AY, Devaux B, Green A., et al. Expression of T cell antigen receptor heterodimers in a lipid-linked form. Science 1990;249:677. 118. Gregoire C, Lin SY, Mazza G, et al. Covalent assembly of a soluble T cell receptor–peptide–major histocompatibility class I complex. Proc Natl Acad Sci U S A, 1996;14:7184. 119. Scott CA, Garcia KC, Carbone FR, et al. Pairing for the production of functional soluble IA MHC class II molecules. J Exp Med 1996;183:2087. 120. Parham P, Alpert BN, Orr HT, et al. Carbohydrate moiety of HLA antigens. Antigenic properties and amino acid sequences around the site of glycosylation. J Biol Chem 1977;252:7555. 121. Wettstein DA, Boniface JJ, Reay PA, et al. Expression of a functional class II MHC heterodimer in a lipid-linked form with enhanced peptide/soluble MHC complex formation at low pH. J Exp Med 1991;174:219. 122. Garboczi DN, Hung DT, Wiley DC. HLA-A2–peptide complexes: refolding and crystallization of molecules expressed in Escherichia coli and complexed with single antigenic peptides. Proc Natl Acad Sci U S A 1992;89:3429. 123. Altman JD, Reay PA, Davis MM. Formation of functional class II MHC/peptide complexes from subunits produced in E. coli. Proc Natl Acad Sci U S A 1993;90:10330. 124. Jackson MR, Song ES, Yang Y, et al. Empty and peptide-containing conformers of class I major histocompatibility complex molecules expressed in Drosophila melanogaster cells. Proc Natl Acad Sci U S A 1992;89:12117. 125. Kozono H, White J, Clements J, et al. Production of soluble MHC class II proteins with covalently bound single peptides. Nature 1994;369:151. 126. Matsui K, Boniface JJ, Reay PA, et al. Low affinity interaction of peptide–MHC complexes with T cell receptor. Science 1991;254:1788. 127. Weber S, Traunecker A, Oliveri F, et al. Specific low-affinity recognition of major histocompatibility complex plus peptide by soluble T-cell receptor. Nature 1992;356:793.


Malmqvist M. Biospecific interaction analysis using biosensor technology. Nature 1993;361:186.

129. Matsui K, Boniface JJ, Steffner P, et al. Kinetics of T cell receptor binding to peptide-MHC complexes: correlation of the dissociation rate with T cell responsiveness. Proc Natl Acad Sci U S A 1994;91:12862. 130. Willcox BE, Gao GF, Wyer JR, et al. TCR binding to peptide-MHC stabilizes a flexible recognition interface. Immunity 1999;10:357. 131. Garcia K, Scott C, Brunmark A, et al. CD8 enhances formation of stable T-cell receptor/MHC class I molecule complexes. Nature 1996;384:577. 132. Wyer JR, Willcox BE, Gao GF, et al. T cell receptor and coreceptor CD8 alpha alpha bind peptide-MHC independently and with distinct kinetics. Immunity 1999;2:219. 133. Renard V, Romero P, Vivier E, et al. CD8 beta increases CD8 coreceptor function and participation in TCR-ligand binding. J Exp Med 1996;184:2439. 134. Xiong Y, Kern P, Chang H, et al. T cell receptor binding to a pMHCII ligand is kinetically distinct from and independent of CD4. J Biol Chem 2001;276:5659. 135. Boniface JJ, Reich Z, Lyons DS, et al. Thermodynamics of T cell receptor binding to peptide/MHC: evidence for a general mechanism of molecular scanning. Proc Natl Acad Sci U S A 1999;96:11446. 136.

Kersh GJ, Allen PM. Essential flexibility in the T-cell recognition of antigen. Nature 1996;380:495.


Jameson SC, Bevan MJ. T cell receptor antagonists and partial agonists. Immunity 1995;2:1.

138. Sykulev Y, Brunmark A, Jackson M, et al. Kinetics and affinity of reactions between an antigen-specific T cell receptor and peptide–MHC complexes. Immunity 1994;1:15. 139. Lyons DS, Lieberman SA, Hampl J, et al. T cell receptor binding to antagonist peptide/MHC complexes exhibits lower affinities and faster dissociation rates than to agonist ligands. Immunity 1996;5:53. 140. Alam SM, Travers PJ, Wung JL, et al. T-cell–receptor affinity and thymocyte positive selection. Nature 1996;381:616. 141. Kessler BM, Bassanini P, Cerottini J-C, et al. Effects of epitope modification on T cell receptor–ligand binding and antigen recognition by seven H-2K d-restricted cytotoxic T lymphocyte clones specific for a photoreactive peptide derivative. J Exp Med 1997;185:629. 142. Holler PD, Lim AR, Cho BK, et al. CD8 - T cell transfectants that express a high affinity T cell receptor exhibit enhanced peptide-dependent activation. J Exp Med 2001;194:1043. 143. McKeithan K. Kinetic proofreading in T-cell receptor signal transduction. Proc Natl Acad Sci U S A 1995;92:5042. 144. Rabinowitz JD, Beeson C, Lyons DS, et al. Kinetic discrimination in T cell activation. Proc Natl Acad Sci U S A 1997;93:1401. 145. Reich Z, Boniface JJ, Lyons DS, et al. Ligand-specific oligomerization of T-cell receptor molecules. Nature 1997;387:617.

146. Vallitutti S, Muller S, Cella M, et al. Serial triggering of many T-cell receptors by a few peptide–MHC complexes. Nature 1995;375:148. 147. Viola A, Lanzavecchia A. T cell activation determined by T cell receptor number and tunable thresholds. Science 1996;273:104. 148. Janeway CA, Jr. The T cell receptor as a multicomponent signalling machine CD4/CD8 coreceptors and CD45 in T cell activation. Annu Rev Immunol 1992;10:645. 149. Wang J-H, Meijers R, Xiong Y, et al. Crystal structure of the human CD4 N-terminal two-domain fragment complexed to a class II MHC molecule. Proc Natl Acad Sci U S A 2001;98:10799. 150. Krummel MF, Davis MM. Dynamics of the immunological synapse: finding, establishing and solidifying a connection. Curr Opin Immunol 2002;14:66. 151. Irvine DJ, Purbhoo MA, Krogsgaard M, et al. Direct observation of ligand recognition by T lymphocytes. Nature 2002;419:845. 152. Wülfing C, Sumen C, Sjaastad MD, et al. Costimulation and endogenous MHC ligands contribute to T cell recognition. Nature Immunol 2002;3:42. 153. Gao GF, Tormo J, Gerth UC, et al. Crystal structure of the complex between human CD8 alpha and HLA-A2. Nature 1997;387:630. 154. Mannie MD, Rosser JM, White GA. Autologous rat myelin basic protein is a partial agonist that is converted into a full antagonist upon blockade of CD4. Evidence for the integration of efficacious and nonefficacious signals during T cell antigen recognition. J Immunol 1995;154:2642. 155. Vidal K, Hsu BL, Williams CB, et al. Endogenous altered peptide ligands can affect peripheral T cell responses. J Exp Med 1996;183:1311. 156. Hampl J, Chien Y, Davis MM. CD4 augments the response of a T cell to agonist but not to antagonist ligands. Immunity 1997;7:1. 157. Madrenas J, Chau LA, Smith J, et al. The efficiency of CD4 recruitment to ligand-engaged TCR controls the agonist/ partial agonist properties of peptide–MHC molecule ligands. J Exp Med 1997;185:219. 158. Davis MM, Bjorkman PJ. T cell antigen receptor genes and T cell recognition. Nature 1988;334:395. 159. Chothia C, Boswell DR, Lesk AM. The outline structure of the T-cell alpha beta receptor. EMBO J 1988;7:3745. 160. Claverie JM, Prochinicka CA, Bouguelert L. Implications of a Fab-like structure for the T-cell receptor. Immunol Today 1989;10:10. 161. Engel I, Hedrick SM. Site-directed mutations in the VDJ junctional region of T cell receptor beta chain cause changes in antigenic peptide recognition. Cell 1988;54:473. 162. Katayama CD, Eidelman FJ, Duncan A, et al. Predicted complementarity determining regions of the T cell antigen receptor determine antigen specificity. EMBO J 1995;14:927. 163.

Jorgensen JL, Esser U, Fazekas de St. Groth B, et al. Mapping T cell receptor/peptide contacts by

variant peptide immunization of single-chain transgenics. Nature 1992;355:224. 164. Jorgensen JL, Reay PA, Ehrich EW, et al. Molecular components of T-cell recognition. Annu Rev Immunol 1992;10:835. 165. Sant’Angelo DB, Waterbury G, Preston-Hurlburt P, et al. The specificity and orientation of a TCR to its peptide-MHC class II ligands. Immunity 1996;4:367. 166. Sun R, Shepherd SE, Geier SS, et al. Evidence that the antigen receptors of cytotoxic T lymphocytes interact with a common recognition pattern on the H-2Kb molecule. Immunity 1995;3:573. 167. Ehrich EW, Devaux B, Rock EP, et al. T cell receptor interaction with peptide/MHC and superantigen/MHC ligands is dominated by antigen. J Exp Med 1993;178:713. 168.

Hennecke J, Wiley DC. T cell receptor–MHC interactions up close. Cell 2001;104:1.

169. Reinherz EL, Tan K, Tang L, et al. The crystal structure of a T cell receptor in complex with peptide and MHC class II. Science 1999;286:1913. 170. Nanda NK, Arzoo KK, Geysen HM, et al. Recognition of multiple peptide cores by a single T cell receptor. J Exp Med 1995;182:531. 171. Mason D. A very high level of crossreactivity is an essential feature of the T-cell receptor. Immunol Today 1998;19:395. 172. Wilson DB, Pinilla C, Wilson DH, et al. Immunogenicity. I. Use of peptide libraries to identify epitopes that activate clonotypic CD4 + T cells and induce T cell responses to native peptide ligands. J Immunol 1999;163:6424. 173. Chien Y, Davis MM. How aß T cell receptors “see” peptide/MHC complexes. Immunol Today 1993;14:597. 174. Garcia KC, Degano M, Pease LR, et al. Structural basis of plasticity in T cell receptor recognition of a self peptide–MHC antigen. Science 1998;279:1166. 175. Reiser JB, Gregoire C, Darnault C, et al. A T cell receptor CDR3beta loop undergoes conformational changes of unprecedented magnitude upon binding to a peptide/MHC class I complex. Immunity 2002;16:345. 176. Hare BJ, Wyss DF, Osburne MS, et al. Structure, specificity, and CDR mobility of a class II restricted single-chain T-cell receptor. Nat Struct Biol 1999;6:574. 177. Garcia KC, Radu CG, Ho J, et al. Kinetics and thermodynamics of T cell receptor–autoantigen interactions in murine experimental autoimmune encephalomyelitis. Proc Natl Acad Sci U S A 2001;8:6818. 178. Boniface JJ, Reich Z, Lyons DS, et al. Thermodynamics of T cell receptor binding to peptide/MHC: evidence for a general mechanism of molecular scanning. Proc Natl Acad Sci U S A 1999;96:11446. 179. Wu LC, Tuot DS, Lyons DS, et al. A two-step binding mechanism for T-cell receptor recognition of peptide MHC. Nature 2002;418:552. 180.

Marrack P, Kappler J. The staphylococcal enterotoxins and their relatives. Science 1990;248:705.


McDonald KR, Acha-Orbea H. Superantigens of mouse mammary tumor virus. Annu Rev Immunol

1995;13:459. 182. Li H, Llera A, Malchiodi EL, et al. The structural basis of T cell activation by superantigens. Annu Rev Immunol 1999;17:435. 183. Sundberg EJ, Li Y, Mariuzza RA. So many ways of getting in the way: diversity in the molecular architecture of superantigen-dependent T-cell signaling complexes. Curr Opin Immunol 2002;14:36. 184. Festenstein H. Immunogenetic and biological aspects of in vitro lymphocyte allotransformation (MLR) in mouse. Transplant Rev 1973;15:62. 185. Kappler JW, Staerz U, White J, et al. T cell receptor Vb elements which recognize Mls-modified products of the major histocompatibility complex. Nature 1988;332:35. 186. Woodland DL, Happ MP, Gollob KJ, et al. An endogenous retrovirus mediating deletion of ab T cells. Nature 1991;349:529. 187. Marrack P, Kushnir E, Kappler J. A maternally inherited superantigen encoded by a mammary tumour virus. Nature 1991;349:524. 188. Choi Y, Kappler JW, Marrack P. A superantigen encoded in the open reading frame of the 3' long terminal repeat of mouse mammary tumour virus. Nature 1991;350:203. 189. Dyson PJ, Knight AM, Fairchild S, et al. Genes encoding ligands for deletion of Vb11 T cells cosegregate with mammary tumour virus genomes. Nature 1991;349:531. 190. Frankel WN, Rudy C, Coffin JM, et al. Linkage of Mls genes to endogenous mammary tumour viruses of inbred mice. Nature 1991;349:526. 191. Janeway CAJ, Yagi J, Conrad PJ, et al. T-cell responses to Mls and to bacterial proteins that mimic its behavior. Immunol Rev 1989;107:61. 192. Stuart PM, Woodward JG. Yersinia enterocolitica produces superantigenic activity. J Immunol 1992;148:225. 193. Abe J, Takeda T, Watanabe Y, et al. Evidence for superantigen production by Yersinia pseudotuberculosis. J Immunol 1993;151:4183. 194. Tomai M, Kotb M, Majumdar G, et al. Superantigenicity of streptococcal M protein. J Exp Med 1990;172:359. 195. Cole BC, Kartchner DR, Wells DJ. Stimulation of mouse lymphocytes by a mitogen derived from Mycoplasma arthritidis. VII. Responsiveness is associated with expression of a product(s) of the V beta 8 gene family present on the T cell receptor alpha/beta for antigen. J Immunol 1989;142:4131. 196. Friedman SM, Crow MK, Tumang JR, et al. Characterization of the human T cells reactive with the Mycoplasma arthritidis–derived superantigen (MAM): generation of a monoclonal antibody against Vb17, the T cell receptor gene product expressed by a large fraction of MAM-reactive human T cells. J Exp Med 1991;174:891. 197. Lafon M, Scott-Algara D, March PN, et al. Neonatal deletion and selective expansion of mouse T cells by exposure to rabies virus nucleocapsid superantigen. J Exp Med 1994;180:1207. 198.

Dobrescu D, Ursea B, Pope M, et al. Enhanced HIV-1 replication in Vß12 T cells due to human

cytomegalovirus in monocytes: evidence for a putative herpesvirus superantigen. Cell 1995;82:753. 199. Yao Z, Maraskovsky E, Spriggs MK, et al. Herpesvirus saimiri open reading frame 14, a protein encoded by T lymphotropic herpesvirus binds to MHC class II molecules and stimulates T cell proliferation. J Immunol 1996;156:3260. 200. Sutkowski N, Conrad B, Thorley-Lawson DA, et al. Epstein-Barr virus transactivates the human endogenous retrovirus HERV-K18 that encodes a superantigen. Immunity 2001;15:57. 201. Denkers EY, Caspar P, Sher A. Toxoplasma gondii possesses a superantigen activity that selectively expands murine T cell receptor V beta 5–bearing CD8 + lymphocytes. J Exp Med 1994;180:985. 202. Held W, Waanders GA, Shakhov AN, et al. Superantigen-induced immune stimulation amplifies mouse mammary tumor virus infection and allows virus transmission. Cell 1993;74:529. 203. Golovkina TV, Chervonsky A, Dudley JP, et al. Transgenic mouse mammary tumor virus superantigen expression prevents viral infection. Cell 1992;69:637. 204. Stauffer Y, Marguerat S, Meylan F, et al. Interferon-induced endogenous superantigen: a model linking environment and autoimmunity. Immunity 2001;15:591. 205. Dalwadi H, Wei B, Kronenberg M, et al. The Crohn’s disease–associated bacterial protein 12 is a novel enteric T cell superantigen. Immunity 2001;5:149. 206. Malchiodi EL, Eisenstein E, Fields BA, et al. Superantigen binding to a T cell receptor ß chain of known three-dimensional structure. J Exp Med 1995;182:1. 207. Sundberg EJ, Li Y, Mariuzza RA. So many ways of getting in the way: diversity in the molecular architecture of superantigen-dependent T-cell signaling complexes. Curr Opin Immunol 2002;14:36. 208. Fields BA, Malchiodi EL, Li H, et al. Crystal structure of a T cell receptor beta-chain complexed with a superantigen. Nature 1996;384:188. 209. Arcus VL, Proft T, Sigrell JA, et al. Conservation and variation in superantigen structure and activity highlighted by the three-dimensional structures of two new superantigens from Streptococcus pyogenes. J Mol Biol 2000;299:157. 210. Li Y, Li H, Dimasi N, et al. Crystal structure of a superantigen bound to the high-affinity, zinc-dependent site on MHC class II. Immunity 2001;14:93. 211. Petersson K, Hakansson M, Nilsson H, et al. Crystal structure of a superantigen bound to MHC class II displays zinc and peptide dependence. EMBO J 2001;20:3306. 212. Li H, Llera A, Tsuchiya D, et al. Three-dimensional structure of the complex between a T cell receptor chain and the superantigen staphylococcal enterotoxin B. Immunity 1998;9:807. 213. Rojo JM, Janeway CAJ. The biologic activity of anti–T cell receptor V region monoclonal antibodies is determined by the epitope recognized. J Immunol 1988;140:1081. 214.

Haas W, Pereira P, Tonegawa S. Gamma/delta cells. Annu Rev Immunol 1993;11:637.


Chien Y-H, Jores R, Crowley MP. Recognition by ?/d T cells. Annu Rev Immun 1996;14:511.


Stingl G, Gunter KC, Tschachler E, et al. Thy-1 + dendritic epidermal cells belong to the T-cell

lineage. Proc Natl Acad Sci U S A 1987;84:2430. 217. Goodman T, Lefrancois L. Expression of the gamma-delta T-cell receptor on intestinal CD8 + intraepithelial lymphocytes. Nature 1988;333:855. 218. Itohara S, Farr AG, Lafaille JJ, et al. Homing of ?d thymocyte subset with hom*ogeneous T-cell receptors to mucosal epithelia. Nature 1990;343:754. 219. Augustin A, Kubo RT, Sim GK. Resident pulmonary lymphocytes expressing the ?/d T-cell receptor. Nature 1989;340:239. 220. Asarnow DM, Kuziel WA, Bonyhadi M, et al. Limited diversity of ?d antigen receptor genes of Thy-1 + dendritic epidermal cells. Cell 1988;55:837. 221. Feeney AJ. Predominance of VH-D-JH junctions occurring at sites of short sequence hom*ology results in limited junctional diversity in neonatal antibodies. J Immunol 1992;149:222. 222. Havran WL, Chien YH, Allison JP. Recognition of self antigens by skin-derived T cells with invariant gamma delta antigen receptors. Science 1991;252:1430. 223. Sim G-K, Augustine A. Dominate expression of BID, an invariant undiversified T cell receptor d chain. Cell 1990;61:397. 224. De Geus B, Van den Enden M, Coolen C, et al. Phenotype of intraepithelial lymphocytes in euthymic and athymic mice: implications for differentiation of cells bearing a CD3-associated ?d T cell receptor. Eur J Immunol 1990;20:291. 225. Cron RQ, Gajewski TF, Sharrow SO, et al. Phenotypic and functional analysis of murine CD3 + , CD4 - , CD8 - TCR-gamma/delta–expressing peripheral T cells. J Immunol 1989;142:3754. 226. Allison JP, Havran WL. The immunobiology of T cells with invariant ?d antigen receptor. Annu Rev Immunol 1991;9:679. 227. Ohga S, Yoshokai Y, Takeda Y, et al. Sequential appearance of ?d- and aß- bearing T cells in the peritoneal cavity during an i.p. infection with Listeria monocytogenes. Eur J Immunol 1990;20:533. 228. Ferrick D, Schrenzel M, Mulvania T, et al. Differential production of interferon-g and interleukin-4 in response to Th1- and Th2-stimulating pathogens by ?d T cells in vivo. Nature 1995;373:255. 229. Kaufmann S. Bacterial and protozoal infections in genetically disrupted mice. Curr Opin Immun 1994;6:518. 230. Carding S, Allan W, Kyes S, et al. Late dominance of the inflammatory process in murine influenza by gamma/delta + T cells. J Exp Med 1990;172:1225. 231. Hiromatsu K, Yoshikai Y, Matsuzaki G, et al. A protective role of ?/d T cells in primary infection with Listeria monocytogenes in mice. J Exp Med 1992;175:49. 232. Roberts SJ, Smith AL, West AB, et al. T-cell alpha beta + and gamma delta + deficient mice display abnormal but distinct phenotypes toward a natural, widespread infection of the intestinal epithelium. Proc Natl Acad Sci U S A 1996;93:11774. 233.

Mombaerts P, Arnoldi J, Russ F, et al. Different roles of aß and ?d T cells in immunity against an

intracellular bacterial pathogen. Nature 1993;365:53. 234. Kaufmann S, Ladel C. Role of T cell subsets in immunity against intracellular bacteria: Experimental infections of knock-out mice with Listeria monocytogenes and Mycobacterium bovis BCG. Immunobiology 1994;191:509. 235. Kaufmann, SHE. ?/d and other unconventional T lymphocytes: what do they see and what do they do? Proc Natl Acad Sci U S A 1996;93:2272. 236. Witherden DA, Rieder SE, Boismenu R, et al. A role for epithelial gamma delta T cells in tissue repair. Springer Semin Immunopathol 2000;22:265. 237. Komano H, Fujiura Y, Kawaguchi M, et al. Homeostatic regulation of intestinal epithelia by intraepithelial ?d T cells. Proc Natl Acad Sci U S A 1995;92:6147. 238. Kaufmann S, Blum C, Yamamoto S. Crosstalk between alpha/beta T cells and gamma/delta T cells in vivo: activation of alpha/beta T-cell responses after gamma/delta T-cell modulation with the monoclonal antibody GL3. Proc Natl Acad Sci U S A 1993;90:9620. 239. McMenamin C, Pimm C, McKersey M, et al. Regulation of IgE responses to inhaled antigen in mice by antigen-specific gamma delta T cells. Science 1994;265:1869. 240. Wen L, Roberts S, Viney J, et al. Immunoglobulin synthesis and generalized autoimmunity in mice congenitally deficient in alpha beta(+) T cells. Nature 1994;369:654. 241. Viney J, Dianda L, Roberts S, et al. Lymphocyte proliferation in mice congenitally deficient in T-cell receptor alpha beta + cells. Proc Natl Acad Sci U S A 1994;91:11948. 242. Fahrer AM, Konigshofer Y, Kerr EM, et al. Attributes of gamma delta intraepithelial lymphocytes as suggested by their transcriptional profile. Proc Natl Acad Sci U S A 2001;98:10261. 243. Shires J, Theodoridis E, Hayday AC. Biological insights into TCR?d + and TCRaß+ intraepithelial lymphocytes provided by serial analysis of gene expression (SAGE). Immunity 2001;15:419. 244. Schild H, Mavaddat N, Litzenberger C, et al. The nature of major histocompatibility complex recognition by ?d T cells. Cell 1994;76:29. 245. Weintraub B, Jackson M, Hedrick S. Gamma delta T cells can recognize nonclassical MHC in the absence of conventional antigenic peptides. J Immunol 1994;153:3051. 246. Sciammas R, Johnson R, Sperling A, et al. Unique antigen recognition by a herpesvirus-specific TCR gamma delta cell. J Immunol 1994;152:5392. 247. Crowley MP, Reich Z, Mavaddat N, et al. The recognition of the nonclassical major histocompatibility complex (MHC) class I molecule, T10, by the gamma delta T cell, G8. J Exp Med 1997;185:1223. 248. Wingren C, Crowley MP, Degano M, et al. Crystal structure of a gamma delta T cell receptor ligand T22: a truncated MHC-like fold. Science 2000;287:310. 249. Van Kaer L, Wu M, Ichikawa Y, et al. Recognition of MHC TL gene products by gamma delta T cells. Immunol Rev 1991;120:89. 250.

Hampl J, Schild H, Litzenberger C, et al. The specificity of a weak gamma delta TCR interaction

can be modulated by the glycosylation of the ligand. J Immunol 1999;163:288. 251. Crowley MP, Fahrer AM, Baumgarth N, et al. A population of murine gamma delta T cells that recognize an inducible MHC class Ib molecule. Science 2000;287:314. 252. Bauer S, Groh V, Wu J, et al. Activation of NK cells and T cells by NKG2D, a receptor for stress-inducible MICA. Science 1999;285:727. 253. Schoel B, Sprenger S, Kaufmann S. Phosphate is essential for stimulation of V gamma 9V delta 2 T lymphocytes by mycobacterial low molecular weight ligand. Eur J Immunol 1994;24:1886. 254. Constant P, Davodeau F, Peyrat M, et al. Stimulation of human gamma delta T cells by nonpeptidic mycobacterial ligands. Science 1994;264:267. 255. Tanaka Y, Sano S, Nieves E, et al. Nonpeptide ligands for human gamma delta T cells. Proc Natl Acad Sci U S A 1994;91:8175. 256. Tanaka Y, Morita C, Tanaka Y, et al. Natural and synthetic non-peptide antigens recognized by human gamma delta T cells. Nature 1995;375:155. 257. Janis EM, Kaufman SHE, Schwartz RH, et al. Activation of ?d T cells in the primary immune response to mycobacterium tuberculosis. Science 1989;244:713. 258. Modlin RL, Pirmez C, Hofman FM, et al. Lymphocytes bearing antigen-specific ?d T-cells receptors accumulate in human infectious disease lesions. Nature 1989;339:544. 259. Vidovic D, Roglic M, McKune K, et al. Qa-1 restricted recognition of foreign antigen by a gamma delta T-cell hybridoma. Nature 1989;340:646. 260. Holosh*tz J, Vila LM, Keroack BJ, et al. Dual antigenic recognition by cloned human ?-d T-cells. J Clin Invest 1992;89:308. 261. Born W, Hall L, Dallas A, et al. Recognition of a peptide antigen by heat shock reactive ?d T lymphocytes. Science 1990;249:67. 262. Loh EY, Wang M, Bartkowiak J, et al. Gene transfer studies of T-cell receptor-?–d recognition. Specificity for staphylococcal enterotoxin-A is conveyed by V-?-9 alone. J Immunol 1994;152:3324. 263. Kim HT, Nelson EL, Clayberger C, et al. Gamma delta T cell recognition of tumor Ig peptide. J Immunol 1995;154:1614. 264. Rock E, Sibbald P, Davis M, et al. CDR3 length in antigen-specific immune receptors. J Exp Med 1994;179:323. 265. Lerner E, Matis L, Janeway CJ, et al. Monoclonal antibody against an Ir gene product? J Exp Med 1980;152:1085. 266. Ozato K, Mayer N, Sachs D. Hybridoma cell lines secreting monoclonal antibodies to mouse H-2 and Ia antigens. J Immunol 1980;124:533. 267. Fisch P, Malkovsky M, Kovats S, et al. Recognition by human V?9/Vd2 T cells of a GroEL hom*olog on Daudi Burkitt’s lymphoma cells. Science 1990;250:1269. 267a.

Seidl K, Ph.D. diss. “Genetic Influences on development of B-1 cell repertoire”. Stanford

University, 1995. 268. Porgador A, Yewdell JW, Deng Y, et al. Localization, quantitation, and in situ detection of specific peptide-MHC class I complexes using a monoclonal antibody. Immunity 1997;6:715. 269. Dafa*glio G, Nelson CA, Deck BM, et al. Characterization and quantitation of peptide-MHC complexes produced from hen egg lysozyme using a monoclonal antibody. Immunity 1997;6:727. 270. Reay PA, Matsui K, Haase K, et al. Determination of the relationship between T cell responsiveness and the number of MHC-peptide complexes using specific monoclonal antibodies. J Immunol 2000;164:5626. 271. Lang F, Peyrat M, Constant P, et al. Early activation of human V?9Vd2 T cell broad cytotoxicity and TNF production by nonpeptidic mycobacterial ligands. J Immunol 1995;154:5986. 272. Morita CT, Beckman EM, Bukaurki JF, et al. Direction presentation of nonpeptide prenyl pyrophosphate antigens to human ?d T cells. Immunity 1995;3:495. 273. Iwashima M, Green A, Bonyhadi M, et al. Expression of a fetal ?d T-cell receptor in adult mice triggers a non–MHC-linked form of selective depletion. Int Immunol 1991;3:385. 274. Cunningham BC, Wells JA. Comparison of a structural and a functional epitope. J Mol Biol 1993;234:554. 275. Clackson T, Wells JA. A hot spot of binding energy in a hormone-receptor interface. Science 1995;267:383. 276. Davis MM, Lyons DS, Altman JD, et al. T cell receptor biochemistry, repertoire selection and general features of TCR and Ig structure. In: The molecular basis of cellular defence mechanisms. Ciba Fdn, John Wiley and Sons, 1997:94–104. 277.

Berek C, Milstein C. The dynamic nature of the antibody repertoire. Immunol Rev 1988;105:5.

278. Patten PA, Gray NS, Yang PL, et al. The immunological evolution of catalysis. Science 1996;271:1078. 279. Xu JL, Davis MM. Diversity in the CDR3 region of VH is sufficient for most antibody specificities. Immunity 2000;13:37. 280. Taylor LD, Carmack CE, Huszar D, et al. Human immunoglobulin transgenes undergo rearrangement, somatic mutation and class switching in mice that lack endogenous IgM. Int Immunol 1994;6:579. 281. Seth A, Stern LJ, Ottenhoff THM, et al. Binary and ternary complexes between T-cell receptor, class II MHC and superantigen in vitro. Nature 1994;369:324. 282. Corr M, Slanetz A, Boyd L, et al. T cell receptor–MHC class I peptide interactions: affinity, kinetics, and specificity. Science 1994;265:946. 283. Sykulev Y, Brunmark A, Tsomides TJ, et al. High-affinity reactions between antigen-specific T-cell receptors and peptides associated with allogeneic and syngeneic major histocompatibility complex class I proteins. Proc Natl Acad Sci U S A 1994;91:11487. 284.

Stern LJ, Wiley DC. The human class II MHC protein HLA-DR1 assembles as empty alpha beta

heterodimers in the absence of antigenic peptide. Cell 1992;68:465. 285. Callahan JE, Herman A, Kappler JW, et al. Stimulation of B10.BR T cells with superantigenic staphylococcal toxins. J Immunol 1990;144:2473. 286. Takimoto H, Yoshikai Y, Kishihara K, et al. Stimulation of all T cells bearing V beta 1, V beta 3, V beta 11 and V beta 12 by staphylococcal enterotoxin A. Eur J Immunol 1990;20:617. 287. White J, Herman A, Pullen AM, et al. The V beta–specific superantigen staphylococcal enterotoxin B: stimulation of mature T cells and clonal deletion in neonatal mice. Cell 1989;56:27. 288. Yagi J, Baron J, Buxser S, et al. Bacterial proteins that mediate the association of a defined subset of T cell receptor: CD4 complexes with class II MHC. J Immunol 1990;144:892. 289. Herman A, Kappler JW, Marrack P, et al. Superantigens: mechanism of T-cell stimulation and role in immune responses. Annu Rev Immunol 1991;9:745. 290. Held W, Shakhov AN, Waanders G, et al. An exogenous mouse mammary tumor virus with properties of Mls-1a (Mtv-7). J Exp Med 1992;175:1623. 291. Goding JW, Harris AW. Subunit structure of cell surface proteins: disulfide bonding in antigen receptors, Ly-2/3 antigens, and transferrin receptors of murine T and B Lymphocytes. Proc Natl Acad Sci U S A 1981;78:4530.

Chapter 9 T-Cell Developmental Biology Fundamental Immunology

Chapter 9 Ellen Rothenberg

T-Cell Developmental Biology

OVERVIEW OF T-CELL DEVELOPMENT Key Molecules: Cell Stage Markers and T-Cell Receptor Genes Narrative of T-Cell Development Regulated Proliferation in T-Cell Development Anatomical Path of T-Cell Development Variations in Thymocyte Development in Ontogeny Thymocyte Development in Species Other than Mouse Plan of Chapter: Close-up Views of Key Events EARLY LINEAGE CHOICES: CLUES TO MOLECULAR MECHANISMS Developmental Potential of Earliest Intrathymic Precursors Molecular Indices of T-Lineage Specification and Commitment Genetic Requirements for T-Lineage Specification and Commitment A REGULATORY UPHEAVAL: ß-SELECTION Multiple Changes at the Transition from T-Cell Receptor–Independent to T-Cell Receptor–Dependent T-Cell Development Triggering Requirements for ß Selection Constituent Events in the ß Selection Cascade Death Mechanisms and Other Checkpoint Controls Significance of ß Selection for Later T-Cell Differentiation THE DIVERGENCE OF T-CELL RECEPTOR aß AND T-CELL RECEPTOR ?d LINEAGE CELLS Choices of Fate within the T-Cell Lineage: Differences between aß and ?d T Cells Generation of T-Cell Receptor ?d Cells Genetic Regulation of T-Cell Receptor aß versus T-Cell Receptor ?d Cell Production Models for the T-Cell Receptor aß:T-Cell Receptor ?d Lineage Choice POSITIVE AND NEGATIVE SELECTION The Double-Positive Thymocyte Stage Time Windows for Positive and Negative Selection Triggering and Results of Positive Selection Strength of Signal versus Distinct Interaction Models for Positive and Negative Selection Another Escape from Autoreactivity in the Thymic Cortex CD4 HELPER T-CELL VERSUS CD8 CYTOTOXIC T-CELL LINEAGE COMMITMENT Major Histocompatibility Complex Restriction Regulates CD4 versus CD8 Lineage Differentiation Models for CD4/CD8 Lineage Divergence Molecules Implicated in the CD4/CD8 Lineage Choice Maturation and Export of CD4 and CD8 Single-Positive Thymocytes Relationships between Positive Selection, Negative Selection, and CD4/CD8 Lineage Choice FRONTIERS FOR THE FUTURE: MYSTERIES AND ALTERNATIVES IN T-CELL DEVELOPMENT Alternative Pathway or Distinct Precursors: The Case of the NK T Cells Variations on a Theme of Tolerance: Regulatory T Cells CONCLUDING REMARKS ACKNOWLEDGMENTS


T-cell development is a composite of overlapping processes in the domains of developmental biology, immunology, and cell biology. It starts with purely hematopoietic developmental mechanisms leading to lymphoid commitment, T-lineage commitment, and later developmental choice points; then gradually, the basis for developmental choices becomes dominated by the immunology of T-cell receptor (TCR) repertoire selection. The underlying mechanisms by which these later choices are made can be understood only in terms of a richly complex cell biology of checkpoint enforcement, defining the two TCR-dependent fate-determination processes of ß selection and positive/negative selection. Repertoire selection is crucial for establishing a functionally competent, mostly self-tolerant population of peripheral T cells, and it has attracted a great deal of interest in isolation from other aspects of T-cell development. In this chapter, we show how this cellular process occurs, on the basis of mechanisms that emerge from a unique and fascinating developmental program. A recurrent theme is how the signals from various TCR complexes come to intertwine with underlying developmental mechanisms to control cell fate at a succession of distinct checkpoints and lineage choices. To begin, this chapter introduces the broad map of T-cell developmental events. The subsequent sections focus in on the mechanisms involved at a few of its most interesting watersheds.

OVERVIEW OF T-CELL DEVELOPMENT In mammals, most circulating T cells develop in the thymus. Bone marrow precursors in small numbers enter the thymus from the blood and undertake a course of proliferation, differentiation, and selection, which converts them into T cells in about 4 weeks (faster in fetal animals). The mature cells then emigrate from the thymus and take up their surveillance roles in the body. Precursors seed the thymus and differentiate into T cells continuously from midgestation throughout adult life. An additional site of development is in the intestinal epithelium, in which T cells that mostly remain associated with the gut epithelium appear to be generated. In either case, T cells distinguish themselves from most hematopoietic cell types by migrating away from the bone marrow in order to carry out their differentiation. It is in the thymus that most TCR gene rearrangement occurs and the cells first acquire their clonal recognition properties. The thymus not only promotes maturation but also rigorously screens each cohort of developing cells to eliminate those with either useless or dangerous TCRs, in a process called “repertoire selection.” Key Molecules: Cell Stage Markers and T-Cell Receptor Genes At any one time after birth, the thymus contains cells in all stages of development, from the earliest precursors to cells that are virtually mature. Understanding of the process of T-cell differentiation has been possible because cells in different stages can be distinguished, and cells of each type can be isolated preparatively without being killed. At least seven developmental stages can be distinguished on the basis of their expression of useful surface molecules. These are introduced in Fig. 1. Key markers for

subdividing the majority of thymocytes are TCRaß, TCR?d, and the co-receptors CD4 and CD8 ( Fig. 1A). These help to identify cells in the later 2 weeks of intrathymic differentiation. The majority of thymocytes, approximately 80%, express both CD4 and CD8 and low levels of surface TCRaß complexes, a constellation of markers that is not seen in general on peripheral T cells ( Fig. 1A). This distinctive population, called “double positive” (DP), is a key developmental intermediate that undergoes “TCR repertoire selection,” the complex process that eliminates cells with either useless or autoreactive TCR specificities. The unique properties of DP cells make TCR repertoire selection possible. Minorities of the cells are CD4 +CD8 - TCRaß high or CD4 -CD8 + TCRaß high, and these “single positive” (SP) thymocytes are the most mature cells.

FIG. 1. Subsets of T-cell precursors: normal development versus. development without T-cell receptor (TCR) gene rearrangement. The major subsets of cells discussed in this chapter, as they appear in a typical flow cytometric analysis. Normal thymocytes are shown on the top ( A and C), whereas thymocytes from recombination activating gene (RAG)–deficient mice, which cannot rearrange any TCR genes, are shown on the bottom ( B and D). The cells are stained with fluorescent antibodies against CD4 and CD8 ( A and B), and the double-negative (DN) cells are further stained with fluorescent antibodies against CD44 and CD25 ( C and D). The axes represent increasing levels of these surface molecules on a 4-decade logarithmic scale: that is, a 10,000-fold range in fluorescent staining intensity. The main populations discussed in Fig. 2 are (DN, DP, CD4 SP, CD8 SP) ( A), and the DN cells are subdivided into DN1, DN2, DN3, and DN4 ( C). Comparison between the upper and lower panels shows that the recombinase-deficient thymocytes are developmentally arrested in the DN stages ( B), with cells accumulating in the DN3 state and blocked from progressing forward to the DN4 state ( D; cf. C). RAG-deficient thymocytes also accumulate only about 1/100 as many cells in the thymus as wild-type thymocytes (˜4 × 10 6 vs. ˜3 × 10 8 ).

FIG. 2. Outline of events in T-cell development. Summary of the events occurring in normal mouse T-cell development, indicating the approximate time taken in each set of transitions, the developmental branch points, and key changes in gene expression and T-cell receptor (TCR) gene rearrangement status. Developmental branch points taken rarely are indicated by broken-line arrows. The two major checkpoints discussed in the chapter, ß selection and positive selection, are indicated. The alternative to positive selection, death, includes both negative selection and death by “neglect,” as discussed in the text. Small curving arrows over the double-negative 1 (DN1)? DN2 and DN4?immature single positive (ISP) stages denote the extensive proliferation at these stages, also suggested by the cartoon at the top. Stages of development in which a majority of cells are seen to be in cycle are indicated by gray filled circles. Cells expressing rearranged TCR? and TCRd genes (TCR?d) and cells expressing rearranged TCRß genes either alone or together with rearranged TCRa genes are indicated above the main diagram. Below the main diagram, bars show the extents of expression of useful cell surface markers other than CD4, CD8, and the TCR complexes. Periods of recombinase expression, specific gene rearrangement, and key developmental events are also indicated by horizontal bars. Broken bars show reduced levels of expression. Common abbreviations of cell stages are given in the text. Hemato, hematopoietic; Lymphoid pre, lymphoid precursor.

Cells in the earlier 2 weeks of differentiation in the thymus lack any TCR expression as well as any expression of CD4 or CD8. Nevertheless, different stages can be distinguished in this “double negative” (DN) or “triple negative” population. In mice, they can be subdivided, on the basis of expression of the interleukin (IL)–2 receptor a-chain CD25 and the adhesion molecule CD44, into progressive developmental stages termed DN1, DN2, DN3, and DN4 (or TN1, TN2, and so forth) ( Fig. 1B). Two other useful markers for these stages are the stem cell growth factor receptor c-kit (CD117), which is coexpressed with CD44, and the small phospholipid-linked heat-stable antigen (HSA, CD24) which is turned on with CD25 and remains on until the latest stages of thymocyte maturation. In the human system, different markers are useful for distinguishing corresponding stages, and they described later. An outline of the progression of mouse precursor cells through these stages is shown in Fig. 2 as a framework for this narrative. TCR gene rearrangement plays a pivotal role in thymocyte fate. Ultimately, thymocytes can survive to maturity only if they successfully carry out combinations of gene rearrangements that will give them in-frame a and ß chains or ? and d chains, to be assembled into TCRaß/CD3 or TCR?d/CD3 complexes. The rules of the process are

therefore worth reviewing. There are four TCR gene loci, each consisting of the constant region exons and multiple variable (V), joining (J), and sometimes diversity (D) segments of the TCRa, ß, ?, and d chain genes (see Chapter 8). The TCRß and TCRd loci have D segments as well as V and J segments to be rearranged, whereas the TCRa and TCR? loci do not. Also, note that the TCRd locus is embedded in the middle of the TCRa locus in such a way that any V-Ja rearrangement automatically deletes the TCRd locus entirely, whether it had undergone rearrangement before or not. These features are important for the regulation of rearrangement and, as described later, for understanding the choice between becoming a TCRaß- or a TCR?d-lineage thymocyte. The rearrangement process is ordered, with D-Jß rearrangements occurring before V-D-Jß and V-Dd rearrangements occurring before V-D-Jd. Narrative of T-Cell Development Figure 2 traces the progress of cells through the best-known stages of T-cell development. The cells that enter the thymus are capable of giving rise to all subsets of T cells plus natural killer (NK) cells and dendritic cells. As discussed later, they may be able to give rise to macrophages and B cells, too. These cells are initially c-kit +, Thy-1 low, CD44 high, CD25 -, and CD24 low. At this stage, the TCR genes are not yet rearranged. These precursor cells form the key component of the subset called DN1 or TN1. Once in the thymus, these cells undergo a major transition, losing much of their ability to give rise to anything but T cells, turning on the expression of multiple T-cell genes, and starting to proliferate. They begin to express Thy-1, CD25, and CD24, and CD44 and c-kit continue to be expressed on the cell surface, although at declining levels. The stage marked by this new phenotype is classified as DN2 (TN2) ( Fig. 1B & Fig. 2). Proliferative expansion during this stage is considerable, approximately 6 to 10 rounds of division. This is the stage when TCR gene rearrangement begins. TCR?, TCRd, and TCRß all appear to be similarly accessible to rearrangement during this initial period, but TCRa is not. At the next stage, DN3, CD44, and c-kit are fully down-regulated; most cell proliferation stops; and rearrangement of the TCRß, ?, and d genes occurs with maximum efficiency. The DN3 stage (Thy-1 + c-kit - CD44 - CD25 + CD24 + cells) is a landmark because, in both adult and fetal thymocytes, it is the first stage when the cells appear to have lost the ability to develop into anything but T cells. It is also the first stage when the protein products of rearranged TCR genes are detected in the cytoplasm. Beyond this stage, the proliferation and survival of the cells depend essentially on interactions mediated by TCR proteins. If they rearrange their TCR genes correctly, they can proceed, often with a burst of proliferation. If they fail, they die. The exact path that the cells follow from this point depends on whether the cells succeed in making D-Jß and V-D-Jß rearrangements to form a productive TCRß open-reading frame before they have completed productive rearrangements of both the ? and d loci. In the first case, they develop into aß T cells; in the second case, they

develop into ?d T cells. The ?d cells mature with little additional proliferation and with few known changes to their surface phenotype other than down-regulation of CD25 and CD24. Cells that rearrange ß, on the other hand, undergo a complex succession of events known as ß selection. These cells proliferate in a rapid burst; down-regulate the DN2/DN3 marker CD25; turn on expression of CD4 and CD8; stop TCRß, TCR?, and TCRd rearrangements; begin rearranging TCRa; and undergo profound functional transformations. Through this cascade of events, the cells are quickly transformed from DN3 to DP cells, through proliferating intermediates called DN4 and immature single positive (ISP) cells, usually CD8 +CD4 -CD3 - ( Fig. 2). DP cells are physiologically peculiar; these peculiarities make them uniquely poised for TCR-dependent repertoire selection. They are therefore key intermediates in the production of a self-tolerant T-cell population. As a rule, the DP cell fate is part of the aß program and not part of the ?d program. Thus, although the choice of aß versus ?d fate is based at least partly on the stochastic success or failure of rearrangements, it results in a real choice between developmental programs. Cells that fail to complete any productive TCRß or TCR? and d gene rearrangements die within a few days. In mutant mice that cannot make rearrangements at all, development cannot proceed beyond the DN3 stage ( Fig. 1D), and death of cells blocked at that point results in a thymus that is only about 1% of the normal cellularity. Besides the choice of TCRaß versus TCR?d, the DN3 stage therefore represents a rigorous developmental checkpoint. The “ß-selection checkpoint” is the first of two checkpoints at which survival is dependent on the TCR. For cells taking the TCRß + CD4 +CD8 + path, rescue from death at the ß-selection checkpoint is only a temporary, conditional reprieve. In these DP cells, TCRß rearrangement must be followed by a successful TCRa gene rearrangement within about 3 days after the proliferative burst subsides, or else the cells die of “neglect.” The selection for cells that have made an acceptable TCRaß complex defines the second TCR-dependent checkpoint in T-cell development: “positive selection.” The criteria for rearrangement success here are more stringent than for ß selection. Any TCRß gene rearrangement that generates a translatable protein coding sequence is adequate for ß selection, but the TCRa rearrangement is evaluated both on the basis of a translatable protein coding sequence and on the basis of the recognition specificity that emerges from the new combination of TCRa chain with the previously fixed TCRß chain. The cells must be able to interact with major histocompatibility complex (MHC) molecules in the microenvironment, but not too well, or else the cells die. The criterion is set so that individual CD4 +CD8 + TCRß + cells have less than a 5% chance of satisfying it. As a result, about 30% of this population dies each day in the young mouse thymus (over 90% die without maturing in the whole 3- to 4-day lifetime of each cell cohort) and must be replaced as a fresh cohort of CD4 +CD8 + cells enters the selection pool. DP cells are actually put through two tests. The first determines whether the newly expressed TCRaß can make sufficiently strong interactions with MHC molecules to be useful, and the second assesses whether the interactions of this TCR with self-antigens in the thymus are weak enough to reduce the danger of autoimmunity. These thresholds are tested in two separable processes: positive selection and negative selection. Cells exceeding the minimum affinity threshold are positively selected, initiating a new

cascade of phenotypic changes and enhancing the viability and functional responsiveness of the cells. Cells that exceed the maximum affinity threshold can be stripped of their receptors or negatively selected by induced apoptosis. Key changes that help trace progress through positive selection are the transient up-regulation of the activation marker CD69, the stepwise increase in TCRaß surface expression from low to intermediate to high, a parallel up-regulation of CD5 and MHC class I molecules, and ultimately the down-regulation of the immature cell marker CD24. Cells remain susceptible to negative selection for several days after the initiation of positive selection, however. They may even encounter the most potent negative selection stimuli in the period after positive selection. Only cells escaping both death by neglect and death by negative selection can complete their maturation and emigrate to the peripheral lymphoid system. Positive selection also appears to drive a choice of maturation fates. It is in the emergence from the DP state that there occurs the first evidence of whether a cell will be a CD4 + helper/regulatory cell or a CD8 + killer cell. As described in detail later, detailed aspects of the TCR–ligand receptor interactions during this process guide or select cells to develop into one type of effector or the other. Cells with TCRs that recognize MHC class II molecules tend to develop as CD4 + cells, whereas those with TCRs that recognize MHC class I molecules develop as CD8 + cells. The basis of this profound differentiation choice is extremely interesting and appears to include subtle quantitative aspects of TCR/co-receptor interaction with MHC and the combination of TCR-generated signals with signals from other pathways. Regulated Proliferation in T-Cell Development Throughout T-cell development, phases of intense proliferation alternate with phases of little or no cycling. The differences in cell cycle activity are dramatic. As shown in Fig. 2, DN2 cells are cycling, DN3 cells halt, and then, after ß selection, cells appear to go through six to eight rounds of division in about 3 days ( 1 , 2 ). After positive selection, in contrast, maturing cells appear to reside in the medulla for approximately 2 weeks without any significant proliferation. Each phase of proliferation tends to be driven by a different mechanism ( 3 ). The growth controllers used at various stages include many of the genes that are essential for progression through the T-cell developmental pathway, as discussed later. In the first DN1 precursors, the initiation of cell division may be controlled by c-kit signaling. In the DN1-DN2 states, proliferation is mostly driven by signals from the interaction of IL-7, a cytokine secreted by the thymic stroma, with IL-7 receptor complexes, composed of IL-7Ra (CD127) and ?c chains (CD132) ( Fig. 2) ( 4 , 5 and 6 ). During ß selection, population expansion is driven by signaling through the pre-TCR (TCRß complex), which is discussed later in detail. The later events of positive selection and maturation do not involve significant proliferation; however, they completely depend on TCRaß for survival signaling. These sequential requirements for survival have overlapping critical periods, so that the mutant phenotypes are slightly leaky but still have powerful quantitative effects ( 7 ). After export from the thymus, mature T-cell proliferation depends on TCRaß triggering and signals through the IL-2, IL-4, or IL-7

receptors. The shift from one kind of proliferative stimulus to another is caused least in part by intrinsic developmental changes in the cells. It is an important factor contributing to the one-way polarity of developmental change. The proliferative bursts are vital for setting up the large excess of precursors that makes it possible to use stringent selection at the checkpoints in T-cell development. The huge losses that occur in these selection processes seem shocking, except that the developmental program also provides for more than 10 5 -fold clonal expansion from each precursor. Harsh selection against useless or autoreactive cells is a price the adaptive immune system must pay for its somatic generation of diversity in recognition structures. This suggests that the mechanisms used by lymphoid precursors to drive developmental proliferation are evolutionarily old and likely to have coevolved with the mechanisms generating clonal diversity. The extents of proliferation at particular stages in T-cell development are somewhat flexible. Proliferation at ß selection can compensate for poor precursor expansion in an earlier phase. Also, although positive selection per se does not involve proliferation, cells in the fetal and early postnatal thymus can undergo several cell cycles during the maturation period after positive selection. This presumably helps to supply T cells to the body quickly, during a period when peripheral T cells are rare. Later, in the adult thymus, maturation occurs with little if any proliferation. Development of T cells depends on proliferation but not in a way that rigidly links certain events to precise numbers of cell cycles. Anatomical Path of T-Cell Development The thymus is made up of lobes, each of which is divided into distinct zones with different stromal cells making up their microenvironments ( 8 ). The largest domain is the cortex, which is packed with DP thymocytes. The cortex surrounds an inner domain called the medulla, where the most advanced SP thymocytes are found. The outermost rim of the cortex (i.e., the subcapsular region) and the region defining the cortical/medullary junction are also specialized in some ways. The organization of the thymus is diagrammed schematically in Fig. 3.

FIG. 3. Summary of migration pathways of thymocytes through the postnatal and fetal thymic microenvironments. The pathway of migration of adult thymocytes from the postcapillary venule (PCV) through the cortex, subcapsular zone, cortex and medulla is shown on the left ( 9 ). For comparison, the entry and migration through the fetal thymus is shown on the right.

From ß selection onward, the pathway of thymocytes through the thymus is well known. Most of the proliferative expansion that follows ß selection occurs in the subcapsular zone. As the cells are pushed away from the rim toward the inner cortex, they stop dividing. DP cells in the cortex continue to sink inward during their 3 days of postmitotic life, but they are not allowed to enter the medulla unless they pass positive selection. Those that do not succeed die in the cortex and are rapidly engulfed by resident macrophages. The path taken from entry into the thymus until ß selection has been less clear. Because of the enormous proliferative expansion during T-cell development, the number of cells in the earliest stages, at any one time, is dwarfed by the number of cells in the later stages. All the DN stages together, representing approximately 2 weeks of developmental change, contribute only approximately 2% to 4% of a typical young adult thymus. The earliest precursors in the DN population have been estimated to be less than one one-hundredth of that frequency. Conceivably, a thymus that turns over approximately 5 × 10 7 cells per day may be resupplied with an input of only 50 to 100 cells per day. The low input numbers have made it hard to trace the path of precursors through intrathymic domains in their early development. Only since 2000 have DN1- and DN2-specific markers been used successfully to track the path of these thymic entrants specifically ( 9 ). This has revealed a more organized role for different subregions of the adult thymus than was previously suspected. It now appears that blood-borne precursors enter the parenchyma of the postnatal thymus by exiting from the medium-sized blood vessels (postcapillary venules) at the cortical/medullary junction. These cells, in an early DN1 state, slowly make their way outward through the thymic cortex to the subcapsular zone, proliferating and differentiating through the DN2 and DN3 states over the next 2 weeks. Those cells becoming ?d cells may not need to progress further than the midcortex. Those continuing on to become aß cells travel all the way to the periphery ( Fig. 3). The ß-selection checkpoint appears to be encountered mostly in the subcapsular zone. Numerous, rapidly proliferating blasts can be seen in the outer cortex; they represent the DN4, ISP, and early DP cells that have just passed this test. As they stop proliferating, DP thymocytes fall back into the cortex, being pushed progressively deeper by products of later cell divisions over the next 2 to 3 days. Those few that are positively selected, along with mature ?d cells, migrate from the cortex to the medulla ( Fig. 3) ( 9 ). Maturation of the surviving cells occurs in the medulla over the following 2 weeks. The medullary epithelial cells are distinct from the cortical epithelial cells. It is not clear exactly how they influence final T-cell maturation, although evidence suggests that they may express a very wide range of self-antigens that may be valuable in negative

selection ( 10 ). Dendritic cells, which are also specifically located in the medulla, present self-antigen to the newly SP cells in the most efficient way for the most stringent form of negative selection. It is only after surviving these encounters that the SP cells complete their maturation, changing their response physiology so that high-affinity interactions with antigen can lead to activation instead of paralysis or death. The compartmentalization of functions in the thymus probably helps guide certain developmental transitions. As the cells change in intrinsic responsiveness to different proliferative signals, their migration may carry them from a zone rich in early stimuli (e.g., IL-7) to a zone that may be rich in later stimuli [possibly Wnt ( 11 , 12 )]. Certain transitions from proliferation to G 1 arrest, such as from DN2 to DN3 or from ß selection to a resting DP state, could result from migration of the cells into a zone where the most recent proliferative stimulus is no longer present. Also, the change in direction of migration of precursors through the cortex, from outbound to inbound, is likely to be a result of a change in the expression of adhesion molecules and chemokine receptors in the DN3 state or during ß selection. There is evidence that could implicate particular integrins (e.g. a 4 ß 1 ) in the outward migration of DN cells ( 13 ), and the roles of these important molecules are likely to become much clearer in the near future. At least one other adhesion molecule, CD44, is expressed very highly in the DN1 cells but then clearly turned off between the DN2 and DN3 state; this could be another participant in the early homing or guidance mechanisms or both. Later, at least one new chemokine receptor is turned on at ß selection, and this may help attract DP cells back toward the interior ( 14 , 15 and 16 ). Migration between domains with different kinds of stromal cells allows a separation between positive and negative selection events in space and time. As discussed in a later section, this can clarify some otherwise confusing features of these processes. Variations in Thymocyte Development in Ontogeny The general outlines of thymocyte development are consistent between postnatal and fetal mice, but there are many differences, both subtle and overt. The mouse fetal thymic stroma develops from the outpocketing of third pharyngeal pouch endoderm, between 10 and 12.5 days of gestation (E10 to E12.5). It may undergo some inductive interaction with the overlying ectoderm and neural crest cells of the third branchial arch ( 17 ). The thymic epithelium, in any case, establishes a distinctive structure and gene expression pattern before any lymphoid precursors arrive ( 17 , 18 ). Seeding begins at about E12 as hematopoietic precursors migrate across the mesenchyme from the subclavian vessels to the tiny, nonvascularized epithelial rudiment ( 19 , 20 ). These cells are CD45 +, expressing certain lymphoid genes and marked by the presence of the lymphoid transcription factor Ikaros in the nucleus. The precursors collect around the thymus and then enter it directly from the outside; this is the one time that the future outer cortex is a point of entry ( Fig. 3, right panel). Lymphoid precursors proliferate exponentially in the fetal murine thymus, increasing from about 10 4 cells at E14.5 dpc to about 5 × 10 5 at E18. In contrast to the kinetics in the postnatal thymus, in the fetal thymus the first TCRaß DP cells are generated at

about E16, the first SP CD4 cells about 2 days later, and CD8 cells a day after that, followed by birth at day 20; the first emigrants are exported to the periphery within 3 days after birth. Thus, the times in the DN state and in the medulla are each cut from approximately 14 days in the postnatal (“adult”) thymus to less than 3 days in the fetal thymus. The precursors that initially seed the fetal thymus are qualitatively different from any that enter it later. Radiation chimera experiments establish that they can generate the same types of T cells that are made from adult bone marrow–derived precursors, but they also have a capacity that adult precursors do not. The first thymic immigrants are uniquely capable of generating two classes of ?d T cells that seed the skin, tongue, and reproductive organ epithelia in late fetal life but are not produced at all after birth ( 21 ). These “early-wave” ?d cells use directed, predetermined V(D)J rearrangements without junctional diversification, so that their recognition specificities are completely invariant. They are, in fact, the first wave of TCR + thymocytes made in the mouse at all, maturing by E15.5 to E16.5, and they have a number of distinctive physiological properties, including growth factor requirements and transcription factor profiles ( 22 ). The precursors that make these cells may themselves be unusual in other ways as well. For example, it is not clear whether they ultimately originate in the intraembryonic hematopoietic tissues at all or from the molecularly unique precursor cells in the yolk sac ( 23 ). At least one additional wave of precursors enters the thymus closer to the time of birth ( 24 ), and then further precursors can continue to enter throughout life. In the meantime, the properties of the major populations of hematopoietic precursors themselves continue to change, shifting from fetal liver to bone marrow and acquiring new molecular properties ( 23 ). Certain gene disruptions have sharply different effects on T-cell development in fetal, postnatal, and adult mice, because of intrinsic differences in the programming of different precursor cohorts ( 25 , 26 , 27 , 28 and 29 ) (detailed in the section on genetic requirements of T-lineage specification and commitment and in Fig. 5B later). It is intriguing that T-cell differentiation takes long enough, in relation to precursor cohort succession, so that considerable overlap could occur in young postnatal animals. Thus, we can even consider that in a 3- to 4-week-old weanling mouse, the bone marrow stem cells, intrathymic DN3 cells, and mature medullary SP thymocytes ready for export at that time could each represent progeny of distinct cohorts of progenitor cells, each with different genetic requirements, molecular expression properties, or both. Distinctions between results in fetal and adult systems are noted throughout this chapter.

FIG. 5. Effects of mutations in growth factor and transcription factor genes on thymus population sizes. A: Evidence for a critical role of interleukin (IL)–7/IL-7R signaling in proliferation at a stage of T-cell development preceding ß-selection. Thymus glands are shown from wild-type and the indicated mutant mice. The recombination activating gene 2 (RAG2) -/- mutation alone results in a 50- to 100-fold reduction in cell numbers because of a developmental block at ß-selection. The IL-7 -/- mutation alone has a similar impact on the size of the thymus. The multiplicative effect of the double mutation shows that IL-7 is needed for the proliferation that normally takes place in double negative (DN) cells in the RAG2 -/- mouse thymus. From( 5 ), with permission. B: Growth of cell numbers in the thymus from fetal life to adulthood in wild-type (wt) mice and mice with inactivating mutations in various genes. From ( 175 ), with permission. T-cell factor (TCF)–1 (TCF), GATA-3, and Ikaros are transcription factors discussed in the text. Ikaros-N -/- is a mutation that creates a dominant negative form; Ikaros-C -/- is a loss of function mutation ( 152 ). Double mutants lacking both c-kit and the common cytokine receptor chain ?c do not generate any T-cell precursors, as discussed in the text. Mutants lacking a 4 -integrin are defective in precursor migration, affecting T cell precursors after birth. The shifts in effects of some mutations between fetal life and postnatal life indicate the distinct molecular requirements for early and later T-cell development.

Thymocyte Development in Species Other than Mouse The overall organization of thymic lobes is conserved from mammals to cartilaginous fish. The overall roles of cortex and medulla for T-cell development seem to have been established early in vertebrate evolution. In chickens and Xenopus, there is also strong evidence for development of distinctive T-cell populations in the tadpole or embryo, unlike those made after hatching. Chicken thymocyte developmental studies have actually provided some of the first evidence for the distinctiveness of early-wave ?d cells. However, markers are just becoming available to distinguish developmental stages in these animals ( 30 , 31 ), and much remains to be learned about hom*ology or

lack of hom*ology of their T-cell developmental pathways with those in mice. Thymocyte development has been studied in some detail in rats and humans as well as in mice. Comparison of features can indicate the most critical events in a developmental process, which tend to be conserved. One surprising finding is the relative lack of conservation of detailed patterns of surface marker expression, even over these short phylogenetic distances. The Thy-1 surface glycoprotein is not expressed in rat or human mature T cells, and CD25, which is distinctively up-regulated in murine DN2 and DN3 stages, is not up-regulated at a corresponding stage in human ( 32 , 33 and 34 ). Rat and human T cells can also express MHC class II molecules on activation, which is confined to non–T cells in mice, whereas CD2, which is T-cell specific in humans, is expressed by murine B and NK cells as well. The rat system has been more closely studied with reference to positive selection and its signaling requirements; some of the details of this process are different in rats than in mice ( 35 ). The timing of CD8 and CD4 expression relative to ß selection is also slightly shifted from the mouse, which is, again, different from the human pattern ( 36 ). hom*ologous surface markers cannot always be assumed to mark hom*ologous developmental stages in these three mammalian species. Against this background of evolutionary variation, however, there are salient points of similarity, listed in Table 1. The human pathway starts with an uncommitted precursor (CD34 +CD38 +) that enters the thymus with a range of developmental potentials that is similar or identical to that of its mouse counterpart. The sequence of T-lineage commitment in relation to onset of T-cell gene expression and TCR gene rearrangement is similar to the timing in the murine system, as is the order of TCR gene rearrangement. The markers useful for distinguishing these stages are different, and CD4 and CD8 are turned on at a different time in relation to ß selection. Nonetheless, the properties of DP thymocytes generated after ß selection are grossly conserved, although they are generally more viable and responsive than their murine counterparts. As in the mouse, TCR high SP thymocytes newly generated through positive selection remain functionally immature at first, which implies that maturation after positive selection is required ( 37 , 38 ). The cell surface markers that provide landmarks for this process are summarized in Table 1.

TABLE 1. T-cell development stage markers: a mouse–human comparison

Plan of Chapter: Close-up Views of Key Events This overview of intrathymic T-cell development gives a sense of the intricacy of the

process, in terms of both the choices the cells must make and the constantly shifting interactions of the cells with their environment. The following sections of this chapter focus on five aspects of T-cell development in depth, in which underlying molecular mechanisms are beginning to be revealed. These areas offer insight into the ways T-cell precursors make the subtle and precise distinctions that are necessary to allow a complex hematopoietic developmental sequence to be governed by TCR specificity and self-tolerance.

EARLY LINEAGE CHOICES: CLUES TO MOLECULAR MECHANISMS Developmental Potential of Earliest Intrathymic Precursors The first question concerns when precursor cells are actually determined to become T cells at all. What is the relationship of this event to TCR gene rearrangement, to entry into the thymus, and to regulatory changes leading to expression of T-cell genes? Answers have emerged from experiments that define the functional properties of the least differentiated cells in the thymus. The most primitive precursors in the thymus have been identified in two ways. One is on the basis of their time of appearance during gestation; that is, the first hematopoietic cells found in the thymus in the fetus. The other is in the adult thymus, through the use of multiple cell surface markers to purify subsets of cells that have immature characteristics (DN, c-kit +) and then assays of their abilities to differentiate into T cells in adoptive transfer experiments. Adoptive transfer of cells to genetically distinct hosts has been the gold standard for proving developmental potential. Finely optimized fetal thymic organ cultures allow any T-lineage progeny to be studied on a smaller scale. In these assays, the most primitive of the precursors are identified as those that give the largest output of descendants per input cell, generate them over the longest time course, and give rise to other verified T-lineage precursors as intermediates in the process. In the adult thymus, the most primitive precursors are found as a subpopulation within the DN1 class. Adoptive transfer into irradiated mice or into fetal thymic organ cultures in vitro reveals that such cells actually have a range of developmental potentials, giving rise to non–T cells as well as T cells. This was shown first at the population level ( 39 , 40 ), and, more recently, it has been confirmed rigorously for single cells. In the murine fetal thymus and the postnatal human thymus, precursors are robust enough to be assayed in single-cell tests for multiple lineage precursor activity ( 41 , 42 and 43 ). These assays agree in showing that many early intrathymic precursors individually have the potential to give rise to dendritic cells, NK cells, and diverse classes of aß and ?d T cells. In contrast, none of the intrathymic precursor populations have been reported to give rise to erythroid, megakaryocytic, or granulocytic cells. Thus, the precursors that enter the thymus are partially restricted in developmental potential but still uncommitted to the T-cell lineage. Oligopotent, partially restricted cells may not be the only cells seeding the thymus. There is strong evidence that some cells can become committed to the T-cell lineage

prethymically. This is indicated by the presence of partial TCR gene rearrangements and by the selective ability to generate T cells in adoptive transfer. Prethymic commitment of some cells to the T lineage is especially evident in the fetus ( 44 , 45 ). Some of these cells may be dedicated precursors of special, extrathymically developing lineages of T cells ( 46 , 47 , 48 and 49 ). Note that there is no known reason why T-lineage commitment could not occur both inside and outside the thymus; NK and mast cells are examples of other hematopoietic cell types that may undergo this step either in the bone marrow or in other sites. The mechanism that enables precursors to home to the thymus from the fetal liver or bone marrow is still poorly understood, but it presumably is a matter of altered expression of particular adhesion molecules and chemokine receptors. Conceivably, these alterations can occur at different times, in relation to other early T-lineage differentiation events, in fetal liver versus adult bone marrow precursor cells. The most controversial question about early intrathymic precursors is whether they have B-cell as well as T-cell precursor activity. Because both B and T cells depend on the unique recombination activating gene (RAG) 1/RAG2–mediated receptor gene rearrangements and checkpoints in their development, it seems attractive to imagine that they are very closely related. In the adult, there are bone marrow common lymphoid precursors that can make colonies in vitro which differentiate selectively to B cells, NK cells, and T-cell precursors ( 50 ). Also, an evolutionarily conserved feature, across most jawed vertebrate classes, is the presence of at least some B cells as well as T cells in the thymus ( 51 , 52 ). The murine thymus itself includes a small number of B cells, as well as T cells, NK cells, and dendritic cells. Results of adoptive transfers and in vitro cultures of bulk cell populations have suggested that the B cells may arise from the same precursors that generate T cells ( 39 , 53 , 54 ). However, single-cell assays of precursors taken from the mouse or human thymus have not confirmed a close B-/T-cell relationship ( 33 , 45 ); the cells that give rise to B cells are not the same as the T-cell precursors, at least not after they have arrived in the thymus. In assays of single cells from fetal liver, cells that show B- and T-cell precursor activity generally also show myeloid precursor activity ( 41 , 55 , 56 ). One explanation seems to be provided by characteristics of the thymic microenvironment. Evidence suggests that many cells going to the thymus could differentiate efficiently into B cells in principle but that much of this activity is masked in the thymic environment by inhibitory signaling through the Notch-1 transmembrane receptor protein. Notch-1 signals are essential for T-cell development from the earliest detectable stages, and the thymic cortical microenvironment is evidently rich in Notch ligands. Notch ligand expression may be one of the conditions that makes the thymus (and some intestinal epithelial domains) uniquely permissive for T-cell development ( 57 , 58 ). This context is intensely suppressive of the development of B cells ( 58 , 59 and 60 ). Any of diverse strategies to inhibit Notch-1 activation allows production of thymic B cells at the expense of T cells ( 61 , 62 , 63 , 64 and 65 ). Thus, regardless of whether the cells populating the thymus have intrinsic B cell potential, this potential would normally be kept inoperative within the thymic cortex as long as the cells express Notch-1. The few B cells that do manage to reside in the thymus tend normally to be confined to the medulla; this may be a domain free of Notch ligands.

The cells gradually lose their developmental alternatives during the first T-lineage–differentiative transitions after precursor entry into the thymus. Single-cell analyses of fetal thymocytes show that many DN1 cells and a few DN2 cells can still differentiate into NK cells, and a large fraction of DN1-DN2 cells can give rise to dendritic cells ( 42 , 66 , 67 and 68 ). Under certain culture conditions, DN2 cells can even generate macrophages ( 69 ). The yield of these non–T cells is lower from DN2 thymocytes than from DN1 cells, and no B-cell potential is detected in the DN2 population. At the DN3 stage, the cells appear to have lost these residual alternatives and give rise only to T cells. The DN3 stage thus marks the completion of T-cell lineage commitment. Molecular Indices of T-Lineage Specification and Commitment Most genes associated with T-lineage development are already activated before commitment is complete. Some of these genes, such as CD3, Lck protein tyrosine kinase, and “sterile” transcripts from certain unrearranged Vß genes, are already expressed in DN1 cells ( 70 , 71 and 72 ). Most others are activated or up-regulated during the DN1?DN2 transition. Strikingly, cytokine and perforin genes used in responses of mature T cells are already expressed or inducible in the DN1 and DN2 cells. By the time the cells have reached the DN2 state, the cells are already expressing a full battery of T-lineage genes. We refer to these cells as “specified” for T-lineage differentiation. However, under appropriate conditions, as many as 50% to 75% of these cells ( 73 ) still remain capable of giving rise to dendritic cells in vitro. In the human system, committed dendritic cells continue to express the “T cell–specific” gene pTa ( 74 ). Thus, specification precedes commitment. Expression of genes associated with a particular cell type before actual commitment is not an anomaly in hematopoietic development. For example, single-cell reverse-transcription polymerase chain reaction assays have shown that uncommitted erythromyeloid precursors and stem cells individually express a multilineage gene profile ( 75 , 76 ). In fact, using green fluorescent protein transgenes under the control of the Rag1 and pTa regulatory sequences, two groups have shown that both of these DN2-stage associated genes can be expressed at a low level even longer before commitment. They are both active in bone marrow precursors that are still pluripotent, and which shut off pTa when they differentiate into B cells and shut off Rag1 when they become NK cells ( 76a, 76b). It is not surprising, therefore, that a population enriched for precursor activity within the DN1 subset expresses genes associated with non–T cell types as well as T cells, such as sterile transcripts of the immunoglobulin heavy chain, the macrophage colony–stimulating factor receptor, and the dendritic cell cytokine TARC ( 70 , 70a). These non–T genes are not expressed in the DN2 and DN3 stages. This is consistent with the evidence from erythromyeloid systems that lineage commitment involves repression of inappropriate genes, in addition to activation of lineage-specific genes. Both the onset of T-lineage gene expression and the shutting off of inappropriate genes are likely to be necessary for T-lineage differentiation. The regulatory genes that are necessary to guide T-cell development may contribute to

either or both of these mechanisms. Genetic Requirements for T-Lineage Specification and Commitment T-cell development can be impaired by mutations of genes that act in specification of precursors, in survival and expansion in the DN2/DN3 states, in ß selection, or in positive selection. None of the genes encoding TCR or pre-TCR components or the molecules that mediate TCR signaling cascades is needed before ß selection. Instead, this early period depends on genes encoding growth factor receptors, their associated signaling components, several key transcription factors, and Notch-1. The developmental transitions dependent on the action of these genes are summarized in Fig. 4, and the effects of some of the knockout mutations on thymus size and cell number are illustrated in Fig. 5.

FIG. 4. Stage-specific requirements for transcription factor and growth factor receptor genes in development of T cells from hematopoietic stem cells. The stages of development at which the indicated genes work are shown, on the basis of the effects of loss or gain of function experiments. Cell types are indicated in bold regular type and genes are indicated in italic type. The genes shown in the figure are discussed extensively in the text. The roles of c-Myb and Runx1 are critical for establishment of definitive-type stem cells, making it difficult to assay any later effects. Ikaros and PU.1 also affect stem cells, at least in postnatal mice. It is not clear exactly which prethymic precursor types are affected by loss of function of these genes, and so two hypothetical pathways are shown, one involving a lymphoid/myeloid precursor, another less characterized. The T-cell factor (TCF) or lymphoid enhancer factor (LEF) is shown primarily acting at the transition from immature single positive (ISP) to double positive (DP), the first stage when these related factors appear to be needed in young mouse thymocytes; however, in older mice, the loss of TCF causes arrest at the transition from double negative (DN) 1 to DN2, indicated here by (TCF/LEF) *. HSC, hematopoietic stem cell; E?, possible role for some E protein, E2A, or HEB or a relative.

The growth factor receptors that are needed in the early stage are IL-7Ra/? c and c-kit ( Fig. 4; cf. Fig. 2). Disruption of ? c, IL-7, IL-7Ra, or c-kit alone causes a decline in the number of viable DN2 and DN3 cells, but a few cells escape to undergo essentially normal differentiation from ß selection onward ( 3 , 4 , 5 and 6 ). Fig. 5A dramatically

illustrates the importance of the IL-7/IL-7R pathway before ß selection: If thymocytes that cannot undergo ß selection, such as RAG2 -/- cells, attempt to differentiate in an IL-7–deficient environment, the thymus remains minute, much smaller than in the mutants of either the IL-7 or the RAG2 gene alone. Thus, the proliferation occurring at ß selection may compensate for a shortage of precursors. The survival of any thymocytes to this point, in the absence of IL-7R signals, appears to be due to the ability of c-kit/stem cell factor interactions to sustain a few cells long enough for them to make a productive TCRß gene rearrangement. Double mutation of both c-kit and the common ? subunit of the IL-7R prevents any detectable lymphoid precursors from appearing in the thymus ( 77 ). Growth factor receptors can be important to development both to provide survival/proliferation signals and, in some cases, to provide “instructive” signals to initiate a cell type–specific gene expression program. For T cell development in general, the roles of c-kit and the IL-7R complex can be explained entirely in terms of survival/proliferation. The only exception is that IL-7R receptor signaling may influence the direction of differentiation in the choice between TCRaß and TCR?d lineages, as discussed later. The generation of precursors with the competence to become T cells depends on regulatory genes that include those coding for the transcription factors Ikaros, PU.1, c-Myb, and GATA3 ( Fig. 4 and Fig. 5) ( 78 , 79 and 80 ). Ikaros, c-Myb, and GATA3 have all been shown to bind to specific target sites in T-cell differentiation genes [TCRa, TCRß, TCR?, TCRd, CD3d, and terminal deoxynucleotide transferase (TdT)], but mutation of any of these genes blocks thymocyte development long before any of these genes is required. PU.1 is not known to be required for any T-cell gene expression, but it may be used specifically in the precursors seeding the thymus in fetal life; loss of PU.1 eliminates all fetal T-cell development. The effects of PU.1 and Ikaros loss-of-function mutations (Ikaros C -/- in Fig. 5B) are quite leaky in postnatal T-cell development, in contrast to fetal T-cell development. This may simply be caused by transcription factor gene redundancy in the postnatal case, rather than true independence of these factors. Both the PU.1 and Ikaros genes are members of small families of related genes with overlapping expression in the postnatal thymus. Ikaros dominant-negative mutations that interfere with all family members result in complete ablation of T-cell development in the adult thymus as well as the fetal thymus. The genes regulated by Ikaros, PU.1, c-Myb, and GATA3 that are critical for early T-cell precursor function remain to be defined ( 78 , 79 , 81 ). An additional set of transcription factors is required during the specification and expansion events in the DN2 and DN3 stage. Here, powerful quantitative effects result from loss-of-function mutations of the basic helix-loop-helix (bHLH) transcription factor E2A or the bHLH repressor Hes-1 ( 82 , 83 ). E2A gene products form heterodimers with related bHLH factors to help drive expression of the pTa and RAG1 genes ( 84 , 85 and 86 ), and this probably accounts for part of the E2A knockout effect. However, the effect of the E2A knockout is most severe in the DN2 and DN3 populations preceding ß selection, before any of these target genes are needed. Thus, it is likely that additional target genes are involved. In the E2A mutants, T-lineage specification per se is not completely blocked; cells that manage to express some form of pre-TCR can undergo ß selection and continue their development. The incomplete effect of E2A mutations may result from some overlap in function with the related bHLH factor HEB, which is

expressed at high levels in T-cell development. It now appears that E2A carries out unique survival functions in lymphoid precursors, complementary to those mediated by IL-7R ( 87 ). These roles in survival of early precursors may explain some of the similarities between the severe early phenotypes of E2A knockouts and those of IL-7R component knockouts. Hes-1 is a transcription factor that is directly induced by Notch signaling. It is therefore a potential component of the mechanism used by the Notch pathway to block B-cell development, enforce T-cell development of lymphoid precursors, or both ( 88 ). The mechanism that blocks progression through the DN2?DN3 stages in Hes-1 mutants also causes more severe effects on subsequent development than in E2A mutants, which is consistent with a direct effect on T-lineage specification. On the other hand, the block is somewhat leaky ( 83 , 89 ). It is possible that complementation by another Hes-related gene is responsible for the ability of a few cells to do without Hes-1, but it may also be that Hes-1 primarily provides a survival function. Mice with the Hes-1 mutation do not survive after birth, and their phenotype has been analyzed thus far only in the fetal thymus. Figure 4 includes a number of factors that are most prominent at later stages, after T-lineage commitment. These include the bHLH transcription factor HEB, a relative of E2A, and the high-mobility group (HMG) box transcription factors T-cell factor 1 (TCF-1) and/or lymphoid enhancer factor (LEF), which have important roles in ß selection. These are not absolutely required in earlier T-lineage specification events, but they probably participate to some extent; for example, HEB can provide a modest compensation for loss of E2A. As shown in Fig. 4, TCF-1 and LEF are not needed until after T-lineage commitment in fetal and young postnatal thymocytes, but the cohorts of precursors that populate the adult thymus need TCF-1 for earlier events, at the DN1-to-DN2 transition ( 25 ). Also shown in Fig. 4 is that transcription factors of the Egr family and an antagonist of bHLH positive regulators, Id3, have major roles later, in the TCR-dependent events of T-cell development, as discussed later. Transcription factors are central players in establishing T-cell identity, for these molecules not only control the differentiation process but also enable the cell to maintain its characteristic pattern of gene expression once its differentiation is complete. The need for contributions from multiple factors is a normal consequence of the way this initial lineage choice works. It is probably a mistake to search for a gene that controls T-cell specification as such. Instead, these regulatory genes each appear to influence distinct developmental choices. The Notch pathway, acting through Hes-1 and other mediators, influences the ability to become a T cell versus a B cell but does not substantially affect the NK or dendritic cell developmental choices ( 90 , 91 ). In contrast, E2A and the genes of its family are crucial for the choice of either T- or B-cell development, as opposed to NK cell or myeloid/macrophage development ( 92 ). T-cell specification is likely to be defined by the combination of regulatory factors that permit lineage progression as they jointly eliminate all other developmental options ( 80 ). Although none of them appear to be “master regulators” of T-cell identity, these genes are used throughout T-cell development. Notch family genes may participate in as many

as three lineage choices within the T-cell pathway. E2A family and Id genes act in ß selection and positive selection ( 93 ). GATA-3 influences not only the CD4/CD8 choice but also the postthymic differentiation of T H1 and T H2 effector subsets of helper T cells ( 94 ). Even Ikaros recurs in specific roles in later T-cell development. Thus, the interplay of these developmentally potent factors is a permanent feature of the T-cell regulatory apparatus, which may help to give mature T cells some of their richly nuanced repertoire of responses to their environment.

A REGULATORY UPHEAVAL: ß-SELECTION Multiple Changes at the Transition from T-Cell Receptor–Independent to T-Cell Receptor–Dependent T-Cell Development The ß-selection process is a watershed in T-cell development that marks the change from events dominated by hematopoietic-like mechanisms to events dominated by TCR interactions with the microenvironment. Understanding of the ß-selection process advanced dramatically in just a few years, and several excellent reviews discuss this event in detail ( 95 , 96 , 97 and 98 ). As an immunological event, it has a significant impact on the eventual T-cell repertoire, but it is also fascinating as a complex, multistep cellular response triggered by a particularly well-studied signaling event. After 2 weeks of TCR-independent growth, ß selection suddenly polarizes the fates of cells that have succeeded or failed at ß-chain rearrangement: The successful ones are rewarded with proliferation and differentiation, and the failures are killed. It thus reflects the imposition of a novel criterion of viability, wresting the cells from their simple “hematopoietic” survival functions and making their futures TCR dependent. The proliferation it triggers sets up the large population of TCRß + cells that is needed to provide enough TCRß diversity and cell numbers so that stringent positive/negative selection criteria can be applied later. This mitotic burst may also include some of the last cell cycles that T-cell precursors undergo, in adults, before they finish development and emerge to the periphery. ß Selection is interesting, overall, in terms of the sweeping regulatory changes it brings about in cell physiology, the relationships among the intricate cascade of processes that it triggers, and its distinctive features in comparison with other TCR-dependent activation responses. A summary of changes occurring in cells during ß selection is presented in Fig. 6 and discussed in the next section. There is, first, a powerful burst of proliferation, beginning before the cells change their DN3 phenotype ( 1 ). Cells responding to ß selection were originally measured to have uncommonly fast cell cycles of only approximately 8 to 9 hours, with minimal G 1 and G 2 phases [reviewed by Rothenberg ( 99 )]. This proliferation is apparently associated with a requirement for new survival functions. There are also changes in cell surface phenotype, with gains of CD8 and CD4 expression and loss of CD25 expression (clearance of CD25 from the cell surface is aided by dilution as the cells proliferate; some CD25 persists in cases in which proliferation is partially blocked). Within the cell, TCRß, TCR?, and TCRd genes become inaccessible to further rearrangement; and in a separable event, TCRa genes

begin to be transcribed and become accessible for rearrangement for the first time.

FIG. 6. Transformations of cell phenotype during ß selection: comparison with later positive selection events. The changes in gene expression, rearrangement accessibility, and cell-surface phenotype at the transition from double negative (DN) 3 cells to double positive (DP) are compared with changes that occur later during positive selection. Changes in cell-surface phenotype where several distinct levels of marker expression are useful to distinguish among developmental states are distinguished by sloping or stepped forms, whereas others are simplified as all-or-none changes. For discussion, see text. Proliferation is shown as starting at a low level on the basis of the properties of DN3 cells that have not yet undergone productive T-cell receptor (TCR) ß gene rearrangement (there is extensive proliferation in earlier DN subsets). Proliferation continues in mouse thymocytes through the DN4 and immature single positive (ISP) stages and into the beginning of the DP stage. CD4 and CD8 expression patterns are shown bifurcating at positive selection to represent the CD4/CD8 single positive (SP) lineage split. Recombination activating gene (RAG) 1 and RAG2 and pTa are transiently shut off during proliferation after ß selection and then expressed again in DP thymocytes. AP-1 loss of function in DP thymocytes is a loss of inducibility of deoxyribonucleic acid (DNA) binding and transactivation activity, leading to broad defects in effector gene inducibility. NF-?B up-regulation is an elevated constitutive level of binding activity. The expression of surface glycoproteins deficient in sialic acid is detected by a sharp increase in binding to the lectin peanut agglutinin (PNA). Fas is a tumor necrosis factor (TNF) receptor family death receptor.

The cells need new survival functions because they undergo rapid changes in their intrinsic survival potential and function that thrust them into a highly vulnerable state ( 100 ). This is associated with the loss of Bcl-2 expression, the onset of expression of the

proapoptotic surface receptor Fas, and the replacement of Bcl-2 survival functions by Bcl-X L and NF-?B ( 101 , 102 , 103 and 104 ). In addition, changes in signaling physiology disable activation of normal AP-1 transcription factor, paralyzing cytokine gene expression and other functional responses that had been established previously in DN cells ( 105 , 106 , 107 and 108 ). At the same time, however, the cells paradoxically become more sensitive to interactions with cell-bound ligands and particularly hypersensitive to low-affinity TCR ligands ( 109 , 110 and 111 ), apparently because of an abrupt change in membrane glycoprotein processing ( 112 , 113 ). The suite of these processes results in DP thymocytes that are suspended in a state close to death and are easy to kill and yet are also uniquely capable of detecting encounters with low-avidity ligands for any TCR complexes they may form after TCRa gene rearrangement. These properties are exploited to the full, in the next 3 to 4 days, for the purposes of TCR repertoire selection. Triggering Requirements for ß Selection ß selection is triggered when a TCRß gene rearrangement generates a sequence that can be translated into a ß-chain protein. It was a mystery for a number of years how this response could occur so efficiently for diverse ß chains, regardless of the binding specificities of their V regions. It now appears that this is the one case in which a TCR-like complex can undergo ligand-independent signaling. The ß chain assembles into a pre-TCR complex with the surrogate a chain pTa and the CD3 components ?, d, e, and ? 2 , which are already being expressed in DN2 and DN3 cells ( 114 , 115 , 116 and 117 ), and thus enters the traffic to the plasma membrane. pTa, a key component of this complex, is an invariant transmembrane glycoprotein that is encoded by a nonrearranging immunoglobulin superfamily gene and expressed very specifically in DN and DP thymocytes ( 115 , 118 ). The TCRß/pTa/CD3 complexes are segregated efficiently into cholesterol-rich lipid microdomains (lipid rafts) on the cell membrane ( 119 ). This is apparently possible because of distinctive features of the pTa transmembrane and submembrane regions, inasmuch as neither conventional TCRaß nor TCR?d complexes appear to partition to the rafts so effectively ( 119 ). Even at low levels of cell surface expression, the pre-TCR complexes cluster with each other in these rafts. The special organization or clustering of these rafts may depend on the Rho/Rac family of guanosine triphosphate (GTP)–binding proteins, which are also required for ß selection ( 120 ). This clustering spontaneously triggers a potent signaling cascade that engages the kinase Lck, the adaptors SLP-76 and LAT, and the activation of protein kinase C, Ras, and MAP kinases to launch the complex ß selection response. Figure 7 shows the mediators and pathways that have been implicated in this signaling response, in comparison with those triggered by the mature TCR in positive selection.

FIG. 7. Signaling cascades activated by pre–T-cell receptor (TCR) signaling during ß selection: comparison with cascades activated by TCRaß complexes in positive selection. A simplified schematic is presented to show relationships discussed in the text. The figure focuses on mediators that are seen to be activated in ß selection and positive selection. Those that appear to be essential for these transitions, on the basis of loss or gain of function experiments, are emphasized by underlining. There are certain to be other key mediators as well. Some differences are seen between the two activation processes. In ß selection, there appears to be some redundancy in the roles for related kinases Lck and Fyn and for ZAP70 and Syk, whereas both Lck and ZAP70 are essential for positive selection. Lck cannot be brought to the pre-TCR by CD4 in ß selection, at least not in mice, which do not express CD4 at this stage (in humans, however, CD4 is expressed before ß selection), but it probably does depend on CD4 to bring it to TCRaß in positive selection. In spite of these differences, there is striking similarity between the two signaling cascades overall. In positive selection, several mediators that are essential in ß selection are not underlined, only because the double positive (DP) thymocytes that are needed to undergo positive selection are not generated when these mediators are absent.

Substantial efforts have been made to find an extracellular ligand that may engage the pre-TCR. Some cross-linking stimulus was thought to be needed because, for years, one of the most effective ways known to induce a wave of ß selection in vivo or in vitro has been to cross-link the sparse CD3e-containing complexes on the surface of RAG-deficient thymocytes ( 121 , 122 ). However, it currently appears that in the normal case, no ligand is needed. Mutant forms of TCRß and of pTa that completely lack extracellular immunoglobulin-like domains can mediate ß selection in place of wild-type forms, at least when expressed as transgenes ( 123 , 124 ). Furthermore, pTa itself does not seem to mediate specific interactions with any distinctive set of signaling molecules, inasmuch as a form of pTa lacking most of its cytoplasmic domain is also capable of complementing a pTa deficiency. These structural perturbation results support the interpretation that assembly of the pTa::TCRß::CD3 complex in lipid rafts itself is sufficient to trigger the signaling that leads to ß selection. The TCRa locus under normal conditions is neither rearranged nor transcribed appreciably until ß selection. However, already-rearranged TCRa transgenes can be expressed early enough to enable the effects of TCRaß complexes to be compared with the effects of pre-TCR (pTa:TCRß) complexes at this transition. Up to a point, rearranged TCRa can replace pTa to mediate ß selection. It certainly supports the differentiative changes induced during ß selection, such as shutoff of CD25 expression and onset of CD4 and CD8 expression, together with a certain amount of proliferation.

TCRaß transgenic mice are capable of generating a DP population that is similar, in most respects, to the DP population in normal mice. Also, pTa -/- mutant mice do generate a small number of DP and SP cells in spite of their inability to make pre-TCR. Such cells appear to be generated through TCRaß signaling, apparently because a few cells initiate TCRa rearrangement precociously. However, there is evidence that TCRaß complexes are less effective than pre-TCR complexes in triggering the major proliferative expansion that normally occurs at ß selection ( 125 ). Constituent Events in the ß Selection Cascade Gene disruption and overexpression experiments show that the constituent events in ß selection can be dissociated from one another. The whole process is therefore a short differentiation program rather than a single response to triggering. Components with distinct genetic requirements can be resolved in several ways: (a) early proliferation, with short-term protection from apoptosis (DN?ISP); (b) later proliferation (ISP?DP); (c) CD4, CD8, TCRa transcriptional activation (and CD25 down-regulation); (d) transient down-regulation of RAG1, RAG2, and pTa; (e) allelic exclusion (i.e., long-term shutoff of Vß and V? rearrangement); and (f) antiapoptotic functions for ISP?DP cells. The roles of different genes in the process are shown vividly by the defects in or blockade of ß selection when they are mutated, or by the ability to bypass the ß-selection checkpoint when they are overexpressed. The differentiation program from DN3 to DP can be triggered efficiently in the absence of pre-TCR expression by activated forms of signaling molecules that are normally mobilized by pre-TCR: activated Lck, activated Ras, or activated Raf ( 126 , 127 , 128 , 129 and 130 ). Any of these mediators allows cells to be generated with the CD4 +CD8 +CD25 - phenotype of DP cells. Examples of these effects are shown in Fig. 8A and B. So far, only certain parts of the mechanism connecting the signaling events to the specific differentiation changes (events c and d) are understood. The immediate-early response transcription factors of the Egr family are prominently activated during ß selection, and these are probably responsible for inducing TCRa germline transcription and the shutoff of RAG1, RAG2, and pTa ( 131 ). However, the transcriptional mechanisms underlying CD4 and CD8 induction are still being studied ( 132 , 132a). Expression of CD4 and CD8 is controlled by complex positively and negatively acting factors, and those acting at ß-selection appear to include some combination of Ikaros, bHLH factors, and Runx transcription factors as well as chromatin remodeling complexes ( 132b, 132c, 132d ). None of these are known to be activated directly by pre-TCR signaling. Regulators of the other major changes resulting in DP cell properties are even less defined.

FIG. 8. Bypassing the ß-selection checkpoint by Ras pathway activation or antagonism of FADD. A: The number and CD4/CD8 expression pattern of thymocytes from recombination activation gene (RAG) 2 -/- mice ( left) are compared with those of thymocytes from RAG2 -/- transgenic mice expressing a constitutively activated Ras transgene (Ras V12) in the thymus. The elevated Ras activity causes a 100-fold expansion of cell numbers and a complete conversion from double negative (DN) to double positive (DP). The cells cannot progress further to CD4 or CD8 SP stages, because they still lack T-cell receptor (TCR) rearrangements and cannot be positively selected. From ( 130 ), with permission. B: A similar experiment is shown in which RAG2 -/mice were bred to express an activated Raf transgene (Raf-CAAX) in the thymus. Activated Raf alone has modest effects on normal thymocyte populations (compare Raf-CAAX, littermate control) but completely transforms the populations seen in RAG2 -/mice (compare RAG2 -/-, RAG2 -/- with Raf-CAAX). From ref. ( 129 ), with permission. C: Not only activation of positive mediators but also competitive antagonism of a checkpoint enforcement function can allow the cells to break through ß selection without a pre-TCR. Here a transgene encoding a dominant negative variant of FADD, which interferes with the normal functions of FADD, causes the appearance of DP and CD8 immature single positive (ISP) thymocytes when expressed in RAG1 -/- thymocytes that would otherwise be blocked at the DN stage. The rescue is not complete (contrast wild-type control, left) but very pronounced in view of the fact that differentiation in this case occurs with little proliferation. From ( 147 ), with permission.

It is easy to explain the termination of the typical activities of DN2/DN3 thymocytes at this point by the shutoff of numerous regulatory genes ( Fig. 6). The Ras signaling pathway can cause some of these changes itself, by activating the bHLH antagonist Id3, which is turned on by Egr-1 ( 133 ). When expressed any earlier, Id3 is a complete inhibitor of the entry of lymphoid precursors into the T lineage and a specific antagonist of the functions needed to prepare cells for ß selection ( 134 , 135 ). Nothing could more emphasize how ß selection terminates the immature stages of T-cell development than

the activation of Id3 at this point. Strangely, some of the obvious differentiative events (c and d) may not really depend on pre-TCR triggering at all. They may be events that the cells have been programmed to undergo during the DN3 stage but with a potent threat attached: that going forward from the DN3 state will lead to certain, rapid death unless a number of protective mechanisms are engaged. One look at the completeness of the block in RAG-deficient thymocytes ( Fig. 1) might make this seem unlikely. However, that block to differentiation is dissipated substantially by a single genetic change: the mutational inactivation of the p53 tumor suppressor gene ( 136 , 137 and 138 ). This p53 gene product has many roles in cell biology, one of which is to impose G 1 arrest and another of which is to induce apoptosis under certain conditions, such as in case of deoxyribonucleic acid (DNA) damage. In pre-TCR–deficient thymocytes (e.g., RAG-knockout, pTa-knockout), the normal role of p53 appears to be mainly to punish differentiation with death, because if p53 function is removed, these cells develop efficiently into DP cells. In certain experimental situations, the death-dealing function of p53 at the ß-selection checkpoint can also be counteracted by signals from the microenvironment ( 136 , 137 , 139 , 140 ). 1 Thus, although some uncertainty remains about how pre-TCR assembly causes differentiation, there is no question that it causes major shifts in susceptibility to apoptosis and proliferation. Proliferation has to be induced and is not just a default, because when p53 is mutated, the differentiating cells remain limited in their ability to proliferate. What is needed to unleash this proliferation may include the removal of a specific brake: the antiproliferative adaptor molecule SOCS-1, which is highly expressed in DN3 cells and is abruptly shut off by pre-TCR triggering ( 141 ). Then, the roles of a succession of transcription factors distinguish early and late stages of ß-selection-associated proliferation (events a and b). Early proliferation (event a) appears to involve the Egr family transcription factors ( 131 , 142 ) and the HMG box transcription factor TCF-1 or LEF, or both ( 25 , 143 /SUP> ). Later proliferation in the transition from ISP to DP (event b) can no longer be driven by Egr activation. Now, the cells depend acutely on TCF/LEF plus ß-catenin (




). The

discontinuity between the two phases may be caused by a switch in the n eed for bHLH transcription factors. In the first stage, proliferation is aided by the bHLH transcription factor antagonist Id3 (


). In the

second phase, proliferation and differentiation depend on one of the molecules that Id3 should antagonize, the bHLH transcriptional activator HEB (



These stage-specific regulators work in collaboration with additional proliferative functions that may have broader roles. One is the proto-oncogene c-Myb, activated at ß-selection by Pim-1 kinase (

). Yet another participant driving proliferation during ß selection is a fascinating one in terms of checkpoint control: the bifunctional mediator FADD (


, 147 ). FADD is an adaptor molecule transducing tumor necrosis factor (TNF) receptor family death signals, but it also seems to be vital for proliferation at 146

ß-selection. It is discussed further later. The ability of pre-TCR/Lck complexes to initiate proliferation is probably mediated by the Ras pathway,

inasmuch as this is sufficient to induce the mitogenic Egr (and Id3) family molecules. However, allelic exclusion (event e) has to be triggered through some mediator other than Ras, because activated Ras alone cannot close the TCRß-chain genes to further rearrangement (

, 130 ). Protein kinase C activity is also induced by pre-TCR signaling, upstream or parallel with Ras, and it has been suggested that that 129

this is the mediator that is responsible for termination of TCRß rearrangement (



Antiapoptotic functions (event f) need to be activated during ß selection because of two kinds of threat: a general loss of survival support, which appears to be a default for DN3 cells, and a differentiation-linked susceptibility to apoptosis. Beyond the DN2/DN3 state, most thymocytes apparently lose their responsiveness to the IL-7R signaling that, until then, has supported most of their proliferation and kept them from death by inducing Bcl-2 gene expression. Furthermore, mitogenic stimulation itself can be risky. As noted previously, one of the molecules stimulating proliferation of DN3 cells, FADD, can also promote apoptosis. Two transcription factors, NF-?B/Rel ( 101 ) and activated TCF/LEF ( 11 , 12 ), appear to turn on the genes that protect cells from apoptosis. NF-?B appears to collaborate with TCF/LEF factors to turn on specific molecules that uncouple FADD from its apoptotic signaling cascade and restrict its effects to promoting growth ( 149 ). NF-?B and TCF/LEF factors activated by ß-catenin can also help to turn on expression of Bcl-X ( Fig. 6), compensating in part for the shutoff of Bcl-2 ( 102 , 103 , 150 ). L

Death Mechanisms and Other Checkpoint Controls Under normal circ*mstances, DN3 thymocytes are probably prevented from spontaneous growth and differentiation by at least three mechanisms. Their ability to grow in response to IL-7R has been arrested by the high levels of the antiproliferative molecule SOCS-1 that they accumulate (

141 ) and probably by a gradual down-regulation of IL-7Ra expression as well. They do not generally express substantial levels of

p53, but through some mechanism not yet understood, any attempt they may make to differentiate is linked with p53 activation, which kills them. Furthermore, FADD plays a role. In addition to its role promoting growth, FADD function is needed to prevent differentiation to DP cells. When a dominant-negative FADD transgene is expressed in cells that cannot make pre-TCR, these cells break through the checkpoint and generate DP cells without TCR ( Fig. 8C). FADD is particularly interesting because it is not only expressed but also required to function when the cells receive their ß-selection signal ( 147 ). It is expressed in concert with multiple receptors of the TNF receptor family that are linked with cell death, although the exact set of receptors expressed shifts from DN3 to DP stages (

). Thus, the pathway involving FADD has the opportunity to be triggered directly from the DN3 stage. There are 147

complex relationships between p53 and the TNF receptor family death receptors, and this may account for thei r interconnected roles. In addition to the death functions enforcing the checkpoint, there are a number of threshold-setting functions that appear to determine the magnitude of signal strength that will be needed to trigger ß selection. One is Csk, the kinase that inhibits Src-family tyrosine kinases, such as Lck, by

phosphorylating their C termini. Mutation of the csk gene allows Lck to be active constitutively, and the result is that cells without pre-TCR can spontaneously differentiate to DP cells (



Another gene product that appears to regulate stimulation thresholds turns out, surprisingly, to be the transcription factor Ikaros. Ikaros is a crucially important gene for all lymphocyte development (

152 ), but it has been difficult to establish the nature of any target genes that it regulates positively; CD8 may be the

first ( 132b). However, striking circ*mstantial evidence has associated Ikaros binding with the silencing of genes in the course of lymphoid development ( 153 ). Thus, it is noteworthy that Ikaros-mutant heterozygous mice show a dramatic breakthrough of pre-TCR–negative cells into the DP stage. Effects on the CD4 and CD8 genes themselves could be involved ( 132b), but this violation of the ß-selection checkpoint, as in the case of Csk deficiency, is associated with a general T-cell hyper-reactivity (


). In

the Ikaros+/- thymus, hyperreactivity is also a prelude to malignant transformation: These animals develop thymic lymphomas at a very high frequency. The gene dosage effect implies that Ikaros levels are tightly correlated with a precise regulator of activation. Significance of ß Selection for Later T-Cell Differentiation As a developmental event, ß-selection is momentous. The approximately 10 2-fold proliferation at ß selection effectively erases the developmental alternatives for TCRaß cells, consummating not only T-lineage commitment but the separation of aß and ?d cell fates. In adult mammals, this is the last significant proliferation that T-cell precursors undergo before being exported to the periphery. Thus, the form in which cells emerge from the various stages of ß selection dictates the defaults for their responses to positive and negative selection signals. Overall, ß selection is a form of activation both in terms of specific gene expression and in terms of its use of stimulatory signaling cascades to trigger proliferation. However, transcriptional repression also seems to participate in differentiation or survival at this stage. The nuclear receptor co-repressor gene N-CoR turns out to be essential for ß selection, because N-CoR -/- mutants are blocked at the DN stage ( 155 ). The target genes that may need to be repressed by N-CoR in the course of ß selection have not yet been identified. However, several transcription factor messenger ribonucleic acids (RNAs) are abruptly silenced in T-lineage precursors during ß selection (

81 , 156 ), which suggests that this could represent a profound upheaval in gene-regulation potential as well as a major physiological change.

In certain ways, the immediate impact of ß selection on the cells is the reverse of the impact of positive selection (see later discussion). In particular, the changes in Bcl-2, Bcl-X , NF-?B, and AP-1 activation, L

glycoprotein sialylation, and signaling thresholds will all be reversed when the cells are positively selected ( Fig. 6). Nevertheless, many of the triggering functions used in ß selection are the same as those used in positive selection ( Fig. 7). Both involve triggering via Lck, SLP-76, LAT, Ras, Raf, Erk, and protein kinase C. Both involve Ca 2+ fluxes and NF-AT transcription factor activation, at least as inferred from

cyclosporine sensitivity (


, 158 and 159 ), as well as induction of Egr family genes and Id3 ( 93 , 131 ,



). The signals may not be instructive but seem to act more as a toggle between alternative physiological states. The same threshold setting functions that limit activation in ß selection are also used 160

again to limit activation in positive selection. Thus, the threshold -setting functions that act at ß selection may turn out to be an important immunological legacy of the process. Of interest is whether the levels of threshold-setting function present in particular cells that undergo ß selection, such as Csk and Ikaros, could be maintained through proliferation and into the DP population. If so, the positive/negative selection thresholds for individual DP cells could depend on the strength of the pre-TCR signals that triggered their ß selection initially.

THE DIVERGENCE OF T-CELL RECEPTOR aß AND T-CELL RECEPTOR ?d LINEAGE CELLS Choices of Fate within the T-Cell Lineage: Differences between aß and ?d T Cells Cells committed to the T-cell lineage continue to make additional developmental choices as to what kind of T cell they will be. T cells that use TCRaß receptors differ in a number of respects from T cells that use TCR?d receptors. At later stages, CD4 SP and CD8 SP T cells exhibit divergent functional properties, and there is increasing evidence that another class of T cells, the NKT cells, represents yet a further discrete lineage. The intrathymic choice between TCRaß and TCR?d fate, like the choice between CD4 + and CD8 + fates, remains controversial because of two problems. The first problem is that the alternative fates are still relatively poorly defined in terms of multiple, independently measurable traits. As long as TCR?d complexes themselves (and failure to acquire CD4 and CD8) are the only clear markers for the ?d cell fate, it is difficult to analyze the role of these complexes in bringing about this fate. The second problem is that the TCR structures on the cells making TCR?d versus TCRaß lineage choices are different in recognition specificity and other properties. Because of this, some of the behavioral differences between subsets could be caused by responses to differential TCR signaling, making it uncertain how much the cells may differ intrinsically. There are two kinds of TCR?d cells that may be produced through distinct pathways. In mice and chickens, at least, the ?d cells appearing first in ontogeny (before birth or hatching) display properties that distinguish them both from adult-type ?d cells and from aß cells. For these early ?d cells, a case can be made that a fundamentally different developmental program is used for these cells and that they arise from a cell type intrinsically different from TCRaß precursors. A summary of the properties of the early ?d, later ?d, and aß classes is presented in Table 2.

TABLE 2. Properties of TCRaß and TCR?d cells: fetal versus postnatal

For adult-type TCR?d cells, the evidence that the precursors are intrinsically different from TCRaß precursors is shakier. In the periphery, TCR?d cells continue to carry out surveillance assignments that are different from those of TCRaß cells, although using many or all of the same effector functions as those used by aß cells. A general difference is that TCR?d cells in mice do not express CD4 or CD8ß, in contrast to TCRaß cells. Because the cytoplasmic tails of CD4 and CD8 are the major known docking sites for Lck, their absence probably alters the way mature TCR?d cells can recruit Lck to lipid rafts with the TCR during antigen recognition. This could have multiple consequences for signaling and could contribute to the distinctive functions of TCR?d cells in the periphery. Developmentally, such differences originate with the separation between TCR?d and precursors of the TCRaß cells in the thymus; this is because TCR?d cells do not go through the full ß-selection process. Most prominently, TCR?d cells are T cells that have succeeded in making both V-J? and V-D-J? gene rearrangements productively before they die or undergo complete ß selection. The TCR?d receptor is both their main distinguishing feature and their apparent cause of divergence from the TCRaß path. In general, the rearrangements of TCR? and nondeleted d genes in TCRaß cells are out of frame for protein translation ( 161 ,


) (see Chapter 8). On the other hand, there are some in-frame TCRß rearrangements

in ?d cells. These data can be interpreted to mean that cells keep “trying” to become TCR?d cells by default; if they fail, they die unless they have been rescued by TCRß rearrangement and ß selection. The first indication that the TCRaß and TCR?d differentiation programs can be separated from the use of these receptors came from analysis of TCRß -/- mice. In these animals, DP thymocytes were generated in small numbers, even though ß-selection as such could not take place (

). The DP thymocytes in this case used TCR?d receptors and could be generated only if the TCR? and d genes were intact. However, 163

DP thymocytes do not use TCR?d receptors in normal mice. A careful analysis has shown that in several transgenic and knockout cases, TCR?d receptors can apparently support development of cells with aß-type characteristics when TCRaß is unavailable, and vice versa (

, 165 ). This is important evidence that cells are assigned to discrete differentiation programs that in some cases can be mismatched with 164

the TCR class they express. In support of this view, researchers have identified several genes that can bias the lineage choice of developing cells to an aß-like or ?d-like program independently of the TCR they express, as described in the following sections. Generation of T-Cell Receptor ?d Cells

For adult-type ?d cells, separation from the TCRaß pathway occurs after the DN2 stage (

166 , 167 ). Successful rearrangement of both TCR? and TCRd genes rapidly down-regulates CD25 expression and

leads to generation of CD25 - CD44 low/int TCR?d + cells. These cells subsequently down-regulate CD24 (HSA) as they complete their maturation. There are multiple points of contrast with the aß pathway. First, there is little or no proliferation associated with TCR?d development, in sharp contrast to ß selection. Second, genes such as TCF-1 and HEB, which are required for completion of ß selection, are dispensable for TCR?d development (

, 144 ). Third, the ?d program does not involve up-regulation of CD4 or CD8ß, and most ?d cells lack CD8a as well. TCR?d cells use Lck for full maturation, but mutation 143

of Lck has little effect on ?d cell numbers (



The lack of proliferation in TCR?d cell development makes it easy to underestimate the percentage of intrathymic precursors taking this path. In steady state, TCR?d + cells constitute only 1% to 2% of thymocytes and circulating peripheral T cells in mice and humans, and so it is easy to regard them as a minor cell type. However, correcting for the approximately 100-fold expansion occurring in TCRaß precursors at ß selection, it can be argued that the absolute number of DN2 cells that will give rise to ?d cells is similar to the absolute number that will give rise to aß cells. What is the relationship between the aß versus ?d lineage choice and ß selection? The TCR?d complex does not usually trigger ß selection, but how its assembly differs functionally from those of the pre-TCR complex, on one hand, and the TCRaß complex, on the other hand, is nonetheless an open question. Some evidence shows that TCR?d is poorer than pre-TCR at spontaneous self-clustering in the pre–T cell membrane (

). This could explain its poorer activity in ß selection. But the ability to generate TCR?d DP cells under conditions in which there is no pre-TCR shows that these receptors can mediate 119

certain aspects of ß selection, at least the antiapoptotic ones. Also, the initial steps in ß selection do not necessarily block the generation of ?d cells. Several groups have found TCRß rearrangements that are apparently in frame in some TCR?d thymocytes ( 162 ,

), and it has been argued that they are enriched above the level expected for random occurrence. Thus, there is a suggestion that some cells can 169

undergo at least some ß-selection–linked clonal expansion and still go on to differentiate into TCR?d cells. These kinds of evidence tend to argue that there is an essential mechanism underlying TCRaß and TCR?d lineage divergence that is different from simple success or failure at triggering ß selection. Most precursors of ?d lineage cells separate from the aß lineage earlier, before the ß-selection checkpoint. As early as the DN2 stage, two populations can be distinguished on the basis of their levels of cell surface markers: one that can give rise to ?d cells as well as aß cells, and one that is mostly or entirely restricted to the aß lineage (


). The cells that retain ?d potential are the highest in IL-7Ra

expression at the DN2 stage and low in pTa/pre-TCR surface expression at the DN3 stage (





Both IL-7R and pTa/pre-TCR could participate in instructive signaling, but the very fact that their expression levels are heterogeneous provides evidence for additional, underlying regulatory differences

that foreshadow the aß/?d lineage choice. Genetic Regulation of T-Cell Receptor aß versus T-Cell Receptor ?d Cell Production Several genes appear to affect TCRaß versus TCR?d lineage choices or lineage-specific survival functions. Their roles appear to be distinct from those that alter the TCRaß:TCR?d ratio simply by enhancing or limiting the extent of proliferation at ß selection. Some of these genes may primarily influence the choice between TCRaß and TCR?d fates in general, whereas others may actually affect the choice between adult-type and first-wave fetal-type T-cell development. Particular growth factor receptor genes are disproportionately important for TCR?d cell production. First-wave fetal TCR?d thymocytes are greatly reduced by mutation or blockade of IL-2Rß/IL-15Rß (CD122), whereas other T-cell subsets are minimally affected (

172 , 173 ) ( Table 2). This correlates with the preferential expansion of these early cells in fetal thymic organ culture in response to moderate doses

of IL-2 or IL-15 (

). Adult-type ?d thymocytes do not have this special response to IL-2 or IL-15, but they require IL-7/IL-7R interactions. More broadly, IL-7Ra (CD127) expression appears to be essential for 174

all TCR?d cell development, both fetal and adult type. The common cytokine receptor chain ?c (CD132) is a component of both receptors and is essential for all TCR?d cell development (

, 176 ). This growth factor receptor dependence has two aspects. One is simply proliferative: Because TCR?d cells cannot 175

expand through ß selection, the earlier IL-7–driven phases of their proliferation account for almost all the TCR?d cells produced during differentiation. If IL-7–driven proliferation fails, there is no way to compensate by excess proliferation at a later stage (


). There is also evidence that something more

instructive may be involved. IL-7R signaling directly facilitates rearrangement of the TCR? genes (

178 ). IL-7R signaling appears to increase transcription of the TCR? genes before rearrangement, possibly via

activation of the transcription factor Stat5 ( 179 ,


). There is evidence that Stat5 can open the TCR? loci

for rearrangement and preferentially target RAG1/RAG2 activity to these sites ( 181 ,



The transcription factor requirements for TCR?d and TCRaß lineage differentiation can also be distinguished. The difference between ?d and aß cells may be regulated in part by the ratio of the E2A bHLH transcription factor to its antagonists of the Id family, especially Id2 and Id3. This regulatory influence emerges dramatically from experiments in which human lymphoid precursors were forced to express high levels of Id3. In early, uncommitted precursors, Id3 expression favors NK cell development and blocks T cell development altogether (


), but in T-lineage–committed precursors, the result is to

block aß cell development while promoting ?d cell generation (

). In mice, a related mechanism may be used naturally, especially to differentiate between fetal ?d cells and adult-type cells of both aß and ?d 134

types. The fetal thymus in general seems to tolerate levels of Id2 expression, in relation to E2A expression, that would be inhibitory to adult TCRaß cell development (


). The reduced net E2A activity

in the fetal thymus appears to help target TCR?d rearrangement preferentially to the unique “first-wav e” V? and Vd gene segments, V?3(V?5), V?4(V?6), and Vd1, while inhibiting adult-type gene rearrangements (


). Adult TCRaß lineage cells in particular may also need higher levels of positively

acting bHLH factors for multiple functions, beyond choosing the correct TCR to rearrange. For example, a dominant negative knock-in mutant form of HEB, which antagonizes both HEB and E2A activity, blocks the generation of DP cells even in the presence of a TCRaß transgene, whereas TCR?d cell development is spared (



Another regulator of the TCRaß:TCR?d ratio appears to be Notch-1. Notch-1 +/- mice have normal thymocyte subsets, but in chimeras of Notch +/- and Notch +/+ cells, a substantially increased percentage of Notch +/- thymocytes develop as TCR?d cells. Thus, reduced Notch signaling, in relation to neighboring cells, favors the TCR?d fate ( 184 ). 2 Workers using a conditional knockout strategy to delete Notch-1 in DN2/DN3-stage thymocytes confirmed that committed T-lineage cells that have lost Notch-1 are specifically inhibited unless they develop as TCR?d cells ( 184a). Although part of its effect may be on TCRß gene rearrangement, Notch-1 signaling seems to act on the function that determines developmental path, irrespective of the TCR gene rearrangement in the cells. For example, although Notch-1 overexpression does not seem to increase the number of TCRaß DP cells generated through ß selection, it significantly enhances the ability of TCR?d + cells to develop into DP cells ( 184 ). Other workers showed that Notch-1–activated transcription factors directly regulate pTa ( pTa expression is correlated with aß lineage bias (


, 186 ) and that high

). Thus, in addition to its essential role in establishing T-lineage precursors, Notch-1 continues to influence their later developmental choices. 171

Models for the T-Cell Receptor aß:T-Cell Receptor ?d Lineage Choice Although TCRaß:TCR?d lineage divergence has not yet been solved, it is useful to compare a few ways of considering it, diagrammed in Fig. 9. The predominant view of the TCRaß:TCR?d choice is as a classic binary choice in which the cells reach a point in their development beyond which they can go forward only as an aß cell in one path or as a ?d cell in the other ( Fig. 9, options A1, A2). According to such a model, cells must actively determine whether to follow the developmental pathway involving ß selection or the one or more distinct pathways that lead to fetal or adult types of ?d cells. The cells then have a problem to solve: how to coordinate their lineage choice accurately with the random success of their rearrangement of TCRß or TCR? and TCRd genes. One way it could work is for the gene rearrangement to occur first and then for the pre-TCR and TCR?d complexes to deliver different instructive signals for differentiation ( 119 ) ( Fig. 9, option A1). Another way it can work is for the lineage choice to occur first and then the recombination accessibilities of TCRß genes versus TCR? and d genes to become biased by some mechanism controlled by that developmental lineage choice. The latter case ( Fig. 9, option A2) would explain why a bias may be detectable even before TCR gene rearrangement, with the surface density of IL-7Ra chains ( 170 ) used as a marker.

FIG. 9. Models for the divergence of T-cell receptor (TCR) aß and TCR?d lineages. The panels depict hypothetical choice points and their effects on the cells according to the models discussed in detail in the text. These changes are proposed to occur during the proliferation and differentiation events from the double negative (DN) 2 stage through the DN3 stage. A: The two versions of this model envision that the cells reach a discrete branch point at which they must decide whether to choose a TCR?d fate or a TCRaß fate; the only difference between these models is whether the choice involves a commitment to a developmental program or a successful gene rearrangement. In these models, the choice made instructively causes the cells to rearrange the appropriate genes or undergo the correct pr ogram thereafter. B: In this model, two stochastic processes are envisioned to overlap: a decreasing ability to make TCR? gene rearrangements and an increasing ability to carry out the specific CD4/CD8 differentiation program associated with the TCRaß lineage. The shaded wedges represent decreasing and increasing expression of interleukin-7R and pTa, respectively. In this model, lineage commitment and TCR gene rearrangement could be matched stochastically without a unique choice point or an instructive process.

An additional model ( Fig. 9B) is suggested by the numerous hints that aß:?d lineage divergence is asymmetrical. For example, IL-7Ra low/- or pTa high DN2/DN3 cells can be shown to have lost TCR?d potential while retaining TCRaß potential, but no comparable TCR?d-committed precursor has been identified. IL-7Ra levels as a whole tend to fall and pTa levels tend to rise as cells differentiate from the DN2 to the DN3 state. The lack of proliferation in DN3 cells before ß selection ( 1 ) could be a sign that the IL-7R signaling system has become completely disengaged, and thus unable to facilitate TCR? rearrangement, by the time TCRß rearrangement is maximal (

166 , 187 ). There is also the asymmetry in gene rearrangement timing and effects in the two pathways. In principle, it should be easier to be ß

selected than to become a TCR?d cell if the two pathways are in even competition and if success at TCRß rearrangement were sufficient to make a cell undergo ß selection. Instead, there are a substantial number of in-frame TCRß rearrangements in TCR?d cells. This suggests that some productive ß-chain rearrangement can occur in precursors that cannot take advantage of it. This may be because they are precommitted to the TCR?d lineage, but it could also be because they are not yet ready to undergo ß selection. Another type of option, then, may be envisioned as shown in Fig. 9B. The main point about such a model

is that the cells never encounter any one, unique TCRaß:TCR?d choice point. Instead, they progress through a continuum of changes that make TCR?d rearrangement less likely while making the cell better and better prepared for the TCRaß/CD4/CD8 differentiation program. A separation in developmental time, and perhaps in thymic microenvironment, would coordinate the correct developmental program with the preferred TCR gene rearrangement. IL-7R begins to be expressed by very primitive T-cell precursors and declines from the DN2 to the DN3 stage. This would bias generation of TCR?d cells to earlier stages of the DN2/DN3 period. Meanwhile, pTa begins to be expressed at the DN1-to-DN2 transition and could accumulate in the DN2/DN3 stages over time. Furthermore, it is possible that additional regulation or signaling, or both, component changes that enhance the proliferative and differentiative capabilities that can be used at ß selection could occur. Over time, the cells would thus become more likely to make vigorous ß-selection responses and less likely to initiate TCR? rearrangement. In the context of the slow migration of DN2/DN3 cells from the cortical/medullary border to the subcapsular zone of the thymus, this model would predict that most TCR?d + precursors would acquire their TCR in deeper parts of the cortex than the aß lineage precursors. The comparison between these models is useful in attempting to interpret the ways that molecules such as Notch-1 may work in this lineage choice. In the first kinds of models—A1 and A2 in Fig. 9—Notch-1 signaling could act synergistically with pre-TCR signaling to promote ß selection (A1). Alternatively, it could act to enhance TCRß rearrangement, in a discrete subset of precursors that will preferentially give rise to TCRaß-lineage progeny (A2). In the second kind of model, Fig. 9B, Notch-1 signaling need only cause the cells to delay responding to early TCR?d signals or to accelerate the clock controlling their progress through the DN2-to-DN3 transition. Notch-1 activation is directly capable of turning on one gene that is critical for the aß-lineage fate, pTa ( 185 ,

). Ho wever, precedents from erythromyeloid systems indicate that it may also control the timing of differentiation and selective responsiveness to growth factors in developing cells (





, 190 and 191 ).

The roles of TCRaß and TCR?d cells in the periphery are increasingly revealed to be distinct. In the end, this divergence of T-cell lineages has great significance for the ability of the immune system to coordinate functions with the innate immune system and to focus the right kind of response for the nature of the threat. One of the great gaps in knowledge of the immune system is the sparse understanding of TCR?d cells. Their development and their divergence from the aß T-cell pathway is likely to become much clearer in the near future.

POSITIVE AND NEGATIVE SELECTION Positive and negative selection became accessible to study as a result of the effects of expressing transgenes encoding pre-rearranged TCRa and TCRß genes in developing T cells. Not only did these

genes impose a predictable recognition specificity on the T cells, blocking most endogenous TCR gene rearrangement and diversity by allelic exclusion, but they also, dramatically, imposed a predictable developmental fate. Transgenes encoding a receptor that recognized some foreign peptide antigen in the context of the same MHC allelic forms expressed in the thymus could give thymocytes a greatly enhanced likelihood of survival (positive selection). Transgenes encoding a receptor that recognized both self-MHC and a self-peptide expressed in the thymus would cause the transgenic TCR + cells to be eliminated (negative selection). Although details of the expression of the TCR transgenes are not normal (they are expressed at an earlier stage, typically, than normal TCR that depend on TCRa rearrangement), they have revealed the overwhelming importance of details of TCR signaling for thymocyte fate determination. This section describes the life/death decisions guided by TCR recognition events, and the next section examines the effect of TCR signaling details on the choice of cells to be CD4 SP or CD8 SP. The Double-Positive Thymocyte Stage The CD4 + and CD8 + thymocytes produced through ß selection are physiologically specialized for undergoing selection on the basis of TCR recognition. As already indicated in the discussion of the changes induced by ß selection, these cells are a paradoxical combination of extreme sensitivity to TCR ligands and extreme functional paralysis. Unable to turn on any of the functional response genes of mature T cells in response to stimulation, they nevertheless do recognize TCR ligands with ultrasensitive dose–response relationships. Antigenic peptides presented on conventional antigen-presenting cells can trigger apoptosis of DP thymocytes with median effective dose values substantially lower (˜10 fold) than those needed to trigger responses of mature T cells with the same TCR (

, 110 and 111 ). This is especially surprising because the cell-surface density of TCR on DP thymocytes, even after productive 109

TCRa rearrangement, is about 10-fold lower than on SP thymocytes. Operationally, this means that DP thymocytes can make responses to peptide/MHC complexes that are low-affinity ligands for their TCR, too low to be stimulatory for mature cells with the same TCR. One important mechanism contributing to this ultrasensitivity is the distinctive glycosylation state of many DP thymocyte surface molecules. These are strikingly deficient in terminal sialylation, in comparison with surface glycoproteins of mature T cells and immature DN cells alike. These distinctive, developmentally regulated glycosylation properties of DP thymocytes were among the first characteristics of these cells to be noticed, in 1976 (

), long before their functional consequences were understood. Lack of sialylation not only gives DP thymocytes a highly specific reactivity with the lectin peanut agglutinin but also reduces 192

electrostatic repulsion between DP thymocytes and other cells with a more typically strong negative surface charge. As a result, CD8 on DP thym ocytes can bind MHC class I independently of class I haplotype or specific TCR recognition, under conditions in which CD8 on mature T cells cannot (


). On DP thymocytes, CD4 can also interact with class II MHC independently of TCR interactions (

, 113

), possibly also enabled by a lowered sialylation level. The DP cells themselves express neither class I nor 193

class II MHC: the lack of class II is normal for murine T cells and the shutoff of class I MHC expression is another unique feature of the DP state ( Fig. 6). Thus, both class I and class II binding by DP thymocyte

CD8 and CD4 force the DP cells to interact with thymic epithelial cells even with low levels of TCR on their surfaces. DP thymocytes have other features that bear on their eventual fates. Because of the regulatory changes that generate them during ß selection, these cells are extremely sensitive to death induced by glucocorticoids, and, even without perturbation, they die quickly outside of the thymic microenvironment. These properties are especially pronounced in the mouse; human and rat DP thymocytes are somewhat more robust. The glucocorticoid sensitivity, exactly coincident with the peanut agglutinin–binding phenotype and the lack of class I MHC, is tightly developmentally regulated; it can be used as an efficient method to deplete DP cells specifically. Even mildly elevated physiological levels of glucocorticoids in vivo shrink the thymus dramatically through loss of DP cells. The role of glucocorticoids in thymocyte homeostasis is complex. Adrenalectomy, which removes a major source of glucocorticoids, does result in an increase in DP cell numbers. However, low levels of glucocorticoids can antagonize DP thymocyte death in response to TCR cross-linking. This has been proposed as one of the mechanisms establishing the thresholds that distinguish positive from negative selection ( 194 , 195 ). On the other hand, a glucocorticoid receptor exon-disruption mutant has been generated, and the mutant mice show no perturbation of T-cell development (

196 , 197 ). Thus, the role of glucocorticoids in selection is still under investigation. Nevertheless, it seems likely that glucocorticoid

sensitivity is one of the physiological mechanisms limiting the life span of postmitotic DP cells that do not get selected, resulting in “death by neglect.” Time Windows for Positive and Negative Selection Throughout the three-day period that is their average life span, DP thymocytes continue actively to carry on V-J rearrangement of the TCRa locus. This can begin early in the proliferation triggered by ß selection ( 198 ), although in most cases it is likely to be aided by the increases in RAG activity that occur after proliferation stops ( 199 ) ( Fig. 2). Individual cells can rearrange the a-chain genes on both chromosomes, not only once but many times, because the locus offers more than 50 possible Ja segments as well as Va segments in a permissive topology. The first a rearrangements often involve Va and Ja segments that are relatively close to each other, separated only by the d locus. Subsequent rearrangements use more 5' Va segments and more 3' Ja segments. There is no allelic exclusion of TCRa gene rearrangement; the process is terminated either by positive selection, which finally shuts off RAG expression, or by cell death ( 200 , 201 and 202 ). Cells enter a thymic microenvironment in which they can be positively selected before they finish the proliferation that follows ß selection. The critical aspect of the microenvironment in this domain is the possibility for intimate interaction with cortical epithelial cells. This specialized stromal cell type provides a rich source of MHC class I and class II surface complexes with a notable lack of co-stimulatory molecules for T cells. This is important because at the DP stage, co-stimulation causes not activation but negative

selection ( 203 , 204 ). The cortical epithelial microenvironment is thus a uniquely forgiving testing ground for newly generated TCR recognition specificities. Many cells are positively sele cted directly from a proliferating DP blast state (

, 206 ). However, the cells remain rescuable even after they become small, postmitotic cells, perhaps as long as they survive, about 3 days after proliferation stops (


). The window of opportunity for positive selection is thus fairly extended. As a result, the same cell might audition for positive selection repeatedly, from its last cell cycle 207

to 2 days later, with continuing TCRa gene rearrangements, so that each attempt tests a different TCR specificity. The effect of this broad window is that there is no necessary size for the DP thymocyte pool. Under normal conditions, the pool contains about 3 days’ accumulation of postmitotic, unselected cells. In disease, however, the DP thymocyte pool can be shrunk by stress-induced glucocorticoid elevation, and in mice with a transgenic TCR, input to the DP compartment can be significantly reduced by early positive selection. Meanwhile, the maximum fraction of cells eligible for positive selection, among cells with identical TCR, may be limited by competition for a finite number of “niches” ( 208 , 209 ). Thus, positive selection can be more efficient when the size of the DP cohort is smaller than the number of relevant peptide/MHC complexes on the whole cortical epithelium. Negative selection has been found to affect thymocytes at two stages. It undoubtedly affects cells shortly after positive selection, aborting their differentiation into mature SP thymocytes or deleting newly made SP thymocytes that are not yet fully mature. At least for CD4 SP cells, negative selection is possible until the late stages after positive selection when the cells down-regulate CD24 (HSA) expression (





By this time, the cells are in the medulla ( Fig. 3). Most negative selection of CD4-lineage cells with class II MHC-restricted TCR can occur at this CD24 + SP stage. In the cases of certain CD8 lineage cells, with class I MHC-restricted TCR, there is evidence that the cells can also be negatively selected before full differentiation into DP cells. Unlike class II MHC, class I MHC is expressed on many cells besides specialized thymic epithelial cells and dendritic cells, including the DN1 to DN3 cells themselves. Any of these could present antigen for negative selection of class I MHC-restricted cells, especially in TCR transgenics, in which TCRaß may be expressed before the DP stages. This shows that in principle, cells can become susceptible to negative selection during or immediately after ß selection, depending on whether the target antigen and appropriate antigen-presenting cells are present as soon as TCRa is expressed. Triggering and Results of Positive Selection The signaling aspects of positive and negative selection have been studied intensively by many groups since 1995 and are discussed in depth in several excellent reviews (

, 213 and 214 ). DP thymocytes are triggered to undergo positive selection when their TCR complexes and their CD4 or CD8 co-receptors 212

engage a peptide/MHC complex presented by cortical epithelial cells. By far, the best TCR-ligand

interactions for this purpose are of low affinity, for reasons to be described. As a result, a single peptide/MHC complex can positively select DP cells with any of numerous different TCR specificities, as long as they cross-react weakly with that complex (

). The heightened sensitivity of DP cells to weak TCR interactions means that the cross-reactivity with some peptide/MHC complex that allows a 215

thymocyte to escape death may become undetectable, or detectable only as competitive antagonism, once the cell has matured into a peripheral T cell. However, mature T cells continue to recognize other peptides in association with the same class I or class II MHC molecule that mediated their positive selection. Thus, the MHC restriction of a population of mature T cells is generally determined by the MHC antigens that were expressed in the thymus where they differentiated. The sequence of events set in train by positive selection begins with activation: the TCR/CD3/co -receptor engagement activates Lck, Ras, Vav, calcineurin, and protein kinase C ( Fig. 7). A major consequence of Ras pathway signaling here is the activation of the MAP kinase, ERK. These signaling mediators in turn must induce the transcription of Egr1 and Id3 ( 133 ,

, 216 ), as in ß selection, but this time there is little if any proliferation that results. Instead, CD69 is up-regulated, and a dramatic transformation begins to 160

unfold. The cells resume the Bcl-2 and class I MHC expression that they had lost at ß selection and begin to recover the functional responsiveness (e.g., through AP-1 inducible effector genes) that had disappeared at that time ( Fig. 6). Concomitantly, TCR complex expression at the cell surface is stabilized by more efficient assembly ( 217 ), and the cells immediately display higher steady-state levels of TCR/CD3. CD5 and TCR/CD3 expression increase in parallel. Glycoprotein processing is altered to a more “normal” pattern, so that new glycoproteins are once again fully sialylated. This restores electrostatic repulsion between DP and other cells and terminates the ability of CD4 and CD8 to interact with MHC independently of TCR. This is also the start of a 1- to 2-week maturation cascade that gradually leads to down-regulation of CD24 (HSA) on the cells as TCR/CD3 levels rise even higher, and CD69 expression finally subsides. These events appear to be common to all positively selected thymocytes. Strength of Signal versus Distinct Interaction Models for Positive and Negative Selection Positive and negative selection contribute to central tolerance because thymocytes with receptors that interact strongly with peptide/MHC complexes in the thymus are deleted, whereas those with receptors that interact weakly with thymic peptide/MHC complex are selected positively. A series of compelling studies from several groups in the early 1990s established these principles by using thymocytes with transgenic TCR of known specificity. Peptide/MHC complexes yielding high-affinity interactions with the TCR, which are good stimulators for mature T cells with that TCR, would induce death of thymocytes. Peptide/MHC complexes yielding low-affinity interactions, which could result in anergy or antagonism of mature T-cell responses, promoted positive selection and maturation of thymocytes. A simple bell-shaped dose–response function could thus be envisioned to govern thymocyte fate, as shown in Fig. 10 (“two-threshold model”).

FIG. 10. Relation of T-cell receptor (TCR) affinity to thresholds for positive and negative selection. Two models are compared to indicate the range of TCR-ligand affinities that thymocytes must be able to distinguish in order to be directed correctly to positive selection (+), negative selection (-), or a failure of both (neglect). Histograms of the number of cells in hypothetical precursor populations with different levels of TCR affinity for major histocompatibility complex (MHC)/peptide ligands in the thymic microenvironment are shown. Most cells have receptors that interact too weakly to be positively selected ( left, stippled part of curves). In the two-threshold model ( top), double positive (DP) cells simultaneously determine whether to be positively selected ( nonstippled portion of curve) and whether to be negatively selected ( right, stippled part of curve). This choice must occur in the cortex, and it means that cells need to compare their TCR affinity with two reference values at once. In the sequential threshold model ( lower two panels ), DP cells in the cortex ( middle panel) first need only compare their TCR affinity with a minimum threshold for positive selection. Those cells being positively selected ( bracket) then differentiate toward the single positive (SP) state, enhancing their TCR expression level and migrating to the medulla, where they encounter highly active antigen-presenting cells ( lower panel). They are then selected for TCR affinities below the threshold for negative selection. One reason the sequential threshold model is needed is that TCR surface expression increases 10-fold between the DP and SP stages, enabling SP cells to interact with ligands with a higher avidity (affinity multiplied by number of interactions) than do DP cells with the same receptor. Cells with TCR that appeared to be innocuous in the DP stage could turn out to be autoreactive with their increased TCR levels in the SP stage, unless they were removed by a medullary negative selection mechanism.

The striking features of this model are the sharp discontinuities in response at two points in a continuum of signal strength: separating nonselection from positive selection and separating positive selection from negative selection. If the cell had its fate determined in a single encounter, it would need extraordinarily precise computation of signal intensity levels. Small alterations in expression of signaling components, such as overexpression of a TCR complex or a downstream mediator in a transgenic mouse, might be expected to send the whole population into negative selection or nonselection catastrophes. However, the system is more robust than this. Moreover, it shows evidence of being tunable. In one model, in which thymocytes were made transgenic for a TCR with several possible ligands, positive selection could be promoted by several different peptide/MHC complexes with different affinities for the same TCR. However, the properties of the cells that matured were different, depending on the ligand that had selected them. In each case, the mature cells were anergic to the selecting ligand but responsive to stronger agonists (

, 219 ). Thus, cellular properties distinct from the primary structure of the TCR ligands could contribute to positive selection thresholds. How could the sharp discontinuities in 218

dose–response function be reconciled with this evidence for plasticity? An important advance has been the demonstration that positive and negative selection involve qualitatively distinct signaling pathways. Through the use of transgenes with the Lck gene proximal promoter to direct thymocyte-specific expression of mutant signaling molecules, the Ras signaling pathway downstream of TCR signaling could be selectively manipulated. Such work showed that Ras and MEK signaling were critical for positive selection but unnecessary for negative selection (

). This result refuted the simple prediction that a stronger TCR-ligand interaction, as needed for negative selection, 220

would necessarily be distinguished by activation of all the pathways used in positive selection plus additional ones. More recently, mutation of specific TCR complex components has been found to disrupt positive selection but not negative selection. The results suggest that positive selection signals uniquely depend on a relay involving a domain of the TCRa chain itself and ultimately, ERK ( 221 ,

222 ) [reviewed by Hogquist ( 214 )]. Mutations of any of these components block positive selection without preventing negative selection. An

implication is that particular substructures within the TCR complex engage a discrete set of downstream signaling components and that those needed for positive selection can be dissected from those used in other responses. In subsequent work, negative selection has been found to depend on mediators of distinct signaling pathways [reviewed by Hogquist (


) and Mak et al. ( 223 )], especially pathways triggered by signals

from professional antigen-presenting cells that would be co-stimulatory for mature T cells (

). Interactions through certain TNF receptor family death receptors are especially potent ways of causing 224

deletion ( 225 , 226 ). Both in the DP stage and in the HSA + SP stage, any of a variety of co-stimulatory ligand-receptor interactions can trigger negative selection. Within the thymocytes themselves, the GTPase Rac1, which helps activate the p38 stress kinase and enhances cytoskeletal reorganization,

apparently promotes negative selection or even converts positive selection responses to death responses when chronically activated (


). Moreover, not all ways to kill immature thymocytes are reflections of the

same negative selection process ( 228 ). In some cases, even IL-2/IL-2 receptor interaction can act as a cofactor for death (

229 ). DP thymocytes die in vivo when animals are injected with anti-CD3 antibodies, at least in part because of TNFa release by mature T cells in the periphery, and this may mimic those forms

of negative selection that are dependent on TNF receptor/Fas/CD40L family co-stimulation but not others. Reflecting the importance of non–TCR-mediated inputs, positive and negative selection can often be mediated by distinct antigen-presenting cell types (

). This was not so clear in initial studies with class I MHC-restricted TCRaß transgenes, which repeatedly tended to promote elimination of DP thymocytes 230

within the cortex when the thymic microenvironment expressed a high-affinity MHC/peptide ligand. However, for cells with class II MHC-restricted TCR, which undergo positive selection to the CD4 + cell lineage, negative selection is triggered primarily in the thymic medulla. In fact, mice that are genetically manipulated to express class II MHC only in the cortex generate autoreactive CD4 + cells that cannot be eliminated and thus may end up causing autoimmune disease ( 231 ). The most effective antigen-presenting cells for negative selection are the hematopoietically derived dendritic cells (the same non-T population that can be derived from a common precursor with T cells). Dendritic cells not only express profuse class I and class II MHC but also display a wide variety of co-stimulatory molecules on their surfaces, from the immunoglobulin superfamily molecules CD80/CD86 to the TNF receptor family molecule CD40 and possibly ligands for CD5 as well. Positive and negative selection can thus take place sequentially; CD4 + cells positively selected in the cortex do not have an occasion to be negatively selected until they migrate to the medulla and encounter dendritic cells there. This sequentiality is important because it relaxes some of the upper limit constraints on affinity that can be used for positive selection. Autoreactive CD4 + cells can afford to be positively selected, because under normal circ*mstances they can be negatively selected later ( Fig. 10, sequential threshold model). This can be demonstrated in thymic organ culture reaggregates by using mixtures of genetically distinct cortical epithelial and dendritic cells with broader or narrower antigen-presentation capacities. In this system, it has been possible to generate mixed microenvironments in which at least 75% of the thymocytes positively selected in the cortex are subsequently destroyed in the thymic medulla (

). The medullary location is important, too, because, unlike the cortex, it is a site for expression or import of a 232

large spectrum of genes used in peripheral somatic tissues (

10 ). Thus, it provides a much better test panel of antigens than does the cortex to detect and eliminate self-reactivity. Sequentiality also frees

positive selection to be mediated by TCR/co-receptor interactions with peptide/MHCs over a wider range of affinities ( Fig. 10, sequential threshold model). As discussed in the later section on CD4 helper T cell versus CD8 cytotoxic T cell lineage commitment, giving the cells a wider spectrum of affinities within which to be positively selected makes it easier to understand some of the aspects of the CD4/CD8 lineage choice.

Another Escape from Autoreactivity in the Thymic Cortex Even if negative selection may normally be most efficient in the medulla, there can still be a penalty for extremely high-affinity interactions with self-MHC antigens in the cortex. Results of one study indicate one mechanism that is seen best by TCR genes in their native chromosomal context, not transgenes. High-avidity (i.e., affinity multiplied by density of interactions) interactions in the cortex do lead to disappearance of cells with the offending receptor, but not only because the cells are committed to negative selection. Instead, it appears that for many of them, such interactions so e fficiently cause internalization of surface TCR that the cells never receive a sustained positive selection signal. Instead, they proceed as though their previous receptor gene rearrangement had yielded no TCRaß complexes and continue with TCRa locus rearrangements, joining a Va segment upstream of the previous one with a Ja segment downstream of the previous one, and so forth, until they receive a positive selection signal or reach the end of their DP cell life span ( 233 ). The ongoing rearrangement that leads to replacement of one TCR specificity by another is called receptor editing. Because of the large number of both V and J gene segments, this process can continue on both chromosomes to generate several rounds of receptor specificities. The result is that not only is the autoreactive TCR lost from the cell surface but also the TCRa gene rearrangement that created it is also commonly lost from the genome. If this model is generally correct, then the first positive selection thresholds would thus be determined as the window between the minimum avidity required to trigger Ras activation and the maximum avidity that allows a significant number of TCR complexes to remain on the cell surface. Negative selection, occurring later in the medulla, offers a refinement in the context of a large spectrum of extrathymic self-peptides, but the cell biology of positive selection versus TCR internalization can provide an initial MHC affinity filter.

CD4 HELPER T-CELL VERSUS CD8 CYTOTOXIC T-CELL LINEAGE COMMITMENT The most challenging aspect of positive selection is its apparent connection with a major developmental lineage decision, over and above the decision to live or die. Some of the positively selected cells become helper (CD4 SP) cells and others become killer (CD8 SP) cells. These subsets differ not only in recognition specificity, effector function, and co-receptor expression pattern but also in a whole host of characteristics regulating every aspect of cellular response from homeostatic mechanisms to growth factor responses. The challenge from a mechanistic point of view is how the apparently simple signal to live or die (or continue receptor editing) is intertwined with signals to undergo divergent pathways to maturation into richly different cell types, as described in detail in the later chapters of this book. Major Histocompatibility Complex Restriction Regulates CD4 versus CD8 Lineage Differentiation The problem of what determines the development of aß T cell lineage cells into either CD4 + helper T

cells or CD8 + cytotoxic T cells has been intensely examined and passionately debated since 1990 but is still not settled ( 214 , 234 ,

235 and 236 ). It is at the DP stage that the association among the CD4 or CD8 co-receptor, aß TCR specificity, and function appears to be established. When cells expressing an aß

TCR that recognizes peptide in the context of the antigen-presentation molecule class I MHC are positively selected, they down-regulate the expression of CD4 and activate the gene program specific to a CD8 cytotoxic T cell. Conversely, cells expressing an aß TCR that recognizes peptide in the context of the antigen-presentation molecule class II MHC down-regulate the expression of CD8 and activate the gene program specific to a CD4 helper T cell. Examples are provided in virtually every thymus of animals with pre-rearranged TCRaß transgenes ( Fig. 11A and B, wild type). Furthermore, CD8 SP cells are not generated in a thymus lacking class I MHC, and CD4 cells are not generated in a thymus lacking class II MHC. 3 It is not a coincidence that CD8 binds to class I MHC and CD4 binds to class II MHC—the simultaneous binding of class I by a class I–specific TCR and CD8 or class II by a class II–specific TCR and CD4 increases the affinity of the interaction and activates signaling pathways inside the T cell. Nevertheless, it is difficult to see how the cell perceives the difference in ligand binding to its TCR and co-receptor and then translates the signal into a choice of divergent differentiation programs.

FIG. 11. CD4/CD8 lineage choice dictated by T-cell receptor (TCR) recognition specificity and by Lck activity levels. CD4/CD8 expression profiles of thymocytes from TCR-transgenic mice show the powerful effect of TCR specificity on CD4/CD8 lineage choice. In these analyses, thymocytes were stained to detect expression of CD4, CD8, and the specific transgenic TCR, and the right portion in each pair shows the CD4/CD8 pattern of the mature single positive (SP) cells expressing high levels of the transgenic TCR. The TCR-transgenic thymocytes are shown both in a normal genetic background ( upper panels) and in genetic backgrounds which decrease or increase Lck activity in thymocytes ( lower panels). A: Transgenic TCRaß that recognizes antigen in association with class II major histocompatibility complex (MHC) directs mature, TCR hi development to the CD4 SP fate ( upper panels). The bias is very strong in comparison to normal thymocytes (compare with Fig. 1, control samples in Fig. 8). In mice with this TCR transgene plus a transgene that reduces thymocyte Lck activity, however, the same transgenic TCR promotes development of CD8 SP cells ( lower panels). B: Transgenic TCRaß that interacts with class I MHC overwhelmingly directs development of mature TCR hi thymocytes to the CD8 SP fate ( upper

panels ). However, when mice with this transgene are crossed with transgenic mice that express elevated levels of Lck in thymocytes, the double transgenics show transgenic TCR hi thymocytes developing as CD4 SP cells ( lower panels). Thus, the level of Lck activity appears to mediate the way the cell distinguishes between recognition of class I or class II MHC. From (


), with permission.

There have been several problems in trying to solve the basis of this choice. First, the severe loss of cells from every DP generation, which is common to most TCR-transgenic models as well as to normal mice, makes accounting impossible. Second, there is a critical shortage of markers that distinguish cells taking the CD4 path from those taking the CD8 path. As a result, much research is focused on the requirements for expression of the CD4 and CD8 molecules themselves. Because these vital co-receptors clearly participate in the process, however, there is an essential circularity in these analyses, which great ingenuity in experimental design has only partially succeeded in overcoming. Furthermore, the developmental expression of these molecules, even at the RNA level, is complex. There is strong evidence that the transit from DP to CD4 or CD8 SP involves not instant repression of the “wrong” co-receptor but rather a period of phenotypic instability going through CD4 loCD8 lo and CD4 + CD8 lo intermediates to both end states ( 237 ,

238 , 239 , 240 , 241 and 242 ) ( Fig. 12A). With few markers capable of distinguishing between pre-CD4 and pre-CD8 intermediates at the earliest stages of their developmental

divergence, studies have had to rely on measuring the output of mature CD4 + versus CD8 + cells. Full maturation is almost certain to rely on specialized survival signals for each committed lineage as well as the initial events that set the cells onto different pathways. Thus, much controversy has revolved around the respective roles of survival versus differentiation processes in the CD4/CD8 lineage choice.

FIG. 12. Models of CD4/CD8 T cell lineage commitment. A: Schematic of changes in thymocyte phenotype during positive selection to the CD4 single positive (SP) or CD8 SP fate. The layout is that in which detailed interpretations of signaling and selection events are presented ( B to D). An idealized flow cytometry plot is shown, with arrows denoting the path of differentiation taken by many CD8 SP precursors ( solid arrow) and by CD4 SP precursors ( dotted arrow). Both CD4 and CD8 expression are initially down-regulated, with transient recovery of CD4 expression, before the pathways of the two cell types are clearly seen to diverge. The following models depict the signaling events proposed to occur in each of the populations indicated. B to D: Symbols for cells that are proposed to be distinctly programmed but similar in surface phenotype are enclosed in gray boxes. B: Selective model. This model posits that before leaving the DP population, cells become programmed for different responses to positive selection signals. The co-receptors are programmed for down-regulation in a pattern that is independent of T-cell receptor (TCR), resulting in diverse combinations of CD4/CD8 phenotype and TCR specificity once positive selection begins. Cells that lose a co-receptor that is required to stabilize their TCR interactions with major histocompatibility complex (MHC) cannot continue to signal. Only cells capable of signaling, either through co-ligation of TCR and co-receptor by MHC or through ligation of a high-affinity TCR that is independent of co-receptor, will survive. C: Instructive model. This model is based on the idea that strong intracellular signals emanating from the co-ligation of a TCR specific for class II MHC and CD4 instruct the cells to silence CD8 and mature into a CD4 + helper T cell, whereas weaker intracellular signals emanating from the co-ligation of a TCR specific for class I MHC and CD8 lead to the silencing of CD4 transcription and the maturation of the thymocyte into a CD8 + cytotoxic T cell. This model does not predict the presence of thymocytes expressing inappropriate co-receptors for their TCR specificity unless signaling pathways are perturbed. D: Kinetic signaling model. This model holds that all thymocytes transit through an initial down-regulation of both co-receptors and then an up-regulation of CD4 to form a CD4 + CD8 lo population ( A, broken arrow and heavy solid arrow). Cells expressing TCR specific for class I MHC cease signaling, because they lack sufficient CD8 to co-ligate

class I MHC with the TCR. The cessation of signaling triggers the silencing of CD4 transcription and the maturation of these cells into CD8 + cytotoxic T cells. Thymocytes expressing a TCR specific for class II MHC continue to receive signals from the co-ligated TCR and CD4 and mature into CD4 + helper T cells.

Models for CD4/CD8 Lineage Divergence Several models have been postulated to explain the correlation of TCR and co-receptor with mature T cell function ( Fig. 12B, C, and D). A selective model postulates that CD4 + CD8 + thymocytes choose a CD4 SP helper versus CD8 SP killer lineage independently of their TCR recognition specificity. On the basis of this prior choice, they immediately begin down-regulating the unwanted co-receptor when positive selection begins. However, because sustained interactions are needed to complete maturation, the only thymocytes that survive are those that continue to express the co-receptor that binds to the same type of MHC molecule as the TCR ( 243 , 244 ). This model predicts that “mismatched” thymocytes would be generated—that is, that there would exist in the thymus transitional thymocytes expressing class I–specific TCR and CD4 or class II–specific TCR and CD8—but that such cells would be unable to finish their differentiation ( Fig. 12B). Evidence for this is seen in a mouse model in which early commitment to the CD8 lineage can be traced by a silencing of a ß-galactosidase gene “knocked into” the CD4 locus. In this mouse, 19% of cells apparently committed to the CD8 lineage express class II MHC-specific TCR, and 11% of cells apparently committed to the CD4 lineage express class I MHC-specific TCR (

). It is not clear whether these mismatched thymocytes die or whether they can reverse this early commitment, 242

but they are not present in the mature T-cell population. The selective model also predicts that these mismatched cells could be rescued by constitutive expression of a co-receptor matching the specificity of the TCR. This has been shown in a mouse with a hom*ozygous deletion of the CD4 silencer, which results in constitutive expression of CD4 in SP thymocytes and T cells and the development of CD8 + CD4 + class II MHC-restricted cytotoxic T cells (

). An argument against the selective model is that analysis of various TCR-transgenic mouse strains shows that the selection efficiency can be much higher than would be expected from the selective model (




The instructive model postulates that the lineage choice is made only in response to the positive selection signal, directed by a difference in signaling between TCR/CD4 recognition of class II MHC and TCR/CD8 recognition of class I MHC ( Fig. 12C). This model is supported by evidence that signaling through the cytoplasmic tail of CD4 (

, 247 ) and CD8 ( 248 , 249 ) leads to the development of CD4 SP and CD8 SP thymocytes, respectively. A difficulty with the instructive theory has been how to account for the 246

differentiation between signals received from the CD4 and CD8 cytoplasmic tails, because both co-receptors associate with the same signaling molecules, most notably the Src family kinase p56 Lck (Lck). The two co-receptors do differ in the numbers of Lck molecules associated with their cytoplasmic tails: The cytoplasmic tail CD4 is associated with approximately 20 times more p56 Lck than is the cytoplasmic tail of CD8a ( 250 ). This difference is exacerbated by the fact that many DP thymocyte CD8a

molecules are truncated so as not to bind any Lck at all. Increases in Lck activity promote the adoption of the CD4 + lineage ( 251 , 252 and 253 ), whereas diminution of p56 Lck activity promotes the adoption of the CD8 + lineage ( 251 ) ( Fig. 11A and B, lower panels). The natural mixture of CD8a Lck-binding and nonbinding forms appears to promote CD8 SP cell development most efficiently, by stabilizing TCR-MHC interactions with the minimum of Lck recruitment ( 254 ). An increased duration of signaling through the TCR is also associated with a commitment to the CD4 + lineage ( 255 ). This is in agreement with results from a suspension culture system in which DP thymocytes can be directed into the CD4 SP or CD8 SP pathway with a combination of ionomycin (a calcium ionophore) and phorbol 12-myristate 13-acetate (PMA), depending on the length of time that the stimulation is applied. In this system, inducing DP thymocytes to commit to the CD4 + lineage requires longer duration of treatment than the duration of treatment required to induce DP thymocytes to commit to the CD8 + lineage ( 256 ). Some effects attributed to signal intensity may actually be signal duration effects. Even in vivo, in transgenic systems, Lck activity levels were most clearly seen to affect CD4/CD8 lineage choice when they were expressed in a sustained way over a period of days or from a promoter with increasing activity after the DP stage (

, 257 ). This would associate the CD4 + lineage choice again with a sustained elevation of Lck signaling, even after the positive selection process had begun. 252

Kinetic signaling models ( Fig. 12D) define this difference in signal duration requirements for the CD4 + and CD8 + fates as being more important than the peak strengths of the signals. This class of model is supported by the observation that most DP thymocytes pass through an initial down-regulation of both co-receptors and then an up-regulation of CD4 to form a CD4 + CD8 lo population before the cells diverge to undergo differentiation to mature CD4 or CD8 SP states. During the CD4 + CD8 lo stage, interactions with class II MHC could continue, whereas interactions with class I MHC would naturally be weakened from loss of co-receptor contribution. The kinetic signaling models posit that continued signaling through TCR and CD4 leads to development into CD4 SP thymocytes; cessation of signaling triggers the down-regulation of CD4 and the up-regulation of CD8, leading to the development of CD8 SP thymocytes ( 258 ). After the basic lineage choice has been made, full maturation of CD8 SP cells appears to depend on non-TCR survival functions: either Notch family signaling (

) or IL-7R signaling ( 258 ). This kind of model can be integrated with signal strength or duration hypotheses, because thymocytes expressing 255

class I MHC-restricted TCR in conjunction with a low level of class I MHC-ligated CD 8 would naturally receive a lower strength signal than would thymocytes expressing class II MHC -restricted TCR in conjunction with high levels of class II MHC-ligated CD4. A kinetic signaling model is supported by an experiment in which a hybrid molecule with a CD8a extracellular and transmembrane domain and a CD4 cytoplasmic tail is expressed in mice lacking class II MHC and CD8a (

259 ). In this background, signals from the hybrid co-receptor promote the development of even more CD8 SP than of CD4 SP class I MHC-restricted thymocytes. In contrast, the instructive

model predicts that the hybrid co-receptor would have preferentially promoted commitment to the CD4 + lineage simply because of its CD4-promoting signal when the hybrid co-receptor co-ligated to class I MHC with a class I MHC-restricted TCR. Indeed, when class II MHC and endogenous CD8a are present, that is what is seen ( 247 ) but not when the occasion for CD4/class II MHC interaction is taken away. To understand how this works, recall that CD8 exists as a dimer (CD8aß) and that the CD8ß chain increases CD8a-associated Lck activity (


). Normally, both CD8a and CD8ß are down-regulated in the CD4

+ CD8 lo

stage. The kinetic signaling model explains the increased numbers of CD8 SP thymocytes produced by the hybrid co-receptor by increased strength of signals during positive selection, increasing the number of cells entering the CD4 + CD8 lo compartment. Once the thymocytes enter the CD4 + CD8 lo compartment, however, endogenous CD8ß is down-regulated, which decreases the signal resulting from class I MHC ligation of the hybrid co-receptor and results in direction of the thymocytes into the CD8 lineage. Thus, the key issues that distinguish these models are (a) whether mismatched cells are generated at all, (b) whether the fates of such cells are to die or to be redirected into a different lineage, and (c) when the choice of fates is actually finalized. These properties of the three kinds of model are contrasted simply in Fig. 13 . Data support the prediction that cells are generated initially that appear “mismatched,” in conflict with the simplest version of an instructive model, but according to the kinetic signaling models, these cells simply have not yet finished receiving instructions about the types of cells they will be. Many, although perhaps not all (

), CD8 cells do appear to develop through a CD4 + CD8 lo intermediate, making the various kinetic signaling models plausible. However, there is still no direct proof that the developmental 242

mechanisms needed for both CD4 and CD8 SP differentiation remain available to individual cells as late as the CD4 + CD8 lo stage. Thus, as in the cases of the earlier lineage choices, models based on selection and differential survival cannot yet be excluded. In any case, the challenge is to understand how these complex differences in intensity and duration of signaling can be translated into different developmental responses.

FIG. 13. Developmental effects of intensity and time courses of T-cell receptor (TCR)/co-receptor signals in three models of CD4/CD8 lineage choice. Plots showing the intensity of predicted TCR/co-receptor signals over time and their effects on cell fate are compared in three models of CD4/CD8 lineage choice. Left: Selective model predicts initial divergence of pre-CD4 single positive (SP) and pre-CD8 SP cells, after which the only choice the cells need to make in response to TCR signals is to live or to die. Termination of signaling, the usual result of down-regulating the wrong co-receptor for the cell’s TCR, leads to death in either lineage. Upper right: Signal strength model predicts that the intensity of TCR/co-receptor signaling instructs the cells whether to take a CD4 SP or a CD8 SP fate. It is assumed that the usable intensity range is limited by negative selection at the high end and by neglect at the low end. Lower right: Signal duration–based models such as the kinetic signaling model shown in Fig. 12D predict that transient signaling is a specific instructive signal for CD8 SP development, whereas sustained signaling is a specific instructive signal for CD4 SP development. In this kind of model, although TCR signaling itself may provide continuous survival functions for CD4 SP cells, additional kinds of survival functions are needed for CD8 SP cells (not shown).

Molecules Implicated in the CD4/CD8 Lineage Choice Signal intensity or duration does not necessarily mean an equal involvement of all the multiple branching pathways activated by TCR ligation ( Fig. 7). Mutant analyses show that the pathway most relevant to the signal duration/strength that will determine CD4 versus CD8 SP fate is the pathway involving Lck and ERK. Members of the Lck signaling cascade are essential for both but favor CD4 SP development over CD8 SP at high levels. Molecules other than TCR/co-receptor signaling components are also implicated in lineage-restricted roles, and this is an important step toward being able to resolve when the CD4 and CD8 gene expression programs actually diverge. To date, however, it is still debated whether the known molecules affect the choice itself or, rather, the efficiency of survival after the choice. Lck activity levels have a powerful impact on the choice, as discussed ( Fig. 11; compare upper and lower panels). From the DP stage onward, Lck is brought to the TCR signaling complex by CD4 or CD8, making the direction of CD4/CD8 lineage choice depend, by implication, on levels of co-receptor involvement in TCR triggering. Thus, the CD4/CD8 co-receptors have two separable roles in positive selection: They stabilize and initiate TCR complex signaling, and they do so either with a minimal amount of Lck—just enough to trigger ZAP70 and the rest of the TCR cascade—or with a “bonus” level of Lck, which CD4 can provide but CD8 cannot ( 244 , 247 ,

, 259 ). For this to work, the cells must be able to distinguish between different levels of Lck signaling even in the context of equal levels of TCR complex/ZAP70 254

signaling. In this connection, it is interesting to note that the ratio of co-receptor–driven signaling to other TCR signaling may be expected to shift during the positive selection process in general, as TCR surface levels rise and the co-receptor-MHC interactions are weakened by increased surface sialylation (


, 113

). This shift would tend to amplify the effects of the CD8 down-regulation at the CD4 + CD8 lo intermediate

stage and enhance a signal duration or kinetic signaling mechanism. One of the mediators that is preferentially activated by Lck is ERK. The involvement of the MAP kinase ERK in CD4 versus CD8 lineage commitment has been suggested by studies in which pharmaceutical inhibitors of ERK selectively block the development of CD4 SP thymocytes but not CD8 SP thymocytes ( 253

, 260 ,


), as do genetic methods to reduce ERK activity, such as use of a dominant-negative MEK transgene (


). However, another study using ERK inhibitors shows a block in development of CD8 SP (

) or a null mutation of ERK ( 264 ). One interpretation is that the disparate results may result from the inhibition of the activity of molecules other than ERK by the pharmaceutical inhibitors used. Alternatively, 263

both lineages may require some ERK signaling, but a higher level or duration of ERK signaling may be critical for CD4 cell commitment or maturation, whereas a transient or low level of ERK signaling may be optimal for CD8 cell development. In general, molecules that increase signaling (without promoting negative selection) favor CD4 cell production, whereas molecules that decrease signaling favor CD8 cell production. Thus, integrin-mediated enhancement of interactions is needed most urgently to make CD4 SP cells (


). A

new DNA-binding protein that promotes CD8 cell development, TOX, appears to do so by reducing overall TCR signal strength (


). There is a role for Notch family molecules, to be discussed, and this

too has been linked with effects on TCR signal strength (


). Even the threshold-setting functions of

Ikaros for T-cell signaling, which were discussed previou sly in the context of ß selection, could play a role again in the CD4/CD8 lineage choice: mice with Ikaros-deficient mutations generate a higher percentage of CD4 SP cells, as opposed to CD8 SP cells ( 154 ). Genes that implement or stabilize the developmental effects of the positive selection signal to bring about CD4 SP or CD8 SP lineage differentiation have been harder to identify so far. Thus, there is a serious shortage of candidates for genes that actually define the long-term lineage identities of CD4 SP or CD8 SP cells. One spontaneous mutation affects a gene that may prove to have a unique role in the development of the CD4 helper lineage. In the “helper-deficient” mutant mouse, class II MHC-restricted thymocytes are directed exclusively to the CD8 lineage, even in the absence of class I MHC ligation of the CD8 co-receptor. Positive selection signals in the helper-deficient mutant thymocytes appear to be normal (


). This suggests that although positive selection, mediated by signals through the TCR, and CD4 or CD8 lineage commitment occur at the same time, they are not the same process. Unfortunately, the gene mutated in this strain is not yet identified. CD8 SP lineage cells apparently require more survival-promoting functions than CD4 SP cells (

29 ). Either transgenic overexpression of Bcl-2 or stimulation of the IL-7R, which is expressed again after the

DP stage ( 269 ), disproportionately enhances CD8 SP yields ( 258 ,

). Thus, some of the genes that enhance CD8 cell development could be doing so simply by providing survival functions. Among the most 270

controversial of these effects is the impact of activation of Notch family genes. Increased signaling by a constitutively activated cytoplasmic domain of the cell surface receptor Notch-1 has been shown to favor the development of CD8 SP thymocytes ( 271 ). The original report was highly important as the first demonstration that overexpression of a non-TCR gene could shift cells into the CD8 SP pathway even when their TCR had interacted with class II MHC. This effect has remained controversial, however, because expression of a larger piece of the intracellular domain of Notch-1, also constitutively activated, promotes the development of CD8 SP and, to a lesser degree, CD4 SP thymocytes (

). In vitro experiments diminishing Notch-1 activity in thymocytes that had already received co-receptor-TCR-MHC interaction signals have revealed that Notch-1 functioned only to 185

promote the postcommitment maturation of CD8 SP thymocytes (

255 ). In addition, a conditional deletion of Notch-1 shows no effect on CD4/CD8 lineage commitment if done after the earliest stages of

thymocyte development (



One way the controversy may be resolved is that other Notch family members expressed in the thymus are able to substitute for Notch-1 in the CD4/CD8 lineage choice. Pharmaceutical inhibition of the enzymatic activation of all three Notch family members molecules in DP thymocytes does selectively decrease the number of CD8 SP thymocytes (


, 65 ). These Notch inhibitor effects are dose and potency

dependent, in such a way that greater inhibition of Notch impairs development of both CD4 + SP and CD8 +

SP thymocytes and inhibition to a lesser degree impairs only the development of CD8 + SP thymocytes.

The disparate effects seen in transgenic mice expressing activated Notch-1 can also be explained by dose and potency dependence. The activated Notch-1 domain that selectively promotes the development of CD8 + SP thymocytes ( 271 ) is shorter and less potent than the activated Notch-1 domain that promotes the development of both CD4 + SP and CD8 + SP thymocytes ( 185 ). The expression of both of these constructs is driven by the Lck proximal promoter, which decreases in activity during the time that DP thymocytes commit to either the CD4 or CD8 lineage. In contrast, expression of the full-length intracellular domain of Notch-1 from a constitutively active retroviral promoter inhibits development of both CD4 and CD8 SP thymocytes through attenuation of signals through the TCR( 267 ). The lower effective dose of activated Notch-1 delivered from a proximal Lck promoter- driven transgene may promote development of CD8 + SP thymocytes by lowering the strength or duration of the signal through the TCR, in keeping with the instructive models of T cell development, or by providing a lineage-specific survival function in committed CD8 precursors ( 255 ). The physiological role that the Notch molecules really play in CD4 and CD8 cell lineage commitment still remains to be determined. Notch activity in DP thymocytes, at least as assessed by the expression of the downstream mediators Hes-1 and Deltex, is very low to undetectable (

). It may be that Notch-1 acts through other downstream mediators in DP thymocytes or that other Notch family members and a different set of downstream mediators play a more important role.


While the role of Notch family molecules and the transcription factors they control is being resolved, indications of other nuclear factors that appear to bias CD4/CD8 lineage choice have begun to appear. GATA-3 and Runx1 (AML1, CBFa2, or PEBP2aB), genes involved in T-cell development from the earliest stages ( Fig. 4), also seem to be able to direct the output of positive selection to disproportionately CD4 SP or disproportionately CD8 SP pathways, respectively ( 272 , 273 and 274 ). These effects suggest a link to the functions that must coordinate CD4 and CD8 expression with effector function and other features of the mature CD4 and CD8 SP populations. Maturation and Export of CD4 and CD8 Single-Positive Thymocytes Commitment to the CD4 + or CD8 + lineage in thymocytes is characterized by the down-regulation of one co-receptor and the maintenance of expression of the other co-receptor, but it is important to remember that the commitment to the CD4 + or CD8 + lineage is also accompanied by a gene expression program specific to helper T cell functions (for CD4 + class II MHC-restricted cells) or cytotoxic T cell functions (for CD8 + class I MHC-restricted cells). The cells can begin expressing some of the effector genes associated with these distinctive pathways in vivo, long before their maturation is complete (




, 275 ,

and 277 ). In the long run, these sublineage-specific molecules that do not participate in antigen binding may provide the best indicators of the timing of the commitment of T-cell precursors to the CD4 and CD8 276

SP lineages. As a general rule, in the rare cases in which the co-receptor is mismatched with the MHC restriction, the function of the cells is matched with the co-receptor, not with the specificity of the TCR (




, 279 ).

Genes integral to the function of CD8 + cytotoxic T cells are expressed starting in the DP thymocytes, increasing only in cells that down-regulate CD4 mRNA (

). This suggests that down-regulation of the CD4 co-receptor and initiation of a program of cytotoxic gene expression occur simultaneously and are 280

normally coordinated. However, CD4 down-regulation is not required for the establishment of a gene expression pattern specific to cytotoxic cells. Deletion of the CD4 silencer prevents the down-regulation of CD4 that normally occurs in the CD8 lineage, but the CD4 + CD8 + cells that result develop as functional cytotoxic T cells (


). Conversely, helper and cytotoxic T cells can develop in the absence of CD4 or

CD8, respectively. In spite of the early flashes of effector gene expression, full functional maturation occurs over a substantially longer time than does the rescue of DP cells from death by neglect. The cells do not show full responsiveness to challenge with a TCR ligand and do not become resistant to apoptosis until the late stages of post–positive-selection processing, when the cells finally down-regulate CD24 (HSA) and CD69 and lose the markers that have distinguished them phenotypically from peripheral T cells. The process through which this occurs is still poorly understood. Perhaps what is needed is just time to reverse the numerous physiological adjustments that give DP thymocytes their unusual responses to stimulation. This

could be a matter of waiting for key inhibitory molecules to decay physically. The process could also be more complex in itself, for example, if it involved a cascade of reciprocal cell–cell interactions with the medullary stroma. Finally, it remains possible that functional maturation includes a component of selection. Can helper or killer differentiation begin in a way that is mismatched with co-receptor or TCR specificity? If so, maturation could include rewarding (with survival) cells that happen to possess the right combinations of effector programming, co-receptor expression, and TCR specificity. In this case, effector maturation would be the actual conclusion of the CD4/CD8 lineage choice. Finally, to detect the success of positive selection, the cells have to receive survival signals that make them permanently long-lived, a dramatic contrast from the DP thymocyte state. Sustained TCR/Lck signaling may provide much of this survival function for CD4 SP cells. For CD8 cells, the discontinuous or hit-and-run TCR/Lck signal may need to be supplemented by additional survival mechanisms. This would account for the enhanced importance of Bcl-2, IL-7R, and possibly Notch-family functions in CD8 SP lineage cells, as we have already noted, especially in the later period after CD8-lineage differentiation has begun. CD4 + and CD8 + mature cells continue to use different homeostatic maintenance and proliferation signals long after they leave the thymus, during their functions in the periphery; therefore, it is not surprising if a divergence is seen in these functions during maturation after positive selection. The mature cells finally leave the thymus in a nonsynchronized manner, about 7 to 14 days after beginning positive selection ( 282 , 283 ). The emigration appears to depend on a G-protein–coupled chemokine receptor, possibly CCR7 (

284 , 285 ). Very little is known yet about the signals that tell the cells when they are ready to leave. The developmental events that occur in the medulla, including the

coordination of maturation and export, currently remain one of the important areas of mystery in T-cell development. Relationships between Positive Selection, Negative Selection, and CD4/CD8 Lineage Choice Positive selection, negative selection, and CD4/CD8 lineage choice all depend to some extent on quantitative aspects of signaling through the TCR. This leads to an obvious question of how the cells can interpret a particular TCR engagement to make the correct decision. If a cell with a class I MHC-restricted TCR has a particularly high affinity for intrathymic class I MHC/peptide complexes, how does that cell determine whether to be deleted or whether to be converted into a CD4 SP cell? The cells do appear to make the positive/negative selection decision independently of the CD4/CD8 decision. In transgenic TCR models, the deletion of self-reactive class I MHC-restricted cells occurs without any obvious redirection into the CD4 lineage (the CD4 SP cells that can be seen in such transgenic individuals generally do not use the transgenic TCR). In principle, it may be possible to obtain four different outcomes from the single variable of signal strength, but the range of values directing negative selection, CD4 positive selection, and CD8 positive selection would each have to be tightly calibrated to distinguish them from one another as well as from death by neglect ( Fig. 14A). Changes in TCR or co-receptor surface density, for example, would need to

be held within sharply limited ranges. As far as can be determined, the system appears much more forgiving. Although TCR and co-receptor levels are finely regulated in normal development, transgenic manipulation can distort these levels quite substantially without altering the direction or predictability of CD4/CD8 lineage choice.

FIG. 14. Three schematic views of the signals that may determine thymocyte developmental fate during positive and negative selection. A: A simple stimulus strength model is depicted, in which the only parameter determining life, death, CD4 single positive (SP) cell fate, or CD8 SP cell fate is the intensity of signals through the T-cell receptor (TCR) and co-receptors. In this model, the cells compute TCR signaling intensity in the cortex and determine their fates immediately. B: A model in which stimulus strength and stimulus duration help the cell to distinguish between CD4 and CD8 cell fates and between these maturation pathways and death by neglect and by negative selection. Note the range of stimulus strengths (˜TCR/co-receptor affinities) that can be used for positive selection in this model, wider than in the simple stimulus strength model ( A) and the overlap in peak stimulus strength values (x-axis) that can give rise to either CD4 or CD8 cells. C: A model in which stimulus strength, stimulus duration, and the presence or absence of co-stimulation distinguish between CD4 and CD8 SP fates, neglect, and deletional and nondeletional forms of tolerance induction. This model is based on the idea that, for many cells, positive and negative selection choices can be encountered sequentially (also see Fig. 10, sequential threshold model). In this figure, high levels of co-stimulation that can lead to negative selection ( gray areas on plots) are encountered only in the medulla, which cells are only allowed to enter if they have first been positively selected as CD4 SP or CD8 SP cells in the cortex.

This, then, is an important attraction of the kinetic signaling/signal duration models: They add a second, independent parameter for the cells to use combinatorially with signal strength to compute the appropriate CD4/CD8 lineage choice ( Fig. 14B). In a similar way, the importance of co-stimulatory interactions in many forms of negative selection gives the cells yet another independent parameter to help even further to distinguish the negative selection threshold from the CD4/CD8 differentiation threshold in vivo ( Fig.

14C). In CD4/CD8 lineage choice, there is a hint that in the repeated contacts the cells make with MHC/peptide ligands over the several days of positive selection, the balance of these pathways can change further. At least in the CD8 cell pathway, this would be caused by the increasing surface density of TCR complexes at the same time that glycosylation changes are reducing the ability of the CD8 to bind class I MHC on its own (see the section on positive and negative selection). A predicted effect would be that the signals delivered at the beginning of positive selection would be dominated more by the co-receptors and their signaling mediators, whereas the signals delivered at the end of positive selection/maturation would be dominated more by the TCR. Thus, depending on when cells actually commit to a CD4 or CD8 SP fate, the composition of the “TCR signal” at the time of positive selection could shift substantially by the time that the maturing cells, now in the medulla, encounter their last negative selection threshold. The need for survival functions after lineage commitment, which may be specialized for each lineage, offers a way to reconcile the more instructive models with key evidence for the selective model. Any experimental manipulation that removes the survival mechanism for CD4 or CD8 cells after lineage commitment would indeed lead to elimination of those cells and could be defined as a mismatch between programming and survival receptors. There is another implication as well. The asymmetrical survival requirements of CD4 and CD8 SP cells mean that failure of survival (i.e. negative selection) could be effected by different mechanisms, too. Thus, the mechanism needed for negative selection may itself be an output of the CD4/CD8 lineage choice.

FRONTIERS FOR THE FUTURE: MYSTERIES AND ALTERNATIVES IN T-CELL DEVELOPMENT There is some evidence that the main stream of thymocyte selection, branched and complex as it is, may not be the only pathway through which T cells mature. In the future, cells that take a somewhat different pathway may be recognized as extremely important for the regulation of the peripher al immune system and thus for human health. It is currently hard to be certain whether all T-lineage cells follow the same program or not. The more than 95% loss of each cohort of cortical thymocytes makes it difficult to track particular precursors from before ß selection to maturation. One kind of evidence, however, suggests that there do exist substreams of T-cell development that are characterized by different surface marker expression patterns than are the majority of thymocytes. Some apparent intermediates in the transition from DN to DP to SP, for example, appear to preserve expression of the DN1/DN2 growth factor receptors c-kit and IL-7R all the way throughout ß selection and positive selection ( +


, 286 ). These c-kit

IL-7R + DP-like thymocytes appear to have better odds of maturing than do the mainstream c-kit - IL-7R

DP thymocytes. It is still unknown whether such cells develop into different kinds of T cells from the mainstream or whether they simply represent a lucky minority that have happened to undergo ß selection -

early, before loss of key survival-enhancing functions.

One reason to consider that these cells could have a distinctive fate is that there are at least two unconventional minority subsets of TCRaß T cells that can be generated in the thymus, with functional properties clearly distinct from the majority. Both cell types appear to play important roles in the regulation of immune responses and the maintenance of peripheral self-tolerance. For this reason, the immunological functions and developmental history of these subsets are currently topics of great interest. Both lineages, it seems, could arise from variations in the mechanisms controlling positive or negative selection or both. Alternative Pathway or Distinct Precursors: The Case of the NK T Cells NK T cells are adult T cells that share characteristics with NK cells, most notably the expression of surface markers CD161 (NK1.1) and CD122 (IL-2Rß); killer inhibitory receptors Ly49A, Ly49C/I, and Ly49G2; and cytolytic molecules perforin and granzyme (

, 288 and 289 ). These cells appear to be important in controlling autoimmunity as well as immune responses to tumors and infectious diseases, 287

perhaps because they produce very high levels of IL-4 and interferon-? upon first encounter with antigen, unlike conventional T cells. They represent a kind of bridge from the adaptive to the innate immune system. More than 80% of the NK T cells in the thymus and liver express an invariant Va14-Ja281; have a bias in Vß usage to Vß8, Vß2, and Vß7; and are specific for glycolipids bound by the nonclassical class I MHC relative CD1d. CD1d tetramers, which bind to the synthetic glycolipid a-galactosyl-ceramide (a-Gal-Cer), have been used successfully to identify these cells independently of NK1.1 (

, 291 ). Most studies have shown that NK T cells do not appear until late in ontogeny, long after the first conventional 290

aß T cells are seen. NK T cells binding the CD1d/glycolipid tetramers are not detectable in the fetal thymus at all and first appear 6 days after birth, increasing about 12- to 14-fold to 0.6% to 0.7% of thymocytes by 5 to 6 weeks of age ( 292 ). A striking thing about these cells is that they are mostly CD4 + or DN, although they recognize a class I MHC-type ligand. Beyond this CD1d-restricted NK1.1 + majority, there are other cells classified as NK T cells that are highly heterogeneous with regard to expression of different surface markers, TCR repertoire, CD1d dependence, tissue localization, and even the NK1.1 expression that has typically been used to define the population. In addition, small populations of CD8 + NK1.1 + and even TCR?d + NK1.1 + cells have also been identified. This heterogeneity and the low abundance of these cells have made the study of NK T-cell function and development particularly difficult. Most DN and CD4 + NKT cells require a thymus for development, although some NK T cells may be generated extrathymically ( 293 ). NK T cells do not develop in neonatally thymectomized mice (


) or

athymic nude mice ( 295 ), whereas they do develop in fetal thymic organ cultures ( 296 ). NK T cells probably represent a separate sublineage (or sublineages), inasmuch as they differ from conventional aß T cells in sensitivity to various mutations. Dominant-negative mutants of Ras and Mek do not appear to affect NK T-cell development ( 220 ), in sharp contrast to their effects on mainstream aß T cells, whereas

NK T cells do not develop in mice with Fyn -/- ( 297 , 298 ) and transcription factor Ets1 -/- ( 299 ), mutations that minimally affect aß T cells in general. NK T cells are also dependent on IL-15/IL-15R signals, like NK cells ( 300 ), and on lymphotoxin (LT)/LTßR (


). On the other hand, NK T cells require all the genetic

functions needed by conventional aß T cells. They are also dependent on the presence of pTa (



which suggests that they may go through ß selection. Positive selection of these cells is unusual; they are selected by interaction with a glycophospholipid antigen presented by a nonclassical class I MHC antigen, CD1d, and the interaction is with CD1d on hematopoietic cells, rather than epithelial cells. CD8 expression apparently makes these cells susceptible to negative selection and is thus generally excluded from the mature NK T-cell population. It may be that the unusual selection of these cells represents one case in which high-affinity class I MHC-restricted T cells are allowed to escape death by conversion to CD4 SP lineage after all. The combination of effector properties they wield, a mixture of T 2 cytokines and NK cell–like cytolytic functions, makes it difficult to H

use these functions as criteria for deciding whether these cells are “naturally” unusual helper-lineage or unusual killer-lineage cells. The developmental origins of NK T cells are controversial, and two models have been proposed [reviewed by Eberl et al. (

)]. According to the “precommitment model,” NK T cells arise from a distinct committed precursor, before TCR rearrangement, which then rearranges a semi-invariant TCR for 303

selection by CD1d molecules ( 304 ). In support of this model, ZAP70-deficient mice, which are blocked at ß selection, were found to accumulate NK1.1 + TCRß - thymocytes that had not undergone D-J or V-D-J rearrangements (


). When purified and stimulated with phorbol ester and ionomycin in neonatal thymic

organ culture, NK1.1 + TCRß + cells were generated, some with Va14-Ja281 rearrangements. In addition, NK1.1 + cells have been detected in mice with CD3?- and/or p56 Lck-deficiencies, which are also blocked at ß selection. These cells were found to have made V-D-J rearrangements in the TCRß locus but express no surface TCRß protein, which suggests that they may be blocked precommitted NK T cells ( ). One problem with this model is that it provides no mechanism to explain why the Va14-Ja281 rearrangement is so common in NK T cells. Although use of a recurrent rearrangement is reminiscent of 306

the fetal-type TCR?d lineages, there is no evidence that Va14-Ja281 rearrangement is directed in NK T cells ( 307 ). The second model is an instructive model in which NK T cells arise from the “mainstream” population of immature T cells with random TCR gene rearrangements that are positively selected by CD1d binding, rather than by conventional class I or II MHC, at the DP stage ( 287 ). The predominance of Va14-Ja281 rearrangements would then be a result of selection for the few cells bearing TCRs with affinity for CD1d, which then diverts the cells to the NK T lineage. CD1d tetramer–binding NK T cells can develop from tetramer-nonbinding CD4 + CD8 + TCRß - thymocytes sorted from 1-week-old mice, and they develop only when injected intrathymically into CD1d + recipients, which demonstrates that NK T cells can arise from DP cells and that they do require some kind of positive selection ( NK T cells undergo this developmental pathway remains to be seen.


). Whether all

Variations on a Theme of Tolerance: Regulatory T Cells Another subset of T cells that are potent suppressors of organ-specific autoimmunity have been id entified and characterized (

, 309 ). These CD4 + aßTCR cells express the activation surface marker CD25 + (IL-2Ra), as do mature peripheral T cells that are anergic as a result of partial stimulation. They constitute 308

approximately 5% of mature thymic CD4 SP cells (

). These cells also clearly require a thymus for their generation and appear to arise at the CD4 SP stage of development, during the time of negative selection 310

and maturation, before export into the periphery ( 311 ,


). When isolated from the thymus, these cells

are capable of suppressive activity similar to that found in peripheral CD4 + CD25 + cells. These T regulatory cells require IL-2 for development or survival, or both (



There is some evidence that the affinity of the TCR to self peptide and MHC is of critical importance in the generation of CD4 + CD25 + cells ( 309 , 313 , 314 ). Antigen presentation by thymic medullary epithelial cells has been shown to be capable of rendering T-cell populations tolerant even when it does not result in deletion ( 315 ). Thus, it is possible that tolerance induction by particular domains of the thymic stroma and distinctive profiles of TCR affinity translate into a distinct anergic and suppressive fate. The characteristic T regulatory cell properties are established in these T cells by their expression of the transcription factor Foxp3 (Scurfin) ( 315a, 315b, 315c). This factor antagonizes conventional T-cell activation responses and not only distinguishes the T regulatory subset but also seems to be essential for its generation. Mutant mice with defects in this gene are subject to lethal autoimmunity as a result. Foxp3 expression alone may even be sufficient to convert conventional naïve T cells to T regulatory-like cells. An explanation for regulatory T cell development may thus emerge if specific cues from the thymic stroma can be linked to Foxp3 induction. In view of the increasing importance of both NK T cells and regulatory T cells in systemic immune responses, filling in the picture of their development and selection is likely to become an active research frontier.

CONCLUDING REMARKS We have focused in depth on several aspects of intrathymic T-cell development that are particularly significant in terms of developmental mechanisms or immunological impact. These are areas in which work since 1995 has offered new glimpses of understanding how a momentous developmental choice or transition will be made. But it is worth returning to the larger picture of T-cell development sketched in Fig. 2 . In overview, it is striking how much functional diversity and lineage choice remain to cells after they have undergone T-lineage commitment. The cells take advantage of the intricate architecture of the thymus to migrate from one domain to another, using a core group of signaling molecules, carefully modulated interactions with environmental ligands, and a persistent set of transcription factors, to refine

progressively what kind of T cell their initial T-lineage commitment will produce. Of course, the interactions that come to dominate the second half of T-cell development are TCR-mediated ones, but the responses they trigger depend on genes such as Notch, GATA-3, IL-7R, bHLH factors, Ikaros, and Lck kinase and its signaling partners, genes that have been used and reused from the earliest stages of T-cell development. T cells equip themselves with a well-stocked developmental toolkit that they use to generate the diverse regulatory functions that characterize our adaptive immune systems.

ACKNOWLEDGMENTS The authors thank Tania Dugatkin for excellent help with the figures and the authors who kindly allowed their work to be reproduced for this chapter. We apologize to our many colleagues whose work and thoughtful interpretations we could not discuss adequately because of space limitations. Work in the authors’ laboratory was supported by grants from the U.S. Public Health Service (AG13108 and CA92033), from the National Science Foundation (MCB-9983129), and from the National Aeronautics and Space Administration (NAG 2-1370), with initial support from a consortium of the Stowers Institute for Medical Research. REFERENCES 1.

Hoffman ES, Passoni L, Crompton T, et al. Productive T-cell receptor ß-chain gene rearrangement: coincident regulation of cell

cycle and clonality during development in vivo. Genes Dev 1996;10:948–962. 2.

Penit C, Lucas B, Vasseur F. Cell expansion and growth arrest phases during the transition from precursor (CD4 -8 -) to

immature (CD4 +8 +) thymocytes in normal and genetically modified mice. J Immunol 1995;154:5103–5113. 3.

Di Santo JP, Rodewald H-R. In vivo roles of receptor tyrosine kinases and cytokine receptors in early thymocyte development.

Curr Opin Immunol 1998;10:196–207. 4.

Maraskovsky E, O’Reilly LA, Teepe M, et al. Bcl-2 can rescue T lymphocyte development in interleukin-7 receptor–deficient

mice but not in mutant RAG-1 -/- mice. Cell 1997;89:1011–1019. 5.

von Freeden-Jeffry U, Solvason N, Howard M, et al. The earliest T lineage-committed cells depend on IL-7 for Bcl-2 expression

and normal cell cycle progression. Immunity 1997;7:147–154. 6.

Candéias S, Muegge K, Durum SK. IL-7 receptor and VDJ recombination: trophic versus mechanistic actions. Immunity

1997;6:501–508. 7.

Di Santo JP, Aifantis I, Rosmaraki E, et al. The common cytokine receptor ? chain and the pre–T cell receptor provide

independent but critically overlapping signals in early a/ß T cell development. J Exp Med 1999;189:563–573. 8.

Anderson G, Moore NC, Owen JJ, et al. Cellular interactions in thymocyte development. Annu Rev Immunol 1996;14:73–99.


Lind EF, Prockop SE, Porritt HE, et al. Mapping precursor movement through the postnatal thymus reveals specific

microenvironments supporting defined stages of early lymphoid development. J Exp Med 2001;194:127–134.


Derbinski J, Schulte A, Kyewski B, et al. Promiscuous gene expression in medullary thymic epithelial cells mirrors the

peripheral self. Nat Immunol 2001;2:1032–1039. 11.

Gounari F, Aifantis I, Khazaie K, et al. Somatic activation of ß-catenin bypasses pre-TCR signaling and TCR selection in

thymocyte development. Nat Immunol 2001;2:863–869. 12.

Ioannidis V, Beermann F, Clevers H, et al. The ß-catenin—TCF-1 pathway ensures CD4 +CD8 + thymocyte survival. Nat

Immunol 2001;2:691–697. 13.

Kitazawa H, Muegge K, Badolato R, et al. IL-7 activates a4ß1 integrin in murine thymocytes. J Immunol 1997;159:2259–2264.


Norment AM, Bogatzki LY, Gantner BN, et al. Murine CCR9, a chemokine receptor for thymus-expressed chemokine that is

up-regulated following pre-TCR signaling. J Immunol 2000;164:639–648. 15.

Uehara S, Song K, Farber JM, et al. Characterization of CCR9 expression and CCL25/thymus-expressed chemokine

responsiveness during T cell development: CD3 highCD69 + thymocytes and ?dTCR + thymocytes preferentially respond to CCL25. J Immunol 2002;168:134–142. 16.

Wurbel MA, Malissen M, Guy-Grand D, et al. Mice lacking the CCR9 CC-chemokine receptor show a mild impairment of early

T- and B-cell development and a reduction in T-cell receptor ?d + gut intraepithelial lymphocytes. Blood 2001;98:2626–2632. 17.

Manley NR. Thymus organogenesis and molecular mechanisms of thymic epithelial cell differentiation. Semin Immunol

2000;12:421–428. 18.

Manaia A, Lemarchandel V, Klaine M, et al. Lmo2 and GATA-3 associated expression in intraembryonic hemogenic sites.

Development 2000;127:643–653. 19.

Reya T, Yang-Snyder JA, Rothenberg EV, et al. Regulated expression and function of CD122 (interleukin-2/interleukin-15 R-ß)

during lymphoid development. Blood 1996;87:190–201. 20.

Suniara RK, Jenkinson EJ, Owen JJ. Studies on the phenotype of migrant thymic stem cells. Eur J Immunol 1999;29:75–80.


Ikuta K, Kina T, NacNeil I, et al. A developmental switch in thymic lymphocyte maturation potential occurs at the level of

hematopoietic stem cells. Cell 1990;62:863–874. 22.

Bain G, Romanow WJ, Albers K, et al. Positive and negative regulation of V(D)J recombination by the E2A proteins. J Exp

Med 1999;189:289–300. 23.

Dzierzak E, Medvinsky A, de Bruijn M. Qualitative and quantitative aspects of haematopoietic cell development in the

mammalian embryo. Immunol Today 1998;19:228–236. 24.

Jotereau F, Heuze F, Salomon-Vie V, et al. Cell kinetics in the fetal mouse thymus: precursor cell input, proliferation, and

emigration. J Immunol 1987;138:1026–1030. 25.

Verbeek S, Izon D, Hofhuis F, et al. An HMG-box–containing T-cell factor required for thymocyte differentiation. Nature

1995;374:70–74. 26.

Wang J-H, Nichogiannopoulou A, Wu L, et al. Selective defects in the development of the fetal and adult lymphoid system in

mice with an Ikaros null mutation. Immunity 1996;5:537–549. 27.

McKercher SR, Torbett BE, Anderson KL, et al. Targeted disruption of the PU.1 gene results in multiple hematopoietic

abnormalities. EMBO J 1996;15:5647–5658. 28.

Nakayama K, Nakayama K, Negishi I, et al. Disappearance of the lymphoid system in Bcl-2 hom*ozygous mutant chimeric

mice. Science 1993;261:1584–1588. 29.

Veis DJ, Sorenson CM, Shutter JR, et al. Bcl-2-deficient mice demonstrate fulminant lymphoid apoptosis, polycystic kidneys,

and hypopigmented hair. Cell 1993;75:229–240. 30.

Ody C, Alais S, Corbel C, et al. Surface molecules involved in avian T-cell progenitor migration and differentiation. Dev

Immunol 2000;7:267–277. 31.

Chretien I, Marcuz A, Courtet M, et al. CTX, a Xenopus thymocyte receptor, defines a molecular family conserved throughout

vertebrates. Eur J Immunol 1998;28:4094–4104. 32.

Res P, Spits H. Developmental stages in the human thymus. Semin Immunol 1999;11:39–46.


Spits H, Blom B, Jaleco AC, et al. Early stages in the development of human T, natural killer and thymic dendritic cells.

Immunol Rev 1998;165:75–86. 34.

Plum J, De Smedt M, Verhasselt B, et al. Human T lymphopoiesis. In vitro and in vivo study models. Ann N Y Acad Sci

2000;917:724–731. 35.

Mitnacht R, Bischof A, Torres-Nagel N, et al. Opposite CD4/CD8 lineage decisions of CD4 +8 + mouse and rat thymocytes to

equivalent triggering signals: correlation with thymic expression of a truncated CD8 alpha chain in mice but not rats. J Immunol 1998;160:700–707. 36.

Hunig T. Cross-linking of the T cell antigen receptor interferes with the generation of CD4 +8 + thymocytes from their

immediate CD4 -8 + precursors. Eur J Immunol 1988;18:2089–2092. 37.

Yang SY, Denning SM, Mizuno S, et al. A novel activation pathway for mature thymocytes. Costimulation of CD2 (T,p50) and

CD28 (T,p44) induces autocrine interleukin 2/interleukin 2 receptor–mediated cell proliferation. J Exp Med 1988;168:1457–1468. 38.

Ramsdell F, Jenkins M, Dinh Q, et al. The majority of CD4 +8 - thymocytes are functionally immature. J Immunol

1991;147:1779–1785. 39.

Wu L, Antica M, Johnson GR, et al. Developmental potential of the earliest precursor cells from the adult mouse thymus. J Exp

Med 1991;174:1617–1627. 40.

Ardavin C, Wu L, Li CL, et al. Thymic dendritic cells and T cells develop simultaneously in the thymus from a common

precursor population. Nature 1993;362:761–763. 41.

Kawamoto H, Ohmura K, Katsura Y. Direct evidence for the commitment of hematopoietic stem cells to T, B, and myeloid

lineages in murine fetal liver. Int Immunol 1997;9:1011–1019. 42.

Res P, Martinez-Caceres E, Jaleco AC, et al. CD34 +CD38 dim cells in the human thymus can differentiate into T, natural killer,

and dendritic cells but are distinct from pluripotent stem cells. Blood 1996;87:5196–5206. 43.

Sanchez MJ, Muench MO, Roncarolo MG, et al. Identification of a common T/natural killer cell progenitor in human fetal

thymus. J Exp Med 1994;180:569–576. 44.

Rodewald HR, Kretzschmar K, Takeda S, et al. Identification of pro-thymocytes in murine fetal blood: T lineage commitment

can precede thymus colonization. EMBO J 1994;13:4229–4240. 45.

Kawamoto H, Ohmura K, Katsura Y. Cutting edge: presence of progenitors restricted to T, B, or myeloid lineage, but absence

of multipotent stem cells, in the murine fetal thymus. J Immunol 1998;161:3799–3802. 46.

Carding SR, Kyes S, Jenkinson EJ, et al. Developmentally regulated fetal thymic and extrathymic T-cell receptor ?d gene

expression. Genes Dev 1990;4:1304–1315. 47.

Lefrancois L, Puddington L. Extrathymic intestinal T-cell development: virtual reality? Immunol Today 1995;16:16–21.


Bendelac A. Mouse NK1 + T cells. Curr Opin Immunol 1995;7:367–374.


Dejbakhsh-Jones S, Garcia-Ojeda ME, Chatterjea-Matthes D, et al. Clonable progenitors committed to the T lymphocyte

lineage in the mouse bone marrow; use of an extrathymic pathway. Proc Natl Acad Sci U S A 2001;98:7455–7460. 50.

Kondo M, Weissman IL, Akashi K. Identification of clonogenic common lymphoid progenitors in mouse bone marrow. Cell

1997;91:661–672. 51.

Litman GW, Anderson MK, Rast JP. Evolution of antigen binding receptors. Annu Rev Immunol 1999;17:109–147.


Miracle AL, Anderson MK, Litman RT, et al. Complex expression patterns of lymphocyte-specific genes during the

development of cartilaginous fish implicate unique lymphoid tissues in generating an immune repertoire. Int Immunol 2001;13:567–580. 53.

Wu L, Li C-L, Shortman K. Thymic dendritic cell precursors: relationship to the T lymphocyte lineage and phenotype of the

dendritic cell progeny. J Exp Med 1996;184:903–911. 54.

Carlyle JR, Michie AM, Furlonger C, et al. Identification of a novel developmental stage marking lineage commitment of

progenitor thymocytes. J Exp Med 1997;186:173–182. 55.

Katsura Y, Kawamoto H. Stepwise lineage restriction of progenitors in lympho-myelopoiesis. Int Rev Immunol 2001;20:1–20.


Lacaud G, Carlsson L, Keller G. Identification of a fetal hematopoietic precursor with B cell, T cell, and macrophage potential.

Immunity 1998;9:827–838. 57.

Wilson A, Ferrero I, MacDonald HR, et al. Cutting edge: an essential role for Notch-1 in the development of both

thymus-independent and -dependent T cells in the gut. J Immunol 2000;165:5397–5400. 58.

Jaleco AC, Neves H, Hooijberg E, et al. Differential effects of Notch ligands Delta-1 and Jagged-1 in human lymphoid

differentiation. J Exp Med 2001;194:991–1002. 59.

Pui JC, Allman D, Xu L, et al. Notch1 expression in early lymphopoiesis influences B versus T lineage determination. Immunity

1999;11:299–308. 60.

Varnum-Finney B, Xu L, Brashem-Stein C, et al. Pluripotent, cytokine-dependent, hematopoietic stem cells are immortalized by

constitutive Notch1 signaling. Nat Med 2000;6:1278–1281. 61.

Radtke F, Wilson A, Stark G, et al. Deficient T cell fate specification in mice with an induced inactivation of Notch1. Immunity

1999;10:547–558. 62.

Koch U, Lacombe TA, Holland D, et al. Subversion of the T/B lineage decision in the thymus by lunatic fringe–mediated

inhibition of Notch-1. Immunity 2001;15:225–236. 63.

Wilson A, MacDonald HR, Radtke F. Notch 1-deficient common lymphoid precursors adopt a B cell fate in the thymus. J Exp

Med 2001;194:1003–1012. 64.

Doerfler P, Shearman MS, Perlmutter RM. Presenilin-dependent ?-secretase activity modulates thymocyte development. Proc

Natl Acad Sci U S A 2001;98:9312–9317. 65.

Hadland BK, Manley NR, Su D, et al. ?-secretase inhibitors repress thymocyte development. Proc Natl Acad Sci U S A

2001;98:7487–7491. 66.

Ikawa T, Kawamoto H, Fujimoto S, et al. Commitment of common T/natural killer (NK) progenitors to unipotent T and NK

progenitors in the murine fetal thymus revealed by a single progenitor assay. J Exp Med 1999;190:1617–1625. 67.

Michie AM, Carlyle JR, Schmitt TM, et al. Clonal characterization of a bipotent T cell and NK cell progenitor in the mouse fetal

thymus. J Immunol 2000;164:1730–1733. 68.

Shortman K, Vremec D, Corcoran LM, et al. The linkage between T-cell and dendritic cell development in the mouse thymus.

Immunol Rev 1998;165:39–46. 69. 69a.

Lee C-K, Kim JK, Kim Y, et al. Generation of macrophages from early T progenitors in vitro. J Immunol 2001;166:5964–5969. Wang H, Rothenberg EV. Gene expression profiles distinguish multiple subsets of DN1 thymocytes, California Institute of

Technology, 2000. 70.

Wang H, Diamond RA, Rothenberg EV. Cross-lineage expression of Ig-ß (B29) in thymocytes: positive and negative gene

regulation to establish T-cell identity. Proc Natl Acad Sci U S A 1998;95:6831–6836. 71.

Chen F, Rowen L, Hood L, et al. Differential transcriptional regulation of individual TCR V beta segments before gene

rearrangement. J Immunol 2001;166:1771–1780. 72.

Shimizu C, Kawamoto H, Yamash*ta M, et al. Progression of T cell lineage restriction in the earliest subpopulation of murine

adult thymus visualized by the expression of lck proximal promoter activity. Int Immunol 2001;13:105–117. 73.

Lucas K, Vremec D, Wu L, et al. A linkage between dendritic cell and T-cell development in the mouse thymus: the capacity of

sequential T-cell precursors to form dendritic cells in culture. Dev Comp Immunol 1998;22:339–349. 74.

Res PCM, Couwenberg F, Vyth-Dreese FA, et al. Expression of a pTa mRNA in a committed dendritic cell precursor in the

human thymus. Blood 1999;94:2647–2657.


Hu M, Krause D, Greaves M, et al. Multilineage gene expression precedes commitment in the hemopoietic system. Genes Dev

1997;11:774–785. 76.

Delassus S, Titley I, Enver T. Functional and molecular analysis of hematopoietic progenitors derived from the

aorta-gonad-mesonephros region of the mouse embryo. Blood 1999;94:1495–1503. 76a.

Gounari F, Aifantis I, Martin C, Fehling HJ, Hoeflinger S, Leder P, von Boehmer H, Reizis B. Tracing lymphopoiesis with the

aid of a pTa-controlled reporter gene. Nat Immunol 2002;3:489–96. 76b.

Igarashi H, Gregory SC, Yokota T, Sakaguchi N, Kincade PW. Transcription from the RAG1 locus marks the earliest

lymphocyte progenitors in bone marrow. Immunity 2002;17:117–30. 77.

Rodewald H-R, Brocker T, Haller C. Developmental dissociation of thymic dendritic cell and thymocyte lineages revealed in

growth factor receptor mutant mice. Proc Natl Acad Sci U S A 1999;96:15068–15073. 78.

Kuo CT, Leiden JM. Transcriptional regulation of T lymphocyte development and function. Annu Rev Immunol

1999;17:149–187. 79.

Staal FJ, Weerkamp F, Langerak AW, et al. Transcriptional control of T lymphocyte differentiation. Stem Cells

2001;19:165–179. 80.

Rothenberg EV, Anderson MK. Elements of transcription factor network design for T-lineage specification. Devel Biol

2002;246:29–44. 81.

Rothenberg EV, Telfer JC, Anderson MK. Transcriptional regulation of lymphocyte lineage commitment. Bioessays

1999;21:726–742. 82.

Engel I, Murre C. The function of E- and Id proteins in lymphocyte development. Nat Rev Immunol 2002;1:193–199.


Tomita K, Hattori M, Nakamura E, et al. The bHLH gene Hes1 is essential for expansion of early T cell precursors. Genes Dev

1999;13:1203–1210. 84.

Reizis B, Leder P. The upstream enhancer is necessary and sufficient for the expression of the pre–T cell receptor a gene in

immature T lymphocytes. J Exp Med 2001;194:979–990. 85.

Kee BL, Murre C. Induction of early B cell factor (EBF) and multiple B lineage genes by the basic helix-loop-helix transcription

factor E12. J Exp Med 1998;188:699–713. 86.

Choi JK, Shen C-P, Radomska HS, et al. E47 activates the Ig-heavy chain and TdT loci in non–B cells. EMBO J

1996;15:5014–5021. 87.

Kee BL, Bain G, Murre C. IL-7Ra and E47: independent pathways required for development of multipotent lymphoid

progenitors. EMBO J 2002;21:103–113. 88.

Osborne B, Miele L. Notch and the immune system. Immunity 1999;11:653–663.


Kaneta M, Osawa M, Sudo K, et al. A role for Pref-1 and HES-1 in thymocyte development. J Immunol 2000;164:256–264.


Radtke F, Ferrero I, Wilson A, et al. Notch1 deficiency dissociates the intrathymic development of dendritic cells and T cells. J

Exp Med 2000;191:1085–1094. 91.

Colucci F, Samson SI, DeKoter RP, et al. Differential requirement for the transcription factor PU.1 in the generation of natural

killer cells versus B and T cells. Blood 2001;97:2625–2632. 92.

Bain G, Murre C. The role of E-proteins in B- and T-lymphocyte development. Semin Immunol 1998;10:143–153.


Engel I, Johns C, Bain G, et al. Early thymocyte development is regulated by modulation of E2A protein activity. J Exp Med

2001;194:733–746. 94.

Murphy KM, Ouyang W, Farrar JD, et al. Signaling and transcription in T helper development. Annu Rev Immunol

2000;18:451–494. 95.

Kruisbeek AM, Haks MC, Carleton M, et al. Branching out to gain control: how the pre-TCR is linked to multiple functions.

Immunol Today 2000;21:637–644. 96.

von Boehmer H, Aifantis I, Feinberg J, et al. Pleiotropic changes controlled by the pre–T-cell receptor. Curr Opin Immunol

1999;11:135–142. 97.

Levelt CN, Eichmann K. Receptors and signals in early thymic selection. Immunity 1995;3:667–672.


Borowski C, Martin C, Gounari F, et al. On the brink of becoming a T cell. Curr Opin Immunol 2002;14:200–206.


Rothenberg EV. The development of functionally responsive T cells. Adv Immunol 1992;51:85–214.


Rothenberg EV, Diamond RA, Chen D. Programming for recognition and programming for response: separate developmental

subroutines in the murine thymus. Thymus 1994;22:215–244. 101.

Voll RE, Jimi E, Phillips RJ, et al. NF-?B activation by the pre–T cell receptor serves as a selective survival signal in T

lymphocyte development. Immunity 2000;13:677–689. 102.

Gratiot-Deans J, Merino R, Nunez G, et al. Bcl-2 expression during T-cell development: Early loss and late return occur at

specific stages of commitment to differentiation and survival. Proc Natl Acad Sci U S A 1994;91:10685–10689. 103.

Linette GP, Grusby MJ, Hedrick SM, et al. Bcl-2 upregulated at the CD4 +CD8 + stage during positive selection and promotes

thymocyte differentiation at several control points. Immunity 1994;1:197–205. 104.

Andjelic S, Drappa J, Lacy E, et al. The onset of Fas expression parallels the acquisition of CD8 and CD4 in fetal and adult

alpha beta thymocytes. Int Immunol 1994;6:73–79. 105.

Chen D, Rothenberg EV. Molecular basis for developmental changes in interleukin-2 gene inducibility. Mol Cell Biol

1993;13:228–237. 106.

Rincon M, Flavell RA. Regulation of AP-1 and NFAT transcription factors during thymic selection of T cells. Mol Cell Biol

1996;16:1074–1084. 107.

Sen J, Shinkai Y, Alt FW, et al. Nuclear factors that mediate intrathymic signals are developmentally regulated. J Exp Med

1994;180:2321–2327. 108.

Chen F, Chen D, Rothenberg EV. Specific regulation of fos family transcription factors in thymocytes at two developmental

checkpoints. Int Immunol 1999;11:677–688. 109.

Davey GM, Schober SL, Endrizzi BT, et al. Preselection thymocytes are more sensitive to T cell receptor stimulation than

mature T cells. J Exp Med 1998;188:1867–1874. 110.

Lucas B, Stefanova I, Yasutomo K, et al. Divergent changes in the sensitivity of maturing T cells to structurally related ligands

underlies formation of a useful T cell repertoire. Immunity 1999;10:367–376. 111.

Vasquez NJ, Kane LP, Hedrick SM. Intracellular signals that mediate thymic negative selection. Immunity 1994;1:45–56.


Moody AM, Chui D, Reche PA, et al. Developmentally regulated glycosylation of the CD8aß coreceptor stalk modulates

ligand binding. Cell 2001;107:501–512. 113.

Daniels MA, Devine L, Miller JD, et al. CD8 binding to MHC class I molecules is influenced by T cell maturation and

glycosylation. Immunity 2001;15:1051–1061. 114.

Wilson A, MacDonald HR. Expression of genes encoding the pre-TCR and CD3 complex during thymus development. Int

Immunol 1995;7:1659–1664. 115.

Bruno L, Rocha B, Rolink A, et al. Intra- and extra-thymic expression of the pre–T cell receptor a gene. Eur J Immunol

1995;25:1877–1882. 116.

Malissen B, Malissen M. Functions of TCR and pre-TCR subunits: lessons from gene ablation. Curr Opin Immunol

1996;8:383–393. 117.

Koyasu S, Clayton LK, Lerner A, et al. Pre-TCR signaling components trigger transcriptional activation of a rearranged TCR

alpha gene locus and silencing of the pre-TCR alpha locus: implications for intrathymic differentiation. Int Immunol 1997;9:1475–1480. 118.

Groettrup M, von Boehmer H. A role for a pre–T-cell receptor in T-cell development. Immunol Today 1993;14:610–614.


Saint-Ruf C, Panigada M, Azogui O, et al. Different initiation of pre-TCR and ?dTCR signalling. Nature 2000;406:524–527.


Corre I, Gomez M, Vielkind S, et al. Analysis of thymocyte development reveals that the GTPase RhoA is a positive regulator

of T cell receptor responses in vivo. J Exp Med 2001;194:903–914. 121.

Chen J, Shinkai Y, Young F, et al. Probing immune functions in RAG-deficient mice. Curr Opin Immunol 1994;6:313–319.


Levelt CN, Ehrfeld A, Eichmann K. Regulation of thymocyte development through CD3. I. Timepoint of ligation of CD3e

determines clonal deletion or induction of developmental program. J Exp Med 1993;177:707–716. 123.

Gibbons D, Douglas NC, Barber DF, et al. The biological activity of natural and mutant pTa alleles. J Exp Med

2001;194:695–703. 124.

Irving BA, Alt FW, Killeen N. Thymocyte development in the absence of pre–T cell receptor extracellular immunoglobulin

domains. Science 1998;280:905–908.


Lacorazza HD, Tucek-Szabo C, Vasovic LV, et al. Premature TCRaß expression and signaling in early thymocytes impair

thymocyte expansion and partially block their development. J Immunol 2001;166:3184–3193. 126.

Mombaerts P, Anderson SJ, Perlmutter RM, et al. An activated lck transgene promotes thymocyte development in RAG-1

mutant mice. Immunity 1994;1:261–267. 127.

Swat W, Shinkai Y, Cheng H-L, et al. Activated Ras signals differentiation and expansion of CD4 +8 + thymocytes. Proc Natl

Acad Sci U S A 1996;93:4683–4687. 128.

Fehling HJ, Iritani BM, Krotkova A, et al. Restoration of thymopoiesis in pTa-/- mice by anti-CD3e antibody treatment or with

transgenes encoding activated Lck or tailless pTa. Immunity 1997;6:703–714. 129.

Iritani BM, Alberola-Ila J, Forbush KA, et al. Distinct signals mediate maturation and allelic exclusion in lymphocyte

progenitors. Immunity 1999;10:713–722. 130.

Gärtner F, Alt FW, Monroe R, et al. Immature thymocytes employ distinct signaling pathways for allelic exclusion versus

differentiation and expansion. Immunity 1999;10:537–546. 131.

Carleton M, Haks MC, Smeele SA, et al. Early growth response transcription factors are required for development of CD4

-CD8 - thymocytes to the CD4 +CD8 + stage. J Immunol 2002;168:1649–1658.


Ellmeier W, Sawada S, Littman DR. The regulation of CD4 and CD8 coreceptor gene expression during T cell development.

Annu Rev Immunol 1999;17:523–554. 132a.

Kioussis D, Ellmeier W. Chromatin and CD4, CD8A and CD8B gene expression during thymic differentiation. Nat Rev

Immunol 2002;2:909–19. 132b.

Harker N, Naito T, Cortes M, Hostert A, Hirschberg S, Tolaini M, Roderick K, Georgopoulos K, Kioussis D. The CD8a gene

locus is regulated by the Ikaros family of proteins. Mol Cell 2002;10:1403–15. 132c.

Taniuchi I, Osato M, Egawa T, Sunshine MJ, Bae S-C, Komori T, Ito Y, Littman DR. Differential requirements for Runx

proteins in CD4 repression and epigenetic silencing during T lymphocyte development. Cell 2002;111:621–33. 132d.

Chi TH, Wan M, Zhao K, Taniuchi I, Chen L, Littman DR, Crabtree GR. Reciprocal regulation of CD4/CD8 expression by

SWI/SNF-like BAF complexes. Nature 2002;418:195–9. 133.

Bain G, Cravatt CB, Loomans C, et al. Regulation of the helix-loop-helix proteins, E2A and Id3, by the Ras-ERK MAPK

cascade. Nat Immunol 2001;2:165–171. 134.

Blom B, Heemskerk MHM, Verschuren MCM, et al. Disruption of aß but not of ?d T cell development by overexpression of

the helix-loop-helix protein Id3 in committed T cell progenitors. EMBO J 1999;18:2793–2802. 135.

Heemskerk MHM, Blom B, Nolan G, et al. Inhibition of T cell and promotion of natural killer cell development by the dominant

negative helix loop helix factor Id3. J Exp Med 1997;186:1597–1602. 136.

Guidos CJ, Williams CJ, Grandal I, et al. V(D)J recombination activates a p53-dependent DNA damage checkpoint in scid

lymphocyte precursors. Genes Dev 1996;10:2038–2054.


Jiang D, Lenardo MJ, Züñiga-Pflücker JC. p53 prevents maturation to the CD4 +CD8 + stage of thymocyte differentiation in

the absence of T cell recepto r rearrangement. J Exp Med 1996;183:1923–1928. 138.

Haks MC, Krimpenfort P, van den Brakel JHN, et al. Pre-TCR signaling and inactivation of p53 induces crucial cell survival

pathways in pre–T cells. Immunity 1999;11:91–101. 139.

Petrie HT, Tourigny M, Burtrum DB, et al. Precursor thymocyte proliferation and differentiation are controlled by signals

unrelated to the pre-TCR. J Immunol 2000;165:3094–3098. 140.

Shores EW, Sharrow SO, Uppenkemp I, et al. T cell receptor–negative thymocytes from SCID mice can be induced to enter

the CD4/CD8 differentiation pathway. Eur J Immunol 1990;20:69–77. 141.

Trop S, De Sepulveda P, Züñiga-Pflücker JC, et al. Overexpression of suppressor of cytokine signaling-1 impairs pre–T-cell

receptor–induced proliferation but not differentiation of immature thymocytes. Blood 2001;97:2269–2277. 142.

Miyazaki T. Two distinct steps during thymocyte maturation from CD4 -CD8 - to CD4 +CD8 + distinguished in the early growth

response (Egr)-1 transgenic mice with a recombinase-activating gene-deficient background. J Exp Med 1997;186:877–885. 143.

Okamura RM, Sigvardsson M, Galceran J, et al. Redundant regulation of T cell differentiation and TCRa gene expression by

the transcription factors LEF-1 and TCF-1. Immunity 1998;8:11–20. 144.

Barndt R, Dai MF, Zhang Y. A novel role for HEB downstream or parallel to the pre-TCR signaling pathway during aß

thymopoiesis. J Immunol 1999;163:3331–3343. 145.

Pearson R, Weston K. c-Myb regulates the proliferation of immature thymocytes following ß-selection. EMBO J

2000;19:6112–120. 146.

Kabra NH, Kang C, Hsing LC, et al. T cell–specific FADD-deficient mice: FADD is required for early T cell development. Proc

Natl Acad Sci U S A 2001;98:6307–6312. 147.

Newton K, Harris AW, Strasser A. FADD/MORT1 regulates the pre-TCR checkpoint and can function as a tumour

suppressor. EMBO J 2000;19:931–941. 148.

Michie AM, Soh J-W, Hawley RG, et al. Allelic exclusion and differentiation by protein kinase C–mediated signals in immature

thymocytes. Proc Natl Acad Sci U S A 2001;98:609–614. 149.

Wu L, Strasser A. “Decisions, decisions…”: ß-catenin–mediated activation of TCF-1 and Lef-1 influences the fate of

developing T cells. Nat Immunol 2001;2:823–824. 150.

Chen C, Edelstein LC, Gelinas C. The Rel/NF-?B family directly activates expression of the apoptosis inhibitor Bcl-x . Mol L

Cell Biol 2000;20:2687–2695. 151.

Schmedt C, Tarakhovsky A. Autonomous maturation of a/ß T lineage cells in the absence of COOH-terminal Src kinase

(Csk). J Exp Med 2001;193:815–826. 152.

Georgopoulos K. Transcription factors required for lymphoid lineage commitment. Curr Opin Immunol 1997;9:222–227.


Brown KE, Guest SS, Smale ST, et al. Association of transcriptionally silent genes with Ikaros complexes at centromeric

heterochromatin. Cell 1997;91:845–854. 154.

Winandy S, Wu L, Wang J-H, et al. Pre–T cell receptor (TCR) and TCR-controlled checkpoints in T cell differentiation are set

by Ikaros. J Exp Med 1999;190:1039–1048. 155.

Jepsen K, Hermanson O, Onami TM, et al. Combinatorial roles of the nuclear receptor corepressor in transcription and

development. Cell 2000;102:753–763. 156.

Anderson MK, Hernandez-Hoyos G, Diamond RA, et al. Precise developmental regulation of Ets family transcription factors

during specification and commitment to the T cell lineage. Development 1999;126:3131–3148. 157.

Adachi S, Amasaki Y, Miyatake S, et al. Successive expression and activation of NFAT family members during thymocyte

differentiation. J Biol Chem 2000;275:14708–14716. 158.

Anderson G, Anderson KL, Conroy LA, et al. Intracellular signaling events during positive and negative selection of CD4

+CD8 + thymocytes in vitro. J Immunol 1995;154:3636–3543.


Aifantis I, Gounari F, Scorrano L, et al. Constitutive pre-TCR signaling promotes differentiation through Ca 2+ mobilization

and activation of NF-?B and NFAT. Nat Immunol 2001;2:403–409. 160.

Shao H, Kono DH, Chen LY, et al. Induction of the early growth response (Egr) family of transcription factors during thymic

selection. J Exp Med 1997;185:731–744. 161.

Dudley EC, Petrie HT, Shah LM, et al. T cell receptor ß chain gene rearrangement and selection during thymocyte

development in adult mice. Immunity 1994;1:83–93. 162.

Burtrum DB, Kim S, Dudley EC, et al. TCR gene recombination and aß-?d lineage divergence: productive TCR-ß

rearrangement is neither exclusive nor preclusive of ?d cell development. J Immunol 1996;157:4293–4296. 163.

Mombaerts P, Clarke AR, Rudnicki MA, et al. Mutations in T-cell antigen receptor genes a and ß block thymocyte

development at different stages. Nature 1992;360:225–231. 164.

Terrence K, Pavlovich CP, Matechak EO, et al. Premature expression of T cell receptor (TCR)aß suppresses TCR?d gene

rearrangement but permits development of ?d lineage T cells. J Exp Med 2000;192:537–548. 165.

Bruno L, Fehling HJ, von Boehmer H. The aß T cell receptor can replace the ?d receptor in the development of ?d lineage

cells. Immunity 1996;5:343–352. 166.

Capone M, Hockett RD Jr, Zlotnik A. Kinetics of T cell receptor ß ? and d rearrangements during adult thymic development: T

cell receptor rearrangements are present in CD44 +CD25 + pro-T thymocytes. Proc Natl Acad Sci U S A 1998;95:12522–12527. 167.

Wilson A, Capone M, MacDonald HR. Unexpectedly late expression of intracellular CD3e and TCR ?d proteins during adult

thymus development. Int Immunol 1999;11:1641–1650. 168.

Penninger J, Kishihara K, Molina T, et al. Requirement for tyrosine kinase p56 lck for thymic development of transgenic ?d T

cells. Science 1993;260:358–361. 169.

Wilson A, MacDonald HR. A limited role for ß-selection during ?d T cell development. J Immunol 1998;161:5851–5854.


Kang J, Volkmann A, Raulet DH. Evidence that ?d versus aß T cell fate determination is initiated independently of T cell

receptor signaling. J Exp Med 2001;193:689–698. 171.

Bruno L, Scheffold A, Radbruch A, et al. Threshold of pre–T-cell receptor surface expression is associated with aß T-cell

lineage commitment. Curr Biol 1999;9:559–568. 172.

Tanaka R, Takeuchi Y, Shiohara T, et al. In utero treatment with monoclonal antibody to IL-2 receptor ß-chain completely

abrogates development of Thy-1 + dendritic epidermal cells. Int Immunol 1992;4:487–491. 173.

Ye S-K, Maki K, Lee H-C, et al. Differential roles of cytokine receptors in the development of epidermal ?d T cells. J Immunol

2001;167:1929–1934. 174.

Leclercq G, Debacker V, De Smedt M, et al. Differential effects of interleukin-15 and interleukin-2 on differentiation of

bipotential T/natural killer progenitor cells. J Exp Med 1996;184:325–336. 175.

Rodewald H-R, Fehling HJ. Molecular and cellular events in early thymocyte development. Adv Immunol 1998;69:1–112.


Berg LJ, Kang J. Molecular determinants of TCR expression and selection. Curr Opin Immunol 2001;13:232–241.


Malissen M, Pereira P, Gerber DJ, et al. The common cytokine receptor ? chain controls survival of ?/d T cells. J Exp Med

1997;186:1277–1285. 178.

Durum SK, Candéias S, Nakajima H, et al. Interleukin 7 receptor control of T cell receptor ? gene rearrangement: role of

receptor-associated chains and locus accessibility. J Exp Med 1998;188:2233–2241. 179.

Kang J, Coles M, Raulet DH. Defective development of ?/d T cells in interleukin 7 receptor–deficient mice is due to impaired

expression of T cell receptor ? genes. J Exp Med 1999;190:973–982. 180.

Lee H-C, Ye S-K, Honjo T, et al. Induction of germline transcription in the human TCR? locus by STAT5. J Immunol

2001;167:320–326. 181.

Ye S-K, Agata Y, Lee HC, et al. The IL-7 receptor controls the accessibility of the TCR? locus by Stat5 and histone

acetylation. Immunity 2001;15:813–823. 182.

Huang J, Durum SK, Muegge K. Cutting edge: histone acetylation and recombination at the TCR gamma locus follows IL-7

induction. J Immunol 2001;167:6073–6077. 183.

Barndt RJ, Dai M, Zhuang Y. Functions of E2A-HEB heterodimers in T-cell development revealed by a dominant negative

mutation of HEB. Mol Cell Biol 2000;20:6677–6685. 184.

Washburn T, Schweighoffer E, Gridley T, et al. Notch activity influences the aß versus ?d T cell lineage decision. Cell

1997;88:833–843. 184a.

Wolfer A, Wilson A, Nemir M, MacDonald HR, Radtke F. Inactivation of Notch1 impairs VDJß rearrangement and allows

pre-TCR-independent survival of early aß lineage thymocytes. Immunity 2002;16:869–79. 185.

Deftos ML, Huang E, Ojala EW, et al. Notch1 signaling promotes the maturation of CD4 and CD8 SP thymocytes. Immunity

2000;13:73–84. 186.

Reizis B, Leder P. Direct induction of T lymphocyte–specific gene expression by the mammalian Notch signaling pathway.

Genes Dev 2002;16:295–300. 187.

Wilson A, Maréchal C, MacDonald HR. Biased Vß usage in immature thymocytes is independent of DJß proximity and pTa

pairing. J Immunol 2001;166:51–57. 188.

Bigas A, Martin DIK, Milner LA. Notch1 and Notch2 inhibit myeloid differentiation in response to different cytokines. Mol Cell

Biol 1998;18:2324–2333. 189.

Li L, Milner LA, Deng Y, et al. The human hom*olog of rat Jagged1 expressed by bone marrow stroma inhibits differentiation of

32D cells through interaction with Notch1. Immunity 1998;8:43–55. 190.

Jones P, May G, Healy L, et al. Stromal expression of Jagged 1 promotes colony formation by fetal hematopoietic progenitor

cells. Blood 1998;92:1505–1511. 191.

Ohishi K, Varnum-Finney B, Serda RE, et al. The Notch ligand, Delta-1, inhibits the differentiation of monocytes into

macrophages but permits their differentiation into dendritic cells. Blood 2001;98:1402–1407. 192.

Reisner Y, Linker-Israeli M, Sharon N. Separation of mouse thymocytes into two subpopulations by the use of peanut

agglutinin. Cell Immunol 1976;25:129–134. 193.

Wiest DL, Yuan L, Jefferson J, et al. Regulation of T cell receptor expression in immature CD4 +CD8 + thymocytes by p56 Lck

tyrosine kinase: basis for differential signaling by CD4 and CD8 in immature thymocytes expressing both coreceptor molecules. J Exp Med 1993;178:1701–1712. 194.

Ashwell JD, King LB, Vacchio MS. Cross-talk between the T cell antigen receptor and the glucocorticoid receptor regulates

thymocyte development. Stem Cells 1996;14:490–500. 195.

Vacchio MS, Lee JY, Ashwell JD. Thymus-derived glucocorticoids set the thresholds for thymocyte selection by inhibiting

TCR-mediated thymocyte activation. J Immunol 1999;163:1327–1333. 196.

Purton JF, Boyd RL, Cole TJ, et al. Intrathymic T cell development and selection proceeds normally in the absence of

glucocorticoid receptor signaling. Immunity 2000;13:179–186. 197.

Cole TJ, Myles K, Purton JF, et al. GRKO mice express an aberrant dexamethasone-binding glucocorticoid receptor, but are

profoundly glucocorticoid resistant. Mol Cell Endocrinol 2001;173:193–202. 198.

Yannoutsos N, Wilson P, Yu W, et al. The role of recombination activating gene ( RAG) reinduction in thymocyte development

in vivo. J Exp Med 2001;194:471–480. 199.

Wilson A, Held W, MacDonald HR. Two waves of recombinase gene expression in developing thymocytes. J Exp Med

1994;179:1355–1360. 200.

Thompson SD, Manzo AR, Pelkonen J, et al. Developmental T cell receptor gene rearrangements: relatedness of the a/ß and

the ?/d T cell precursors. Eur J Immunol 1991;21:1939–1950. 201.

Petrie HT, Livak F, Schatz DG, et al. Multiple rearrangements in T cell receptor a chain genes maximize the production of

useful thymocytes. J Exp Med 1993;178:615–622. 202.

Petrie HT, Livak F, Burtrum D, et al. T cell receptor gene recombination patterns and mechanisms: cell death, rescue, and T

cell production. J Exp Med 1995;182:121–127. 203.

Page DM. Cutting edge: thymic selection and autoreactivity are regulated by multiple coreceptors involved in T cell activation.

J Immunol 1999;163:3577–3581. 204.

Kishimoto H, Sprent J. Several different cell surface molecules control negative selection of medullary thymocytes. J Exp Med

1999;190:65–73. 205.

Lucas B, Vasseur F, Penit C. Production, selection, and maturation of thymocytes with high surface density of TCR. J

Immunol 1994;153:53–62. 206.

Guidos CJ, Weissman IL, Adkins B. Intrathymic maturation of murine T lymphocytes from CD8 + precursors. Proc Natl Acad

Sci U S A 1989;86:7542–7546. 207.

Lundberg K, Shortman K. Small cortical thymocytes are subject to positive selection. J Exp Med 1994;179:1475–1483.


Huesmann M, Scott B, Kisielow P, et al. Kinetics and efficacy of positive selection in the thymus of normal and T cell receptor

transgenic mice. Cell 1991;66:533–540. 209.

Itano A, Robey E. Highly efficient selection of CD4 and CD8 lineage thymocytes supports an instructive model of lineage

commitment. Immunity 2000;12:383–389. 210.

Kishimoto H, Sprent J. Negative selection in the thymus includes semimature T cells. J Exp Med 1997;185:263–271.


MacDonald HR, Lees RK. Programmed death of autoreactive thymocytes. Nature 1990;343:642–644.


Sebzda E, Mariathasan S, Ohteki T, et al. Selection of the T cell repertoire. Annu Rev Immunol 1999;17:829–874.


Sprent J, Kishimoto H. The thymus and central tolerance. Philos Trans R Soc Lond B Biol Sci 2001;356:609–616.


Hogquist KA. Signal strength in thymic selection and lineage commitment. Curr Opin Immunol 2001;13:225–231.


Ignatowicz L, Kappler J, Marrack P. The repertoire of T cells shaped by a single MHC/peptide ligand. Cell 1996;84:521–529.


Rivera RR, Johns CP, Quan J, et al. Thymocyte selection is regulated by the helix-loop-helix inhibitor protein, Id3. Immunity

2000;12:17–26. 217.

Bonifacino JS, McCarthy SA, Maguire JE, et al. Novel post-translational regulation of TCR expression in CD4 +CD8 +

thymocytes influenced by CD4. Nature 1990;344:247–251. 218.

Kawai K, Ohashi PS. Immunological function of a defined T-cell population tolerized to low-affinity self antigens. Nature

1995;374:68–69. 219.

Sebzda E, Kundig TM, Thomson CT, et al. Mature T cell reactivity altered by peptide agonist that induces positive selection. J

Exp Med 1996;183:1093–1104.


Alberola-Ila J, Hogquist KA, Swan KA, et al. Positive and negative selection invoke distinct signaling pathways. J Exp Med

1996;184:9–18. 221.

Backstrom BT, Muller U, Hausmann B, et al. Positive selection through a motif in the alphabeta T cell receptor. Science

1998;281:835–838. 222.

Werlen G, Hausmann B, Palmer E. A motif in the aß T-cell receptor controls positive selection by modulating ERK activity.

Nature 2000;406:422–426. 223.

Mak TW, Penninger JM, Ohashi PS. Knockout mice: a paradigm shift in modern immunology. Nature Rev Immunol

2001;1:11–19. 224.

Li R, Page DM. Requirement for a complex array of costimulators in the negative selection of autoreactive thymocytes in vivo.

J Immunol 2001;166:6050–6056. 225.

Wang J, Chun T, Lo JC, et al. The critical role of LIGHT, a TNF family member, in T cell development. J Immunol

2001;167:5099–5105. 226.

Wang ECY, Thern A, Denzel A, et al. DR3 regulates negative selection during thymocyte development. Mol Cell Biol

2001;21:3451–3461. 227.

Gomez M, Kioussis D, Cantrell DA. The GTPase Rac-1 controls cell fate in the thymus by diverting thymocytes from positive

to negative selection. Immunity 2001;15:703–713. 228.

Page DM, Roberts EM, Peschon JJ, et al. TNF receptor-deficient mice reveal striking differences between several models of

thymocyte negative selection. J Immunol 1998;160:120–133. 229.

Bassiri H, Carding SR. A requirement for IL-2/IL-2 receptor signaling in intrathymic negative selection. J Immunol

2001;166:5945–5954. 230.

Laufer TM, Glimcher LH, Lo D. Using thymus anatomy to dissect T cell repertoire selection. Semin Immunol 1999;11:65–70.


Laufer TM, DeKoning J, Markowitz JS, et al. Unopposed positive selection and autoreactivity in mice expressing class II MHC

only on thymic cortex. Nature 1996;383:81–85. 232.

Anderson G, Partington KM, Jenkinson EJ. Differential effects of peptide diversity and stromal cell type in positive and

negative selection in the thymus. J Immunol 1998;161:6599–6603. 233.

McGargill MA, Derbinski JM, Hogquist KA. Receptor editing in developing T cells. Nat Immunol 2000;1:336–341.


Basson MA, Zamoyska R. The CD4/CD8 lineage decision: integration of signalling pathways. Immunol Today

2000;21:509–514. 235.

Singer A. New perspectives on a developmental dilemma: the kinetic signaling model and the importance of signal duration

for the CD4/CD8 lineage decision. Curr Opin Immunol 2002;14:207–215. 236.

Benoist C, Mathis D. T-lymphocyte differentiation and biology. In: Paul WE, ed. Fundamental immunology, 4th ed.

Philadelphia: Lippincott-Raven Publishers, 1999:367–409. 237.

Lucas B, Germain RN. Unexpectedly complex regulation of CD4/CD8 coreceptor expression supports a revised model for

CD4 +CD8 + thymocyte differentiation. Immunity 1996;5:461–477. 238.

Lundberg K, Heath W, Köntgen F, et al. Intermediate steps in positive selection: differentiation of CD4 +8 int TCR int

thymocytes into CD4 -8 + TCR hi thymocytes. J Exp Med 1995;181:1643–1651. 239.

Lucas B, Vasseur F, Penit C. Stochastic coreceptor shut-off is restricted to the CD4 lineage maturation pathway. J Exp Med

1995;181:1623–1633. 240.

Kydd R, Lundberg K, Vremec D, et al. Intermediate steps in thymic positive selection. Generation of CD4 -8 + T cells in

culture from CD4 +8 +, CD4 int8 +, and CD4 +8 int thymocytes with up-regulated levels of TCR-CD3. J Immunol 1995;155:3806–3814. 241.

Suzuki H, Punt JA, Granger LG, et al. Asymmetric signaling requirements for thymocyte commitment to the CD4 + versus

CD8 + T cell lineages: a new perspective on thymic commitment and selection. Immunity 1995;2:413–425. 242.

Correia-Neves M, Mathis D, Benoist C. A molecular chart of thymocyte positive selection. Eur J Immunol 2001;31:2583–2592.


Chan SH, Cosgrove D, Waltzinger C, et al. Another view of the selective model of thymocyte selection. Cell

1993;73:225–236. 244.

Davis CB, Killeen N, Casey Crooks ME, et al. Evidence for a stochastic mechanism in the differentiation of mature subsets of

T lymphocytes. Cell 1993;73:237–247. 245.

Leung RK, Thomson K, Gallimore A, et al. Deletion of the CD4 silencer element supports a stochastic mechanism of

thymocyte lineage commitment. Nat Immunol 2001;2:1167–1173. 246.

Itano A, Salmon P, Kioussis D, et al. The cytoplasmic domain of CD4 promotes the development of CD4 lineage T cells. J

Exp Med 1996;183:731–741. 247.

Seong RH, Chamberlain JW, Parnes JR. Signal for T-cell differentiation to a CD4 cell lineage is delivered by CD4

transmembrane region and/or cytoplasmic tail. Nature 1992;356:718–20. 248.

Irie HY, Mong MS, Itano A, et al. The cytoplasmic domain of CD8 beta regulates Lck kinase activation and CD8 T cell

development. J Immunol 1998;161:183–191. 249.

Itano A, Cado D, Chan FK, et al. A role for the cytoplasmic tail of the ßchain of CD8 in thymic selection. Immunity

1994;1:287–290. 250.

Campbell KS, Buder A, Deuschle U. Interactions between the amino-terminal domain of p56 lck and cytoplasmic domains of

CD4 and CD8a in yeast. Eur J Immunol 1995;25:2408–2412. 251.

Hernandez-Hoyos G, Sohn SJ, Rothenberg EV, et al. Lck activity controls CD4/CD8 T cell lineage commitment. Immunity

2000;12:313–322. 252.

Legname G, Seddon B, Lovatt M, et al. Inducible expression of a p56 Lck transgene reveals a central role for Lck in the

differentiation of CD4 SP thymocytes. Immunity 2000;12:537–546. 253.

Sharp LL, Hedrick SM. Commitment to the CD4 lineage mediated by extracellular signal-related kinase mitogen-activated

protein kinase and lck signaling. J Immunol 1999;163:6598–6605. 254.

Salmon P, Mong M, Kang X-J, et al. The role of CD8 a' in the CD4 versus CD8 lineage choice. J Immunol

1999;163:5312–5318. 255.

Yasutomo K, Doyle C, Miele L, et al. The duration of antigen receptor signalling determines CD4 + versus CD8 + T-cell

lineage fate. Nature 2000;404:506–510. 256.

Iwata M, Kuwata T, Mukai M, et al. Differential induction of helper and killer T cells from isolated CD4 +CD8 + thymocytes in

suspension culture. Eur J Immunol 1996;26:2081–2086. 257.

Wolfer A, Bakker T, Wilson A, et al. Inactivation of Notch 1 in immature thymocytes does not perturb CD4 or CD8 T cell

development. Nat Immunol 2001;2:235–241. 258.

Brugnera E, Bhandoola A, Cibotti R, et al. Coreceptor reversal in the thymus: signaled CD4 +8 + thymocytes initially terminate

CD8 transcription even when differentiating into CD8 + T cells. Immunity 2000;13:59–71. 259.

Bosselut R, Feigenbaum L, Sharrow SO, et al. Strength of signaling by CD4 and CD8 coreceptor tails determines the number

but not the lineage direction of positively selected thymocytes. Immunity 2001;14:483–494. 260.

Bommhardt U, Basson MA, Krummrei U, et al. Activation of the extracellular signal-related kinase/mitogen-activated protein

kinase pathway discriminates CD4 versus CD8 lineage commitment in the thymus. J Immunol 1999;163:715–722. 261.

Shao H, Wilkinson B, Lee B, et al. Slow accumulation of active mitogen–activated protein kinase during thymocyte

differentiation regulates the temporal pattern of transcription factor gene expression. J Immunol 1999;163:603–610. 262.

Mariathasan S, Ho SS, Zakarian A, Ohashi PS. Degree of ERK activation influences both positive and negative thymocyte

selection. Eur J Immunol 2000;30:1060–1068. 263.

Alberola-Ila J, Forbush KA, Seger R, et al. Selective requirement for MAP kinase activation in thymocyte differentiation.

Nature 1995;373:620–623. 264.

Pages G, Guerin S, Grall D, et al. Defective thymocyte maturation in p44 MAP kinase (Erk 1) knockout mice. Science

1999;286:1374–1377. 265.

Schmeissner PJ, Xie H, Smilenov LB, et al. Integrin functions play a key role in the differentiation of thymocytes in vivo. J

Immunol 2001;167:3715–3724. 266.

Wilkinson B, Chen JY, Han P, et al. TOX: an HMG box protein implicated in the regulation of thymocyte selection. Nat

Immunol 2002;3:272–280. 267.

Izon DJ, Punt JA, Xu L, et al. Notch1 regulates maturation of CD4 + and CD8 + thymocytes by modulating TCR signal

strength. Immunity 2001;14:253–264. 268.

Keefe R, Dave V, Allman D, et al. Regulation of lineage commitment distinct from positive selection. Science

1999;286:1149–1153. 269.

Akashi K, Kondo M, von Freeden-Jeffry U, et al. Bcl-2 rescues T lymphopoiesis in interleukin-7 receptor–deficient mice. Cell

1997;89:1033–1041. 270.

Petrie HT, Strasser A, Harris AW, et al. CD4 +8 - and CD4 -8 + mature thymocytes require different post-selection processing

for final development. J Immunol 1993;151:1273–1279. 271.

Robey E, Chang D, Itano A, et al. An activated form of Notch influences the choice between CD4 and CD8 T cell lineages.

Cell 1996;87:483–492. 272.

Nawijn MC, Ferreira R, Dingjan GM, et al. Enforced expression of GATA-3 during T cell development inhibits maturation of

CD8 single-positive cells and induces thymic lymphoma in transgenic mice. J Immunol 2001;167:715–723. 273.

Hayashi K, Abe N, Watanabe T, et al. Overexpression of AML1 transcription factor drives thymocytes into the CD8

single-positive lineage. J Immunol 2001;167:4957–4965. 274.

Telfer JC, Laurent MN, Rothenberg EV. Runx1 expression influences both the aß/?d T cell and CD4/CD8 T cell lineage

decisions. Submitted 2002. 275.

Vandekerckhove BAE, Barcena A, Schols D, et al. In vivo cytokine expression in the thymus. CD3 high human thymocytes are

activated and already functionally differentiated into helper and cytotoxic cells. J Immunol 1994;152:1738–1743. 276.

Vanhecke D, Verhasselt B, Debacker V, et al. Differentiation to T-helper cells in the thymus—gradual acquisition of T-helper

cell–function by CD3 +CD4 + cells. J Immunol 1995;155:4711–4718. 277.

Wang H, Diamond RA, Yang-Snyder JA, et al. Precocious expression of T-cell functional response genes in vivo in primitive

thymocytes before T-lineage commitment. Int Immunol 1998;10:1623–1635. 278.

Corbella P, Moskophidis D, Spanopoulou E, et al. Functional commitment to helper T cell lineage precedes positive selection

and is independent of T cell receptor MHC specificity. Immunity 1994;1:269–276. 279.

Matechak EO, Killeen N, Hedrick SM, et al. MHC class II–specific T cells can develop in the CD8 lineage when CD4 is

absent. Immunity 1996;4:337–347. 280.

Bhandoola A, Kithiganahalli B, Granger L, et al. Programming for cytotoxic effector function occurs concomitantly with CD4

extinction during CD8 + T cell differentiation in the thymus. Int Immunol 2000;12:1035–1040. 281.

Zou YR, Sunshine MJ, Taniuchi I, et al. Epigenetic silencing of CD4 in T cells committed to the cytotoxic lineage. Nat Genet

2001;29:332–336. 282.

Tough DF, Sprent J. Thymic emigration—a reply. Immunol Today 1995;16:273–274.


Scollay R, Godfrey DI. Thymic emigration: conveyor belts or lucky dips? Immunol Today 1995;16:268–273.


Ueno T, Hara K, Willis MS, et al. Role for CCR7 ligands in the emigration of newly generated T lymphocytes from the

neonatal thymus. Immunity 2002;16:205–218. 285.

Chaffin KE, Perlmutter RM. A pertussis toxin–sensitive process controls thymocyte emigration. Eur J Immunol

1991;21:2565–2573. 286.

Akashi K, Reya T, Dalma-Weiszhausz D, et al. Lymphoid precursors. Curr Opin Immunol 2000;12:144–150.


Bendelac A, Rivera MN, Park SH, et al. Mouse CD1-specific NK1 T cells: development, specificity, and function. Annu Rev

Immunol 1997;15:535–562. 288.

Godfrey DI, Hammond KJ, Poulton LD, et al. NKT cells: facts, functions and fallacies. Immunol Today 2000;21:573–583.


Wilson SB, Byrne MC. Gene expression in NKT cells: defining a functionally distinct CD1d-restricted T cell subset. Curr Opin

Immunol 2001;13:555–561. 290.

Benlagha K, Weiss A, Beavis A, et al. In vivo identification of glycolipid antigen–specific T cells using fluorescent CD1d

tetramers. J Exp Med 2000;191:1895–1903. 291.

Matsuda JL, Naidenko OV, Gapin L, et al. Tracking the response of natural killer T cells to a glycolipid antigen using CD1d

tetramers. J Exp Med 2000;192:741–754. 292.

Gapin L, Matsuda JL, Surh CD, et al. NKT cells derive from double-positive thymocytes that are positively selected by CD1d.

Nat Immunol 2001;2:971–978. 293.

MacDonald HR. NK1.1 + T cell receptor-a/ß + cells: new clues to their origin, specificity, and function. J Exp Med

1995;182:633–638. 294.

Hammond K, Cain W, van Driel I, et al. Three day neonatal thymectomy selectively depletes NK1.1 + T cells. Int Immunol

1998;10:1491–1499. 295.

Coles MC, Raulet DH. NK1.1 + T cells in the liver arise in the thymus and are selected by interactions with class I molecules

on CD4 +CD8 + cells. J Immunol 2000;164:2412–2418. 296.

Bendelac A, Killeen N, Littman DR, et al. A subset of CD4 + thymocytes selected by MHC class I molecules. Science

1994;263:1774–1778. 297.

Eberl G, Lowin-Kropf B, MacDonald HR. Cutting edge: NKT cell development is selectively impaired in Fyn-deficient mice. J

Immunol 1999;163:4091–4094. 298.

Gadue P, Morton N, Stein PL. The Src family tyrosine kinase Fyn regulates natural killer T cell development. J Exp Med

1999;190:1189–1196. 299.

Walunas TL, Wang B, Wang CR, et al. Cutting edge: the Ets1 transcription factor is required for the development of NK T

cells in mice. J Immunol 2000;164:2857–2860. 300.

Ohteki T, Ho S, Suzuki H, et al. Role for IL-15/IL-15 receptor ß-chain in natural killer 1.1 + T cell receptor-aß + cell

development. J Immunol 1997;159:5931–5935. 301.

Elewaut D, Brossay L, Santee SM, et al. Membrane lymphotoxin is required for the development of different subpopulations

of NK T cells. J Immunol 2000;165:671–679. 302.

Eberl G, Fehling HJ, von Boehmer H, et al. Absolute requirement for the pre–T cell receptor a chain during NK1.1 + TCRaß

cell development. Eur J Immunol 1999;29:1966–1971. 303.

MacDonald HR. Development and selection of NKT cells. Curr Opin Immunol 2002;14:250–254.


Sato H, Nakayama T, Tanaka Y, et al. Induction of differentiation of pre-NKT cells to mature Va14 NKT cells by

granulocyte/macrophage colony–stimulating factor. Proc Natl Acad Sci U S A 1999;96:7439–7444. 305.

Iwabuchi K, Iwabuchi C, Tone S, et al. Defective development of NK1.1 + T-cell antigen receptor aß + cells in zeta-associated

protein 70 null mice with an accumulation of NK1.1 + CD3 - NK-like cells in the thymus. Blood 2001;97:1765–1775. 306.

Baur N, Nerz G, Nil A, et al. Expression and selection of productively rearranged TCRß VDJ genes are sequentially regulated

by CD3 signaling in the development of NK1.1 + aß T cells. Int Immunol 2001;13:1031–1042. 307.

Shimamura M, Ohteki T, Beutner U, et al. Lack of directed Va14-Ja281 rearrangements in NK1 + T cells. Eur J Immunol

1997;27:1576–1579. 308.

Sakaguchi S, Sakaguchi N, Asano M, et al. Immunologic self-tolerance maintained by activated T cells expressing IL-2

receptor alpha-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases. J Immunol 1995;155:1151–1164. 309.

Shevach EM. Regulatory T cells in autoimmmunity. Annu Rev Immunol 2000;18:423–449.


Itoh M, Takahashi T, Sakaguchi N, et al. Thymus and autoimmunity: production of CD25 +CD4 + naturally anergic and

suppressive T cells as a key function of the thymus in maintaining immunologic self-tolerance. J Immunol 1999;162:5317–5326. 311.

Asano M, Toda M, Sakaguchi N, et al. Autoimmune disease as a consequence of developmental abnormality of a T cell

subpopulation. J Exp Med 1996;184:387–396. 312.

Papiernik M, de Moraes ML, Pontoux C, et al. Regulatory CD4 T cells: expression of IL-2Ra chain, resistance to clonal

deletion and IL-2 dependency. Int Immunol 1998;10:371–378. 313.

Jordan MS, Boesteanu A, Reed AJ, et al. Thymic selection of CD4 +CD25 + regulatory T cells induced by an agonist

self-peptide. Nat Immunol 2001;2:301–306. 314.

Suto A, Nakajima H, Ikeda K, et al. CD4 +CD25 + T-cell development is regulated by at least 2 distinct mechanisms. Blood

2002;99:555–560. 315.

Ramsdell F, Lantz T, Fowlkes BJ. A nondeletional mechanism of thymic self tolerance. Science 1989;246:1038–1041.


Hori S, Nomura T, Sakaguchi S. Control of regulatory T cell development by the transcription factor Foxp3. Science

2003;299:1057–61. 315b.

Fontenot JD, Gavin MA, Rudensky AY. Foxp3 programs the development and function of CD4 +CD25 + regulatory T cells.

Nat Immunol 2003;4:330–6. 315c.

Khattri R, Cox T, Yasayko SA, Ramsdell F. An essential role for Scurfin in CD4 +CD25 + T regulatory cells. Nat Immunol

2003;4:337–42. 316.

Plum J, De Smedt M, Verhasselt B, et al. In vitro intrathymic differentiation kinetics of human fetal liver CD34 +CD38 -

progenitors reveals a phenotypically defined dendritic/T-NK precursor split. J Immunol 1999;162:60–68. 317.

Blom B, Verschuren MC, Heemskerk MH, et al. TCR gene rearrangements and expression of the pre–T cell receptor complex

during human T-cell differentiation. Blood 1999;93:3033–3043. 318.

van Ewijk W, Wang B, Hollander G, et al. Thymic microenvironments, 3-D versus 2-D? Semin Immunol 1999;11:57–64.


Jiang R, Lan Y, Chapman HD, Shawber C, Norton CR, Serreze DV, Weinmaster G, Gridley T. Defects in limb, craniofacial,

and thymic development in Jagged2 mutant mice. Genes Dev 1998;12:1046–57. In a recombinase-deficient mutant thymus, cells become quiescent and arrest differentiation completely at the DN3 stage, but in adoptive transfer of recombinase-deficient cells to a wild-type thymus, some cells can pass through the DN4/ISP stages before dying in the DP state. The difference is likely to be the signals provided by the microenvironment. There is bidirectional communication between developing lymphocytes and the microenvironment at different stages ( 318 ), and the passage of a cohort of normal TCR + cells appears to cause the microenvironment to provide better mitogenic and signals to support “spontaneous” differentiation of DN3 cells. It is not known whether the level of checkpoint control that seems to be bypassed by stromal signals is the mechanism mediated by p53 or by TNF receptor/Fas + FADD. Here an exception may be the “first wave” TCR?d cells, since a large fraction of fetal ?d cell development depends on the Notch ligand Jagged-2 ( 319 ). The exception is one unconventional subset of CD4 SP cells that is actually selected by nonclassical class I MHC family molecules; this is the NK T-cell subset, discussed in the section on Frontiers for the Future.

Chapter 10 Peripheral T-Lymphocyte Responses and Function Fundamental Immunology

Chapter 10 Marc Jenkins

Peripheral T-Lymphocyte Responses and Function

INTRODUCTION NAÏVE T CELLS Generation Recirculation Survival IN VIVO PRESENTATION OF PEPTIDE–MHC LIGANDS Dendritic Cells Presentation of Low-Affinity Self-Peptide–MHC Ligands Presentation of High-Affinity Foreign Peptide–MHC Ligands T-CELL ACTIVATION Signal Transduction IL-2 Production and Proliferation Effector Cells Memory Cells SUMMARY ACKNOWLEDGMENTS REFERENCES

INTRODUCTION A thymus-derived (T) lymphocyte becomes activated when its antigen receptor (TCR) binds to a major histocompatibility complex (MHC)–encoded protein containing a specific peptide on the surface of an antigen-presenting cell (APC) ( 1 ). The ensuing signal transduction and functional changes that occur in T cells have been studied extensively in culture systems and will be discussed in great detail in other chapters. The goal of this chapter is to provide the reader with a “big picture” of how T-cell activation occurs in vivo and how this activation results in immune memory. Exposure of a normal host to virtually any foreign protein will activate a few naïve T cells that express TCRs with high affinity for peptide–MHC combinations produced from the antigen. Because of the vast diversity of the T-cell repertoire, the T cells expressing TCRs specific for a single foreign antigen are so rare that their activation cannot be detected after addition of a foreign antigen to an in vitro culture of blood or lymphoid tissue cells from a naïve host. On the other hand, in vitro culture with an antigen that the host has been exposed to in the past results in detectable signs of T-cell activation,

such as proliferation, lymphokine production, and cell-mediated cytotoxicity ( 2 ). This ex vivo approach has been the mainstay for studying the ability of foreign antigens to induce T-cell activation in vivo. However, in addition to being too insensitive to detect the activation of naïve T cells in the polyclonal repertoire, cell culture approaches cannot reproduce the complex microenvironments in which T cells are activated in vivo and are indirect measures of the products of activated T cells not direct measures of the T cells themselves. These limitations have been remedied by new systems that allow direct ex vivo or in situ monitoring of antigen-specific T cells. One method relies on fluorochrome-labeled, hom*ogenous, multimeric peptide–MHC I or II complexes ( 3 , 4 and 5 ). Peptide–MHC multimers bind to the T cells that express an appropriate TCR, allowing direct detection of antigen-specific T cells by flow cytometry or immunohistology. The strength of the peptide–MHC multimer approach is that it can theoretically measure all potentially responsive T cells in the normal repertoire. However, since the frequency of T cells specific for most peptide–MHC complexes in naïve individuals is below the limit of detection of flow cytometry ( 6 , 7 and 8 ), peptide–MHC multimers cannot currently be used to study the earliest events in T-cell activation that occur before proliferation. One way to solve the clonal infrequency problem is adoptive transfer of T cells from TCR transgenic mice into syngeneic normal recipients ( 9 ). This maneuver produces a traceable naïve T-cell population of known peptide–MHC specificity, comprising ˜0.2% of cells in the secondary lymphoid organs of the recipient. The transferred cells can be distinguished from those of the recipient with antibodies specific for the transgenic TCR clonotype or an allelic marker such as Thy 1 or CD45. The earliest events in T-cell activation in vivo can be studied with this method because the antigen-specific T cells are abundant enough to be detected by flow cytometry or immunohistology before proliferation. A potential disadvantage is that even though only a small number of T cells are transferred, the resulting frequency of antigen-specific T cells is still higher than normal. This can create a situation in which the transferred T cells outcompete endogenous T cells of similar specificity under certain conditions ( 10 , 11 and 12 ). However, the kinetics and relative magnitude of proliferation and loss reported for transferred T cells after in vivo exposure to antigen are identical to those described for endogenous T cells tracked with peptide–MHC multimers ( 9 ). This chapter rests heavily on studies involving peptide–MHC multimers or TCR-transgenic T-cell adoptive transfer methods because they provide the most physiologically relevant information on the in vivo immune response.

NAÏVE T CELLS Generation As aß T cells develop in the thymus, TCR a and ß gene segments are rearranged such that each clone eventually expresses a unique TCR ( 13 ). Developing thymocytes that produce a surface TCR express CD4 and CD8 co-receptors and undergo one of three fates, depending on the specificity and affinity of their TCRs for self-peptide–MHC ligands. Thymocytes that express TCRs with no affinity for self-peptide–MHC molecules

die by a programmed cell death mechanism. Potentially harmful thymocytes that express TCRs with strong affinity for the self-peptide–MHC ligands expressed on cells in the thymus are eliminated via physical deletion ( 14 ), functional inactivation ( 15 ), or receptor editing ( 16 ). Only thymocytes that express TCRs with a low but significant affinity for self-peptide–MHC ligands on thymic stromal cells differentiate into mature T cells ( 17 ). Thymocytes expressing TCRs with a low affinity for self-peptide-class I MHC molecules (MHC I) become mature T cells that express CD8 but not CD4, whereas thymocytes expressing TCRs with a low affinity for self-peptide–class II MHC molecules (MHC II) mature into cells that express CD4 but not CD8. This process is termed “positive selection.” Recirculation CD4 and CD8 T cells that survive positive selection leave the thymus and enter the secondary lymphoid organs. T cells that have not yet encountered a foreign peptide–MHC ligand for which their TCR has a high affinity are referred to as “naïve” T cells. These cells account for the majority of T cells in the secondary lymphoid organs in healthy young adults. Naïve T cells recirculate continuously through the secondary lymphoid organs, which include the spleen, lymph nodes, and mucosal lymphoid organs (e.g., the Peyer’s patches of the intestines) ( 18 , 19 ). Secondary lymphoid organs are defined based on the presence of segregated T- and B-cell–rich regions and specialized blood vessels that facilitate the entry of naïve lymphocytes ( 20 ). Naïve T cells are found only in the T-cell–rich areas known as the paracortex in the lymph nodes and mucosal lymphoid organs, and the periarteriolar lymphoid sheath (PALS) in the spleen ( 21 , 22 ). The predilection for lymph nodes is explained by the fact that naïve T cells express a unique set of receptors that bind ligands expressed on the specialized blood vessels of the lymph nodes and mucosal lymphoid organs known as high endothelial venules (HEV) ( 23 ). Naïve T cells use CD62L or a4ß7 integrin, CC chemokine receptor (CCR) 7, and LFA-1 (CD11a/CD18) for rolling, adhesion, and extravasation through the HEV in peripheral lymph nodes and mucosal lymphoid organs. HEV are the only blood vessels in the body that display all of the ligands for these receptors ( 23 ). Naïve T cells move from the blood into the spleen because blood is emptied from terminal branches of the central arteriole into marginal sinuses or directly into the red pulp ( 20 ). The T cells then move into the PALS by a poorly understood CD62L-independent, G protein–dependent mechanism ( 24 ). Naïve T cells are retained in the T-cell areas of the spleen for about 5 hours and the lymph nodes for about 1 day ( 18 ), in part via CCR7 signaling in response to CCL21 (SLC) and CCL19 (ELC) chemokines produced by stromal cells in the T-cell areas ( 20 ). Naïve T cells leave the lymph nodes via efferent lymphatic vessels that eventually merge into the blood stream via the thoracic duct, or the spleen via the splenic vein. Once in the blood, a naïve T-cell will quickly enter a new secondary lymphoid organ, repeating the processes described above. An exception to this scenario occurs for a short period after birth when naïve T cells recirculate through nonlymphoid tissues ( 25 ). This behavior may exist to induce tolerance in those naïve T cells that express TCRs specific for self-peptide–MHC combinations displayed only outside the thymus.

Survival It is estimated that an individual naïve T-cell will on average circulate through the secondary lymphoid organs for several months ( 26 , 27 ). To have this normal lifespan, naïve CD4 T cells must be exposed to MHC II molecules and CD8 T cells to MHC I molecules ( 28 , 29 ). Thus, it is likely that T-cell survival is maintained by low-affinity TCR recognition of self-peptide–MHC complexes. This recognition results in a subset of the signals that emanate from the TCR when bound by a high-affinity ligand, including partial phosphorylation of the CD3-zeta chain ( 30 ). IL-7 is also required to maintain the survival of naïve T cells, as evidenced by the findings that IL-7 receptor–deficient naïve T cells have a short life span ( 31 ) and normal T cells survive poorly in IL-7–deficient recipients ( 32 ). Although signals through the TCR and IL-7 receptor are required for the survival of naïve T cells, these signals do not cause the T cells to proliferate in hosts containing normal numbers of T cells. In contrast, naïve T cells proliferate when transferred into T-cell–deficient hosts ( 33 , 34 , 35 and 36 ). This “homeostatic” proliferation also depends on IL-7 ( 31 , 32 ) and low-affinity TCR recognition of self-peptide–MHC complexes—probably the self-peptide–MHC complex that caused the T-cell to undergo positive selection in the thymus ( 34 , 35 ), but not IL-2 or the CD28 co-stimulatory receptor ( 37 ). Thus, the same signaling events that cause naïve T cells to survive in interphase in T-cell–sufficient hosts, cause these cells to proliferate in T-cell–deficient hosts, but using a program different from that engaged during the T-cell response to high-affinity TCR ligands. Survival and proliferation could both contribute to control of the number of naïve T cells in normal hosts. In young individuals, new naïve T cells are constantly produced by the thymus and exported to the secondary lymphoid organs to replace senescent naïve T cells. Since the secondary lymphoid organs would constantly be full under these conditions, the resident naïve T cells would survive in interphase. In contrast, in older individuals in whom thymic output is reduced or absent, and are thus susceptible to lymphopenia, senescent cells may be replaced by proliferation of remaining naïve T cells.

IN VIVO PRESENTATION OF PEPTIDE–MHC LIGANDS As naïve T cells percolate through the T-cell areas of secondary lymphoid organs, they encounter a dense network of large, irregularly shaped dendritic cells that constitutively express the highest levels of MHC molecules of any cell in the body ( 38 ). Given their location in the T-cell areas, surface molecule repertoire, and potent capacity to stimulate naïve T cells in vitro ( 38 ), it is likely that dendritic cells play an important role in the presentation of low-affinity self-peptide–MHC ligands that maintains naïve T cells, as well as the initial presentation of high-affinity foreign peptide–MHC ligands that stimulates the proliferation and differentiation of naïve T cells. Dendritic Cells

Dendritic cells exist in several subsets ( 39 , 40 ). In the mouse, the CD11c integrin appears to mark most large, MHC high cells, although CD11c is also expressed on monocytes and a subset of antigen-experienced CD8 T cells ( 41 ). Three types of CD11c +, MHC + dendritic cells are found in the spleen and lymph nodes ( Fig. 1). One type expresses the myeloid marker CD11b and variably expresses CD4, but does not express CD8a or the CD205 integrin, and is often referred to as the “myeloid dendritic cell.” These dendritic cells are found mainly in the red pulp or marginal zones of the spleen and the outer edges of the paracortex in the lymph nodes. A second type lacks CD11b and CD4 but expresses CD8a and CD205, and is often referred to as the “lymphoid dendritic cell.” Lymphoid dendritic cells are located primarily in the PALS of the spleen and the central paracortex of the lymph nodes. Although the names imply derivation from separate lineages, recent work indicates that both myeloid and lymphoid dendritic cells can be derived from either common myeloid or lymphoid precursors ( 42 , 43 ). The common origin of these dendritic cells is supported by the recent finding that highly purified CD11b +, CD205 -, CD8a - myeloid dendritic cells give rise to CD11b -, CD205 +, CD8a + lymphoid dendritic cells after adoptive transfer ( 44 ). Myeloid and lymphoid dendritic cells survive for about 10 days in the secondary lymphoid organs ( 45 , 46 ).

FIG. 1. A: Dendritic cell subsets of the spleen. B: Dendritic cell subsets of the lymph nodes. Molecules expressed by each dendritic subtype are shown. For certain molecules, the level of expression is indicated as high ( + +), intermediate ( +), or low ( +/ -). Arrows indicate cellular movements. The dashed arrow indicates that the migration of monocytes and subsequent conversion to CD11c - dendritic cells probably only operates in the presence of infection or tissue damage.

The third type of CD11c + dendritic cell in the spleen and lymph nodes lacks CD11b but expresses the B220 and Gr-1 molecules normally expressed by B cells and granulocytes, respectively ( 47 ). These cells have plasmacytoid morphology and are concentrated near the HEV ( 47 ) in the lymph nodes. The fact that the number of these

dendritic cells is greatly reduced in the lymph nodes of CD62L-deficient mice ( 47 ) indicates that they enter the lymph nodes from the blood through the HEV. Plasmacytoid dendritic cells are potent producers of IFN-a, which plays a role in the generation of IFN-?–producing memory T cells in humans ( 48 ). The lymph nodes contain several additional dendritic populations ( 45 , 49 , 50 ) ( Fig. 1B). All lymph nodes contain CD11c + cells that express CD11b and lack CD8a and are thus similar to myeloid dendritic cells, but also express CD205, albeit at a lower level than lymphoid dendritic cells. The absence of these cells in the spleen or Peyer’s patches, which lack afferent lymphatic vessels, implies that these cells are migrants that move into the lymph node from interstitial tissue via afferent lymphatic vessels. These cells will be referred to as “interstitial dendritic cell migrants.” The superficial lymph nodes that drain the skin contain another population of dendritic cells that expresses high levels of CD205 and CD11b and intermediate levels of CD8a. These cells also express langerin ( 50 ), a protein that is expressed primarily by Langerhans cells ( 51 ). Therefore, it is very likely that CD11c +, CD205 +, CD11b +, CD8a intermediate cells in the superficial lymph nodes are Langerhans cells that recently migrated from the skin. These Langerhans cell migrants are more long-lived than myeloid and lymphoid dendritic cells, as they survive for months in the secondary lymphoid organs ( 45 ). The migration of dendritic cells from nonlymphoid organs to the lymph nodes, or red pulp of the spleen into the PALS, is driven by inflammatory stimuli such as LPS ( 52 ) or inflammatory cytokines including IL-1 and TNF-a ( 53 , 54 ). Dendritic cell migration from nonlymphoid organs into the secondary lymphoid organs is associated with a maturation process that results in changes in antigen processing and presentation potential ( 55 ). This maturation process can be mimicked by culturing immature nonlymphoid tissue–derived cells or their precursors in the presence of inflammatory cytokines ( 55 ). Immature dendritic cells efficiently engulf particles including apoptotic cells and extracellular fluid, and produce many peptide–MHC complexes from the ingested proteins, especially in the presence of inflammatory mediators such as TNF-a ( 56 ). Immature dendritic cells also display MHC II molecules that turn over rapidly ( 57 ). In contrast, mature dendritic cells that have been exposed to inflammatory cytokines for several days are inefficient at antigen uptake and processing and display stable MHC II molecules ( 57 ). Together, these observations have led to the idea that in vivo antigen presentation to CD4 T cells depends heavily on immature dendritic cells that acquire antigens in nonlymphoid tissues, migrate to the secondary lymphoid organs, and in the process produce stable peptide–MHC II complexes. Inflammatory stimuli enhance this process, probably by increasing TNF-a and IL-1, and cause the migrated dendritic cells to increase expression of molecules in involved in co-stimulation (CD80, CD86, and CD40) ( 52 ). Presentation of Low-Affinity Self-Peptide–MHC Ligands

As mentioned above, naïve T cells must recognize low-affinity self-peptide–MHC ligands to have a normal life span. Dendritic cells probably play a role in this process because they are constantly in contact with T cells in the secondary lymphoid organs ( 58 59 , ), and expression of MHC II molecules under the control of the dendritic cell-specific CD11c promoter is sufficient to maintain the survival of naïve CD4 T cells ( 58 ). In the absence of infection or tissue damage, all dendritic cell populations in the secondary lymphoid organs exist in a resting state characterized by low expression of co-stimulatory molecules such as CD80 and CD86 ( 52 ). Recent discoveries indicate that this resting state is actively maintained by cytokines that are produced by phagocytes as they engulf senescent apoptotic cells. Engulfment of apoptotic cells is mediated by receptors that recognize unique molecules on apoptotic cells ( 60 ), for example, the phosphatidyl serine receptor ( 60 , 61 ), which trigger the production of TGF-ß1, IL-10, and prostaglandins ( 62 ). These molecules are known to inhibit dendritic cell maturation and expression of co-stimulatory molecules ( 63 ). Together these results suggest a scenario, in which the dendritic cells that present self-peptide–MHC molecules to naïve T cells and maintain their survival are in a suppressed state brought about by anti-inflammatory cytokines produced by themselves or other phagocytes as they engulf senescent apoptotic cells, either in the lymphoid organs or in the nonlymphoid organs before migration. It is not clear which of the dendritic cell types found in the secondary lymphoid organs play this role. Recent reports of autoimmune disease and immunopathology in mice with mutations that prevent disposal of apoptotic cells ( 64 , 65 ) suggest that the purpose of this inhibitory pathway is to prevent dendritic cells from activating self-reactive T cells in the absence of inflammation. Presentation of High-Affinity Foreign Peptide–MHC Ligands Unlike the presentation of low-affinity self-peptide–MHC ligands that maintains the survival of naïve T cells in interphase, the presentation of high-affinity foreign peptide–MHC ligands induces the specific T cells to produce lymphokines and proliferate. These more dramatic biological effects occur because the responding T cells express TCRs with high affinity for the foreign peptide–MHC ligands, and thus receive stronger or more durable signals through the TCR ( 66 ). In addition, foreign antigens naturally enter the body during infection or tissue damage and the accompanying inflammation improves the quantity, quality, and type of cells that present foreign peptide–MHC ligands to T cells ( 67 , 68 ). As detailed below ( Table 1), the type of APC that presents foreign peptide–MHC ligands to naïve T cells is also influenced by the physical properties of the antigen and the site where it enters the body.

TABLE 1. Dendritic cell subsets that present peptide–MHC ligands in vivo Particulate Antigens The presentation of particulate antigens has been studied by tracking the fate of fluorescent microbeads ( 69 ). After subcutaneous injection, the beads are engulfed by monocytes that quickly enter the injection site from the blood, probably in response to inflammatory mediators made as a consequence of tissue damage resulting from the injection. Within 24 hours, cells containing more than one bead and expressing high amounts of MHC I and II molecules and the dendritic cell-specific molecules recognized by the MIDC-8 and 2A1 antibodies but low amounts of CD11c, appear in the draining lymph nodes. Monocytes that engulf particles and migrate in vitro across an artificial endothelial cell layer in an ablumenal-to-lumenal direction, also acquire these phenotypic characteristics ( 70 ). Together these results suggest that monocytes that enter tissue sites and engulf particles receive signals to differentiate into dendritic cells in the process of migrating from the tissue into the lumen of an afferent lymphatic vessel. These dendritic cells must appear in the lymph nodes transiently after tissue damage or change phenotype after migrating because CD11c low, MHC high cells are rare in the secondary lymphoid organs under normal conditions. Although it is reasonable to assume that these cells would be capable of presenting peptide–MHC ligands to T cells after arriving in the T-cell areas, this remains to be determined. CD11c high cells containing one bead also appear in the draining lymph nodes after subcutaneous injection of fluorescent beads. It is therefore possible that these cells, which are probably interstitial dendritic cell migrants, could be the important APC for particulate antigens that enter the dermis. In either case, the bead-containing cells that appear in the lymph nodes after subcutaneous injection are most likely important for presentation of peptide–MHC II ligands to CD4 T cells because, as described next, a different dendritic cell type is responsible for presentation of peptide–MHC I ligands derived from particulate antigens. Naïve CD8 T cells require signals through the TCR, CD28, and IL-12 receptor to proliferate maximally and differentiate into cytotoxic effector cells ( 71 , 72 and 73 ). Thus, although other cells express MHC I molecules in the T-cell areas, only dendritic cells express the ligands for these receptors and produce IL-12 ( 74 ), implying that dendritic cells play an important role as APC for naïve CD8 T cells. The TCR ligands for CD8 T cells are normally generated from endogenous proteins produced in the cytoplasm of the APC ( 75 ). This would be the case for particulate antigens, such as viruses that directly infect dendritic cells and replicate in the cytoplasm. However, if dendritic cells are necessary APC for naïve CD8 T cells, and peptide–MHC I complexes can only be produced from endogenous proteins, it is less clear how CD8 T-cell responses are initiated in cases where the antigen does not replicate directly within dendritic cells. The answer to this question is provided by the existence of dendritic cells that violate the classical rules of antigen processing and are capable of producing peptide–MHC I complexes from exogenous antigens. This capacity is sometimes called cross-presentation or cross-priming ( 76 , 77 ), and is particularly prominent for particulate antigens. Recent evidence indicates that the cross-priming APC is a CD11c +, CD8a + dendritic cell. Intravenous injection of ovalbumin-pulsed, MHC I-deficient splenocytes into normal mice

leads to the activation of ovalbumin peptide–MHC I-specific CD8 T cells ( 78 ). Because the injected cells are incapable of directly presenting ovalbumin peptide–MHC I complexes, cross-priming APC from the recipient must acquire ovalbumin from the injected cells and present ovalbumin peptide–MHC I complexes to the T cells. These cells are probably lymphoid dendritic cells because CD11c +, CD8a high but not CD11c +, CD8a - cells isolated from the spleens of injected mice stimulate ovalbumin peptide–MHC I-specific CD8 T cells in vitro. The simplest explanation for these results is that lymphoid dendritic cells are one of the few cell types in the body that deliver exogenous antigens into the cytosol and then the proteosome- and TAP-dependent MHC I processing pathway ( 79 , 80 ). This capacity of CD8a + dendritic cells is enhanced by CD4 T cells via a CD40-dependent mechanism ( 81 , 82 ). The facts that CD8a + dendritic cells are better IL-12 producers than CD8a - dendritic cells ( 74 ) and that IL-12 enhances the proliferation of CD8 T cells ( 73 ) provide further evidence that CD8a + dendritic cells are important APC for CD8 T cells, at least when CD4 T cells are also present. Skin-Surface Antigens A rich literature suggests that Langerhans cells are involved in the presentation of antigens that enter the body through the epidermis ( 38 , 83 , 84 ). This contention is based largely on experiments done with reactive haptens that, when applied to the skin surface covalently attach to soluble proteins within the epidermis. Langerhans cells efficiently take up hapten-labeled proteins and could produce haptenated peptide– MHC I and II complexes via the exogenous antigen-processing pathway mentioned above ( 85 , 86 and 87 ). Alternatively, haptens could couple directly to self-peptide–MHC I or II molecules already on the surface of Langerhans cells ( 88 ). Chemically reactive haptens stimulate the production of inflammatory cytokines such as IL-1 and TNF-a within the epidermis ( 83 , 89 ). As mentioned above, these cytokines cause Langerhans cells to leave the epidermis ( 53 , 54 ), enter local afferent lymphatic vessels and migrate into the T-cell areas of the draining lymph nodes where they can be found about 24 hours after application of the hapten ( 45 , 83 ). Hapten-labeled interstitial dendritic cells from the dermis also arrive in the lymph nodes but only after 48 hours, probably because they are located further from the skin surface than Langerhans cells ( 45 ). Since inflammatory cytokines are also inducers of the maturation process, Langerhans cells or interstitial dendritic cells that migrate from inflamed skin express elevated levels of co-stimulatory molecules and are potent stimulators of CD4 and CD8 T cells ( 90 , 91 and 92 ). The finding that manipulations that reduce Langerhans cell density or function reduce the amount of T-cell priming induced by reactive haptens is evidence that migrating Langerhans cells present antigen to T cells in vivo ( 93 ). Soluble Antigens Along their length, mucosal surfaces have lymph node-like organs that abut the intestinal epithelium on the albumenal side ( 94 ). The best studied of these mucosal lymphoid organs are the Peyer’s patches of the intestinal mucosa, although all mucosal surfaces probably have similar structures. The side of the mucosal lymphoid organ that attaches to the epithelium contains M cells that bring lumenal contents into the organ ( 94 ). Since mucosal lymphoid organs do not contain afferent lymphatic vessels, M cell sampling is the major mechanism whereby antigens enter these organs. Within the Peyer’s patch, just beneath the M cells, lies the subepithelial dome, which is rich in CD11c +, CD11b +, CD8a - dendritic cells ( 95 ). CD11c +, CD11b -, CD8a + dendritic cells are concentrated in the T-cell area, beneath the subepithelial dome. A

third CD11c + population that lacks CD11b and CD8a, and found only in mucosal lymphoid organs, is located in the subepithelial dome and the T-cell area. The fact that CD11c +, CD11b +, CD8a - dendritic cells are situated near the M cells where antigen enters the Peyer’s patches suggests that these cells present peptide–MHC II ligands derived from antigens that enter through mucosal surfaces. This contention is further supported by the fact that CD11c +, CD11b +, CD8a - dendritic cells migrate into the central T-cell areas in response to inflammation ( 95 ). The relationship between CD11c +, CD11b +, CD8a - dendritic cells in mucosal lymphoid organs and those in other secondary lymphoid organs is unknown. However, since these cells express CD11b and lack CD8a and could not have migrated from nonlymphoid tissue via afferent lymphatic vessels, they are probably myeloid dendritic cells. CD11c +, CD11b +, CD8a - dendritic cells have also been implicated in the presentation of soluble antigens that enter lymph nodes from the mucosal surface of the trachea. Following instillation of labeled ovalbumin into the trachea, labeled CD11c +, MHC high cells appear in the lung-associated lymph nodes. When isolated, these dendritic cells stimulate ovalbumin-specific CD4 T cells in vitro ( 96 ). Since these cells express CD11b and CD205, but most lack CD8a, they are probably interstitial dendritic cell migrants. Manickasingham and Reis e Sousa ( 97 ) used a monoclonal antibody specific for a chicken lysozyme peptide–MHC II complex to show that CD8a - dendritic cells also produce peptide–MHC II complexes derived from subcutaneously injected lysozyme. It is not clear from their studies whether the CD8a - dendritic cells that produce peptide–MHC II ligands in this case are myeloid dendritic cells or interstitial dendritic cell migrants because CD205 and CD11b expression was not assessed. Manickasingham and Reis e Sousa ( 97 ) also showed that CD8a + dendritic cells produce lysozyme peptide–MHC II complexes if lysozyme is injected with lipopolysaccharide (LPS). Although expression of CD8a indicates that these cells are lymphoid dendritic cells, they could have been Langerhans cell migrants if cells with intermediate levels of CD8a were included in the gate used to identify CD8 + cells. LPS-induced inflammation may cause CD8a + lymphoid dendritic cells that acquire lysozyme within the T-cell areas to produce peptide–MHC II complexes from this material, or stimulate the migration of Langerhans cells from the subcutaneous injection site. B cells in the follicles also produce peptide–MHC II complexes from lysozyme within several hours of injection ( 98 ). However, these peptide–MHC II complexes are probably inaccessible to naïve CD4 T cells, which reside only in the T-cell areas ( 21 , 22 ). The site where the dendritic cells in the aforementioned studies acquired antigen is not clear. They could have acquired antigen at the point of entry into the body before migrating to the lymph node, or they could have acquired lymph-borne antigen after arriving in the lymph node. Subcutaneously injected proteins are deposited in the extracellular fluid of the tissue at the injection site. This fluid, also know as lymph, is constantly siphoned from the tissues into blind-ended afferent lymphatic capillaries that are present in most organs. Thus, antigens that are injected into tissues would be drawn in the lymph into an afferent lymphatic vessel and then into the subcapsular sinus of a connected lymph node ( 99 ). Thin conduits connect the subcapsular sinus to perivenular spaces that surround the HEV that pass through the lymph node ( 100 ) ( Fig. 1B). The conduits are composed of thin fibers of extracellular matrix proteins wrapped continuously on the outside with a coating composed of 90% reticular fibroblasts and 10% other cells, including dendritic

cells ( 100 ). The lumen of each conduit is not completely filled with the fibers because soluble molecules pass from the subcapsular sinus through the conduits and into the perivenular space. Surprisingly, however, soluble molecules do not pass in large amounts from the subcapsular sinus or conduits into the T-cell–rich paracortex where naïve T cells reside ( 99 ). Thus, the resident dendritic cells that would have the best access to antigens present within the conduits would be those that coat the conduits and are exposed to high concentrations of antigen within the conduit lumen, or those that are near the conduits to take up the small amount of antigen that leaks out. CD11c +, CD11b + dendritic cells are the best candidates because they are concentrated in the outer paracortex ( 101 ) where the conduit network is most dense ( 99 ). Expression of CD11b indicates that these cells are Langerhans cell migrants, interstitial dendritic cell migrants, or myeloid dendritic cells. If these cells produce peptide–MHC complexes from antigen that leaks from the conduits, then this implies that Langerhans cell migrants and interstitial dendritic cells migrants are still capable of antigen uptake and processing even after migrating. If on the other hand, lymph-borne free antigen within the conduits is not accessible to APC in the lymph node, then the APC must acquire antigen before migrating to the lymph node. In this case, the CD11c +, CD8a - cells identified as APC in the studies of Vermaelen et al. ( 96 ) and Manickasingham and Reis e Sousa ( 97 ) are probably interstitial dendritic cells that acquired the antigen in the tissue where it entered the body. CD11b +, CD8a - dendritic cells are also involved in the splenic presentation of peptide–MHC II ligands derived from intravenously injected antigens. Pooley et al. ( 102 ) found that CD8a - dendritic cells isolated from the spleens of mice injected intravenously with ovalbumin were better CD4 T-cell stimulators than CD8a + dendritic cells. These CD8a - dendritic cells are likely to be myeloid or plasmacytoid dendritic cells because these are the only CD8a - populations in the spleen. Since fluorescent-labeled ovalbumin was taken up by splenic CD8a + dendritic cells in this case, Pooley et al. ( 102 ) attributed the failure of these cells to stimulate CD4 T cells to an inability to produce peptide–MHC II complexes from the internalized antigen. This finding is reminiscent of in vitro results in which immature dendritic cells took up antigen but did not produce peptide–MHC II complexes unless exposed to an inflammatory stimulus ( 56 ). However, this property does not explain the failure of CD8a + dendritic cells to produce peptide–MHC II complexes from the internalized antigen, since addition of LPS to the injected ovalbumin did not correct the failure ( 102 ). In contrast, another study that used a monoclonal antibody specific for a chicken lysozyme peptide–MHC II ligand showed that CD8a - and CD8a + dendritic cells in the spleen participate in the presentation of peptide–MHC complexes derived from intravenously injected lysozyme ( 103 ). The number of CD8a - and CD8a + dendritic cells displaying lysozyme peptide–MHC II complexes in the spleen is greatly enhanced in the presence of LPS-induced inflammation ( 103 ). Inflammation may stimulate the migration of dendritic cells from the red pulp into the T-cell areas as described by DeSmedt et al. ( 52 ). As in the case of particulate antigens, CD11c +, CD8a + dendritic cells in the spleen produce peptide–MHC I ligands from intravenously injected antigen ( 102 ), again suggesting that these cells are capable of producing peptide–MHC I ligands from exogenous material.

T-CELL ACTIVATION Signal Transduction In vitro experiments have shown that high-affinity TCR ligation activates protein tyrosine kinases such as Lck, which stimulate signaling cascades that elevate intracellular calcium, convert Ras into its active form, and activate the extracellular signal-regulated kinases (ERK1 and ERK2) and stress-activated protein kinases (Jun kinase and p38 mitogen-activated protein kinase) ( 104 ). These pathways culminate in the nuclear translocation and DNA binding of transcription factors that regulate lymphokine gene expression ( 105 ). Very little is known about early signaling events in naïve T cells in vivo because the assays used to measure most of these events rely on cell lines and in vitro culture methods. However, intracellular staining with antibodies that recognize the active forms of the c-Jun transcription factor and the p38 mitogen-activated protein kinase has been used to show that TCR signaling is initiated in antigen-specific naïve CD4 T cells in the spleen within minutes of intravenous injection of the relevant peptide ( 59 ). This rapid response is likely explained by the fact that the majority of naïve CD4 T cells are in contact with MHC II–expressing dendritic cells, at all times ( 59 ). Since the peptide used in this experiment does not require antigen processing ( 106 ), it would be able to immediately bind to MHC II molecules on dendritic cells, and activate the interacting antigen-specific T cells. Although these results show that in vivo TCR signaling commences very quickly after recognition of peptide–MHC complexes, this process would take longer in cases where the relevant APC must process the antigen and/or migrate into the T-cell areas from another location. IL-2 Production and Proliferation Naïve CD4 and CD8 T cells produce IL-2 in vivo within the first day after TCR ligation ( 107 108 , ). In vitro experiments indicate that cell division by naïve, antigen-stimulated T cells is driven by autocrine production of IL-2 ( 109 ). Surprisingly, however, antigen-driven proliferation of naïve T cells is minimally dependent on IL-2 in vivo ( 107 , 110 111 112 , , and 113 ). Therefore, other signals or growth factors must be capable of driving T-cell proliferation in vivo, although IL-2 may contribute. As noted below, IL-2 plays an important role in the elimination of activated T cells. The dual function of IL-2 as both a T-cell growth factor early in the response and an inhibitory factor later, may make it difficult to reveal the growth factor activity of IL-2 in IL-2–deficient animals. Naïve CD4 and CD8 T cells shown signs of DNA replication and cell division as early as 48 hours after exposure to antigen in vivo ( 21 , 108 , 114 ). These events are followed by an exponential increase in the number of antigen-specific T cells over the next several days. Depending on the stimulus, the number of antigen-specific T cells reaches its highest level in the relevant secondary lymphoid organs, 3 to 7 days after antigen enters the body ( 6 , 7 , 21 , 115 , 116 , 117 , 118 and 119 ) ( Fig. 2). Recently it has been estimated that naïve mice contain about 200 CD8 T cells specific for a given peptide–MHC I complex ( 120 ). Since antigen-specific CD8 T cells specific for a single peptide–MHC I complex can

increase to 10 7 cells at the peak of the primary response ( 121 ), it follows that CD8 T cells can expand 500,000-fold in vivo. Although naïve CD4 T cells are also capable of dramatic clonal expansion when stimulated appropriately, their burst size appears to be less than CD8 T cells ( 121 ).

FIG. 2. A: Kinetics, quantities, and phenotypes of antigen-specific CD4. B: Kinetics, quantities, and phenotypes of antigen-specific CD8. T cells during the primary immune response. RO and RA denote the CD45RO and CD45RA isoforms, respectively. LT denotes the set of molecules that are required for migration through HEV—CCR7, CD62L, and CD11a/CD18. NLT denotes the set of molecules involved in migration into nonlymphoid tissues, such as sPSGL-1, ß1 and ß7 integrins, and CCR5. Although a hypothetical situation is depicted, the number of cells shown is based on the work of Homann et al. ( 121 ), in which the number of CD4 or CD8 T cells specific for single peptide–MHC II or peptide–MHC I ligands was monitored during the course of a viral infection.

Several factors influence the magnitude of in vivo T-cell proliferation. One is the size of the starting naïve population. The degree of proliferation is inversely correlated with the starting frequency of responding cells. In cases where the starting frequency is relatively high, for example after transfer of TCR transgenic T cells, the clonal burst size of this population is relatively low, probably as a result of competition between the T cells for peptide–MHC complexes ( 10 , 11 ). In vivo T-cell proliferation is also regulated by co-stimulatory signals from APC. The proliferation of antigen-stimulated CD4 or CD8 T cells is reduced by two- to ten-fold in mice in which CD28 cannot interact with its ligands CD80 and CD86 ( 107 , 114 , 122 , 123 ).

CD40 ligand deficiency has a similar effect on T-cell expansion ( 124 , 125 and 126 ), which may be related to the fact that CD40 signaling in APC induces CD80 and CD86 ( 127 ). Co-stimulatory signals regulate T-cell proliferation by enhancing growth factor production. Antigen-driven IL-2 production is greatly impaired when CD28 signaling is eliminated ( 107 ). Although it has been proposed that CD28 acts by promoting TCR aggregation in the synapse at the point of contact between the T-cell and APC ( 128 ), recent work indicates that CD28 is actually recruited into the synapse after it forms ( 129 ). CD28 then transduces signals that enhance lymphokine mRNA production and stability and promote T-cell survival by augmenting Bcl-XL production ( 130 ). Members of the TNF receptor family, such as OX40, CD27, and 4-1BB are induced on activated T cells several days into the primary response ( 131 ). These molecules bind ligands of the TNF family on the surface of APC and transduce signals that sustain the proliferation or survival of antigen-stimulated T cells ( 131 , 132 ). Enhancement of co-stimulatory signals may underlie the observation that in vivo T-cell proliferation is also influenced by inflammation at the time of initial antigen presentation. The effect of inflammation is easily observed in the case of soluble antigens, where the magnitude of T-cell proliferation is several fold greater if antigen is administered with an adjuvant that induces inflammation or with inflammatory cytokines such as TNF-a, IL-1, or IL-12 ( 21 , 133 , 134 and 135 ). Adjuvant molecules are recognized by pattern recognition receptors ( 136 ), for example, Toll-like receptors (TLR) on cells of the innate immune system. The expansion of CD4 T cells in response to antigen plus complete Freund’s adjuvant is deficient in mice that lack a functional TLR signaling pathway ( 137 ). The defect is probably related to the fact that TLR signaling stimulates tissue macrophages to produce TNF-a ( 138 ), which in turn stimulates dendritic cells to migrate from nonlymphoid tissues into the T-cell areas. TLR signaling causes APC to express higher levels of ligands for CD28 and produce inflammatory cytokines ( 139 ). Thus, adjuvants could enhance proliferation by driving more dendritic cells into the T-cell areas to present antigen, or by increasing the co-stimulatory capacity of the dendritic cells. It is also likely that inflammatory cytokines stimulate proliferation by acting directly on the T cells. Support for this possibility comes from in vitro experiments that show that the proliferation of highly purified CD4 T or CD8 T cells in response to plastic surfaces coated with TCR and CD28 ligands is augmented by IL-1 or IL-12, respectively ( 73 , 140 ). These cytokines probably act by enhancing T-cell responsiveness to growth factors. For example, IL-1 has been shown to enhance IL-4–driven proliferation of CD4 Th2 clones ( 141 ). Effector Cells Antigen-specific T cells that are present at the time when the number of antigen-specific T cells reaches its peak express effector functions, and thus are sometimes to referred to as “effector cells” ( 142 ). Effector cells are blasts, express a different set of adhesion receptors, and possess different functional capabilities than naïve T cells. The functional properties that effector cells acquire are influenced by the presence of inflammatory cytokines and co-stimulatory ligands on APC present at the time of initial antigen

presentation. At least two types of antigen-specific effector CD4 T cells can be identified in the draining lymph nodes of mice injected subcutaneously with antigen and cholera toxin based on expression of CD62L and the functional, sialyated form of P-selectin ligand (sPSGL-1): CD62L +, sPSGL-1 + cells and CD62L-, sPSGL-1- cells ( 143 ) ( Fig. 2A). The CD62L +, sPSGL-1 + cells are poor helpers of antibody production by B cells but are capable of IFN-? production and cause delayed-type hypersensitivity (DTH) skin reactions when transferred into naïve recipients that are challenged with antigen. The DTH potential of these cells is explained by the fact that sPSGL-l is critical for T-cell migration through CD62P-expressing blood vessels into inflamed skin ( 144 ). Once in the skin, IFN-? production by CD62L +, sPSGL-1 + effector T cells likely causes some of the manifestations of the DTH reaction ( 145 ). Expression of sPSGL-1 on effector T cells is induced by exposure to IL-12 ( 146 ). The fact that IL-12 also controls acquisition of IFN-? production capacity ( 147 ) probably underlies the finding that IFN-?–producing effector cells are targeted preferentially to tissues like the skin that contains CD62P-expressing blood vessels ( 144 ). Effector CD4 T cells capable of migrating into nonlymphoid tissues have been identified in several other types of immune responses. When antigen is initially presented in the mucosal lymphoid organs, the nonlymphoid trafficking population of effector CD4 T cells is induced to express the a4/ß7 integrin instead of sPSGL-1 ( 148 ), and would be expected to migrate to mucosal tissues instead of the skin. CD4 T cells capable of rapid IFN-? production after challenge with antigen are found in the liver, lungs, thymus, salivary gland, and intestines of mice injected intravenously with antigen plus LPS ( 22 ). In addition, effector CD4 T cells capable of IL-4 production during Leishmania infection are found in the lungs ( 149 ). The fact that effector CD4 T cells produce IFN-? after exposure to antigen plus cholera toxin or LPS, or IL-4 after to exposure to Leishmania, is probably related to differences in the early production of IL-12 or IL-4 by cells of the innate immune system. LPS stimulates IL-12 production, which favors the differentiation of IFN-?–producing T cells, whereas Leishmania organisms stimulate IL-4 production, which favors the differentiation of IL-4–producing T cells ( 147 ). Together, these results suggest that antigenic stimulation within the secondary lymphoid organs produces a subset of IFN-?- or IL-4–producing effector CD4 T cells that migrate into inflamed nonlymphoid organs and mediate immune reactions there. Such reactions lead to macrophage and granulocyte activation, which is an efficient means of eliminating microbes and parasites. The CD62L -, sPSGL-1 - effector CD4 T cells found in the lymph nodes after injection of antigen and cholera toxin are efficient helpers of antibody production by B cells, do not cause DTH, and are poor IFN-? producers ( 143 ). These may be the CD4 T cells that migrate into the B-cell–rich follicles during the primary response ( 21 , 117 , 143 , 150 , 151 ). Follicular migration is controlled by CXCR5, which is specific for the CXCL13 (BLC) chemokine produced by follicular stromal cells ( 20 ). CXCR5 expression is induced on T cells several days after in vivo exposure to antigen and adjuvant, but not antigen alone ( 152 ), probably because CXCR5 induction and follicular migration are dependent on

signals through CD28 and OX40, the ligands for which are induced on dendritic cells by inflammation. Migration into follicles allows effector CD4 T cells to interact with and provide helper signals to antigen-specific B cells that display the relevant peptide–MHC complexes ( 125 , 153 , 154 ). Surprisingly, CD62L -, sPSGL-1 - effector CD4 T cells do not produce IL-4 ( 143 ), which is thought to be a critical component of T-cell–mediated promotion of antibody production ( 147 ). In addition, although the CD62L -, sPSGL-1 effector CD4 T cells express CD40 ligand, another molecule that is critical for B-cell help, so do the CD62L +, sPSGL-1 + effector CD4 T cells that lack this activity ( 143 ). Therefore, the molecular basis for the potent B-cell helper function of CD62L -, sPSGL-1 effector CD4 T cells is unclear. It is also not clear how the two different types of effector CD4 T cells are produced simultaneously in the same secondary lymphoid organs during the primary response. Effector CD8 T cells also differ from their naïve precursors with respect to surface markers, function, and trafficking properties ( Fig. 2B). Effector CD8 T cells that are generated during microbial infections express slightly lower levels of CD8 and more surface O-glycans than naïve cells ( 155 ), and in the human some effector CD8 T cells lose CD27 and CD28 but retain CD45RA ( 156 ). Unlike naïve cells, these effector CD8 T cells express perforin and granzymes, which are required for efficient cytolytic function ( 155 ). Expression of perforin and granzymes contributes to the defining feature of effector CD8 T cells, that is, the ability to directly kill target cells that display the appropriate peptide–MHC I complexes. Interestingly, although large numbers of antigen-specific CD8 T cells accumulate in mice injected with a heat-killed microbe, these T cells do not acquire cytolytic function ( 157 ). Since the T cells undergo fewer cell divisions under these conditions, it is possible that CD8 T cells must divide many times before becoming effector cytolytic cells. Effector CD8 T cells gain the capacity to produce IFN-?, but lose the capacity to produce IL-2, thus becoming dependent on IL-2 from CD4 T cells for further proliferation ( 158 ). The loss of CD28 function by effector CD8 T cells may contribute to the loss of IL-2 production capacity ( 159 , 160 ). Effector CD8 T cells migrate out of the T-cell areas and into many nonlymphoid tissues, particularly inflamed sites of antigen deposition, such as the lungs during influenza infection ( 161 , 162 ) and the gut during vesicular stomatitis virus infection ( 163 ). In vitro experiments indicate that exposure to IL-2 is an important factor in the generation of nonlymphoid tissue-homing effector CD8 T cells ( 164 ). The migratory capacities of effector CD8 T cells correlate with loss of receptors involved in lymph node migration (CCR7 and CD62L) and acquisition of receptors such as a4ß7 integrin ( 165 ), which binds to MadCAM-1 expressed on blood vessels in mucosal organs. The migration of effector CD8 T cells with cytotoxic potential into nonlymphoid organs is an effective way of eliminating cells that display peptide–MHC I complexes from all parts of the body. The number of effector T cells in the secondary lymphoid organs falls dramatically after the peak of proliferation ( 6 , 7 , 21 , 115 , 116 , 117 , 118 and 119 ) ( Fig. 2). Some of this loss is due to the emigration of effector cells into nonlymphoid tissues as mentioned above ( 22 ). However, much of the loss must be due to cell death because apoptotic antigen-specific effector T cells can be identified in the secondary lymphoid organs ( 166

), and because the total number of cells in the nonlymphoid organs declines shortly after its peak ( 22 ). The molecular basis for the death of effector T cells varies depending on the nature of the antigenic stimulus. The loss of effector CD8 T cells after the peak of proliferation in response to a single injection of antigen has been shown to be Fas-independent and Bcl-2 sensitive ( 167 ). This type of cell death has been observed in situations where cells are deprived of growth factors ( 168 ). This is a reasonable scenario because antigen-specific T cells stop making lymphokines at least 1 day before effector cells begin to disappear in hosts injected once with antigen ( 107 ). On the other hand, if antigen is presented chronically, TCR-mediated activation-induced cell death may occur ( 169 ). This type of apoptosis is dependent on Fas and is poorly inhibited by Bcl-2 ( 168 ). This scenario is plausible because chronic activation causes expression of Fas on T cells ( 170 ). In addition, a death pathway involving Fas could explain the role of IL-2 in activation-induced death of effector cells, because IL-2 prevents the activation of FLICE inhibitor protein, which normally inhibits Fas signaling ( 171 ). Yet another form of T-cell death has been described in studies of superantigen-induced T-cell activation. Superantigen-stimulated effector T cells die after peak proliferation by a mechanism that involves internal production of reactive oxygen species, but not Fas, TNF receptors, or caspases ( 172 ). Reactive oxygen species may damage mitochondrial membranes leading to metabolic dysfunction and apoptosis. The death of effector T cells is regulated by inflammation. In the absence of inflammation, the loss of antigen-specific T cells from the secondary lymphoid and nonlymphoid organs after the peak of proliferation is nearly complete ( 22 ). In contrast, many more cells survive the loss phase in both types of organs after injection of antigen or superantigen plus adjuvants such as LPS or IL-1 ( 21 , 22 , 134 , 173 ). This sparing effect can be induced by injection of LPS 24 hours after superantigen injection ( 173 ), and equally well in normal and CD28-deficient mice ( 174 ). Because lymphokine production by antigen-stimulated T cells is CD28 dependent ( 107 ), it is unlikely that this is the target of this late adjuvant effect. It is possible that LPS promotes survival by protecting T cells from the toxic effects of reactive oxygen species by inducing the Bcl-3 survival protein ( 175 ). Memory Cells Although the vast majority of effector cells die after the peak of proliferation, a stable population of antigen-experienced T cells survives for long periods of time if the antigen was initially presented in an inflammatory context ( 142 ). These long-lived “memory” cells are capable of very rapid responses that can produce protective immunity to a later challenge with a microbe ( 176 ). Memory cells can be distinguished from effector cells in that most memory cells are not blasts, are not in the cell cycle, and many are not directly cytolytic or producing lymphokines ( 142 ). In many ways, memory cells can be thought of as effector cells that have returned to a basal activation state. Indeed, several lines of evidence suggest that effector cells are precursors of memory cells ( 177 , 178 ). Memory CD8 T Cells Antigen-specific memory CD8 T cells have been studied

extensively in viral and bacterial infections. The number of naïve antigen-specific CD8 T cells in the secondary lymphoid organs increases manyfold during the first week after infection, falls dramatically as effector cells die, and achieves a stable level about 2 weeks after infection that is lower than the peak level but higher than the starting level ( 6 7 121 , , ) ( Fig. 2B). The number of antigen-specific CD8 T cells then does not change for the life of the host, at least in the case of one viral infection in mice that have a life span of about 2 years ( 121 , 179 ). Unlike naïve CD8 T cells, memory CD8 T cells do not depend on MHC I molecules for survival ( 180 ). Thus, memory CD8 T-cell survival cannot be explained by chronic TCR signaling as a result of recognition of peptide–MHC I complexes derived from persistent antigen. Whereas most memory CD8 T cells are not cycling, a small fraction of the memory CD8 T population is proliferating in an MHC I–independent fashion at all times ( 111 , 180 ). This proliferation must be balanced by death since the total number of antigen-specific memory CD8 T cells does not change over time. Several observations suggest that IL-15 plays a role in this process. The antigen-independent proliferation of memory CD8 T cells is accelerated by injection of IL-15 ( 181 ) and blocked by injection of antibodies that prevent IL-15 from binding to its receptor ( 111 ). In addition, memory CD8 T cells are diminished in IL-15–deficient mice ( 182 ). Since IL-15 is produced by non–T cells during the innate immune response, it is possible that memory CD8 T cells are maintained as a consequence of IL-15 produced in response to other infections ( 181 , 183 ). Memory CD8 T cells are heterogeneous. Human memory CD8 T cells can be divided into at least three subsets with the following phenotypes: CD45RA -, CCR7 +; CD45RA -, CCR7 -; and CD45RA +, CCR7 - ( 156 , 184 ) ( Fig. 2B). The CD45RA -, CCR7 + memory cells also express CD62L and therefore are expected to recirculate through secondary lymphoid organs including lymph nodes and mucosal lymphoid organs ( 184 ). CD45RA -, CCR7 + memory CD8 T cells lack perforin and thus would not be expected to be directly cytotoxic ( 184 ). Virus antigen-specific CD8 T cells with these features are present in the lymphoid organs of mice beginning several weeks after viral infection ( 163 ). The two CCR7 - subsets also lack CD62L ( 184 ) and therefore could not enter lymph nodes and mucosal lymphoid organs through HEV. On the other hand, subsets within these populations express high levels of ß1 and ß7 integrins, sPSGL-1, and CCR5 ( 184 ); these molecules facilitate migration into nonlymphoid tissues, especially in the presence of inflammation ( 185 ). Both of the CCR7 - subsets contain perforin ( 184 ) and are thus likely to be cytotoxic. The CD45RA +, CCR7 - subset possesses especially high levels of perforin and the direct ex vivo cytotoxic function of these cells has been demonstrated ( 156 , 184 ). Both of the CCR7 subsets produce IFN-? rapidly after in vitro stimulation ( 184 ). All things considered, the CCR7 - subsets of memory CD8 T cells are very similar to effector CD8 T cells, and have in fact been referred to as effector memory cells ( 185 ). The finding of virus antigen-specific CD8 T cells with ex vivo cytotoxic function in the nonlymphoid organs of mice weeks after viral infection lends credence to the existence of these effector memory cells ( 163 ). It is possible that the subset of cycling memory CD8 T cells observed in murine studies are the effector memory cells. Memory CD4 T Cells The number of antigen-specific CD4 T cells in the body drops sharply several days after the peak accumulation of effector cells, to a level that is lower than the peak and greater than the starting level ( 21 , 116 , 117 , 121 ) ( Fig. 2A). The

antigen-specific CD4 T cells that are present at this time are not cycling blasts and thus can be considered memory cells. In one type of viral infection, the number of virus antigen-specific CD4 T cells then continues to fall at a slow rate over the next year ( 121 ), indicating that memory CD4 T cells are not indefinitely maintained as are memory CD8 T cells ( Fig. 2A). This possibility is supported by other evidence of instability, including the findings that memory CD4 T cells revert some surface markers to the naïve phenotype over time ( 186 , 187 and 188 ), and lose enhanced helper function in hosts that contain normal numbers of T cells ( 189 ). Memory in the CD4 compartment may wane because cells that die are not replaced by proliferation of other memory T cells from the same cohort as in the case of memory CD8 T cells. This possibility is supported by the finding that the IL-15 growth factor does not enhance the proliferation of memory CD4 T cells ( 181 ). It should be noted that other experiments indicate that memory CD4 T cells are just as stable as memory CD8 T cells. For example, antigen-specific CD4 T cells that are stimulated in vitro with antigen and then transferred into T-cell–deficient hosts survive for months even in the absence of MHC II molecules ( 190 191 , ). Similarly, antigen-specific CD4 T cells retain the CD44 high phenotype and the capacity to produce IFN-? for months after exposure to antigen in hosts that lack T cells with other specificities ( 192 ). Although these experiments indicate that memory CD4 T cells can survive indefinitely in the absence of antigen, it is possible that persistence is related to the “space-filling” homeostatic proliferation that occurs in hosts that lack other T cells. As in the case of CD8 T cells, the population of memory CD4 T cells that survives after the death of effector cells is heterogeneous. Humans have at least two populations of memory CD4 T cells in peripheral blood, both lacking CD45RA (and presumably expressing CD45RO); one expresses CCR7 and the other lacks CCR7 ( 184 ) ( Fig. 2A). The CD45RA -, CCR7 + cells produce IL-2 rapidly when stimulated with anti-CD3 antibody in vitro, but do not produce IFN-? or IL-4 ( 184 ). The cells in this population express high levels of CD62L and thus would be expected to circulate through secondary lymphoid organs including lymph nodes, although subsets express CCR4, CCR6, and CXCR3 and thus could migrate into certain sites of inflammation ( 184 ). The existence of such lymphoid tissue-seeking memory cells is supported by the presence of antigen-specific CD4 T cells capable of rapid IL-2 but not IFN-? production in the lymph nodes of mice several months after exposure to antigen ( 22 ). CD45RA -, CCR7 - memory CD4 T cells differ from CD45RA -, CCR7 + memory cells with respect to function and trafficking. CD45RA -, CCR7 - memory CD4 T cells produce IFN-?, IL-4, and IL-5 rapidly when stimulated with anti-CD3 antibody in vitro, but are poor producers of IL-2 under these conditions ( 184 ). These cells express low or variable levels of CD62L and high levels of fPSGL-1, and/or ß1 and ß7 integrins ( 184 ). This expression pattern predicts that these cells would be excluded from lymph nodes but could enter nonlymphoid sites of inflammation. This possibility is supported by the fact that the nonlymphoid tissues, especially liver, lungs, and gut are major reservoirs of antigen-experienced CD4 T cells in mice after effector cells disappear in a response induced by intravenous injection of antigen plus adjuvant ( 22 ). Like CD45RA -, CCR7 human CD4 T cells, the murine memory CD4 T cells in nonlymphoid tissues are potent IFN-? producers but produce IL-2 poorly ( 22 ). Because Mackay et al. ( 193 ) found that memory T cells are constantly coming out of tissues and into afferent lymphatic vessels, it is possible that memory CD4 T cells in nonlymphoid tissues are not fixed there but

recirculate through the spleen and/or nonlymphoid tissues. The relationship between the antigen-specific effector CD4 T cells that are present at the peak of the response and the memory cells that survive is unclear. The lymphoid tissue-seeking memory cells ( 22 ) are similar to the CD62L -, sPSGL-1 - effector CD4 T cells ( 143 ) with respect to poor IFN-? and IL-4 production, and thus could be derived from these cells. If so, then the lymphoid tissue-seeking memory cells may be potent B cell helpers like their CD62L -, sPSGL-1 - effector precursors ( 143 ). However for this scenario to be correct, the CD62L -, sPSGL-1 - effector CD4 T cells must re-express CD62L to be able to recirculate through the lymph nodes as memory cells. The similar production of IFN-? or IL-4 but not IL-2 by nonlymphoid–tissue seeking memory cells ( 22 ) and CD62L +, sPSGL-1 + effector CD4 T cells ( 143 ) suggests that the former derive from the latter. If this is correct, then the CD62L +, sPSGL-1 + effector CD4 T cells must lose CD62L as they become memory cells. It is also possible that effector cells give rise to lymphoid tissue-seeking CD45RA -, CCR7 + central memory cells, which in turn give rise to nonlymphoid tissue-seeking CD45RA -, CCR7 - effector memory cells. This possibility is supported by the finding that human CD45RA -, CCR7 + memory cells lose CCR7 after 10 days of in vitro stimulation and acquire the capacity to produce IFN-? ( 184 ). Confirmation of this linear relationship has been hampered by a lack of anti-mouse CCR7 antibodies. Thus, it has not been possible to analyze the CCR7 expression on a defined population of antigen-specific CD4 T cells at precise times and locations after a primary and secondary exposure to antigen.

SUMMARY What follows is an attempt to unify the information presented above into a hypothetical sequence of events that occurs in the lives of antigen-specific CD4 and CD8 T cells from the time that they first encounter antigen as naïve cells until they become memory cells. Since this process is not completely understood, certain aspects of this sequence are speculative. Educated guesses have been made to marry analogous information from studies of mice and humans, assuming that T cells from these species behave similarly. A naïve T cell spends its life of about 2 months, in a series of 1-day stops in the T-cell areas of different secondary lymphoid organs with intervening trips through the blood. While in the T-cell area, a naïve T cell receives survival signals through the IL-7 receptor as it binds to IL-7 made by stromal cells, and the TCR as it binds to the relevant selecting self-peptide–MHC ligand on the surface of an APC, probably a dendritic cell. In the absence of infection or tissue damage, these dendritic cells exist in a semisuppressed state characterized by low expression of co-stimulatory ligands, and caused by suppressive cytokines made by phagocytes as they engulf apoptotic senescent cells during normal homeostasis. A naïve T cell is roused from its survival program when it encounters an APC bearing the foreign peptide–MHC ligand for which its TCR has a high affinity. This APC will be a resident dendritic cell that captured free antigen in the T-cell area as it flowed in from the afferent lymph or blood, or a dendritic cell that acquired the antigen in a

nonlymphoid tissue and then migrated into the T-cell area, depending on the nature and entry point of the antigen. If the antigen is part of a microbe or is administered with an adjuvant, then signals from the innate immune system will directly or indirectly activate dendritic cells from their semisuppressed state, enhancing their rate of migration into the T-cell areas and increasing antigen processing, stabilization of peptide–MHC complexes on the cell surface, and expression of co-stimulatory ligands. Since dendritic cells are one of the few types in the body that are capable of producing both peptide–MHC I and peptide–MHC II ligands from exogenous antigens, they are uniquely suited for antigen presentation to naïve CD8 and CD4 T cells expressing the appropriate TCRs. Naïve CD4 T cells produce IL-2 within several hours of encountering an activated dendritic cell expressing the appropriate peptide–MHC II complexes and increased levels of co-stimulatory ligands. IL-2, other unknown T-cell growth factors, and co-factors such as IL-1 then stimulate the CD4 T cells to proliferate extensively, eventually leading to the development of CD62L +, sPSGL-1 + and CD62L -, sPSGL -1effector cells. The CD62L -, sPSGL - effector cells gain expression of CXCR5, allowing them to sense the follicular chemokine CXCL13 and migrate into the follicles to provide helper signals to antigen-specific B cells. The CD62L +, sPSGL-1 + effector cells acquire the capacity to produce IFN-? or IL-4, depending on the cytokines produced by innate immune cells, and then leave the secondary lymphoid organs through the efferent lymphatic vessels, enter the blood, and migrate into inflamed tissues where they produce IFN-? or IL-4 in response to antigen presentation by tissue APC. The activating effects of these lymphokines and antibodies on the microbicidal activities of macrophages and granulocytes, lead to elimination of the antigen. At this point, most of the effector cells die by apoptosis. However, some of the effector cells return to a resting state and survive as memory cells. CD62L +, sPSGL-1 + effector cells give rise to CCR7 - memory CD4 T cells that recirculate via the blood through the spleen, or the spleen and nonlymphoid tissues. This recirculation pattern would enable these memory cells to produce IFN-? or IL-4 rapidly during secondary immune responses in nonlymphoid tissues where antigens enter the body. The CD62L -, sPSGL - effector cells with B cell helper function may give rise to CCR7 + memory CD4 T cells that re-express CD62L and recirculate via blood and efferent lymph through spleen and lymph nodes like naïve cells. These memory cells may help memory B cells produce antibody in the lymphoid tissues during secondary immune responses, or they may proliferate to produce more effector cells. If antigen does not enter the body a second time, both populations of memory cells may disappear slowly over time because they do not proliferate to renew themselves. Naïve CD8 T cells also produce IL-2 within several hours of encountering an activated dendritic cell expressing the appropriate peptide–MHC I ligands. IL-2, other unknown T-cell growth factors, and co-factors such as IL-12 then stimulate the CD8 T cells to proliferate extensively. Since CD8 T cells rapidly lose the ability to produce IL-2, their proliferation is aided by IL-2 produced by CD4 T cells. The proliferating CD8 T cells develop into perforin-expressing cytotoxic effector cells, many of which rapidly migrate into nonlymphoid tissues. These effector CD8 T cells then kill cells in nonlymphoid tissues that display the relevant peptide–MHC I complexes. As such cells are eliminated, most of the effector CD8 T cells die by apoptosis. However, some of the

effector cells survive in nonlymphoid tissues for long periods of time as CCR7 - memory cells and retain their cytotoxic potential. Other noncytotoxic CCR7 + memory CD8 T cells survive in the lymphoid tissues. All memory CD8 T cells constitutively express the IL-15 receptor and use it to proliferate periodically in response to IL-15, which in turn is perhaps made in response to unrelated immune responses. This proliferation replaces memory cells that die and results in a constant number of memory CD8 cells for the life of the host, even in the absence of MHC I molecules. When exposed to antigen a second time, the memory CD8 T cells in nonlymphoid tissues rapidly kill cells displaying peptide–MHC I molecules, whereas the memory CD8 T cells in lymphoid organs cells proliferate extensively and rapidly acquire cytotoxic potential.

ACKNOWLEDGMENTS I am grateful to Drs. Mathew Mescher, Daniel Mueller, Andre Itano, and Stephen Jameson for their helpful suggestions. REFERENCES 1. Davis MM, Boniface JJ, Reich Z, et al. Ligand recognition by alpha beta T cell receptors. Annu Rev Immunol 1998;16:523. 2. Rothman AL, Yamada Y, Jameson J, et al. Assessment of human CD4+ and CD8+ T lymphocyte responses in experimental viral vaccine studies. Dev Biol (Basel) 1998;95:95. 3. Altman JD, Moss PAH, Goulder JR, et al. Phenotypic analysis of antigen-specific T lymphocytes. Science 1996;274:94. 4. Kozono H, White J, Clements J, et al. Production of soluble MHC class II proteins with covalently bound single peptides. Nature 1994;369:151. 5. O’Herrin SM, Lebowitz MS, Bieler JG, et al. Analysis of the expression of peptide–major histocompatibility complexes using high affinity soluble divalent T cell receptors. J Exp Med 1997;186:1333. 6. Murali-Krishna K, Altman JD, Suresh M, et al. Counting antigen-specific CD8 T cells: a reevaluation of bystander activation during viral infection. Immunity 1998;8:177. 7. Busch DH, Pilip IM, Vijh S, et al. Coordinate regulation of complex T cell populations responding to bacterial infection. Immunity 1998;8:353. 8. Kotzin BL, Falta MT, Crawford F, et al. Use of soluble peptide-DR4 tetramers to detect synovial T cells specific for cartilage antigens in patients with rheumatoid arthritis. Proc Natl Acad Sci U S A 2000;97:291. 9. Jenkins MK, Khoruts A, Ingulli E, et al. In vivo activation of antigen-specific CD4 T cells. Annu Rev Immunol 2001;19:23. 10. Kedl RM, Rees WA, Hildeman DA, et al. T cells compete for access to antigen-bearing antigen-presenting cells. J Exp Med 2000;192:1105. 11.

Smith AL, Wikstrom ME, Fazekas de St Groth B. Visualizing T cell competition for peptide/MHC

complexes: a specific mechanism to minimize the effect of precursor frequency. Immunity 2000;13:783. 12. Laouar Y, Crispe IN. Functional flexibility in T cells: independent regulation of CD4+ T cell proliferation and effector function in vivo. Immunity 2000;13:291. 13.

Davis MM. T cell receptor gene diversity and selection. Annu Rev Biochem 1990;59:475.

14. Kappler JW, Wade T, White J, et al. A T cell receptor Vb segment that imparts reactivity to a class II major histocompatibility complex product. Cell 1987;49:263. 15. Ramsdell F, Lantz T, Fowlkes BJ. A nondeletional mechanism of thymic self tolerance. Science 1989;246:1038. 16. McGargill MA, Derbinski JM, Hogquist KA. Receptor editing in developing T cells. Nat Immunol 2000;1:336. 17. Jameson SC, Hogquist KA, Bevan MJ. Positive selection of thymocytes. Annu Rev Immunol 1995;13:93. 18.

Ford WL, Gowans JL. The traffic of lymphocytes. Semin Hematol 1969;6:67.


Mackay CR. Homing of naive, memory and effector lymphocytes. Curr Opin Immunol 1993;5:423.


Cyster JG. Chemokines and cell migration in secondary lymphoid organs. Science 1999;286:2098.

21. Kearney ER, Pape KA, Loh DY, et al. Visualization of peptide-specific T cell immunity and peripheral tolerance induction in vivo. Immunity 1994;1:327. 22. Reinhardt RL, Khoruts A, Merica R, et al. Visualizing the generation of memory CD4 T cells in the whole body. Nature 2001;410:101. 23. Campbell JJ, Butcher EC. Chemokines in tissue-specific and microenvironment-specific lymphocyte homing. Curr Opin Immunol 2000;12:336. 24. Cyster JG, Goodnow CC. Pertussis toxin inhibits migration of B and T lymphocytes into splenic white pulp cords. J Exp Med 1995;182:581. 25. Alferink J, Tafuri A, Vestweber D, et al. Control of neonatal tolerance to tissue antigens by peripheral T cell trafficking. Science 1998;282:1338. 26.

Sprent J, Tough DF. Lymphocyte life-span and memory. Science 1994;265:1395.

27. Ferreira C, Barthlott T, Garcia S, et al. Differential survival of naive CD4 and CD8 T cells. J Immunol 2000;165:3689. 28. Rooke R, Waltzinger C, Benoist C, et al. Targeted complementation of MHC class II deficiency by intrathymic delivery of recombinant adenoviruses. Immunity 1997;7:123. 29. Tanchot C, Lemonnier FA, Perarnau B, et al. Differential requirements for survival and proliferation of CD8 memory T cells. Science 1997;276:2057–62. 30. Dorfman JR, Stefanova I, Yasutomo K, et al. CD4+ T cell survival is not directly linked to self–MHC-induced TCR signaling. Nat Immunol 2000;1:329. 31.

Schluns KS, Kieper WC, Jameson SC, et al. Interleukin-7 mediates the homeostasis of naive and

memory CD8 T cells in vivo. Nat Immunol 2000;1:426. 32. Tan JT, Dudl E, LeRoy E, et al. IL-7 is critical for homeostatic proliferation and survival of naive T cells. Proc Natl Acad Sci U S A 2001;98:8732. 33. Goldrath AW, Bevan MJ. Low-affinity ligands for the TCR drive proliferation of mature CD8+ T cells in lymphopenic hosts. Immunity 1999;11:183. 34. Bender J, Mitchell T, Kappler J, et al. CD4+ T cell division in irradiated mice requires peptides distinct from those responsible for thymic selection. J Exp Med 1999;190:367. 35. Ernst B, Lee DS, Chang JM, et al. The peptide ligands mediating positive selection in the thymus control T cell survival and homeostatic proliferation in the periphery. Immunity 1999;11:173. 36. Kieper WC, Jameson SC. Homeostatic expansion and phenotypic conversion of naive T cells in response to self peptide/MHC ligands. Proc Natl Acad Sci U S A 1999;96:13306. 37. Prlic M, Blazar BR, Khoruts A, et al. Homeostatic expansion occurs independently of costimulatory signals. J Immunol 2001;167:5664. 38.

Banchereau J, Steinman RM. Dendritic cells and the control of immunity. Nature 1998;392:245.

39. Vremec D, Shortman K. Dendritic cell subtypes in mouse lymphoid organs: cross-correlation of surface markers, changes with incubation, and differences among thymus, spleen, and lymph nodes. J Immunol 1997;159:565. 40. Pulendran B, Lingappa J, Kennedy MK, et al. Developmental pathways of dendritic cells in vivo: distinct function, phenotype, and localization of dendritic cell subsets in FLT3 ligand–treated mice. J Immunol 1997;159:2222. 41. Huleatt JW, Lefrancois L. Antigen-driven induction of CD11c on intestinal intraepithelial lymphocytes and CD8+ T cells in vivo. J Immunol 1995;154:5684. 42. Traver D, Akashi K, Manz M, et al. Development of CD8alpha-positive dendritic cells from a common myeloid progenitor. Science 2000;290:2152. 43. Manz MG, Traver D, Miyamoto T, et al. Dendritic cell potentials of early lymphoid and myeloid progenitors. Blood 2001;97:3333. 44. del Hoyo GM, Martin P, Arias CF, et al. CD8alpha(+) dendritic cells originate from the CD8alpha( - ) dendritic cell subset by a maturation process involving CD8alpha, DEC-205, and CD24 up-regulation. Blood 2002;99:999. 45. Ruedl C, Koebel P, Bachmann M, et al. Anatomical origin of dendritic cells determines their life span in peripheral lymph nodes. J Immunol 2000;165:4910. 46. Kamath AT, Pooley J, O’Keeffe MA, et al. The development, maturation, and turnover rate of mouse spleen dendritic cell populations. J Immunol 2000;165:6762. 47. Nakano H, Yanagita M, Gunn MD. CD11c(+)B220(+)GR-1(+) cells in mouse lymph nodes and spleen display characteristics of plasmacytoid dendritic cells. J Exp Med 2001;194:1171. 48. Rissoan MC, Soumelis V, Kadowaki N, et al. Reciprocal control of T helper cell and dendritic cell differentiation. Science 1999;283:1183.

49. Anjuere F, Martin P, Ferrero I, et al. Definition of dendritic cell subpopulations present in the spleen, Peyer’s patches, lymph nodes, and skin of the mouse. Blood 1999;93:590. 50. Henri S, Vremec D, Kamath A, et al. The dendritic cell populations of mouse lymph nodes. J Immunol 2001;167:741. 51. Valladeau J, Ravel O, Dezutter-Dambuyant C, et al. Langerin, a novel C-type lectin specific to Langerhans cells, is an endocytic receptor that induces the formation of Birbeck granules. Immunity 2000;12:71. 52. DeSmedt T, Pajak B, Muraille E, et al. Regulation of dendritic cell numbers and maturation by lipopolysaccharide in vivo. J Exp Med 1996;184:1413. 53. Cumberbatch M, Kimber I. Dermal tumour necrosis factor-alpha induces dendritic cell migration to draining lymph nodes, and possibly provides one stimulus for Langerhans’ cell migration. Immunology 1992;75:257. 54. Shornick LP, Bisarya AK, Chaplin DD. IL-1beta is essential for Langerhans cell activation and antigen delivery to the lymph nodes during contact sensitization: evidence for a dermal source of IL-1beta. Cell Immunol 2001;211:105. 55. Banchereau J, Briere F, Caux C, et al. Immunobiology of dendritic cells. Annu Rev Immunol 2000;18:767. 56. Turley SJ, Inaba K, Garrett WS, et al. Transport of peptide–MHC class II complexes in developing dendritic cells. Science 2000;288:522. 57. Cella M, Engering A, Pinet V, et al. Inflammatory stimuli induce accumulation of MHC class II complexes on dendritic cells. Nature 1997;388:782. 58. Brocker T. Survival of mature CD4 T lymphocytes is dependent on major histocompatibility complex class II–expressing dendritic cells. J Exp Med 1997;186:1223. 59. Zell T, Khoruts A, Ingulli E, et al. Single-cell analysis of signal transduction in CD4 T cells stimulated by antigen in vivo. Proc Natl Acad Sci U S A 2001;98:10805. 60.

Savill J, Fadok V. Corpse clearance defines the meaning of cell death. Nature 2000;407:784.

61. Fadok VA, Bratton DL, Rose DM, et al. A receptor for phosphatidylserine-specific clearance of apoptotic cells. Nature 2000;405:85. 62. Fadok VA, Bratton DL, Konowal A, et al. Macrophages that have ingested apoptotic cells in vitro inhibit proinflammatory cytokine production through autocrine/paracrine mechanisms involving TGF-beta, PGE2, and PAF. J Clin Invest 1998;101:890. 63. Yamaguchi Y, Tsumura H, Miwa M, et al. Contrasting effects of TGF-beta 1 and TNF-alpha on the development of dendritic cells from progenitors in mouse bone marrow. Stem Cells 1997;15:144. 64. Scott RS, McMahon EJ, Pop SM, et al. Phagocytosis and clearance of apoptotic cells is mediated by MER. Nature 2001;411:207. 65. Lu Q, Lemke G. Homeostatic regulation of the immune system by receptor tyrosine kinases of the Tyro 3 family. Science 2001;293:306.

66. Rosette C, Werlen G, Daniels MA, et al. The impact of duration versus extent of TCR occupancy on T cell activation: a revision of the kinetic proofreading model. Immunity 2001;15:59. 67. Janeway CA Jr. Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harb Symp Quant Biol 1989;54:1. 68.

Matzinger P. Tolerance, danger, and the extended family. Annu Rev Immunol 1994;12:991.

69. Randolph GJ, Inaba K, Robbiani DF, et al. Differentiation of phagocytic monocytes into lymph node dendritic cells in vivo. Immunity 1999;11:753. 70. Randolph GJ, Beaulieu S, Lebecque S, et al. Differentiation of monocytes into dendritic cells in a model of transendothelial trafficking. Science 1998;282:480. 71. Whitmire JK, Ahmed R. Costimulation in antiviral immunity: differential requirements for CD4(+) and CD8(+) T cell responses. Curr Opin Immunol 2000;12:448. 72. Gajewski TF, Renauld JC, Van Pel A, et al. Costimulation with B7-1, IL-6, and IL-12 is sufficient for primary generation of murine antitumor cytolytic T lymphocytes in vitro. J Immunol 1995;154:5637. 73. Curtsinger JM, Schmidt CS, Mondino A, et al. Inflammatory cytokines provide a third signal for activation of naive CD4+ and CD8+ T cells. J Immunol 1999;162:3256. 74. Reis e Sousa C, Hieny S, Scharton-Kersten T, et al. In vivo microbial stimulation induces rapid CD40 ligand-independent production of interleukin 12 by dendritic cells and their redistribution to T cell areas. J Exp Med 1997;186:1819. 75. Pamer E, Cresswell P. Mechanisms of MHC class I-restricted antigen processing. Annu Rev Immunol 1998;16:323. 76. Bevan MJ. Cross-priming for a secondary cytotoxic response to minor H antigens with H-2 congenic cells which do not cross-react in the cytotoxic assay. J Exp Med 1976;143:1283. 77. Heath WR, Carbone FR. Cross-presentation, dendritic cells, tolerance and immunity. Annu Rev Immunol 2001;19:47. 78. den Haan JM, Lehar SM, Bevan MJ. CD8(+) but not CD8(-) dendritic cells cross-prime cytotoxic T cells in vivo. J Exp Med 2000;192:1685. 79. Kovacsovics-Bankowski M, Rock KL. A phagosome-to-cytosol pathway for exogenous antigens presented on MHC class I molecules. Science 1995;267:243. 80. Brossart P, Bevan MJ. Presentation of exogenous protein antigens on major histocompatibility complex class I molecules by dendritic cells: pathway of presentation and regulation by cytokines. Blood 1997;90:1594. 81. Bennett SR, Carbone FR, Karamalis F, et al. Induction of a CD8+ cytotoxic T lymphocyte response by cross-priming requires cognate CD4+ T cell help. J Exp Med 1997;186:65. 82. Schoenberger SP, Toes RE, van der Voort EI, et al. T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature 1998;393:480. 83.

Lappin MB, Kimber I, Norval M. The role of dendritic cells in cutaneous immunity. Arch Dermatol

Res 1996;288:109. 84. Romani N, Ratzinger G, Pfaller K, et al. Migration of dendritic cells into lymphatics-the Langerhans cell example: routes, regulation, and relevance. Int Rev Cytol 2001;207:237. 85. Nalefski EA, Rao A. Nature of the ligand recognized by a hapten- and carrier-specific, MHCrestricted T cell receptor. J Immunol 1993;150:3806. 86. Honda S, Zhang W, Kalergis AM, et al. Hapten addition to an MHC class I-binding peptide causes substantial adjustments of the TCR structure of the responding CD8(+) T cells. J Immunol 2001;167:4276. 87. Franco A, Yokoyama T, Huynh D, et al. Fine specificity and MHC restriction of trinitrophenyl-specific CTL. J Immunol 1999;162:3388. 88. Horton H, Weston SD, Hewitt CR. Allergy to antibiotics: T-cell recognition of amoxicillin is HLA-DR restricted and does not require antigen processing. Allergy 1998;53:83. 89. Flint MS, Dearman RJ, Kimber I, et al. Production and in situ localization of cutaneous tumour necrosis factor alpha (TNF-alpha) and interleukin 6 (IL-6) following skin sensitization. Cytokine 1998;10:213. 90. Nuriya S, Yagita H, Okumura K, et al. The differential role of CD86 and CD80 co-stimulatory molecules in the induction and the effector phases of contact hypersensitivity. Int Immunol 1996;8:917. 91. Jones DA, Morris AG, Kimber I. Assessment of the functional activity of antigen-bearing dendritic cells isolated from the lymph nodes of contact-sensitized mice. Int Arch Allergy Appl Immunol 1989;90:230. 92. De Creus A, Van Beneden K, Taghon T, et al. Langerhans cells that have matured in vivo in the absence of T cells are fully capable of inducing a helper CD4 as well as a cytotoxic CD8 response. J Immunol 2000;165:645. 93. Lynch DH, Gurish MF, Daynes RA. Relationship between epidermal Langerhans cell density ATPase activity and the induction of contact hypersensitivity. J Immunol 1981;126:1892. 94. Neutra MR, Mantis NJ, Kraehenbuhl JP. Collaboration of epithelial cells with organized mucosal lymphoid tissues. Nat Immunol 2001;2:1004. 95. Iwasaki A, Kelsall BL. Localization of distinct Peyer’s patch dendritic cell subsets and their recruitment by chemokines macrophage inflammatory protein (MIP)-3alpha, MIP-3beta, and secondary lymphoid organ chemokine. J Exp Med 2000;191:1381. 96. Vermaelen KY, Carro-Muino I, Lambrecht BN, et al. Specific migratory dendritic cells rapidly transport antigen from the airways to the thoracic lymph nodes. J Exp Med 2001;193:51. 97. Manickasingham S, Reis e Sousa C. Microbial and T cell-derived stimuli regulate antigen presentation by dendritic cells in vivo. J Immunol 2000;165:5027. 98. Zhong G, Sousa CR, Germain RN. Antigen-unspecific B cells and lymphoid dendritic cells both show extensive surface expression of processed antigen-major histocompatibility complex class II complexes after soluble protein exposure in vivo or in vitro. J Exp Med 1997;186:673. 99. Gretz JE, Norbury CC, Anderson AO, et al. Lymph-borne chemokines and other low molecular weight molecules reach high endothelial venules via specialized conduits while a functional barrier limits

access to the lymphocyte microenvironments in lymph node cortex. J Exp Med 2000;192:1425. 100. Ebnet K, Kaldjian EP, Anderson AO, et al. Orchestrated information transfer underlying leukocyte endothelial interactions. Annu Rev Immunol 1996;14:155. 101. Steinman RM, Pack M, Inaba K. Dendritic cells in the T-cell areas of lymphoid organs. Immunol Rev 1997;156:25. 102. Pooley JL, Heath WR, Shortman K. Intravenous soluble antigen is presented to CD4 T cells by CD8(-) dendritic cells, but cross-presented to CD8 T cells by CD8(+) dendritic cells. J Immunol 2001;166:5327. 103. Reis e Sousa C, Germain RN. Analysis of adjuvant function by direct visualization of antigen presentation in vivo: endotoxin promotes accumulation of antigen-bearing dendritic cells in the T cell areas of lymphoid tissue. J Immunol 1999;162:6552. 104. Kane LP, Lin J, Weiss A. Signal transduction by the TCR for antigen. Curr Opin Immunol 2000;12:242. 105. Crabtree GR. Generic signals and specific outcomes: signaling through Ca2+, calcineurin, and NF-AT. Cell 1999;96:611. 106. Sette A, Buus S, Colon S, et al. I-Ad-binding peptides derived from unrelated protein antigens share a common structural motif. J Immunol 1988;141:45. 107. Khoruts A, Mondino A, Pape KA, et al. A natural immunological adjuvant enhances T cell clonal expansion through a CD28-dependent, interleukin (IL)-2-independent mechanism. J Exp Med 1998;187:225. 108. Veiga-Fernandes H, Walter U, Bourgeois C, et al. Response of naive and memory CD8+ T cells to antigen stimulation in vivo. Nat Immunol 2000;1:47. 109. Schorle H, Holtschke T, Hunig T, et al. Development and function of T cells in mice rendered interleukin-2 deficient by gene targeting. Nature 1991;352:621. 110. Kneitz B, Herrmann T, Yonehara S, et al. Normal clonal expansion but impaired Fas-mediated cell death and anergy induction in interleukin-2-deficient mice. Eur J Immunol 1995;25:2572. 111. Ku CC, Murakami M, Sakamoto A, et al. Control of homeostasis of CD8+ memory T cells by opposing cytokines. Science 2000;288:675. 112. Leung DT, Morefield S, Willerford DM. Regulation of lymphoid homeostasis by IL-2 receptor signals in vivo. J Immunol 2000;164:3527. 113. Lantz O, Grandjean I, Matzinger P, et al. Gamma chain required for naive CD4+ T cell survival but not for antigen proliferation. Nat Immunol 2000;1:54. 114. Gudmundsdottir H, Wells AD, Turka LA. Dynamics and requirements of T cell clonal expansion in vivo at the single-cell level: effector function is linked to proliferative capacity. J Immunol 1999;162:5212. 115.

Rocha B, von Boehmer H. Peripheral selection of the T cell repertoire. Science 1991;251:1225.

116. McHeyzer-Williams MG, Davis MM. Antigen-specific development of primary and memory T cells in vivo. Science 1995;268:106.

117. Gulbranson-Judge A, MacLennan I. Sequential antigen-specific growth of T cells in the T zones and follicles in response to pigeon cytochrome c. Eur J Immunol 1996;26:1830. 118. Luther SA, Gulbranson-Judge A, Acha-Orbea H, et al. Viral superantigen drives extrafollicular and follicular B cell differentiation leading to virus-specific antibody production. J Exp Med 1997;185:551. 119. Zimmerman C, Brduscha-Riem K, Blaser C, et al. Visualization, characterization, and turnover of CD8+ memory T cells in virus-infected hosts. J Exp Med 1996;183:1367. 120. Blattman JN, Antia R, Sourdive DJ, et al. Estimating the precursor frequency of naive antigen-specific CD8 T cells. J Exp Med 2002;195:657. 121. Homann D, Teyton L, Oldstone MB. Differential regulation of antiviral T-cell immunity results in stable CD8+ but declining CD4+ T-cell memory. Nat Med 2001;7:913. 122. Kearney ER, Walunas TL, Karr RW, et al. Antigen-dependent clonal expansion of a trace population of antigen-specific CD4 + T cells in vivo is dependent on CD28 costimulation and inhibited by CTLA-4. J Immunol 1995;155:1032. 123. Suresh M, Whitmire JK, Harrington LE, et al. Role of CD28-B7 interactions in generation and maintenance of CD8 T cell memory. J Immunol 2001;167:5565. 124. Grewal IS, Xu J, Flavell RA. Impairment of antigen-specific T-cell priming in mice lacking CD40 ligand. Nature 1995;378:617. 125. Garside P, Ingulli E, Merica RR, et al. Visualization of specific B and T lymphocyte interactions in the lymph node. Science 1998;281:96. 126. Buhlmann JE, Gonzalez M, Ginther B, et al. Sustained expansion of CD8+ T cells requires CD154 expression by Th cells in acute graft versus host disease. J Immunol 1999;162:4373. 127. Ranheim EA, Kipps TJ. Activated T cells induce expression of B7/BB1 on normal or leukemic B cells through a CD40-dependent signal. J Exp Med 1993;177:925. 128. Viola A, Schroeder S, Sakakibara Y, et al. T lymphocyte costimulation mediated by reorganization of membrane microdomains. Science 1999;283:680. 129. Bromley SK, Iaboni A, Davis SJ, et al. The immunological synapse and CD28-CD80 interactions. Nat Immunol 2001;2:1159. 130. Salomon B, Bluestone JA. Complexities of CD28/B7: CTLA-4 costimulatory pathways in autoimmunity and transplantation. Annu Rev Immunol 2001;19:225. 131. Watts TH, DeBenedette MA. T cell co-stimulatory molecules other than CD28. Curr Opin Immunol 1999;11:286. 132. Rogers PR, Song J, Gramaglia I, et al. Ox40 promotes bcl-xl and bcl-2 expression and is essential for long-term survival of CD4 T cells. Immunity 2001;15:445. 133. Schmidt CS, Mescher MF. Adjuvant effect of IL-12: conversion of peptide antigen administration from tolerizing to immunizing for CD8+ T cells in vivo. J Immunol 1999;163:2561. 134.

Pape KA, Khoruts A, Mondino A, et al. Inflammatory cytokines enhance the in vivo clonal

expansion and differentiation of antigen-activated CD4+ T cells. J Immunol 1997;159:591. 135. Sun S, Kishimoto H, Sprent J. DNA as an adjuvant: capacity of insect DNA and synthetic oligodeoxynucleotides to augment T cell responses to specific antigen. J Exp Med 1998;187:1145. 136. Medzhitov R, Janeway C Jr. Innate immune recognition: mechanisms and pathways. Immunol Rev 2000;173:89. 137. Schnare M, Barton GM, Holt AC, et al. Toll-like receptors control activation of adaptive immune responses. Nat Immunol 2001;2:947. 138. Jones BW, Means TK, Heldwein KA, et al. Different Toll-like receptor agonists induce distinct macrophage responses. J Leukoc Biol 2001;69:1036. 139. Kaisho T, Takeuchi O, Kawai T, et al. Endotoxin-induced maturation of MyD88-deficient dendritic cells. J Immunol 2001;166:5688. 140. Joseph SB, Miner KT, Croft M. Augmentation of naive, Th1 and Th2 effector CD4 responses by IL-6, IL-1 and TNF. Eur J Immunol 1998;28:277. 141. Lichtman AH, Chin J, Schmidt JA, et al. Role of interleukin 1 in the activation of T lymphocytes. Proc Natl Acad Sci U S A 1988;85:9699. 142.

Dutton RW, Bradley LM, Swain SL. T cell memory. Annu Rev Immunol 1998;16:201.

143. Campbell DJ, Kim CH, Butcher EC. Separable effector T cell populations specialized for B cell help or tissue inflammation. Nat Immunol 2001;2:876. 144. Borges E, Tietz W, Steegmaier M, et al. P-selectin glycoprotein ligand-1 (PSGL-1) on T helper 1 but not on T helper 2 cells binds to P-selectin and supports migration into inflamed skin. J Exp Med 1997;185:573. 145. Fong TA, Mosmann TR. The role of IFN-gamma in delayed-type hypersensitivity mediated by Th1 clones. J Immunol 1989;143:2887. 146. Xie H, Lim YC, Luscinskas FW, et al. Acquisition of selectin binding and peripheral homing properties by CD4(+) and CD8(+) T cells. J Exp Med 1999;189:1765. 147.

Murphy KM. T lymphocyte differentiation in the periphery. Curr Opin Immunol 1998;10:226.

148. Campbell DJ, Butcher EC. Rapid Acquisition of Tissue-specific Homing Phenotypes by CD4(+) T cells activated in cutaneous or mucosal lymphoid tissues. J Exp Med 2002;195:135. 149. Mohrs M, Shinkai K, Mohrs K, et al. Analysis of type 2 immunity in vivo with a bicistronic IL-4 reporter. Immunity 2001;15:303. 150. Fuller KA, Kanagawa O, Nahm MH. T cells within germinal centers are specific for the immunizing antigen. J Immunol 1993;151:4505. 151. Zheng B, Han S, Kelsoe G. T helper cells in murine germinal centers are antigen-specific emigrants that downregulate Thy-1. J Exp Med 1996;184:1083. 152. Ansel KM, McHeyzer-Williams LJ, Ngo VN, et al. In vivo-activated CD4 T cells upregulate CXC chemokine receptor 5 and reprogram their response to lymphoid chemokines. J Exp Med 1999;190:1123.

153. Cyster JG, Goodnow CC. Antigen-induced exclusion from follicles and anergy are separate and complementary processes that influence peripheral B cell fate. Immunity 1995;3:691. 154. Fulcher DA, Lyons AB, Korn SL, et al. The fate of self-reactive B cells depends primarily on the degree of antigen receptor engagement and availability of T cell help. J Exp Med 1996;183:2313. 155. Harrington LE, Galvan M, Baum LG, et al. Differentiating between memory and effector CD8 T cells by altered expression of cell surface O-glycans. J Exp Med 2000;191:1241. 156. Hamann D, Baars PA, Rep MH, et al. Phenotypic and functional separation of memory and effector human CD8+ T cells. J Exp Med 1997;186:1407. 157. Lauvau G, Vijh S, Kong P, et al. Priming of memory but not effector CD8 T cells by a killed bacterial vaccine. Science 2001;294:1735. 158. Matloubian M, Concepcion RJ, Ahmed R. CD4+ T cells are required to sustain CD8+ cytotoxic T-cell responses during chronic viral infection. J Virol 1994;68:8056. 159. Finch RJ, Fields PE, Greenberg PD. A transcriptional block in the IL-2 promoter at the -150 AP-1 site in effector CD8+ T cells. J Immunol 2001;166:6530. 160. Tham EL, Mescher MF. Signaling alterations in activation-induced nonresponsive CD8 T cells. J Immunol 2001;167:2040. 161. Flynn KJ, Belz GT, Altman JD, et al. Virus-specific CD8+ T cells in primary and secondary influenza pneumonia. Immunity 1998;8:683. 162. Hendriks J, Gravestein LA, Tesselaar K, et al. CD27 is required for generation and long-term maintenance of T cell immunity. Nat Immunol 2000;1:433. 163. Masopust D, Vezys V, Marzo AL, et al. Preferential localization of effector memory cells in nonlymphoid tissue. Science 2001;291:2413. 164. Weninger W, Crowley MA, Manjunath N, et al. Migratory properties of naive, effector, and memory CD8(+) T cells. J Exp Med 2001;194:953. 165. Lefrancois L, Parker CM, Olson S, et al. The role of beta7 integrins in CD8 T cell trafficking during an antiviral immune response. J Exp Med 1999;189:1631. 166. Liblau RS, Tish R, Shokat K, et al. Intravenous injection of soluble antigen induces thymic and peripheral T cell apoptosis. Proc Natl Acad Sci USA 1996;93:3031. 167. Petschner F, Zimmerman C, Strasser A, et al. Constitutive expression of Bcl-xL or Bcl-2 prevents peptide antigen-induced T cell deletion but does not influence T cell homeostasis after a viral infection. Eur J Immunol 1998;28:560. 168. Lenardo M, Chan KM, Hornung F, et al. Mature T lymphocyte apoptosis—immune regulation in a dynamic and unpredictable antigenic environment. Annu Rev Immunol 1999;17:221. 169. Refaeli Y, Van Parijs L, Abbas AK. Genetic models of abnormal apoptosis in lymphocytes. Immunol Rev 1999;169:273. 170.

Hiromatsu K, Aoki Y, Makino M, et al. Increased Fas antigen expression in murine

retrovirus-induced immunodeficiency syndrome, MAIDS. Eur J Immunol 1994;24:2446. 171. Refaeli Y, Van Parijs L, London CA, et al. Biochemical mechanisms of IL-2-regulated Fas-mediated T cell apoptosis. Immunity 1998;8:615. 172. Hildeman DA, Mitchell T, Teague TK, et al. Reactive oxygen species regulate activation-induced T cell apoptosis. Immunity 1999;10:735. 173. Vella AT, McCormack JE, Linsley PS, et al. Lipopolysaccharide interferes with the induction of peripheral T cell death. Immunity 1995;2:261. 174. Vella AT, Mitchell T, Groth B, et al. CD28 engagement and proinflammatory cytokines contribute to T cell expansion and long-term survival in vivo. J Immunol 1997;158:4714. 175. Mitchell TC, Hildeman D, Kedl RM, et al. Immunological adjuvants promote activated T cell survival via induction of Bcl-3. Nat Immunol 2001;2:397. 176. Ahmed R, Gray D. Immunological memory and protective immunity: understanding their relation. Science 1996;272:54. 177. Opferman JT, Ober BT, Ashton-Rickardt PG. Linear differentiation of cytotoxic effectors into memory T lymphocytes. Science 1999;283:1745. 178. Jacob J, Baltimore D. Modelling T-cell memory by genetic marking of memory T cells in vivo. Nature 1999;399:593. 179. Jamieson BD, Ahmed R. T cell memory. Long-term persistence of virus-specific cytotoxic T cells. J Exp Med 1989;169:1993. 180. Murali-Krishna K, Lau LL, Sambhara S, et al. Persistence of memory CD8 T cells in MHC class I-deficient mice. Science 1999;286:1377. 181. Zhang X, Sun S, Hwang I, et al. Potent and selective stimulation of memory-phenotype CD8+ T cells in vivo by IL-15. Immunity 1998;8:591. 182. Kennedy MK, Glaccum M, Brown SN, et al. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice. J Exp Med 2000;191:771. 183. Lodolce JP, Burkett PR, Boone DL, et al. T cell-independent interleukin 15Ralpha signals are required for bystander proliferation. J Exp Med 2001;194:1187. 184. Sallusto F, Lenig D, Forster R, et al. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions. Nature 1999;401:708. 185. Sallusto F, Mackay CR, Lanzavecchia A. The role of chemokine receptors in primary, effector, and memory immune responses. Annu Rev Immunol 2000;18:593. 186. Bell EB, Sparshott SM. Interconversion of CD45R subsets of CD4 T cells in vivo. Nature 1990;348:163. 187. Tough DF, Sprent J. Turnover of naive- and memory-phenotype T cells. J Exp Med 1994;179:1127. 188. Merica R, Khoruts A, Pape KA, et al. Antigen-experienced CD4 T cells display a reduced capacity for clonal expansion in vivo that is imposed by factors present in the immune host. J Immunol

2000;164:4551. 189. Gray D, Matzinger P. T cell memory is short-lived in the absence of antigen. J Exp Med 1991;174:969. 190. Swain S. Generation and in vivo persistence of polarized Th1 and Th2 memory cells. Immunity 1994;1:543. 191. Swain SL, Hu H, Huston G. Class II-independent generation of CD4 memory T cells from effectors. Science 1999;286:1381. 192. Garcia S, DiSanto J, Stockinger B. Following the development of a CD4 T cell response in vivo: from activation to memory formation. Immunity 1999;11:163. 193. Mackay C, Marston W, Dudler L. Naive and memory T cells show distinct pathways of lymphocyte recirculation. J Exp Med 1990;171:801.

Chapter 11 T-Lymphocyte Activation Fundamental Immunology

Chapter 11 Arthur Weiss

T-Lymphocyte Activation

INTRODUCTION EXPERIMENTAL MODELS USED TO STUDY T-CELL ACTIVATION Responding T Cells and Antigen-Presenting Cells Stimuli: Complex Antigens and Peptides Stimuli: Superantigens Stimuli: Lectins Stimuli: Monoclonal Antibodies Pharmacologic Agents REQUIREMENTS FOR THE INITIATION OF T-CELL ACTIVATION Primary Signal for T-Cell Activation: Requirement or Dependency for TCR Involvement CD4 and CD8 Co-receptors Contribute to Primary Activation Signal Accessory Molecules Increase Avidity of T-Cell–APC Interaction Co-stimulatory Signal Is Required for T-Cell Activation SIGNAL TRANSDUCTION BY T-CELL ANTIGEN RECEPTOR Complex Structure and Signal Transduction Function of TCR TCR ITAMs and Cytoplasmic Protein-Tyrosine Kinases Src PTKs Involved in TCR Signal Transduction Function of SH2 Domains in Signal Transduction Pathways Protein-Tyrosine Phosphatase (PTPase) CD45 Plays Critical Role in TCR Signal Transduction Consequences of TCR-Mediated PTK Activation CONSEQUENCES OF EARLY SIGNAL TRANSDUCTION EVENTS Early Biochemical Events Cellular Responses Gene Activation Events TERMINATING T-CELL RESPONSES T-CELL INACTIVATION CONCLUSION REFERENCES

INTRODUCTION The immune system has evolved to provide a flexible and dynamic mechanism to respond specifically to a wide variety of antigens. In order for a response to occur following antigen challenge, antigen must not only be recognized by antigen-specific lymphocytes but such recognition must be translated into signal transduction events that are responsible for the initiation of cellular responses. T-lymphocytes, together with B-lymphocytes, represent the two antigen-specific components of the cellular immune system. The activation of resting T cells is critical to most immune responses and allows these cells to exert their regulatory or effector capabilities. During activation these

relatively quiescent T cells in the G 0 stage of the cell cycle undergo complex changes resulting in cell differentiation and proliferation. Since each T cell expresses T-cell antigen receptors (TCR) of a single antigen specificity, only a small subset of T cells is activated by any particular antigen (clonal selection). This results in the clonal expansion of antigen-reactive T cells that acquire differentiated functional capacities. However, the activation of T-lymphocytes is actually a consequence of multiple ligand–receptor interactions that occur at the interface of the T cell and an antigen-presenting cell (APC). In sum, these interactions initiate intracellular biochemical events within the T cells that culminate in cellular responses. It is clear that a large number of different cell-surface molecules on the T-lymphocyte and the APC, only some of which are depicted in Fig. 1, may participate in the complex cell–cell interaction that occurs during antigen presentation. In view of the specificity of T-cell responses, antigen-induced T-lymphocyte activation must be directed by T-cell antigen receptors (TCR). The ligand for the TCR is a short peptide antigen fragment, derived by proteolysis from a larger molecule, which is bound to a syngeneic major histocompatibility complex (MHC) molecule (see Chapter 19 and Chapter 20). The antigen receptor is a multichain structure derived from at least six genes ( Fig. 2). On most T cells, it contains at least one disulfide-linked a/ß heterodimer responsible for antigen recognition (see Chapter 8). A small subset of T cells that recognize antigen with a ?/d heterodimer, may preferentially play a role in immune responses in epithelial tissues. The a/ß or ?/d heterodimer is noncovalently associated with invariant chains derived from the ? and CD3?, d, and e genes that are responsible for coupling the receptor to intracellular signal-transduction components ( 1 , 2 ) (see below).

FIG. 1. Schematic representation of some of the ligand–receptor interactions that occur during the interaction of a T cell with an antigen-presenting cell (APC) or target cell.

FIG. 2. The T-cell antigen receptor. Illustrated schematically is the antigen-binding subunit comprised of an aß heterodimer, and the associated invariant CD3 and ? chains. Acidic (-) and basic (+) residues located within the plasma membrane are indicated. The open rectangular boxes indicate motifs (see Fig. 3) within the cytoplasmic domains that interact with cytoplasmic protein-tyrosine kinases.

FIG. 3. Interactions between a T cell and antigen-presenting cell (APC) that lead to IL-2 production. The sequential interactions of the TCR with peptide antigen/MHC complex lead to the induced expression of CD40L on the T cell that interacts with CD40 on the APC. This induces the expression of B7 molecules on the APC that can then stimulate CD28, the co-stimulatory receptor on the T cell. The two signals induced by the TCR and CD28 lead to IL-2 gene expression.

Antigen-induced stimulation of the TCR delivers the primary signal in initiating activation. For naïve resting T cells, stimulation of the TCR alone is insufficient to induce proliferative responses by purified resting G 0 T cells, but may be sufficient to induce activation of more differentiated T-cell populations or to induce a state of unresponsiveness, termed “anergy” (reviewed in Schwartz [ 3 ]) (see also Chapter 29). Other cell-surface molecules expressed on T cells, by binding to their respective

ligands, play a role in antigen-specific activation by functioning as accessory molecules in the initial antigen-specific events occurring between an APC and T cell. These accessory molecules may contribute to the initiation of cellular activation by (a) functioning as adhesion molecules, strengthening the interaction between the T cell and APC (e.g., LFA-1 and CD2); (b) modifying the transmembrane signal initiated via the antigen receptor (e.g., CD4 and CD8); and/or (c) initiating their own transmembrane signaling events, distinct from those of the TCR, which are necessary for cellular responses (e.g., CD28 and the interleukin-1 receptor). These latter signal transduction events are responsible for the requisite second signal, or co-stimulatory signal, required to activate resting T cells. A more detailed discussion of the structure and function of some of these accessory molecules is presented in Chapter 13. The interaction of the TCR with its ligand, or the co-stimulatory receptor with its ligand, initiates cellular activation by inducing a series of rapid biochemical changes that have a number of important consequences. For example, activation of protein kinases results in phosphorylation of specific substrates. These phosphorylations can serve as binding sites for additional signaling molecules. This process of protein–protein interaction results in recruitment of a number of molecules, the creation of signaling complexes, and activation of signaling molecules. Among the latter are enzymes that induce formation of intracellular biochemical mediators called “second messengers.” These second messengers as well as multiple activated enzymes can function to initiate or influence cellular response pathways. In resting T cells, such signals can alter a multitude of intracellular events. In differentiated effector T cells, such signals can initiate the activation of the cytolytic mechanism, a stimulus coupled-secretory response in which exocytosis of previously synthesized and packaged proteins involved in the cytolytic apparatus occurs. During the process of T-cell activation, there are early responses occurring within minutes or hours after the initiation of signal transduction, while others may only occur days after the stimulating event. The early cellular responses may directly or indirectly be the result of TCR- or other receptor-mediated signal transduction. During the early phase of T-cell activation, T cells undergo enormous changes characterized by protein phosphorylation, membrane lipid and cytoskeletal changes, membrane reorganization, ion fluxes, cyclic nucleotide alterations, increased or decreased RNA synthesis of constitutive and newly activated gene products, and cell volume increases (blast transformation). The later cellular responses, such as proliferation, generally result from a complex cascade of gene activation events and the coordinated sequential influence of the products of these genes. For instance, stimulation of the TCR can drive a resting G 0 T cell into G 1 , where it expresses lymphokine receptors, but further progres